JOURNAL OF NEUROTRAUMA Volume 7, Number 2, 1990 Mary Ann Liebert, Inc., Publishers

Characterization of Axonal Injury Produced by Controlled Cortical Impact LIGHTHALL,1 HARRY G. GOSHGARIAN,2 and CHRISTOPHER R. PINDERSKI3

JAMES W.

ABSTRACT Axonal injury and behavioral changes were evaluated 3-7 days after traumatic brain injury. Previous research from this laboratory demonstrated that clinical central nervous pathology is produced by dynamic brain compression using a stroke-constrained impactor. We wanted to determine if the technique also would produce diffuse axonal injury after recovery from the procedure. The experiments were performed at Wayne State University School of Medicine using aseptic techniques while assuring analgesic care. Impacts were performed at 4.3 m/sec or 8.0 m/sec, with 10% compression (2.5 mm). Extensive axonal injury was observed at 3 and 7 days postinjury using both velocity-compression combinations. Regions displaying axonal injury were the subcortical white matter, internal capsule, thalamic relay nuclei, midbrain, pons, and medulla. Axonal injury also was evident in the white matter of the cerebellar folia and the region of the deep cerebellar nuclei. Behavioral assessment showed functional coma lasting up to 36 h following 8.0 m/sec impacts, with impaired movement and control of the extremities over the duration of the postinjury monitoring time. These experiments confirm that the cortical impact model of traumatic brain injury mimics all aspects of traumatic brain injury in humans and can be used to investigate mechanisms of axonal damage and prolonged behavioral suppression. —

INTRODUCTION

Thedamage.

in humans is a reflection of the extent and severity of axonal However the mechanism underlying the development of axonal injury remains to be defined. Current neural trauma research has demonstrated that the severity of dysfunction following traumatic head injury is correlated with the degree of diffuse axonal injury as indicated by the presence of axonal beading and retraction balls. The presence of axonal damage following brain trauma has been described in both the experimental and clinical literature (Nevin, 1967;Gennarellietal., 1982; Pilz, 1983;Povlishocketal., 1983; Adams and Doyle, 1984; Adams et al., 1985, 1986;Povlishock, 1985; Lighthall, 1988; Cortezetal., 1989). morbidity of traumatic brain injury

'Biomédical Science Department, General Motors Research Laboratories, Warren, Michigan.

department of Anatomy, and 3medical student School of Medicine, Wayne State University, Detroit, Michigan. 65

LIGHTHALL ET AL.

Controversy regarding the onset and development of axonal injury remains. It may be that a subtle and progressive axonal change following mild head injury eventually will lead to swelling of axons and disruption of the neuraxis without initial tearing of the axons at the moment of trauma (Kontos et al., 1980; Povlishock et al., 1983; Povlishock, 1985; Dixon et al., 1987; Cheng and Povlishock, 1988). On the other hand, axonal injury may occur at the moment of trauma by the shear or tensile forces of the traumatic event physically disrupting the axons, leading to membrane retraction and extrusion of axoplasm, forming large reactive swellings (Gennarelli etal., 1982; Pilz, 1983; Adams and Doyle, 1984; Adams etal., 1985,1989). In addition to the onset of axonal injury, a question remains as to what level of mechanical input to the central nervous system is required to set in motion a progression of events that lead to axonal damage (Gennarelli and Thibault, 1982; Gennarelli et al., 1982; Adams et al., 1985; Povlishock, 1985; Lighthall et al., 1989; Lighthall and Pinderski, 1989). The purpose of the present study was to determine if nonfatal, diffuse brain injury could be produced reliably through controlled mechanical compression of the brain. The experiments were designed to determine the evolution of neural injury at light microscopic levels and to ascertain whether or not reactive axonal changes and functional coma are produced in conjunction with focal contusion of the cortex. The influence of contact velocity and level of forced compression on the evolution of the pathological alterations were addressed. Portions of this study have been reported in abstract form (Lighthall and Pinderski, 1989). MATERIALS AND METHODS

Surgical Procedures Adult male ferrets (Mustela putorius furo)[ weighing 1.0-1.5 kg were obtained from Marshall Farms, a long-established closed colony. The ferrets were fasted for 12 h prior to an experiment. Anesthesia was induced and maintained throughout the duration of the study with a mixture of ketamine HC1 (100 mg/ml; 25 mg/kg) and xylazine (20 mg/ml; 4 mg/kg) injected i.m. The plane of anesthesia was determined by the withdrawal reflex, palpebral response to light touch, and ECG. The anesthetic was supplemented by administering 20% of the initial dose of ketamine every 30 min if needed; xylazine was supplemented by administering 50% of the initial dose every 2 h or as indicated by neurological signs and ECG. Lidocaine (20 mg/ml) was infused subcutaneously in all areas where incisions were made. Careful attention was given to sterile technique, postsurgical wound closure, and care over the period of observation. All surgical procedures and recovery observations were conducted in the sterile surgery facilities in the Division of Laboratory Animal Resources, School of Medicine, Wayne State University, Detroit, MI. Animals received chloramphenicol (1000 IU/ml; 50 mg/kg i.m.) twice a day, beginning the evening before surgery and continued for 3 days after surgery. The scalp and dorsal neck were shaved and treated with a depilatory to remove any remaining hair. The surgical field was then prepared with povidone-iodine (Betadine) scrub and washed twice with alcohol. Core body temperature was monitored by a rectal thermistor probe and maintained at 39.5 ± 1.0°C by a circulating water heating pad. Sterile stainless steel needle electrodes were inserted subcutaneously to record lead II ECG on a neonatal monitor. The ferrets were mounted in a supine position in a custom built stereotaxic device. The animals were secured by ear bars and an orbitomaxillary clamp with incisor bar. All pressure points were infiltrated with lidocaine (20 mg/ml), and the ear bar tips were coated with lidocaine paste before insertion in the external acoustic meatus. A 1.5 cm diameter midline craniotomy was performed between bregma and lambda. The bone flap was stored in sterile saline during brain contusion and then replaced at the end of the experiment.

'The Research Biomédical Laboratory of GM Research Laboratories is accredited by the American Association for the Accreditation of Laboratory Animal Care (AAALAC). The rationale and experimental protocol for use of an animal model in this study have been reviewed and approved by both the Research Laboratories and Wayne State University School of Medicine Animal Research Committees and are in compliance with federal, state, and local laws and regulations and in accordance with the NIH Guide (DHEW Publication NIH 85-23). 66

AXONAL INJURY BY CORTICAL IMPACT

Controlled Impact Procedures The controlled impact was delivered using a stroke-constrained pneumatic impactor. The impact device and impact procedures have been described elsewhere (Lighthall, 1988). Briefly, the device consists of a stroke-constrained stainless steel pneumatic cylinder with a 5.0 cm stroke. The cylinder is mounted in a vertical position on a crossbar that can be adjusted in the vertical axis. The impactor tip (i.e., the end of the shaft that comes into contact with the exposed dura mater) is constructed from 9 mm diameter aluminum rod, 1.50 cm in length, with a flat interface and a rounded (2.0 mm radius) perimeter edge. For each trial, the pneumatic cylinder shaft with the impactor head in place is first extended the full stroke length. The crossbar holding the cylinder is then lowered until the impactor tip compresses the dura and underlying subdural space to contact the surface of the cerebral cortex. This is used as the zero point from which the depth of controlled compression is measured. The cylinder shaft is then returned to the fully retracted position, and the crossbar holding the pneumatic cylinder is lowered the desired level of compression, locked in place, and verified with

vernier micrometer. The impactor tip and surgical instruments were autoclaved or gas sterilized prior to surgery. The surgery and impact were performed aseptically, and the surgical site was not open for more than 1 h. Following impact, the bone flap was replaced and secured with bone wax, and the wound was closed with nonabsorbable suture. The animals were then removed from the stereotaxic device and monitored until recovery from anesthesia was complete. The analgesic pentazocine (Talwin) (50 mg/ml; 2.0 mg/kg i.m.) was administered every 6-8 h for 3 days to prevent discomfort. To determine the recovery time from deep anesthesia, control animals were given the same amount of anesthesia, including any supplement, and monitored until recovery was complete. This enabled control, nonimpact recovery times to be determined and compared with postimpact recovery times. The experiments were divided between mild severity and moderate severity injury categories consisting of a 2.5 mm deformation at contact velocities of 4.3 m/sec and 8.0 m/sec, respectively. The two injury severity groups were subdivided into two postinjury time point groups, 3 days and 7 days. Histological examination was performed on tissue taken from 3 surviving animals in each test group. Tissue was also examined from 2 sham surgery preparations. a

Postinjury Observation Procedures For the purpose of medical monitoring during the postinjury period, the animals the extent of their recovery from surgery and anesthesia.

were

staged according to

Stage 4:

Stage Stage Stage

Animal unconscious or semiconscious; unable to sit up or maintain sternal recumbancy. No response to external stimuli. 3: Animal can maintain sternal recumbancy but can not stand. Limited response to external stimuli. 2: Animal can stand and move about; not eating and drinking normally. 1: Animal active, alert, eating and drinking normally.

At the end of the selected postinjury observation time, the ferrets were prepared for transcardial perfusion fixation under ketamine HC1 (25 mg/kg)/xylazine (4 mg/kg) anesthesia. A midsternal thoracotomy was performed to expose the heart, and a mixture of heparin (1000 USP units/ml; 0.10 ml/kg) and lidocaine (20 mg/ml; 2 mg/kg) was injected into the left ventricle to prevent blood coagulation and arteriole constriction.

Transcardial perfusion was performed with cold (4°C) 0.1 M phosphate buffer (pH 7.4) until atrial return was clear. This was followed by 1000 ml of cold fixative consisting of 10% neutral buffered formalin in phosphate buffer. The brain, brainstem, and cervical spinal cord were removed and reimmersed in cold fixative for 12 h. The tissue was transferred to fresh phosphate buffer for refrigerated storage.

Histological Procedures During dissection, the presence of subdural hematoma, subarachnoid hemorrhage, and contusion or laceration of the dura mater, brain, and spinal cord were noted. After the dorsal and ventral surfaces of the 67

LIGHTHALL ET AL. intact brain were photographed, the rostral brainstem was transected at the pontomedullary junction, and serial 2 mm thick sagittal sections were made through the cerebral hemispheres. The cerebellum was sectioned on a plane 90° perpendicular to the folia. A composite photograph was taken of all sections. Sections were then dehydrated through xylene and embedded in paraffin. Microtome sections, 8 |xm thick, were cut from the region of the impact, the brainstem, and spinal cord and stained with either hematoxylin and eosin or a modified Palmgren technique for nerve fibers (Goshgarian, 1977).

RESULTS Acute

Impact Pathology

The 4.3 m/sec-2.5 mm velocity-compression impacts produced subdural hematoma (SDH) at the impact site. The SDH formed from the posterior region of the craniotomy to spread anteriorly and laterally to occupy the entire craniotomy site. In one case, the impact resulted in a tear in the superior sagittal sinus near the confluens sinuum. Hemorrhage was controlled using Gelfoam and was stopped before closure of the craniotomy. None of the impacts performed at this velocity-compression combination were fatal. The severity of the 8.0 m/sec-2.5 mm velocity-compression impacts was immediately evident by rapid SDH formation. Hemorrhage could be controlled, and the SDH was evacuated to facilitate closure of the craniotomy. In all but one trial, the high-velocity impact produced a tear in the dura mater in the posterior region of the craniotomy site similar to that produced by the low-velocity impact. Disrupted cortical tissue was evident in two cases, but hemorrhage was controllable, and the injury was not fatal. In four trials, the impact was fatal within minutes.

Functional Recovery

Excluding the four immediate fatalities, no deaths occurred during the recovery period. The low severity velocity-compression combination (4.3 m/sec-2.5 mm) did not produce behavioral suppression. In general, the ferrets in the low severity (4.3 m/sec-2.5 mm) impact group were judged to be in stage 4 only during their recovery from anesthesia. This was determined by anesthetizing uninjured ferrets using identical supplementation schedules and observing them in parallel during recovery. The low severity injured ferrets progressed through stages 3 and 2 more slowly than did the uninjured controls. However, none required special attention. Stage 1 was achieved within 6-10 h after the surgical procedures (in this series of experiments, the scalp sutures remained in place throughout the postinjury observation time period). The high severity (8.0 m/sec-2.5 mm) velocity-compression combination produced a more profound change in functional recovery. The same parallel anesthetic recovery comparison with controls was employed. Of five ferrets that survived the impact, two remained in stage 4 for approximately 24-36 h. During this time, i.v. supplementation with saline-dextrose was given. The ferrets displayed an absence of withdrawal reflex and corneal reflex during this period. Stage 3 was maintained through the 3-day observation period in one of the animals. The other was observed for 7 days and eventually reached a level between stage 2 and stage 1. The animal could eat and drink unassisted if the food and water were placed near the mouth, and veterinary nursing care was no longer needed. All ferrets in the high severity impact group displayed motor deficits, as indicated by paresis of the hindlimbs or hemiparesis of the right or left side. Circling was observed contralateral to the side displaying the paresis. None of the animals in the high-velocity impact group made full functional recovery. At 7 days, all showed some form of neurological deficit, although they were alert to external stimulus.

Gross

Pathology

Gross examination of the brains that received low severity impact (4.3 m/sec-2.5 mm) revealed cortical contusions (Fig. 1A, B). Contusions were restricted to the impact site and consisted of several (2-5) punctate lesions forming a circular pattern circumscribing the perimeter of the craniotomy site. In general, the appearance of the cerebellum, brainstem, and spinal cord was unremarkable in the low injury severity group 68

AXONAL INJURY BY CORTICAL IMPACT

FIG. 1.

A Contusion pattern produced by a moderate severity impact (velocity 4.3 m/sec; deformation 2.5 mm) following a 7-day postinjury observation time (cerebellum and brainstem on left). The open circle delineates the craniotomy/impact site. Cal bar 1 cm. B Frontal sections, 2 mm thick, showing the severity of the cortical contusion (taken from A on right), as indicated by arrowheads. C Extensive cortical contusion produced by a high severity impact (velocity 8.0 m/sec; deformation 2.5 mm) followed by 7-day recovery. Cal bar 1 cm. D Frontal sections, 2 mm thick, of high severity impact injury shown in C. Note the depth of the contusion (indicated by arrowheads) extending through the subcortical white matter. High severity impact results in near total disruption of the cortical gray and white matter immediately beneath the impact site. The corpus callosum does appear to be intact. Also note petechial hemorrhage in the brainstem and hippocampus in the lower left section. =

=

(Fig. 1 A). Gross frontal sections (2 mm thick) revealed the extent of the cortical contusion (Fig. IB). The low severity impact contusions generally were restricted to the cortical gray matter. However, contusions were seen to extend into the subcortical white matter in some cases. The remnants of subdural hemorrhage were noted in all cases to be confined to the impact/craniotomy site. The high severity impact group all displayed gross contusions and hemosiderin-lined cavities in the cerebral cortex in the region of the impact (Fig. IC, D). Frontal sections taken through the impact site of the high severity impact group confirmed the extent of damage to the cortex (Fig. ID), revealing total dissolution of the cortical structure. In some cases, the cavity extended to the depth of the corpus callosum but left this structure intact. Petechial hemorrhage was apparent in the hippocampus and forebrain structures in two cases.

Histopathology Microscopic examination of material from both low and high severity impact groups revealed consistent and extensive axonal injury (AI), as indicated by the presence of axonal retraction balls, beaded axons, and 69

LIGHTHALL ET AL.

sinusoidal, swollen axons (Figs. 2A, B, C, D and 3A, B, C, D). In general, AI within the subcortical white (SCWM) was restricted to the region of the impact and the corpus callosum. AI was observed

matter

infrequently in other SCWM regions. The extent of AI adjacent to and beneath the impact site was dependent the size of the contusion. As stated, the size and severity of the contusion were a function of the impact severity. High severity impact always produced a more extensive lesion (compare Fig. 1 A, B with Fig. IC, D) and, therefore, greater axonal injury in the region of the contusion. This observation was true also for the level

on

of AI observed in the corpus callosum. Both impact severities produced AI in thalamic relay nuclei, fibers of the internal capsule, pul vinar, pretectal regions, and the red nucleus. The precise location of the AI is difficult to specify because of the diffuse nature of the injury pattern (i.e., retraction balls were visible in various adjacent thalamic nuclei). The appearance of the injured axons ranged from small =S 10 pm round balls scattered through apparently normal parenchyma to large =* 30 pm diameter terminals occupying large voids. The high severity impacts produced what appeared to be a greater number of injured axons in any given section when compared to identical regions

FIG. 2. A Photomicrograph of pontomedullary junction displaying single large axonal retraction ball in the center of field. B Cerebellar gray matter showing numerous axonal retraction balls (arrowheads) of various sizes and appearing singly and in clusters. C Region of anterior thalamus showing extensive axonal injury, indicated by arrowheads. D Dorsolateral thalamus exhibiting axonal retraction balls occurring in string of beads. Cal bars 50 p.m. =

70

AXONAL INJURY BY CORTICAL IMPACT

FIG. 3. A Cerebellar white matter displaying axonal retraction balls (arrowheads). Extensive axonal injury was present in the deep cerebellar white matter extending into the superior cerebellar peduncle. Cal bar 50 ¡xm B Region of the cervical spinal cord displaying axonal injury as beaded, sinusoidal axons (indicated by arrowheads) and a string of single retraction balls (lower half of field). Cal Bar 50 |xm. C Subcortical white matter immediately beneath the impact site (Fig. 1C). Axonal retraction balls of various sizes are seen scattered throughout the field (indicated by arrowheads). D Axonal injury is also present in the corpus callosum, just medial to the region shown in C. In some cases, the retraction balls appear to be surrounded by apparently normal parenchyma, whereas some occupy voids created by the disrupted 100 u,m. axon. Cal bars for C and D =

=

=

severity impact group. Increasing postinjury observation times improved the likelihood of observing AI, yielding a trend that was similar to comparing the low and high severity groups. Longitudinal sections of the cerebellum and brainstem including the pons, medulla, and upper cervical spinal cord showed evidence of AI. Sinusoidal axons displaying beads and varicosities were observed coursing through these regions. It was difficult to observe AI in all sections taken from the brainstem, perhaps due to the orientation of the axons with respect to the plane of section. The most significant AI was observed in the white matter of the cerebellum and the region of the deep cerebellar nuclei (Figs. 2B and 3 A). Numerous axonal beads and varicosities were seen in the superior cerebellar peduncle and the main body of the cerebellar white matter involving the deep cerebellar nuclei (Fig. 3A). in the low

71

LIGHTHALL ET AL.

The number of reactive axonal changes was dependent not only on the severity of the impact but also on the length of the postinjury observation times. The high severity, 7-day postinjury group demonstrated more AI when compared to the low severity, 3-day postinjury group. In addition to qualitatively greater AI in the high severity group, AI could be seen radiating out into the cerebellar folia. In some sections, single retraction balls were observed in the granular cell layer and adjacent to Purkinje cell bodies in the anterior folia of the cerebellar vermis. The AI in the cerebellum was observed in the absence of hematoma or petechial hemorrhage.

DISCUSSION

Controversy regarding the onset and development of reactive axonal changes in traumatic brain injury still exists. However, it is generally agreed that the severity of behavioral neurological dysfunction produced by traumatic brain injury is directly related to the presence and extent of diffuse axonal injury. Neurological dysfunction, as shown by the duration of coma, neurological signs, and final outcome, is a direct indication of the extent of axonal injury and parallels the severity of the trauma (Nevin, 1967; Oppenheimer, 1968; Gennarelli et al., 1982; Pilz, 1983; Povlishock et al., 1983; Adams and Doyle, 1984; Povlishock, 1985; Adams et al., 1985, 1986, 1989). Clinical observation and experimental studies suggest that the immediate, persistent coma associated with severe head trauma is probably coupled to extensive and severe axonal damage distributed throughout the neuraxis (Gennarelli et al., 1982; Pilz, 1983; Adams and Doyle, 1984; Adams et al., 1989). Diffuse axonal injury may, therefore, functionally disconnect cortical from subcortical structures, resulting in a behavioral deficit such as coma (Povlishock, 1985). Subdural hemorrhage, cortical contusion, and/or brainstem laceration and contusion may not contribute to the onset of coma but rather affect the duration and final outcome of the injury through secondary metabolic complications, such as hypoxia and edema (Sullivan et al., 1976; Graham, et al., 1978; Kontos et al., 1980; Povlishock, 1985; Adams et al., 1985). Contusional hemorrhage (as depicted in Fig. 1) and edema may contribute to the onset of a prolonged comatose state by elevating intracranial pressure. It is most likely that the sequelae of focal traumatic brain injury in conjunction with widespread axonal injury contribute to the observed functional suppression of behavioral and motor responses. To understand the mechanisms underlying the development of neural injury and dysfunction following brain trauma in humans, an experimental model that mimics those changes must first be developed. Preliminary results from this laboratory indicate that cortical impact produces acute vascular and neuronal pathology that is similar to observations in currently accepted experimental models and clinical injury (Lighthall, 1988). Controlled impacts using the pneumatic impactor produce a range of injury severities that are a function of both contact velocity and level of deformation. This model is unique in its ability to produce graded cortical contusion, subcortical injury, and, in high severity impacts, brainstem contusion. Moreover, diffuse axonal injury, the hallmark of diffuse brain injury in humans, was observed. This would imply, based on clinical and experimental observation, that rapid mechanical deformation of the brain can produce behavioral suppression or functional changes similar to coma. In the present experiments, functional changes were observed, particularly in the high severity impact group. It was, however, difficult to judge with certainty if a persistent comatose state was produced. Since functional changes interact with recovery from surgical anesthesia and are progressive in nature, it was necessary to monitor recovery from anesthesia in the injured group in parallel with uninjured controls to document subchronic functional changes produced by the impact. Coma may be described as a profound state of unconsciousness from which one cannot be aroused (Langfitt, 1978). We believe that test subjects displaying a prolonged stage 4 recovery period were in a state of coma, based on the absence of withdrawal reflex, corneal reflex, righting reflex, and no response to deep pain stimulus. While in stage 4, the high severity impact group also exhibited extended, rigid limbs and rigid postural muscles indicative of decerebrate rigidity. Taking into account the other functional changes, the high severity group displayed all the clinical signs of a closed head injury that would correspond to a motor score rating of 3 to 5 on the Glasgow Coma Scale (Langfitt, 1978). Preinjury behavioral training, although not performed in these tests, would be useful in determining postinjury functional outcome, in addition to a more refined 72

AXONAL INJURY BY CORTICAL IMPACT

neurological or electrophysiological assessment using brainstem auditory evoked or electroencephalographic responses. The similarity in appearance of the axonal injury reported here to that reported in primate, subprimate, and human autopsy material is remarkable. We believe that the results obtained with both the hematoxylin and eosin and the Palmgren silver stain for axons have demonstrated that this model is appropriate for further investigation of the biomechanics of axonal injury. There is some question as to whether the axonal injury produced by controlled cortical impact is diffuse in nature, as originally defined in head acceleration experiments and clinical brain injury (Adams et al., 1989). The major difference between cortical impact injury and head acceleration injury is the lack of AI in regions of the subcortical white matter other than the impact site in the former case and the absence of hemorrhagic lesion associated with AI (Gennarelli etal., 1982; Adams etal., 1985, 1986,1989). The cortical impact model did produce AI, however, in various other anatomic regions of the neuraxis and more extensively than reported in mild fluid-percussion brain injury (Povlishock et al., 1983; Cheng and Povlishock, 1988; Cotiez

etal., 1989).

More extensive AI observed in the cortical impact model may be due to the open skull nature of the loading conditions placed on the brain. In the cortical impact model, the brain is loaded through an open craniotomy as opposed to the fluid-percussion model, where the boundary of the craniotomy is sealed and the intracranial space is continuous with the pressure column in the fluid-percussion cannula. This closed system produces generally uniform hydrostatic loading of the brain. In fluid-percussion, the only outlet for the hydrostatic pressure load is through the foramen magnum, producing the greatest tissue deformations in the brainstem resulting in the AI seen there by Povlishock et al. In contrast, the open skull system allows the pressure to be released through the craniotomy and the foramen magnum, producing a nonuniform mechanical deformation of the brain. This enables tissue deformations to occur remote from the impact site and in regions other than the brainstem. If the regional magnitude of the tissue deformation is sufficient, it will produce AI. This hypothesis is supported by the observation of AI following controlled cortical impact in regions other than the brainstem. The regions displaying AI following cortical impact included corpus callosum, cerebral peduncles, numerous thalamic nuclei, pretectum, pons, cerebellar peduncles, deep cerebellar nuclei, and cerebellar white matter. In some cases, the locations of the reactive axonal changes were in areas traversed by long-tract efferent systems originating in the cerebral cortex. Since contusion of the cerebral cortex was present in all cases, it is conceivable that the reactive axonal changes may be a result of primary damage to the soma and initial segment of the neuron, leading to progressive secondary pathology of the axon remote from the impact site. The contention for the cause and development of the reactive axonal changes is not supported by the present study. Numerous axonal swellings, beads, and varicosities were observed in the white matter of the cerebellum. Because the cerebellum neither receives nor projects fibers directly to the cerebral cortex, the axonal pathology observed consistently in the cerebellar white matter is likely due to primary focal damage to the axon rather than its soma of origin. Although extensive neuronal injury was not observed in the pontine region, it is conceivable that somata damage in pontine nuclei could be responsible for the axonal degeneration described in the cerebellum. Recent experimental evidence tends to refute clinical speculation that physical disruption of the axon occurs entirely at the time of impact rather than developing over time. Experiments performed by Povlishock et al. demonstrated that minor head injury in animals (i.e., no vascular damage or contusion/laceration) results in the genesis of reactive swellings within 12-24 h of the insult. However, when the brains of the animals were examined during the first 12 h posttrauma, no reactive swellings nor axonal disruptions were noted. These experiments demonstrated that the traumatic event (mild fluid-percussion brain injury) did not tear or shear axons to immediately form retraction balls but rather induced an early, subtle form of axon change that over time became progressively severe. Before complete transection of the fiber, axonal transport was disrupted, as indicated by a massing of cellular organdíes. Early changes in the axon were apparent only at an electron microscopic level (Povlishock etal., 1983; Povlishock, 1985; Cheng and Povlishock, 1988). It is conceivable that a moderate to severe trauma may indeed produce some immediate stretching or shearing of the axons, but it appears most likely that progressive axonal damage must also be responsible for the genesis of retraction balls seen in moderate to severe trauma. 73

LIGHTHALL ET AL.

Currently, the fluid-percussion method is the most widely accepted technique for producing experimental brain trauma, including reactive axonal injury, in subprimate species, such as cat and rat (Sullivan et al., 1976; Povlishock et al., 1978, 1983; Dixon et al., 1987, 1988; Hayes et al., 1987; Mclntosh et al., 1987, 1989; Stalhammar et al., 1987 ; Clifton et al., 1989). The fluid-percussion technique reproduces some of the features associated with moderate head injury in humans, characterized clinically by transient unconsciousness with prolonged alteration of mental status, a low incidence of hematomas, and prolonged neuropsychological deficits (Rimel et al., 1982; Clifton et al., 1989). The severity of behavioral and motor suppression observed in the cortical impact model may be attributed to the extent of contusional hemorrhage and concomitant accumulation of edema fluid and elevated intracranial pressure. However, this hypothesis has not yet been examined. Recently, Mclntosh et al. showed that the fluid-percussion, when used off midline, also produces cortical contusion, a histological end point previously unreported using this technique (Cortez et al., 1989; Mclntosh et al., 1989). Although the fluid-percussion technique may be suitable to study the physiology and pharmacology of moderate severity head injury, it is not well suited for the study of the biomechanics of trauma. Experimental evidence taken from high speed x-ray movies of the fluid-percussion event demonstrates a consistent though complex movement of fluid within the cranial cavity in the rat (Dixon et al., 1988; Lighthall et al., 1989). The interaction of the fluid pulse with the cranial contents does not lend itself to accurate biomechanical analysis of tissue deformations that produce the injury. The cortical impact model presented here can be used for analytic modeling of brain injury because the level of deformation can be controlled and measured independently. Velocity of impact is controlled, is reproducible, and can be verified independently using high-speed cineradiography. Experimental advantages inherent to the cortical impact model are further enhanced by its ability to reliably produce more extensive axonal injury than fluid-percussion in addition to nonfatal cortical contusion. In addition to anatomical injury, this model produces prolonged functional coma rather than transient unconsciousness. Confirmation of long-term functional outcome, including coma, and of progressive development of axonal pathological changes argue strongly in favor of the clinical relevance of this experimental injury model for studies of human brain injury mechanisms. The experimental results confirmed that the cortical impact model of brain injury can be used to investigate the mechanisms of axonal damage and the suppression of behavioral and motor function. This model should form the basis for analytical modeling of brain trauma biomechanics and lead to an enhanced understanding of how brain injury occurs and progresses in humans. ACKNOWLEDGMENTS The authors would like to express their sincere thanks to the veterinary staff of the Division of Laboratory Animal Resources at the School of Medicine, Wayne State University. Our appreciation is extended to Drs. T.E. Anderson, J. W. Melvin, and D.C. Viano and Mr. S. Ridella for helpful comments and thoughtful review of the manuscript. The authors would also like to thank Mr. J. Hiben for his assistance in the surgical procedures and Ms. A. Brady for her expertise in histological procedures.

REFERENCES ADAMS, J.H., and DOYLE, D. (1984). Diffuse brain damage in non-missile head injury, in: Recent Advances in Histopathology, 12th ed. P.P. Anthony and R.N.M. MacSween (eds). Churchill Livingstone: Edinburgh, pp. 241-257.

ADAMS, J.H., GRAHAM, D.I., and GENNARELLI, T.A. (1985). Contemporary neuropathological considerations regarding brain damage in head injury, in: D.P. Becker and Povlishock, J.T. (eds). Central Nervous System Trauma Status Report—1985. NINCDS Publication, pp. 443-452. ADAMS, J.H., DOYLE, D., FORD, I., et al. (1989). Diffuse axonal injury in head injury: Definition, diagnosis and

grading. Histopathology 15, 49-59.

ADAMS, J.H., DOYLE, D., GRAHAM, D.I., et al. (1986). Gliding contusions in nonmissile head injury in humans. Arch. Pathol. Lab. Med. 10, 485-488. 74

AXONAL INJURY BY CORTICAL IMPACT

CHENG, C.L.Y., and POVLISHOCK, J.T. (1988). The effect of traumatic brain injury

morphologic characterization of reactive axonal change. J. Neurotrauma 5, 47-60.

on

the visual system: A

CLIFTON, G.L., LYETH, B.G., JENKINS, L.W., et al. (1989). Effect of D, a-tocopheryl succinate and polyethylene

glycol on performance tests, after fluid percussion brain injury. J. Neurotrauma 6, 71-81. CORTEZ, S.C., McINTOSH, T.K., and NOBLE, L.J. (1989). Experimental fluid percussion brain injury: Vascular disruption and neuronal and glial alterations. Brain Res. 482, 271-282. DIXON, CE., LIGHTHALL, J.W., and ANDERSON, T.E. (1988). Physiologic, histopathologic andcineradiographic characterization of a new fluid-percussion model of experimental brain injury in rat. J. Neurotrauma 5, 105-121. DIXON, CE., LYETH, B.G., POVLISHOCK, J.T., et al. (1987). A fluid percussion model of experimental brain injury in the rat: Neurological physiological, and histopathological characterizations. J. Neurosurg. 67, 110-119. GENNARELLI, T.A., and THIBAULT, L.E. (1982). Biomechanics of acute subdural hematoma. J. Trauma 22, 680-686.

GENNARELLI, T. A., THIBAULT, L.E., ADAMS, J.H., et al. (1982). Diffuse axonal injury and traumatic coma in the primate. Ann. Neurol. 12, 564-574. GOSHGARIAN, H.G. (1977). A rapid silver impregnation for central and peripheral nerve fibers in paraffin and frozen sections. Exp. Neurol. 57, 296-301. GRAHAM, D.T., ADAMS, J.H., and DOYLE, D. (1978). Ischémie brain damage in fatal non-missile head injuries. J. Neurol. Sei.

39, 213-234.

HAYES, R.L., STALHAMMAR, D., POVLISHOCK, J.T., et al. (1987). A new model of concussive brain injury in the cat produced by extradural fluid volume loading: II. Physiological and neurophysiological observations. Brain Injury 1, 93-112.

KONTOS, H.A., WEI, E.P., and POVLISHOCK, J.T. (1980). Pathophysiology of vascular consequences of experimental concussive brain injury. Clinc. Assoc. Trans. 92, 111. LANGFITT, T.W. (1978). Measuring the outcome from head injuries. J. Neurosurg. 48, 673-678. LIGHTHALL, J.W. (1988). Controlled cortical impact: A new experimental brain injury model. J. Neurotrauma 5,1-17. LIGHTHALL, J.W., and PINDERSKI, CR. (1989). Axonal injury produced by controlled cortical impact: A light microscopic study. Sixth Annual Neurotrauma Society Meeting Abstract. J. Neurotrauma 6, 49. LIGHTHALL, J.W., DIXON, CE., and ANDERSON, T.E. (1989). Experimental models of brain injury. J. Neurotrauma 6, 83-99.

MCINTOSH, T.K., FADEN, A.I., BENDALL, M.R., et al. (1987). Traumatic brain injury in the rat: Alterations in brain lactate and pH as characterized by 'H and 31P nuclear magnetic resonance. J. Neurochem. 49, 1530-1540. MCINTOSH, T.K., VINK, R., NOBLE, L., et al. (1989). Traumatic brain injury in the rat: Characterization of a lateral

fluid-percussion model. Neuroscience 28, 233-244. NEVIN, N.C. (1967). Neuropathologic changes in the white matter following head injury. J. Neuropathol. Exp. Neurol. 26, 77. OPPENHEIMER, D.R. (1968). Microscopic lesions in the brain following head injury. J. Neurol. Neurosurg. Psychiatry 31, 299. PILZ, P. (1983). Axonal injury in head injury. Acta Neurochir. 32, 119. POVLISHOCK, J.T. (1985). The morphopathologic responses to experimental head injuries of varying severity, in: D.P. Becker and Povlishock, J.T. (eds). Central Nervous System Trauma Status Report—1985. NINCDS Publication, pp. 443-452.

POVLISHOCK, J.T., BECKER, D.P., and SULLIVAN, H.G. (1978). Vascular permeability alterations to horseradish peroxidase in experimental brain injury. Brain Res. 153, 223-239. POVLISHOCK, J.T., BECKER, D.P., CHENG, C.L.Y., et al. (1983). Axonal change in minor head injury. J. Neuropathol. Exp. Neurol. 42, 225-242. RIMEL, R.W., GIORDANI, B., BARTH, J.T., etal. (1982). Moderate head injury: Completing the clinical spectrum of brain trauma.

Neurosurgery 11, 344-351.

75

LIGHTHALL ET AL.

STALHAMMAR, D., GALINAT, B.J., ALLEN, A.M., et al. (1987). A new model of concussive brain injury in the cat

produced by extradural fluid volume loading: I. Biomechanical properties. Brain Injury 1, 79-91. SULLIVAN, H.G., MARTINEZ, J., BECKER, D.P., et al. (1976). Fluid percussion model of mechanical brain injury in the cat. J. Neurosurg. 45, 520-534.

Address reprint requests to: Dr. J. W. Lighthall Biomédical Science Department General Motors Research Laboratories Warren, Ml 48090

76

Characterization of axonal injury produced by controlled cortical impact.

Axonal injury and behavioral changes were evaluated 3-7 days after traumatic brain injury. Previous research from this laboratory demonstrated that cl...
10MB Sizes 0 Downloads 0 Views