Effects of Cage Change and Nest Transfer on Artificial Intelligence-Measured Behaviors in Male C57BL/6 Mice

Abstract

Cage changes can be a stressful period in a mouse’s life due to the rapid change of environment. It is essential to study the behavioral effects of a cage change on mice, including drinking, eating, aggression, and sleep, as these can affect research variability. This study evaluated the effects of a cage change on behavior and whether transferring nesting material could mitigate the negative effects of a cage change. Using home cage monitoring with computer vision artificial intelligence, 40 C57BL/6 mice were continuously monitored for 2 consecutive cage changes. Mice received either new nesting material (control) or transferred soiled nesting material at cage change. Data were analyzed before and after a cage change and reported for at least 3 days or until baseline behavior resumed. Parameters collected included time spent drinking, eating, sleeping, and aggression. Two hours after a cage change, drinking and eating significantly increased, and sleep significantly decreased compared with pre-cage change values. Long-term effects ranged from decreased sleep in the first 24 hours to decreased eating for 3 days post-cage change. Total distance traveled was increased up to day 4 post-cage change, and rearing time was increased for 3 days after a cage change. All significant behaviors gradually stabilized by day 5. Transfer of old nesting material had no significant impact at any time point. These results show that there are both long-term and short-term effects of a cage change, and nest transfer may not influence these parameters in C57BL/6 mice.

Introduction

Transferring a mouse from a soiled cage to a clean cage, as done during a cage change, inevitably alters the animal’s microenvironment, resulting in acute stress and elevated cortisol levels.¹ Cage change disrupts the pheromone profile and other scent patterns of the cage, resulting in increased anxiety-like behaviors and disruption to the social hierarchy.^2–4 Aggression of cage mates, especially between male mice, is a documented example of the consequences of a cage change, as mice fight to reestablish the lost hierarchy.^2,5–7 Cage changes can increase physiologic parameters in rodents, including corticosterone levels, heart rate, and mean arterial pressure, as well as reproductive outcomes such as pup cannibalism and pup mortality.^5,8–10

Sleep duration is another example of a behavioral change associated with cage changes. Mice often undergo cage changes during the daylight cycle, which coincides with their sleep period, and can lead to sleep disturbances for up to 3 hours in C57BL/6 mice.¹¹ In prior studies using a Digital Ventilated Caging (DVC) recording system that tracked locomotor activity, C57BL/6 mice exhibited elevated activity for several days after a cage change, with levels returning to baseline by day 5¹² or by the second week.¹⁰ These changes in rodent behaviors associated with a cage change have the potential to introduce variability into research studies and negatively influence experimental reproducibility.¹³ As such, there is interest in reducing the impact of cage change through enrichment items or cage modifications.

Nesting material is an enrichment item commonly used in rodents to reduce stress and improve animal welfare. The provision of nesting material is a popular enrichment offering as it provides numerous benefits, including increased activity level, thermoregulation, a location for offspring, species-specific nest building behavior, sleeping, and refuge.^14,15 Mice offered nesting materials have been reported to have increased body weight and consume less food,¹⁶ as the nesting material helps regulate their body temperature and conserve energy balance.¹⁶ Moreover, nesting material is noted to be preferred by mice over other sheltering structures, such as tubes or nest boxes, and it is favored to be present in the cage rather than absent.^15,17–19 This preference is thought to occur because mice manipulate the nesting material and use it to hide and build, giving animals control over their own environment.^15,18 Previous behavioral studies have demonstrated that enriched housing conditions are associated with decreased levels of glucocorticoid stress hormones.¹⁴ In addition, olfactory cues present in rodent nesting material are important for maintaining social hierarchies within a group.^2,7

As the cage change process is associated with increased stress in mice and instability in social hierarchy,² the transfer of nesting material, by moving the formed nest from the soiled cage to the clean cage upon cage change, may provide a buffer against associated stress. Transferring nesting material is a practice that has been empirically cited to reduce some of the impacts of cage change, particularly around anxiety and aggression.⁷ Nest transfer has been found to reduce fighting behavior in BALB/c mice,⁷ and subsequently reduce corticosterone.²⁰ In contrast, the transfer of soiled bedding material, rather than nesting material, was found to increase aggression in male BALB/c mice.²⁰ Nesting material, unlike bedding, is often kept clear of urine and feces and contains hormones from the plantar glands of mice.⁷ These hormones enable recognition of group members, which is associated with reduced aggression.⁷ With prior research focusing on the impact of nest transfer on aggression, there is also potential for other behavioral changes experienced by mice post-cage change that have not been investigated.

In this study, continuous automated home cage monitoring (HCM) with artificial intelligence (AI) based behavioral analysis was used to examine the potential consequences of a cage change and nest transfer in C57BL/6 mice. With automated HCM becoming more popular and available, there is an opportunity to closely observe animals without introducing human-observer interference. This advanced technology allows for experiments to be conducted uninterrupted over an extended period and is more easily reproducible across other institutions by eliminating experimenter bias.^21–23 HCM is especially beneficial for behavioral experiments, as mouse activity can be analyzed without the stress of human handling and the presence of behavioral apparatuses.²⁴ Studying mice in their home environment has been shown to result in more naturalistic observations and reduce novelty-related interferences.²⁵

For our experiment, we compared 2 common methods of cage change. In our control condition, all components of the mouse cage were changed at the same time, and the mice were placed into a new cage with sterile cage components. In our experimental cage change condition, we transferred part of the nesting material from the soiled cage into the clean cage. We hypothesized that animals would fight more, sleep less, and consume less food and water after a cage change, and these behavioral changes would be mitigated by the transfer of nesting material from the former cage to the new cage.

Materials and Methods

Ethics statement.

All procedures were approved by the Massachusetts General Hospital IACUC. All mice were maintained in compliance with the Guide for the Care and Use of Laboratory Animals,²⁶ and this study was conducted at a facility accredited by AAALAC International.

Animals and housing.

Forty male C57BL/6NCrl mice were acquired from Charles River Laboratories (Wilmington, MA) at 5 weeks old. Upon arrival, mice were housed 4 mice per cage and maintained in the same housing groups throughout the duration of the study.

Mice were housed in solid-bottom polysulfone, individually ventilated cages (IVCs) (Jag 75 MicroVent IVC Racks; Allentown, Allentown, NJ). Mice were housed on Sani-Chip bedding (P.J. Murphy, Montville, NJ), given acidified water (pH 2.5-2.8) in water bottles, and fed an irradiated diet (Prolab IsoPro RMH 3000; LabDiet, St. Louis, MO) ad libitum. The housing room was maintained at 19-23 °C with 30%-70% humidity under a 12-hour light/12-hour dark cycle. Each cage had L0110 Biofresh comfort bedding as nesting material (ScottPharma Solutions, Marlborough, MA) present throughout the study. HCM was conducted using Digital Online Monitoring Equipment (DOME), available commercially from Olden Labs (South San Francisco, CA). DOME lids provide monitoring of the home cage, using computer vision and machine learning to automatically measure animal activity continuously (Figure 1). The DOME lids use infrared cameras at a resolution of 640 × 480 pixels in grayscale format, stored at 10 frames per second. The computer vision tracking AI pipeline (MOT - Multi-Organism Tracker) was used to track individual animals in a group setting.²⁷

Figure 1.

IVC Rack Housing for New Nest (Control) and Old Nest (Experimental) Conditions. (A) Standard IVC cage bottoms with DOME lids, Olden Labs food hoppers, and cage card holders. (B) Underside of DOME lid before assembly into the rack. (C) Cage partially slid out of the rack with the filter top and food hopper.

All cages were housed in the same room and on the same IVC rack for the duration of the study. After a 7-day acclimation to our facility, the animal’s standard polysulfone cage lid was replaced with a DOME lid, with an accompanying food hopper, water bottle, and filter top (Olden Labs, South San Francisco, CA). Standard cage bottoms were kept the same, and the animals inhabiting the cages were not touched or manipulated during lid exchange. The cages were placed back into the original IVC rack slot location with the new DOME lids. Preliminary evaluation of mouse behavior after the DOME lid placement showed mice took up to 36 hours to acclimate to the new lids, food hopper, and water. This was noted by the increased activity levels seen on video monitoring (eg, movement, sleeping, climbing, drinking) returning to baseline levels.

Microisolator preparation technique was maintained throughout both rounds of cage changes. This was done by performing the change in a laminar flow hood and spraying the handler’s gloves or any objects with Peroxigard (Virox Technologies, Leominster, MA) to minimize the risk of cross-contamination. The animals were assessed for fight wounds and any other abnormal appearances during this period. All animals were transferred from their soiled cage bottom to a clean cage bottom. The clean cage bottom was covered with 1 cup of autoclaved Sani-Chip bedding along with ½ cup of nesting material. The provision of bedding and nesting material is standard in rodent cages at our institution, with the nesting material serving as the enrichment substrate. Nest transfer was an experimental variable for this study, as detailed in Experimental design.

Cages were maintained in a specific pathogen-free barrier facility. Excluded pathogens included ectromelia virus, epizootic diarrhea of infant mice virus, Hantaan virus, LDH-elevating virus, lymphocytic choriomeningitis virus, minute virus of mice, mouse adenovirus-FL, mouse adenovirus-K87, mouse hepatitis virus, mouse parvovirus, murine cytomegalovirus, pneumonia virus of mice, polyoma virus, reovirus, Sendai virus, Theiler murine encephalomyelitis virus, Bordetella bronchiseptica, Filobacterium rodentium (formerly cilia-associated respiratory bacillus), Citrobacter rodentium, Clostridium piliforme, Corynebacterium kutscheri, Mycoplasma spp., Salmonella enteritidis, Salmonella typhimurium, Streptobacillus moniliformis, Streptococcus pneumoniae, Aspiculuris tetraptera, Eimeria falciformis, Encephalitozoon cuniculi, Eperythrozoon coccoides, Giardia muris, Klossiella muris, Myobia musculi, Myocoptes musculinus, Notoedres musculi, Ornithonyssus spp., Polyplax spp., Psorergates simplex, Radfordia spp., Rodentolepis spp., and Syphacia obvelata.

Experimental design.

Ten cages (4 mice to a cage) of intact, male, 5-week-old C57BL/6 mice were housed in IVCs for the study duration. Two rounds of cage changes were completed 14 days apart for all 10 cages; round 1 cage change was on day 14, and round 2 was on day 28 (Figure 2). All cages of study animals were recorded continuously by the DOME lids from 7 days before the first cage change until day 31. The DOME lids are capable of recording 23 digital measures. A digital measure is defined as quantitative data that are collected continuously from animals that are neither restrained nor instrumented, using digital in vivo technology.²⁸ For the purpose of the study, only 18 digital measures were used to analyze behavior before and after a cage change (Table 1).

Figure 2.

Flowchart of the Experimental Design. A crossover study was performed with both groups receiving either the control (½ cup of new nesting material), or experimental (¼ cup of old nesting material plus ¼ cup of new nesting material) conditions.

Table 1.

List of DOME Digital Measures with Definitions of Each Measure

DOME digital measures		Definition	Unit
1	Aggression events	Count of discrete aggression events (chasing, mounting, fighting, or a combination of the 3) involving ≥1 animals during the selected observation period	Count
2	Aggression time	Percentage of time 2 or more animals are engaged in aggression (chasing, mounting, fighting, or a combination of the 3)	%
3	Average movement acceleration	Mean linear acceleration while the animal is moving (excludes quiescent/sleep periods)	cm/s²
4	Average movement angular velocity	Mean animal angular velocity while moving (angular velocity refers to the rate of change of angular position of a rotating body)29	°/s
5	Average overall speed	Mean linear speed over the selected observation period (bin time), including inactivity (speed = 0)	cm/s
6	Average speed when awake	Mean linear speed over the selected observation period, excluding inactivity (speed = 0)	cm/s
7	Cage center occupancy	Percentage of time spent in the center zone (≥2 cm from all cage walls)	%
8	Cage front half occupancy	Percentage of time spent in the front half of the cage	%
9	Climbing events	Count of discrete climbing bouts on the food-hopper metal grid, defined as both feet off the floor and reaching the upper portion of the hopper	Count
10	Inactivity bouts	Number of discrete inactivity bouts lasting ≥2 min	Count
11	Maximum movement speed	Mean of the top 1% of frame-wise movement speeds during active periods (reduces noise from single-frame spikes)	cm/s
12	Rearing events	Count of rears: defined as animal standing with 2 feet against one of the 3 side walls; excludes rears against the food hopper, red tube, or in the cage center	Count
13	Time spent climbing	Percentage of time spent climbing on the food-hopper metal grid; defined as both feet off the floor and reaching the upper portion of the hopper	%
14	Time spent drinking water	Percentage of time spent interacting with the bottle nozzle or automated water spout	%
15	Time spent eating	Percentage of time spent interacting with food at the hopper; includes “grinding” (chewing without ingestion)	%
16	Time spent inactive	Percentage of time classified as sleep (continuous inactivity ≥2 min)	%
17	Time spent rearing	Percentage of time spent rearing as defined above; excludes rearing against the food hopper, red tube, or in the cage center	%
18	Total distance traveled	Total travel path length over the observation period	cm

Digital measures include the 4 primary measures analyzed in the study, that is, aggression, sleeping (inactivity bouts and time spent inactive), eating, and drinking.

All mice in the study were separated into 2 equal groups of 5 cages and randomly assigned to group 1 or 2. All mice experienced one standard cage change cycle with ½ cup of new nesting material present in their clean cage (control condition) and one cage change cycle with nest transfer (experimental condition). The experimental condition included ¼ cup of new nesting material plus ¼ cup of their old nesting material (½ cup total) transferred over to their clean cage. A crossover study design was used, with half the cages experiencing the control state first, group 1, and the other half experiencing the experimental state first, group 2 (Figure 2). Each group then experienced the alternative condition at the second cage change. The control condition was referred to as new nest, and the experimental condition was referred to as old nest.

The new nest cages involved placing a measured ½ cup of new nesting material into the front left of the clean cage, without transferring any previous nesting materials from the old cage (Figure 3A). The old nest condition received ¼ cup of new nesting material in the back right of the cage and ¼ cup of old nesting material transferred to the front left of the clean cage (Figure 3B). This arrangement was designed to confirm that the mice used both the old and new nesting materials. All the measurements were done via a plastic cup (Figure 4), which was soaked in Peroxigard for at least 1 minute in between each cage change to ensure that no leftover pheromones or contaminants from the mice were transferred into the subsequent cage. Care was taken to ensure that minimal fecal material was transferred from the old nest. Water bottles and feed were replaced at each cage change.

Figure 3.

Clean Cage Setup for (A) New Nest (Control) Condition and (B) Old Nest (Experimental) Condition. Each cage was prepared with Sani-Chip bedding on the floor and new nesting material on the front left of the cage. The experimental condition in (B) had additional old nesting material placed on the back right of the cage.

Figure 4.

Transferred Old Nesting Material in a ¼ Measuring Cup with Minimal Amount of Fecal Material

Standard daily health checks were performed throughout the study; each cage was evaluated for the presence of food and water, the health of the mice, and the overall condition of the cage, while ensuring minimal disruption to the mice.

Statistical analysis.

A power analysis was conducted using the resource equation approach,³⁰ which indicated that the number of animals used reflected the minimum anticipated to achieve sufficient statistical power. Recorded data from the DOME lids were exported from the Olden Labs user portal into an Excel spreadsheet for recordkeeping and analysis (Excel, Microsoft, Redmond, WA).

Data were exported into hourly increments over 2-hour periods immediately after a cage change for the short-term assessments. Pre-cage change data were exported over 2-hour periods for 3 previous days at the same time the cage change occurred. The 3 days of pre-cage change datasets were then averaged together to represent the standard baseline pre-cage change measurements.

For the multiday assessments, data were exported in hourly increments over 24 hours, or 3 days immediately before and after a cage change. Exported data from the user portal automatically summarized the hourly averages of each digital measure per mouse in Excel. The data of these measures were then manually averaged for all 4 mice in each cage. The first 24 hours immediately after the time of cage change were regarded as day 0. Each subsequent day was a 24-hour period following day 0. The experiment was originally designed to monitor animals from the time of cage change to 3 days post-cage change. However, the scope of the study was expanded and analyzed beyond day 3 for group 1 when the behavioral changes seen did not subside at the 3-day observation mark for either group.

All graphs and statistical analyses were performed using Prism 10.6.0 (GraphPad Software, Boston, MA). Each cage was evaluated as the unit of analysis for statistical comparison. Analysis of each of the digital measures (Table 1) was performed using paired t tests to compare behaviors from pre-cage change and post-cage change for both groups. A 2-way mixed ANOVA model and paired t tests were used to compare post-cage change behavior in the new nest (control) and the old nest (experimental) condition for each of the digital measures. The 2 nest conditions were analyzed using the post-cage change data, which were normalized to baseline with the pre-cage change data. Significance was set at a P value of <0.05.

Results

Data.

Two cages (n = 2) had falsely elevated digital measures for time spent drinking water (measure no. 14) during the round 1 post-cage change, as the mice built their nests in front of the water bottle spout. The time spent drinking water digital measure was excluded for that time point for those 2 cages. The mice gradually relocated their nests toward the center of the cage before the second cage change, so the time spent drinking water measure was included for these 2 cages in the round 2 analysis.

Short-term effects of cage changes.

Mice were monitored by automated HCM using DOME lids to determine the effect of cage change on the 18 digital measures (Table 1). Digital data 2 hours before and after a cage change were analyzed for group 1 in round 1, and group 2 in round 2 (Figure 2), both under the new nest (control) conditions.

For group 1, 13 out of the 18 digital measures for behavior were statistically different 2 hours after a cage change compared with the averaged baseline measurements before a cage change (Table 2). The same analysis was repeated using group 2, round 2, for new nest (control) cages (Figure 2). Results from this repeated analysis using group 2 cages were equivalent and repeatable across all measures, except group 2 had 2 additional significant measures that increased after a cage change (Table 2), that is, time spent climbing and climbing events (time spent, P = 0.031; events, P = 0.010). Group 1 did not have significant differences in the climbing measures, but the mean increased 2 hours after a cage change for time spent climbing (0.041) and climbing events (0.517), compared with the mean before a cage change (0.000 for both measures). For groups 1 and 2, respectively, there was an increase in time spent eating (P = 0.038; 0.018), time spent drinking (P = 0.027; 0.018), and reduced sleeping (time spent inactive, P = 0.011, 0.005; inactivity bouts, P = 0.009; P = 0.0007) post-cage change. There was no significant difference in aggression time (P = 0.333; 0.530) and aggression events (P = 0.309; 0.530) for groups 1 and 2 before and after a cage change, respectively.

Table 2.

Comparison of Mean and P Values for Each Digital Measure over 2 h Before and After a Cage Change for Groups 1 and 2 with the New Nest (Control) Condition

Digital measures	Round 1: group 1 (P value)	Round 2: group 2 (P value)	Rounds 1 and 2: groups 1 and 2 combined (P value)
Aggression events	0.309	0.530	0.130
Aggression time	0.333	0.530	0.169
Average movement acceleration	0.0008	0.013	0.00002
Average movement angular velocity	0.0005	0.002	<0.000001
Average overall speed	0.0008	0.004	0.000001
Average speed when awake	0.0009	0.004	0.000002
Cage center occupancy	0.830	0.177	0.169
Cage front half occupancy	0.001	0.274	0.004
Climbing events	0.180	0.010	0.019
Inactivity bouts	0.009	0.0007	<0.000001
Maximum movement speed	0.003	0.006	0.000001
Rearing events	0.007	0.005	0.000002
Time spent climbing	0.210	0.031	0.046
Time spent drinking water	0.027	0.018	0.00005
Time spent eating	0.038	0.018	0.001
Time spent inactive	0.011	0.005	0.000003
Time spent rearing	0.011	0.010	0.00001
Total distance traveled	0.0007	0.004	0.000001

Groups 1 and 2, both individually and combined, had a significant decrease in inactivity bouts and time spent inactive after a cage change compared with before a cage change. All the other significant digital measures had increased post-cage change compared with baseline pre-cage change. P ≤ 0.05 was considered significant.

When combining all the new nest conditions for groups 1 and 2 in both rounds, 15 out of the 18 digital measures were significantly different pre- and post-cage change. Over 2 hours, time spent eating (P = 0.001) and time spent drinking (P < 0.00005) were significantly higher following a cage change, and sleep decreased significantly (time spent inactive, P < 0.000003; inactive bouts, P < 0.000001). The total distance traveled (P < 0.000001) significantly increased over 2 hours after a cage change, with an average of 668.8 cm/h before cage change compared with 9,167 cm/h after cage change. Average aggression time (P = 0.169) and number of aggression events (P = 0.130) were not significantly different during this period. However, the relative change in average aggression events (0.06986/0.004722) was >14-fold greater post-cage change compared with baseline. The relative average aggression time also increased by a factor of 30 (0.1926/0.0064) post-cage change compared with pre-cage change.

Long-term effects of cage changes.

Data were collected from group 1 in round 1, and group 2 in round 2, from Figure 2. For the round 1 new nest (control) condition, cage change was evaluated from day −3 pre-cage change to day 5 post-cage change, where day 0 represented the day of cage change (Figure 5). Baseline measurements before a cage change were the averages of days −1, −2, and −3 combined.

Figure 5.

Effects of a Cage Change on Individual Digital Measures (Table 1) from Day −3 to Day 5 Post-Cage Change Across All Cages in Group 1, Round 1, over the Specified Point in Time (Day). Day 0 is the day of cage change. Data represent means with SD of summed measures per day. (A) Number of aggression events per day. (B) Percentage of time exhibiting aggressive behavior per day. (C) Movement acceleration per day. (D) Movement angular velocity per day. (E) Overall speed per day. (F) Speed during awake time per day. There was a significant increase in average movement angular velocity and average movement acceleration on post-cage change days 0, 1, and 2 compared with pre-cage change baseline data (days −1, −2, and −3). Average overall speed and average speed when awake were significantly increased on post-cage change days 0, 1, 2, 3, and 4 compared with pre-cage change baseline data (days −1, −2, and −3). *P ≤ 0.05.

For group 1, the average time spent rearing (P = 0.019) and rearing events (P = 0.001) were significantly higher on days 0, 1, and 2 post-cage change compared with baseline. Average speed when awake (P = 0.004), average speed overall (P = 0.0011), and total distance traveled (P = 0.0005) were significantly higher on days 0, 1, 2, 3, and 4 compared with baseline and became statistically insignificant on day 5 (P = 0.861, 0.878, and 0.836, respectively). The mean distance traveled before a cage change was 1,977 cm/h, compared with an increase on days 0, 1, and 2 post-cage change at 3,083 cm/h.

For group 1, the mean time spent drinking per hour was lower overall for the 5-day observed post-cage change compared with pre-cage change (1.861), with the lowest mean on day 1 (1.237). However, this change in time spent drinking was not statistically significant (Figure 5N, day 1, P = 0.851). Average aggression for group 1 (time spent, P = 0.996; events, P = 0.996) and time spent eating (P = 0.980) were not significant for any of the days post-cage change compared with baseline.

Figure 5 (cont.) panels M–R.

Effects of a cage change on individual digital measures (Table 1) from day −3 to day 5 post-cage change across all cages in Group 1, Round 1, over the specified point in time (day). Day 0 is the day of cage change. Data represents means with SD of summed measures per day. (M) Maximum movement speed per day. (N) Percentage of time spent drinking water per day. (O) Number of rearing events per day. (P) Percentage of time spent rearing per day. (Q) Percentage of time spent eating per day. (R) Total distance traveled in centimeters per day. There was a significant increase in time spent rearing and rearing events post-cage change at day 0, 1, and 2 compared with pre-cage change baseline data (days −1, −2, and −3). Total distance traveled was significant post-cage change at day 0, 1, 2, 3, and 4 compared with pre-cage change baseline data (days −1, −2, and −3). *, P ≤ 0.05.

Climbing measures were not significantly different before and after a cage change for group 1 (Figure 5I and J, time spent, P = 0.223; events, P = 0.135). However, the mean time spent climbing and climbing events were highest on days 0, 1, and 2 post-cage change, with peak averages on day 0 (time spent, 0.024; events, 0.275) compared with baseline (time spent, 0.002; events, 0.023). On day 0, the relative change for mean climbing events (0.252/0.023) was 11-fold higher after a cage change, and time spent climbing (0.0217/0.002) was 11-fold higher post-cage change compared with the pre-cage change baseline.

Figure 5 (cont.) panels G–L.

Effects of a cage change on individual digital measures (Table 1) from day −3 to day 5 post-cage change across all cages in Group 1, Round 1, over the specified point in time (day). Day 0 is the day of cage change. Data represents means with SD of summed measures per day. (G) Percentage of time spent in the center of the cage per day. (H) Percentage of time spent in the front half of the cage per day. (I) Number of climbing events per day. (J) Percentage of time spent climbing per day. (K) Number of inactivity bouts per day. (L) Percentage of time spent inactive per day. There was a significant decrease in sleep (Inactivity bouts and time spent inactive) on post-cage day 0 compared with pre-cage change baseline data (days −1, −2, and −3). *, P ≤ 0.05.

Sleep measures for group 1 were significant on day 0 (time spent inactive, P = 0.052; inactivity bouts, 0.046) but became insignificant on day 1 (P = 0.825; P = 0.825) through day 5. Mean amount of time spent inactive and mean number of inactivity bouts were lower for 5 days post-cage change compared with pre-cage change (time spent, 11.70; bouts, 0.372), with the lowest averages on day 0 (time spent, 7.957, bouts, 0.246), but neither of the digital measures was statistically significant after day 0 (Figure 5K and L).

The same analysis was repeated using group 2, round 2, new nest (control) cages (Figure 2) from day −3 to day 2, with 0 being the day of cage change and baseline pre-cage change being days −1, −2, and −3. When comparing groups 1 and 2, the P values with significant repeatable measures were higher average speed when awake (P = 0.029), average overall speed (P = 0.021), and total distance traveled (P = 0.021). Group 2 also had a lower time spent drinking mean on days 0, 1, and 2 (0.942) compared with baseline on days −1, −2, and −3 (1.149), but this was not statistically significant (P = 0.200), which was a similar trend seen with group 1. The mean total distance traveled for group 2 was comparable to group 1, at 1,870 cm/h before a cage change, and increased to 2,432 cm/h on days 0, 1, and 2 post-cage change. In addition, the resulting P value for group 2 on days 0, 1, and 2 was not significant for aggression (time spent, P = 0.683; events, P = 0.464) before and after a cage change, also seen with group 1.

Group 2 did not show a significant increase in rearing over days 0, 1, and 2 post-cage change, as demonstrated by group 1, but there was still a significant increase on day 0 (time spent, P = 0.016; events, P = 0.006). As identified with group 1, the mean for group 2 was higher on post-cage change days 0, 1, and 2 (time spent, 0.228; events, 0.838) compared with baseline (time spent, 0.1606; events, 0.518).

In group 2, there was a significant difference in time spent eating that was not evident in group 1. Eating was significantly less on days 0, 1, and 2 (P = 0.019) compared with baseline, with the mean before a cage change (9.062) higher than after a cage change (6.887).

The number of climbing events in group 2 was significant on day 0, with increases detected post-cage change (P = 0.018) compared with pre-cage change, and became insignificant on the subsequent days (0.200). This was different from group 1, where there was no significance on any of the days post-cage change. For group 2, in contrast to the significant number of climbing events on day 0, time spent climbing on day 0 was not significant (P = 0.092). The relative change for time spent climbing (0.005/0.001) was 5-fold higher post-cage change than pre-cage change. Group 2 had similar averages for climbing on days 0, 1, and 2 as group 1, with a higher mean post-cage change (time spent, 0.006; events, 0.054) than pre-cage change (time spent, 0.001; events, 0.011).

Sleep measures were not significant for group 2 on days 0, 1, and 2 (time spent inactive, P = 0.260; inactivity bouts, P = 0.220); however, similar to group 1, the mean was lower post-cage change (time spent inactive, 12.55; inactivity bouts, 0.386) compared with the pre-cage change mean (time spent inactive, 14.44; inactivity bouts: 0.431).

Short-term effects of nest transfer.

The old nest conditions from both groups were combined for one analysis. Similarly, the new nest conditions from both groups were combined for analysis. Cages were analyzed for differences between the 2 conditions, old and new nest, and the effects on behavior in response to a cage change (Figure 6A and B). Analysis revealed no significant difference in 14 out of the 16 digital measures (all measures except no. 10 and no. 16 from Table 1) between old and new nest 2 hours after a cage change (P > 0.05). Time spent inactive and inactivity bouts (conditions no. 10 and no. 16 from Table 1) for sleep conditions were unable to be analyzed because the raw values were 0 for both measures. Digital measures that were not statistically significant included time spent eating (P = 0.982), time spent drinking (P = 0.945), aggression events (P = 0.978), and aggression time (P = 0.978).

Figure 6.

Effects of Old and New Nesting Material on Behavior 2 h After a Cage Change (Normalized). Data represent the mean digital behavioral measure scores with SD for old nest (experimental) condition (Table 1) and new nest (control) condition (Table 1). P < 0.05. Data were analyzed using a (A) 2-way mixed measures ANOVA with post-cage change data normalized to baseline pre-cage change data and (B) individualized paired t tests. None of the digital measures was statistically significant between old and new nest conditions.

When group 1 mice were compared with themselves, there was no statistical difference between old and new nest conditions 2 hours after a cage change, including time spent eating (P = 0.995), time spent drinking (P = 0.998), aggression time (P = 0.806), and aggression events (P = 0.806). Inactivity time and time spent inactive for sleep analysis could not be assessed due to all the data being 0s for both groups.

Assessments for group 2 for old nest and new nest conditions were analyzed in the same way as in group 1, with round 1 having the old nest condition and round 2 having the new nest condition. When group 2 was compared with themselves, there was no significant difference between old nest and new nest conditions 2 hours after a cage change, except for cage front half occupancy (P = 0.013), which was higher for the old nest condition. Time spent eating (P = 0.999) and time spent drinking (P = 0.118) were insignificant. Individual t tests were performed for the digital measures with values of 0s in the raw data, and aggression time (P = 0.846), aggression events (P = 0.846), climbing events (P = 0.352), climbing time (P = 0.515), time spent inactive (P = 0.846), and inactivity bouts (P = 0.846) were not significant.

Long-term effects of nest transfer.

The effects of old and new nesting materials on various behaviors were evaluated over 3 days or until baseline, both before and after a cage change (Figure 7). The data for each cage were normalized by dividing the post-cage change data by the pre-cage change data. There was no significance in any of the digital measures between the 2 nesting conditions, including aggression time (P = 0.977), aggression events (P = 0.807), time spent eating (P = 0.977), time spent drinking (P = 0.798), inactivity bouts (P = 0.977), and time spent inactive (P = 0.973).

Figure 7.

Effects of Old and New Nesting Material on Behavior 3 d Before and After a Cage Change (Normalized). Digital measures were recorded in the home cage. Data represent means with SD of summed measures across 2 nesting conditions, with 2 wk between each condition and cage change. Each cage was normalized by post-cage change data divided by the pre-cage change data. Data were analyzed using a 2-way mixed measures ANOVA. n = 10 cages of 4 mice per cage for each nesting condition. P < 0.05. None of the digital measures was statistically significant between old and new nest conditions.

For group 1, the effects of old nest compared with new nest conditions on the digital measures were not significant over the 3 days following a cage change. The effects of nesting material on the digital measures for group 2 were also not significant when comparing old and new nest conditions over 3 days post-cage change.

Discussion

HCM is becoming increasingly popular due to the benefits of reproducibility across institutions and decreased human bias. HCM has been used in previous studies for behavior analysis and cage changes. In an earlier report using DVC systems, there were significant differences found in the circadian activity of 3 different strains, C57BL/6NCrl, BALB/cAnNCrl, and CRL:CD1.³¹ The same authors performed a subsequent study with DVC systems, analyzing the home-cage spatial preference and circadian home-cage behavioral phenotypes through 3 consecutive intervals of weekly cage changes. This study found that C57BL/6NCrl mice have more pronounced thigmotaxis within the first hour post-cage change in early adolescence as compared with in adulthood.²⁴ Prior studies have also used DVC systems to assess in-cage activity post-cage change, and the correlation between ammonia levels and animal activity during 5 consecutive 2-week cage change cycles in C57BL/6 mice.¹⁰ Another study evaluated the effect of enrichment on aggression and locomotor activity over 6 weeks using DVC systems.⁵

Advancements to automated HCM through the development of computer vision AI technology have led to capturing a wider range of mouse behaviors measured in a home cage. These digital measures include continuously tracking the food and water intake, aggression, sleep frequency, and multiple other species-specific behaviors of individual mice. To our knowledge, the effects of a cage change and nest transfer in C57BL/6 mice monitored under continuous automated AI video cages have not yet been reported.

We hypothesized that animals would fight more, sleep less, and consume less food and water after a cage change. Our results indicated that 2 hours after a cage change, C57BL/6 mice slept less but consumed more food and water. In addition, the percentage of time spent on aggressive behavior and the number of aggression events were not significant post-cage change compared with pre-cage change. The mean aggression time and events were higher 2 hours after a cage change compared with before, but because the number of aggression events was so low overall, this result was inconclusive. We suspect the low incidence of aggression was due to C57BL/6 being a less aggressive strain than other more aggressive strains such as C3H, CBA, DBA, FVB, and CD1 mice.²

The 2 digital measures, that is, time spent inactive and inactivity bouts, were used to quantify sleep. Not only did mice sleep less in the 2 hours immediately post-cage change, but in group 1, they continued to sleep less significantly over 24 hours post-cage change. Both groups also had a lower daily mean sleep measurement for up to 5 days after a change compared with their baseline before the cage change. The low mean results correlated with cage changes directly modulating the sleep behaviors of mice, due to the transition from a sleep state to an awakened state during a cage change. Our findings align with past publications noting reduced sleep after a cage change in C57BL/6 mice,^10–12,32 and increased activity by 50% during the day when animals typically rest, gradually decreasing until the following cage change.¹⁰

A likely contributing factor to the loss of sleep was the provision of new nesting material. When a cage is replaced with fresh new nesting material during a cage change, C57BL/6 mice will demonstrate nest-building behavior during the light and late dark phases, which is the period mice will typically sleep.^33–35 Two hours after a cage change, not only was sleep impacted, but drinking and eating increased for both groups, contrary to the original hypothesis that they would decrease. We speculate that the observed correlation between increased food and water consumption immediately after a cage change may be due to the animals being awake during their regular sleep cycle hours from the disturbance of a cage change. Instead of sleeping, they were engaging in other activities at this time, including eating and drinking.

The short-term results 2 hours after a cage change for group 1 and group 2 new nest conditions were consistent across all digital measures, except for cage front half occupancy, cage center occupancy, and climbing. Group 2, assessed alone, did not demonstrate significance for cage center and front half occupancy similar to group 1, but it did show significance for time spent climbing and climbing events that group 1 did not exhibit. Out of the 5 cages in group 2, 2 cages had their nests situated in the back of the cage, which may be why we did not see a significant difference in these 2 parameters when statistical analysis was repeated only on group 2. In addition, the cage location for group 2 was in the lower half of the rack, while group 1 cages were in the top half. The difference in cage location on the IVC rack may have impacted how much time the mice spent at the front of the cage and the placement of their nest location. Different locations of the cage create slight differences in light levels and proximity to other equipment in the room (such as the laminar flow hood), which could impact how much time mice spend at the front of the cage.

The other distinction between the 2 groups was climbing, with group 1 not showing significant differences in time spent climbing or the number of climbing events 2 hours after a cage change, as observed in group 2. For group 2, the time spent climbing continued to be significantly increased up to 24 hours after a cage change. Despite not being significant, group 1 did have an increased mean for time spent climbing and climbing events post-cage change compared with pre-cage change. The group 1 climbing measures were likely not significant due to the increased variance between the cages, as one cage was an outlier and mice did not climb as much as the other cages in that group. With the small n, this one cage may have affected the analysis for the whole group. However, when both groups were combined for analysis for the climbing measures, thus increasing the value of n, both the time spent climbing and the climbing event measures were significant.

When groups 1 and 2 with the new nest conditions were combined for analysis, 15 out of the 18 digital measures were significantly altered after a cage change. The only insignificant measures were aggression time, number of aggression events, and cage center occupancy. Most likely, the lack of aggression events and the 2 cages of mice that placed their nesting material in the back portion of the cage explain why changes in these measures were insignificant. The total number of significant differences seen in the digital measures for both groups put together highlights the wide scope of behavioral changes that occur in mice after a cage change.

The long-term effects of a cage change were also explored. Results for time spent on aggression and the number of aggression events did not indicate a correlation between aggression and cage changes. These results are suspected to be due to the same reasoning for the 2-hour window assessments; that is, there were not enough aggression events to accurately assess the validity of the findings. Food consumption significantly decreased over 3 days for group 2. Eating for group 1 did not significantly change for the 3-day assessments. Although it was not significant, the mean time spent drinking water after a cage change decreased until day 5 for group 1 and over 3 days for group 2. The differences in eating and drinking could be due to the stressful impacts of a cage change. This fluctuation in food and water intake can be particularly valuable when scheduling experimental work or knowing how a routine cage change might impact consumption of a water- or diet-based treatment.

Other significant digital measures having long-term effects after a cage change were average speed when awake, total distance traveled, and average overall speed. These digital measures were significantly increased on days 0, 1, and 2 for both groups, and in addition up to day 3 and 4 in group 1, stabilizing to baseline on day 5. This corresponds with prior studies that showed increased locomotor activity over multiple days after a cage change.^10,12 Male C57BL/6 mice have been shown to have a greater response than females to increased activity during the lights-on period the day after a cage change, followed by a decreased response over multiple days, and returning to baseline by day 5,¹² which is in line with our findings in our male C57BL/6 mice.

Time spent rearing and rearing events increased in the first 2 hours post-cage change, and increases were still significant on days 0, 1, and 2 after change for group 1, and on day 0 for group 2. Our findings may suggest that rearing in the initial days after a cage change could have been one of the possible locomotor behaviors discussed in previous publications.^10,12 The results of the rearing digital measures may be important regarding animal surgeries, as increased rearing in mice over a period of days might increase susceptibility to postoperative incision complications, particularly with ventral abdominal incisions.

The total distance traveled had increased by an average of at least 1,000 cm per hour per cage in both groups over 3 days after a cage change. There was a noticeably higher distance traveled during the daylight hours post-cage change compared with the pre-cage change baseline. We suspect this change was due to the animals moving around more frequently in the cage to build their new nest and establish their scent and territory over the days following a cage change. The total distance traveled, being significantly greater up to day 4 post-cage change, may need to be accounted for in metabolic studies, as the calorie expenditure could be increased. The increase in total distance traveled up to 4 days could be attributed to C57BL/6 mice being known for their exploratory behavior and activity, as they demonstrate significant exploratory behavior in open fields and increased total distance traveled compared with BALB/c, CBA/J, 129SvEv, and CD1 strains.³⁶

In another vein, an additional aim of this study was to investigate the potential effects of transferring soiled nesting material into a clean cage during cage change. In this study, we elected to use Biofresh comfort bedding, a commonly used nesting material. The transfer of nesting material has only been directly studied in BALB/c mice to date.⁷ The C57BL/6 strain was selected for this study to increase the chances of reproducibility across institutions regarding cage changes and nest transfer, since C57BL/6 mice are the most widely used strain in laboratory animal research.³⁷ As discussed previously, the provision of nesting material is a popular enrichment offering as it provides numerous benefits to reduce stress and increase welfare.

With nesting material being being widely regarded as essential for the wellbeing of mice, we sought to examine the benefits of nest transfer. We hypothesized that transferring nesting material from the former cage to the new cage would mitigate the behavioral changes in mice caused by a cage change. Our results rejected this hypothesis, indicating that there was no significant difference in behavior after a cage change when old nesting material was transferred into the clean cage. Old nest transfer is presumed to add familiar scent pheromones to the new cage, which could decrease the associated stress for the mice when exposed to all clean cage components.⁷ If cortisol levels or other physiologic stress parameters had been examined, it is possible that nest transfer would be associated with significant physiologic impact, even though mouse behavior was not significantly different. The insignificant changes observed for aggression may be due to C57BL/6 mice being less aggressive than BALB/c mice, the strain previously analyzed for nest transfer.⁷ To better understand the correlation between nest transfer and aggression, it would be beneficial to use HCM to investigate strains that are more commonly known for aggression.

We suspect further rounds of cage changes are warranted to more accurately assess the effects of nest transfer, as the age of the mice could also be a contributing factor. The mice were 5 weeks old at the start of the study, and the last round ended with mice at 9 weeks of age. In a previous HCM study, C57BL/6 mice had a high behavioral variability from prepuberty until early adulthood.²⁴ As our study used juvenile mice,³⁸ monitoring mice at the adult stage using HCM is also an area for future study. In addition, while this study only involved male mice, which are generally more aggressive than females,² evaluation of female behaviors would be valuable. Another aspect that was not explored was how nest transfer affects breeding, which is essential for many colonies.

Another limitation of this study was that the mice studied were naive and of adequate health. Future studies examining postoperative surgical models or mice with transgenic backgrounds that may exhibit altered behaviors would be beneficial to assess more experimental conditions of mice typically housed in research institutions. Lastly, the cage setup for the old nest conditions may have contributed to the unnoticeable differences observed between the 2 conditions, as the old nest cages still received ¼ cup of new nesting material in the cage to enable the mice to display nest-building behaviors. If solely old nesting material had been transferred during a cage change, more marked changes might have been observed.

One downside to transferring old nesting material is that it introduces an additional step to cage change, which can add time and resources to the overall husbandry of research mice. In addition, transferring old nesting material could make it more challenging to eliminate pathogens. Given the drawbacks of old nest transfer and its minimal impact on behavior after cage changes in C57BL/6 mice, we recommend further evaluation of its applicability as a standard practice across various mouse strains maintained in IVCs. The transfer of nesting material adds an average of 30 seconds to each cage change, multiplied by 500 cage changes per staff member every 2 weeks; this is estimated to add ∼2.1 hours of work per husbandry staff member per week or 109.2 hours per staff member per year based on our institution’s standard work.

In summary, our findings demonstrate that the standard husbandry practice of a cage change can impact a wide variety of behaviors in mice both in the immediate hours and days following a cage change. Our data show increased eating and drinking, and decreased sleep in the initial hours after a cage change. Longer-term impacts include an increase in total distance traveled, rearing, and a decrease in sleep and eating after a cage change. While nest transfer is beneficial for decreasing aggression in strains such as BALB/c mice, our findings support the conclusion that nest transfer does not have an impact on C57BL/6 mice on the 18 digital measures of behavior we evaluated. Further research is needed to investigate the impact of nest transfer on aggression in other mouse strains. A better understanding of the behavioral impacts of a cage change and nest transfer can help maximize efficiency in husbandry practices and inform researchers about the potential impacts of cage change on their experiments.

Conflict of Interest

The authors have no conflicts of interest to declare.

Funding

This work was internally funded.