The Use of Ampicillin-Medicated Diet to Treat Segmented Filamentous Bacteria in a Mouse (Mus musculus) Colony
Segmented filamentous bacteria (SFB), or Candidatus savagella, are gram-positive, spore-forming, filamentous commensal bacteria that colonize the ilea of mice and rats. SFB can impact certain research studies due to their effect on the innate and adaptive immune system. Therefore, there is a need to eliminate SFB from some rodent colonies. Antibiotics are an effective treatment method that have been shown to be successful at eradicating this bacterium. Commonly, antibiotics are supplied to rodents via water. However, antibiotics in water have several limitations, depending on the type of water and the length of time in water, and administration is labor intensive. Therefore, this study sought to create a rodent diet containing the antibiotic ampicillin. The feed was constructed to counter an expected ampicillin concentration decrease from the pelleting process and irradiation. Mice were placed on the ampicillin feed for 4 weeks, and SFB was measured via PCR fecal analysis. After one week of feeding the ampicillin diet, all mice in the treatment group were negative for SFB. These mice remained negative for 2 additional weeks after treatment cessation. This study shows ampicillin diet is an effective method to eliminate SFB from mouse colonies.
Introduction
Segmented filamentous bacteria (SFB), or Candidatus savagella, are gram-positive, spore-forming, filamentous commensal bacteria that colonize the ilea of mice and rats right before weaning.1 These bacteria are exclusively present in high numbers in the ilea shortly after weaning, as there are only low levels present in adult mice.2 SFB have attracted the attention of researchers in recent years due to their effect on the innate and adaptive immune system, because they are involved in the maturation and function of the host gut immune barrier.1,3 SFB are known to induce the differentiation of T-helper 17 cells in the small intestine.4,5 Specifically, in non-obese diabetic (NOD) mice, a model for type 1 diabetes, SFB have been shown to cause a T-helper 17 cell response, which protects the mice from developing type 1 diabetes and subsequently decreases the incidence of disease.5,6 Therefore, at multiple animal care programs, there is an interest in eliminating this bacterium from NOD mouse colonies due to its significant effect on research data. In addition, the need to maintain SFB-negative mouse colonies extends to other models, as SFB has other effects on the immune system, including limiting the colonization and proliferation of harmful enteric pathogens.1 Finally, SFB exacerbates the promotion of the experimental autoimmune encephalomyelitis model.7,8 Currently, no rodent vendors offer barrier mice that are confirmed to exclude SFB; therefore, to produce SFB-negative mouse colonies, treatment in-house is required.9 Once treated, institutions maintain SFB colonies by either using entry order restrictions or isolator housing.10
Treatment of bacterial infection usually involves the use of antibiotics. One case study used ampicillin provided in the water to successfully eliminate SFB from an animal room housing multiple NOD strains.6 Treatment success was verified by negative fecal SFB PCR analysis; however, the concentration of the antibiotic was not measured in the water nor in the plasma of the mice.6 Previous publications have shown that there are multiple limitations when providing antibiotics to mice in the water. These include instability of certain medications in water, the requirement for an adequate amount of water to be consumed to be efficacious, especially when the treatment can affect the palatability, and that adequate concentrations of the antibiotic must be maintained in the plasma to achieve treatment efficacy.11 In one study,11 it was found that the plasma concentrations of mice treated with water containing either enrofloxacin, doxycycline, amoxicillin, or trimethoprim sulfamethoxazole were all below the plasma concentrations necessary for treatment efficacy against multiple common bacteria. In addition, the type of water used when providing antibiotic treatment, be it reverse osmosis, acidified, or tap, can affect the antibiotic concentration.12 This was highlighted in a study that assessed amoxicillin-clavulanic acid and trimethoprim-sulfamethoxazole concentrations compounded in these 3 water types.12
Due to these limitations and the fact that the use of medicated water is labor intensive, as it needs to be replaced on a frequent basis, we decided to assess a treatment using a medicated diet for an SFB-positive colony composed predominantly of NOD mice. Specifically, we used ampicillin, as this antibiotic was previously shown to be effective against SFB.6 In addition, relative PCR copy numbers for each positive finding were evaluated to estimate organism load and compared among groups to assess treatment efficacy. Finally, we performed HPLC to assess the concentration of ampicillin in the irradiated diet. We hypothesized that diet medicated with ampicillin would be effective at eliminating SFB in the mouse colony of concern.
Ethical review.
All animal care and use was conducted in accordance with federal polices and guidelines and was approved by the University of Chicago’s IACUC. The University of Chicago has a PHS assurance with OLAW and is accredited by AAALAC International.
Materials and Methods
Animals and husbandry.
All animals in this study were housed within an Animal Resources Center facility at the University of Chicago on three IVC racks within one animal housing room (RRID: SCR_021806). Cages were from one investigator’s colony and contained adult mice of varying ages, sexes, and strains; although a large majority were from a NOD background, one cage in the treatment group was of C57BL/6 background. Mice were group-housed in solid-bottom polysulfone IVC (19.69 × 30.48 × 16.51 cm; Jag 75 Micro-Barrier IVC; Allentown, Allentown, NJ) with shredded pine shavings (NEPCO no. 326.2; NEPCO, Warrensburg, NY) for bedding and cotton nestlet enrichment (NES3600; Ancare, Bellmore, NY) in breeder cages. They were provided with chlorinated reverse osmosis water through an automatic watering system (Avidity Science, Waterford, WI) ad libitum. The control animals were fed irradiated rodent diet (Teklad 2918; Inotiv. West Lafayette, IN). All cages, bedding, and enrichment were autoclaved before use. Cages were changed every 7 days per investigator request within a Class II Type A2 biosafety cabinet (NuAire, Plymouth, MN).
Animal rooms were maintained on a 12:12-hour light:dark cycle with humidity ranging from 30% to 70% and temperatures ranging from 68 to 76 °F (20 to 24 °C) in compliance with the Guide for the Care and Use of Laboratory Animals.13 Mice were checked daily by the animal care staff to assess their health status and the availability of appropriate food, water, and cage conditions.
Routine colony health monitoring was performed quarterly by exhaust dust testing via PCR as previously described.14 Excluded agents were Sendai virus, pneumonia virus of mice, mouse hepatitis virus, mouse parvoviruses, reovirus, mouse rotavirus, mouse encephalomyelitis virus, ectromelia virus, lymphocytic choriomeningitis virus, murine adenovirus, murine cytomegalovirus, K virus, polyoma virus, mouse thymic virus, hantavirus, lactate dehydrogenase-elevating virus, Filobacterium rodentium, Mycoplasma pulmonis, Salmonella spp., Citrobacter rodentium, Clostridium piliforme, Streptobacillus moniliformis, Corynebacterium kutscheri, and endo- and ectoparasites such as Giardia muris, Myobia musculi, Myocoptes musculinus, Radfordia affinis, Syphacia spp., and Aspicularis tetraptera. Mouse norovirus, Rodentibacter spp., and Helicobacter spp. were endemic in the vivaria.
Medicated diet.
Teklad global 18% protein diet (Teklad 2018; Inotiv, West Lafayette, IN) was used as the control diet and to produce the ampicillin diet. The irradiated version, Teklad 2918 (Inotive, West Lafayette, IN), was used for the control. The non-irradiated Teklad 2018 diet was used as a base diet to create a pelleted diet containing 1.0 mg ampicillin (ampicillin sodium; A-301-25; Gold Biotechnology, Olivette, MO) per 1.0 g of diet created with assistance of the vendor nutritionists. The pellets were then irradiated using gamma irradiation at a dose of 20-50 kGy via cobalt 60 (TD.220046 - 0.1% ampicillin sodium diet [2018]), similar to Teklad 2918. This concentration was chosen based on efficacious doses for mice that range from 20 to 200 mg/kg6,15–17 with 100 mg/kg per day more frequently reported, and with an expected approximate 40% drug loss post-irradiation and pelleting. The anticipated 40% loss was extracted from a previous study that reported 60% of amoxicillin remained post-pelleting and irradiation of rodent chow.12 After a 40% reduction in the ampicillin concentration, the expected dose for a 30.0 g mouse eating 5 g of feed a day would be 100 mg/kg per day. The ampicillin diet was provided to each experimental cage for 4 weeks. The recommended expiration date of the ampicillin feed is 6 months post-mill date. For this experiment, the mice were fed the diet between the fourth and fifth months after the post-manufacture date. However, an extension was cofirmed as appropriate by the manufacturer for a 9-month expiration post-mill date.
Ampicillin extraction.
A pelleted diet was submitted to 2 different laboratories with HPLC capability and knowledge (University of Illinois and MD Anderson Cancer Center) to quantify the concentration of ampicillin in the diet after being exposed to pelleting and irradiation. The first lab analyzed the diet at approximately 2 months post-mill date, and the second lab at about 8 months post-mill date. Both laboratories used similar procedures to extract the ampicillin concentration. In summary, the rodent diet was ground, and 1.0 g was put into a 5.0-mL tube with 4.0 mL of 0.1 M sodium phosphate buffer (pH 6.5). The solution was homogenized and shaken at 4 °C overnight. The mixture was spun down, and the supernatant was filtered. Filtrate was collected and spun down again. Ampicillin in the filtrate was measured by HPLC using spiked samples as standards.
Experimental design.
There were 115 cages initially screened for SFB by fecal PCR. Positive cages that were approved by the investigator's lab were used in this study (23 cages). The authors of this study had no input regarding which cages were screened or selected for use in this study. Mice were never rearranged between cages; all littermates remained the same throughout the study.
Fecal samples for PCR testing came from a sample group. Sample groups contained either 1 or 2 cages. If a cage had 2 or fewer mice, it was combined with another cage to create one sample group. In some cases, a cage with 3 mice was combined with another cage to create one sample group. In other words, fecal pellets from 1 cage, or 2 combined cages, were submitted for PCR testing. A total of 17 samples (total n) were submitted for PCR testing.
The cages were randomly assigned to the control or ampicillin treatment group in a stratified manner by the authors before SFB copy numbers were known. They were placed in a treatment group in a fashion that created an equal number of sample groups per sex, as well as an equal number of sample groups that contained 2 cages. Each treatment group had 3 male sample groups and 2 female sample groups. Each treatment group also had 3 sample groups that were of 2 cages combined, and the remainder of the sample groups for each treatment group were from one cage. Breeder cages were also assigned to each treatment group with more breeders in the ampicillin group as this group was more important in finding if treatment worked, as well as if treated parents resulted in negative offspring. Each breeder cage was considered its own sample group. Treatment groups resulted in the following: control diet with 6 sample groups (n = 6, 1 breeder cage, 5 non-breeder cages) or ampicillin diet with 11 sample groups (n = 11, 6 breeder cages, 5 non-breeder cages).
Two breeder sample groups were removed from the study before the third week of testing due to euthanasia by the lab, one from the control group and one from the treatment group. At the sixth week of testing, another breeder group from the treatment group was removed for the same reason. During the course of the 4-week treatment, mice of weanling age (28 days old) were separated by sex, and a maximum of 5 mice were placed into new cages. Littermates were then assigned as a new sample group combining both the male and female cages. Two groups of weanlings were created from the treatment sample group, and 2 groups of weanlings from the control sample group were created during the treatment period. After the treatment had ended, but before the study was complete, 2 more groups of weanlings were also created, and both were from parents of the ampicillin diet treatment group. All new cages that were created during and after treatment have results reported separately from the other study groups. The authors collecting the fecal samples were not blind to the groups. The commercial diagnostic laboratory that received the fecal samples for PCR analysis was blind to the treatment and control groups.
All IVC racks within the housing room were replaced with sanitized IVC racks after cages had been on the ampicillin diet for 2 weeks and before the third week to eliminate any remaining bacteria or spores within the racks. Racks were sanitized in a rack washer (Better Built R-600; Delta, British Columbia, Canada). The rack washing procedure involved spraying plenums with a water hose and GP-100 (Sanitation Strategies, Holt, MI) and allowing 3-5 minutes of contact time before again spraying with water. The GP-100 was sprayed on plenums again and allowed another 3-5 minutes of contact time before using a brush to scrub inside plenums. The racks were then sent through the rack washer on a 30-minute cycle. To ensure that an appropriate sanitation temperature (180 °F [82.2 °C]) was achieved, a temperature-recording strip (TempTape 180; Pharmacal Research Laboratories, Waterbury, CT) was placed in the cage washer at the start of each day. Cages were changed in the room starting with the known SFB negative racks and ending with the known positive racks. The ampicillin diet ceased after 4 weeks, and mice were then put back on their standard diet (Teklad 2918; Inotiv, West Lafayette, IN).
SFB fecal PCR testing.
Fecal samples were collected from each cage, and sample groups with 2 cages were pooled. This was done weekly for the first 4 weeks. All fecal samples were fresh and collected directly from one mouse in each cage in microcentrifuge tubes during regular work hours when the lights were on. After treatment, fecal samples were collected again from the same remaining cages and pooled per sample group at 6 weeks post-treatment. The person collecting the samples thoroughly sanitized gloves with chlorine dioxide (Labsan C-Dox; no. 64983345; Sanitation Strategies, Holt, MI) between cages and changed gloves between each sample group.
All fecal samples were submitted for SFB screening at the Charles River Research Animal Diagnostic Services (Frederick, MD). Copy numbers were determined using real-time TaqMan PCR, which evaluates target template amplification throughout the thermocycling process. During cycling, the number that was reported by the PCR software for each sample is the point at which the amplification for a test sample level crossed the cycle threshold. The cycle threshold score was then rounded to the next whole number and compared with the cycle threshold score obtained from a plasmid-based positive control for that same run. The positive control was prepared to represent 100 copies. The estimated samples are provided with a relative quantity rounded to the next whole number. Relative PCR copy numbers for each positive finding were evaluated to estimate organism load.
Statistical analysis.
Power analysis for group size was calculated in R 4.4.0 using data from previous literature of a related study.3 Power was calculated using a 2-sample proportion test and assuming 80% power and 6 control cages; at least 8 ampicillin cages are needed. Statistical analysis was performed using Stata 17 by the University’s Biostatistics Laboratory and Research Computing Group (StataCorp LLC, College Station, TX). Statistical significance was defined as P < 0.05. Data were summarized using median values for continuous measures and first measurements to summary measurements using the Wilcoxon rank-sum test. Interquartile range was used for data dispersion.
Results
Ampicillin extraction.
The initial concentration of ampicillin in the diet was formulated to be 1.000 mg (1.0 mg/g) of ampicillin per 1.0 g of diet per the manufacturer (Figure 1). After pelleting, irradiation, and approximately 2 months of time, the concentration of the ampicillin in the diet was 0.396 mg per 1.0 g of diet (Figure 1). After 8 months, 2 months past the recommended expiration, the concentration of ampicillin was 0.311 mg per 1.0 g of diet (Figure 1).
Citation: Journal of the American Association for Laboratory Animal Science 2025; 10.30802/AALAS-JAALAS-25-111
SFB PCR fecal testing and copy numbers.
Before ampicillin diet treatment began, there was no significant difference in the median SFB copy numbers between the ampicillin-treated and control groups (P = 0.152; Figure 2). After one week of consuming ampicillin-medicated diet, PCR analysis of the fecal pellets from the treated group was negative at zero copy numbers of SFB. The treatment group continued to be negative for SFB each week throughout the 4 weeks of treatment, and remained negative for 2 weeks after treatment ceased (Figure 2). The control group was consistently positive for SFB at each weekly PCR analysis with the median copy number fluctuating each week (Figure 2). Some sample groups within the control group had weeks with zero copies of SFB, but none of those sample groups continued to be negative through the remainder of the 6 weeks (data not shown). There was a significant decrease (P ≤ 0.05) in the ampicillin-treated group median SFB copy numbers compared with the median control group copy numbers when comparing the SFB results from week zero to week 6 (Figure 2).
Citation: Journal of the American Association for Laboratory Animal Science 2025; 10.30802/AALAS-JAALAS-25-111
Weanling cages originating from the ampicillin treatment sample groups, both during the treatment period (2 cages) and after (2 cages), were all negative for SFB at each week of fecal PCR testing. The weanling cages (2) from the control groups were positive for SFB at each remaining week (median SFB copy numbers weeks 3, 4, 6; 4085.5, 2033.5, 2441). Two months after the end of this study, 2 weanling cages from ampicillin-treated parents were tested again and were still negative for SFB (data not shown).
Discussion
To our knowledge, this study is the first hypothesis-based project assessing the use of an ampicillin diet to treat SFB in mice. We found that an ampicillin diet (1 mg/g) provided over a 4-week treatment was efficacious at eliminating SFB and maintaining the animals free of SFB post-treatment. We were able to track treatment efficacy by using copy number data. After just one week of ampicillin treatment, the copy numbers for the ampicillin-treated mice decreased to zero and were negative for SFB. This treatment efficacy is consistent with a previous case report that used ampicillin medicated water to treat SFB.6 However, we measured the actual concentration of the ampicillin-medicated diet by HPLC, which was not performed in the previous case report.6 We found that the ampicillin concentration in the diet was lower than expected; however, the treatment was still efficacious.
The ampicillin concentration in the diet post-pelleting and post-irradiation, and after 2 months, was reduced by approximately 60% from the initial concentration of 1.0 mg/g. It is known that pelleting and/or irradiation can decrease antibiotic concentrations in medicated diet, as a 40% reduction was also seen with amoxicillin and about 60% for trimethoprim and sulfamethoxazole concentrations.12 The concentration of the ampicillin further decreased over the following 6 months by 21.5% as compared with 2 months, for a total reduction from the original concentration pre-pelleting of 69%. For a 30.0-g mouse eating 5.0 g of food per day, the ampicillin dose after about 2 months was 66.0 mg/kg. After 8 months, the dose would have been 51.8 mg/kg. However, the diet was fed to the treated mice between the fourth and fifth months after the post-manufacturer date. Reported efficacious doses for ampicillin in mice range from 20 to 200 mg/kg.6,15–17
We chose to combine cages for non-breeder cages to form sample groups instead of testing individual cages of less than 3 mice because of the known variability of SFB intestinal colonization and shedding.2 If we had used all individual cages, we would have expected a higher variability in SFB copy numbers, and the purpose of this study was to assess the effectiveness in eliminating SFB, not the variability between cages.
We found that SFB shedding fluctuates, as shown by the variation in copy number. All cages enrolled in the study initially tested positive for SFB. In the control group without any treatment, copy numbers increased and decreased at different time points. At some time points, the copy numbers were even zero in the control group. This could indicate either that not all the mice are infected with SFB even when cohoused or that we inadvertently collected samples from the negative mice during the weeks with results showing zero copy numbers. However, any control cages that tested negative had subsequent positive tests in the following weeks and therefore did not remain negative for long. This has been corroborated by a researcher at our institution who found that after experimentally infecting mice with SFB by gavage and then cohousing with other mice, mice in each cage show varying SFB levels, and some mice do not have detectable SFB by PCR analysis (D. Esterhazy, oral communication, 2024). Alternatively, within each shedding, SFB may rise and fall for unknown reasons, which impacts the copy number results. It is expected that these results would translate to other species with SFB, as ampicillin is tolerated across species and SFB colonizes a wide range of mammals.
Limitations to the level of efficacy of this treatment option for SFB can include the reliance on mice to consume the appropriate or expected amount of feed needed to produce therapeutic levels of ampicillin. To compare the level of efficacy, one could do a serum and fecal pharmacokinetic study. It is important to include fecal concentrations since ampicillin is not fully absorbed and can remain in the gastrointestinal tract, where it can act directly on bacteria.18,19 Other recommendations for future studies would be to preferably use the same laboratory to conduct high-performance liquid chromatography at different time points to eliminate any laboratory variabilities in the ampicillin concentration analysis. It would be interesting to conduct a study that focuses more on the copy numbers of SFB and comparing different types of groups such as sex, breeding, or age differences as well as testing mice further out after the end of treatment. The control breeder group for this study had the highest and greatest difference in copy numbers compared with any other sample group, and not the weanling mice. Furthermore, one might better understand the spread of SFB within one cage by testing each mouse in the cage, instead of one sample per cage. Finally, understanding the presence of SFB spores within laboratory equipment such as the IVC racks and biosafety cabinets would also be useful. A previous study determined that the use of a tunnel washer that used water temperature of at least 180 °F to sanitize caging prevents the transmission of SFB to naïve immunocompromised mice.20
Conclusions.
In conclusion, ampicillin-medicated feed was successful at eliminating SFB in mice. SFB has been shown to affect the immune system of mice and may alter research data relating to the immune system and type 1 diabetes. As expected, the pelleting and irradiation process significantly decreased the ampicillin concentration in rodent feed, and with time it slowly decreased further. Accounting for ampicillin loss, the chosen ampicillin medicated diet of 1 mg/g over a 4-week period was efficacious at eliminating SFB and maintained the colony of NOD mice free of SFB post-treatment.
Ampicillin Concentration in Diet over Time. Month 0 represents the concentration before pelleting and irradiation of diet. The ampicillin diet has a recommended 6-mo expiration postmanufacture date.
Contributor Notes
Current affiliation: University of North Carolina at Charlotte, Charlotte, NC.