Acta Neuropathol (2000) 100 : 13–22

© Springer-Verlag 2000

R E G U L A R PA P E R

Mohammad Farooque

Spinal cord compression injury in the mouse: presentation of a model including assessment of motor dysfunction

Received: 2 December 1998 / Revised, accepted: 14 October 1999

Abstract The purpose of this study was to develop a spinal cord injury model in the mouse. Various degrees of extradural compression were used to induce mild, moderate or severe compression injuries. Furthermore, a locomotor rating scale was developed by which the functional outcome of the spinal cord injury could be assessed. The introduction of such a model will be useful for further studies on the pathogenesis and treatment strategies of spinal cord injury. To assess hindlimb motor function, a 10-point scale was used. Initially, the animals were allowed to move freely in an open field and were rated 0–5, 0 being no movement and 5 being almost normal. Animals scoring a 5 were then assessed using steel bars with decreasing widths from 2 cm to 5 mm. For each bar successfully crossed over, they gained additional points. Before injury the hindlimb motor function score (MFS) in all the animals was 10. In mice with mild compression, MFS was decreased slightly on day 1 and recovered to 9 ± 0.6 on day 14. For mice with moderate compression, the MFS decreased to 4.6 ± 0.4 on day 1 after injury and gradually improved to 8.1 ± 0.6 on day 14. Severe injury resulted in paraplegia of the hindlimbs day 1 after injury with a score of 0.6 ± 0.2. By day 14 after injury, these animals gradually recovered to 3.9 ± 0.1, could bear the weight on the hindlimbs and walk with a severe deficit. There was a 3%, 9% and 19% decrease in the total cross-sectional area of the spinal cord 14 days after mild, moderate and severe injury, respectively. Microtubule-associated protein immunostaining revealed that the gray matter decreased to 61 ± 7% in moderately injured animals, while severe compression resulted in a complete loss of gray matter. White matter decreased to 86 ± 6% in moderately injured animals and 29 ±11% in severely injured animals. This study shows that the mouse can be used to achieve repro-

M. Farooque (쾷) Laboratory of Neuropathology, Department of Genetics and Pathology, Uppsala University Hospital, 75185 Uppsala, Sweden e-mail: [email protected], Tel.: +46-18-663838, Fax: +46-18-502172

ducible spinal cord compression injuries of various degrees of severity. The force of the impact correlates well with the neurological and light microscopic outcome. The motor function test presented in this paper and the computerized quantification of tissue damage can be used to evaluate the efficacy of different treatment strategies. Key words Spinal cord injury model · Mouse · Locomotion assessment

Introduction Experimental models of spinal cord trauma have, to a great extent, increased our understanding of the pathophysiology associated with this type of lesion [29, 34]. In the past, studies have used various species including cats, dogs, and rats. For walking behaviour studies, cats have been the animal of choice. The rat is easy to breed and, as such, has been used for molecular and regeneration studies as well as kinematic and behavioural studies [5, 36]. Using a model of compression trauma in rat, secondary injuries such as the formation of swelling, the effects of various drugs on the composition of extracellular edema [7, 9, 10] as well as axonal and dendrite alterations [20, 23] have been studied. In the past, reproducible compression models for inducing graded spinal cord injury in the mouse have been lacking. However, recently, a mouse model of spinal cord injury using a weight-drop technique has been presented [16]. Mice have also been used to study the effect of radiation injuries of the spinal cord [18], wound healing after crush injury [37], and inflammatory response in partially transected mouse spinal cord [31]. Genetically manipulated (transgenic and knockout) mice also appear to be very useful for further investigations on the pathophysiology of secondary injuries following trauma to the spinal cord. Corresponding strains of rats do not exist. The purpose of this study was to develop a spinal cord trauma model in mouse by which injuries of different degrees of severity can be produced. Since studies on the ef-

14

fect of various therapeutic measures on spinal cord trauma require recording of various chemical, morphological and functional outcomes, a locomotor rating scale was developed by which the functional outcome of the spinal cord injury can be assessed.

Materials and methods Female mice (B6CBAF1 hybrids) between 8–12 weeks of age with an average weight of 25 g were used. Food and water were provided ad libitum before and after the experiments. The mice were kept at a temperature of 20 °C controlled by a thermostat and exposed to alternate light and dark periods of 12 h. This study was approved by the Uppsala ethical committee for animal research. Spinal cord compression trauma The animals were anesthetized with a mixture of fluanisone (0.5 mg/ml) and midazolam (0.25 mg/ml) in distilled water. A total volume of 0.5 ml of the mixture was given intraperitonially. A laminectomey of T8 vertebra was performed leaving the dura intact. The animals were placed in a stereotaxic apparatus. Two adjustable forceps were applied to the spinous processes of one vertebra proximal and distal to the laminectomy to stabilize the spinal cord (Fig. 1, 1c). The compression was applied to the spinal cord for 5 min, using a rectangular plate, which was longitudinally oriented over the spinal cord (Fig. 1). The plate was then removed and the skin sutured.

Fig. 1 The compression was applied to the spinal cord for 5 min using a rectangular plate (a and b). The compression device consisted of a 2 × 1-mm rectangular plastic plate at the end of a 4-cm rod. A platform was fitted on its upper part on to which different weights could be placed (main panel, c). The middle part of the rod was held by a steel clamp attached to an arm of a stereotaxic device. The compresion plate was adjusted at the spinal cord with the micro manipulators of the stereotaxic device

The compression device consisted of a 2 × 1-mm rectangular plastic plate at the end of a 4-cm rod (Fig. 1, 1a, 1b). The upper part of the rod was hollow to fit a round plastic platform on to which different weights could be placed (total weight of plate+rod+ platform was 0.4 g and this was added to the weight applied over the platform to calculate the final weight of compression). The middle part of the rod was held by a steel clamp attached to an arm of a stereotaxic device. Using micromanipulators, it was possible to adjust the compression plate at a desired location over the spinal cord. By gently releasing the clamp around the rod which is carrying a weight, the plate was allowed to compress the cord (Fig. 1c).

Postoperative care The mice were kept under a heating lamp until they regained consciousness. No pre- and postoperative antibiotics were given. Bladder function was observed during the first 24 h after trauma to observe for signs of retention. Animals were housed in groups of seven to ten per cage.

Experimental groups First, a number of pilot experiments were performed to choose appropriate weights to induce mild, moderate and severe compression injury. Three different weights were chosen for the final experiments based on functional and the light microscopical outcomes. To induce mild injury, a weight of 2 g/mm2 on the compression surface was used. To produce moderate and severe injury, 5 and 10 g/mm2, respectively, were used.

15 Table 1 The number of animals in the control groups or groups subjected to different degrees of compression injury of the spinal cord and their survival time

Procdure

Intact controls Laminectomy only Compression Compression Compression Transection Resection (2 mm)

Compression

Nil Nil 2 g/mm2 5 g/mm2 10 g/mm2 – –

Twelve other mice (female and male) were used to evaluate the inter-observer mean difference regarding assessment of motor function. Furthermore, the spinal cords of six mice were transected and in four other animals a 2-mm piece of the cord was resected. These mice were allowed to survive 12 weeks. Non laminectomized and laminectomized animals without compression injury served as controls. The groups and the survival times are presented in Table 1. Hindlimb motor function rating A scoring system (hindlimb motor function score, MFS), by which functional outcome can be assessed, was developed. In addition, righting reflex and locomotor activity as described below were tested, but these two determinations were not included as parameters in the scoring system. Hindlimb locomotive ability was evaluated before and on days 1, 4, 7, 10 and 14 after injury. A scale from 0–10 points was created to assess the locomotive ability. First, animals were allowed to move freely in an open field and were rated 0–5 according to the scale presented below. Animals were placed one at a time on a (0.7 × 0.9 m) paper-covered mattress on a table. The first motor score was calculated by observing each animal for 1 min in the open field. If there were no noticeable movements at all. the animals were rated 0. If there were barely visible movements at any hindlimb joints (hip, knee or ankle) the animals were scored 1. The animal scored a 2 if there were obvious movements of one or more hindlimb joints (hip, knee or ankle) in one or both limbs, but no co-ordination, alternate stepping movements or weight bearing were observed. This included movements ranging from only one joint to movements in all the joints of the lower limbs. The animal was graded 3 if there were alternate stepping and forward propulsive movements of the hindlimbs, but no weight bearing. In this case the hindlimbs were externally rotated and the animal used the hind limbs for forward propulsion. There was no plantar placement of the feet. Mice scoring 3 used the tips (nails) of the middle three toes for forward movement. Animals were scored 4 if they were able to bear weight on their hindlimbs and could walk with some deficit. There was plantar placement of the feet. Deficit included slight external rotation of one or both limbs and/or hip instability. If it was not possible to differentiate the injured from a normal animal except reduced mobility they were scored 5. Mice demonstrating normal movement (score 5) were then assessed using steel bars (50 cm long and 3 mm thick) with decreasing widths: 2 cm, 1.5 cm, 1 cm, 7 mm and 5 mm. The mice were required to walk on the bars and the narrowest bar they could traverse without any slips, in at least two trials, was recorded. If the mice, instead of walking, tried to climb under the bar it was considered as a failure, i.e., an inability to walk on the bar. If the animals could walk on the 2-cm bar they were scored 6, on the 1,5-cm bar then 7, on the 1-cm bar then 8, on the 7-mm bar then 9, and on the 5-mm bar then 10 (Table 2). Uninjured animals would walk on a 5-mm-wide bar (score 10) before injury. Some animals were able walk on a 3-mm-wide bar before injury, but this was not a consistent finding. In the initial stages of scale development, wider bars of 3, 4 and 6 cm were used. However, it was found that even animals with mild deficit

Injury grade

– – Mild Moderate Severe – –

Survival time 0h

2 days

2 weeks 12 weeks

3 – – – – – –

– 3 3 3 3 – –

– 7 9 8 9 – –

– – – – – 6 4

Table 2 Hindlimb motor function scoring system of mouse 0 No movement of the hindlimbs 1 Barely perceptible movement of any hindlimb joints (hip, knee, or ankle) 2 Brisk movements at one or more hindlimb joints (hip, knee, or ankle) in one or both limbs but no co-ordination 3 Alternate stepping and propulsive movements of hindlimbs but no weight bearing 4 Weight bearing and can walk with some deficit 5 Normal walking 6 Normal walking and can walk on a 2-cm-wide bar 7 Can walk on a 1.5-cm-wide-bar 8 Can walk on a 1-cm-wide-bar 9 Can walk on a 0.7-cm-wide-bar 10 Can walk on a 0.5-cm-wide-bar

(score 4) sometimes could walk on the wider bars and the wider bars were excluded. Repeated evaluations showed that animals with mild deficit (score 4) were unable to cross a 2-cm-wide bar, so this was selected as maximum width. Righting reflex This was assessed by dropping the animals on a cushion. The animals were held upside down, dropped and rated as 0, no righting reflex; 1, attempt to right itself; 2, rights itself during the drop; or 3, rights itself immediately after the drop. Locomotive activity In an open field the mice were observed and graded normal (3 points), reduced (2 points), or minimal (1 point) activity according to their mobility. Tissue preparation 2 or 14 days after compression of the spinal cord, the mice were reanesthetized as described above. The chest was opened and a blunt stainless steel 21-gauge cannula was inserted into the left ventricle after which the right atrium was opened to permit exsanguination. The mice were then perfused with 10 ml of phosphate buffer in saline (PBS) at room temperature followed immediately by 10 ml of 4% PBS-buffered formalin. The spinal cord was excised and placed in the same fixative for 24 h. Transverse sections of spinal cord were taken at the site of compression, proximal and distal to the trauma, and 5 mm away from the site of injury. The pieces were then dehydrated in ethanol and embedded in paraffin. Transverse sections of 5 µm were cut. Additionally, longitudinal sections were prepared from one animal of each group. To keep the spinal cords straight they were placed on filter paper and im-

16 mersed in 95% alcohol overnight. They were then dehydrated in ethanol and embedded in paraffin. To characterize the tissue damage after injury, the sections were stained with hematoxylin and eosin, Luxol fast blue-cresyl violet, and by microtuble-associated protein 2 (MAP2) immunohistochemistry [19]. MAP2 immunohistochemistry After deparaffination, the sections were treated in the following order: Microwave oven for 10 min in citrate buffer pH 6.0, in 1% H2O2 in PBS for 30 min, 20% rabbit serum in PBS for 30 min, and then incubated overnight with a monoclonal antibody against MAP2 (Amersham International, Amersham, UK, code RPN 1194, batch 10) at a dilution of 1: 1,000 in 0.1% rabbit serum. The sections were then exposed to rabbit anti-mouse IgG (Dako) for 30 min, and the reaction product was visualized by the avidin-biotin-peroxidase complex (Dako ABC complex/HRP) method using 3, 3,-diaminobenzidine tetrahydrochloride as the chromogen. To intensify the reaction product we applied the nickel-enhancement procedure combined with the glucose-glucose oxidase method [14, 32]. For control purposes, the primary antibody was omitted and thereafter the sections were treated as those in which the MAP2 antiserum had been applied. Image analysis Total cross-sectional area was measured to determine the extent of atrophy. The cross-sectional area immunostained with MAP2 was calculated using image analysis to evaluate the damage to the gray matter since MAP2 is present only in the nerve cell bodies and dendrites and not in the axons of the longitudinal tracts. The crosssectional area stained with Luxol was also calculated by image analysis to evaluate damage of the longitudinal myelinated tracts. Respective areas in laminectomized animals were taken as 100%. The loss in total area, gray matter, and white matter after injury was described as % of total area stained in laminectomized animals. Statistics Factorial analysis of variance and Fisher’s PLSD for posthoc testing (Statview, Abacus Concepts, Berkeley, Calif.) were applied to compare groups. Analysis of variance repeated measures was used to find overall differences between the groups for motor function evaluation. Linear regression coefficients were calculated for correlation between extent of damage and behavioral evaluation. Differences with a P-value < 0.05 were considered statistically significant. Values in the figures and text are given as the mean ± SEM.

Results After surgery, the general condition of the mice was good and none of the animals developed any obvious infection. Depending upon the degree of compression, the graded outcome was evident from neurological tests and light microscopic examination as described below. Hindlimb motor function rating The hindlimb MFS was ten in all the animals before injury and all could traverse a 0.5 cm wide bar. Some animals were able to walk on a 3-mm wide bar, but this was not a consistent finding. Depending upon the severity of

Fig. 2 Hindlimb motor function was scored before injury and on days 1, 4, 7, 10, and 14 after laminectomy, mild, moderate, and severe injury. An asterisk indicates significant difference between laminectomized animals and other groups. a denotes a significant difference between mild and moderate or mild and severe. b indicates significant difference between moderate and severe

injury, hindlimb motor function decreased substantially (Fig. 2). No animal below score 5 in the open field could walk the widest bar (2 cm). In laminectomized controls, compared to preinjury values, the motor function score was slightly decreased on day 1 to 9.1 ± 0.1 (P = 0.16) and on day 14 after injury it was 9.8 ± 0.2. Mild injury Compared to laminectomized animals, the MFS for mice with mild injury was decreased slightly to 8.2 ± 0.7 (P = 0.1) on day 1 and recovered to 9 ± 0.6 (P = 0.24) on day 14. Compared to preinjury values, the MFS was significantly lower on day 1.

Righting reflex

17

Fig. 3 Hindlimb motor function was scored before injury and on days 1, 4, 7, 10, 14, 28, 56, and 84 after resection and transection. The score remained below 2

Moderate injury MFS decreased to 4.6 ± 0.4 on day 1 after injury and gradually improved to 8.1 ± 0.6 on day 14 (Fig. 2). Compared to laminectomized animals, the MFS was significantly lower from day 1 until day 14. Compared to mildly injured animals, the MFS was significantly lower on days 1, 4, and 7. Severe injury This degree of injury resulted in paraplegia of the hindlimbs 1 day after injury. The hindlimb MFS decreased to 0.6 ± 0.2. The animals gradually recovered to 3.9 ± 0.1 2 weeks after injury. Eight of the nine mice could bear their weight on the hindlimbs and walk with some deficit (score 4). One animal was not able to bear weight (score 3). Compared to laminectomized controls, mildly and moderately injured animals, the MFS was significantly lower from day 1 until day 14. Complete spinal cord transection and resection Complete transection or 2-mm resection of the spinal cord resulted in severe motor functional deficit, and all the animals became paraplegic for the entire observation period of 3 months. The score remained below 2 (no meaningful movements) (Fig. 3). Animals exhibited some movements in the hindlimbs during open field examination; however, there was no co-ordination and no weight bearing. Inter-rater difference The motor functions were evaluated on three occasions after moderate injury by two observers on days 4, 7, and

Fig. 4 Diagrammatic presentation of righting reflex in different groups. The animals were held upside down, dropped, and rated as 0, no righting reflex; 1, attempts to right itself; 2, rights itself during the drop; or 3, right itself immediately after the drop

14. Mean inter-rater difference was 0, 0.08, and 0.08 at first, second, and third occasions, respectively. Righting reflex and mobility The righting reflex remained normal after surgery in laminectomized and mildly injured rats (Fig. 4). Righting reflex was considerably impaired up to day 4 in moderately injured animals, but normalized on day 7. In severely injured mice, the righting reflex was absent 1 day after injury, started to improve on day 4, and recovered to 50% of the normal level after 2 weeks. The righting reflex was not present in animals in which the spinal cord was transected or resected. No effect on mobility was noted in laminectomized animals tested in an open field (Fig. 5). Mobility was slightly reduced on day 1 in mildly injured animals and became normal by day 7. Mobility was decreased to 50% in moderately injured animals 1 day after injury. The mice gradually regained mobility and approched close to normal values after 2 weeks. There was a 65% decrease in mobility in severely injured animals throughout the observation period of 2 weeks. Mobility was severely impaired in transected and resected animals during the observation period of 3 months.

18

mostly macrophages and lymphocytes. The architecture of the spinal cord was distorted. However, the subpial regions of the white matter were often spared (Fig. 6, top row: d). MAP2 immunostaining

Fig. 5 Diagrammatic presentation of mobility in different groups. In an open field the mice were observed and their locomotive activity was graded normal (3 points), reduced (2 points), or minimal (1 point). Asterisks indicate a significant difference between laminectomized controls and injured animals

MAP2 immunostaining was used to evaluate the extent of damage to nerve cell bodies and dendrites. The spinal cord of normal mice exhibited MAP2 staining in the dendrites and nerve cell bodies of the ventral horns, dorsal horns, and central gray area, whereas axons of the longitudinal tracts remained completely unstained (Fig. 6, middle row: a). Image analysis showed that 45 ± 2% of the total cross-sectional area was stained in laminectomized controls. After injury, depending upon the degree of impact, there was a decrease of the immunoreactive area when compared to laminectomized controls. The area stained was 96 ± 8% in mildly injured animals (Fig. 6, middle row: b). In moderately injured animals, there was an extensive loss of immunoreactivity and the area stained was only 61 ± 7% (Fig. 6, middle row: c; Fig. 7). Severe compression resulted in a complete loss of immunoreactivity of the entire cross-sectional area (Fig. 6, middle row: d). There was a significant difference between mild and moderate (P < 0.0006), mild and severe (P < 0.0001), moderate and severe (P < 0.0001) injury groups.

Light microscopy

Luxol staining

Atrophy of the spinal cord was noted 14 days after mild, moderate and, severe injury. Computer-assisted area measurements showed that, when compared to laminectomized animals, there was 3%, 9% and 19% decrease in the total cross-sectional area of the spinal cord after mild, moderate, and severe injury, respectively. No cavity formation in any animals was noted up to 14 days post injury.

The cords were stained also with Luxol fast blue to assess the extent of damage to the white matter. Myelinated tracts of the dorsal, ventral, and lateral columns were stained blue with Luxol in normal controls. Image analysis showed that the total cross-sectional area stained with Luxol was 58 ± 1% in laminectomized controls (Fig. 6, bottom row: a). No decrease in stained area was noted in mildly (Fig. 6, bottom row: b) injured mice. Compared to laminectomized controls in moderately injured animals, there was a considerable loss of luxol staining and the stained area was reduced to 86 ± 6% (Fig. 6, bottom row: c; Fig. 7). In severely injured mice the stained area was only 29 ± 11% (Fig. 6, bottom row: d; Fig. 7). There was a significant difference between mild and moderate (P < 0.04), mild and severe (P < 0.0001), and moderate and severe (P < 0.0001) injury groups.

Routine staining At 48 h after mild injury, there was some swelling at the site of compression as indicated in sections stained with hematoxylin and eosin; otherwise the spinal cord appeared normal. After moderate injury there was extensive swelling and bleeding at the site of compression and the adjacent proximal and distal segments showed many shrunken cells with pyknotic nuclei. Severe injury resulted in extensive swelling and bleeding of the spinal cord at the site of injury at 48 h. At 14 days after mild injury, there were only a few inflammatory cells at the site of injury (Fig. 6, top row: b); however, there were many inflammatory cells in the traumatized part after moderate injury (Fig. 6, top row: c). At 14 days after severe injury most of the spinal cord cross section at the site of injury contained inflammatory cells,

Relations between motor function recovery and morphology A strong correlation was present between hindlimb MFS, righting reflex and mobility. Spared white matter area and spared gray matter also showed a linear correlation with hindlimb MFS, righting reflex and mobility (Table 3).

19

Fig. 6 Top row: Hematoxylin and eosin staining, 14 days survival. The laminectomized control showed few shrunken neurons, otherwise the cord appeared normal (a). There were a few inflammatory cells and macrophages in mice with mild compression (b) and many in moderately compressed cords (c). After severe compression a large central necrosis occurred. However, the subpial regions of the white matter were often spared (d). Middle row: MAP2 immunostaining, 14 days survival. MAP2 immunostaining was used to evaluate the extent of damage to nerve cell bodies and dendrites. The spinal cord of normal mice exhibited MAP2 staining in the dendrites and nerve cell bodies (a). Depending upon the

degree of compression, there was a decrease of the immunoreactive area. In mice with mild compression, there was minimal loss of MAP2 immunoreactivity of nerve cell bodies and dendrites (b). There was an extensive loss of immunoreactivity in moderately injured animals (c), while severe compression resulted in a complete loss of immunoreactivity of the entire cross-sectional area (d). Bottom row: Luxol staining, 14 days survival. The cords were also stained with the Luxol fast blue to assess the extent of damage to the white matter. Depending upon the severity of the injury, there was increasing damage to myelinated tracts, control (a), mild (b), moderate (c) and severe (d) injury

20

Area spared (%)

Laminectomy Moderate Severe

White matter

Gray matter

Fig. 7 Graphic presentation of percentage of area of white and gray matter spared compared to laminectomized controls 14 days after moderate or severe injury. Since no gray matter was immunostained after severe injury, it is not presented in the graph. Luxol was used to stain white matter, while MAP2 was used to immunostain neurons and dendrites

Discussion This study demonstrates that the mouse can be used to achieve reproducible spinal cord compression trauma of various degrees of severity. The force of the impact correlates well with the neurological and light microscopic outcomes, which parallels previous results in a similar model in rat spinal cord [7, 23, 26]. The motor function test presented here and the computerized quantification of tissue damage in MAP2- and luxol-stained sections can be used to evaluate the efficacy of different treatment strategies. The most commonly used method to induce spinal cord injury is the weight-drop technique introduced by Table 3 Correlation between spared gray or white matter at the site of compression vs MFS, righting reflex and mobility, at different days. Correlation between the MSF, righting reflex and mobility is also shown (MFS motor function score)

Gray matter vs MFS White matter vs MFS Gray matter vs righting reflex White matter vs righting reflex Gray matter vs mobility White matter vs mobility Righting reflex vs MSF Mobility vs MSF

Allen in the dog [1], and afterwards adapted to other animals including the rat [4, 6, 13, 28]. In other experiments, the cord is compressed for a designated period of time with a clip [12, 30, 35], a weight [27] or a balloon [33]. For this study, compression was used to induce spinal cord injury for several reasons. The compression model allows production of lesions with different degrees of severity and is a common method used with the rat [9, 15, 21]. The model also mimics a clinical situation, in which a compression is caused by bony fragments or extruded disc material. Using a longitudinally oriented rectangular plate, it is also possible to obtain a large longitudinal region of the cord which has been exposed to the same degree of compression. This enables us to take large numbers of serial sections with the similar magnitude of damage for histology. Furthermore, the perifocal zone is well defined. This region is important for studies of various secondary injuries [7, 8, 19–22, 24]. When comparing the mouse to the rat, with regard to studies on spinal cord trauma, several advantages and disadvantages have been found. Since the mouse is more active than the rat it is easier to perform the behavioral tests applied in this study. However, due to their small size, detailed kinematic analysis is more difficult in the mouse. When using the mouse, computer-assisted kinematic analysis of video tapes is crucial for studying various segments of movements, but such studies are laborious and require expensive equipment. Open field examination was originally introduced in the dog as a method for studying the functional outcome after spinal cord trauma [33], and afterwards modified for investigations in rat [3, 13]. Walking on a bar [17, 35] and grid walking [17] have also been used in rat to evaluate motor function deficit. The test presented here to assess motor dysfunction (MFS) is easy to perform and requires minimal training for the examiner. By testing the animals on the bar, it is possible to detect minor deficits that are not easily detected in open field testing. Furthermore, inter-rater reliability of the MFS scale has been found to be high. Early in the development of the MFS, 2 cm was identified as the bar width of choice as only animals initially scoring 5 were able to walk on it. This was done to preclude the possibility of overlapping the two tests. The degree of functional deficit defined as motor dysfunction, loss of righting reflex, and changes of mobility strongly correlated with the magnitude of white or gray 1 day

4 days

7 days

10 days

14 days

0.87 0.84 0.93 0.90 0.79 0.65 0.95 0.80

0.87 0.88 0.90 0.79 0.87 0.73 0.89 0.84

0.87 0.85 0.88 0.94 0.85 0.79 0.83 0.94

0.82 0.81 0.88 0.94 0.91 0.86 0.81 0.93

0.81 0.77 0.82 0.93 0.91 0.88 0.69 0.76

21

matter loss (Table 3). However, righting reflex and mobility were not used in the motor function scoring to avoid overlapping in locomotor rating. Recently, a mouse model of spinal cord injury was presented [16]. The major difference between our model and that presented by Kuhn and Wrathall [16] is the use of a compression plate versus a weight drop to induce injury. They also applied a variety of behavioral tests to assess motor function and found a correlation between magnitude of injury and histological damage [16]. In their study various functional tests, motor score, platform hang, mesh decent, bar grab and rope walk were strongly correlated to each other. However, in their study, other tests such as righting response, extension withdrawal, pain withdrawal, were not good indicators of white matter sparing [16]. In this study the MFS and righting reflex were strongly correlated with the white or gray matter damage. Cavitation has been reported in the rats after spinal cord injury [25]; however, no cavitation was noted up to 14 days after spinal cord injury in our mouse model. These findings are consistent with mouse model of contusive spinal cord injury [16] in which Kuhn and Wrathall noticed a consistent lack of central cavitation. Similarly, extradural crush injury in mouse also did not produce central cavitation [37]. Thus, the finding that no central cavity develops in mice in our compression model, the contusion model [16] or the crush injury model [37] supports the hypothesis [37] that there is a difference in cellular response in mice compared to other vertebrate models [2, 11]. Thus, the question arises as to whether the mouse model of spinal cord injury is relevant to human spinal cord injury. In this respect, further investigations are required to define the differences and similarities between mouse and human spinal cord injury. However, the mouse models present an opportunity to evaluate the use of genetically manipulated animals for further understanding of spinal cord injury mechanisms. In this study, mice having complete transection or 2mm resection exhibited some movements of the hindlimbs in the open field. These movements could be attributed to local spinal cord reflexes, as the MFS remained below 2, i.e., there were movements in the hindlimbs during open field examination but no co-ordination or weight bearing was noted. The outcome, seen by light microscopy of sections stained by hematoxylin and eosin as well as Luxol fast blue-cresyl violet, is very similar to that obtained in experiments in the rat [20]. Thus, mild injury produced only minor cell changes and minimal necrosis in some animals. Moderate injury produced regions of hemorrhage and necrosis, preferentially in the gray matter of the cord. Severe compression caused extensive hemorrhages and necrosis. The use of MAP2 immunostaining revealed that alterations actually took place in nerve cell bodies and dendrites of the gray matter even in mice with mild compression, which is in line with the subtle functional injury which we recorded. With increasing degree of compression, the loss of MAP2 immunostaining and Luxol stain‘

ing were more pronounced. In moderately injured animals (Fig. 6, panel 3c) MAP2 staining was lost in the layers IV–VII with preservation of staining in layers I–III, similar patterns have been noted after moderate injury of rat spinal cord [21]. Computer-assisted image analysis of sections stained by the MAP2 and the Luxol method will prove of use in further studies on treatment of spinal cord trauma in the model introduced here. Acknowledgements The author would like to thank Madeleine Jarild for her excellent technical help that made this work possible. I am grateful to Dr. Inga Hansson, Department of Pathology, Uppsala University, for giving me the idea and inspiration that mouse can be a suitable model for spinal cord injury. I am also grateful to Mats Linder, Pharmacia and Upjohn, for the diagrammatic presentation of model in Fig. 1. Finally, would like to thank Cara McKee, OTR/L, Team Leader, Spinal Cord Injury Department, HealthSouth Sunrise Rehabilitation Hospital, Ft. Lauderdale, Florida, for her very helpful comments on the manuscript and revision of the text. This work was supported by grants from the Swedish Association of Neurologically Disabled, Fredrik and Ingrid Thurings Stiftelse, Tore Nilsons Stiftelse, Åhlén’s Stiftelsen and Svenska Läkaresällskapet.

References 1. Allen AR (1911) Surgery of experimental lesion of spinal cord equivalent to crush injury of fracture dislocation of spinal column. JAMA 57 : 878–880 2. Balentine JD (1978) Pathology of experimental spinal cord trauma. 1. The necrotic lesion as a function of vascular injury. Lab Invest 39 : 236–253 3. Bresnahan JC, Beattie MS, Todd FD, Noyes DH (1987) A behavioral and anatomical analysis of spinal cord injury produced by a feedback-controlled impaction device. Exp Neurol 95 : 548–570 4. Bresnahan JC, Beattie MS, Stokes BT, Conway KM (1991) Three-dimensional computer-assisted analysis of graded contusion lesions in the spinal cord of the rat. J Neurotrauma 8 : 91– 101 5. Broton JG, Nikolic Z, Suys S, Calancie B (1996) Kinematic analysis of limb position during quadrupedal locomotion in rats. J Neurotrauma 13 : 409–416 6. Constantini S, Young W (1994) The effects of methylprednisolone and the ganglioside GM1 on acute spinal cord injury in rats. J Neurosurg 80: 97–111 7. Farooque M, Zhang Y, Holtz A, Olsson Y (1992) Exudation of fibronectin and albumin after spinal cord injury in rats. Acta Neuropathol 84 : 613–620 8. Farooque M, Olsson Y, Holtz A (1994) Effect of the 21aminosteroid U74006F and methylprednisolone on motor function recovery and oedema after spinal cord compression in rats. Acta Neurol Scand 89 : 36–41 9. Farooque M, Hillered L, Holtz A, Olsson Y (1996) Changes of extracellular levels of amino acids after graded compression trauma to the spinal cord: an experimental study in the rat using microdialysis. J Neurotrauma 13 : 535–546 10. Farooque M, Hillered L, Holtz A, Olsson Y (1996 Effects of methylprednisolone on extracellular lactic acidosis and amino acids after severe compression injury of rat spinal cord. J Neurochem 66 : 1125–1130 11. Farooque M, Isaksson J, Jackson DM, Olsson Y (1999) Clomethiazole (ZENDRA, CMZ) improves hind limb motor function and reduces neuronal damage after severe spinal cord injury in rat. Acta Neuropathol 98 : 22–30 12. Fehlings MG, Tator CH, Linden RD (1989) The relationships among the severity of spinal cord injury, motor and somatosensory evoked potentials and spinal cord blood flow. Electroencephalogr Clin Neurophysiol 74 : 241–259

22 13. Gale K, Kerasidis H, Wrathall JR (1985) Spinal cord contusion in the rat: behavioral analysis of functional neurologic impairment. Exp Neurol 88 : 123–134 14. Hancock MB (1984) Visualization of peptide-immunoreactive processes on serotonin-immuno reactive cells using two-color immunoperoxidase staining. J Histochem Cytochem 32 : 311– 314 15. Holtz A, Nystrom B, Gerdin B (1989) Spinal cord blood flow measured by 14C-iodoantipyrine autoradiography during and after graded spinal cord compression in rats. Surg Neurol 31 : 350–360 16. Kuhn P, Wrathall JR (1998) A mouse model of graded contusive spinal cord injury. J Neurotrauma 15 : 125–140 17. Kunkel Bagden E, Dai HN, Bregman BS (1993) Methods to assess the development and recovery of locomotor function after spinal cord injury in rats. Exp Neurol 119 : 153–164 18. Lavey RS, Taylor JM, Tward JD, Li LT, Nguyen AA, Chon Y, McBride WH (1994) The extent, time course, and fraction size dependence of mouse spinal cord recovery from radiation injury. Int J Radiat Oncol Biol Phys 30 : 609–617 19. Li GL, Farooque M (1996) Expression of ubiquitin-like immunoreactivity in axons after compression trauma to rat spinal cord. Acta Neuropathol 91 : 155–160 20. Li GL, Farooque M, Holtz A, Olsson Y (1995) Microtubule-associated protein 2 as a sensitive marker for dendrite lesions after spinal cord trauma. Rest or Neurol Neurosci 8 : 189–197 21. Li GL, Brodin G, Farooque M, Funa K, Holtz A, Wang WW, Olsson Y (1996) Programmed cell death and expression of Bcl2 after compression trauma to rat spinal cord. J Neuropathol Exp Neurol 55 : 280–289 22. Li GL, Farooque M, Holtz A, Olsson Y (1996) Increased expression of growth-associated protein 43 immunoreactivity in axons following compression trauma to rat spinal cord. Acta Neuropathol 92 : 19–26 23. Li GL, Farooque M, Holtz A, Olsson Y (1997) Effects of αphenyl-N-tert-butyl nitrone (PBN) on compression injury of rat spinal cord. Free Rad Res 27 : 187–196 24. Li GL, Farooque M, Holtz A, Olsson Y (1997) Expression of the ubiquitin carboxyl-terminal hydrolase PGP 9.5 in axons following spinal cord compression trauma. An immunohistochemical study in the rat. APMIS 105 : 384–390 25. Noble LJ, Wrathall JR (1985) Spinal cord contusion in the rat: morphometric analyses of alterations in the spinal cord. Exp Neurol 88 : 135–149

26. Nystrom B, Berglund JE (1988) Spinal cord restitution following compression injuries in rats. Acta Neurol Scand 78 : 467–472 27. Nystrom B, Berglund JE, Bergquist E (1988) Methodological analysis of an experimental spinal cord compression model in the rat. Acta Neurol Scand 78 : 460–466 28. Panter SS, Yum SW, Faden AI (1990) Alteration in extracellular amino acids after traumatic spinal cord injury. Ann Neurol 27 : 96–99 29. Povlishock JT (1993) Pathobiology of traumatically induced axonal injury in animals and man. Ann Emerg Med 22 : 980– 986 30. Rivlin AS, Tator CH (1978) Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat. Surg Neurol 10 : 38–43 31. Schnell L, Schneider R, Berman MA, Perry VH, Schwab ME (1997) Lymphocyte recruitment following spinal cord injury in mice is altered by prior viral exposure. Eur J Neurosci 9 : 1000– 1007 32. Shu S, Ju G, Fan L (1988) The glucose oxidase-DAB-nickel method in peroxidase histochemistry of the nervous system. Neurosci Lett 85 : 169–171 33. Tarlov IM, Klinger H, Vitale S (1953) Spinal cord compression studies. 1. Experimental techniques to produce acute and gradual compression. Arch Neurol Psychiatry 70 : 813–819 34. Tator CH, Fehlings MG (1991) Review of the secondary injury theory of acute spinal cord trauma with emphasis on vascular mechanisms. J Neurosurg 75 : 15–26 35. Von Euler M, Akesson E, Samuelsson EB, Seiger A, Sundstrom E (1996) Motor performance score: a new algorithm for accurate behavioral testing of spinal cord injury in rats. Exp Neurol 137 : 242–254 36. Yakovleff A, Roby Brami A, Guezard B, Mansour H, Bussel B, Privat A (1989) Locomotion in rats transplanted with noradrenergic neurons. Brain Res Bull 22 : 115–121 37. Zhang Z, Fujiki M, Guth L, Steward O (1996) Genetic influences on cellular reactions to spinal cord injury: a wound-healing response present in normal mice is impaired in mice carrying a mutation (WIdS) that causes delayed Wallerian degeneration. J Comp Neurol 371 : 485–495