The Effect of Noise, Vibration, and Light Disturbances from Daily Health Checks on Breeding Performance, Nest Building, and Corticosterone in Mice

Abstract

At our institution, the techniques that technicians use for health checks vary for mice housed in cages on individually ventilated caging (IVC) racks. If the mice cannot be adequately visualized, some technicians partially undock the cage whereas others use an LED flashlight. These actions undoubtedly alter the cage microenvironment, particularly with regard to noise, vibration, and light, which are known to affect multiple welfare and research-related parameters in mice. The central aim of this study was to assess the effects of partial cage undocking and LED flashlight use during daily health checks on fecundity, nest building scores, and hair corticosterone concentrations in C57BL/6J mice to determine the least disturbing method of performing these health checks. In addition, we used an accelerometer, a microphone, and a light meter to measure intracage noise, vibration, and light under each condition. Breeding pairs (n = 100 pairs) were randomly assigned to one of 3 health check groups: partial undocking, LED flashlight, or control (in which mice were observed without any cage manipulation). We hypothesized that mice exposed to a flashlight or cage undocking during daily health checks would have fewer pups, poorer nest building scores, and higher hair corticosterone levels than did the control mice. We found no statistically significant difference in fecundity, nest building scores, or hair corticosterone levels between either experimental group as compared with the control group. However, hair corticosterone levels were significantly affected by the cage height on the rack and the amount of time on study. These results indicate that a short duration, once-daily exposure to partial cage undocking or to an LED flashlight during daily healthy checks does not affect breeding performance or wellbeing, as measured by nest scores and hair corticosterone levels, in C57BL/6J mice.

Introduction

Animals must be assessed each day to ensure their wellbeing and to comply with regulations.²⁹ Husbandry technicians assess each mouse daily to verify their health, the availability of water and food, and appropriate cage conditions. However, our technicians use one of 3 techniques to perform health checks. First, if the mice and cage can be fully assessed visually, the cage is not disturbed. Second, if mice cannot be adequately viewed, some technicians partially undock cages from the rack to allow adequate visualization (without opening the cage lid). Third, technicians are supplied with a rechargeable cool-white LED (light-emitting diode) flashlight that can be used to help with mouse visualization. While both cage undocking and flashlights facilitate in visualization of the mice and the cage, these actions could also confound research results and alter animal wellbeing by altering noise, vibration, and light in the cage environment.

Chronic noise and vibration exposure due to disruptions such as construction and nearby train activity have been shown to affect reproductive efficiency,⁵³ cardiovascular parameters,^4,35 fecal corticosterone metabolites,² and intestinal mucosal morphology⁵ in rodents. Similarly, variations in photoperiod, light intensity, and light spectral quality have been shown to affect various experimental parameters and measures of wellbeing.^{3,16,45,60,62} For example, rats housed under dim lighting conditions or exposed to a long dark phase displayed lower resting heart rates,³ certain strains of mice exposed to blue LED light experienced beneficial effects on circadian regulation of neuroendocrine, metabolic, and physiologic parameters,⁹ and exposure to light at night caused adverse changes in metabolism, immune function, anxiety, and cognition.^{7,8,10,11,17-19,25,44,49} The Guide for the Care and Use of Laboratory Animals recommends controlling noise, vibration, and light levels, and even provides recommendations for appropriate light level ranges in animal housing rooms.²⁹ However, these variables may be overlooked when implementing husbandry practices, and are not assessed or documented regularly in the same way that environmental temperature and humidity are strictly controlled.⁶³ Furthermore, ambient sound, vibration, and light levels vary significantly based on room, rack, and activities performed in the room,⁶ highlighting the need for each institution to perform its own systemic evaluation of these parameters.

Exposure to a pulse of light during the dark phase is well-known to affect nocturnal rodents’ circadian clock by either shifting the clock forward or backward, depending on the duration, timing, and intensity of light pulse.^18,41,57 However, relatively little information is available on whether a pulse of light during the light phase affects animal wellbeing, stress, fecundity, or various parameters linked to the circadian clock. Therefore, the effects of a cool-white LED flashlight during daily health checks on mouse welfare, reproductive performance, and experimental variables are not known. Similarly, because most published studies focus on chronic exposure to noise and vibration (for example, during construction),^2,53,59 little is known about the effect on mice of an acute increase in noise and vibration, as would occur with cage undocking,. Intracage vibration levels are recommended to remain below 25 milli-g (RMS, root mean square), as this level has been shown to cause increased fecal corticosterone metabolites and overt behavioral reactions in female mice.^2,20,63 However, those studies involved protracted exposure to vibration (that is repeated multiple times per day for multiple minutes) as compared with that produced by partially undocking individual cages. Intracage noise levels should remain below 70 dB SPL to prevent damage to the auditory system and to prevent a cascade of stress responses.^63,64 The act of connecting ventilated cages to a rack has been previously shown to produce noise intensities in the 85- to 100-dB range, which is loud enough to produce an acoustic startle response in mice.^30,63 However, whether an acute, once-daily cage undocking would affect welfare or experimental outcomes is unknown. With a recent availability of automated systems that provide electronic remote monitoring of rodent cages, examination of the potential consequences of manual daily health checks is important for making evidence-based decisions about how daily health checks are performed.

The central aim of this study was to assess the effects of partial cage undocking and flashlight use during daily health checks on fecundity, nest building, and hair corticosterone in mice. Noise, vibration, and light monitoring equipment was used to measure average intracage noise and vibration levels associated with cage undocking, and to determine average intracage illumination levels associated with cool-white LED flashlight use during daily health checks, as compared with control conditions. We hypothesized that the mice exposed to a flashlight or cage undocking during daily health checks would have lower fecundity, poorer nest building, and elevated hair corticosterone levels compared with the control mice, which were not exposed to a flashlight or cage undocking during daily health checks.

Materials and Methods

Animals and housing.

Mice were housed in the University of Chicago ARC facilities RRID:SCR_021806. The University of Chicago’s animal program is accredited by AAALAC International, and all animal work was approved by the University of Chicago’s IACUC. This project adheres to the ARRIVE Essential 10 Guidelines.⁵⁰ Male and female C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME) arrived at 6 wk of age and were allowed to acclimate for 1 wk prior to the initiation of experiments. During the acclimation period, mice from the same shipping container were housed in groups of 5 per cage. C57BL/6J mice were chosen for the experiment due to their common use both at our institution and in biomedical research overall.

Mice were housed on a single-sided rack (Jag 75 Micro-VENT Environmental System IVC racks, Allentown, Allentown, NJ) with the motor positioned above the rack. Solid-bottom polycarbonate IVC cages (19.69 × 30.48 × 16.51 cm; Allentown Jag 75 Micro-Barrier, Allentown NJ) were used with 60 air changes per hour. Mice were housed on corncob bedding (1/4-in. [.64-cm]; Teklad 7097, Envigo, Indianapolis, IN), provided acidified water (pH, 2.8 to 3.2) in water bottles, and fed an irradiated diet (Teklad 2918, Envigo), provided ad libitum. Mice received 4 g of specialty shredded paper (Bed-r’Nest, Lab Supply, North Lake, TX) for enrichment. All cages, bedding, and enrichment were autoclaved prior to use. Cages were changed every 14 d in a class-II type A2 biosafety cabinet (NuAire, Plymouth, MN). Mouse rooms were maintained on a 12:12-h light:dark cycle, with lights on at 0600 and lights off at 1800, with humidity ranging from 30% to 70% and temperatures ranging from 68 to 76 °F (20.0 to 24.4 °C), in compliance with the Guide for the Care and Use of Laboratory Animals.²⁹ The mean illumination level at 1 m above the floor in the center of the housing room was 293 ± 3 lux (n = 8 measurements). Illumination was provided by broad-spectrum (300- to 750-nm) fluorescent lighting with the largest spectral power peaks occurring at approximately 540 nm and 612 nm. The bulbs were 48-in. (122-cm) fluorescent bulbs (FO32/835/ECO, color temperature: 3500 Kelvin, Sylvania, Wilmington, MA) with one bulb per ballast in a room that held 8 ballasts.

Intracage ambient illuminance during the lights-on period was measured in 6 different positions within (back right, back left, middle right, middle left, front right, front left) a cage that was located at eye level in the center of the rack. The mean illumination level of these 6 measurements was 14 lux. Illuminance was measured using the light meter described below. No light contamination was present during the dark phase.

Excluded agents, as assessed by using PCR testing of exhaust dust, as described previously,³⁸ were Sendai virus, pneumonia virus of mice, mouse hepatitis virus, mouse parvoviruses, reovirus, epizootic diarrhea of infant mice, 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 Hymenolepis spp., Giardia muris, Encephalitozoon cuniculi, Myobia musculi, Myocoptes musculinus, Radfordia affinis, Psoregates simplex, Syphacia spp., and Aspiculuris tetraptera.

Study design.

Breeding pairs were set up at 7 wk of age and randomly assigned to 1 of 3 experimental groups using an online random number generator (GraphPad, San Diego, CA). A total of 100 breeding pairs (n = 100; 200 mice) were used for this experiment, based on a one-way ANOVA power analysis performed before study initiation. To limit environmental disruptions due to the activity of other investigators, these mice were the only animals housed in this specific housing room. The 3 experimental groups represented 3 methods of performing daily health checks: the undocking group (n = 33), the flashlight group (n = 33), and the control group (n = 34) (Figure 1). Cages in the undocking group were health checked each day by partially undocking the cage and sliding the cage halfway off the rack so that half of the cage length was exposed. The cage was partially undocked by sliding it forward so that half the cage length was exposed over 1 s, the mice and environment were evaluated for 5 s, and the cage was then redocked by sliding it back into its original position over 1 s. Cages were undocked and redocked by placing one hand on each side of the cage front; the cage lid was not opened during this process. Cages in the flashlight group were exposed to a cool-white, LED, rechargeable, 40-lumen flashlight that is routinely used by husbandry staff to perform daily health checks of rodent cages (Energizer WeatherReady, St Louis, MO). These cages were not touched or partially undocked. The LED flashlight provided broad-spectrum light but was high in emission of blue-appearing light, with intensities peaking at approximately 455 nm (see Figure 2). The flashlight was held 6 in. from the front of the cage directed just below the cage card, shining at a 90° angle to the cage front for a total of 5 s while observing the mice, and then switching the flashlight off and moving to the next cage. Cages in the control group were health checked by observing the mice and cage environment for 5 s without touching the cage or using a flashlight. A step stool was used to evaluate the top 2 rows of cages. All cages were health checked between 0800 to 1000 every day for 6 mo, including weekends, by the primary investigator (BC) or, rarely, by one other individual (RT) who was trained by the primary investigator.

Figure 1.

Breeding pairs of C57BL/6J mice were randomly assigned to 1 of 3 groups; The partial undocking group (n = 33 breeding pairs), the flashlight group (n = 33), or the control group (n = 34). Cages were either (A) partially undocked so that half the cage length was exposed for 5 s during health checks, (B) exposed to an LED flashlight held 6 in. from the cage front for 5 s, or (C) observed for 5 s without cage interaction.

Figure 2.

Relative spectral power distribution of the LED flashlight used in the study. Curve was drawn based on spectral data provided by the manufacturer. The light is broad spectrum, with peak intensity at approximately 455 nm (blue-appearing light).

During health checks, each cage was evaluated for the presence of food and water, the health of mice, and the state of the cage environment. Daily health checks could not be performed in a blind manner, as each cage required a specific type of health check (cage undocking, flashlight use, or control) based on the assigned group, and each cage had a brightly colored cage card to denote the experimental group to which it belonged. Two single-sided IVC racks were placed equidistant from the biosafety cabinet, which was located on the opposite side of the room. Although rack 1 was located closer to the room door, any room equipment, such as cage change supplies, were wheeled completely across the room to the opposite side, so both racks were exposed to approximately the same amount of husbandry activity. In addition, both racks were housed against the same back wall, which is adjacent to a quiet facility hallway that has low foot traffic. Therefore, the rack position in the room is unlikely to serve as a confounding variable. The animal room had no light contamination during the dark phase. Rack 1 housed the control group on the right side and the flashlight group on the left side, with construction paper installed to divide the 2 groups, thus protecting the control group from light produced by the flashlight (Figure 3). Before initiation of the experiment, cages in the control group were tested with a light sensor (described below) to verify that light from the flashlight group did not affect light levels in the control cages. In addition, ambient light within cages did not change after the paper barrier was constructed. Rack 2 housed the undocking group to avoid exposing the control and flashlight groups to noise and vibration from the undocking of cages. All 3 groups had an approximately equal number of cages on each rack row (each group contained 3 to 4 cages on each row of the rack). Cages housed on rows 1 to 3 were considered “high,” rows 4 to 7 were considered “mid,” and 8 to 10 were considered “low” for statistical analysis to determine if cage height on the rack affected any of the parameters measured.

Figure 3.

IVC rack that housed the flashlight and control groups. A construction paper barrier, indicated by red dashed line, was installed with masking tape between the flashlight group (left) and the control group (right) to prevent light from the flashlight from reaching control mice.

Noise, vibration, and light monitoring.

An empty sham cage was set up with a Sensory Sentinel device (Turner Scientific, Jacksonville, IL) which includes an accelerometer, microphone, light sensor, and temperature/humidity sensor (Figure 4). The cage contained standard corncob bedding and shredded paper enrichment. The food hopper was filled with standard diet, and a water bottle was placed in order to approximate the weight of a typical mouse cage. The accelerometer was placed flush with the polycarbonate cage floor and secured to the cage floor with masking tape. The microphone was secured off the cage bottom by resting it on a square of foam padding, per manufacturer recommendations. The accelerometer, microphone, and light sensors were placed in the front third of the cage, because this is where most mice are observed to build their nests, sleep, and spend most of their time at our institution. The temperature and humidity monitor was placed in the center of the cage. Wires from all 4 monitoring devices were tunneled out the back of the cage through the automatic watering grommet and were connected to the Sensory Sentinel tablet. The sham cage was rotated every 7 d between the 3 experimental group locations and was treated in exactly the same way as mouse cages in that group during daily health checks. The sham cage was moved weekly among the 3 groups for the total duration of the study (6 mo), resulting in a total of n = 43 noise, vibration, and light measurements taken per group, or n = 129 measurements total. One empty cage position was present for each group, located on the 5^th row, one cage spot in from the outer column, for placement of the sensory sentinel device. Environmental data were recorded 24 h a day and automatically saved to the Sensory Sentinel device. An activity log was placed on the door to allow husbandry staff to record activities performed in the room, such as cleaning and restocking supplies, so that erroneous noise and vibration at specific times could be linked to certain husbandry activities.

Figure 4.

Setup of sham cage with environmental monitoring equipment. Microphone was supported by foam that was taped to the cage floor to keep microphone off of the cage floor. Accelerometer was taped to the cage floor so that the base of the accelerometer rested flush to the cage bottom with no bedding between the accelerometer and cage floor. Light sensor rested directly on top of bedding.

An ultrasonic microphone (PCB Piezotronics, Depew, NY; preamplifier model 426A11, microphone tip model 377C01, adapter model 079A02) purchased as part of the Sensory Sentinel system (Turner Scientific, Jacksonville, IL) was used. The microphone was set to capture and record noise levels, including ultrasonic noise, within the hearing range frequency of mice, 900 to 96,000 Hz, and had a sensitivity of 2 mV/Pa. Whenever the sham cage was moved to a new location (once a week for the duration of the study), the microphone was calibrated using an acoustic calibrator that played a sample tone at 94 dB,.

The accelerometer (PCB Piezotronics, Depew, NY; model 352C33) measured vibration in the frequency range of 2 to 500 Hz and as the root mean squared in the z axis.⁶³ The z axis is appropriate because most vibration in an animal facility that reaches animals tends to occur in the z axis.⁶³ The accelerometer had a sensitivity of 101.3 mV/g and a measurement range of ± 50 g pk. It was calibrated by PCB Piezotronics directly before its purchase in June 2021.

The light sensor (Yoctopuce, Cartigny, Switzerland, model Yocto-light-V2) is a USB light-sensing chip that measures illumination levels in the range of 0.01 to 65,000 lux with a refresh rate of 10 Hz. The sensor uses 2 distinct detectors, sensitive to different spectrums, and their responses can be combined to compute an estimated human eye perception. The sensor uses a transformation function to provide a value in lux based on the value measured by the 2 detectors, taking into account both the estimated spectrum and the measured intensity. This light sensor is appropriate for measuring multiple light types, including LED and fluorescent lighting, and measures light in the frequency range between 300 and 1100 nm. Illumination (units of lux) measures the amount of light that is perceived on a surface by the human eye. Measurements in lux are generally accepted for human daytime vision, but not for quantifying light-phase vision in animals, or for quantifying the light that would affect circadian or neurobehavioral responses in mammals.^8,25,37 Radiometric values of irradiance (μW/cm²) are appropriate for quantifying light for these purposes.^8,9,25,37 Many researchers who study behavior and neurobiology also recommend reporting the spectral power distribution of light sources.^37,49 Adhering to this recommendation is important because all species have unique spectral sensitivities to light, and animals, including mice, have both visual and nonvisual responses to light exposure that arise from ocular photoreceptors, which include rods, cones, and the recently discovered melanopsin-containing intrinsically photosensitive retinal ganglion cells (ipRGCs).^9,13,36,49 Therefore, the lux value of illumination was converted to a radiometric approximation. We also included calculations of estimated lux-derived units based on the photopigments of the mouse retina, based on a rodent irradiance toolbox provided for this purpose (available from: www.ndcn.ox.ac.uk/team/stuart-peirson). Specifically, the light source (LED light for flashlight group, white fluorescent for control group) and mean intracage photopic illuminance level (in lux) at the time of daily health checks were entered into the toolbox for both the flashlight group and the control group. The ‘approximate’ mode in the toolbox was used to calculate irradiance, photon flux, and mouse retinal photopigment values. Because the approximate mode assumes general spectra of the light based on the light type, these calculations are estimates and not exact conversions.

At the conclusion of the project, noise, vibration, and light were measured at select cage positions on the rack to determine ambient environmental parameters at various cage heights. Four cage slots each on rows 1 and 2 were designated ‘high,’ on rows 5 and 6 were designated ‘mid,’ and on rows 9 and 10 were designated ‘low.’ Noise, vibration, and light recordings were taken for 30 s in each cage position when no other activity was occurring in the room in order to document the ambient intracage environmental parameters at each rack position.

Breeding performance.

Breeding performance was chosen as a central outcome measure in this study because many researchers have concerns regarding breeding performance of sensitive lines of mice. In addition, breeding mice are particularly susceptible to noise and vibration.^2,53 Each cage was observed every day between 0800 to 1000 for the presence of new litters. The date of birth of each litter was recorded on a breeding card, and each litter was weaned at 21 d of age. At weaning, the weight and sex of each pup were recorded. If a litter was born and no longer present at weaning, or if only dead pups or pup remnants were noted in the cage, the litter was designated as a ‘lost litter.’ The percentage of lost litters, latency to first surviving litter born, average pup weight (grams) at weaning per breeding pair, and breeding index were all calculated for each breeding cage. The breeding index was calculated as the average number of pups weaned per week that the breeding pair was housed together on study (breeding index = total # of pups weaned/# of weeks on study). For statistical analysis, a ‘low’ breeding index was considered to be ≤1, and a ‘high’ breeding index was considered to be >1. This cutoff value was chosen due to the fact that approximately half of all breeders had a breeding index > 1. Breeding pairs were assessed for breeding performance for 6 mo (24 wk).

Hair corticosterone.

Hair corticosterone was chosen as the most appropriate measure of corticosterone levels for this study for multiple reasons. Corticosterone and its metabolites in mice are commonly measured in blood, saliva, urine, and feces, but these samples represent circulating corticosterone levels that were present in blood between a few minutes to a few days before sampling.^40,47 However, hair samples can accumulate corticosterone over a period of weeks to months, depending on the time points chosen for sample collection.^15,55 The use of hair has been validated as a measure of long-term changes in corticosterone in rodents,^15,31,58,67 and a growing body of evidence supports the use of hair corticosterone as a measure of chronic or repeated stressors.^26,67 Because this study is examined the effects of daily health checks over a 6-mo period, hair corticosterone was the optimal way to detect chronic stress. In addition, hair samples can be collected noninvasively, diurnal variations in corticosterone are not relevant for this method, and the corticosterone in the hair shaft is relatively stable over time when protected from light after collection.^26,67 To ensure that the hair sample contains enough actively growing hairs that will contain representative quantities of corticosterone metabolites, the “shave-re-shave” method is recommended, in which a certain area is shaved at the beginning of the time period of interest, and the regrown hair is again shaved at the end of this period.^14,26,40 At the beginning of the study, we used electric clippers (Philips Norelco Multipurpose Trimmer, Philips, NV) to shave a 3 × 3-cm square area over the caudodorsal rump of each male and female breeder mouse. These hair samples were disposed of and not used for the study.

After 3 mo on study, hair regrowth from the previously shaved area was shaved again and collected into a microfuge tube (Eppendorf, Hamburg, Germany) that was then wrapped in tin foil to prevent degradation of corticosterone via light exposure. Samples were shipped to the Oregon National Primate Research Center Endocrine Technologies Lab (Beaverton, OR) for the corticosterone assay. After 6 mo on study, hair was again shaved from the same area and submitted for corticosterone analysis. Hair from both male and female breeders was collected and analyzed at both the 3- and 6-mo time points. The 3-mo and 6-mo time points were chosen because at least 5 mg of hair was required to run the corticosterone analysis, and we wanted to ensure that each mouse had grown back enough hair to run the analysis. Hair regrowth varies significantly among individual mice; a recent study showed that at 9 wk after shaving, hair regrowth in ICR mice ranged from 0% to 100%.³⁴ While most mice in the current study grew back enough hair within 1 mo, the 3-mo time window was necessary for a small portion of the mice to regrow sufficient hair for analysis.

Hair corticosterone concentrations were measured by radioimmunoassay in the Endocrine Technologies Core (ETC) at the Oregon National Primate Research Center (ONPRC). Hair samples were washed with isopropanol, drained through P8 filter paper (Thermo Fisher Scientific, Waltham, MA), dried, and weighed. Because the hair was already in sections as small as if they had been chopped or ground, these steps were not performed. Hair was weighed into a microtube, 1 mL methanol was added to hair samples, and corticosterone was extracted overnight with gentle shaking. Samples were then centrifuged to collect hair and 400 µL of supernatant was transferred to 13 × 100 glass tubes and dried under nitrogen in a RapidVap Vertex evaporator (Labonco, KS City, MO). Samples were then dissolved in 0.1% gel-PBS and assayed for corticosterone using an inhouse radioimmunoassay. A standard curve ranging from 5 to 1000 pg/tube was created using ³H-corticosterone (American Radiolabeled Chemicals, St. Louis, MO). The antibody used was a commercially available anticorticosterone antibody (Abcam, Cambridge, MA). Hormone values were corrected for extraction losses determined by radioactive trace recovery at the same time as sample extraction (95.8%). The sensitivity was 5 pg/tube. The intra-assay variation was 2.8%, and interassay variation was 2.9% (n = 2 assays).

Cage change and nest scoring.

At the time of cage change (once every 14 d), a nest score was recorded using previously developed scoring systems,^22,27 which take into account the completeness of a nest and the height of each side. Briefly, each nest was scored as a 0 for no manipulation of the nesting material, 1 for manipulation of the nesting material but not shaped into a nest with 4 complete sides, or 2 for a nest with 4 complete sides. For all nests that scored a 2, the maximum height of each side of the nest was measured using a ruler, and an average nest height was calculated. This average nest height was added to the original nest score to achieve a final nest score for each cage. Nests were scored immediately prior to moving the mice to their new cage in order to avoid disturbing the nest architecture and height. The nest scoring was scored only at the time of cage change in order to avoid introducing confounding effects on the study. The nest and/or nesting material was not removed or disturbed for the 2 week duration between cage changes, and 4 g of fresh specialty shredded paper nesting material (Bed-r’Nest, Lab Supply, North Lake, TX) was placed in each new cage at the time of cage change. Old nesting material was not moved to clean cages to ensure that available quantity of nesting material was constant throughout the study. Cage change and nest scoring were performed only by the primary investigator (BC). Blind scoring was not possible because each cage was clearly marked to denote its group for health checks. Mice were transferred to fresh caging by using forceps dipped in Clidox (Pharmacal, Waterbury, CT) to grasp the base of the tail, as this was the standard practice for cage change at our institution at the time of this study. Cages containing mice with litters that were less than 7 d old were not changed until the next cage change date, unless the cage was excessively soiled.

Statistical analysis.

Data were recorded into spreadsheets for record keeping (Excel, Microsoft, Redmond, WA). Average peak noise, vibration, and light levels during daily health checks were presented as descriptive statistics (mean ± SD) and were compared with the control using a student t test. All analyses were performed in R-citation (R Foundation for Statistical Computing, Vienna, Austria). A P value of 0.05 or less was considered significant. Nest scores across time were analyzed using a linear mixed effects model to account for correlation over time of measurements for each cage. Package lme4 was used to perform the analysis. Time was modeled using a cubic model that included interactions between group and time. When comparing nest scores at each time point, each intervention group was compared with the control using the Wilcox rank-sum test, with a Bonferroni correction of 0.05/12, approximately 0.004. Breeding index and average weight of pups at weaning were compared using linear regression, which also allowed inclusion of rack height to ensure that rack height was not a confounding factor. Wilcox rank-sum test was also used to compare the proportions of litters lost between the control and each intervention group. Corticosterone concentrations in hair were analyzed across time using a linear mixed effects model. This approach permitted analysis of correlation of corticosterone levels for each mouse with both rack height and breeding index to determine whether they were or were not relevant factors. Package lme4 was used to perform the analysis. Wilcox rank-sum test was used to compare each intervention group to the control at each time point.

Results

Mice.

Three breeding pairs were excluded from the study due to health issues unrelated to experimental manipulation (imperforate vagina, spontaneous hemothorax due to cardiac anomaly, and dehydration due to malfunctioning water bottle valve), for a total of n = 97 breeding pairs. Mice began the study at 7 wk of age and finished the study 6 mo later at 31 wk (approximately 8 mo) of age.

Noise, vibration, and light monitoring.

Average peak illumination level during health checks for the control group was significantly lower (P < 0.001) than the flashlight group, at 29 ± 4 lux for the control group and 536 ± 165 lux for the flashlight group. Illuminance measurements were taken 43 times throughout the study for each group. Using the Rodent Irradiance Toolbox mentioned above, the average peak illuminance was used to approximate irradiance, photon flux, and mouse retinal photopic illuminance values for both the flashlight and control groups (see Table 1).

Table 1.

Measured photopic illuminance and approximated irradiance, photon flux, and mouse retinal photopic illuminance values for the flashlight and controls. Rodent Irradiance toolbox was used to convert measured cage-level illuminance into all other approximated values below. The control group values also represent the ambient fluorescent room-level light as measured at cage level.

	Radiometric and Photometric Values			Mouse Retinal Photopic Illuminance (α-opic lux) (approximated)
	Photopic illuminance (lux) (measured)	Irradiance (µW/cm2) (approximated)	Photon flux (cm2/s) (approximated)	S Cone	Melanopsin ipRGC	Rod	M Cone
Flashlight Group (LED Flashlight)	535.8	166.6	47.0 E^13	0.2	359.2	385.5	403.6
Control Group (Fluorescent Room Light)	28.6	8.5	2.35 E^13	0.8	15.7	18.0	19.5

Average peak noise level during health checks for the control group was significantly lower (P < 0.001) than for the undocking group, at 55 ± 9 dB SPL for the control group and 84 ± 7 dB SPL for the undocking group. Average peak vibration level during health checks for the control group was also significantly lower (P < 0.001) than the undocking group at 6.3 ± 2.5 milli-g for the control group and 260 ± 98 milli-g for the undocking group.

The ambient noise, vibration, and light levels at various cage heights on the rack are summarized in Table 2. Statistically, the ‘high’ rows had significantly higher noise (P = 0.007), vibration (P = 0.002), and light levels (P = 0.003) than the ‘low’ rows. The ‘mid’ rows had significantly lower noise (P = 0.01) and higher light levels (P < 0.001) than the ‘low’ rows, but not statistically higher vibration levels (P = 0.36).

Table 2.

Comparison of mean environmental parameters at various cage positions on the rack. Data are presented as mean ± SD. Measurements were taken for 30 s at each cage position, and mean noise, vibration, and light levels were calculated. Four cages were measured on each of rows 1, 2, 5, 6, 9, and 10. Rows 1 and 2 were considered “top,” rows 5 and 6 were considered “mid,” and rows 9 and 10 were considered “low.”

Cage position	Noise (DB SPL)	Vibration (milli-g)	Light (lux)
Top	59.8 ± 3.0	3.8 ± 1.2	89.4 ± 47.7
Mid	55.1 ± 0.6	2.1 ± 0.5	30.3 ± 4.4
Low	56 ± 0.6	1.8 ± 0.4	16.0 ± 3.4

Breeding performance.

The mean breeding indices (number of pups/breeding pair/week) for the control, flashlight, and undocking groups were 1.06 ± 0.34, 1.07 ± 0.36, and 1.05 ± 0.25, respectively, with no statistically significant differences when comparing the flashlight group (P = 0.91) or undocking group (P = 0.92) to the control group. The mean pup weights at weaning for the control, flashlight, and undocking group were 8.9 ± 1.2, 8.8 ± 1.0, and 8.8 ± 0.7 g, with no statistically significant difference when comparing the flashlight group (P = 0.47) or the undocking group (P = 0.51) to the control group. The mean percentage of litters lost for the control, flashlight, and undocking groups were 19 ± 18%, 19 ± 15%, and 21 ± 17% respectively, with no statistically significant differences when comparing the flashlight group (P = 0.43) or the undocking group (P = 0.75) with the control group (Figure 5). The mean latency to first litter born for the control, flashlight, and undocking groups were 24 ± 7, 23 ± 9, and 24 ± 6 d respectively, with no statistically significant difference when comparing the flashlight group (P = 0.75) or the undocking group (P = 0.51) to the control group. Finally, an interaction was not detected between rack height and any of the breeding parameters measured (breeding index: P = 0.689; pup weight at weaning: P = 0.875; percentage of litters lost: P = 0.772).

Figure 5.

Box plots of various breeding parameters for flashlight use, partial cage undocking, and control cages during daily health checks. (A) The mean breeding index (no. of pups/breeding pair/week) for the control, flashlight, and undocking group were 1.06, 1.07, and 1.05 respectively. No statistically significant differences were detected when comparing the flashlight and control groups (P = 0.91) or the undocking and control groups (P = 0.92). (B) The mean pup weight at weaning for the control, flashlight, and undocking group were 8.9, 8.8, and 8.8 g, respectively. No statistically significant differences were detected when comparing the flashlight and control groups (P = 0.47) or the undocking and control groups (P = 0.51). (C) The mean percentage of litters lost for the control, flashlight, and undocking group were 18.6%, 19.2%, and 21.1% respectively. No statistically significant differences were detected when comparing the flashlight and control groups (P = 0.43) or the undocking and control groups (P = 0.75). Box and whiskers indicate interquartile range, min, and max. Line and ‘x’ indicate median and mean, respectively.

Hair corticosterone.

Male and female breeders in each experimental group had no significant differences in hair corticosterone levels as compared with the control group at either time point measured (Figure 6). However, sex, time point, breeding index, and rack position (high, mid, or low) all had a significant effect on corticosterone levels. These effects were analyzed using all mice, regardless of group. Females had significantly higher corticosterone levels (42 pg/mg ± 38) than did with males (31 pg/mg ± 23;) (P < 0.001). Females had significantly higher corticosterone levels (60 pg/mg ± 46) at the 3-mo time point than they did at the 6-mo time point (24 pg/mg ± 10) (P < 0.001). Males also had significantly higher corticosterone levels (41 pg/mg ± 29) at the 3-mo time point than they did at 6-mo time point (20 pg/mg ± 5) (P = 0.02). The breeding index at 3 mo had no significant effect on corticosterone (P = 0.94), but females with a high breeding index had significantly higher corticosterone levels (25 pg/mg ± 10) at 6 mo compared with those with a low breeding index (22 pg/mg ± 10) (P = 0.048). At the 3-mo time point, both males and females on both the top and mid rows of the rack had significantly higher corticosterone levels than did mice on the bottom rows (Figure 7). However, hair corticosterone concentrations were not different among the top, mid, and bottom rows at the 6-mo time point.

Figure 6.

Box plots of hair corticosterone levels at both 3 and 6 mo time points, comparing flashlight use, partial cage undocking, and control cages during daily health checks. (A) For males, 3 mo hair corticosterone levels (mean = 41 pg/ng) were significantly higher than 6 mo levels (mean = 20 pg/ng, P < 0.001). No statistically significant differences were detected when comparing either experimental group and the control at either 3 mo (flashlight compared with control, P = 0.313; undocking compared with control, P = 0.739) or 6 mo (flashlight compared with control, P = 0.789; undocking compared with control, P = 0.331). (B) For females, 3 mo hair corticosterone levels (mean = 60 pg/ng) were significantly higher than 6 mo levels (mean = 24 pg/ng, P < 0.001). No statistically significant differences were detected when comparing either experimental group with the control group at either 3 mo (flashlight compared with control, P = 0.712; undocking compared with control, P = 0.724) or 6 mo (flashlight compared with control, P = 0.267; undocking compared with control, P = 0.074). Box and whiskers indicate interquartile range, min, and max. Line and ‘x’ indicate median and mean, respectively.

Figure 7.

Bar graphs of mean hair corticosterone levels at both 3 and 6 mo time points, comparing top (rows 1–3), mid (rows 4–7), and low (rows 8–10) rack positions. (A) Male mice in the low rack position had significantly lower corticosterone levels (mean = 19 pg/ng) than the mid position (mean = 37 pg/ng, P = 0.002) and the top position (mean = 66 pg/ng, P < 0.001) at the 3-mo time point. At the 6-mo time point, no statistically significant differences were detected between the low rack position (mean = 21 pg/ng) and either the mid rack position (mean = 20 pg/ng, P = 0.40) or the top rack position (mean = 19 pg/ng, P = 0.29). (B) Female mice in the low rack position had significantly lower corticosterone levels (mean = 27 pg/ng) than those in the mid position (mean = 50 pg/ng, P = 0.015) and the top position (mean = 100 pg/ng, P < 0.001) at the 3-mo time point. At the 6-mo time point, no statistically significant differences were detected between the low rack position (mean = 25 pg/ng) and either the mid rack position (mean = 21 pg/ng, P = 0.10) or the top rack position (mean = 25 pg/ng, P = 0.99). Females (B) had significantly higher corticosterone levels (42 pg/mg ± 38) than did males (A) (31 pg/mg ± 23) (P < 0.001).

Nest building.

Nest building scores were not significantly different between the control group and either experimental group at any of the 12 time points measured (Table 3). There was no interaction between rack height and nest scores (P = 0.235).

Table 3.

Comparison of mean nest scores and P values. Data are presented as mean ± SD. Nest scores were taken at the time of cage change, once every 2 wk, for a total of 24 wk, for each cage. P values are considered significant if < 0.004 due to a Bonferroni correction. There were no significant differences in nest scores between either of the experimental groups compared with the control at any time point measured.

	Mean Nest Score			P value
Week	Undocking ( n = 32 )	Flashlight ( n = 32 )	Control ( n = 33 )	Undocking compared with Control	Flashlight compared with Control
2	6.0 ± 1.5	6.3 ± 1.8	6.5 ± 2.2	0.396	0.747
4	7.2 ± 0.7	7.3 ± 0.7	7.0 ± 0.6	0.583	0.624
6	5.6 ± 2.2	5.8 ± 1.8	5.8 ± 1.8	0.669	0.541
8	6.4 ± 1.0	6.3 ± 0.7	6.4 ± 1.0	0.227	0.152
10	6.5 ± 0.9	6.4 ± 0.7	6.5 ± 1.0	0.141	0.589
12	7.1 ± 0.9	7.3 ± 0.9	6.9 ± 1.0	0.667	0.964
14	7.2 ± 0.9	6.9 ± 1.0	7.3 ± 0.9	0.837	0.523
16	7.2 ± 1.0	7.4 ± 0.8	7.1 ± 0.9	0.339	0.012
18	7.6 ± 1.1	7.8 ± 1.0	7.6 ± 0.9	0.418	0.096
20	7.3 ± 1.0	7.5 ± 0.9	7.0 ± 0.9	0.714	0.232
22	7.7 ± 1.0	8.1 ± 1.0	7.6 ± 1.1	0.842	0.367
24	7.9 ± 1.1	7.8 ± 1.0	7.6 ± 1.0	0.066	0.025

Discussion

The goal of this study was to examine the effect of partial cage undocking or LED flashlight use during daily health checks on breeding performance, nest building scores, and hair corticosterone in C57BL/6J mice. We hypothesized that partial cage undocking and LED flashlight use would result in poorer breeding performance, lower nest building scores, and higher hair corticosterone levels compared with the control. However, we found was no significant differences in any of the outcome parameters when comparing either partial cage undocking or LED flashlight use to the control. These results indicate that partial cage undocking or LED flashlight use during daily health checks does not impact breeding performance or certain measures of wellbeing in C57BL/6J mice.

The act of partially undocking and redocking cages during health checks produced peak noise levels of approximately 84 dB SPL and vibration levels of approximately 260 milli-g. These levels of exposure did not affect breeding performance, nest building scores, or hair corticosterone levels. Most current research that has identified significant effects of increased noise and vibration in rodents has focused on more protracted exposure lasting from 1-h to 24-h,^{23,24,32,33,42,53,59} rather than the 5-s exposures used in our health checks. The possible repercussions of these daily 5-s exposures to increased noise and vibration on various research parameters, from behavioral assays to circadian parameters to neoplastic growth, are unknown, and were beyond the scope of the current study. Noise levels above 100 dB SPL have been documented to cause permanent damage to the cochlea resulting in hearing loss, while intermittent exposures between 85 to 95 dB can “toughen” the auditory system and make it less prone to future damage.^46,64 In the current study, sound exposure was near the range for “toughening,” but we did not determine whether daily partial cage undocking had any long-term effects on the auditory system. An important caveat is the well-known differences in hearing among strains of mice. For example, C57BL/6, DBA/2, and BALB/c mice are susceptible to progressive hearing loss and audiogenic seizures.⁶⁴ Different strains of mice may be affected by noise exposure differently, and the current study only evaluated C57BL/6J mice. Some evidence indicates that mice habituate to the startle response evoked by either repeated noise or vibration exposure.^20,51,52,56 Such habituation may explain why we found no effects on breeding performance, nest building scores, or hair corticosterone levels, given that health checks were performed within the same time window each day and the mice likely habituated to the daily exposure.

The Guide for the Care and Use of Laboratory Animals recommends that light levels in animal housing rooms remain below 325 lux at 1 m above the floor in the center of the room in order to prevent phototoxic retinopathy in albino rodents.²⁹ Other sources recommend maintaining intracage illumination at much lower levels, approximately 40 lux, based on mouse preference testing.⁴³ However, these recommendations apply to chronic light exposure for the duration of the light phase, not for a single exposure lasting 5 s, once a day. In the current study, exposure to a daily flash of cool white LED light for 5 s, at an intensity of approximately 535 lux, did not affect breeding performance, nest building scores, or hair corticosterone levels. However, the study was performed in C57BL/6J mice, and LED flashlight exposure may have caused significant effects in other strains, such as albino or melatonin-producing strains of mice. A growing body of literature shows that exposure to blue-enriched LED light during the light phase can alter nighttime melatonin production and the circadian regulation of metabolism and physiology in certain strains of rats and mice.^9,12,13 However, the rodents in these studies were exposed to 12 h of LED light daily, not to a single, 5-s exposure, so whether a flash of cool white LED light during the light phase would have a significant impact on circadian-related parameters is unknown. In addition, most research mouse strains, including C57BL/6 mice, do not produce melatonin, and therefore circadian effects on physiology and metabolism caused by exposure to blue-enriched LED light are not prominent in these strains.⁹ The present study demonstrates that a flash of cool-white LED light did not affect breeding performance or multiple measures of wellbeing, but care should be taken in interpreting these results because we did not evaluate behavioral or circadian parameters, and melatonin producing strains of mice, such as C3H, may be more sensitive to LED light exposure.

Neither partial cage undocking nor LED flashlight use during health checks had any impact on breeding performance. Breeding performance was evaluated based on the breeding index, latency to first litter born, percent of lost litters, and average pup weight at weaning. Percentage of lost litters was examined because pup mortality is a significant problem in mouse breeding, and C57BL/6 mice have a relatively high litter mortality rate, reported to be approximately 32%.⁶⁵ Compared with this reported mortality rate, the average litter mortality rate in our study was approximately 19% to 21% (see Figure 5 C), and we saw no significant difference in lost litters between either of the experimental groups and the control. This was surprising because breeding mice are particularly susceptible to elevations in noise and vibration exposure.^53,54,63 As discussed previously, this outcome may be due to the short duration (5 s) of undocking and flashlight use, such that mice were able to habituate to this exposure.

Nest building scores were measured for each cage every 2 wk throughout the duration of the 6-mo study, with no statistically significant difference between either experimental group and the control at any time point. Because research mice are highly motivated to build nests when given the proper materials, nest building is commonly used as a sensitive measure of wellbeing in mice.^21,22,27,28 Therefore, the current study indicates that performing daily health checks with an LED flashlight or by partially undocking the cage did not affect mouse wellbeing, as measured by nest scores. While most nest score studies have been performed using 8 g of nesting material, the current study used 4 g of shredded paper nesting material. We used this amount because 4 g of shredded paper is standard for our institution; increasing the amount of nesting material would likely reduce the amount of light that the mice were exposed to from both the room lighting and the flashlight light and would no longer replicate our standard mouse environment. While the comparatively smaller amount of nesting material likely reduced nest quality in our study in comparison to others, this would not be expected to affect the validity of the measure. We still saw a robust variation in nest building scores between cages, with scores varying between 1 and 8.5, so any variation due to experimental changes would likely have been captured. The presence of litters in the breeding cages may also have been a confounding variable for nest scores, particularly when the litters were approaching weaning age, because the older litters may have caused flattening of the nest, thus potentially reducing nest scores. However, because breeding pairs from each group produced approximately the same number of litters over the 6-mo study, the presence of litters was not likely to have affected nest scores disproportionately in any one group compared with the others.

Neither partial cage undocking nor LED flashlight use during daily health checks altered hair corticosterone levels at either 3 mo or 6 mo, for either male or female breeders. Corticosterone concentrations in hair collected at 3 mo into the study reflect circulating corticosterone that was deposited in the hair shaft between 0 to 3 mo, whereas hair collected at 6 mo reflects circulating corticosterone levels between 3 to 6 mo. Because hair corticosterone is used as a measure of chronic or repeated stressors,^26,67 these results indicate that partial cage undocking and LED flashlight use during daily health checks did not increase chronic stress. Both males and females had significantly less hair corticosterone at the 6-mo time point as compared with the 3-mo time point. A possible explanation for this reduction is the acclimation of the mice to their environment over time. Although stress responsiveness and circulating corticosterone levels decrease in aged mice, that reduction occurred only in mice that reached approximately 2 y of age.⁴⁸ The mice in the present study reached 8 mo of age, so an age-related decrease in hair corticosterone is likely not relevant. We examined the effect of high breeding performance on hair corticosterone levels in female mice because high performing performance likely results in higher energy demand. High breeding performance was not related to hair corticosterone levels at the 3-mo time point, but it was associated with higher hair corticosterone levels at 6 mo. Perhaps at the 3-mo time point, the mice had high corticosterone levels due to other factors (for example, acclimation stress), but by the 6-mo time point, background stress had resolved that a slight elevation in corticosterone levels was detectable in the high performing female breeders.

Although cage position on the rack (high, mid, or low) did not affect breeding performance or nest building scores, it did affect hair corticosterone. For both male and female mice, higher cage position on the rack was associated with significantly higher hair corticosterone levels at the 3-mo time point. This finding suggests that some aspect of the elevated cage position on the rack caused a chronic stress response as compared with cages located lower on the rack. Measurements of ambient environmental parameters across various cage positions on the rack revealed more illumination, noise, and vibration on the “high” rows. These differences were likely due to the higher rows being closer to the light source of the room, and closer to the rack motor that is located on top of the rack. This is the first report of elevated corticosterone levels in mice housed in a higher cage position on the rack. However, multiple studies report various ways in which higher cage positions may affect animal welfare and research outcomes, including behavioral measures of stress,¹ incidence of aggression,⁶¹ and incidence and rate of tumor growth.^39,66

To conclude, the use of a cool-white, LED flashlight or partial cage undocking during daily health checks did not affect breeding performance or mouse wellbeing as measured by nest building scores and hair corticosterone in C57BL/6J mice. Due to these results, programs may consider using partial cage undocking or LED flashlights during daily health checks under conditions in which visualization of the cage environment may otherwise be difficult. However, caution should be taken in interpreting the results of this study because the effects of these practices on various research outcomes are currently unknown. This study provides valuable information on the impact, or lack thereof, of short duration daily exposure to increased light, noise, and vibration, and adds to the body of literature on the possible effects of environmental exposures on mouse breeding performance and welfare. This study also highlights the importance of considering cage position on the rack as a confounding variable and a source of stress. Further research is needed to elucidate the cause of this position-related stress response and to identify strategies that will diminish the effects of cage position.