Spatial And Temporal Damage Evolution after Hemi-Crush Injury in ...

0 downloads 0 Views 230KB Size Report
Dec 23, 2007 - Injury in Rat Spinal Cord Obtained by High b-Value q-Space. Diffusion Magnetic ...... mental spinal cord injury. Brain Res. 579, 32–42. FENYES ...
Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

JOURNAL OF NEUROTRAUMA Volume 24, Number 3, 2007 © Mary Ann Liebert, Inc. Pp. 481–491 DOI: 10.1089/neu.2006.0158

Spatial and Temporal Damage Evolution after Hemi-Crush Injury in Rat Spinal Cord Obtained by High b-Value q-Space Diffusion Magnetic Resonance Imaging REVITAL NOSSIN-MANOR,1 REVITAL DUVDEVANI,2 and YORAM COHEN1

ABSTRACT Spinal cord injury (SCI) is a major cause of disability for many living persons. Therefore, several experimental models and handful of techniques were developed to study and characterize the damage evolution following SCI. In the present study, high b-value q-space diffusion-weighted imaging (DWI) was used to follow the spatial and temporal damage evolution in excised rat spinal cords following hemi-crush injury. The DWI results were correlated with behavioral testing. It was found that the damage depends, as expected, on the severity of the insult. Significant spontaneous recovery was observed, six weeks following the insult, only for the mild hemi-crush injury but not following the severe injury. The damage was found to be more severe in the area caudal to the trauma site as compared to the rostral section of the cord. Key words: DWI; MRI; hemi-crush; q-space diffusion; spinal cord injury

INTRODUCTION

S

PINAL CORD INJURY (SCI) causes devastating consequences of life-long disability and significant economic costs (Profyris et al., 2004), and therefore provides an impetus for numerous experimental studies aiming to find new repair strategies (Jeffery and Blakemore, 1999). The functional decline following SCI is attributed to the direct mechanical trauma, named the primary injury, but much more to a subsequent cascade of events leading collectively to a progressive wave of secondary degeneration, which develops minutes to days after the injury (Profyris et al., 2004). Among the macroscopic secondary damages one can observe the formation of hemorrhage

in the gray matter (GM) and edema predominantly in the white matter (WM) (Guth et al., 1999; Tator, 1995). The microscopic barrage of noxious pathophysiological mechanisms include the partially reversible disturbance of homeostasis, axonal swelling, and moreover an increase in the concentration of extracellular glutamate and reactive oxygen species, which eventually leads to necrosis and apoptosis, and hence to degeneration of cell bodies and axonal fiber tracts left intact by the initial trauma (Beattie et al., 2002; Profyris et al., 2004, Schwab and Bartholdi, 1996; Liu et al., 1997). Injury to the central nervous system (CNS) can be classified according to whether the damage is initiated at the cell body (GM) or axon (WM). The process of degener-

1School

of Chemistry, The Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Ramat Aviv, Tel Aviv, Israel. 2D-Pharm Company, Rehovot, Israel.

481

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

NOSSIN-MANOR ET AL. ation following axonal injury is relatively delayed, therefore provides a window of opportunity for treatment (Schwartz et al., 1999). At present, there is no universal accepted treatment for reducing the ongoing spinal cord (SC) damage following injury. In general, the different treatment approaches can be divided into two broad groups: methods used to induce regeneration (Horner and Gage, 2000) and those used to prevent and stop the spread of damage (Schwartz and Kipnis, 2001). The evaluation of new therapies heavily depends on the employed experimental models, which differ by the clinical injury they mimic and the extent and nature of the subsequent spontaneous recovery (De La Torre, 1984; Jeffery and Blakemore, 1999). In principle, SC recovery can be documented by morphological (Blight, 1991), behavioral (Basso et al., 1995), electrophysiological (Fehlings and Tator, 1992), and imaging techniques (Weirich et al., 1990). Conventional magnetic resonance imaging (MRI) techniques (T1, T2, and T2* weighted images), as well as more recent diffusion-weighted MRI (DWI) and diffusion tensor imaging (DTI) have been employed to evaluate acute SCI (Schwartz and Hackney, 2003; Nevo et al., 2001; Fraidakis et al., 1998; Ford et al., 1994; Weirich et al., 1990). Since the initial demonstration of the utility of DWI for early detection of acute stroke (Moseley et al., 1990), diffusion has become a powerful non-invasive mechanism for studying and evaluating the integrity of WM tracts in the CNS (Stark and Bradley, 1999; Le Bihan, 1995; Schwartz and Hackney, 2003; Beaulieu, 2002). In the last few years, DWI and its multidirectional tensor modality, DTI, have been used to study the morphology and pathology of normal and injured SC in ex vivo (Schwartz et al., 2003; Nevo et al., 2001; Inglis et al., 1997; Ford et al., 1994) and in vivo (Franconi et al., 2000; Fenyes and Narayana, 1999; Fraidakis et al., 1998) experimental models as well as in human subjects (Holder et al., 2000; Ries et al., 2000). DWI and DTI have also been employed to evaluate new therapies for SCI and were proven to provide non-invasive measurements that may reflect spontaneous and non-spontaneous axon regrowth and neuroprotection (Schwartz et al., 2003; Nevo et al., 2001). Until recently, most DWI applications have used Eq. 1, derived from the work of Stejskal and Tanner (1965), for analyzing the diffusion data. ln(E/Eo)  2g22(  /3)D  bD

(1)

According to this equation the normalized signal decay, E/Eo, as a function of b, the diffusion weighting, is monoexponential. Furthermore, , , and g represent the duration, time separation and strength of the pulsed magnetic field gradients, while  is the gyro-magnetic ratio,

and D is the self-diffusion coefficient. Indeed, all the abovementioned and many other diffusion studies have used relatively low b-values (bmax  2500 s mm2), where a single apparent diffusion coefficients (ADC) describes the water signal decay in the intact and damaged tissue. However, it is well known now that, when water diffusion is studied with sufficient high b-values, the signal decay is no longer mono-exponential (Niendorf et al., 1996; Assaf and Cohen, 1996, 1998; Inglis et al., 2001; Stanisz et al., 1997, 2001). Indeed, in recent years, different approaches have been suggested to analyze such diffusion data (Szafer et al., 1995; Stanisz et al., 1997; Ford et al., 1998; Chin et al., 2002). One approach for analyzing such DWI data is the qspace analysis, which is based on the Fourier transformation relationship between the echo attenuation, E(q), and q, the “reciprocal spatial vector,” defined as (2)1g, as shown in Eq. 2, E(q)  Ps(R,)exp(i2q  R)dR

(2)

where  P (R,) is the displacement distribution profile of the averaged propagator vector (R  rr0) during the diffusion time, ( (Callaghan et al., 1990; Cory and Garroway, 1990). This analysis provides a displacement distribution profile, which can be characterized by a mean displacement and probability for zero displacement. Not long ago, we suggested expending this approach to imaging of the CNS, as it was used to study SC maturation (Assaf et al., 2000) and degeneration (Assaf et al., 2003). Since then, the q-space approach was used to study multiple sclerosis (MS) patients (Assaf et al., 2002; Cohen and Assaf, 2002), experimental allergic encephalomyelitis (EAE) (Biton et al., 2005), and the effect of glutamate (Biton et al., 2004) in excised SC. Preliminary studies have shown that a hemi-crush of rat SC caused primarily a paralysis of one hind limb. Functional testing, such as grid walking and swimming, revealed an improvement of this paralysis, without any treatment, during the first 6 weeks post-trauma. In a preliminary study, we showed that high b-value q-space diffusion MRI is a useful tool for studying the hemi-crush trauma model in excised rat SC (Nossin-Manor et al., 2002). In the present study, we present the spatial and temporal evolution of the damage following hemi-crush injury in rat SC using high b-values q-space diffusion MRI. Spinal cords subjected to mild (15 sec) and severe (60 sec) unilateral trauma were examined ex vivo by qspace DWI. In general, the damage depended on the severity of the insult, and its severity decreased as the distance from the trauma site increased. In the case of severe injury, the damage was found to be more severe caudally to the trauma site. Spontaneous recovery was observed only in the case of mild injury at 6 weeks

482

HIGH b-VALUE q-SPACE DIFFUSION MRI IN SCI following the insult. Behavioral testing showed similar results.

METHODS

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

Subjects Thirty-six male Sprague-Dawley rats aged about 3 months on the day of the trauma were used in this study. All animal experiments were evaluated and approved by the IIBR Animal Care Committee. The rats were housed in wire cages, three or four rats per cage, and exposed to a 12 h/12 h light/dark cycle. Food and water were provided ad libitum. The rats were trained for grid walking and were then assigned randomly to one of three experimental groups: (i) mild unilateral SC trauma (15 sec hemi-crush; n  13); (ii) severe unilateral SC trauma (60 sec hemi-crush; n  16); and (iii) sham-operated SC (n  7). Of the 36 rats used, only 33 were analyzed, since one sham-operated SC was damaged when removed from the vertebral column, and in two rats subjected to severe injury, the crush extended well beyond one half of the SC.

Hemi-Crush Injury On completion of the behavioral training, all rats were deeply anesthetized with xylazine (15 mg/kg) and ketamine (70 mg/kg) injected intraperitoneally. A midline skin incision was made in the mid-thoracic region. The rats were subjected to a mild (15 sec) or severe (60 sec) unilateral spinal cord crush using “jewelers” forceps as described previously (Nossin-Manor et al., 2002). Following surgery, the wound was sutured. Sham-operated controls were anesthetized and their skin was incised, but no SCI was inflicted. Immediately after surgery, the rats were treated for 10 days to prevent infections, especially bladder infections, with 0.5 mL/kg Resprim™ (Teva Pharmaceutical Ltd., Israel) administered p.o. daily, and by mechanical emptying of the bladder twice a day.

Behavioral Testing for Motor Movement Rats were tested for their simple motor reflexes and their ability to walk over a grid prior to the insult and 5 days, 10 days, and 6 weeks post-injury. Neurological assessment of their sensory motor system following hemi-crush SCI included tests for their righting, grasping, and toe spreading reflexes (De La Torre, 1984). Locomotor scores were obtained for these animals using graded neurological scale ranging from 0 (no performance at all) to 2 (complete reflex performance) in a way that full implementation of all three reflexes was assigned the score 6. To test the ability of the rats to walk over a wire grid, we used a method previously described (Behrmann et al., 1992). For both train-

ing and testing, the rats were placed on the middle of the grid (100  70 cm with grid holes of 5  5 cm) and allowed to walk for 2 min. The number of times that each hind limb slipped was counted. Training consisted of one trial every day until the animal could walk on the grid, with each hind limb slipping not more than twice over the entire testing period, a criterion usually met within 10 days.

Tissue Preparation Following completion of behavioral studies, the rats were anesthetized with pentobarbital (45 mg/kg i.p.) and perfused through the heart with cold (4°C) saline followed by cold (4°C) fixative (4% paraformaldehyde [PFA] aqueous solution, 0.1M phosphate buffer, pH of 7.4). The spinal cords were removed and post-fixed in the same PFA solution until analyzed by MRI.

Magnetic Resonance Imaging Protocol In vitro MRI experiments were performed on excised rats spinal cords immersed in Flourinert (FC-77; Sigma, St. Louis, MO) using an 8.4-Tesla nuclear magnetic resonance (NMR) spectrometer (Bruker, Karlshruhe, Germany) equipped with a micro5-imaging probe (Bruker) capable of producing pulse gradients of up to 190 gauss cm1 in each of the three directions. The MRI protocol was based on a previously used protocol (Nossin-Manor et al., 2002), and included acquisition of axial multi-slice T1 (TR/TE  500/13 msec), T2 (TR/TE  4000/90 msec) and heavily diffusion-weighted images. All images were acquired as 256/128 pixel matrices and were transformed to 256  256 matrices after FT. First, a sagittal T1 image was collected to allow selection of five axial 2-mm slices with a 1-mm gap between them and FOV of 1.5  1.5 cm. Efforts were made to place the central slice (slice no. 3) on the crush site (n). Therefore, we could examine 14 mm of the spinal cord, 6 mm rostral (the remote, n2, and adjacent, n1, slices) and caudal (the adjacent, n1, and remote, n2, slices) to the trauma site. Diffusion was measured perpendicular to the long axis of the spinal cord. To compute q-space images, 16 axial heavily diffusion-weighted images were acquired with a stimulated-echo diffusion pulse sequence with the following experimental parameters: TR/TE//  2000/ 20/150/2ms. Accordingly, diffusion gradient strength, g, was incremented from 0 to 150 g/cm in 16 steps. Therefore, the maximal b and q values used in these experiments were 1  107 s cm2 and 1277 cm1, respectively.

Computation of the q-Space Magnetic Resonance Images The q-space MR images were obtained by previously described procedure (Assaf et al., 2000; Nossin-Manor et

483

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

NOSSIN-MANOR ET AL. al., 2002; Cohen and Assaf, 2002). The basic principle of this method is to arrange a three-dimensional (3D) array (256  256  16) from 16 diffusion-weighted images, acquired with linearly increasing q values in a way that the x and y coordinates are the image axes and the z direction is that of the q values. The signal decay in each pixel of the 256  256 matrix was transformed into displacement distribution profiles using Eq. (2) by a Fourier transformation of the signal decay with respect to q using an in-house Matlab® program. Two parameters characterize these pixel-by-pixel displacement distribution profiles; namely, the mean displacement (taken as 0.425 of the full width at half-height) and the probability of zero displacement (taken as the height of the displacement profile at zero displacement). Finally, the program was used to extract these two parameters and construct two sub-images based on these two on a pixel-by-pixel basis. These images are referred to as q-space displacement and probability MR maps.

Data Analysis To analyze damage evolution, a previously published concept (Kipnis et al., 2003) was used to construct a new Matlab program, which was designed to quantify the extent of damage in the acquired slices. To evaluate injury severity as a function of the distance from the trauma site and time following injury, a damage index (DI) was defined by the following expression: DI  (1-NInt)*(LA/TSCA)*100

(3)

where NInt is the normalized intensity (i.e., the mean intensity of the unilateral lesion divided by the mean intensity of a normal contralateral area), LA is the lesion area and TSCA is the total spinal cord area. Using Matlab, the probability MR maps were processed to calculate the TSCA automatically. The LA was determined by a user define procedure. Finally, the normalized intensity, NInt, was calculated by dividing the matching lesion probability by the probability of the non-injured contralateral side, and DI was determined.

Statistical Analysis The averaged results are expressed as means  SEM, and were analyzed using the independent t-test.

RESULTS High b-values q-Space Diffusion MRI after Spinal Cord Hemi-Crush Injury Figure 1 shows a series of q-space probability images of all five slices scanned of three different representative

rats spinal cords fixed 5 days, 10 days, and 6 weeks following mild (15 sec) hemi-crush injury. Figure 2 shows the same data sets for representative spinal cords following a severe (60 sec) hemi-crush injury. These two figures clearly demonstrate the injury spatial and temporal evolution for the two insult severities. Injury is manifested as a decrease in the probability for zero displacement values, and an increase in the mean displacement (data not shown). In principle, the larger these changes, the more severe the insult. Trauma was inflicted on less than one half of the SC. Accordingly, most of the damaged area can be observed at the lesion side in the lateral and ventrolateral WM. Figures 1 and 2 show that the injury is most prominent at the trauma site, and decreases rostrally and caudally as the distance from the site increases. Figure 1 demonstrates that, following mild hemicrush, the injury encompasses mainly the lesion center and is less evident in rostral or caudal slices. As expected, 5 days following unilateral trauma, the lesion covers almost one half of the central slice, and the probability for zero displacement in this area dramatically decreases. Subsequently, an additional slight deterioration in tissue state occurs after 5 more days (10 days post-injury; Fig. 1b). However, 6 weeks following mild hemi-crush injury, the SC appears to be nearly normal (Fig. 1c). On the contrary, in the case of a severe hemi-crush injury, pronounced damage is still evident 6 weeks following trauma (Fig. 2c). In this case, at all time points, the lesion can be vividly seen at the insult location itself (n), covering one half of the slice, as well as at adjacent and more distant rostral (n 1 and n 2, respectively) and caudal (n1 and n2, respectively) slices.

Spatial and Temporal Damage Evolution after Spinal Cord Hemi-Crush Injury Using q-Space DWI Parameters To quantify the degree of SC MRI abnormalities following injury, a user-defined Matlab programs was used to calculate a damage index, DI, which took into account the percentage of area affected by trauma and the change in magnitude of the examined parameter. By analyzing the q-space probability MR maps, we were able to estimate damage severity as a function of distance from the trauma site and time elapsed since injury. Figure 3 presents the damage index, calculated by the user-defined recognition analysis, versus the distance from trauma site 5 days, 10 days, and 6 weeks following mild and severe hemi-crush injury, respectively. As expected, DI was found to be greatest at the insult location (slice n) for both types of injury. In general, the damage severity decreased as the distance from the trauma site increased. A closer look revealed that 6 weeks following severe hemi-crush injury, the rostral portion of the cord

484

FIG. 1. High b-value q-space probability MR images of three different representative rat spinal cords 5 days (a), 10 days (b), and 6 weeks (c) after mild (15 sec) hemi-crush injury. The trauma site was placed in the central slice (slice no. 3, n).

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

FIG. 2. High b-value q-space probability MR images of three different representative rat spinal cords 5 days (a), 10 days (b), and 6 weeks (c) after severe (60 sec) hemi-crush injury. The trauma site was placed in the central slice (slice no. 3, n).

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

HIGH b-VALUE q-SPACE DIFFUSION MRI IN SCI

FIG. 3. The damage index (DI), calculated by the user-defined recognition analysis, as a function of the distance from the trauma site 5 days, 10 days, and 6 weeks following mild (a) and severe (b) hemi-crush injury. Asterisk indicates damage indices that are statistically different (p  0.05) from the damage index obtained 5 and 10 days following mild hemi-crush injury, or 5 days following severe hemi-crush injury.

was slightly less damaged than the caudal portion (Fig. 3b). This trend was not, however, statistically significant (p  0.06 for the two remote rostral and caudal slices, n 2 and n-2, respectively). We found that the rostral damage increases towards the trauma site and decreases more gradually caudally further away from the site. The temporal evolution of the damage following injury seems to indicate a significant recovery at the trauma site after 6 weeks in the mild hemi-crush model (p  0.05), although significant damage could be detected 5 and 10 days following trauma (Fig. 3a). In the severe hemi-crush model, on the other hand, recovery at the trauma site was found to be statistically insignificant (Fig. 3b). Interestingly, in this case, a trend of a spontaneous recovery 10 days post-injury and renewed damage deterioration about 5 weeks later (6 weeks post-injury) was observed in caudal and, more prominently, in rostral slices. This could be the result of competing secondary injury processes. Comparing Figures 3a and 3b at the different time points following hemi-crush injury revealed a significantly greater damage in the trauma site (slice n) as well as in the adjacent caudal slice (n-1) 6 weeks post-injury as the severity of injury was increased from mild to severe. Significant greater damage was also found in rostral and caudal slices adjacent to the trauma site (n 1 and n1, respectively) 5 days post-injury.

days, and 6 weeks following mild and severe hemi-crush injury, are depicted in Figure 4. Figure 4a demonstrates significant lower locomotor scores than that in controls 5 days, 10 days, and 6 weeks following severe injury. The results obtained 5 days and 6 weeks following mild injury are similar. However, locomotor score observed 10 days following mild injury was not significantly different from that of controls. Accordingly, these results suggest that there is a significant locomotor recovery 10 days post-injury and recurring damage deterioration about 5 more weeks after that (6th week). Figure 4b, which shows the number of footfalls in a period of 2 min, demonstrates that the number of footfalls was significantly higher for both insult severities than for controls 5 and 10 days following injury. This was also found to be true 6 weeks following severe hemi-crush injury. However, rats subjected to mild hemi-crush injury showed normal performance after the same time period; the number of footfalls observed was no significantly different from controls. Theses results suggest a spontaneous functional recovery following mild hemi-crush injury, only.

DISCUSSION Spinal Cord Hemi-Crush Injury and MRI Parameters

Behavioral Tests and Temporal Damage Evolution after Spinal Cord Hemi-Crush Injury The results obtained from behavioral tests of simple motor reflexes and grid walking performance 5 days, 10

In this study, we used a unilateral SC crush model and q-space DWI to assess the spatial and temporal damage evolution following SCI. Partial crush injury leads to

487

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

NOSSIN-MANOR ET AL.

FIG. 4. Behavioral tests. Simple motor reflexes, assigned by the locomotor scores running from 0 to 6 (a), and grid walking performance, deduced form the average number of footfalls in 2 min (b), of rats subjected to mild and severe hemi-crush injury 5 days, 10 days, and 6 weeks following SCI. The solid lines designate the performance of the sham-operated controls, whereas the dashed lines present the SEM of their performance. Asterisk indicates the locomotor scores and number of footfalls that are statistically different (p  0.05) as compared to controls.

measurable primary and secondary degeneration that can be distinguished from one another. This is of particular relevance since, for example, myelinated axons are known to be particularly vulnerable to pressure and can, therefore, be affected by secondary damage mechanisms (Gruner et al., 1996). Moreover, the hemi-crush model has an advantage in that both injured and non-injured reference cord tissue come from the same slice. DWI has been widely used in animal research as a method to evaluate the integrity of the WM in the injured SC (Schwartz and Hackney, 2003). Since it measures the Brownian motion of water molecules, DWI provides structural information beyond the spatial resolution of conventional MRI. Conventional based MRI techniques (T1, T2, T2*) may underestimate the degree of injury, as their contrast primarily reflects anatomical changes, water content (edema) and hemorrhage, respectively, which do not always correlate with the neurological and histological severity of SCI (Ford et al., 1994). Additionally, these MRI techniques do not necessarily reflect the WM pathophysiologycal state and consequently the functional lose caused by the injury. Therefore we suggest to add high b-value q-space DWI, that has been found to be very useful in several WM experimental models and human associated disorders (Assaf et al., 2000, 2002, 2003; Cohen and Assaf, 2002; Biton et al., 2004, 2005) to the MRI method to used to study SCI. This provide a more accurate delineation of the damage in the WM.

q-Space DWI, and the Spatial and Temporal Damage Evolution The observed decrease in probability for zero displacement and the concomitant increase in displacement values (data not shown) in the injured lateral and vetrolateral WM indicate that water molecules can diffuse more freely perpendicular to the SC WM fibers following SCI. In the acute stage, this may be due to direct mechanical disruption of the axonal membrane and myelin sheath at the lesion epicenter, and axonal swelling further away from it. In a later stage, the increase in the mean displacement could be the result of degenerative secondary events, like Wallerian degeneration, leading to a long-lasting destruction of spared WM tracts. The spatial damage evolution as revealed by q-space DWI maps and damage analyses depends, as expected, on the severity of the induced SCI. While most of the damage in the case of mild hemi-crush injury is located at the trauma site itself, when severe hemi-crush is induced damage is apparent in all five slices acquired. Both mild and severe hemi-crush injuries result in maximal damage at the trauma site and decreasing damage as the distance from the insult location increases. This finding is expected, as the spread of indirect injury mechanisms should decrease with increasing distance from the lesion site. Indeed, our results are in line with previous ex vivo and in vivo low b-value diffusion studies, following dif-

488

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

HIGH b-VALUE q-SPACE DIFFUSION MRI IN SCI ferent experimental models of SCI (Ford et al., 1994; Nevo et al., 2001; Schwartz et al., 2003; Fraidakis et al., 1998). In the present study the severe hemi-crush model shows that the cord portion situated caudally to the lesion center is more vulnerable to compression injury; the damage found in this tissue decreases more gradually than damage observed in the rostral portion of the cord. A recent MRS study of different models of trauma in the sciatic nerve also showed greater damage in the caudal portion of the nerve than in its rostral portion, following complete 60-sec crush injury (Stanisz et al., 2001). There is was found that the portion of the nerve rostral to the site of injury exhibits MR characteristics similar to those of normal nerves, while the caudal portion showed significant differences in all MR parameters. The differences between our findings and those of Stanisz et al. can be attributed first to the higher sensitivity of high b-value qspace DWI to restricted diffusion in the axonal milieu as compared to diffusion data obtained with lower b-values, and secondly to the fact that diffusion MRS evidently provides only averaged information on the entire system studied, whereas the use of DWI provides diffusion parameters for each pixel in the image. In addition, the different diffusion times used in the two studies may also affect the results by causing the slow (restricted) and fast (free) diffusing water molecule to be filtered differently (Cohen and Assaf, 2002; Nossin-Manor et al., 2005). The use of different tissue types may also be responsible for the different results. The degree of damage following SCI depends on the severity and level of injury to the cord (Profyris et al., 2004). In the present study, injury was consistently induced at the same cord level. Consequently, 6 weeks postinjury, both MRI parameters and grid walking performance results provide evidence for significantly greater damage in the severe hemi-crush model than in the mild one. In the latter model at this time-point DI at the trauma site is significantly different from the ones measured 5 and 10 days post-injury, and the number of footfalls does not differ significantly from control values. Accordingly, the results of the present study strongly suggest a structural and functional spontaneous recovery of the SC after mild hemi-crush injury. This significant recovery in q-space MR parameters was much less apparent and statistically insignificant in the 60-sec hemi-crush model. Stanisz et al. have demonstrated significant recovery of MRS parameters and histology of the caudal portion of a sciatic nerve about 4 weeks following a complete 60sec crush. In our study, however, significant spontaneous recovery was observed 6 weeks after 15-sec hemi-crush injury, but not after 60-sec injury. The results obtained by the behavioral tests of grid walking performance also

show significant spontaneous recovery 6 weeks following mild hemi-crush injury, only. In fact, there are currently numerous behavioral and histological studies showing that partial SC injuries in rats are followed by spontaneous functional recovery (Basso et al., 1995; Gruner et al., 1996; Harris et al., 1994). In general, the temporal patterns of motor recovery reveled by these functional tests manifests a significant improvement in behavioral outcome starting about 4 weeks post-injury. Furthermore, studies evaluating compression, contusion or transection SC injuries have shown that sparing of as little as 5–10% of descending axonal fibers (motor tracts) may be enough to lead to substantial spontaneous locomotor recovery (You et al., 2003; Fehlings and Tator, 1992; Blight, 1991). These previous finding together with our results have important consequence for therapeutic studies, as recovery following treatment should be distinguished from the spontaneous recovery characterizing the experimental model used, in order to truly evaluate the efficacy of new treatments in any incomplete SCI model including the hemi-crush injury model.

CONCLUSION In the present study, high b-value q-space diffusionweighted MRI was used to follow the spatial and temporal damage evolution in excised rat spinal cords following hemi-crush injury. Fixed rat spinal cords assigned to two experimental groups differing in their unilateral trauma severity, namely, mild (15 sec) and severe (60 sec), were scanned by MRI 5 days, 10 days, and 6 weeks following hemi-crush induction. Spatial damage evolution and the extent of chronic damage were found to depend upon the severity of insult. Both mild and severe hemi-crush injuries caused maximal damage at the trauma site and decreasing damage as the distance from the insult location increased. However, while most of the damage following mild hemi-crush injury was located at the trauma site itself, when severe hemi-crush was induced damage was apparent in all five slices acquired. In the latter injury model, the caudal portion of the SC seemed to be more vulnerable to compression injury. Furthermore, 6 weeks post-injury the results showed significantly greater damage in the severe hemi-crush model than in the mild one. Interestingly and most importantly, q-space diffusion MRI findings and behavioral tests pointed toward significant spontaneous recovery, especially at the trauma site, in this experimental trauma model when the hemi-crush was inflicted for 15 sec. In the severe 60-sec hemi-crush model, recovery at the trauma site was much less apparent and statistically insignificant. This study demonstrates the occurrence of

489

NOSSIN-MANOR ET AL. spontaneous recovery following hemi-crush injury and should, therefore, be taken into account when using this model to evaluate the efficacy of new drugs or therapies for spinal cord injury.

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

ACKNOWLEDGMENTS

BLIGHT, A.R. (1991). Morphometric analysis of a model of spinal cord injury in guinea pigs, with behavioral evidence of delayed secondary pathology. J. Neurol. Sci. 103, 156–171. CALLAGHAN, P.T., MACGOWAN, D., PACKER, K.J., and ZELAYA, F.O. (1990). High resolution q-space imaging in porous materials. J. Magn. Reson. 90, 177–182.

This research was supported by grants from the Israel Science Foundation (ISF; grant 522/03) and by the U.S.-Israel Binational Science Foundation (BSF; grants 353/2003).

CHIN, C.L., WEHRLI, F.W., HWANG, S.N., TAKAHASHI, M., and HACKNEY D.B. (2002). Biexponential diffusion attenuation in the rat spinal cord: computer simulation based on anatomic images of axonal architecture. Magn. Reson. Med. 47, 455–460.

REFERENCES

COHEN, Y., and ASSAF, Y. (2002). High b-value q-space analysis diffusion-weighted MRS and MRI in neuronal tissues—a technical review. NMR Biomed. 15, 516–542.

ASSAF, Y., and COHEN, Y. (1996). Detection of different water populations in brain tissue using 2H single- and doublequantum-filtered diffusion NMR spectroscopy. J. Magn. Reson. B 112, 151–159.

CORY, D.G., and GARROWAY, A.N. (1990). Measurement of translational displacement probabilities by NMR: an indicator of compartmentation. Magn. Reson. Med. 14, 435–444.

ASSAF, Y., and COHEN, Y. (1998). Non-mono-exponential attenuation of water and N-acetyl aspartate signals due to diffusion in brain tissue. J. Magn. Reson. 131, 69–85. ASSAF, Y., MAYK, A., and COHEN, Y. (2000). Displacement imaging of spinal cord using q-space diffusion-weighted MRI. Magn. Reson. Med. 44, 713–722. ASSAF, Y., BEN-BASHAT, D., CHAPMAN, J., et al. (2002). High b-value q-space analyzed diffusion-weighted MRI: application to multiple sclerosis. Magn. Reson. Med. 47, 115–126. ASSAF, Y., MAYK, A., ELIASH, S., SPEISER, Z., and COHEN, Y. (2003). Hypertension and neuronal degeneration in excised rat spinal cord studied by high-b value q-space diffusion magnetic resonance imaging. Exp. Neurol. 184, 726–736. BASSO, D.M., BEATTIE, M.S., and BRESNAHAN, J.C. (1995). A sensitive and reliable locomotor rating scale for open field testing in rats. J. Neurotrauma 12, 1–21. BEATTIE, M.S., HERMANN, G.E., ROGERS, R.C., and BRESNAHAM, J.C. (2002). Cell death in models of spinal cord injury. Prog. Brain Res. 137, 37–47. BEHRMANN, D.L., BRESNAHAN, J.C., BEATTIE, M.S., and SHAH, B.R. (1992). Spinal cord injury produced by consistent mechanical displacement of the cord in rats: behavioral and histologic analysis. J. Neurotrauma 9, 197–217. BEAULIEU, C. (2002). The basis of anisotropic water diffusion in the nervous system—a technical review. NMR Biomed. 15, 435–455. BITON, I.E., MAYK, A., ASSAF, Y., and COHEN, Y. (2004). Structural changes in glutamate cell swelling followed by multiparametric q-space diffusion MR of excised rat spinal cord. Magn. Reson. Imaging 22, 661–672. BITON, I.E., MAYK, A., KIDRON, D., ASSAF, Y., and COHEN, Y. (2005). Improved detectability of experimental allergic encephalomyelitis in excised swine spinal cords by high b-value q-space DWI. Exp. Neurol. 195, 437–446.

DE LA TORRE, J.C. (1984). Spinal cord injury models. Prog. Neurobiol. 22, 289–344. FEHLINGS, M.G., and TATOR, C.H. (1992). The effect of direct current field polarity on recovery after acute experimental spinal cord injury. Brain Res. 579, 32–42. FENYES, D.A., and NARAYANA, P.A. (1999). In vivo diffusion characteristics of rat spinal cord. Magn. Reson. Imaging 17, 717–722. FORD, J.C., HACKNEY, D.B., ALSOP, D.C., et al. (1994). MRI characterization of diffusion coefficient in rat spinal cord injury model. Magn. Reson. Med. 31, 488–494. FORD, J.C., HACKNEY, D.B., LAVI, E., PHILLIPS, M., and PATEL, U. (1998). Dependence of apparent diffusion coefficient on axonal spacing, membrane permeability, and diffusion time in spinal cord white matter. J. Magn. Reson. Imaging 8, 775–782. FRAIDAKIS, M., KLASON, T., CHENG, H., OLSON, L., and SPENGER, C. (1998). High-resolution MRI of intact and transected rat spinal cord. Exp. Neurol. 153, 299–312. FRANCONI, F., LEMAIRE, L., MARESCAUX, L., JALLET, P., and LE JEUNE, J.J. (2000). In vivo quantitative microimaging of rat spinal cord at 7T. Magn. Reson. Med. 44, 893–898. GRUNER, J.A., YEE, A., and BLIGHT, A.R. (1996). Histological and functional evaluation of experimental spinal cord injury: evidence of stepwise response to graded compression. Brain Res. 729, 90–101. GUTH, L., ZANG, Z., and STEWARD, O. (1999). The unique histopathological responses of the injured spinal cord. Implications for neuroprotective therapy. Ann. N.Y. Acad. Sci. 890, 366–384. HARRIS, R.M., LITTLE, J.W., and GOLDSTEIN, B. (1994). Spared descending pathways mediate locomotor recovery after subtotal spinal cord injury. Neurosci. Lett. 180, 37–40.

490

HIGH b-VALUE q-SPACE DIFFUSION MRI IN SCI HOLDER, C.A., MUTHUPPILLAI, R., MUKUNDAN, S., EASTWOOD, J.D., and HUDGINS, P.A. (2000). Diffusionweighted MR imaging of the normal human spinal cord in vivo. Am. J. Neuroradiol. 21, 1799–1806.

Downloaded by Tel Aviv University from www.liebertonline.com at 12/23/07. For personal use only.

HORNER, P.J., and GAGE, F.H. (2000). Regenerating the damaged central nervous system. Nature 407, 963–970.

SCHWARTZ, E.D., SHUMSKY, J.S., WEHRLI, S., TESSLER, A., MURRAY, M., and HACKNEY, D.B. (2003). Ex vivo MR determined apparent diffusion coefficients correlate with motor recovery mediated by intraspinal transplants of fibroblasts genetically modified to express BDNF. Exp. Neurol. 182, 49–63.

INGLIS, B.A., YANG, L., WIRTH III, E.D., PLANT, D., and MARECI, T.H. (1997). Diffusion anisotropy in excised normal rat spinal cord measured by NMR microscopy. Magn. Reson. Imaging 15, 441–450.

SCHWARTZ, E.D., and HACKNEY, D.B. (2003). Diffusionweighted MRI and the evaluation of spinal cord axonal integrity following injury and treatment. Exp. Neurol. 184, 570–589.

INGLIS, B.A., BOSSART, E.L., BUCKLEY, D.L., WIRTH III, E.D., and MARECI, T.H. (2001). Visualization of neural tissue water compartments using biexponential diffusion tensor MRI. Magn. Reson. Med. 45, 580–587.

SCHWARTZ, M., and KIPNIS, J. (2001). Protective autoimmunity: regulation and prospects for vaccination after brain and spinal cord injuries. Trends Mol. Med. 7, 252–258.

JEFFERY, N.D., and BLAKEMORE, W.F. (1999). Spinal cord injury in small animals 1. Mechanisms of spontaneous recovery. Vet. Rec. 144, 407–413. LE BIHAN, D. (1995). Diffusion and Perfusion Magnetic Resonance Imaging. Raven Press: New York. LIU, X.Z., XU, S.M., HU, R., et al. (1997). Neuronal and glial apoptosis after traumatic spinal cord injury. J. Neurosci. 17, 5395–5406. MOSELEY, M.E., COHEN, Y., MINTOROVITCH, J., et al. (1990). Early detection of regional cerebral ischemia in catscomparison of diffusion-weighted and T2-weighted MRI and spectroscopy. Magn. Reson. Med. 14, 330–346. NEVO, U., HAUBEN, E., YOLES, E., et al. (2001). Diffusion anisotropy MRI for quantitative assessment of recovery in injured rat spinal cord. Magn. Reson. Med. 45, 1–9. NIENDORF, T., DIJKHUIZEN, R.M., NORRIS, D.G., VAN LOOKEREN CAMPAGNE, M., and NICOLAY, K. (1996). Biexponential diffusion attenuation in various states of brain tissue: implications for diffusion-weighted imaging. Magn. Reson. Med. 6, 847–857. NOSSIN-MANOR, R., DUVDEVANI, R., and COHEN, Y. (2002). q-Space high b value diffusion MRI of hemi-crush in rat spinal cord: evidence for spontaneous regeneration. Magn. Reson. Imaging 20, 231–241. NOSSIN-MANOR, R., DUVDEVANI, R., and COHEN, Y. (2005). Effect of experimental parameters on high b-value qspace MR images of excised rat spinal cord. Magn. Reson. Med. 54, 96–104. PROFYRIS, C., CHEEMA, S.S., ZANG, D., AZARI, M.F., BOYLE, K., and PETRATOS, S. (2004). Degenerative and regenerative mechanisms governing spinal cord injury. Neurobiol. Dis. 15, 415–436.

SCHWARTZ, M., YOLES, E., and LEVIN, L.A. (1999). “Axogenic” and “Somagenic” neurodegenerative diseases: definitions and therapeutic implications. Mol. Med. Today 5, 470–473. STANISZ, G.J., SZAFER, A., WRIGHT, G.A., and HENKELMAN, R.M. (1997). An analytical model for restricted diffusion in bovine optic nerve. Magn. Reson. Med. 37, 103–111. STANISZ, G.J., MIDHA, R., MUNRO, C.A., and HENKELMAN, R.M. (2001). MR properties of rat sciatic nerve following trauma. Magn. Reson. Med. 45, 415–420. STARK, D.D., and BRADLEY, W.G. (1999). Magnetic Resonance Imaging, 3rd ed. Mosby Company: New York. STEJSKAL, E.O., and TANNER, J.E. (1965). Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J. Chem. Phys. 42, 288–292. SZAFER, A., ZHONG, J., and GORE, J.C. (1995). Theoretical model for water diffusion in tissues. Magn. Reson. Med. 33, 697–712. TATOR, C.H. (1995). Update on the pathophysiology and pathology of acute spinal cord injury. Brain Pathol. 5, 407–413. WEIRICH, S.D., COLTER, H.B., NARAYANA, P.A., et al. (1990). Histopathological correlation of magnetic resonance imaging in a spinal cord injury model. Spine 15, 630–638. YOLES, E., HAUBEN, E., PALGI, O., et al. (2001). Protective autoimmunity is a physiological response to CNS trauma. J. Neurosci. 21, 3740–3748. YOU, S.W., CHEN, B.Y., LIU, H.L., et al. (2003). Spontaneous recovery of locomotion induction by remaining fibers after spinal cord transection in adult rat. Restor. Neurol. Neurosurg. 21, 39–45.

RIES, M., JONES, R.A., DOUSSET, V., and MOONEN, C.T. (2000). Diffusion tensor MRI of the spinal cord. Magn. Reson. Med. 44, 884–892. SCHWAB, M.E., and BARTHOLDI, D. (1996). Degeneration and regeneration of axons in the lesioned spinal cord. Physiol. Rev. 76, 319–369.

491

Address reprint requests to: Yoram Cohen, Ph.D. School of Chemistry The Sackler Faculty of Exact Sciences Tel Aviv University Ramat Aviv, Tel Aviv 69978, Israel E-mail: [email protected]