Ultrastructural scoring of graded acute spinal cord injury in the rat

Transcription

1 J Neurosurg (Spine 1) 97:49 56, 2002 Ultrastructural scoring of graded acute spinal cord injury in the rat ERKAN KAPTANOGLU, M.D., SELCUK PALAOGLU, M.D., PH.D., H. SELCUK SURUCU, M.D., PH.D., MUTLU HAYRAN, M.D., M.S., AND ETEM BESKONAKLI, M.D., PH.D. Department of Neurosurgery, Hacettepe University Institute of Neurological Sciences and Psychiatry, Spinal Research Laboratory, Ankara Numune Education and Research Hospital; and Departments of Neurosurgery and Anatomy, and Department of Preventive Oncology, Oncology Institute, Hacettepe University, Faculty of Medicine, Ankara, Turkey Object. There is a need for an accurate quantitative histological technique that also provides information on neurons, axons, vascular endothelium, and subcellular organelles after spinal cord injury (SCI). In this paper the authors describe an objective, quantifiable technique for determining the severity of SCI. The usefulness of ultrastructural scoring of acute SCI was assessed in a rat model of contusion injury. Methods. Spinal cords underwent acute contusion injury by using varying weights to produce graded SCI. Adult Wistar rats were divided into five groups. In the first group control animals underwent laminectomy only, after which nontraumatized spinal cord samples were obtained 8 hours postsurgery. The weight-drop technique was used to produce 10-, 25-, 50-, and 100-g/cm injuries. Spinal cord samples were also obtained in the different trauma groups 8 hours after injury. Behavioral assessment and ultrastructural evaluation were performed in all groups. When the intensity of the traumatic injury was increased, behavioral responses showed a decreasing trend. A similar significant negative correlation was observed between trauma-related intensity and ultrastructural scores. Conclusions. In the present study the authors characterize quantitative ultrastructural scoring of SCI in the acute, early postinjury period. Analysis of these results suggests that this method is useful in evaluating the degree of trauma and the effectiveness of pharmacotherapy in neuroprotection studies. KEY WORDS electron microscopy scoring spinal cord injury ultrastructure N UMEROUS substances have been tested in experimental neurotrauma models. Behavioral tests in which the results were difficult to interpret have often been used as the main outcome measurement, without careful and quantitative histological examination. The absence of such data has made the testing of certain promising drugs in clinical trials problematic. 23 Many histological methods have been used to evaluate experimental SCI to obtain information on injury-related pathophysiology and the effects of neuroprotective agents on the spinal cord. Histological methods by which recovery from experimental SCI is assessed comprise posttraumatic cyst measurements, 6,12,18,27 axonal tracer determined cell counts in the cerebrum and midbrain, 9,19,20,22,26 cell counts in the pyramidal tract of the injured spinal cord, 3,25 and nonquantified ultrastructural evaluation. 1,13 A posttraumatic cystic cavity develops in the spinal cord of a rat subjected to severe spinal cord contusion. 12 In studies on neuroprotection, the maximum cross-sectional area and the volume of the cavities have been estimated for Abbreviations used in this paper: HRP = horseradish peroxidase; SCI = spinal cord injury; SEM = standard error of the mean; TEM = transmission electron microscopy. J. Neurosurg: Spine / Volume 97 / July, 2002 quantitative evaluation. In this evaluation, tissues are serially sectioned; every 10th section of tissue is photographed and magnified; and measurements are performed using an image analysis system. Thus, numerical values can be used for statistical analysis. 18 Axonal tracers such as HRP have been used in the quantitative evaluation of SCI. After HRP administration, coronal sections obtained in the cerebrum and midbrain are processed for HRP reactivity. Labeled corticospinal and rubrospinal neurons are counted to determine a cortical score and red nucleus score. The degree of preservation of dorsal columns including the corticospinal tracts at the injury site correlates with the cortical score, whereas the red nucleus score is related to the degree of the preservation of the lateral columns. 19 Fluorogold, an axonal tracer, has also been demonstrated to label rubrospinal neurons after SCI. Rubrospinal neuron counts correlate well with functional scores. 20 The concentration of axons distal to the injury site has been shown to diminish markedly after acute spinal cord transection or compression injury. Axons are counted in sections after being stained with the Holmes stain. 25 The quantitative analysis of axons in acute SCI has also been performed ultrastructurally. 3 In that study, investigators 49

2 E. Kaptanoglu, et al. used axon counting to evaluate axonal pathophysiology following trauma. Although a variety of ultrastructural pathological features of SCI have been described previously, a quantitative ultrastructural assessment of the changes in the postinjury spinal cord in conjunction with behavioral evaluation has not been performed. We previously reported for the first time that using electron microscopy to score SCI is very useful in neuroprotection studies. 15,16 In the present study we investigate the differential effects of graded SCI on spinal cord ultrastructure. Our goal was to determine if ultrastructural scoring of acute SCI in the early posttraumatic period shows correlation with trauma-related intensity and clinical findings after weight-drop contusion injury. Materials and Methods Adult male Wistar rats, weighing 190 to 230 g, were used for the study. The surgical procedure was performed after induction of general anesthesia (10 mg/kg xylazine and 60 mg/kg ketamine hydrochloride intramuscularly). Rats were placed in the prone position. After making a T6 10 midline skin incision, the paravertebral muscles were dissected. The T7 9 spinous processes and laminar arcs were removed, and a laminectomy was performed. The meninges were not disrupted. The weight-drop method 2 was used to produce 10-, 25-, 50-, and 100-g/cm spinal cord contusion injuries in the trauma groups. The graded force was applied using stainless-steel rods (3-mm-diameter tip, weighing 2, 5, 10, and 20 g) by dropping vertically through a calibrated tube. Injury apparatus included a 5-cm guide tube that was positioned perpendicular to the center of the spinal cord. After surgical and traumatic interventions, silk sutures were used to close the wound in layers. Experimental Groups The rats were randomly divided into five groups of 10 rats each; they were perfused and killed 8 hours postoperatively after undergoing clinical evaluation. In Group 1 rats (sham-operated controls) only laminectomy was performed, and a 1-mm-thick transverse section of nontraumatized spinal cord tissue was obtained after clinical evaluation. The Groups 2 to 5 rats (trauma-induced animals) underwent surgical and traumatic interventions. In these trauma groups, rats were subjected to 10-, 25-, 50-, and 100-g/cm impact injury, respectively. A 1-mm transverse section of traumatized spinal cord tissue was also obtained 8 hours postsurgery after clinical evaluation. Functional Evaluation Clinical behavior was assessed 8 hours after trauma by using the inclined-plane technique 21 and the Basso-Beattie-Bresnahan scoring system. 5 This scale includes 21 different levels of hind limb movements. Mean Basso-Beattie-Bresnahan scores of both legs were examined. Observers undertook the evaluations in a blinded fashion. FIG. 1. Drawing representing a transverse section of rat spinal cord. Letters in boxes indicate the areas in which samples were obtained (W = white matter [corticospinal tract, ventral portion of dorsal column; 50 axons/sample evaluated]; G = gray matter [anterior motor neurons; 20 neurons/sample were evaluated]). Obtaining Samples From the Spinal Cord Eight hours after surgery, following clinical evaluation, anesthesia was reinduced and animals underwent perfusion. To guarantee that equal quantities of tissue were obtained in the different groups, samples were taken from the spinal cord as follows: after performing T-7, T-8, and T-9 laminectomies, the area between the T-6 and T-10 laminar arcs was measured. At the level of T-8 trauma was produced in the middle of the area where laminectomy was performed. Additionally, a nonabsorbable No. 3-0 suture was inserted to paraverterbal muscles adjacent to lesion site to confirm the center of the traumatized area. A 1-mm cross-sectional area of spinal cord at the epicenter of trauma was removed. Samples for white matter were obtained from the dorsal column of the spinal cord where the corticospinal tract is located, and for gray matter samples were taken from the anterior gray horns where motor neurons are primarily located (Fig. 1). Spinal cord samples were collected in randomly numbered containers, and given to the authors as each sample in its own numbered container. After evaluating the numbered tissues by a person blinded to the study, the results were collected in the appropriate group lists. Sample Preparation for Electron Microscopy The tissues used for TEM were obtained after transcardiac perfusion with phosphate-buffered 2.5% glutaraldehyde/2% paraformaldehyde solution. The tissue samples were kept in the perfusion solution for 24 hours, postfixed with phosphate-buffered 2% OsO 4 for 1 hour and then dehydrated in a graded series of alcohol. After araldite embedding, 1- to 2- m semithin sections were obtained using an ultratome, stained with toluidine blue, and visualized using a light microscope. The same ultratome was used to obtain 60- to 90- nm-thick sections, which were contrast stained with uranyl acetate and lead citrate, and visualized using an electron microscope. TABLE 1 Grading system for quantitative evaluation of ultrastructural findings Category Score axonal myelin normal myelin layers 0 vesiculated myelin 1 cracked myelin layers 2 honeycomb & extruded vesicles 3 general axonal normal 0 light edema 1 mild edema 2 severe edema & loss of structure 3 intracytoplasmic edema absent 0 light 1 mild 2 severe (cell membrane defect) 3 nucleus normal 0 clumping 1 sparse chromatin 2 severe damage 3 vascular endothelium normal 0 light edema 1 mild edema 2 severe edema 3 50 J. Neurosurg: Spine / Volume 97 / July, 2002

3 Ultrastructural scoring of experimental SCI TABLE 2 Clinical functionality scores according to intensity of SCI* Percentiles Score Intensity (g/cm) Means SEMs degree of inclined plane Basso-Beattie-Bresnahan score *Ten samples were evaluated for each intesity level of SCI. p (Spearman nonparametric correlation test). Samples were evaluated according to the following grading system. Electron Microscopic Evaluation Grading System The grading system, used for quantitative evaluation, was based similarly on the principles of the methods used for evaluating different tissue samples 17,24 and the spinal cord 15,16 (Table 1). Fifty axons, 20 neurons, and 10 capillaries for each sample were evaluated. In each group, axonal degeneration, axon myelin splitting, intracytoplasmic edema, nuclear damage, and vascular endothelial damage were evaluated, counted, and presented as mean scores SEMs. Statistical Analysis Statistical analysis was performed using one-way analysis of variance and the Bonferroni t-test. Data are expressed as the means SEMs in Table 2. A probability value less than 0.05 is considered statistically significant. The correlation between the clinical functional scores and the trauma intensity was analyzed using the Spearman nonparametric correlation test. The trauma group TEM scores were analyzed using the nonparametric test for trends across ordered groups. 8 Results Functional Findings Inclined-plane scores obtained in the laminectomy-only group were higher than in all other groups. There was a decreasing trend in inclined plane scores with increasing trauma intensity. There was no significant difference between the 25- and 50-g/cm-injured groups (p 0.05). All other groups showed statistically significant differences (p 0.05). The Basso-Beattie-Bresnahan scores also showed a similar trend. Laminectomy-only rats showed the highest scores; the scores decreased with increasing trauma intensity. All groups showed statistically significant differences (p 0.05), except the 25- and 50-g/cm trauma groups (p 0.05). Ultrastructural Findings In the control (laminectomy-only) group, the spinal cord gray and white matter was normal and intact (Fig. 2). In the 10-g/cm weight-drop group, the neuronal nuclei appeared almost intact, although some marginalization of the chromatin material was observed. The cytoplasm was J. Neurosurg: Spine / Volume 97 / July, 2002 FIG. 2. Electron microscopic studies. Upper: Gray matter obtained in control (laminectomy-only) group, showing the normal appearance of nucleus (N), nuclear membranes (Nm), and axons (A). Lower: White matter of control (laminectomy-only) group demonstrating the normal appearance of axon, myelin sheath (S), mitochondria (M), and neurofilaments (Nf). Original magnification 6000, bar = 1 m. 51

4 E. Kaptanoglu, et al. FIG. 3. Electron microscopic studies. Upper: Gray matter obtained in the 10-g/cm trauma group, showing near-normal nucleus (N) but fewer organelles in the cytoplasm. Lower: White matter obtained in the 10-g/cm trauma group, demonstrating greater damage than gray matter. The axons show some dispersion in the myelin lamella (Ad) and the mitochondria are edematous. Intact endothelium (En) and patent capillary lumen (L) are seen. Original magnification 6000, bar = 1 m. less condensed with fewer organelles than normal (Fig. 3 upper). The white matter was more affected by the trauma. The axons showed some dispersion in the myelin lamellae and edema in the mitochondria. The vascular endothelium was intact and the capillary lumina were patent (Fig. 3 lower). In the 25-g/cm weight-drop group, the gray matter was also affected. There were numerous neuronal bodies with sparse chromatin and few organelles in the cytoplasm, including some swollen mitochondria (Fig. 4 upper). The damage in the white matter also increased with edematous spaces inside the axons, cracked myelin layers, and edematous areas in the periaxonal cytoplasm. The endothelial cells of the capillaries showed some swelling with narrowing of the lumen (Fig. 4 lower). In the 50-g/cm weight-drop group, the nuclear damage advanced with large empty spaces in the cytoplasm (Fig. 5 upper). In the white matter, the myelin layers of the axons were nearly destroyed with edematous areas in between (Fig. 5 lower). In the 100-g/cm weight-drop group, the neuronal nuclei showed severe damage and the cytoplasm was composed FIG. 4. Electron microscopic studies. Upper: Gray matter obtained in the 25-g/cm trauma group, showing sparse chromatin and few organelles in the nucleus. Lower: White matter obtained in the 25-g/cm trauma group, showing edematous axons with cracked myelin layers. Endothelial cells are swollen and lumen shows narrowing. Original magnification 6000, bar = 1 m. of a large empty space with no organelles. The vascular endothelium was swollen, and the lumen was completely blocked (Fig. 6 upper). In the white matter, there were few axons, large edematous spaces, and some erythrocytes due to bleeding (Fig. 6 lower). Correlation of Behavioral and Ultrastructural Results The descriptive characteristics of the functionality variables and the TEM score variables are presented in Tables 2 and 3, respectively. When the trauma intensity was increased, the degree of inclined plane that the rats could climb decreased (correlation coefficient: 0.860, p 0.001; Fig. 7 left). A similar significant negative correlation was observed between Basso-Beattie-Bresnahan scores and trauma intensity (correlation coefficient: 0.901, p 0.001; Fig. 7 right). The nonparametric trend test for ordered groups showed that axonal, neuronal, and capillary damage increased with each increasing level of trauma severity, as reflected by each of the five scores (p 0.01 for all) obtained in the TEM evaluation (Fig. 8). Discussion Graded SCI produced using the Allen technique in the rat has been confirmed by our results. 2 The rats subjected 52 J. Neurosurg: Spine / Volume 97 / July, 2002

5 Ultrastructural scoring of experimental SCI FIG. 5. Electron microscopic studies. Upper: Gray matter obtained in the 50-g/cm trauma group, showing advanced nuclear damage (N) with edema (E). Lower: White matter obtained in the 50-g/cm trauma group, showing severe axonal damage (A); myelin fragments (Mf), and edema (E) are seen. Original magnification 6000, bar = 1 m. to different weights dropped from the same height exhibited a graded functional deficit as measured by Basso- Beattie-Bresnahan scores 5 and the mean angle demonstrated in the inclined-plane test of Rivlin and Tator. 21 Ultrastructural results also indicated the production of graded lesions. These findings are well correlated with the results presented by Wrathall, et al., 27 in which graded injury produced using the Allen technique was correlated well with clinical results and light microscopy. Basso, et al., 5 also produced graded SCI by weight-drop technique. They demonstrated that their scale can distinguish differences in neurological outcomes in rats subjected to different levels of traumatic injury and that the behavioral outcome is highly correlated with histological damage. Behavioral testing revealed changes in functional deficits in relation to the degree of trauma. Significant differences were demonstrated among the laminectomy-only, 10-g/cm trauma, and 25-g/cm trauma groups, as well as between the 50- and 100-g/cm groups (Table 2). Results observed in the 25- and 50-g/cm trauma groups were very similar. Hence, we classified contusion injuries in the experimental rat model of SCI as mild (10-g/cm), moderate (25 50 g/cm), and severe (100 g/cm). We found similar clinical results in the laminectomy-only group pre- and postoperatively. Because we have also previously demon- J. Neurosurg: Spine / Volume 97 / July, 2002 FIG. 6. Electron microscopic studies. Upper: Gray matter obtained in the 100-g/cm trauma group, showing severe nuclear damage: empty spaces in the nucleus with loss of organelles (N), ruptured mitochondria (M), and edema (E). Swollen endothelium (En) and completely blocked lumen (L) are also seen. Lower: White matter obtained in the 100-g/cm trauma group. Large area of edema (E), complete axonal damage, erythrocytes due to red blood cell bleeding (RBC). Original magnification 6000, bar = 1 m. strated that laminectomy itself has no effect on the ultrastructure of the spinal cord, 15 results of nonoperated rats were excluded. Electron microscopy showed a gradual increase (worsening) in ultrastructural scores in relation to the increase of trauma intensity. The general axonal score was the most affected by traumatic injury, whereas vascular endothelium was the least (Table 3). Although all ultrastructural components showed gradual worsening in relation to trauma intensity, only nucleus-related damage remained unchanged when injury was increased from moderate (50 g/cm) to severe (100 g/cm). This finding is in contrast to with that found in the light microscopy study reported by Wrathall, et al., 27 who showed that a 50-g/cm or greater traumatic injury produces gray matter disappearance at the epicenter of the lesion 4 weeks after injury. This may be the result of the ischemia-related chronic effect on gray matter. Results reported by Balentine, et al., 4 support our findings of early axonal damage. They showed that axons in the lesioned segment of the cord resulted in necrosis, by 8 hours postinjury in most samples and in all samples by 1 day. Wound healing in the central nervous system results in 53

6 E. Kaptanoglu, et al. TABLE 3 Trauma intensity scores determined using TEM Percentiles Score Intensity (g/cm) Means SEMs axonal myelin score general axonal score intracytoplasmic edema nucleus vascular endothelium * Statistically significant at p 0.01 (nonparametric test for trends across ordered groups). the formation of cysts. Guizar-Sahagun, et al., 12 have demonstrated that there are three postinjury stages in the formation of cysts: a stage of necrosis, seen from Day 1 to 1 to 2 weeks after injury; a stage of repair, from 1 to 2 weeks postsurgery; and a stage of stability, from 8 to 15 weeks to 1 year postlesioning. Microcystic cavities begin to develop as early as 3 days postinjury in the white and gray matter by parenchymatous hemorrhage, vascular thrombosis, edema, axonal segmentation, and inflammatory infiltration. Two to 3 weeks after contusion, macrophages, which absorb the necrotic tissue, disappear from the lesion site, leaving cavities of variable size. Four to 5 weeks after contusion, cysts in the trabecular system are usually well defined. Although there are some disadvantages to measuring the cystic cavities for example, shrinkage of cysts during histological processing or rupture of cysts, lack of detailed histological information, and inability to use this method for early histological examination after acute SCI the method is very useful for determining the pharmacological effects in chronic neuroprotection studies. 6,10,14,18 The disadvantage of our present method may be the use of the ultrastructural scoring system in the chronic experimental SCI. Here, if the magnitude of the trauma is great enough to produce a cystic cavity within the chronic stage, it may not be possible to obtain equivalent tissue samples from the epicenter of the traumatic injury site, as described previously. In our study, we observed no cystic cavity, probably because of the early time period postinjury. It is also possible that mildto-moderate injuries may not result in the formation of huge cysts. Hence, we propose that the present method is very effective in the early postinjury period in cases of mild-to-severe injuries, although it may not be as effective in severe, chronic SCI experiments. Labeling of cortico- and rubrospinal neurons by retrograde tracers is a well-described and widely used method to assess experimental SCI. 9,19,20,22,26 Midha, et al., 19 have demonstrated that counting corticospinal neurons, retrogradely labeled with HRP, is related to the degree of preservation of the dorsal columns, whereas counting red nucleus neurons is related to the degree of preservation of the lateral columns. Naso and coworkers 20 have shown that retrogradely HRP-labeled rubrospinal neuron counts did correlate well with functional scores in experimental SCI. Theriault and Tator 26 have proposed that red nucleus neurons exert a long-term survival effect following SCI. They recommended the use of fluorogold rather than HRP to label neurons retrogradely. Although all these methods that provide lesion site related indirect information are excellent for showing the intensity of and recovery from the trauma for dorsal and lateral columns, they yield no data concerning spinal cord gray matter. They can be used in acute as well as chronic studies, although they may pose a critical disadvantage in neuroprotection studies. Giehl and Tetzlaff 11 have demonstrated that some neurotropic factors can prevent axotomy-induced death of corticospinal neurons in the rat. Their results proved that in neuroprotection studies the pharmacological action may affect axotomy-induced death of cortico- or rubrospinal neurons. Therefore, with the counting of brain and midbrain neurons in experimental drug studies, we may not be able to determine whether the drug exerts a neuroprotective effect on the lesion site of the injured spinal cord or prevents the death of cerebral or midbrain neurons. In the 54 J. Neurosurg: Spine / Volume 97 / July, 2002

7 Ultrastructural scoring of experimental SCI FIG. 7. Graphs demonstrating the negative association observed between function and trauma intensity. Left: The degree of inclined plane decreases with increasing trauma intensity (correlation coefficient: 0.860, p 0.001). Right: The Basso-Beattie-Bresnahan (BBB) scores show similar significant negative correlation with increasing trauma intensity (correlation coefficient: 0.901, p 0.001). ultrastructural scoring method described in our report, it is possible to observe the direct pharmacological action on the lesion site by examining the quantitative and subcellular data. Spinal cord injury can also be assessed by counting axons in the pyramidal tract of rats. 25 The axonal counting technique provides quantitative information concerning the severity of injury and is well correlated with clinical status. This method has been used to evaluate acute (15- minute), and chronic (6-week) SCIs. 3,9 Fehlings and Tator 9 have shown that after transversely sectioning and processing the injury site containing segment of spinal cord, cell counts can be performed as long as 6 weeks after cord injury. These data support the idea that TEM scoring of spinal cord injury may also be possible in studies of chronic disease. Ultrastructural changes in neuroprotection studies have been widely used in acute (30-minute) and chronic (10- week) studies. Electron microscopy examination provides very detailed information about injury site, protection of FIG. 8. Graph showing the change in the mean TEM scores according to intensity of the traumatic injury. The scores of axonal myelin, nucleus, vascular endothelium, axon, and edema are increased with each increasing level of trauma intensity (nonparametric trend test, p 0.01 for all). Higher values are associated with increasing severity of SCI. J. Neurosurg: Spine / Volume 97 / July,

8 E. Kaptanoglu, et al. neurons, cell membranes, axons, subcellular elements, and edema. 1,4,7,13,15,16 Although most of these studies lack quantitative data, it seems to be possible to evaluate the SCI site after 10 weeks. 13 These data also support the use of ultrastructural scoring of SCI in chronic studies. We previously determined that TEM scoring of SCI is a useful method in pharmacological studies. 15,16 We not only found that ultrastructural evaluation yielded ultrastructural information about the lesion site but that it allowed for numeric comparison of control and trauma groups. It was also possible to compare the effects of drug actions on different subcellular organelles. 16 In those studies, we did not have clinical data. In the present study, we demonstrated that, when the trauma intensity was increased, behavioral responsiveness showed a decreasing trend. A similar negative correlation was observed between trauma intensity and ultrastructural scoring. In the present study, we aimed to classify trauma intensity in rats in which the weight-drop technique was used to produce SCI. Mild injury was produced using 10-g/cm trauma, moderate injury by using 25 to 50 g/cm trauma, and severe injury by using 100-g/cm trauma according to clinical findings and quantitative ultrastructural scoring. We showed that there is a correlation between trauma intensity and quantitative ultrastructural scoring. We propose that this method is useful when studying different trauma intensities. It can be used in experimental SCI in the early postinjury period to evaluate detailed subcellular changes; by using the scoring system it is possible to compare the effects of neuroprotective agents. References 1. Akpek EA, Bulutcu E, Alanay A, et al: A study of adenosine treatment in experimental acute spinal cord injury. Effect on arachidonic acid metabolites. Spine 24: , Allen AR: Surgery of experimental lesion of spinal cord equivalent to crush injury of fracture dislocation of spinal column. A preliminary report. JAMA 57: , Anthes DL, Theriault E, Tator CH: Characterization of axonal ultrastructural pathology following experimental spinal cord compression injury. Brain Res 702:1 16, Balentine JD, Paris DU: Pathology of experimental spinal cord trauma, II. Ultrastructure of axons and myelin. Lab Invest 39: , Basso DM, Beattie MS, Bresnahan JC: Graded histological and locomotor outcomes after spinal cord contusion using the NYU weight-drop device versus transection. Exp Neurol 139: , Braughler JM, Hall ED, Means ED, et al: Evaluation of an intensive methylprednisolone sodium succinate dosing regimen in experimental spinal cord injury. J Neurosurg 67: , Coskun K, Attar A, Tuna H, et al: Early protective effects of iloprost after experimental spinal cord ischemia in rabbits. Acta Neurochir (Wien) 142: , Cuzick J: A Wilcoxon-type test for trend. Stat Med 4:87 90, Fehlings MG, Tator CH: The relationships among the severity of spinal cord injury, residual neurological function, axon counts, and counts of retrogradely labeled neurons after experimental spinal cord injury. Exp Neurol 132: , Fujimoto T, Nakamura T, Ikeda T, et al: Effects of EPC-K1 on lipid peroxidation in experimental spinal cord injury. Spine 25: 24 29, Giehl KM, Tetzlaff W: BDNF and NT-3, but not NGF, prevent axotomy-induced death of rat corticospinal neurons in vivo. Eur J Neurosci 8: , Guizar-Sahagun G, Grijalva I, Madrazo I, et al: Development of post-traumatic cysts in the spinal cord of rats subjected to severe spinal cord contusion. Surg Neurol 41: , Ildan F, Polat S, Oner A, et al: Effects of naloxone on sodiumand potassium-activated and magnesium-dependent adenosine 5 -triphosphate activity and lipid peroxidation and early ultrastructural findings after experimental spinal cord injury. Neurosurgery 36: , Kao CC, Chang LW, Bloodworth JM: The mechanism of spinal cord cavitation following spinal cord transection. Part 2. Electron microscopic observations. J Neurosurg 46: , Kaptanoglu E, Caner HH, Surucu SH, et al: Effect of mexiletine on lipid peroxidation and early ultrastructural findings in experimental spinal cord injury. J Neurosurg (Spine 2) 91: , Kaptanoglu E, Tuncel M, Palaoglu S, et al: Comparison of the effects of melatonin and methylprednisolone in experimental spinal cord injury. J Neurosurg (Spine 1) 93:77 84, Kirkali Z, Esen AA, Hayran M, et al: The effect of extracorporeal electromagnetic shock waves on the morphology and contractility of rabbit ureter. J Urol 154: , Means ED, Anderson DK, Waters T, et al: Effect of methylprednisolone in compression trauma to the feline spinal cord. J Neurosurg 55: , Midha R, Fehlings MG, Tator CH, et al: Assessment of spinal cord injury by counting corticospinal and rubrospinal neurons. Brain Res 410: , Naso WB, Cox RD, McBride JP, et al: Rubrospinal neurons and retrograde transport of fluoro-gold in acute spinal cord injury a dose-response curve. Neurosci Lett 155: , Rivlin AS, Tator CH: Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat. Surg Neurol 10:38 43, Ross IB, Tator CH, Theriault E: Effect of nimodipine or methylprednisolone on recovery from acute experimental spinal cord injury in rats. Surg Neurol 40: , Schwab ME, Bartholdi D: Degeneration and regeneration of axons in the lesioned spinal cord. Physiol Rev 76: , Tasdemir O, Katircioglu F, Kucukaksu S, et al: Warm blood cardioplegia; ultrastructural and hemodynamic study. Ann Thorac Surg 56: , Tator CH, Rivlin AS, Lewis AJ, et al: Effect of acute spinal cord injury on axonal counts in the pyramidal tract of rats. J Neurosurg 61: , Theriault E, Tator CH: Persistence of rubrospinal projections following spinal cord injury in the rat. J Comp Neurol 342: , Wrathall JR, Pettegrew RK, Harvey F: Spinal cord contusion in the rat: production of graded, reproducible, injury groups. Exp Neurol 88: , 1985 Manuscript received August 27, Accepted in final form February 26, Address reprint requests to: Selcuk Palaoglu, M.D., Department of Neurosurgery, Hacettepe University, School of Medicine, Sihhiye, Ankara, Turkey. palaoglu@hacettepe.edu.tr. 56 J. Neurosurg: Spine / Volume 97 / July, 2002