Abstract
Recently, several capsular stroke models have been reported with different targets of destruction. This study was performed to establish an accurate internal capsule (IC) target for capsular stroke modeling in rats. We injected adeno-associated virus serotype 5 (AAV)-CaMKII-EYFP into forelimb motor cortex and AAV-CaMKII-mCherry into hindlimb motor cortex (n = 9) to anterogradely trace the pyramidal fibers and map their somatotopic distribution in the IC. On the basis of the neural tracing results, we created photothrombotic infarct lesions in rat forelimb and hindlimb motor fiber (FMF and HMF) areas of the IC (n = 29) and assessed motor behavior using a forelimb-use asymmetry test, a foot-fault test, and a single-pellet reaching test. We found that the FMFs and HMFs were primarily distributed in the inferior portion of the posterior limb of the IC, with the FMFs located largely ventral to the HMFs but with an area of partial overlap. Photothrombotic lesions in the FMF area resulted in persistent motor deficits. In contrast, lesions in the HMF area did not result in persistent motor deficits. These results indicate that identification of the somatotopic distribution of pyramidal fibers is critical for accurate targeting in animal capsular stroke models: only infarcts in the FMF area resulted in long-lasting motor deficits.
Introduction
Human stroke disorders can be modeled in small animals to help understand the pathophysiology of stroke and to develop novel rehabilitative and therapeutic strategies for stroke recovery. In animals, cortical infarct models involving gray matter in motor and somatosensory cortices have been used extensively. However, in human stroke, white matter is frequently involved following the occlusion of cerebral arteries. The degree of damage in the white matter influences the severity of motor impairment and the clinical outcome,1–3 especially when the damage is to the posterior limb of the internal capsule, which contains the pyramidal tract. Nonetheless, only a few studies have investigated subcortical capsular infarct, owing to a lack of pertinent rodent models. Occlusion of the internal carotid artery or anterior choroidal artery generated an infarct lesion in and around the internal capsule, but long-term follow-up studies showed a failure to produce persistent motor deficits in this model.4,5 Several recent reports have described selective destruction of capsular fibers in the posterior limb of the internal capsule (PLIC), leading to long-lasting motor deficits and histological findings compatible with white-matter stroke.6–8 The essence of these motor-deficit stroke models is accurate destruction of part of the corticospinal tract at a specific target level; however, although these PLIC models all produced motor deficits, they used very different stereotactic coordinates to target the internal capsule. This inconsistency in stereotactic coordinates may be ascribed to a lack of knowledge of the somatotopic organization of the corticospinal tract in the internal capsule.
The corticospinal tract conveys motor signals from higher cortical structures to the peripheral musculature. In the human corticospinal tract, pyramidal fibers are organized somatotopically and the somatotopic mapping of the human internal capsule was a focus of debate for several decades. 9 Recent studies using diffusion tensor magnetic resonance imaging (DTI) and diffusion tensor tractography (DTT) have produced relatively unanimous results regarding the location of the corticospinal tract in human PLIC, and the locations of the hand and foot areas within the PLIC.10–12 However, there are few studies on the somatotopic organization of the corticospinal tract in rats, specifically the organization of forelimb and hindlimb areas within the internal capsule. Thus, there is a very real possibility of incorrect targeting in rat capsular infarct models. 13 To generate a capsular infarct model that has long-lasting motor deficits while avoiding unwanted neurological sequelae, selective destruction of the pyramidal tract is critical. For example, for a rat stroke model to produce forelimb motor deficits, the infarct lesion must accurately target the forelimb area in the internal capsule. Otherwise, the desired motor deficit may not occur or be long-lasting, despite a large area of destruction in the PLIC, and unwanted neurological and behavioral symptoms may occur instead. To address these issues, we localized the forelimb and hindlimb areas in the internal capsule by injecting adeno-associated-virus–fluorophore constructs into the forelimb and hindlimb areas of motor cortex and anterogradely traced the pyramidal tract to the internal capsule. We subsequently created selective capsular infarct lesions in the identified forelimb and hindlimb areas and observed the patterns of behavioral deficits. Our results suggest that targeting based on neural tracing is a useful strategy for generating a consistent and accurate rodent capsular infarct model that produces persistent motor deficits.
Materials and methods
Animals
Animal experiments were performed according to GIST (Gwangju Institute of Science and Technology) institutional guidelines and all procedures were approved by the Institutional Animal Care and Use Committee at GIST. ARRIVE guidelines were followed in the preparation of the manuscript. Rats were housed two per cage in a controlled animal facility at 21 ± 1 ℃ with water ad libitum. The animal care unit was maintained on a 12-h light–dark cycle with lights on at 7:00 am.
Thirty-eight Sprague-Dawley male rats (300–400 g) were used in this study. Nine rats were used for neural tracing of pyramidal fibers and 29 rats were used for capsular infarct lesioning and subsequent behavioral evaluation (10 for the forelimb-target group, 10 for the hindlimb-target group, and 9 for the control group).
Neural tracing of pyramidal fibers
Neural tracing of forelimb and hindlimb pyramidal fibers was conducted with adeno-associated virus serotype 5 (AAV5)-CaMKII-EYFP and AAV5-CaMKII-mCherry. Rats were anaesthetized with an intramuscular injection of ketamine hydrochloride (100 mg/kg) and xylazine (7 mg/kg), then head-fixed in a small-animal stereotactic frame. Rectal temperature was maintained at 37 ± 0.5 ℃. Our previous experiments have shown that forelimb and hindlimb motor responses can be evoked by electrical stimulation of rostrolateral and caudomedial motor cortex, respectively,
14
so we selected these areas as our targets for cortical viral injections. Small holes were drilled through the skull above the forelimb (AP + 2.0 from bregma; ML 2.5 from the midline) and hindlimb (AP −2.0 from bregma; ML 1.5 from the midline) areas of motor cortex (Figure 1). AAV5-CaMKII-EYFP (virus core facility, KIST, Seoul, Korea) was used for neural tracing of the forelimb pyramidal fiber and AAV5-CamKII-mCherry (virus core facility, KIST, Seoul, Korea) was used for tracing of the hindlimb pyramidal fibers. One microliter of AAV virus was stereotactically injected into layer 5 of the motor cortex at a rate of 0.1 µl/min with a 30G Hamilton syringe connected to an UltraMicroPump (WPI, Sarasota, FL, USA). After injection, the needle was left in place for an additional 10 min before being slowly retracted. After the scalp wound was secured, rats were released from the stereotactic frame and transferred to a recovery chamber with ketoprofen (2 mg/kg, i.m.) for pain control. After waiting 3 weeks for viral expression, the rats were sacrificed and transthoracic cardiac perfusion was performed with 1% paraformaldehyde (PFA) solution for 5 min followed by 4% PFA for 25 min. Then, the brains were removed and further fixed by immersion in 4% PFA for 24 h before immersion in a 30% sucrose solution until saturated. Brains were coronally (n = 4) or sagitally (n = 5) sectioned at 40 µm thickness and mounted onto slides. Images were acquired and analyzed with a DM3000 microscope (Leica, Germany) and iSolution DT Software (OEM-optical.com, Roseville, CA, USA). EYFP expression and mCherry expression were drawn manually onto the stereotactic atlas to show the relative locations of the pyramidal fibers. These drawings were independently confirmed by three reviewers.
Schematic diagram showing the location of the virus injection sites in the motor cortex. Green represents AAV-CaMKII-EYFP injection into forelimb motor cortex and red represents AAV-CaMKII-mCherry injection into caudal hindlimb motor cortex.
Photothrombotic infarct lesioning
On the basis of the neural tracing results, we performed photothrombotic infarct lesioning in the forelimb area and the hindlimb area of the PLIC using techniques similar to those previously described.8,15 Briefly, after head-fixation in a stereotactic frame, a scalp incision was made along the midline. After drilling a small hole in the skull and performing a durotomy, an insulated optical fiber was inserted into the forelimb area of the PLIC (AP −2.04 from bregma; ML ± 3.0 from the midline; DV −7.8; n = 10) or the hindlimb area of the PLIC (AP −2.04 from bregma; ML ± 3.8 from midline; DV −6.2; n = 10). Rose bengal dye (20 mg/kg) was injected into the tail vein followed by 1.5 min of laser irradiation (532 nm green laser) through the insulated optical fiber, with a laser light intensity of 3.7 mW at the tip of the fiber. The sham group (n = 9) received identical treatment except that normal saline was injected instead of rose bengal dye. After the scalp wound was secured, rats were released from the stereotactic frame and transferred to a recovery chamber with ketoprofen (2 mg/kg, i.m.) for pain control. The 29 rats that underwent capsular infarct lesioning were sacrificed after behavioral evaluation and processed for histological staining as described previously. 16 Serial sections were Nissl stained to measure the extent of the lesions. In addition, the sections were immunostained for glial fibrillary acidic protein (GFAP) and myelin (not shown here).
Behavioral testing
Three types of behavioral tests were used in this study: the forelimb-use asymmetry test (cylinder test) and the foot-fault test (ladder-rung walking test) were used to assess unskilled motor behavior and the single-pellet reaching task (SPRT) was used to assess skilled motor behavior. All tests were carried out 6 days/week for 2 weeks after the infarct lesioning.
In the forelimb-use asymmetry test, rats were placed in a cylinder for 2 min to assess the frequency of usage of the ipsilesional and contralesional forelimbs to support an upright body posture. The test score was calculated as the percentage of ipsilesional and contralesional forelimb usage in each session.
17
For the foot-fault test, subjects were placed on an elevated ladder and motivated with food reward to transverse a 1 m grid with irregularly spaced openings. The 46 bars, each 10 cm long, were spaced irregularly (1 to 3 cm) in order to avoid habituation. Videotape analysis was used to calculate the forelimb and hindlimb contralesional error rate:
The foot-fault score was calculated separately for the forelimb and hindlimb and performance was assessed 6 days/week for 2 weeks.
We used SPRT training to assess skilled motor behavior, as previously described.
18
Rats were food restricted to 90% of the starting body weight to motivate them to complete the task. Rats were placed inside a clear Plexiglas box (40 cm × 45 cm × 13.1 cm wide) and trained to reach for a sucrose pellet (Bio-Serve, Frenchtown, NJ, USA) before (∼1 week) and after (6 days/week for 2 weeks) infarct lesioning. Subjects were given 20 pellets per session and each trial was considered successful only if the rat used the contralesional paw to reach for, grasp, and place the pellet at its mouth without dropping it. The performance was evaluated as the following score:
Statistical analysis
Data were analyzed with Origin Pro v9.1 software (Origin Lab, Northampton, MA, USA). One-way analyses of variance (ANOVAs) were used to assess the effects of group and time in the forelimb-use asymmetry, SPRT, and foot-fault data. Tukey's post hoc comparison were performed to assess where significant differences occurred between groups. p values < 0.001 were considered significant.
Results
Neural tracing of forelimb and hindlimb motor fibers (HMFs)
In this study, we used a viral anterograde tracing technique to examine the course and relationship of motor fibers originating from forelimb and hindlimb areas of motor. Neural tracing revealed that the area of cortical AAV-EYFP expression (forelimb cortical injection site) was confined to the motor cortex (Figure 2(a) to (c)). However, the area of AAV-mCherry expression (hindlimb cortical injection site) was located predominantly in motor cortex but also included a small area of sensorimotor cortex (Figure 2(d) to (f)) because the injection site was close to the caudal pole of motor cortex. Despite this, the expression of AAV-mCherry largely represented the HMFs.
Serial sagittal sections of rat brain showing viral expression. (a–c) show the course of the forelimb motor fibers (FMFs), (d–f) show the course of the hindlimb motor fibers (HMFs), and (g–i) show the merged images. The white lines indicate the position of the internal capsule and numbers indicate the distance from midline. IC: internal capsule.
After leaving the cortex, the forelimb motor fibers (FMFs) traveled anterogradely and obliquely in a craniocaudal and dorsoventral direction, forming the corona radiate (Figure 2(b) and (c)), whereas HMFs formed a descending tract with a steeper angle. The two types of motor fibers did not overlap while descending from the cortex through the striatum but the thalamic branches did partially overlap (Figure 2(g) to (i)). After reaching the PLIC, the FMFs occupied the ventral portion of the PLIC and the HMFs occupied the dorsal portion of the PLIC, with partial overlap in between (Figure 3). Interestingly, the FMFs occupied the full thickness of the capsular tract in anterior PLIC, whereas the HMFs occupied only the medial part of the PLIC dorsal to the location of the FMFs (Figure 3). The dorsolateral portion of the PLIC lacked forelimb and HMFs.
Serial coronal sections of rat brain showing viral expression (a–e) and schematic diagrams showing the distribution of forelimb motor fibers (FMFs, green) and hindlimb motor fibers (HMFs, red) in the posterior limb of the internal capsule. Numbers indicates the distance from bregma. IC: internal capsule; TH: thalamus; EP: entopeduncular nucleus; opt: optic tract.
Extent of the photothrombotic capsular infarcts
Serial sections showed the infarct lesions were circumscribed in the PLIC with minimal invasion of neighboring tissues such as the thalamus, hypothalamus, and striatum. The HMF infarct was located dorsal to the FMF infarct and had a more elliptical shape (Figure 4). The mean infarct volumes, including the areas of necrosis and surrounding demyelination, were 1.67 ± 0.64 mm3 for the FMF lesion and 1.29 ± 0.37 mm3 for the HMF lesion. The FMF infarct invaded almost the full thickness of the ventral PLIC and partially extended into the HMF area, and the HMF infarct involved the mediodorsal portion of the PLIC as well as part of the FMF area (Figure 5). Further, the FMF infarct partially invaded the dorsal portion of the entopeduncular nucleus but the HMF infarct did not (Figure 5).
Serial coronal sections of rat brain showing the extent of capsular infarct (arrows) after lesioning of the forelimb motor fiber area (a) or the hindlimb motor fiber area (b). Note that the circumscribed destruction of capsular fibers was confined to the internal capsule without encroaching into the neighboring striatum or thalamus. Numbers indicates the distance from bregma. Relationship between the extent of forelimb motor fiber (black-dashed line) and hindlimb motor fiber (red-dashed line) infarcts, and the location of the entopeduncular nucleus (EP). (a–c) show the infarct lesion of forelimb motor fibers and (d–f) show the infarct lesion of hindlimb motor fibers. (g–i) shows the relationship between the neural tracing of motor fibers (FMFs, green and HMFs, red), the extent of infarct, and the location of EP. Note the partial invasion of dorsal EP by the FMF infarct. Numbers indicates the distance from bregma. IC: internal capsule.
Behavioral outcomes following capsular infarctions
Figure 6 shows how the postlesional behavior depended on the location of the infarct lesion. FMF infarcts resulted in a significant decline in both skilled and unskilled behavioral performance, deficits that persisted throughout the testing period. All the behavioral tests, including the SPRT, cylinder test, and foot-fault test, showed a significant decrease in sensorimotor function for the contralesional limb compared with the sham-operated and HMF-infarct groups. However, HMF-infarct animals did show a significant decline in performance in the SPRT task during the first week post-infarct, compared with sham animals, but this motor deficit was not seen in the second week post-infarct (Figure 6(a)). The FMF infarct also impaired motor performance of the contralesional hindlimb over 10 days post-infarct, as shown by the results for the foot-fault test; this was likely due to the involvement of the HMF area in the FMF infarct (Figure 6(c) and (d)). In contrast, the HMF infarct did not show a significant decline in the performance of unskilled behaviors for the contralesional limb, as tested by the cylinder and foot-fault tests (Figure 6(b) to (d)). The sham-operated group had no deficits in behavioral performance on any of the tests, compared with pre-infarct baseline levels.
Behavioral evaluation in the forelimb motor fiber (FMF) area and the hindlimb motor fiber (HMF) area infarct groups, and in the sham-operated group. (a) Single pellet reaching task. The FMF infarct group showed a significant decrease in performance in the contralesional limbs compared with the sham-operated group (*p < 0.001). The HMF infarct group showed a significant decrease in performance compared with the sham-operated group during the first week (
Discussion
This is the first study to compare patterns of rodent behavior after lesioning identified PLIC subregions. Our neural tracing study demonstrated that FMFs and HMFs were located predominantly in the inferior portion of the PLIC, with the FMFs located ventral to the HMFs. Photothrombotic capsular infarct lesioning of the PLIC FMF region successfully resulted in persistent motor deficits in the corresponding forelimb. However, infarct lesioning in the HMF region resulted in only a temporary decline in motor performance, rather than a persistent motor deficit. These results suggest that destruction of the FMF area is ideal for modeling capsular stroke in rodents and should eradicate confusion in capsular targeting.
Our study shows that viral neural tracing is a useful technique for identifying the somatotopic distribution and course of pyramidal fibers from the cortex to the internal capsule. Our tracer findings were consistent with the results of axonal tracing with biotinylated dextran. 19 The pyramidal tract is the most critical white matter structure in which permanent stroke damage is directly related to the severity of the subsequent motor deficit. Although the course and somatotopic distribution of the pyramidal tract have been extensively explored in humans with diffusion tensor imaging and diffusion tensor tracking techniques, similar studies on the anatomy of the pyramidal tract in rodents are extremely rare.10,20 This lack of knowledge has led to differences in the stereotactic coordinates used to generate capsular infarct models in the rat.6–8 Viral neural tracing is thus useful for clarifying this confusion and determining appropriate targets when generating new models for stroke.
Our results shows that capsular infarct lesioning in the PLIC FMF area, as based on neural tracing, can demonstrate persistent motor deficits with high reliability. However, capsular infarct lesioning in the HMF area failed to show a permanent motor deficit. Instead, HMF-area lesions induced an immediate decline in motor function but the motor deficits resolved after 1 week. It is not clear why the motor deficits were less prominent after destruction of hindlimb fibers than forelimb fibers but there are several possible explanations. For example, the assessment of hindlimb motor function is difficult in an animal model because of the limited availability of accurate testing methods.21–24 The SPRT for assessing forelimb motor function tests complex finger dexterity and can shed light on post-stroke deficits and recovery.25,26 In contrast, rats do not display an equivalent skilled hindlimb movement, making it impossible to evaluate the same quality of distal movement in the hindlimb as in the forelimb. At present, assessment of hindlimb motor function is largely based on locomotion, in which deficits can be easily compensated.21,23 Interestingly, discrepancies exist here between human patients and the animal model. For example, human patients with capsular infarct are likely to show hemiplegia or hemiparesis, whereas animals may not show a hindlimb deficit. 3 This discrepancy has both anatomical and physiological explanations. First, the pyramidal tract in rodents has extensive, complex, connections with the striatum, thalamus, subthalamic nuclei, midbrain, brain stem, and spinal cord, whereas the pyramidal tract in humans has simpler connections to fewer other structures, such as the striatum and spinal cord. 27 This is likely to reduce the chance of compensatory motor recovery in humans. Second, pyramidal tract damage in human patients, as in capsular infarct, is commonly accompanied by damage to the corticobulbar tract, which is involved in the programming and execution of voluntary movements. Selective destruction of the corticospinal tract, as in our experiment, is feasible only in animal stroke models.
PLIC FMF-area lesions also involved part of the entopeduncular nucleus (EP) corresponding to internal globus pallidus (GPi) in human, raising a concern that motor deficits from an FMF infarct are in fact related to injury of two distinct anatomical areas: the internal capsule and the EP. Although the GPi has been known to be involved in basal ganglia circuits and to participate in motor coordination in humans, the role of the EP in motor function has been shown to be minor in rats.28,29 Unilateral lesion of the EP does not affect locomotor function, motivation, motor skills, sensorimotor gating, or learning and memory. 29 Therefore, involvement of the EP in the capsular infarct is unlikely to affect the severity of the motor impairment. Instead, we should heed damage to neighboring gray-matter structures such as the thalamus and striatum. For example, the FMF infarct area is located near thalamic motor nuclei, such as the ventral anterior and ventral lateral nuclei, 30 and near the striatum, which is a key structure in the basal ganglia circuit that mediates motor function, cognition, executive function, and learning. Any injury to these structures may influence the integrity of the motor circuit.
We previously reported using the photothrombotic technique to create a capsular infarct model. This technique cannot reproduce human capsular infarct exactly, but it has great merit over other techniques because it selectively destroys the white matter without damaging the neighboring gray-matter structures. This is because light scattering is more than 4 times higher in white matter than in gray matter, so the use of low laser light intensities prevents light from reaching the neighboring gray-matter structures. 8 The vasoconstrictor endothelin-1 (ET-1) has recently been used to create a capsular stroke model. 6 ET-1 injections can be targeted to the desired site; however, ET-1 is likely to diffuse out into the neighboring gray-matter tissue, creating undesirable damage in thalamic areas. 7 Furthermore, ET-1 may influence the function of neurons and glia through activation of ET-1 receptors expressed on astrocytes and microglia. 31 Therefore, the photothrombotic infarction technique remains the most suitable for modeling chronic capsular infarct in animals.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Basic Science Research Program through NRF of Korea funded by the Ministry of Science, ICT and future Planning (2013R1A2A2A01067890 & NRF-2016R1A2B3009660).
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Authors' contributions
HS and H-IK designed the experiment; HS, WJ, EL, and JYP performed the experiment; MCL, MSK, and HIK analyzed the data; HS and HIK wrote the manuscript.