Reinventing the wheel: comparison of two wheel cage styles for assessing mouse voluntary running activity

Abstract

Voluntary wheel cage assessment of mouse activity is commonly employed in exercise and behavioral research. Currently, no standardization for wheel cages exists resulting in an inability to compare results among data from different laboratories. The purpose of this study was to determine whether the distance run or average speed data differ depending on the use of two commonly used commercially available wheel cage systems. Two different wheel cages with structurally similar but functionally different wheels (electromechanical switch vs. magnetic switch) were compared side-by-side to measure wheel running data differences. Other variables, including enrichment and cage location, were also tested to assess potential impacts on the running wheel data. We found that cages with the electromechanical switch had greater inherent wheel resistance and consistently led to greater running distance per day and higher average running speed. Mice rapidly, within 1–2 days, adapted their running behavior to the type of experimental switch used, suggesting these running differences are more behavioral than due to intrinsic musculoskeletal, cardiovascular, or metabolic limits. The presence of enrichment or location of the cage had no detectable impact on voluntary wheel running. These results demonstrate that mice run differing amounts depending on the type of cage and switch mechanism used and thus investigators need to report wheel cage type/wheel resistance and use caution when interpreting distance/speed run across studies.

NEW & NOTEWORTHY The results of this study highlight that mice will run different distances per day and average speed based on the inherent resistance present in the switch mechanism used to record data. Rapid changes in running behavior for the same mouse in the different cages demonstrate that a strong behavioral factor contributes to classic exercise outcomes in mice. Caution needs to be taken when interpreting mouse voluntary wheel running activity to include potential behavioral input and physiological parameters.

INTRODUCTION

Collecting and analyzing wheel running data in animal models are the foundation of many research projects. Animal running activity is used to assess physiological function but is also used to track behavioral changes that yield vital information on drug treatment efficacy, altered genetics, and disease states. One of the most common tools used for this data collection is the voluntary exercise wheel. It is important that the exercise wheel data be robust and reliable in describing rodent behavior to accurately assess experimental outcomes. This data are sometimes confounded by factors including, but not limited to, genetic and environmental factors (3, 7–9, 16, 18).

Wheel cage running data disparities, including significant differences in distance run and time run between data collection sources, are present throughout the literature (2, 12, 14, 16, 19–21). A potential explanation for these discrepancies is the simple fact that all wheel cages are not the same. Even when all other factors are controlled, such as food, water, lighting, temperature, ambient noise, age and gender of subjects, genetics, etc., the one variable often overlooked is the wheel cage itself. While many publications describe the use of the wheel cages, only some of them describe the actual physical attributes of the wheel, the cage, or both. There are many manufacturers of wheel cages and activity wheel monitoring systems, and numerous researchers have custom built wheel data collection systems from readily available, inexpensive materials. Because there is so much variety, there is no “standard” for reference. With so much variation in the primary means of collecting voluntary activity data, it is not surprising that the data vary from study-to-study.

This study was designed to determine if the wheel cage switch type (i.e., counting mechanism) contributes to variations in measured mouse wheel cage activity data. The following variables were examined: wheel cage switch type, cage placement, and cage enrichment to best discern factors that might influence wheel cage running data. Results from this investigation demonstrate that the type of wheel cage counting switch (magnetic vs. electromechanical) was a significant contributor to voluntary wheel cage activity and that measurable differences in running behavior likely result from the resistance inherent in the different switch types. These data establish that mice can exhibit significant differences in running distance and average running speed depending on the type of switch used. Study design should integrate the type of wheel cage with the specific animal model to appropriately address the experimental questions, as we have found that resistance within the wheel can significantly alter running outcome measures. These findings highlight the importance of accurately reporting wheel cage attributes to appropriately assess and interpret mouse wheel cage data between laboratories.

MATERIALS AND METHODS

Animals.

All animal procedures were conducted in compliance with the guidelines of the Association for Assessment and Accreditation of Laboratory Animal and were approved by the Institutional Animal Care and Use Committee at University of Kentucky. Age-matched, male, C57BL/6 mice were randomized into two groups (n = 6 mice in each group). Mice were 8 wk of age at the beginning of the study and were in wheel cages for a total of 12 wk. Mice were acclimated to the wheel cages for 2 wk before the beginning of the study. The following 2 wk assessed cage location (light box or open shelf). This was followed by 3 wk with either the electromechanical switch or the magnetic switch. The mice in cages with the electromechanical switch were then placed in cages with the magnetic switch, and magnetic switch cage mice were placed in cages with the electromechanical switch for the following 3 wk. Cage size was assessed for 1 wk with all mice being placed in the two cage sizes with the electromechanical switch. Finally, enrichment was assessed for 1 wk. Both groups of mice had ad libitum access to food (Harlan Teklad 2018) and water. Cages contained equal amounts of hardwood chip bedding and were monitored daily. Lighting conditions were 12:12-h light-dark, and room temperature was constant.

Wheel running cages.

Two types of voluntary wheel cages for mice were used for comparison in this study. The first cage (the electromechanical switch) was originally purchased from Phenome Technologies and is currently available from Actimetrics (model: PT2-MCR1). It consists of a Tecniplast model 1144B cage bottom (33.2 × 15 × 13cm) and wire bar lid. The wheel is stainless steel, 11-cm inside diameter, 5.4-cm wide, with 1.2-mm wide bars placed 7.5 mm apart. The wheel is mounted in the cage by a shaft running through its center that rests in holes drilled in each side of the cage bottom. A modified wing nut mounted on the end of the shaft serves as the counting trigger as each rotation causes it to strike a switch mounted on the outside of the cage. The switch is wired to a computer using the laboratory ClockLab software (Actimetrics) to record total number of revolutions/counts (Fig. 1A).

Fig. 1.

A: shown is the electromechanical switch wheel cage with interface cables. This cage is smaller in dimension. B: shown is the magnetic switch wheel cage with interface cables. This cage has more available “living space.”

The second cage (the magnetic switch cage) was a Tecniplast 1284 (36.5 × 20.7 × 14 cm) with a wire bar lid. The wheel is a Philips Respironics Mini-mitter, 11-cm inside diameter, 5.3-cm wide, with 1-mm bars placed 10 mm apart. The wheel is suspended from the wire bar lid, supported by an external bracket with a shaft running through the center of the wheel and into each side of the bracket. A small magnet mounted on the side of the wheel triggers a reed switch mounted on the bracket of the wheel with the data cable interfaced to a computer using the laboratory ClockLab software to record total number of counts (Fig. 1B).

Wheel cage resistance was measured similar to protocols outlined previously (1, 22). Briefly, calibration of wheel resistance was performed by placing a mass on the horizontal moment arm of the wheel to determine the wheel load in grams to initiate spin. The resistance inherent in the electromechanical switch was 5.54 ± 0.15 g, which was significantly greater than that measured for the magnetic switch, 1.95 ± 0.05 g (P < 0.0001).

Statistics.

A two-way repeated measures ANOVA was employed to compare daily wheel cage activity patterns. A two tailed t-test was utilized to compare groups and a two tailed, paired t-test was used to compare the same mice in the two different cages. This analysis was performed with GraphPad Prism software.

RESULTS

To determine if open or closed environmental housing had any impact on voluntary wheel cage activity, mice were placed in magnetic and electromechanical switch wheel cages and housed on open shelving units within the mouse room with standard fluorescent lighting, or in our custom light boxes with green LED lighting. The light boxes were maintained on the exact same light cycle as the mouse room. Cages were disturbed as little as possible, with cage changes and weighing of food, water bottles, and mice occurring during the room change, the same day each week. The mice were maintained with continuous recording of wheel activity data and weekly recording of body weight, food, and water consumption throughout the experimental protocol. Position on the rack or in the light box did not influence the running data collected from either the electromechanical or magnetic wheel cages as they were not significantly different from counterparts housed in different locations (Fig. 2 A, B, and C). This was confirmed when mice from the light box were placed on the open shelves with the other cages. The running data remained stable after the move.

Fig. 2.

A: average run time (h/day) was not significantly different for electromechanical switch wheel cages or magnetic switch wheel cages in open or closed housing (n = 6). B: the average distance run (km/day) was not significantly different for electromechanical switch wheel cages or magnetic switch wheel cages in open or closed housing (n = 6). C: the average speed (km/h) was not significantly different for electromechanical switch wheel cages or magnetic switch wheel cages in open or closed housing (n = 6). A two-tailed t-test was performed to compare the effect of open and closed housing in each of the cages on run time, average speed, and average distance.

Significant differences in both run distance per day and average run speed were demonstrated between the cages containing the electromechanical switch and the magnetic switch. Although total run time was not different between the electromechanical and magnetic switch cages (Fig. 3A), the mice in the mechanical switch wheel cages ran at a faster speed (1.0 ± 0.06 km/h electromechanical switch; 0.6 ± 0.08 km/h magnetic switch; P = 0.0018; Fig. 3B), for a longer distance (9.9 ± 0.98 km/day electromechanical switch; 5.5 ± 0.72 km/h magnetic switch; P = 0.0045; Fig. 3, C and D). To further delineate the cause of running differences between the mice in the electromechanical and magnetic switch wheel cages, mice in electromechanical switch wheel cages were switched to magnetic switch wheel cages and vice versa. Cage positions on the racks were maintained, and no other variables were changed. Mice placed in the magnetic switch wheel cages (5.8 ± 0.15 km/day) ran less than when in electromechanical switch wheel cages (9.8 ± 0.28 km/day; P < 0.0001), and mice placed in electromechanical switch wheel cages (9.7 ± 0.27 km/day) ran more than they had previously (5.5 ± 0.15 km/day; P < 0.0001; Fig. 4, A, B, and C).

Fig. 3.

A: the average run time (h/day) was not significantly different between electromechanical switch wheel cages or magnetic switch wheel cages (n = 6). B: the average speed (km/h) was significantly reduced in the cages with the magnetic switch (n = 6). C: average daily run distance (km/day) is shown for 7 days. Mice in the electromechanical switch wheel cages ran significantly more than those in the magnetic switch wheel cages (n = 6). D: a summary graph of average daily run distance (km/day) is shown (n = 6). A two-tailed t-test was performed to compare the effect of electromechanical and magnetic switch wheel cages on run time, average speed, and average distance (*P < 0.05; ** P < 0.01). A two-way ANOVA was employed to compare the average daily distance between the 2 cage types (*P < 0.05; **P < 0.01; ***P < 0.001).

Fig. 4.

A: a representative double-plotted actogram demonstrates raw activity counts for a mouse placed in an electromechanical switch wheel cage and then moved to a magnetic switch wheel cage. The arrow marks the day of cage swap. B: a representative double-plotted actogram demonstrates raw activity counts for a mouse placed in a magnetic switch wheel cage and then moved to an electromechanical switch wheel cage. The arrow marks the day of cage swap. C: summary data showing the daily running activity of mice in each of the cages (n = 6). The arrow marks the day of cage swap.

Since the cages were slightly different in size, we tested for this by placing all mice in the smaller Tecniplast model 1144B cages. Half of the cages maintained the electromechanical switch wheel, while the other half had the magnetic switch wheel. In a pattern similar to that shown in Figs. 3 and 4, the size of the cage did not matter as the mice ran more in the cages with the electromechanical switch (9.2 ± 1.3 km/day electromechanical switch; 4.2 ± 0.55 km/h magnetic switch; P = 0.0044).

To break the data down further, we extracted bout data from the ClockLab software. A bout is a defined period of wheel running that can be further analyzed by the number of wheel revolutions over a period of time. For the mouse running activity to be considered a bout required that the mouse run at a minimum of five wheel revolutions per minute for a minimum of 5 min; anything less than this was not considered in the analysis. When looking at the bout data, we found that the average length of the bouts (112.4 ± 15.02 min electromechanical switch vs. 73.5 ± 6.6 min magnetic switch; P = 0.0365; Fig. 5A) and total number of counts/bout (5,742 ± 1,079 counts electromechanical switch vs. 2072 ± 258.5 counts magnetic switch; P = 0.007; Fig. 5B) were higher in cages with the electromechanical switch. In addition, like the average speed data, the peak rate within a bout was greater in the cages with the electromechanical switch (59.9 ± 3.99 counts/min electromechanical switch vs. 38.2 ± 1.93 counts/min magnetic switch; P = 0.021; Fig. 5C). The total number of bouts per day was not significantly different between groups (5.0 ± 0.65 bouts/day electromechanical switch vs. 6.1 ± 0.5 bouts/day magnetic switch; P = 0.23; Fig. 5D).

Fig. 5.

Activity bout data was examined for mice in the magnetic and electromechanical switch wheel cages (n = 6). Bout length (min; A), counts/bout (counts; B), peak rate (counts/min; C), and bouts/day (D) were studied. Magnetic switches were significantly lower when compared with electromechanical switch cages for bout length, counts/bout and peak rate. A two-tailed, paired t-test was performed (*P < 0.05; **P < 0.01).

We also questioned whether the presence of additional “living space” in the Tecniplast 1284 cage could serve as enrichment. For this reason all testing of enrichment was performed in the smaller Tecniplast model 1144B. Mice were placed in the Tecniplast model 1144B cage with the electromechanical switch. An aspen chew stick was placed in half of the cages. The chew stick was tethered to maintain position at the end of the cage away from the wheel, so it could not interfere with wheel activity. Care was taken to ensure it did not interfere with access to food or water either. Wheel activity data did not differ between cages with or without enrichment. Thus enrichment in the form of a chew stick was not enough to significantly change average wheel cage data (Fig. 6, A–D; 8.3 ± 0.83 km/day no enrichment; 9.7 ± 1.1 km/day enrichment; P = 0.3).

Fig. 6.

Mice were placed in electromechanical switch wheel cages with and without enrichment in the form of a chew stick (n = 6). No significant differences were observed for average run time (h/day; A), average speed (km/h; B), or average distance (km/day; C). D: summary data showing the daily average distance with and without enrichment is shown. A two-tailed t-test was performed to compare the effect of access to enrichment in the form of a chew stick on run time, average speed, and average distance. A two-way ANOVA was used to compare the average daily distance with and without enrichment.

DISCUSSION

Voluntary running wheels are frequently used as a means to assess physical activity in rodents. Most mice will readily run on wheels when present in a cage, but there is significant variation among mouse strains (2, 4, 7, 8, 16). It is critical that exercise wheel data be robust and reliable to accurately assess experimental outcomes. To date, no one has performed a comprehensive comparison of the various commercially available voluntary wheels and switch systems. To our knowledge, this is the first study to address this important issue. The main finding of this investigation is that the type of wheel cage counting switch (magnetic switch vs. electromechanical switch) was responsible for significantly different running distance/day and average speed with the same cohort of mice. Cage placement and the presence of enrichment in the form of a chew stick did not significantly impact running wheel data collected. This study underscores the need for recognizing the contribution of behavioral inputs to total running activity in mice as well as highlighting the need for precise reporting of wheel cage attributes. In addition, it suggests that care should be taken when selecting voluntary wheel cages for studies as added resistance has been demonstrated to impact muscle strength (13), slow sarcopenia, alter mitochondrial function, fiber type and autophagy in aging (22), impact innervation and neurogenesis (15), and muscle strength in the mdx mice, a model of Duchenne Muscular Dystrophy (1).

In many studies that use a voluntary exercise wheel, the exercise wheel is the enrichment item. Environmental enrichment appears to have an overall positive effect on mice (5). Wheel cages are generally not provided with shelters, toys, or other enrichment items because the design of the wheel cage does not allow for enrichment. The addition of nesting material or other enrichment items has the potential for the mice to physically block the wheel by interacting with the enrichment, thus leading to inaccurate or complete lack of activity counts. It is known that the mice run voluntarily, even to the point of abandonment of other activities (10). Mice can experience anxiety and other behavioral changes if the access to a wheel is suddenly withdrawn (17). They will also interact with other forms of enrichment when given the choice (14). Specifically, access to enrichment in the form of something to chew seems to calm mice, reducing reaction to novel exposure (11). Data presented in this study are in agreement with previous research suggesting that mice prefer running to enrichment (6, 14). The presence of enrichment in the form of a chew stick employed in this study did not significantly impact running wheel data. Further studies need to be completed examining alternative enrichment strategies, but it seems reasonable that if mice have a preferred alternative activity, they will run less. The impact of any given form of enrichment will be dependent on how much the mice favor the alternative to voluntary wheel running.

This study is the first to demonstrate that the use of an electromechanical switch for data interface vs. a magnetic switch significantly influenced the distance a mouse ran per day as well as the average speed. Mice ran faster for a longer distance in both large and small cages with the electromechanical switch. One potential explanation is that the audio/tactile input of the electromechanical switch could be acting as a cue modifier, encouraging the mice to run more. However, based on examination of bout data, as well as, direct measurement of the resistance inherent in the two switch types, we demonstrate that the electromechanical switch provides more resistance to the running wheel compared with the magnetic switch. With the magnetic switch there is little to catch or slow down the revolutions of the wheel. A wheel in free-spin holds potential to rotate faster than the mouse is capable of running. Thus the typical running pattern of mice in a more free-spinning wheel would be shorter, interrupted bouts. Wheels with slight resistance, due to the electromechanical switch, do not lead to the rate of wheel rotation exceeding the speed of running so these mice will run without interruption, and the bouts tend to be much longer. Previous studies with wheel cages that have resistance show this pattern (13).

The results of this study provide a first step in highlighting the need for the development of standardization for voluntary wheel data collection in mice. Our findings clearly demonstrate that running distance, running speed, and bouts of running are all different when mice are on wheels with the electromechanical switch. There are still other issues not addressed in this study including whether this is seen in both female and male mice and/or rats. Until standardization of wheels/cages is fully performed across systems, investigators are recommended to be cautious about interpretations of cardiovascular or musculoskeletal physiology when comparing data across different voluntary running wheel systems. Care should be employed when designing studies to be sure that the type of wheel cage when combined with a specific animal model will appropriately address the experimental questions, as unintended variation in wheel resistance can significantly alter results. We may not need to reinvent the wheel, but we can better understand how it can influence our data.

GRANTS

This study was supported by National Institute on Aging Grant R01-AR-066082 (to K. A. Esser, Principal Investigator).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.A.E. conceived and designed research; T.S. and B.D.H. performed experiments; T.S. and E.A.S. analyzed data; B.D.H., K.A.E., and E.A.S. interpreted results of experiments; E.A.S. prepared figures; E.A.S. drafted manuscriptT.S., B.D.H., K.A.E., and E.A.S. edited and revised manuscript; T.S., B.D.H., K.A.E., and E.A.S. approved final version of manuscript.