The mouse jumping apparatus
We have designed and constructed the MJ (Mouse Jump) apparatus; a plyometric training device for one mouse that induces box jump motion using a green flashing light emitting diode (LED) followed by a 175 VRMS, 60 Hz AC voltage on the platform as the jump trigger. The device (Fig. 5a) has two movable platforms and the mouse is placed on one platform at the beginning of the cycle. Controlled by a custom Python algorithm, the platform hosting the mouse at the start of the cycle is lowered, while the second platform is raised to the desired height. As the platform stops, the green LED is activated to signal the incoming electric shock. The motorized gate opens prior to the jump and closes when the jump is complete (Fig. 5b). An infrared transmitter paired with a photodiode receiver on each platform senses whether the mouse is present on the elevated platform after the jump, and triggers the gate closure. Subsequently, the platforms are inverted so that the higher platform (which now hosts the mouse) is lowered and the mouse can repeat the jumps for the chosen number of repetitions. In Fig. 6 we show close-up views of the MJ apparatus. In Fig. 6a the mouse is sitting on the lower platform, ready to jump as soon as the green LED is activated and the gate is open. Figure 6b shows a view from the lower platform level, at the time when the green light is on and the gate is open to allow the jump to the upper platform. Supplementary videos SV1 and SV2 show a mouse jumping during the experiment, as seen from the bottom and from the top, respectively.
a Full view of the MJ apparatus imaged on the lab table during the experiment. In panel b we show the basic mechanism of the apparatus. A mouse is sitting on the lower platform (1) when a green LED light (not shown) turns on. Shortly after that the gate opens and a mild electric shock is applied to the platform. The mouse learns that as soon as the gate opens she should jump onto the higher platform 2. When the mouse reaches platform 2, the gate closes, and the two platforms move (using motors M1 and M2) so that platform 1 is raised and platform 2 is lowered, and the cycle continues for the next jump, until the desired number of jumps is achieved. The platforms are located in the central (dark) part of the machine shown in panel a.
a A mouse is sitting on the lower platform as the platform is being moved down by the motor (red box on the right). The upper platform is located on the opposite side of the apparatus (not visible in this figure). b A mouse is sitting on the lower platform as the green LED turns on to signal the incoming shock. The gate to the upper platform is open and the mouse can jump up to the upper platform (visible above the green LED level). See videos of a jump in the Supplementary Videos SV1 and SV2 showing views from the bottom and the top, respectively.
In the engineering phase of the project we tested MJ using slowly increasing voltages, making sure that the mouse would feel only a very mild electric shock. At the beginning of the experiment, we observed that the current is only relevant in the first training phase for each mouse, since the animal quickly learns to jump as soon as they see the LED light turning on and/or the gate opening.
The platform elevation is variable, with the elevation of each platform being controlled independently by a stepper motor driven linear stage. This control allows jump height to be changed via software over a range of 0–300 mm, and allows to carry out standardized, well controlled exercise regimens. The MJ software platform control can be set up to allow for jumps from lower to higher elevations or from higher to lower elevations. For this experiment, we only focused on jumping up to reduce the risk of overloading. Jump data is recorded by software. Collected data include the number of successful jumps, the jump height for each jump, and the number of failed jumps. A failed jump is defined as an event in which the mouse does not reach the upper platform before the hatch is closed. We observed a few of those events at the beginning of the training week, but none were recorded during the 8-week experiment.
Mice
Animal treatment and care followed the rules and regulation of OLAW and the requirements of Carnegie Institution (vertebrate animal assurance # A3861-01). Procedures and husbandry of mice used in this project are described in animal protocol #163 and approved by Carnegie Institutional Animal Care and Use Committee (IACUC). The following descriptions of animal use are organized according to the ARRIVE Essential 10 guidelines 2.0. The study design was to test whether jumping exercise (JUMP) helps improve knee cartilage health. The jumping mice were compared to control (sedentary) and hind limb suspended (HLS) mice (also specified in the main text). The experimental unit is each mouse. C57BL/6J female mice were purchased from the Jackson Laboratory (JAX) at 2 months (mon) of age, acclimated to the new housing environment for one month, and used at 3 mon old. Total of 15 mice were assigned to 6 control mice, 5 JUMP mice and 4 HLS mice (see Experimental design section below for details). The sample sizes were determined based on prior published studies. The JUMP protocol is detailed in the Plyometric training protocol section below. Due to their same genetic background and sex, no specific criteria was applied to their exclusion. As such, data collected from all animals were included in tables and figures, including the n number for each group. Mice assigned to each group were chosen at random after receiving the shipment from JAX. To minimize environmental counforders, all mice were housed in the same room, with the same food, water, and bedding. Their health was monitored every other day (see blow). Due to experimental design, data recording was not blinded. However, at least two people had independently examined and analyzed the data for conclusions. Outcome measures included Micro-CT, histology, and body weight (details in relevant sections below and in the main text). The primary outcome measure was the articular cartilage health (thickness and OA score). Statistical methods and presentations are described below in the Statistical analyses section, as well as in figures and figure legends. The choice of pure bred C57BL/6J female mice of the same age (3 mon) was to ensure proper comparison among groups, and also because of their widespread use in research related to human health. Details for specific experimental procedures are described in several sections below. The results are reported in the main text and figures with means, standard deviations, and p values.
Hindlimb suspension
We designed and constructed hindlimb suspension (HLS) cages to host 4 mice. Mice in the HLS group underwent a minor surgical procedure to implant a wire loop near the base of the tail for suspension, following the procedure described in ref. 49, similar to the NASA recommended procedure50 using the rat model. Briefly, one week before the HLS group reached 3 months of age, they were anesthetized by Avertin (Sigma; at 0.025%, 10 μl per gram body weight, via intraperitoneal injection) to reduce pain. Their tails then were cleaned with 70% ethanol before implanting 2-0 surgical steel sutures (Ethicon) through tail vertebrae. Extra-lengths of steel sutures were braided into extension loops (for suspension clips) above surgical sites, and surgical tapes were wrapped around the tails after being sprayed with Bactine, which is an antiseptic and pain reliever. After one week of recovery, the mice were subjected to HLS leading to unloading of the hind limbs. The steel loop on the tail was hooked to the clip at the end of an adjustable threaded plastic rod, which is attached to a horizontal PVC bar. The bar has rollers along the railway at the top of our custom-designed mouse tail suspension cage. For each mouse, the hindlimbs were lifted off the cage bottom at ~30° angle by adjusting the length of the plastic rod.
Each cage had two chambers (25 cm (W) × 25 cm (L) × 30 cm (H) each) divided by a clear acrylic divider with small holes that allowed nose-touch and social interaction between 2 animals. Water bottle, food tray, and woodchip bedding were placed inside each chamber and changed weekly.
The health status of the mice was monitored throughout the experiment. No signs of paralysis based on the movement of suspended hindlimbs and non-suspended forelimbs were seen. The urine color remained yellowish throughout (i.e. no accumulation of reddish porphyrin). There was likely stress caused by suspension, based on animals’ immediate weight loss and lagged recovery compared to other groups.
Experimental design
A group of 4 mice were hindlimb suspended (HLS group) using custom-made suspension cages. A group of 5 mice were trained with jumps (JUMP group) three times a week, for a total of nine weeks, including one week at the beginning in which the mice got accustomed to the MJ machine. The jumping exercise protocol utilizes the standard progressive overload method to increase the stimulus during the program. We define the volume V of a training session as the height of the jump (i.e. the vertical distance between the two platforms in the MJ machine, h) and the number of jumps N for each mouse, in each session. The volume V = h * N is smaller at the beginning, and it is progressively increased every other week. More details on the protocol are given below in the appropriate section. After each training session, the JUMP mice were put back in their regular cages. A control group of six mice (CON) was housed in their regular cages for the total time of the experiment (9 weeks). All mice were in the same room at all times. No changes in the activity of the CON and HLS groups were made during the nine week duration of the experiment. Figure 1 shows a schematic of these 3 groups.
Plyometric training protocol
Mice in the JUMP group exercised in the MJ apparatus three times a week (Mondays, Wednesdays and Fridays) for a total of nine weeks. The first week (Week 0) was dedicated to teaching the mice in the MJ apparatus, starting with a jump height of 5 cm. After 2–5 trials the mice learned how to avoid the electrical shock and jump to the elevated platform when the light flashed. When the mouse jumped confidently, we increased the jump height by 2 cm each time, until we reached a height difference of 15 cm between the platforms. This procedure was performed over four consecutive days, cycling through all of the five mice in the JUMP sample, until the mice confidently jumped at least 15 cm for at least 5 consecutive times. During the training week, the total number of jumps performed by each mouse was 23–41. In between each training session, the JUMP mice were housed in their regular cages.
The exercise program started in week 1 and all JUMP mice followed a predefined progressive overload protocol. Progressive overload is a standard procedure implemented in human exercise programs for the purpose of gradually increasing the neuromuscular demand to facilitate further adaptations over time46. Exercise volume (V) per session was: V = N * h, where N is the number of jumps per session and h is the height (in cm). During the first week each of the mice performed 10 jumps with h = 15 cm, for a volume of 150. In Fig. 7 we show the progressive volume per session in each week, up to a value of V = 300 on week 8 (15 jumps, h = 20). Jump frequency depends on the time the platforms take to reach the predetermined height, and it was in the range ~0.01–0.05 Hz.
The plot shows the progressive change in exercise volume plotted against the week, from week 1 to 8 (end of the experiment). The volume is calculated as the number of jumps per session (i.e. each training day) multiplied by jump height in cm.
Jumping form and efficiency did not decrease over time, which could have signaled muscle-level damage. The mice quickly learned to jump over the first week of training and progressively improved their ability to reach greater heights and performed more jumps per session, as the protocol progressed to higher volumes.
Tissue harvesting
Control, JUMP, and HLS groups were harvested on the same day at the end of the training period. Mice were anesthetized by Avertin (Sigma; at 0.025%, 10 μl per gram body weight, via intraperitoneal injection, prior to being sacrificed by cervical dislocation. The two-step euthanasia was used to reduce stress and pain. Afterwards, one hindlimb was dissected out and used for histological analysis, and the contralateral hindlimb prepared for micro-CT. Given there is no laterality of limb usage in rodents, left and right hind limbs were randomly chosen for either histological analysis or micro-CT.
Histology
For histology, each removed hindlimb was immersed in 10% neutral buffered formalin (Polysciences) overnight, followed by decalcification in 10% EDTA/PBS (w/v) for 10 days, and dehydration and embedding in paraffin. Samples were sectioned at 5 μm thickness (frontal plane) and processed for hematoxylin-eosin staining per manufacturer’s (Surgipath) recommendations, and mounted in Micromount (Surgipath) with coverslip (VWR). At least 10 sections per mouse were examined. Images were taken using a Canon (EOS T3i) camera mounted on a Nikon E800 upright scope, under a 10X objective.
We assessed the severity of osteoarthritis on the medial tibial plateau in each section utilizing the Articular Cartilage Structure (ACS) score for osteoarthritis, which has been validated for use with hematoxylin-eosin staining29,30. We analyzed 10–14 mid-coronal sections from each group, with an average between 2 and 3 slices per mouse. Scoring was performed in a blind manner, by randomizing the images and concealing the group each image belongs to. Only after scoring all images the group was revealed. Scoring of three expert classifiers was then averaged. We tested for potential differences between the three groups as described in the Section Statistical Analysis.
Contrast-enhanced Micro-Computed Tomography (micro-CT) and image analysis
To evaluate cartilage by micro-CT19, tibiae were fixed in 10% neutral buffered formalin for 3 days and further incubated in 5% Phosphotungstic acid solution at room temperature for 7 days. The samples were washed in PBS before scanning. The samples were imaged using an ex vivo high-resolution micro-CT imaging system (Bruker Micro-CT, Skyscan 1275, anode current 200μA; voltage 50 kV, use of 1 mm of an aluminum filter, scanning exposure time 218 ms with 0.3o stem rotation, frame averaging of 6 with an image resolution of 5μm/pixel). After scanning, the image projections were reconstructed by the NRecom software (Bruker). CTVox and CTAan (Bruker) software were used for further analysis and visualization of reconstructed images. All images were oriented to align the knee-to-knee long axis in Dataviewer (Bruker). Tibial cartilage volume was measured at 200μm radii circular volume of interest of medial femur-tibial contact point within the cartilage (transaxial view of scans with a fixed number of slices for all samples). Tibial cartilage thickness was measured manually from the lower to upper margin of the cartilage coronal sections.
Bone-associated parameters (Tibial BMD, Trabecular BV, Trabecular thickness), to understand the effect on trabecular and cortical tibial bone were also assessed using the manufacturer software. For tibial cortical analysis, the mid-diaphyseal region was located using growth plate and notch after initial reorientation of all samples in the same position in Dataviewer software. Within the defined region of interest slices (ROI, approximately 60 slices)51, the cortical bone ROI is manually drawn including the outer periosteal surface. The cortical bone parameters were measured within a threshold of 90–255. For trabecular measurement, from the aligned scanned images, the growth plate and notch were noted initially. An offset of 25 slices from the growth plate was used and the analyses were performed between 150 and 200 slices for all the samples51,52. Next, the trabecular region of interest was manually drawn for all the slices of the samples. From the trabecular ROI and binary image verification, the bone parameters were measured with global thresholding between 110 and 255.
Statistical analysis
Statistical analysis of the micro-CT measurements was performed using R53. The level of statistical significance was set at α = 0.05. Normal distribution was tested with the Shapiro-Wilk test (shapiro.test in R53) and homogeneity of variances with the LeveneTest (leveneTest54 in R). Both tests returned a p-value that does not allow for rejection of the null hypothesis, thus allowing the use of ANOVA (oneway.test in R53) to test for group (JUMP, HLS, CON) equality in cartilage and bone measurements. Additionally, we performed a pairwise post-hoc analysis to determine which group differs from each of the others using a Holm post-hoc method (pairwise.t.test in R53). All animal data were included for analyses., i.e. no exclusion criteria. The same statistical method was utilized to investigate differences in the ACS score for osteoarthritis and body weight between groups.
link