Antibiotic Administration in the Drinking Water of Mice

Abstract

Although antibiotics frequently are added to the drinking water of mice, this practice has not been tested to confirm that antibiotics reach therapeutic concentrations in the plasma of treated mice. In the current investigation, we 1) tested the stability of enrofloxacin and doxycycline in the drinking water of adult, female C57BL/6 mice; 2) measured the mice's consumption of water treated with enrofloxacin, doxycycline, amoxicillin, or trimethoprim–sulfamethoxazole; and 3) used HPLC to measure plasma antibiotic concentrations in mice that had ingested treated water for 1 wk. Plasma concentrations of antibiotic were measured 1 h after the start of both the light and dark cycle. The main findings of the study were that both enrofloxacin and nonpharmaceutical, chemical-grade doxycycline remained relatively stable in water for 1 wk. In addition, mice consumed similar volumes of antibiotic-treated and untreated water. The highest plasma antibiotic concentrations measured were: enrofloxacin, 140.1 ± 10.4 ng/mL; doxycycline, 56.6 ± 12.5 ng/mL; amoxicillin, 299.2 ± 64.1 ng/mL; and trimethoprim–sulfamethoxazole, 5.9 ± 1.2 ng/mL. Despite the stability of the antibiotics in the water and predictable water consumption by mice, the plasma antibiotic concentrations were well below the concentrations required for efficacy against bacterial pathogens, except for those pathogens that are exquisitely sensitive to the antibiotic. The findings of this investigation prompt questions regarding the rationale of the contemporary practice of adding antibiotics to the drinking water of mice for systemic antibacterial treatments.

The use of antibiotics to treat bacterial infections is a standard of care in veterinary medicine. In many species, the administration of antibiotics is a routine procedure with proven efficacy. Unfortunately, this is not the case for laboratory mice used in biomedical research, where the delivery of antibiotics may be associated with stress to animals and where confirming that drugs reach therapeutic concentrations in the blood has proven challenging.

The administration of antibiotics to mice either parenterally or bolused enterally involves handling of the mice and induces stress in the animal.³ To ameliorate this handling-associated stress, medications—including antibiotics—frequently are added to the drinking water. This method is time-efficient for laboratory animal personnel and is thought to be of added therapeutic benefit, because it provides continuous accessibility to the medication. As with any route, there are potential limitations to this route of delivery to mice: first, the antibiotic must remain stable in the drinking water and be available for consumption by the mouse; second, mice must drink predictable volumes of treated water; and, last, sufficient concentrations of antibiotic must be maintained in the bloodstream to achieve systemic antibacterial efficacy.

To date, few studies have tested the stability of antibiotics in the drinking water of laboratory mice. For example, one study¹⁶ that tested the stability of amoxicillin–clavulanic acid and trimethoprim–sulfamethoxazole (TMS) in acidified and reverse-osmosis (RO)–treated water found that amoxicillin was stable in RO water but had an immediate drop in concentration to approximately 50% in acidified water, whereas clavulanic acid dropped to 40% in RO over 7 d and immediately was degraded in acidified water. TMS showed variability over the course of 7 d, making reliable dosing with this drug difficult.¹⁶ The cited study did not measure the consumption or systemic absorption of the antibiotics in the mice; therefore although these antibiotics exhibit variable stability in the drinking water, it is unknown whether these drugs reach concentrations sufficient to eliminate pathogenic bacteria.

When antibiotics are administered in drinking water, it is challenging to estimate accurately the total amount of water consumed by the mice. Many factors complicate this measurement, including: spillage of water from the bottle into the bedding; altered taste of the antibiotic-treated water, which may alter the daily water consumption by the mice; ill or unhealthy mice, which may consume less water than would clinically healthy animals, resulting in dehydration and inadequate antibiotic ingestion; and the diurnal pattern of water intake in laboratory mice, which tend to consume most of their daily water intake at the beginning of the dark cycle, creating potential circadian changes in the ingestion of the antibiotics, with the highest dosing occurring at night.^7,19 Therefore, despite the common practice of adding drugs to drinking water, using this route for dosage of antibiotics to mice is unpredictable.

Even when consumption is sufficient, it is complicated to determine whether the antibiotics reach plasma concentrations adequate to eliminate the pathogenic bacteria responsible for the infection. When inadequate antibiotic concentrations occur, there is an increased risk of selecting for drug-resistant pathogens and eliminating normal flora. The risks of failure to achieve adequate drug concentrations are augmented further in laboratory mice, given that many mice are immunosuppressed due to genetic manipulation, radiation, or pharmacologic immunosuppression. These laboratory mice will be almost totally dependent on the bactericidal activity of the antibiotic, with little contribution of the immune system, in the resolution of an infection. The minimum inhibitory concentration (MIC) is the lowest concentration of antibiotic that will effectively inhibit bacterial growth.^10,24 Another important parameter when discussing antimicrobial susceptibility of a given bacterial species is the MIC₉₀. The MIC₉₀ represents the MIC value at which 90% of the bacterial strains within a test population containing multiple independent isolates of the same species are inhibited.

An important factor that potentially limits the efficacy of antibiotics in mice is allometric scaling (also referred to as scaling). Scaling relates to the change in physiologic parameters in species in relationship to body size. Obvious examples of scaling can be recognized as changes in heart rate, gestation, and life expectancy with body size. Factors influencing drug pharmacokinetics, such as metabolic rate, glomerular filtration rate, and hepatic blood flow also scale relative to overall body size, resulting in increased metabolic clearance and decreased drug half lives in small species, such as mice.^4,17,22 Pharmacokinetic features of both enrofloxacin and doxycycline have been demonstrated to have significant scaling effects,^6,21 and amoxicillin is reported to have potential scaling effects.²² The drugs that comprise TMS do not demonstrate evidence of a significant scaling effect, although this drug combination has yet to be tested against scaling parameters other than body size, which may yield a more accurate representation of the effects of scaling on the metabolism of these drugs.^14,15 This increase in the metabolic clearance of antibiotics may limit their ability to achieve the necessary plasma concentrations in mice required for antibiotic efficacy.

The purpose of the current study was to analyze the limitations of the administration of 4 commonly used antibiotics in the drinking water of mice. The antibiotics studied were doxycycline, which typically is administered for the control of gene expression in genetically manipulated mice,²⁸ and TMS, amoxicillin, and enrofloxacin, which are broad-spectrum antibiotics that have been added to the drinking water of mice.^5,16,25 The first experiment tested the stability of 2 of the antibiotics, enrofloxacin and doxycycline, in tap and acidified water, and enrofloxacin in hyperchlorinated water, all of which are commonly used in laboratory mouse vivaria. The second experiment measured the consumption of the 4 antibiotics from the drinking water of mice and the plasma concentrations of the antibiotics that were achieved. In light of anecdotal evidence of a positive therapeutic effect of antibiotics administered in the drinking water of mice, we hypothesized that 1) the antibiotics would be stable in the drinking water, 2) the treated water would be consumed normally by the mice, and 3) therapeutic plasma concentrations would be achieved.

Materials and Methods

Experiment 1: stability of enrofloxacin and doxycycline in drinking water.

The dosages and concentrations of the antibiotics in the drinking water were based on published antibacterial doses and the projected daily consumption of 5 mL by an adult mouse (Table 1). The products used were injectable enrofloxacin (Baytril 100 mg/mL, Bayer HealthCare Animal Health Division, Shawnee Mission, KS); oral pharmaceutical-grade doxycycline calcium (Vibramycin calcium 5 mg/mL, Pfizer Labs, New York, NY); and chemical, nonpharmaceutical-grade doxycycline HCl (Research Products International, Mt Prospect, IL).

Table 1.

Antibiotic doses and concentration of antibiotic in the drinking water

Drug	Daily oral dose (mg/kg; [reference])	Antibiotic concentration (mg/mL) in water
Enrofloxacin	50 (5)	0.25
Doxycycline	10 (18)	0.05
Amoxicillin	50 (23)	0.25
TMS	160 (16)	0.8

Sample collection.

The stability of enrofloxacin in tap, acidified, and hyperchlorinated water was tested over 7 d, and that of doxycycline in tap and acidified water was tested over 7 d (Table 2), in light of the water systems available at the University of Pennsylvania. The antibiotic-treated water was maintained in a standard, clear mouse water bottle (265 mL; Polysulfone Water Bottles, Ancare, Bellmore, NY) and was placed in a complete mouse cage setup that was empty of animals in a mouse holding room. Facility temperatures were maintained at 22.2 ± 1.1 °C (72 ± 2 °F); humidity was between 30% and 70% with 10 to 15 air changes hourly, as recommended by the Guide.¹² The cages were 7.5 in. × 11.5 in. × 5 in. polycarbonate, static isolation cages (Ancare, Bellmore, NY) with 1/4-in. corn cob bedding (Animal Specialties and Provisions, Quakertown, PA). All treated water bottles were shaken daily by the research staff. Samples of treated water were collected on days 0 and 7. At the time of sample collection, 10 mL of treated water was collected by syringe from the end of the sipper tube from each water bottle. Care was taken to disturb the water bottle as little as possible before the sample was collected, to obtain a representative sample of the water that would be available to the mice from the sipper tubes. The samples were stored in centrifuge tubes and frozen at −80 °C until analysis. The effect of the antibiotics on the pH of the water was tested by measuring its pH (pH 510 Benchtop Meter, Oakton Instruments, Vernon Hills, IL) before and after the addition of the antibiotic.

Table 2.

Antibiotic concentrations (mg/mL± SE; n = 4 water bottles) during experiment 1

	Tap water		Acidified water		Hyperchlorinated water
	Day 0	Day 7	Day 0	Day 7	Day 0	Day 7
Enrofloxacin	0.239 ± 0.006	0.237 ± 0.006	0.250 ± 0.004	0.248 ± 0.005	0.246 ± 0.002	0.155 ± 0.019
Doxycycline
Pharmaceutical grade	0.017 ± 0.001	0.017 ± 0.003	0.048 ± 0.001	0.043 ± 0.001	not tested	not tested
Chemical grade	0.052 ± 0.004	0.032 ± 0.001	0.037 ± 0.002	0.042 ± 0.002	not tested	not tested

Analysis of antibiotic concentrations in water.

Water samples were analyzed by using Shimadzu (Columbia, MD) liquid chromatography with a diode array detector. Water samples containing enrofloxacin were diluted 1:10 with 0.5% formic acid containing 10% acetonitrile. Doxycycline-containing water samples were diluted 1:1 into the same diluent. Control samples of both antibiotics were prepared by dissolving in methanol to obtain a concentration of 1 mg/mL of free drug. Standards were prepared that reflected the expected concentrations for each drug: enrofloxacin standards were 0, 5, 10, 25, and 50 µg/mL; and doxycycline standards were 0, 10, 50, 100, and 200 µg/mL. Enrofloxacin was analyzed by using water and acetonitrile (20:80, both containing 0.1% formic acid) in an isocratic run with a flow rate of 0.6 mL/min. Water samples containing doxycycline were analyzed by gradient chromatography using 0.1% formic acid with 0.005 M EDTA and acetonitrile at a flow rate of 0.8 mL/min; the gradient was as follows: 20% acetonitrile for first 2 min, ramp to 70% acetonitrile over 1 min and then held constant for 3 min, and back to original conditions over 1 min. The system was equilibrated for 5 min prior to the next injection. The diode array detector was monitored from 190 to 320 nm, with quantification done at 280 nm for enrofloxacin and at 265 nm for doxycycline. Acetonitrile, methanol, formic acid, and EDTA were purchased from Thermo-Fisher Scientific (Fair Lawn, NJ). Enrofloxacin and doxycycline hyclate standards were obtained from Sigma-Aldrich (St Louis, MO).

Experiment 2: consumption of treated water and serum antibiotic concentrations.

Young adult (6 to 10 wk) female C57BL/6J mice (Mus musculus, Jackson Laboratory, Bar Harbor, ME) were used in this investigation. The mice were housed in polycarbonate cages with bedding, as described earlier, with free access to autoclaved food (Lab Diet 5010, Animal Specialties and Provisions, Quakertown, PA) and were maintained on a 12:12-h light:dark cycle. Prior to the start of the study, the mice were allowed at least 1 wk to acclimate to the housing facility and conditions. Sentinel mice were tested routinely and were free of pinworms by cecal exam and of fur mites by fur pluck and were antibody-negative for tested pathogens including mouse hepatitis virus, mouse parvoviruses, rotavirus, ectromelia virus, Sendai virus, pneumonia virus of mice, Theiler murine encephalomyelitis virus, reovirus, Mycoplasma pulmonis, lymphocytic choriomeningitis virus, mouse adenovirus, and polyomavirus. All aspects of the current investigation were approved by the University of Pennsylvania IACUC.

Mice were pair-housed and randomly assigned to receive 1 of the 4 antibiotics (n = 8 mice for each antibiotic) during the study. The antibiotics tested were enrofloxacin and chemical-grade doxycycline as in experiment 1, amoxicillin (50 mg/mL, Sardoz, Princeton, NJ), and TMS (48 mg/mL, Hi-Tech Pharmacal, Amityville, NY; Table 1). The mice were weighed at the start of the study, and daily water consumption was measured by weighing the water bottles for each pair of mice for 7 d. On day 7, the mice were weighed again, and the antibiotic was added to fresh tap water. The water bottles were shaken daily, and water consumption was measured over an additional 7 d. At the day 14 endpoint, the mice were weighed, and a random half of the mice (n = 4 per group) underwent blood collection into heparinized tubes at 0700; the remaining mice had blood collected at 1900. All blood collection in the study was by retroorbital bleeding under isoflurane anesthesia. Mice were induced at 3% isoflurane until they lost the righting reflex, after which they were maintained at 2.25% for 3 min. This protocol allowed sufficient anesthesia time after removing the mice from the anesthetic to safely collect the blood. Approximately 200 µL blood was collected into heparinized centrifuge tubes at each time point. The mice then were allowed to recover and were returned to their home cages. Two days later, the mice underwent a terminal blood collection at either 0700 or 1900, so that blood was collected from each mouse during both the morning and evening. The blood sample was centrifuged and the plasma separated and frozen at −80 °C until analysis.

To detect the highest possible plasma enrofloxacin concentration, 2 additional groups of 4 mice each were studied. In these mice, the blood was collected at 0100, in an effort to measure the concentration when mice are likely to recently have consumed the greatest water volume (and thus largest therapeutic dose of antibiotic). In addition, the enrofloxacin dose was increased in one group of mice to increase the plasma antibiotic concentration. Specifically, one group of 4 mice received the 50-mg/kg daily dose used in the previous mice, and remaining mice received 100 mg/kg daily.

Analysis of antibiotic concentrations in plasma.

Due to the higher sensitivity required for plasma samples compared with water samples, plasma samples were analyzed by using an API 4000 (AB Sciex, Foster City, CA) liquid chromatography–tandem mass spectrometry system. The system was equipped with a Luna C18 (150 × 4.6 mm, 5-µm particle size) analytical column (Phenomenex, Torrance, CA). For each sample, 50 µL plasma was mixed with 0.1 mL acetonitrile containing 1% formic acid. The mixture was vortexed, centrifuged, and filtered through a 0.22-µm nylon filter prior to analysis. Plasma samples containing doxycycline, enrofloxacin, or TMS were analyzed by using a gradient run with 0.1% formic acid and 85% methanol containing 0.1% formic acid at a flow rate of 0.5 mL/min. The gradient conditions were as follows: methanol for the first 2 min, ramp to 95% methanol over the next 3 min, hold at 95% methanol for 4.5 min, return to the original conditions over 0.5 min, and then hold for 4 min. Samples containing amoxicillin were analyzed by using the same gradient conditions but with 0.1% formic acid and acetonitrile instead of methanol. The following ion transitions were selected to quantitate each antibiotic: doxycycline, 445.4 /154; enrofloxacin, 360/316.2; TMS, 291/261.1; and amoxicillin, 366/143.9. The test samples were quantified against curves obtained by analyzing control bovine serum spiked with antibiotic in concentrations ranging from 0.001 to 0.5 µg/mL. Methanol was purchased from Thermo-Fisher Scientific, and control bovine serum was obtained from Sigma-Aldrich.

Statistical analysis.

Antibiotic concentrations in water were compared by ANOVA (SigmaPlot 12.3, Systat Software, San Jose, CA). Morning and evening plasma antibiotic concentrations were compared by repeated-measures ANOVA. Statistical significance was defined as a P value of less than 0.05.

Results

Enrofloxacin.

Enrofloxacin remained stable in both tap and acidified water throughout the 7-d test period (Table 2). The addition of the injectable enrofloxacin immediately and dramatically increased the pH of both the tap and acidified water (Table 3). In the hyperchlorinated water, a precipitate rapidly formed over the first 24 h. This precipitate was absent from the untreated tap and acidified water, and once formed, the precipitate remained throughout the entire 7-d period. The time 0 sample had the expected antibiotic concentration; however at day 7, only 62% of the antibiotic remained in solution and was available to mice.

Table 3.

pH of antibiotic-treated water after drug addition

	Tap water (pH = 6.99 before addition)	Acidified water (pH = 3.29 before addition)
Enrofloxacin	9.54	8.78
Doxycycline	7.02	3.26
Amoxicillin	6.89	3.50
TMS	6.92	3.39

Note the profound effect of the addition of injectable enrofloxacin on the pH of both the tap and acidified water.

Doxycycline.

The pharmaceutical-grade oral doxycycline immediately dissolved in the acidified water and remained at stable concentrations for the entire 7-d period (Table 2). However when the drug was added to tap water, a precipitate immediately formed and quickly settled to the bottom of the water bottle. HPLC analysis of the water samples revealed that the concentrations of doxycycline at days 0 and 7 were approximately 30% of the expected value (Table 2). The sample then was acidified to a pH of 3.0 with hydrochloric acid and remeasured. This action resulted in a doxycycline concentration that was 90.8% of that expected, indicating that the majority of the active ingredient was present but unavailable for consumption by the mice because it was in the precipitate at the bottom of the water bottle.

The nonpharmaceutical, chemical-grade doxycycline powder was tested in both tap and acidified water. In acidified water, the chemical-grade doxycycline immediately dropped to approximately 75% of the expected concentration and then remained stable over 7-d period. There was no significant difference between the day 0 and day 7 doxycycline measurements. In tap water, the initial concentration was approximately 100% of the expected value, dropping significantly (P < 0.05) to 64% of the expected concentration at day 7. The addition of chemical-grade doxycycline had little effect on the pH of either the tap or the acidified water (Table 3).

Consumption of antibiotic-treated water and measurement of body weight.

The pairs of mice in experiment 2 consumed 9.6 ± 0.2 mL water daily when no antibiotic was added (Table 4). Neither baseline body weight nor water consumption differed between any of the groups. Only the enrofloxacin group had a significant (P < 0.05) change in water consumption during the week of antibiotic administration; consumption increased from 9.7 ± 0.4 mL/d to 11.4 ± 0.2 mL/d per pair of mice. Initial body weight did not differ between any of the groups at the start of the experiments, and mice in all 4 groups gained weight over the next 2 wk.

Table 4.

Water consumption (mL; mean ± SE; n = 4 cages) and plasma antibiotic concentration (ng/mL; mean ± SE)

	Consumptiona		Plasma antibiotic concentrationb at
Antibiotic	Control water	Antibiotic water	0700	1900	0100
Enrofloxacin	9.7 ± 0.8	11.4 ± 0.3c	112.2 ± 11.7	140.1 ± 10.4	117.5 ± 16.9
Doxycycline	9.3 ± 0.3	10.1 ± 0.9	56.6 ± 12.5	42.9 ± 7.8	not tested
Amoxicillin	9.3 ± 0.4	8.8 ± 0.8	299.2 ± 64.1	275.2 ± 50.2	not tested
TMS	10.1 ± 1.3	11.2 ± 1.5	5.7 ± 2.3	5.9 ± 1.2	not tested

^aConsumption data represent 2 mice per cage.

^bn = 8 mice per antibiotic, except for the 0100 enrofloxacin sample (n = 4).

^cValue significantly (P < 0.05) different from that for consumption of control water.

Plasma antibiotic concentrations.

Plasma concentrations (Table 4) showed no significant differences between the morning and evening sampling time points for any of the antibiotics. To maximize the measured plasma enrofloxacin concentration, 2 additional groups of mice were tested at 0100, during the dark cycle. The 0100 plasma enrofloxacin concentrations for the group receiving the 50-mg/kg dose was similar to those of the earlier time points; and the plasma antibiotic concentration of the group that received 100 mg/kg enrofloxacin was 174.8 ± 55.5 ng/mL.

Discussion

Achieving therapeutic concentrations of antibiotic in patients is critical to the efficacy of any antibiotic, independent of species. The current investigation demonstrates that, although the antibiotics tested remained stable in the drinking water and the mice consumed predictable volumes of antibiotic-treated water, plasma concentrations above the reported MIC values for most common pathogenic bacteria (Table 5) were not attained. These findings question the rationale for the common practice of antibiotic administration in the drinking water of mice.

Table 5.

MIC of various antibiotics for common bacteria

	Bacteria	MIC90 (ng/mL; [reference])
Enrofloxacin	E. coli	30-125 (20)
	S. aureus	120-250 (20)
	Enterococcus spp.	1000-2000 (20)
	Pseudomonas aeruginosa	1000-8000 (20)
	P. multocida	500 (27)
Doxycycline	M. pneumonia	500 (26)
	Pasteurella spp.	125 (27)
Amoxicillin	S. aureus	50 (20)
	E. coli	5000 (20)
	S. pseudintermedius	2000 (20)
	C. perfringens	50 (20)
	P. multocida	250 (27)
TMS	S. xylosus	>2000 (25)
	K. pneumonia	<500 (20)
	E. coli	<500 (20)
	β-hemolytic streptococci	2000 (20)
	Pasteurella spp.	250 (27)

The bacteria–antibiotic combinations represent common pathogenic bacteria in veterinary medicine. MIC₉₀ values for ampicillin were used interchangeably with those for amoxicillin. Systemic infections with bacteria in bold can be treated reasonably effectively with the corresponding antibiotic. Note that most of the bacteria isolated during common murine infections lack published MIC₉₀ values for various antibiotics.

The efficacy of antibiotics depends on the pharmacodynamics of the antibiotic–bacteria interaction. Antibiotics are commonly divided into 2 groups according to their pharmacodynamic characteristics: time-dependent, such as β-lactam drugs, in which the efficacy of the drug is determined by the total time the plasma antibiotic concentration is above the MIC of the organism being targeted, and concentration-dependent, such as fluroquinolones, in which efficacy is associated with the peak plasma concentration of the antibiotic.^1,20 Administering antibiotics in the drinking water of mice will optimize the performance of the time-dependent antibiotics, maintaining elevated concentrations of antibiotics in the blood stream whenever mice drink water. In the current study, the amoxicillin plasma concentrations were similar at the start of both the light and dark cycles. However, dosing by water resulted in plasma concentrations that were well below the MIC of most common bacterial pathogens, so that only exquisitely sensitive organisms would be effectively killed by this route of dosing (Table 5).^2,20 The results of the current study are similar to those reported previously¹⁶ regarding plasma levels achieved after antibiotic administration in animals’ food.

Providing enrofloxacin in the drinking water failed to achieve effective plasma concentrations. Enrofloxacin is a concentration-dependent drug, which means that the peak serum concentration (C_max) achieved has been shown to be a critical factor in the efficacy of bactericidal activity. An C_max:MIC value greater than 10 has been shown to predict efficacy.^10,24 Oral bolus dosing of enrofloxacin in dogs has been shown to achieve C_max values of 2.1 to 5.2 µg/mL, whereas the plasma concentration measured in the current murine experiment were only 112.2 ± 11.7 ng/mL at 0700 and 140.1 ± 10.4 ng/mL at 1900. We hypothesized that the peak plasma concentration would occur in the middle of the dark cycle, when mice tend to drink the most water,^7,13 so we measured plasma enrofloxacin concentrations in mice in the middle of the dark cycle and found that, surprisingly, this value (117.5 ± 16.9) was lower than the 1900 value. In an effort to maximize C_max, a second group of mice for which the enrofloxacin dose was doubled were tested in the middle of the dark cycle, but this adjustment resulted in an average plasma concentration of only 174.8 ± 55.5 ng/mL. Considering that the goal is to achieve a C_max:MIC ratio of 10 or greater, these findings indicate that providing enrofloxacin in the drinking water of mice likely will be ineffective against most pathogenic bacteria.²⁰

The plasma concentrations of both TMS and doxycycline were well below the MIC₉₀ values (Table 5) for common pathogenic bacteria, indicating that the administration of these antibiotics by this route for the treatment of systemic infections in mice should be discouraged. The doses reported in the literature for mice are similar to those used in other species. This dosing regimen fails to take into account the effects of allometric scaling on drug metabolism, which as discussed earlier, will result in an increase in drug metabolism and a subsequent decrease in plasma concentration. Both doxycycline and TMS are used frequently with success in laboratory mice for purposes other than systemic bacterial infections. Doxycycline is used most often in genetically manipulated mice in the control of ‘Tet-on’ gene expression, by using a tetracycline-sensitive promoter gene to control either the expression or inhibition of gene expression.²⁸ TMS frequently is added to the drinking water of mice after ionizing irradiation to prevent bacterial sepsis by reducing the number of potential pathologic bacteria within the gastrointestinal tract.⁸

Increasing the amount of antibiotic consumed by the mice can be accomplished by either increasing the concentration of antibiotic in the drinking water or by increasing the amount of water consumed by the mice.¹¹ However, according to our findings, plasma concentrations would need to be increased by 10-fold to achieve effective plasma concentrations through the drinking water or those that are achieved with oral bolus dosing in other species. Further compounding these difficulties are that the plasma concentrations may not increase linearly with increasing doses, meaning it may take more than a 10-fold increase in the amount of antibiotic consumed to achieve the desired increase in plasma concentration.

A potential use of administration of antibiotics in the drinking water of mice involves the treatment of localized infections in mice. Both amoxicillin and enrofloxacin are concentrated in the urine due to renal excretion.¹⁸ This concentration may enable these antibiotics to achieve sufficient urinary concentrations to be effective for the treatment of cystitis and renal infections in mice. The results of the first experiment indicate care must be taken to ensure the stability of the antibiotic–water combination. Enrofloxacin is stable in both tap and acidified water but radically alters hyperchlorinated water, making the addition of this drug to hyperchlorinated water a poor option. When in either tap or acidified water, this antibiotic appeared to be minimally affected by light over the brief time period studied, given that the water concentrations remained stable over the 7-d period. Ultimately, however, bolus dosing of some antibiotics, particularly enrofloxacin (which works in a concentration-dependent fashion), is more likely to achieve effective plasma concentrations.

Several different water treatments are used to prevent the exposure of immunosuppressed mice to bacterial pathogens, particularly Pseudomonas aeruginosa. These include acidification, hyperchlorination, and reverse-osmosis.⁹ The stability and solubility of the antibiotics we tested was dependent on the type of water and the formulation of the antibiotic. Preliminary experiments used a pharmaceutical-grade, oral doxycycline suspension and showed that it dissolved into and was stable in acidified water but remained as a suspension in tap water. We then tested a chemical nonpharmaceutical-grade doxycycline powder that is used by many research laboratories for control of gene expression using the Tet promoter, and the drug demonstrated mild degradation over the 7-d observation period. This distinction is an important one to make for institutional committees that review the use of doxycycline for research purposes, because investigators typically are expected to justify the use of nonpharmaceutical chemical-grade products for research animals. This difference in solubility would be a scientific justification for investigators to choose the chemical-grade product over the pharmaceutical grade product. Similar findings occurred with enrofloxacin, which was soluble in both acidified and tap water but precipitated in hyperchlorinated water. Finally, injectable enrofloxacin had a profound effect on the pH of the water, both acidified and tap, whereas the other antibiotics had little effect on this parameter. The effects of drugs on the drinking water's pH is an important consideration, particularly when adding drugs to acidified water, given that a loss of acidification may favor the growth of Pseudomonas spp. in the water of vulnerable immunosuppressed mice. The finding that the mice drank more of the enrofloxacin-treated water when compared with the untreated control was surprising, considering that enrofloxacin is reported to have a bitter taste. It is possible that the novel taste of the water appealed to the mice and promoted increased drinking during the week of treatment. Future work examining the taste preferences of mice will be valuable in an effort to increase their consumption of medicated water.

The findings of the current study demonstrate that the administration of antibiotics in the drinking water of mice does not result in plasma antibiotic concentrations that are effective against most pathogenic bacteria. Although this oral administration route may be adequate for treatment of some bacterial infections, such as when the antibiotic is concentrated at the site of infection, it is inappropriate for general systemic bacterial infections in which the sensitivity of the pathogenic bacteria has not been identified.