Background

The use of voluntary running wheels in laboratory rodent cages is a common approach for investigating the impact of exercise on brain function and neurodegeneration (Liu et al., 2019). Employing a voluntary exercise study design eliminates the need for a stressful stimulus, such as a foot shock or investigator handling, to encourage the animal to exercise. Studies utilizing in-cage voluntary running wheels have traditionally housed rodents individually to generate accurate running distances and capture individual variations in exercise amounts (Goh and Ladiges, 2015). Individual running data can then be correlated with other intra-individual datasets. For example, total distance run over a 2-6 hour period positively correlates with brain-derived neurotrophic factor (BDNF) gene expression in the hippocampus (Oliff et al., 1998). However, a disadvantage of individual housing is the potential for stress induced by social isolation. Since mice are social animals, their innate social behaviors (whether in wild or captive environments) develop most naturally in social arrangements similar to those found in wild colonies (Reimer and Petras, 1967). The controlled laboratory environments required for accurate data collection frequently result in housing conditions that challenge the formation of natural social structures. The lack of a social environment when mice are individually housed must be considered not only for rodent welfare but also for its impact on the quality of data produced and the interpretation of findings (Kappel et al., 2017; Arakawa, 2018).

Prior research has demonstrated that individual housing of mice can impact brain function in a number of ways. It can alter hypothalamic-pituitary-adrenocortical (HPA) axis function, specifically glucocorticoid regulation and feedback (Hawkley et al., 2012). Individual housing can also confound performance on various behavioral tests, including those assessing anxiety (Koike et al., 2009) and learning and memory (Okada et al., 2015). Finally, single housing can lower the expression of neuroplasticity-related genes in the hippocampus and prefrontal cortex (Ieraci et al., 2016). Social isolation stress in singly housed rodents is therefore an important variable to consider when interpreting studies that assess the neurobiological effects of voluntary exercise.

On the other hand, group housing of mice can invoke male-on-male aggression. This usually emerges in rodents after the onset of puberty and tends not to be present during juvenile and adolescent developmental stages (Terranova et al., 1998). Female mice have a lower tendency to exhibit this aggression, even post-puberty (Hayes, 2000). Social isolation in juvenile rodents adversely impacts myelination in the medial prefrontal cortex (Makinodan et al., 2012), whereas social play in juveniles can enhance neural plasticity in this region (Himmler et al., 2013). Moreover, juvenile mice use the body temperature of their cage mates for thermoregulation (Batchelder et al., 1983). This underscores the importance of paired or grouped housing for maintaining basal body temperature, particularly in the setting of shifting metabolic demands with exercise. Therefore, in studies linking juvenile voluntary wheel running with neural function and behavior, a group housing-based approach may be preferred to eliminate isolation stress.

Individual home-cage rodent tracking has been accomplished through the use of video (Krynitsky et al., 2020) (Poffe et al., 2018) (Wang et al., 2018), subcutaneously implanted RFID microchips (Peleh et al., 2019) (Frahm et al., 2018), a combination of the two, or passive infrared sensors (Matikainen-Ankney, Garmendia-Cedillos et al., 2019). Current protocols utilizing subcutaneously implanted RFID microchips require prolonged restraint and anesthesia, which may produce undesired stress and pain. Unified Information Devices details a protocol that describes safely implanting a mouse RFID microchip (UID UC-1485) without anesthesia on postnatal day 12. This approach is used primarily for mouse identification at a single point in time, such as for taking rapid mouse inventory within a group-housed cage, but its use has not been demonstrated for live, continuous tracking. Another factor is the size of the RFID chip, which must be large enough to be detected by the antenna. For juvenile mice, typically weighing 6-9 grams, this size requirement prohibits administration without the use of anesthesia. In video tracking, wire cage tops obstruct continuous top-down video recording but are required for many home-cage running wheel systems. Finally, existing home-cage RFID activity-tracking systems currently on the market use a matrix of RFID antennae arranged throughout the base of the cage or in strategic locations to monitor baseline ambulatory activity (Voikar and Gaburro, 2020). However, if the objective of the experiment is to track voluntary wheel-running, only one antenna with a read range precisely limited to the area inside the wheel is required.

Our apparatus uses strong, low-profile RFID ear tags and one RFID antenna at the base of each cage to present a minimally invasive alternative to existing video-based and implantation-based tracking systems (Figure 1). The major advantage of our protocol is the ability to individually track running wheel activity of pair-housed juvenile mice starting at the age of weaning (postnatal day 21). We reduce isolation stress and issues with thermoregulation by pair-housing the mice and providing nesting material. Indeed, this model can be scaled up to >2 mice per cage if preferred. This procedure has been tested in cages containing mice of the same sex, but we did not investigate mixed-sex population effects on individual running activity. Our experimental design allows for the comparison of individual running distances between female and male pairs of mice within the same cage. Although there may be territorial issues once social hierarchy is established in cages housing two or more male mice, a precise quantity of exercise from each individual mouse in a group-housed cage can be correlated with any desired experimental readout.

A minor limitation of this apparatus is that in our hands, about 5-10% of running activity is unable to be reconciled; however, we believe that our 90-95% accuracy of allocating individual running data in group-housed environments is an acceptable yield. Although this method does not encounter the same issues with accuracy-diminishing antenna cross-talk as observed in RFID antenna matrices (Voikar and Gaburro, 2020), the metal wheel in very close proximity to the antenna may have the same interfering effect (thus necessitating a large ear tag). Our design effectively limits the read range of the antenna to the wheel area only.

Investigating the effects of voluntary physical activity has numerous implications for increasing our understanding of its benefits toward brain function and behavior. Determining individualized amounts of exercise for specific outcomes or targets requires accurate monitoring of running distances, which can now be performed as described in this protocol. Our approach can track individual mice by recognizing the unique RFID tags of pair-housed mice coupled with time-stamped data obtained from a magnetic sensor on a home-cage running wheel. Importantly, the procedure can be scaled up to tracking running distances of multiple mice in a cage sharing access to one running wheel. Our method brings RFID applications to juvenile mouse exercise studies, while minimizing stress from isolation and restraint, eliminating the need for anesthesia, and producing highly precise and individualized running data.