To investigate the co-morbid Alzheimer's disease (AD) and stroke condition in a novel model, three behavior tasks are described that assess both motor control and cognitive behaviors. These tasks include the beam-walk task, cylinder task and Morris water maze.
Alzheimer's disease (AD) is a debilitating neurodegenerative disease that results in neurodegeneration and memory loss. While age is a major risk factor for AD, stroke has also been implicated as a risk factor and an exacerbating factor. The co-morbidity of stroke and AD results in worsened stroke-related motor control and AD-related cognitive deficits when compared to each condition alone. To model the combined condition of stroke and AD, a novel transgenic rat model of AD, with a mutated form of amyloid precursor protein (a key protein involved in the development of AD) incorporated into its DNA, is given a small unilateral striatal stroke.
For a model with the combination of both stroke and AD, behavioral tests that assess stroke-related motor control, locomotion and AD-related cognitive function must be implemented. The cylinder task involves a cost-efficient, multipurpose apparatus that assesses spontaneous forelimb motor use. In this task, a rat is placed in a cylindrical apparatus, where the rat will spontaneously rear and contact the wall of the cylinder with its forelimbs. These contacts are considered forelimb motor use and quantified during video analysis after testing. Another cost-efficient motor task implemented is the beam-walk task, which assesses forelimb control, hindlimb control and locomotion. This task involves a rat walking across a wooden beam allowing for the assessment of limb motor control through analysis of forelimb slips, hindlimb slips and falls. Assessment of learning and memory is completed with Morris water maze for this behavioral paradigm. The protocol starts with spatial learning, whereby the rat locates a stationary hidden platform. After spatial learning, the platform is removed and both short-term and long-term spatial reference memory is assessed. All three of these tasks are sensitive to behavioral differences and completed within 28 days for this model, making this paradigm time-efficient and cost-efficient.
Alzheimer's Disease (AD) is the most prevalent form of dementia in the elderly population and a debilitating neurodegenerative disease. Histopathologically, AD presents itself as amyloid plaques, neurofibrillary tangles and neuronal loss. Amyloid plaques consist primarily of beta-amyloid peptide (Aβ) that has been produced through an altered proteolytic cleavage of amyloid precursor protein (APP) by β-secretase and γ-secretase enzymes1,2,3. The cleavage product, Aβ, deposits in the brain creating pathological amyloid plaques and has toxic effects on the brain that can lead to the characteristic learning impairments and memory loss. All of these steps together are referred to as the "amyloid cascade hypothesis"3,4. While this hypothesis is important when investigating AD, other cellular changes have been found to precede these plaque formations that strays from the original amyloid cascade pathway. These other cellular changes are thought to contribute to early memory loss, learning impairments and other cognitive dysfunctions involved in AD prior to plaque formation3,5,6.
With AD becoming increasingly prevalent, risk factors for developing AD are becoming an extremely important focus of research. Although age is the principal risk factor for sporadic forms of AD, other risk factors have been identified, including stroke7,8. Stroke is not only a risk factor, but it can also exacerbate already present dementias. For example, clinically, the progression of AD has been shown to be worse in patients who had previously experienced stroke9. Moreover, increased APP expression and Aβ accumulation has been found in experimental animal models of Aβ toxicity combined with induced stroke10,11. Since there is this important interaction between stroke and AD, it is essential that these two pathologies be further researched together in co-morbid models to better understand pathophysiology and behaviors implicated in both conditions.
To investigate co-morbid conditions, an appropriate model had to be developed in which a stroke could interact with Aβ to produce AD-like pathology. For the first time, an APP21 transgenic rat that has a mutated human APP gene incorporated into its DNA was used to achieve an appropriate model of AD. The mutations are the Swedish double missense and Indiana single missense mutations, which have both been implicated in familial forms of AD4,8,12. In the absence of an additional insult, this rat model ages without developing the characteristic Aβ plaques or neurofibrillary tangles12. Therefore, in an effort to induce AD-like behavioral pathology, a small stroke is introduced into the right striatum to mimic the small subcortical strokes often present in dementia patients9. The stroke in the APP21 transgenic rat embodies the co-morbid condition and allows investigation of various types of behavioral changes implicated in both disease conditions. In particular, this induction of AD-like pathology and cognitive deficits in the adult rat allows us investigate the earliest molecular and cognitive changes preceding AD.
Since the goal is to determine the first signs of behavioral changes and since both stroke and AD have very distinct behavioral pathologies, when studying the co-morbid model, behavior tasks need to assess a variety of behavioral phenotypes. There are a battery of relatively sensitive tests that can be done to analyze motor and cognitive behavior in rodent models that involve a variety of paradigms and equipment. To specifically analyze forelimb and hindlimb motor function, the cylinder task and beam-walk task have been implemented to detect motor deficits and monitor locomotion in this model. Other sensitive tasks designed to specifically assess fine forelimb motor skill (i.e. the staircase task and single pellet reaching task) require food deprivation11,13,14. To avoid any of the known effects of food deprivation on disease pathologies15,16,17, these tests have been deemed unsuitable for this study. The cylinder task assesses the spontaneous use of the rat's forelimbs during rearing in a novel environment and can detect asymmetry between forelimbs in rats with unilateral stroke10,18. A major benefit to this task is that the apparatus can be utilized for other behavior tasks, such as the Porsolt forced swim task19. Contrary to the cylinder task, the beam-walk task also allows analysis of hindlimb and forelimb motor control, in addition to locmotion10,14. Beam walking includes a locomotor component, a balance component and skilled foot placement. Both of these tests are cost-efficient, straightforward, and time-efficient and elucidate the effects of stroke and AD on differences in limb functioning.
Aside from changes in motor function, AD involves memory deficits that can present in early stages of the disease progression. When addressing AD-like pathologies in a rodent model, it is crucial that hippocampal dependent learning and memory is assessed because the hippocampus is an important brain structure largely affected in AD2. The hippocampus is an essential brain region for spatial learning and memory and its function can be tested using various maze paradigms in rodents. One of the mostly widely used maze tasks for rodent models of different diseases is the Morris water maze20. The Morris water maze utilizes spatial cues to assist the rat in locating a stationary hidden platform and tests spatial reference memory when the platform is removed. A valuable advantage of the water maze setup is that it is highly adaptable depending on the proposed research question20.
For the first time, the techniques described have been used to assess motor and cognitive function in a novel co-morbid rat model of stroke and AD. involving small strokes in an APP21 transgenic rat model. Co-morbidity was achieved by inducing vasoconstriction of the blood vessels in the striatum to produce a small stroke in APP21 transgenic rats. This stroke model has been well established as a co-morbid condition in an alternative rat model of AD11. Advancement into this novel APP21 transgenic rat model was intended to produce a more translationally valuable model. While the behavioural tasks are described using a co-morbid stroke and AD rat model, these tasks can be further applied to other models of stroke or models of other neurological diseases (i.e. Parkinson’s disease). The general methodology described will be widely applicable to these other disease states, but behavior timelines and paradigms may require alteration based on the proposed research question and model. In addition to being adaptable, the tasks described are effective in demonstrating minor deficits, while also being cost and time-efficient.
The appropriate institutional animal ethics committee should approve all of the behavioral procedures prior to starting experimentation. All animal work described here was approved by Western University Animal Use Subcommittee and follows the Canadian Council on Animal Care guidelines. These animal experiments were performed during the light phase.
1. Cylinder Task for Gross Forelimb Motor Assessment
2. Beam-walk Task for Gross Motor Assessment
3. Morris Water Maze for Hippocampal-dependent Spatial Learning and Reference Memory
The behavioral tasks described were used to demonstrate the effects of stroke in an APP21 transgenic rat model of Alzheimer's disease. The combination of stroke and the APP21 transgene is expected to result in greater motor deficit in the affected limbs, as well as increased memory deficits.
The cylinder task assessed gross forelimb motor function and is represented as the use of the affected forelimb. Furthermore, the beam-walk task was used to specifically assess hindlimb motor function and locomotion. Since the stroke was induced in the right striatum, the left forelimb is expected to show a motor deficit if one is present. The data presented in both Figure 2 and Figure 3a and 3b does not statistically demonstrate that co-morbid rats have a forelimb or hindlimb deficit, respectively. While these animals do not appear to have forelimb or hindlimb motor deficits, they do appear to have minor differences in motor function pertaining to locomotion. In the beam-walk task total steps were significantly increased in the co-morbid transgenic rats with stroke (p <0.05, Figure 3c), suggesting that the beam-walk task is sensitive enough to pick up minor changes in gait and locomotion. The small striatal stroke model used produces small strokes that are likely too small to produce any major motor deficits, but deficits have been demonstrated before in other stroke models with these two tasks10,14. Here, these tasks can simply monitor motor function and locomotion when investigating the parameters presented.
Hippocampal dependent spatial learning and reference memory can be effectively assessed using Morris water maze. There were no apparent differences in learning between groups (Figure 5a, Figure 5b), therefore differences in learning cannot account for any differences in memory performance. APP21 transgenic rats with stroke demonstrated a robust long-term reference memory deficit compared to transgenic rats without stroke and wild type rats with stroke (p <0.05, Figure 5c).
Cued learning is completed to ensure that rats have equal ability to use visual spatial cues to locate the platform in the Morris water maze. As demonstrated in Figure 6a and Figure 6b, no differences were observed in the latency or path length to reach the platform during cued learning between groups. Furthermore, the average swim speed was consistent between groups (Figure 6c) and demonstrates that the motivation to escape and swim abilities was equal between groups.
Figure 1: Timeline for motor and cognitive behavior assessments. Stroke-inducing surgery day is assigned day 0 and all testing days are in reference to this day. (A) Pre-surgery and post-surgery testing for cylinder task (C) and beam-walk task (BM). (B) Morris water maze spatial learning, probe testing and cued learning. Please click here to view a larger version of this figure.
Figure 2: Cylinder Task. Percent use of affected forelimb was calculated using the equation in step 1.3.4 and standardized to day -3 baseline values. The red dotted line annotates the 1.0 value that represents equal use of each forelimb in the cylinder task. Wild type is abbreviated as WT and transgenic is abbreviated as TG. Animal numbers are as follows: WT + saline (n = 7), WT + stroke (n = 8), TG + saline (n = 8), TG + stroke (n = 6). All values are presented as mean ± SEM. (Two-way ANOVA, Tukey's post hoc). Please click here to view a larger version of this figure.
Figure 3: Beam-walk task. The sum of all slips made with the (A) affected and (B) unaffected hindlimb across 5 trials is presented as a ratio of the total number of steps taken to cross the beam. Values were standardized to baseline values as follows: post-surgery ratio – pre-surgery ratio. (C) The total steps to cross the beam on day -7 and day 21. Wild type is abbreviated as WT and transgenic is abbreviated as TG. Animal numbers are as follows: WT + saline (n = 6), WT + stroke (n = 6), TG + saline (n = 6), TG + stroke (n = 5). All values are presented as mean ± SEM and asterisks indicate statistical significance. (One-way ANOVA, Tukey's post hoc, p <0.05). Please click here to view a larger version of this figure.
Figure 4: Program view demonstrating sections required for setup and running an experiment. A few highlighted features of the tracking program used for Morris water maze. The apparatus view with the video of the pool featured above will only appear as presented when in the Tests tab. Please click here to view a larger version of this figure.
Figure 5: Spatial learning and reference memory in Morris water maze. Spatial learning was measured by (A) latency and (B) path length to reach the platform. (C) Reference memory was measured as percent change of latency to first entry into the target zone on the day 19 probe test compared to the day 12 probe test. Wild type is abbreviated as WT and transgenic is abbreviated as TG. Animal numbers are as follows: WT + saline (n = 8), WT + stroke (n = 7), TG + saline (n = 7), TG + stroke (n = 8). All values are presented as mean ± SEM and asterisks indicate statistical significance. (Two-way ANOVA for spatial learning and one-way ANOVA for probe testing, Tukey's post hoc, p <0.05). Please click here to view a larger version of this figure.
Figure 6: Cued learning in Morris water maze. (A) Latency and (B) path length to reach platform are stated as an average of all eight trials of learning. (C) Swim speed is presented as an average of all eight trials of learning. Wild type is abbreviated as WT and transgenic is abbreviated as TG. Animal numbers are as follows: WT + saline (n = 8), WT + stroke (n = 7), TG + saline (n = 7), TG + stroke (n = 8). Values are presented as the mean ± SEM. (One-way ANOVA, Tukey's post hoc). Please click here to view a larger version of this figure.
The combination of stroke and Alzheimer's disease results in very distinct behavioral pathologies that can affect both motor and cognitive function depending on the severity of each condition. Thus, it is necessary to make use of a variety of behavioral tasks to determine the individual contributions of these conditions, as well as give some insight into the combined and potentially interactive effects in the co-morbid condition. The data presented demonstrate three cost-efficient, time-efficient and sensitive behavior tasks to assess motor function and hippocampal dependent spatial learning and reference memory in a novel co-morbid APP21 transgenic rat model with stroke. In addition to the data presented, these tasks have been verified in more severe stroke models10,14,18, as well as in models of AD8,16 and should be widely applicable to various models of both diseases.
That being said, no task is without limitations. For the motor tasks specifically, some troubleshooting may be required if rats become habituated to the cylinder and the beam. In the cylinder task, a motivation may be required to ensure the rat achieves the appropriate amount of forelimb contacts with the wall of the cylinder. To achieve this result, the application of a non-toxic scent on the perforated lid or wall of the cylinder can motivate the rat to rear and contact the wall of the cylinder with its forelimbs. For example, a little bit of peanut butter or vanilla extract can be smeared on the inside wall close to the top of the cylinder to further encourage sedentary rats to explore the cylinder walls. A ring of coloured tape can also be applied on the inside of the cylinder ¾ from the bottom of the cylinder. Another way of promoting rearing involves removing a sedentary rat and re-introducing it to the cylinder after some time has passed, which has been achieved with an inter-trial interval in the protocol. Furthermore, depending on the height of the cylinder compared to the length of the rat, removing the perforated lid halfway through a trial can prompt rearing and forelimb contact with the cylinder wall. This should not be done if there is a possibility the rat will be able to escape from the cylinder during testing. These motivations may help increase the number of rears, but should not influence the use of the left and right forepaws during wall contact. If there is still a major concern about an inadequate number of forelimb contacts after implementing these suggestions, a trial can be ended after the rat has made a total of 10 rears regardless of the length of time. This would be different from the 5 minute trials completed in the protocol above, as each rat would reach a total of 10 rears at different times. With a model of stroke and AD co-morbidity, it is important to be aware that changes in cognitive performance, anxiety and activity levels may develop in these rats. While such changes do not directly affect the primary outcome of the cylinder task (spontaneous forelimb use during rearing), recording other observations may constitute valuable evidence for non-motor-related behavioral changes that may develop. In the cylinder task, such behaviors can potentially be observed by analyzing increased or decreased rearing behavior, as rearing is a form of motivated exploratory behavior. Furthermore, other measures such as time spent grooming, turning and landing can provide insight on anxiety and other physical deficits, respectively.
In regards to the beam-walk task, using the home cage enrichment tubing is usually enough motivation for the rat to walk across the beam. If a rat continues to stop midway along the beam, a fragrant food reward or treat, such as peanut butter or precision sugar pellets, could be introduced in addition to the enrichment tubes at the opposite end. If treats are given, ensure that all rats receive about the same amount of treats on testing day, independent of their performance. Additionally, rather than running all six trials for a rat before moving onto the next rat, trials could be staggered. For example, the first two trials could be completed for all rats before starting the next two. This may prevent the increased number of stops midway across the beam that can occur on later trials when some rats become too comfortable with the beam environment. Another possible solution if frequent stopping occurs is to decrease the number of trials. This could be implemented by placing a mirror behind the beam to analyze both left and right limbs in each trial, thus allowing the number of trials to be decreased. Furthermore, habituation to the beam may occur after multiple testing sessions and can result in rats refusing to cross the beam and remain seated on the beam, despite all efforts to motivate the animal. Due to this problem, beam assessments are not ideal for repeated testing in long-term experiments. In protocols that utilize repeated testing, compensation can also become a concern in addition to motivation. To overcome issues of compensation while crossing the beam, a tapered beam can be used rather than a regular wooden beam21.
Again, cognitive changes, including increased anxiety and changes in overall activity levels and motivation, may occur in this animal model. Therefore, it is important to take note of any irregularities among the experimental groups regarding the animals' motivation to cross the beam as well as their speed and non-locomotor behavior (stopping, sitting, trembling, orientation) while traversing the beam.
Behavioral Alzheimer's disease research requires short-term and long-term memory testing, which has been achieved here using the Morris water maze. Many protocols consider 24 hr after spatial learning to be long-term memory21, but with this protocol a timeline of 24 hr after spatial learning is considered short-term memory and one week is considered long-term memory. During this one-week time period in-between probe trials, the rats should not be exposed to other behavioral tasks or unnecessary stressors to avoid interference with Morris water maze spatial reference memory.
In regards to spatial learning, the experimenters should be persistent in requiring the rat to sit on the platform without jumping for a certain amount of time. If the rat successfully sits on the platform but jumps off as the experimenters come into sight, the rat should be placed or guided back to the platform and required to sit on the platform until the experimenter retrieves the animal. To enforce the learning of the platform as the only way to escape the water, rats should not be picked up while swimming, unless in a probe trial setting. If rats manage to jump from the platform onto the rim of the pool, consider moving the platform further towards the center of the pool to discourage learning of an alternate route of escape.
To ensure there are no spatial biases during spatial and cued learning, the start positions and cued platform positions for each rat in a treatment group should be randomized. Each treatment group should have a representative number of rats starting from each starting position for spatial learning or following the same start and platform position paradigm for cued learning. For assigning start positions, Vorhees and Williams present a very detailed set of randomized start positions for spatial learning and randomized start and platform positions for cued learning20. This can be used directly or as a guideline to assign positions to each rat prior to beginning Morris water maze testing.
For the analysis of Morris water maze data, the suggestions in the protocol above represent how the data presented here was acquired. The data collected in step 3.6 of the above Morris water maze protocol can be used to calculate various outcome measures that may be useful to describe a cognitive deficit. For example, beyond the numerical data set, tracking software also offers the experimenter the opportunity to analyze track plots, which can give additional insights into the animals' search strategies. Additionally, the percentage of time spent or distance traveled within the target quadrant in relation to all quadrants can be used as a measure of reference memory in probe trials. It is important to keep in mind that these co-morbid animal models may exhibit some motor deficits. Comparing the swim speed among the experimental groups can give an indication whether a potential motor deficit is impacting the animals' ability to perform in the Morris water maze. Furthermore, to exclude motor performance from becoming a confounding factor in water maze outcome, it is recommended to look at path length to reach platform in addition to latency and swim speed as demonstrated in Figure 5 and 6. If swimming ability is compromised in any way, path length is the most accurate measure to assess hippocampal function.
While there are various different timelines that can be applied to all of these behavioral tasks described, the methodology of running each experiment should remain the same. The data presented here was achieved with a 21 day post-stroke recovery time point, which captures the early events after stroke and their potential interaction with Aβ metabolism in the brain. While the data presented here was in a co-morbid model of stroke and Alzheimer's disease, these tasks can be applied to or adapted to suit various research questions and models. While the cylinder test is a less-modifiable standard procedure, the beam walk task can be somewhat adjusted to the severity of the expected motor deficit by choosing appropriate beam widths. The water maze is the most versatile test of all paradigms mentioned herein. For example, choosing various intervals between the end of the acquisition phase and the spatial reference memory probe trial can easily test short-term and long-term memory. Working memory and strategy shifting, two components of executive function, can also be tested using the water maze setup. For assessing working memory, the inter-trial interval can be reduced to under 1 min during acquisition learning of a new platform location. Furthermore, having animals learn a second, new location of the platform after successful acquisition of an initial platform location can test mental flexibility or strategy shift. Considering all of these potential modifications, there is a lot of flexibility with these behavior tasks, which is another major benefit in addition to all the aforementioned benefits.
The authors have nothing to disclose.
This work was made possible with funding from the Canadian Institutes of Health Research (CIHR) and the authors would like to thank CIHR for their funding support.
Cylinder | Western University | Plexiglas Cylinder Cylinder Diameter: 23 cm Cylinder Height: 40 cm Platform Height: 30 cm Mirror Length: 35 cm Mirror Width: 26.5 cm |
This was made specifically by Western University Machine Services for our lab. Please contact your own chosen manufacturer to design this product. |
Handheld Video Camera | JVC | GZ-E200 | |
Video Camera Tripod | Slik | F163 | |
AM/FM Clock Radio | Sylvania | SCR1388 | |
White Board | Walmart | Width: 71.12 cm Height: 55.88 cm |
These can be purchased at any store (i.e. Walmart, University bookstores) |
Dry-erase Marker | Expo | Expo Dry-erase Original Marker | These can be purchased at any store (i.e. Walmart, University bookstores) |
Wooden Beam | Rona | Plywood Width: 2 cm Length: 100cm |
|
VWR General Purpose Laboratory Tape | VWR Intl. | 89097-920 | |
iMovie '11 | Apple Inc. | Version 9.0.9 (1795) | This is the version used in this manuscript, but any other iMovie version or video editing software could be used. |
Morris Water Maze Pool with Platform | Stoelting Co. | 60136/60035 | These are not the exact products used in the video, but these are essentially identical. |
Platform Cue | – | – | The platform cue used was created using a small metal stand and white spherical foam ball. These can likely be purchased at any store with home improvement materials (i.e. Walmart, Rona etc.) |
Mainstays 71" Floor Lamp | Walmart | HW-F0377SLV | |
ANY-Maze Behavioural Tracking Software | Stoelting Co. | 60000 | There are ANY-maze® bundles that include the camera with or without a computer and accessories. |
Compact Video Camera | Logitech | V-U0023 | |
Laptop | Hewlett-Packard | HP® Pavilion dv6 Notebook PC | Laptop Specifics: AMD A6-3420M APU with Radeon™ HD Graphics 1.50 GHz, 6.00 GB RAM, 64-bit operating system. |
Americana Non-toxic Acrylic Paint | DecoArt | DAO67-9 | This can be ordered on the DecoArt site or purchased in store at DecoArt® retailers. |
Poster Board | Walmart | PA-1961 |