Trendelenburg-Like Gait, Instability and Altered Step Patterns in a Mouse Model for Limb Girdle Muscular Dystrophy 2i

Limb-girdle muscular dystrophy type 2i (LGMD2i) affects thousands of lives with shortened life expectancy mainly due to cardiac and respiratory problems and difficulty with ambulation significantly compromising quality of life. Limited studies have noted impaired gait in patients and animal models of different muscular dystrophies, but not in animal models of LGMD2i. Our goal, therefore, was to quantify gait metrics in the fukutin-related protein P448L mutant (P448L) mouse, a recently developed model for LGMD2i. The Noldus CatWalk XT motion capture system was used to identify multiple gait impairments. An average galloping body speed of 35 cm/s for both P448L and C57BL/6 wild-type mice was maintained to ensure differences in gait were due only to strain physiology. Compared to wild-type mice, P448L mice reach maximum contact 10% faster and have 40% more paw surface area during stance. Additionally, force intensity at the time of maximum paw contact is roughly 2-fold higher in P448L mice. Paw swing time is reduced in P448L mice without changes in stride length as a faster swing speed compensates. Gait instability in P448L mice is indicated by 50% higher instances of 3 and 4 paw stance support and conversely, 2-fold fewer instances of single paw stance support and no instance of zero paw support. This leads to lower variation of normal step patterns used and a higher use of uncommon step patterns. Similar anomalies have also been noted in muscular dystrophy patients due to weakness in the hip abductor muscles, producing a Trendelenburg gait characterized by “waddling” and more pronounced shifts to the stance leg. Thus, gait of P448L mice replicates anomalies commonly seen in LGMD2i patients, which is not only potentially valuable for assessing drug efficacy in restoring movement biomechanics, but also for better understanding them.


Introduction
(supplemental video 1). P448L mice also require more support from other legs while galloping, which limits step patterns and promotes a radial "giraffe walk" or pacing pattern. This gait substantially differs from that of wild-type mice and provides multiple metrics for assessing drug efficacy in this important animal model and possibly in other dystrophic mice or even muscular dystrophy patients.

Animal Care
Experiments were performed in male and female wild-type C57Bl/6 and P448L congenic mice, which also possess a C57Bl/6 genetic background, bred and housed in an IACUC accredited facility at Washington State University. Animals were fed and watered ad libitum and housed in an environmentally controlled room under a 12 h light/dark cycle. All handling and experimental protocols were preapproved by an Institutional Animal Care and Use Committee in accordance with the Guide for the Care and Use of Laboratory Animals. Two month old mice were used as this is a common age used in drug and gene therapeutic studies with other dystrophic lines and because at this age, both wild-type and P448L mice have similar body weights and lengths. Although hydrocephaly and other developmental defects were previously noted in P448L mice, none of these defects was detected in any of the mice used in this study.

Catwalk Setup
Six male and female mice of each genotype were run through the Catwalk XT system (Noldus Information Technology, Wageningen, NL) that has a camera mounted underneath a glass walkway to capture individual print, footfall and force profiles measured using a light intensity equivalent as described [28]. The camera was set 9 inches away from the walkway before capture and the green light threshold that captures print outlines and intensities on the walkway was set between 0.10 and 1.00 with a frame capture rate of 100/sec. Mice were coaxed to run by gently tapping on the walkway or blowing and runs longer than 5 seconds and shorter than 1.5 seconds were automatically excluded in order to maintain consistent speeds. Furthermore, each mouse was assessed individually for 3 consecutive viable runs and those that resulted in speed variance exceeding 20% were also excluded. Mice completed 3-4 step cycles in frame per run allowing the system to create average run statistics from the 3 individual runs. Each of these run averages was then used to create mean statistics between the mouse models.

Gait Cycle
The Catwalk system provides several measurement variables automatically (indicated by italics in the text). Swing time is the time a paw is airborne from one footfall to the next and stride length is the distance covered per swing. Dividing stride length by swing time results in the average swing speed of the paw. The step cycle is the time it takes to complete a single cycle of a step, effectively the sum of stance and swing time. The duty cycle then explains what percentage of the whole step cycle is encompassed by the stance phase: where stance (referred to as stand by the equipment) is the duration of the stance phase and step cycle is the duration of stance and swing phases combined (Fig 1A). showing print measurements and recorded intensities for the left hind paw of a P448L mouse. Light displacement leaves a green paw print on the walkway as shown and is used to derive print surface area metrics and intensity. The latter is an indication of weight supported by the paw. Intensity is measured in real time, graphed and used to calculate the highest intensity pixel in the print, Max Intensity (black line), and the Mean Intensity of all pixels (green area under curve). (C) Summary of support metrics. Solid prints and arrows represent a support pair, the opposite pairs are represented as fragmented prints and arrows. (D) Common step patterns include the alternate patterns AA and AB and the cruciate patterns CA and CB. Numbers and arrows indicate step order. The right front paw was arbitrarily chosen as the initial step.

Paw Contact
The stance phase is the time from initial to final paw contact where the transition between braking and propulsion occurs. During stance, the paw contact length, width and area are recorded over the entire stance phase as print length, width, and area, respectively, and at maximum contact area (Fig 1A and 1B). Stand index is the speed at which the paw leaves the walkway. It is measured using the following equation where "a" is the slope of the line and "X0" is the camera frame rate when the paw is at max contact: Time to max contact measures the time maximum contact is recorded relative to the beginning of the stance phase and is determined by the following equation: where Max contact At is the time of stance phase when the maximum foot area is in contact with the glass, Initial contact is the start time of the stand phase (when the foot first touches the walkway) and Stand is the stance time duration.

Paw Intensity
In addition to measuring timing and paw metrics during the stance phase, the Catwalk system also quantifies step intensities based on the amount of scattered light within a scale of 0 to 255. Intensity measurements are reported at specific instances, such as at maximum contact or are integrated throughout the stance period. These intensities reflect weight over the paws when they are in contact with the walkway, which the machine correlates to the force exerted by the paws. Max contact max intensity measures the mean of the highest intensity pixel of the print image at the maximum print contact for each step. Max contact mean intensity is the mean for all pixels in the print image during maximum print contact ( Fig 1B). Max intensity refers to the mean of the highest recorded pixel intensity through the entire stance phase while time to max intensity (referenced as max intensity at by Noldus) quantifies the time required to reach max intensity. Support Support from each limb in contact with the surface, while other limbs are in swing, can be measured during each step cycle and as a percentage of the whole recorded run. Initial and terminal dual stance measure the amount of time that both contralateral paws are on the walkway at the start and end of the step cycle, respectively. The distance between contralateral paws during initial and terminal stance is referred to as base of support (BOS). Single stance refers to the time a paw in swing is supported by the contralateral paw. The percent of support throughout the run can be parsed into the total number of paws as well as special instances with only 2-paw support (i.e. diagonal, contralateral, and girdle) ( Fig 1C).
Step Patterns Quadrupeds can potentially use multiple step sequence patterns when galloping and walking [29]. The Catwalk system records the actual order of footfalls as they occur and categorizes them based on common and uncommon sequence patterns. The number of different patterns used and the regularity index reflect the percentage of defined gait patterns used during the run according to: where normal step patterns is the total amount the defined AA, AB, CA, or CB gait patterns (see below) were used and paw placement is the total paw placement patterns recorded. The frequency of the four most common quadruped step patterns can then be determined and include the following: (AA) RF-RH-LF-LH, (AB) LF-RH-RF-LH, (CA) RF-LF-RH-LH and (CB) LF-RF-LH-RH and can be visualized in the summary figure ( Fig 1D).

Statistical Analysis
Statistical analyses were performed using Prism (GraphPad Software; La Jolla, CA). Initial tests assessed sex as an independent variable and identified no effect of sex in any parameter. The results from both sexes were, therefore, combined (n = 12) for each group and significant differences between means (-/+ SEM) were determined using a student's t-test (p 0.05).

Body Speed and Variance
As speed is an underlying factor that affects gait, maintaining body speed and its variation between mouse groups was critical. Body speeds were maintained at a gallop with no significant differences between groups (wild-type: front paws 35.45±8.80 cm/s, hind paws 35.77±8.03 cm/s; P448L: front paws 33.99±3.89 cm/s, hind paws 34.37±4.65 cm/s) [30,31]. Additionally, there were no differences in body weights between the two groups independent of sex (data not shown). These results indicate that all proceeding differences in gait are due to gait mechanics and not speed.

Gait Cycle
Swing time is less in P448L mice, indicating that they spend significantly less time with their paws in the air and this is consistent with a higher forepaw swing speed (Fig 2A-2D). However, these differences did not lead to longer a stride length or slower step cycle (Fig 2E-2H). This suggests that although the velocity and cadence between the groups were similar, the stance phase in the P448L must be longer relative to swing as this compensates for less swing time. This is confirmed in the duty cycle measurement as P448L mice proportionally spend more time in the stance phase compared to swing phase (Fig 2I and 2J). This is not to say that stance time is significantly different between mouse groups, but the percentage of the step cycle spent in the stance phase is higher.

Paw Contact
The stance phase of the step cycle begins when paws contact the walkway and ends when paws leave and initiate swing. Although stance was similar in both groups (Fig 3A and 3B), the time at which paws come into maximal contact with the walkway, which represents the transitionary phase between braking and propulsion, is reached in P448L mice roughly 10% faster than wild-type (Fig 3C and 3D). This in turn lead to a faster speed of the paws leaving the walkway for P448L mice as indicated by a larger stand index (Fig 3E and 3F). The total surface contact area of the paws was 2.5 times larger in P448L mice (Fig 3G and 3H). Indeed, print length and width throughout the entire stance phase are roughly 30% larger in both front and hind paws of P448L mice, resulting in 70% larger total print area (Fig 4A and 4F). This indicates that P448L mice gallop with a more "foot slap" gait, which could account for the shortened braking period and the need for a longer propulsion phase during stand (Fig 3C and 3D).

Paw Intensities
The step weight placed on the paws by P448L mice, as indicated by the intensity of light scatter, was significantly higher than that of wild-type regardless of metric. Maximum and mean intensities at the time of maximum contact were 2-fold higher (Fig 5A-5D), consistent with the maximum intensities of the whole paw throughout the stance phase (Fig 5E and 5F). This is likely due to a higher shift in weight between paws during stance and an extended propulsion phase in P448L mice. Time to max intensity refers to the time it takes paws to reach the largest weight exerted on the walkway in proportion to the stance phase, much like the time to max contact metric. These two percentiles are normally identical or very similar, as seen with the wild-type paws. However, time to max contact (Fig 3C and 3D) and time to max intensity (Fig 5G and 5H) are uncoupled in P448L hind paws. The P448L hind paws exhibit a disconnect between max contact and max intensity, with the max contact reached at 23% of the total stance phase as opposed to 37% for the max intensity. Thus, the mice are braking during the stance phase with a larger weight being placed on the walkway and braking faster than wild-type in their hind legs. This means that proportionally more weight is being used to stop momentum in the P448L mice than during the beginning of the propulsion phase, which transitions roughly at the point of max contact as seen in the wild type mice. This is again consistent with a foot slap gait, but additionally suggests that gait pathologies are more apparent in P448L hind paws.

Stance and Support
Step cycle differences in the support by contralateral limbs suggest that P448L mice require more support when running. The time spent with one paw in swing and the contralateral paw on the surface (single stance) was less in P448L mice (Fig 6A and 6B). Conversely, initial and terminal dual stance measures in P448L mice, measures of simultaneous contralateral support, were double those of wild-type (Fig 6C-6F). These data partially explain why swing time is shorter in P448L mice, which compensate their instability by rapidly accelerating limbs through the step cycle in order to provide greater support with more paw contacts. The base of support in hind paws was larger in P448L mice, although forepaw support was smaller (Fig 6G and 6H). This is not particularly surprising as hind paws carry more proportional mass than forepaws.
Similar differences were also detected when support metrics assessed all paws over the course of the entire run. At no time did P448L mice gallop with all paws in the air and their relative time with only one paw on the surface was approximately half that of wild-type mice ( Fig  7A), which is consistent with lower single stance measures (Fig 6A and 6B). By contrast, the relative time with 3 or 4 paws on the surface was 63% and 125-fold higher, respectively, in P448L mice (Fig 7A), and this is again consistent with higher dual stance measures (Fig 6C-6F). Multi-paw support in P448L mice was not distributed equally as the frequency of diagonal support (e.g. front left and right hind) was similar in both groups (Fig 7B). The frequency of girdle support (both paws on a side), however, tended to be higher in P448L mice and although this difference was not significant, the frequency of lateral support was less than a third of the wild-type percentage (Fig 7B). The latter results do not conflict with a higher incidence of dual stance in P448L mice as lateral support metrics are collected when only two paws touch the surface and P448L mice gallop predominantly with three or even four paws touching.
Step Patterns P448L mice used fewer step sequence patterns than wild-type mice, although the regularity index was similar in both groups indicating that they predominantly used the four most common patterns (Fig 8A and 8B). The primary difference in patterns used by wild-type and P448L mice were in the frequency of radial patterns. P448L mice used the AA "giraffe walk" pattern 2.5-fold more often than the wild-type mice that used, as is typical, the AB pattern 40% more often than did P448L mice (Fig 8C and 8D). Alternating contralateral step CA/CB patterns were used at the same frequency between strains and at smaller frequency than the radial AA/AB patterns (Fig 8E  and 8F). Thus, the varied gait mechanics and need for greater support in P448L mice is accompanied by the use of fewer, albeit common, step sequence patterns and a reliance on the AA radial gait pattern that would result in a similarly higher use of girdle support.

Discussion
Understanding the adverse effects on gait in a disease state that causes loss of ambulation has the potential to influence patient management and even the development of novel treatment strategies and therapeutics. This highlights the importance of our studies as they are the first to quantify abnormal gait patterns in any dystroglycanopathy model, regardless of specific disease type, and are by far the most comprehensive assessment of gait in any dystrophic animal model to date. Compared to wild-type mice, the "waddling" or Trendelenburg-like gait of P448L mice is exemplified by a lesser time to maximal paw contact, greater paw surface   contact, higher paw intensity measures, smaller paw swing times and the requirement of greater limb support, which is consistent with a higher frequency of irregular step patterns and pacing (summarized in Fig 9A-9C). These results correspond with some human muscular dystrophy gait metrics [22,[32][33][34][35][36] and despite the obvious differences in bipedal versus quadruped gait, they illustrate the translational utility of assessing gait in P448L mice. Indeed, these studies identified several novel biomechanical disease markers that would facilitate the testing of drug candidates and other therapeutic interventions in whole animals.
The distinctive P448L gait is readily distinguishable (see S1 Movie of mice on a treadmill). It is also consistent with a Trendelenburg gait [26,27,37] that commonly develops with different muscular dystrophies and myopathies and results from weakened muscles of the hip abductors as in LGMD2i patients [2,24,38]. This causes the hips to dip excessively as subjects walk and often results in a characteristic "waddle" and "foot slap" [39][40][41], which also occur in P448L mice. Although the P448L hip abductors have yet to be functionally analyzed, hind limb retraction time is compromised in 2 week old P448L mice as is hind limb grip strength [7,15], all of which is consistent with limb and hip muscle dysfunction.
The Trendelenburg-like gait in conjunction with flat step patterns partly explain why paw intensity metrics are higher in P448L mice. These metrics reflect paw intensities generated during the step cycle and although it seems counterintuitive that dystrophic animals generate more force with each step, this could be explained by a less fluid gait pattern causing hard footfalls (i.e. foot slap) that require higher weight shifts during a faster braking period and, therefore, an extended propulsion period. This is supported by the fact that max intensity in P448L mice is reached before max contact and this coincides with higher instability and need for more support (i.e. swing time is lower). However, intensity only truly measures proportional weight exerted on the walkways and does not provide a quantitative measure of actual exerted force, which is a limitation of the measure.
It is logical to assume that a Trendelenburg-like gait in a quadruped might require heightened support from contralateral limbs. This indeed occurs in P448L mice and coincides with a faster swing speed and less swing time (Fig 2). In fact, a slower swing time results from added stress on the hip abductors in patients with a bilateral Trendelenburg gait [42]. The P448L Step patterns are also significantly affected as P448L mice rely heavily on just a few patterns, especially radial step patterns that are uncommon in wild-type mice (Fig 9). One shortcoming in the Catwalk system is that it does not measure the joint kinematics that can also influence gait mechanics [43,44]. Such technologies have recently been developed for hindlimb assessments and could potentially be used to confirm the Trendelenburg-like gait and especially the kinematics of pelvic tilt [45]. It is also possible that problems in neural developmental may have contributed to differences in gait mechanics as FKRP is expressed in muscle and brain. We suspect this is highly unlikely as none of the mice used in this study displayed the previously reported neural abnormalities, in fact none have occurred in our breeding colony, and more importantly, the Trendelenburg gait also occurs in human patients with other forms of muscular dystrophy. The phenotype described is, therefore, a consequence of muscle weakness rather than neural deficits.
Early diagnosis of muscular dystrophies allows for earlier preventative treatment and extends patient ambulatory times especially as minor gait issues are often among the first noticed symptoms [22,32,46]. This underscores the need to perform gait analyses in other muscular dystrophy animal models and in human patients as such studies, especially in mice, are extremely rare. The use of gene therapeutics to treat various forms of muscular dystrophy has shown significant promise. This includes studies using mdx mice to replace mutated dystrophin with microdystrophin and those using L276I mice to restore wild-type FKRP [16,[47][48][49][50][51]]. The Catwalk system provides an opportunity to assess these and other dystrophic mice in vivo using a "real world" metric-gait-which our studies demonstrate is similarly affected in P448L mice and human LGMD2i subjects. A better mechanistic understanding of gait abnormalities in P448L mice could, therefore, help to better understand the biomechanical nature of similar pathologies in humans and possibly more important, to aid in the screening of novel drugs and or gene therapeutics for treating LGMD2i.
Supporting Information S1 Movie. Treadmill Running Gait. C57BL/6 wild-type and P448L mice were recorded running on the Columbus Instruments Oxymax FAST treadmill system. The treadmills were set at 12 m/min and the video was slowed to 60% speed. (MOV)

Acknowledgments
This study was made possible with assistance from current and former members of the Rodgers Lab includuing Jennifer Eldridge, Naisi Li, Jordan Rehwaldt,and Denali Kagel. We would additionally like to thank Joe Harding for technical assistance with the Noldus Catwalk System.
proportion of the cycle, maximum contact is reached earlier and swing time is less in P448L mice while stride length is unaffected due to compensatory faster swing speeds. (B) Visual representation of differences in initial dual stance (IDS), single stance (SS), and terminal dual stance (TDS) between WT and P448L mice (gray boxes = paw in contact with surface; lines = periods of IDS, SS or TDS). The longer IDS and TDS and shorter SS in P448L mice reflect longer periods of contralateral paws touching the walkway during the gait cycle and, conversely, less time with only one contralateral paw touching. This was also reflected in greater instances of 3 and 4 paw support in P448L mice. (C) Hildebrand plot of the common gait patterns: AA (RF-RH-LF-LH), AB (LF-RH-RF-LH), CA (RF-LF-RH-LH) and CB (LF-RF-LH-RH). Percentages indicate relative use of each step pattern in each mouse group. The most common step pattern in both was the alternating AB pattern, although P448L mice used it significantly less and compensated with an increase use of the radial AA pattern. doi:10.1371/journal.pone.0161984.g009