Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Are the Pathobiological Changes Evoked by Traumatic Brain Injury Immediate and Irreversible?

Are the Pathobiological Changes Evoked by Traumatic Brain Injury Immediate and Irreversible? Anatomy and Neurological Surgery, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA 23298-0709, USA Deptartment of Anesthesiology, University of Texas Medical Branch a t Galveston, Galveston, TX 77555-0830, USA more complex and progressive than previously envisioned, and as such, rejects many of the previous beliefs regarding the pathobiology of traumatic brain injury. Introduction Traumatic brain injury has long been thought t o evoke immediate and irreversible damage t o the brain parenchyma and its intrinsic vasculature. In this review, w e call into question the correctness of this assumption by citing t w o traumatically related brain parenchymal abnormalities that are the result of a progressive, traumatically induced perturbation. In this context, w e first consider t h e pathogenesis of traumatically induced axonal damage t o show that it is not the immediate consequence of traumatic tissue tearing. Rather, w e illustrate that it is a delayed consequence of complex axolemmal and/or cytoskeletal changes evoked by the traumatic episode which then lead to cytoskeletal collapse and impairment of axoplasmic transport, ultimately progressing t o axonal swelling and disconnection. Second, w e consider the traumatized brain's increased neuronal sensitivity t o secondary ischemic insult by showing that even after mild traumatic brain injury, C A I neuronal cell loss can be precipitated by the induction of sublethal ischemic insult within 24 hrs of injury. In demonstrating this increased sensitivity t o secondary insult, evidence is provided that it is triggered by the neurotransmitter storm evoked b y traumatic brain injury, allowing for sublethal neuroexcitation. In relation t o this phenomenon, the protective effect of receptor antagonists are discussed, as well as the concept that this relatively prolonged posttraumatic brain hypersensitivity offers a potential window for therapeutic intervention. Collectively, it is felt that both examples of the brain parenchyma's response t o traumatic brain injury show that the resulting pathobiology is much Corresponding author Dr John T Povlishock, Department of Anatomy, Virginia Commonwealth University, P O Box 980709, Richmond, VA 23298-0709. USA Tel + I 1804) 828-9535, Fax + I (804) 828-6293 Until recently most clinicians who routinely deal with traumatic brain injury have assumed that the contact and inertial phenomena associated with trauma physically tore neural and vascular elements, causing immediate structural damage and functional failure. Because of the perceived immediate and irreversible nature of these traumatically induced events, most felt that traumatic brain injury did not constitute an entity amenable to routine therapeutic intervention, and thus, for many, traumatic brain injury was viewed as a hopeless condition, in which the clinician is but a bystander. Despite the aura of negativism that has clouded the clinical impression of traumatic brain injury, recent evidence emerging from both the laboratory and clinical setting, suggests that these previously held concepts are no longer correct. Simply stated, it appears that, in the case of both neuronal and vascular elements, the forces associated with traumatic brain injury do not necessarily trigger immediate or irreversible damage. Rather, it appears that the forces of injury evoke more subtle cell membrane, cytoskeletal and synaptic changes that translate into a progression of reactive alterations, ultimately culminating in damage or dysfunction of the respective neural and/or vascular elements. In the current review, we focus on two examples of a delayed progression of traumatically induced reactive events that culminate in irreversible damage. First, we consider the pathobiology of traumatically induced axonal damage to demonstrate that such damage is not the consequence of immediate transection of the axon cylinder but rather the result of a progressive axonal change that leads to disconnection. Second, we consider the traumatized brain's hypersensitivity to delayed secondary cerebral ischemia to show that the neuronal cell loss found after such combined traumatic/ischemic insult is, in itself, a delayed event that can occur even when the onset of ischemia is postponed for up to 24 hrs post- J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Figure 1 (a) With anterograde peroxidase, focal axonal swelling (arrow) is recognized in an otherwise intact axon within 1 hr of the traumatic event (Dark Field Microscopy). XI ,000; (b)Twenty-four hours postinjury, one can see that the focal impairment of axoplasmic transport has led t o local axonal swelling and detachment. Note that the formed reactive axon swelling (RS) is encompassed by a distended and thinned myelin sheath (arrowheads). Also, note that many of the surrounding axons (AX) are unaltered. (Plastic thick section) X I ,000; (c) This section is taken from a control animal receiving intrathecal horseradish peroxidase. Note that the tracer diffuses through the extracellular space to concentrate at the Nodes of Ranvier (arrowheads) from where its passage into the axon is blunted by the intact axolemma. [plastic thick section.1 XI ,000; (d) Again, using intrathecally administered peroxidase, this thick section shows evidence of peroxidase passage into the axon cylinder following traumatic brain injury. Note the presence of peroxidase containing axons (block arrows) as well as other axons (AX) in the field showing no evidence of peroxidase uptake. X1.000. traumatic injury. These findings suggest that traumatic brain injury, alone, can perturb neurons sufficiently, without their physical disruption, to result in their altered sensitivity to a subsequently applied secondary ischemic insult. Taken together, these findings strongly suggest that, for the most part, the pathobiology of traumatic brain injury is not one of immediate and irreversible renting of neural elements. What follows are overviews of these axonal and neuronal abnormalities to fully illustrate these examples of evolving/ delayed damage occurring with traumatic brain injury. Traumatically Axonal Injury Induced Axonal Change/Diffuse Historically, it has been well recognized that axonal damage can be found in humans who sustain traumatic brain injury (2,3,4,5,6,52,63,64,65,45,46, 47,48,66). Such axonal damage is most commonly seen scattered throughout the brain of patients who show significant morbidity, without evidence of mass lesion formation or secondary insult, and these findings have given rise to the characterization of the condition as diffuse axonal injury (2,22), with the suggestion that the damaged axons are causative of the observed morbidity (2,22). The recognition of such axonal damage has been based primarily upon histological studies performed during postmortem analysis. In the first days postinjury, evidence of axonal damage has been based upon the histological recognition of swollen and/or truncated axonal segments which have been termed reactive axonal swellings or “retraction” balls (1). With continued posttraumatic survival, these swellings and balls are replaced by microglia, followed over time by Wallerian degeneration downstream from these sites (1). Because these reactive, swollen axons looked reminiscent of those profiles seen by Cajal and others J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Frequency Distrbuiton Interf-hment Distance Inter(iiamen1 thatmce (nm) Figure lie) This electron micrograph represents a color-enhanced image taken from an image analysis unit. Note that the neurofilaments (NF) are tightly packed and that they appear smooth showing no evidence of prominent sidearms. [The central bar was used for measurement.] X500.000 (f) This represents a frequency distribution plot of the interfilament distances in the control [non-peroxidase-containing] versus the experimental [peroxidase containing] axons. Note that in the experimental [peroxidasecontaining] axons, the frequency at which decreased interfilament distances occur is significantly increased in comparison to the controls; (9) This coronally sectioned electron micrograph represents a color-enhanced image taken from an image analysis unit. Note that in this axon, which displays no alteration in its axolemmal permeability, the neurofilaments (dots) and microtubules [boxes] demonstrate normal intra-axonal alignment and density. X45.000; (h) In contrast, this color-enhanced image, taken 5 mins after an injury, which evoked alterations in axolemmal permeability, n o w displays obvious compaction of neurofilaments as well as an apparent loss of microtubules. The hexagons were used in quantitative studies. X45.000. following physical transection of nervous tissue, it was assumed that the tensile forces of injury immediately tore axons, causing them to retract and expel their axoplasm to form the reactive swelling or "retraction" ball. The correctness of this assumption was first called into question by our laboratory in the early 1980's when we attempted to better understand the pathobiology of this process utilizing well-controlled animal models of injury (54). Using fluid-percussion brain injuries in cats, we devised multiple strategies to dissect out if the traumatic event was capable of immediately tearing the axon cylinder and,extruding its axoplasmic mass. In these studies, we relied primarily upon anterograde tracers which were delivered to the somata of the major projectional systems from where they were moved into the axon cylinder via anterograde axonal transport (54). While such transport was underway, the same animals were then subjected to mild traumatic brain injury, and the intra-axonal course of the anterogradely transported tracer was followed. It was assumed that, if the injury was capable of immediately tearing the axon and extruding its axoplasmic mass, we would be able to recognize, at some point along the axon's length, a focus showing disruption, with a burst of anterogradely transported tracer forming a truncatediswollen ball. Despite exhaustive efforts to identify such sites, no evidence of such direct tearing with axoplasmic extrusion was found. Rather, using light and electron microscopy, we recognized that the traumatic episode resulted in a perturbation of axoplasmic transport, at one point, along the axon's length (54). Such impaired transport could be recognized within the first hour postinjury (Fig. la), and over time, this site of impaired transport was associated with increased axonal swelling followed by local lobulation of the axon. Over time, 3-6 hours, J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury this lobulated segment was associated with further swelling, with ultimate axonal detachment at this site over a 12-24 hour posttraumatic period (54). With continued survival, the detached segment, in continuity with the staining soma, underwent continued expansion and swelling due to delivery of material via anterograde transport, and by 24 hours, the proximal disconnected segment now formed a grossly swollen reactive axon (Fig. l b ) entirely consistent with the reactive swellings/”retraction balls” long associated with traumatic brain injury. While this observation was made initially in the case of mild traumatic brain injury, it was subsequently confirmed in moderate and severe traumatic brain injury and was also confirmed in axons of both the motor and special sensory systems (11,19,56). Subsequent studies conducted by other investigators using different species and models further confirmed the correctness of our observations by demonstrating a comparable repertoire of impaired axoplasmic transport, axonal swelling, and disconnection (21,23,41). Gennarelli and colleagues (21,23), utilizing an inertial injury model in nonhuman primates and a guinea pig optic nerve stretch model further characterized by Maxwell and colleagues (41), found a similar progression of reactive change. That the progressive/evolving axonal changes found in experimental animals, are found in brain-injured humans has also been confirmed. Although one cannot directly assess anterograde transport in traumatically brain-injured man, immunocytoskeletal markers of cytoskeletal movement and/or other anterogradedly transported materials have been used to evaluate the progression of traumatically induced axonal change. Specifically, using antibodies targeted to the . cytoskeletal neurofilament (NF) subunits, we have confirmed, in traumatically braininjured humans, a comparable sequence of local axonal swelling, lobulation and disconnection, with subsequent continued swelling of the detached axon (13,ZS). Similarly, in elegant studies by Sherriff et al. (59,60,61) and Gentleman et al. (24), utilizing antibodies to the P-amyloid precursor protein, a substance known to move by anterograde axoplasmic transport, a comparable progression of events has been shown through the light microscopic examination of this protein at varying survival periods postinjury. Clearly, based upon the above, there is now consensus, in both animals and man, that traumatic brain injury can impair axoplasmic transport and result in a progression of changes leading to continued axonal swelling ultimately followed by axonal detachment. The key questions that now remain regard precisely what triggers or initiates this impaired axoplasmic transport and what then influences the temporal progression of the abovedescribed events. Znitiating mechanisms of axonal injury. In regards to those factors capable of initiating impaired axoplasmic transport, multiple processes may be implicated. On one hand, one could argue that the traumatic episode focally stretches the axon and thereby alters the axolemma. This could then allow for the influx of normally excluded ions, causing a disruption of intra-axonal ionic homeostasis to result in impairment of axoplasmic transport. Conversely, one could argue that the tensile forces of injury directly stretch the cytoskeletal framework of the axon cylinder, causing its disorganization and/or collapse to thereby elicit an impairment of axoplasmic transport. Fortunately, within the last two years, considerable insight has been gained into each of these mechanisms and their potential role in the genesis of the reactive axonal change associated with traumatic brain injury. In regards to the potential for traumatically induced alteration in the axolemma, evidence from our laboratory suggests that such axolemmal change occurs, and as such, may be a player in the pathobiology of reactive axonal change. Following upon the seminal observation of Maxwell et al. (43), that, in the most severe forms of traumatic brain injury, a subpopulation of fine caliber axons reveals direct axolemmal renting, we investigated the possibility for more subtle axolemmal change in axons, harvested from animals sustaining less severe injury (49,SO). In this approach, we utilized extracellular tracers normally excluded by the intact axolemma to determine if, with injury, these tracers were capable of moving from the extracellular compartment to the axoplasmic front. Utilizing intrathecally administered horseradish peroxidase and microperoxidase which are normally confined to the extracellular compartment (Fig. lc), we observed, with moderate to severe traumatic brain injury, focal alterations in the axolemma’s permeability to these tracers (Fig. Id). These tracers were seen in the axon cylinder within minutes of the traumatic episode and were observed at nodal, paranodal and internodal sites (49). In this sequence of events, it also appeared that the traumatic episode disrupted the normal tight alignment of the myelin sheath with the axolemma, permitting the tracer to move along the axolemma to the precise site of maximal axolemmal perturbation. When these sites were sampled for ultrastructural analysis, the tracers were identified within the axoplasm, where they were associated with a distinct subset of intra-axonal changes. Specifically, within minutes of the traumatic episode, the influx of tracer was associated with local mitochondrial swelling as well as an apparent compaction of the local neurofilament network (49). This compaction was confirmed by computer assisted reconstruction, wherein the neurofilaments revealed a dramatic reduction in interfilament distance, together with a loss of the neurofilament sidearms (49) (Figs. le,f). Interestingly, this mitochondrial swelling and neurofilament compaction remained relatively stable J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury over a several hour posttraumatic period with swellings developing, over time, rostra1 and caudal to this site of NF compaction. Following upon our initial description of neurofilament compaction and sidearm loss, made in sagittally sectioned axons, we repeated these studies in coronally sectioned axons both to gain better insight into the nature of these neurofilament changes and to better appreciate any related microtubular changes which are best seen when the microtubules are viewed en face (50,53). Once again, utilizing computer-assisted reconstruction of the axon cylinder, with counting of the neurofilament and microtubular density, we confirmed that the traumatic event was associated with focal changes in axolemmal permeability that resulted in immediate neurofilament compaction, with sidearm loss (50,531 (Figs. lg, h). Moreover, through the use of these coronally sectioned profiles, we also observed a rapid loss of the related microtubular network, suggesting that the axonal injury was associated with a complete collapse of the local cytoskeleton (50). As noted above, this cytoskeletal collapse remained relatively stable over a several hour period, and then was associated with an upstream and downstream swelling of the axon. This was most likely related to the fact that the cytoskeletal collapse disrupted those axoplasmic constituents (the microtubules) known to be related to anterograde and retrograde transport (28). Collectively, then, in the case of moderate to severe injury, evidence points to perturbation of axolemmal permeability which either directly or indirectly leads to disruption of those subcellular mechanisms involved in axoplasmic transport. The correctness of this belief was further substantiated by recent work conducted by Dr. William Maxwell and colleagues in an optic nerve stretch model of injury (40). Unlike the large caliber, heavily myelinated axons assessed in our studies, Maxwell and collaborators (40) focused on optic nerve fibers which, by their nature, are fine caliber, thinly myelinated axons, in which microtubules rather than neurofilaments predominate. Subjecting these axons to tensile loading by stretch, Dr. Maxwell performed a quantitative analysis of the cytoskeleton and observed a striking loss of microtubules in those axons that ultimately went on to show impaired axoplasmic transport, swelling and disconnection (40). While Maxwell and colleagues failed to show neurofilament compaction comparable to that described in our studies (49,50,51,53), this is not surprising. As these fine caliber axons possess proportionally few neurofilaments, the previously described neurofilament changes may be reserved only for neurofilament rich, large caliber axons. Irrespective of these differences, however, both studies point to traumatically induced alteration of the cytoskeleton which most likely, through microtubular damage, translates into impaired axoplasmic transport and axonal swelling. The key question that remains, in relation to the above-described pathobiology, is precisely what extracellular factor(s) migrate(s) through the altered axolemma to trigger the above-described cytoskeletal changes. In most people's mind, the obvious candidate is calcium, with the suggestion that an increase in intra-axonal calcium would lead to the activation of proteases which, in turn, would lead to the dissolution of the cytoskeleton. While such arguments are credible, it is important to note that, although the traumatic brain injury can result in rapid change within the axonal cytoskeleton, it does not typically result in rapid cytoskeletal dissolution or granulation as described with ischemic insult or other CNS pathologies, linked with calcium influx and calcium-mediated proteolysis (7,8,18,58,67,78, 69). Because we do not routinely observe such rapid cytoskeletal dissolution in the case of traumatic brain injury, we initially posited that the pathobiology associated with axonal injury was not calcium linked or alternatively, if calcium was involved, it was operating through previously unidentified mechanisms (56). However, because of the apparent differences between our views in traumatic brain injury and those described for other forms of CNS insult, we naturally questioned whether some of our observations were correct, or rather reflected some unique aspect of the chosen experimental animal model. This concern was obviated by several recent studies wherein the above-described sequence of rapid neurofilament compaction and sidearm loss were described using neurofilament rod and sidearm specific antibodies in addition to Western blotting in lampreys subjected to spinal cord transection (27). Also, using these same antibodies targeted to the neurofilament sidearm specific and rod domains, we have been able to confirm the same sequence for neurofilament compaction and sidearm loss in a rodent impact acceleration model of traumatic brain injury (57). In these same rodent studies, we also confirmed that the focus of neurofilament compaction and sidearm loss as well as microtubular change were directly related to sites of altered axolemmal permeability (57). Thus, from multiple species and animal models, there is compelling evidence that the traumatic episode can result in neurofilament compaction without overt dissolution of the cytoskeleton, although microtubular loss can be identified. Calcium could be implicated in this process as the rapid influx of calcium could lead to microtubular disruption and neurofilament sidearm cleaving and neurofilament compaction, with the compacted neurofilaments, remaining resistant to further calcium-mediated proteolysis (27). Alternatively, it is also possible that the flux of calcium may be transient and thereby incapable of causing further degradation of the cytoskeleton. In the case of traumatic brain injury, the issue of calciummediated damage may be more complex than in other experimental conditions and perhaps J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Figure 2 (a) This plastic thick section demonstrates that, with mild traumatic brain injury, reactive axonal change (RA)can occur without any evidence of altered axolemmal permeability to intrathecally applied horseradish peroxidase (arrowheads). XI ,000. (b) In this schematic illustration, w e attempt to portray those axonal events ongoing with traumatic brain injury of varying severity that lead to the formation of reactive axonal swellings. In the first panel, one can see a normal axon profile (A) and an enlargement of its basic cytoskeletal detail showing aligned neurofilaments (NF) with prominent side arms and microtubules (MT). With mild to moderate injury (B&C), one can see that the traumatic episode focally deforms the axon and translates into primary cytoskeletal (neurofilament and microtubule) misalignment, which then translates into focal impairment of axoplasmic transport. Continued anterograde transport of axoplasm and organelles leads to accumulation of these components a t the site of cytoskeletal injury, with progressive convergence of the axolemma (not shown), and disconnection of the proximal axonal segment (D) from its distal appendage. In this process, it is important to note that, although the traumatic event initially can be associated with infolding of the axolemma, there is no evidence of overt axolemmal perturbation. With more severe injury (B&E), focal intra-axonal abnormalities again occur; however, in this case, they involve concomitant changes within the axolemma and the related cytoskeleton. As illustrated, the traumatic episode is associated with altered axolemmal permeability (arrows), as well as rapid neurofilament side arm loss and subsequent neurofilament compaction (enlargement). Further, in the same field, microtubules are lost, and the local mitochondria swell. It is envisioned that all of these axonal changes are a direct reflection of consequences triggered by the altered axolemmal permeability which is assumed to initiate a generalized cytoskeletal collapse leading to a focal impairment of axoplasmic transport, localized swelling of the axon, with pinching off of the axon at the swollen site, and subsequent disconnection (F). In this illustration, the related myelin sheaths are not portrayed. J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury constitutes a manifestation not only of the presence of calcium, but also its concentration and duration of delivery via the altered axolemma. Recently, we have shown that some traumatically injured axons can remain permeable to extracellular tracers for several hours postinjury, suggesting that there may be a continuous and, perhaps, slow leak of the membrane, allowing for a continuing influx of calcium (51, 5 3 ) . While, obviously, more remains to be done in relation to the calcium story in traumatic brain injury, other recent studies serve to further illustrate the complexity of the issues at hand. Specifically, Maxwell et al. (42), have shown, in the stretched optic nerve, that injury results in the loss of the calcium-ATPase activity at the nodal axolemma and the loss of the ecto-Ca-ATPase in the outer aspect of the myelin sheath. Collectively, these events were associated 4-6 hours later with a delayed dissolution of the axonal cytoskeleton. Although traumatically induced injury is seen here to disrupt the cytoskeleton, it still constitutes a delayed event in contrast to other experimental conditions. Further, since it involves the perturbation of multiple mechanisms involved in the of flux and compartmentalization of calcium, the potential complexity of traumatically induced damage becomes even more apparent. thereby translate into impaired axoplasmic transport. Therefore, it appears that there are at least two initiating mechanisms in the pathobiology of reactive axonal change, one involving axolemmal/ cytoskeletal interaction and the other involving direct cytoskeletal perturbationhisalignment (Fig. 2b). The validity of this in vivo interpretation is indirectly substantiated by in vitru studies which show that differential tensile loading of isolated axons can result in a repertoire of change comparable to that described above (20). Specifically, with low tensile loading, axons, assessed in vitru, show axoplasmic disruption independent of any change in axolemmal integrity, entirely consistent with those changes described with mild traumatic brain injury. With more severe injuries, on the other hand, these same axons now reveal axolemmal change that correlates with dramatic axoplasmic failure (20). As these differential in vitru findings are consistent with our observations made in vivo, they tend to confirm our belief that the initiating pathobiology of traumatically induced reactive axonal change is varied and may be a function of the severity of tensile loading to which the axon is initially subjected. Possible diversity of the axonal response to injury. Although, in the case of moderate to severe traumatic brain injury, the evidence is mounting in support of the above-described scenario of axolemmal perturbation and cytoskeletal collapse, it is important to note that these same initiating factors may not be operant across the full spectrum of traumatic brain injury. In this regard, it is important to note that, in the same experiment described above using extracellular peroxidase, we also observed, with milder traumatic brain injuries, a population of axons that demonstrated reactive change without any evidence of axolemmal perturbation reflected in peroxidase passage from the extracellular to axoplasmic compartment (49) (Fig. 2a). Further in these axons, no evidence of neurofilament compaction or microtubular loss were observed. Rather, it appeared that the traumatic episode caused a direct focal misalignment of the neurofilament network which, over time, translated into impaired axoplasmic transport, swelling and disconnection. In these axons, the neurofilament misalignment and disorganization could be seen within 12 minutes of injury, and these sites were soon associated with a focal accumulation of organelles long linked to impaired axoplasmic transport (49). Precisely how this neurofilament disorganization translates into impaired axoplasmic transport is unknown. However, as the neurofilaments and microtubules are structurally aligned and integrated, it would seem logical that misalignment of one cytoskeletal element would cause comparable misalignment of others and Temporal course of reactive axonal change. One of the most puzzling aspects of reactive axonal change in both animals and man, centers on the temporal progression of this change from the moment of initiation to its termination as a fully formed reactive axonal swelling. While typically we have described the progression of reactive axonal change over a 3-6 hour time period, with fully mature reactive swellings seen within 12-24 hours of injury, we must admit that this temporal framework is not cast in stone, and indeed shows variation. In fact, Maxwell and colleagues have shown disconnection and the formation of mature reactive swellings in the optic nerve of the guinea pig with 6 hours of injury (41). Unfortunately, at present, little additional experimental evidence exists to shed light on to this issue. To add to this problem, it is our impression that the temporal sequence of reactive axonal change may reflect the species used (56). Although not supported by statistical studies, we have the strong impression that the temporal progression of reactive axonal change is more accelerated in rats than in micropigs and humans (56). It is unclear if this is an issue of species specificity, or alternatively a function of the distance between the sustaining soma and the reactive axon, which would naturally be much shorter in the rat than in the micropig or human. For example, reactive axonal change in the corticospinal tract of a rat will involve a pathway with a projection distance a fraction of that seen in higher order animals or man. Also related to this issue of the temporal progression is the possibility that this progression may be influenced by the severity of the injury. That is to say that the more severely injured J.T. Povlishock and L W. Jenkins: Pathobiology of Traumatic Brain Injury axons may undergo a more rapid progression of change than less severely injured axons. To date, little evidence exists to support either of these assumptions, and obviously this is an area which mandates continued investigation. Site of axonal failure. No consensus has been reached on precisely what axonal site is most vulnerable to the forces of injury. Many advocate that the nodal region is exquisitely sensitive due to the forces of injury, and thus, constitutes the most characteristic site of axolemmal and cytoskeletal failure (5,23). Our observations, on the other hand, do not support this assumption, as we have consistently found reactive axonal change, reflected in axolemmal and/or cytoskeletal abnormality within nodal, paranodal and internodal domains (11,19,49,53,54,55,56). In our estimation, axons fail at the site of maximal tensile loading and not at any preordained site. It should be noted that most of the fibers undergoing reactive axonal change are large caliber, long-track axons that are most likely subjected to maximal strain at one point along their length. Typically, in these fibers, reactive axonal swellings are seen at points where the axons decussate, cross blood vessels, or change their intraaxial course, suggesting that these sites are most prone to failure under maximal tensile loading. This implies, therefore, that the actual site of maximal tensile loading is more important than the presence or absence of a node of Ranvier. While we cannot preclude the fact that the node may be the most sensitive site for potential failure, we would argue that large caliber axons which are separated by large internodal distances most commonly do not fail at the nodal domains. Evidence to support.such an argument is also found in in vitro studies wherein mechanical stretching of peripheral axons has been shown to result in their collapse at internodal and paranodal segments rather than at nodal sites (Saatman, personal communication). Evidence for the Traumatized Brain's Hypersensitivity t o Delayed Secondary Cerebral Ischemia the posttraumatic secondary insults exacerbate ongoing pathologies found after traumatic brain injury, our studies have taken a slightly different focus, arguing that traumatic brain injury can also alter the neurons in such a fashion to display an increased sensitivity to a secondary ischemic insult which would have been well tolerated in the nontraumatized state. In the introduction, it was noted that the pathobiological sequelae of traumatic brain injury are not always immediate and irreversible and in this context, the delayed and progressive pathology associated with traumatically induced axonal injury was considered. Now, we turn our attention to the issue of the brain parenchyma's sensitivity to secondary ischemic insult. Many have recognized that secondary insults contribute substaqtially to the morbidity associated with human traumatic brain injury (9,26,37,44,26,53), and moreover, laboratory evidence has also shown that secondary insults such as hypertension, hypoxia, and global ischemia worsen the bioenergetic, electrophysiologic and behavioral status of animals subjected to traumatic brain injury (16,30,31,32). While most advocate that Evidence for the traumatically injured brain's hypersensitivity to delayed cerebral ischemia. Although this issue was first considered in cats (39, it was most critically assessed in rodent models of traumatic brain injury in which mild traumatic brain injury was complicated by imposed forebrain ischemia (33,34,36). In this approach, fasted rats were subjected to mild fluid percussion brain injury (17) of sufficient severity to evoke behavioral suppression while eliciting no overt change within the brain parenchyma, as defined by traditional histopathological studies of both the neuronal somata and their axonal appendages. These mildly injured animals were then subjected, one hour later, to the onset of 6 min of transient forebrain ischemia, accomplished by bilateral carotid clamping and systemic hypotension. In addition to assessing each of these insults in combination, each insult was individually assessed to fully appreciate the respective pathobiology of each insult. In this context, we observed, as anticipated, that mild traumatic brain injury was associated with no histological change within the brain parenchyma. Similarly, when the 6 min of transient forebrain ischemia was studied in isolation, it, too, was observed to result in no morphological change. However, when the traumatic insult was combined with an ischemic insult one hour postinjury, extensive bilateral CA1 hippocampal cell loss was found at 7 days postinjury, but not within the first day of survival (33,34,36). Through this paradigm, we recognized that mild brain injury, in some fashion, enhanced selectively vulnerable and regionally restricted ischemic delayed neuronal cell death over 7 days in the absence of any initial overt pathological change within these loci. Mechanisms involved in the induction of the brain's hypersensitivity to secondary cerebral ischemia. Following upon our observation that mild traumatic brain injury can enhance the potential for delayed neuronal cell death following ischemic insult, the obvious question arose as to what could be causing the brain's hypersensitivity to this secondary cerebral ischemia. At that time, it was becoming well established that a mild traumatic brain injury, not capable of evoking overt structural cell change, was able to elicite a net phase of neuroexcitation in which multiple transmitters were released to participate in receptor- mediated pathology throughout the brain (29). Against this backdrop, it J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury was postulated that the increased regional, delayed neuronal cell death found in our studies was consistent with a sublethal excitotoxic process which occurred with mild traumatic brain injury, and which was then potentiated to lethality by a second, normally sublethal-ischemic initiated excitotoxic process. Additionally, since the observed selected neuronal vulnerability was similar to the pattern of delayed ischemic neuronal necrosis found in pure forebrain ischemia in fasted rats (14) in which there is substantial evidence for excitotoxicity and calcium mediated mechanisms (15,62), there appeared to be compelling evidence for the role of excitotoxic processes in our observed pathobiological responses. This issue was critically addressed by the repetition of the above-described experiments, now using multiple receptor antagonists (34). For reasons, which will be detailed below, we first used lmg/kg of scopolamine and 4 mg/kg of phencyclidine which were administered 15 mins prior to the induction of mild traumatic brain injury, followed one hour later by secondary cerebral ischemia as described previously. Through this approach, we found the use of the competitive muscarinic (scopolamine) and noncompetitive NMDA (phencyclidine) antagonist provided substantial protection against the previously observed delayed CA1 neuronal death typically seen at 7 days after a traumatic injury complicated by cerebral ischemia (34). Interestingly, in this approach, it was also found that the drug action was effective whether it was given prior to or during both trauma and the imposed forebrain ischemia. The rationale for antagonizing both receptor populations, i.e. the muscarinic and NMDA receptors, was based upon evidence that the ion channel controlled by the NMDA receptor is voltage sensitive and that maximum calcium influx via this channel occurs when the membrane potential is decreased or depolarized. The magnesium block is removed from the NMDA channel in proportion to the membrane voltage decrease. Obviously, then by blocking the action of either acetylcholine acting on the muscarinic receptor or other excitatory transmitters which contribute to CA1 depolarization, one may inhibit the removal of the NMDA channel magnesium block and thereby blunt calcium entry via voltage sensitive calcium channels. Paralleling these traumatically induced changes, it is also well established that during ischemia, there are extensive ionic shifts which include the loss of calcium homeostasis. This is related to energy failure which inactivates energy dependent ion membrane pumps abolishing neuronal membrane potential and resulting in calcium influx via voltage operated calcium channels (VSCC) (62). Further, the ischemiamediated massive release of neurotransmitters also results in calcium influx via receptor operated calcium channels (ROCC). All these events result in loss of neuronal calcium homeostasis. In models of pure neuronal ischemia, this calcium dysregulation may not be of sufficient magnitude and duration to result in immediate death during the ischemic insult but may require further calcium entry during the postischemic period before lethal, levels of calcium toxicity are obtained and neuronal death ensues. It has been postulated that delayed neuronal death following cerebral ischemia is due to subsequent calcium entry via ROCCs and possibly VSCCs due to increased neuronal excitability to excitatory neurotransmitters (12,15,62). In contrast to ischemia, the immediate ionic events during traumatic brain injury are less well documented (29). It is our hypothesis that mechanical impact results in sufficient physical brain deformation to alter ion channel permeability and result in the transient loss of neuronal ion homeostasis. This, in turn, results in the excessive release of neurotransmitters which causes additional supra-physiological neuronal calcium entry possibly via both ROCCs and VSCCs and net excitation of certain neuronal populations. These normal populations may have long lasting changes in excitability thresholds and abnormal neuronal function as a consequence of aberrant stimulation of key regulatory calcium response elements. Such change may, in turn, render some neuronal populations more vulnerable to subsequent challenges that produce further loss of ion homeostasis and calcium regulation. In the traumatic or ischemic insults employed in our double insult paradigm neither insult alone is sufficient to produce over immediate or delayed neuronal death. Yet, when these insults are combined within an, as yet undefined, period of time, delayed neuronal death occurs in fasted rats. It is our hypothesis that this is the result of a cumulative calcium signal that reaches a lethal threshold during the postischemic period that can at least be partially attenuated by blocking muscarinic and NMDA receptors and reducing receptor operated and possibly voltage operated calcium entry as well. Duration of the posttraumatic increased ischemic sensitivity. Having determined that the traumatized brain is hypersensitive to a secondary ischemic insult, the question naturally arose regarding the temporal window during which the brain retained this increased posttraumatic sensitivity to secondary cerebral ischemia. To address this issue, the previously described studies of traumatic brain injury followed by cerebral ischemia were repeated in fasted rats, with the caveat that the induction of sublethal ischemia was now delayed for 24 hrs following traumatic brain injury in contrast to the one hour interval previously described. In such an approach, it was again recognized that CA1 cell loss occurred within the hippocampus (34). In this new paradigm which employed the induction of secondary ischemia 24 hrs postinjury, the use of J .T. Povlishock and L W Jenkins Pathobiology of Traumatic Brain Injury receptor antagonists also proved efficacious. Once again, scopolamine and phencyclidine administered 15 min pretrauma resulted in a significant reduction in the delayed CA1 neuronal cell death, although the degree of protection was not as great as seen in those animals subjected to traumatic brain injury followed by ischemia one hour later. Collectively, these studies utilizing cerebral ischemic insult initiated 24 hrs posttrauma plus receptor antagonism confirm that the traumatically induced hypersensitivity t o secondary cerebral ischemia persists for at least up to 24 hrs postinjury. Moreover, once again, it appears to implicate a receptor-mediated pathobiology, most likely triggered by the traumatic event. As the pharmacokinetics of the receptor antagonists utilized indicate that they have a relatively short half life, this also argues that their protective effect was exerted primarily on the traumatic event itself rather than some long-term effect on the delayed ischemic insult administered 24 hrs posttraumatic brain injury. What is unknown in this issue of delayed posttraumatic ischemia, centers on t h e brain's response to insults occurring beyond the 24 hrs time window. This is not an inconsequential issue as a number of laboratories evaluating postischemic sensitivity t o imposed secondary ischemia have shown that such strategies may evoke either hypersensitivity or induced tolerance dependent upon the temporal window between the primary and secondary ischemic insult (38,39). Altered postischemic sensitivity t o secondary imposed ischemia has been shown to have a receptormediated component, and interestingly, this response varies with time. When the secondary insult is imposed within 24 hrs of the primary ischemic injury, the brain appears hypersensitive to this insult. While, if more than 24 hrs elapse between insults, the brain becomes hyposensitive (38,39). Although there are considerable differences between the receptor-mediated injury that occurs with cerebral ischemia compared t o traumatic brain injury (29), such changes in posttraumatic ischemic sensitivity over time may also occur with traumatic brain injury and obviously warrant continued investigation. Summary and Conclusions Regarding Traumatically Induced Axonal Injury and the Brain's Hypersentivity to Secondary Ischemic Insult remain t o be clearly identified, so that therapies aimed at blunting these can be administered postinjury to block their damaging consequences. In the case of the brain's delayed hypersensitivity to cerebral ischemia, the window of vulnerability spans to 24 hours postinjury. As traumatically brain-injured patients are one of the few clinical populations in which ischemic events can be anticipated due to the development of secondary insults such as intracranial hypertension, it is obvious that more efforts must be directed at controlling and preventing these and other related secondary insults. In sum, against the backdrop of these two examples of traumatically induced pathobiology, it would seem most likely that many other events occuring after traumatic brain injury are also associated with evolving neuronal change and thus may be amenable t o therapeutic intervention. Therefore, we could argue that there is now legitimate hope for the better care and management of traumatically brain-injured humans, whose long-range outcome can no longer be envisioned t o be cast in stone at the moment of impact. Acknowledgements This work was supported by NIH grants NS20193 and R01-NS19550 and the Commonwealth Center for the Study of Brain Injury. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Brain Pathology Wiley

Are the Pathobiological Changes Evoked by Traumatic Brain Injury Immediate and Irreversible?

Brain Pathology , Volume 5 (4) – Oct 1, 1995

Loading next page...
 
/lp/wiley/are-the-pathobiological-changes-evoked-by-traumatic-brain-injury-AsHIztMXuZ

References (75)

Publisher
Wiley
Copyright
Copyright © 1995 Wiley Subscription Services, Inc., A Wiley Company
ISSN
1015-6305
eISSN
1750-3639
DOI
10.1111/j.1750-3639.1995.tb00620.x
Publisher site
See Article on Publisher Site

Abstract

Anatomy and Neurological Surgery, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA 23298-0709, USA Deptartment of Anesthesiology, University of Texas Medical Branch a t Galveston, Galveston, TX 77555-0830, USA more complex and progressive than previously envisioned, and as such, rejects many of the previous beliefs regarding the pathobiology of traumatic brain injury. Introduction Traumatic brain injury has long been thought t o evoke immediate and irreversible damage t o the brain parenchyma and its intrinsic vasculature. In this review, w e call into question the correctness of this assumption by citing t w o traumatically related brain parenchymal abnormalities that are the result of a progressive, traumatically induced perturbation. In this context, w e first consider t h e pathogenesis of traumatically induced axonal damage t o show that it is not the immediate consequence of traumatic tissue tearing. Rather, w e illustrate that it is a delayed consequence of complex axolemmal and/or cytoskeletal changes evoked by the traumatic episode which then lead to cytoskeletal collapse and impairment of axoplasmic transport, ultimately progressing t o axonal swelling and disconnection. Second, w e consider the traumatized brain's increased neuronal sensitivity t o secondary ischemic insult by showing that even after mild traumatic brain injury, C A I neuronal cell loss can be precipitated by the induction of sublethal ischemic insult within 24 hrs of injury. In demonstrating this increased sensitivity t o secondary insult, evidence is provided that it is triggered by the neurotransmitter storm evoked b y traumatic brain injury, allowing for sublethal neuroexcitation. In relation t o this phenomenon, the protective effect of receptor antagonists are discussed, as well as the concept that this relatively prolonged posttraumatic brain hypersensitivity offers a potential window for therapeutic intervention. Collectively, it is felt that both examples of the brain parenchyma's response t o traumatic brain injury show that the resulting pathobiology is much Corresponding author Dr John T Povlishock, Department of Anatomy, Virginia Commonwealth University, P O Box 980709, Richmond, VA 23298-0709. USA Tel + I 1804) 828-9535, Fax + I (804) 828-6293 Until recently most clinicians who routinely deal with traumatic brain injury have assumed that the contact and inertial phenomena associated with trauma physically tore neural and vascular elements, causing immediate structural damage and functional failure. Because of the perceived immediate and irreversible nature of these traumatically induced events, most felt that traumatic brain injury did not constitute an entity amenable to routine therapeutic intervention, and thus, for many, traumatic brain injury was viewed as a hopeless condition, in which the clinician is but a bystander. Despite the aura of negativism that has clouded the clinical impression of traumatic brain injury, recent evidence emerging from both the laboratory and clinical setting, suggests that these previously held concepts are no longer correct. Simply stated, it appears that, in the case of both neuronal and vascular elements, the forces associated with traumatic brain injury do not necessarily trigger immediate or irreversible damage. Rather, it appears that the forces of injury evoke more subtle cell membrane, cytoskeletal and synaptic changes that translate into a progression of reactive alterations, ultimately culminating in damage or dysfunction of the respective neural and/or vascular elements. In the current review, we focus on two examples of a delayed progression of traumatically induced reactive events that culminate in irreversible damage. First, we consider the pathobiology of traumatically induced axonal damage to demonstrate that such damage is not the consequence of immediate transection of the axon cylinder but rather the result of a progressive axonal change that leads to disconnection. Second, we consider the traumatized brain's hypersensitivity to delayed secondary cerebral ischemia to show that the neuronal cell loss found after such combined traumatic/ischemic insult is, in itself, a delayed event that can occur even when the onset of ischemia is postponed for up to 24 hrs post- J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Figure 1 (a) With anterograde peroxidase, focal axonal swelling (arrow) is recognized in an otherwise intact axon within 1 hr of the traumatic event (Dark Field Microscopy). XI ,000; (b)Twenty-four hours postinjury, one can see that the focal impairment of axoplasmic transport has led t o local axonal swelling and detachment. Note that the formed reactive axon swelling (RS) is encompassed by a distended and thinned myelin sheath (arrowheads). Also, note that many of the surrounding axons (AX) are unaltered. (Plastic thick section) X I ,000; (c) This section is taken from a control animal receiving intrathecal horseradish peroxidase. Note that the tracer diffuses through the extracellular space to concentrate at the Nodes of Ranvier (arrowheads) from where its passage into the axon is blunted by the intact axolemma. [plastic thick section.1 XI ,000; (d) Again, using intrathecally administered peroxidase, this thick section shows evidence of peroxidase passage into the axon cylinder following traumatic brain injury. Note the presence of peroxidase containing axons (block arrows) as well as other axons (AX) in the field showing no evidence of peroxidase uptake. X1.000. traumatic injury. These findings suggest that traumatic brain injury, alone, can perturb neurons sufficiently, without their physical disruption, to result in their altered sensitivity to a subsequently applied secondary ischemic insult. Taken together, these findings strongly suggest that, for the most part, the pathobiology of traumatic brain injury is not one of immediate and irreversible renting of neural elements. What follows are overviews of these axonal and neuronal abnormalities to fully illustrate these examples of evolving/ delayed damage occurring with traumatic brain injury. Traumatically Axonal Injury Induced Axonal Change/Diffuse Historically, it has been well recognized that axonal damage can be found in humans who sustain traumatic brain injury (2,3,4,5,6,52,63,64,65,45,46, 47,48,66). Such axonal damage is most commonly seen scattered throughout the brain of patients who show significant morbidity, without evidence of mass lesion formation or secondary insult, and these findings have given rise to the characterization of the condition as diffuse axonal injury (2,22), with the suggestion that the damaged axons are causative of the observed morbidity (2,22). The recognition of such axonal damage has been based primarily upon histological studies performed during postmortem analysis. In the first days postinjury, evidence of axonal damage has been based upon the histological recognition of swollen and/or truncated axonal segments which have been termed reactive axonal swellings or “retraction” balls (1). With continued posttraumatic survival, these swellings and balls are replaced by microglia, followed over time by Wallerian degeneration downstream from these sites (1). Because these reactive, swollen axons looked reminiscent of those profiles seen by Cajal and others J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Frequency Distrbuiton Interf-hment Distance Inter(iiamen1 thatmce (nm) Figure lie) This electron micrograph represents a color-enhanced image taken from an image analysis unit. Note that the neurofilaments (NF) are tightly packed and that they appear smooth showing no evidence of prominent sidearms. [The central bar was used for measurement.] X500.000 (f) This represents a frequency distribution plot of the interfilament distances in the control [non-peroxidase-containing] versus the experimental [peroxidase containing] axons. Note that in the experimental [peroxidasecontaining] axons, the frequency at which decreased interfilament distances occur is significantly increased in comparison to the controls; (9) This coronally sectioned electron micrograph represents a color-enhanced image taken from an image analysis unit. Note that in this axon, which displays no alteration in its axolemmal permeability, the neurofilaments (dots) and microtubules [boxes] demonstrate normal intra-axonal alignment and density. X45.000; (h) In contrast, this color-enhanced image, taken 5 mins after an injury, which evoked alterations in axolemmal permeability, n o w displays obvious compaction of neurofilaments as well as an apparent loss of microtubules. The hexagons were used in quantitative studies. X45.000. following physical transection of nervous tissue, it was assumed that the tensile forces of injury immediately tore axons, causing them to retract and expel their axoplasm to form the reactive swelling or "retraction" ball. The correctness of this assumption was first called into question by our laboratory in the early 1980's when we attempted to better understand the pathobiology of this process utilizing well-controlled animal models of injury (54). Using fluid-percussion brain injuries in cats, we devised multiple strategies to dissect out if the traumatic event was capable of immediately tearing the axon cylinder and,extruding its axoplasmic mass. In these studies, we relied primarily upon anterograde tracers which were delivered to the somata of the major projectional systems from where they were moved into the axon cylinder via anterograde axonal transport (54). While such transport was underway, the same animals were then subjected to mild traumatic brain injury, and the intra-axonal course of the anterogradely transported tracer was followed. It was assumed that, if the injury was capable of immediately tearing the axon and extruding its axoplasmic mass, we would be able to recognize, at some point along the axon's length, a focus showing disruption, with a burst of anterogradely transported tracer forming a truncatediswollen ball. Despite exhaustive efforts to identify such sites, no evidence of such direct tearing with axoplasmic extrusion was found. Rather, using light and electron microscopy, we recognized that the traumatic episode resulted in a perturbation of axoplasmic transport, at one point, along the axon's length (54). Such impaired transport could be recognized within the first hour postinjury (Fig. la), and over time, this site of impaired transport was associated with increased axonal swelling followed by local lobulation of the axon. Over time, 3-6 hours, J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury this lobulated segment was associated with further swelling, with ultimate axonal detachment at this site over a 12-24 hour posttraumatic period (54). With continued survival, the detached segment, in continuity with the staining soma, underwent continued expansion and swelling due to delivery of material via anterograde transport, and by 24 hours, the proximal disconnected segment now formed a grossly swollen reactive axon (Fig. l b ) entirely consistent with the reactive swellings/”retraction balls” long associated with traumatic brain injury. While this observation was made initially in the case of mild traumatic brain injury, it was subsequently confirmed in moderate and severe traumatic brain injury and was also confirmed in axons of both the motor and special sensory systems (11,19,56). Subsequent studies conducted by other investigators using different species and models further confirmed the correctness of our observations by demonstrating a comparable repertoire of impaired axoplasmic transport, axonal swelling, and disconnection (21,23,41). Gennarelli and colleagues (21,23), utilizing an inertial injury model in nonhuman primates and a guinea pig optic nerve stretch model further characterized by Maxwell and colleagues (41), found a similar progression of reactive change. That the progressive/evolving axonal changes found in experimental animals, are found in brain-injured humans has also been confirmed. Although one cannot directly assess anterograde transport in traumatically brain-injured man, immunocytoskeletal markers of cytoskeletal movement and/or other anterogradedly transported materials have been used to evaluate the progression of traumatically induced axonal change. Specifically, using antibodies targeted to the . cytoskeletal neurofilament (NF) subunits, we have confirmed, in traumatically braininjured humans, a comparable sequence of local axonal swelling, lobulation and disconnection, with subsequent continued swelling of the detached axon (13,ZS). Similarly, in elegant studies by Sherriff et al. (59,60,61) and Gentleman et al. (24), utilizing antibodies to the P-amyloid precursor protein, a substance known to move by anterograde axoplasmic transport, a comparable progression of events has been shown through the light microscopic examination of this protein at varying survival periods postinjury. Clearly, based upon the above, there is now consensus, in both animals and man, that traumatic brain injury can impair axoplasmic transport and result in a progression of changes leading to continued axonal swelling ultimately followed by axonal detachment. The key questions that now remain regard precisely what triggers or initiates this impaired axoplasmic transport and what then influences the temporal progression of the abovedescribed events. Znitiating mechanisms of axonal injury. In regards to those factors capable of initiating impaired axoplasmic transport, multiple processes may be implicated. On one hand, one could argue that the traumatic episode focally stretches the axon and thereby alters the axolemma. This could then allow for the influx of normally excluded ions, causing a disruption of intra-axonal ionic homeostasis to result in impairment of axoplasmic transport. Conversely, one could argue that the tensile forces of injury directly stretch the cytoskeletal framework of the axon cylinder, causing its disorganization and/or collapse to thereby elicit an impairment of axoplasmic transport. Fortunately, within the last two years, considerable insight has been gained into each of these mechanisms and their potential role in the genesis of the reactive axonal change associated with traumatic brain injury. In regards to the potential for traumatically induced alteration in the axolemma, evidence from our laboratory suggests that such axolemmal change occurs, and as such, may be a player in the pathobiology of reactive axonal change. Following upon the seminal observation of Maxwell et al. (43), that, in the most severe forms of traumatic brain injury, a subpopulation of fine caliber axons reveals direct axolemmal renting, we investigated the possibility for more subtle axolemmal change in axons, harvested from animals sustaining less severe injury (49,SO). In this approach, we utilized extracellular tracers normally excluded by the intact axolemma to determine if, with injury, these tracers were capable of moving from the extracellular compartment to the axoplasmic front. Utilizing intrathecally administered horseradish peroxidase and microperoxidase which are normally confined to the extracellular compartment (Fig. lc), we observed, with moderate to severe traumatic brain injury, focal alterations in the axolemma’s permeability to these tracers (Fig. Id). These tracers were seen in the axon cylinder within minutes of the traumatic episode and were observed at nodal, paranodal and internodal sites (49). In this sequence of events, it also appeared that the traumatic episode disrupted the normal tight alignment of the myelin sheath with the axolemma, permitting the tracer to move along the axolemma to the precise site of maximal axolemmal perturbation. When these sites were sampled for ultrastructural analysis, the tracers were identified within the axoplasm, where they were associated with a distinct subset of intra-axonal changes. Specifically, within minutes of the traumatic episode, the influx of tracer was associated with local mitochondrial swelling as well as an apparent compaction of the local neurofilament network (49). This compaction was confirmed by computer assisted reconstruction, wherein the neurofilaments revealed a dramatic reduction in interfilament distance, together with a loss of the neurofilament sidearms (49) (Figs. le,f). Interestingly, this mitochondrial swelling and neurofilament compaction remained relatively stable J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury over a several hour posttraumatic period with swellings developing, over time, rostra1 and caudal to this site of NF compaction. Following upon our initial description of neurofilament compaction and sidearm loss, made in sagittally sectioned axons, we repeated these studies in coronally sectioned axons both to gain better insight into the nature of these neurofilament changes and to better appreciate any related microtubular changes which are best seen when the microtubules are viewed en face (50,53). Once again, utilizing computer-assisted reconstruction of the axon cylinder, with counting of the neurofilament and microtubular density, we confirmed that the traumatic event was associated with focal changes in axolemmal permeability that resulted in immediate neurofilament compaction, with sidearm loss (50,531 (Figs. lg, h). Moreover, through the use of these coronally sectioned profiles, we also observed a rapid loss of the related microtubular network, suggesting that the axonal injury was associated with a complete collapse of the local cytoskeleton (50). As noted above, this cytoskeletal collapse remained relatively stable over a several hour period, and then was associated with an upstream and downstream swelling of the axon. This was most likely related to the fact that the cytoskeletal collapse disrupted those axoplasmic constituents (the microtubules) known to be related to anterograde and retrograde transport (28). Collectively, then, in the case of moderate to severe injury, evidence points to perturbation of axolemmal permeability which either directly or indirectly leads to disruption of those subcellular mechanisms involved in axoplasmic transport. The correctness of this belief was further substantiated by recent work conducted by Dr. William Maxwell and colleagues in an optic nerve stretch model of injury (40). Unlike the large caliber, heavily myelinated axons assessed in our studies, Maxwell and collaborators (40) focused on optic nerve fibers which, by their nature, are fine caliber, thinly myelinated axons, in which microtubules rather than neurofilaments predominate. Subjecting these axons to tensile loading by stretch, Dr. Maxwell performed a quantitative analysis of the cytoskeleton and observed a striking loss of microtubules in those axons that ultimately went on to show impaired axoplasmic transport, swelling and disconnection (40). While Maxwell and colleagues failed to show neurofilament compaction comparable to that described in our studies (49,50,51,53), this is not surprising. As these fine caliber axons possess proportionally few neurofilaments, the previously described neurofilament changes may be reserved only for neurofilament rich, large caliber axons. Irrespective of these differences, however, both studies point to traumatically induced alteration of the cytoskeleton which most likely, through microtubular damage, translates into impaired axoplasmic transport and axonal swelling. The key question that remains, in relation to the above-described pathobiology, is precisely what extracellular factor(s) migrate(s) through the altered axolemma to trigger the above-described cytoskeletal changes. In most people's mind, the obvious candidate is calcium, with the suggestion that an increase in intra-axonal calcium would lead to the activation of proteases which, in turn, would lead to the dissolution of the cytoskeleton. While such arguments are credible, it is important to note that, although the traumatic brain injury can result in rapid change within the axonal cytoskeleton, it does not typically result in rapid cytoskeletal dissolution or granulation as described with ischemic insult or other CNS pathologies, linked with calcium influx and calcium-mediated proteolysis (7,8,18,58,67,78, 69). Because we do not routinely observe such rapid cytoskeletal dissolution in the case of traumatic brain injury, we initially posited that the pathobiology associated with axonal injury was not calcium linked or alternatively, if calcium was involved, it was operating through previously unidentified mechanisms (56). However, because of the apparent differences between our views in traumatic brain injury and those described for other forms of CNS insult, we naturally questioned whether some of our observations were correct, or rather reflected some unique aspect of the chosen experimental animal model. This concern was obviated by several recent studies wherein the above-described sequence of rapid neurofilament compaction and sidearm loss were described using neurofilament rod and sidearm specific antibodies in addition to Western blotting in lampreys subjected to spinal cord transection (27). Also, using these same antibodies targeted to the neurofilament sidearm specific and rod domains, we have been able to confirm the same sequence for neurofilament compaction and sidearm loss in a rodent impact acceleration model of traumatic brain injury (57). In these same rodent studies, we also confirmed that the focus of neurofilament compaction and sidearm loss as well as microtubular change were directly related to sites of altered axolemmal permeability (57). Thus, from multiple species and animal models, there is compelling evidence that the traumatic episode can result in neurofilament compaction without overt dissolution of the cytoskeleton, although microtubular loss can be identified. Calcium could be implicated in this process as the rapid influx of calcium could lead to microtubular disruption and neurofilament sidearm cleaving and neurofilament compaction, with the compacted neurofilaments, remaining resistant to further calcium-mediated proteolysis (27). Alternatively, it is also possible that the flux of calcium may be transient and thereby incapable of causing further degradation of the cytoskeleton. In the case of traumatic brain injury, the issue of calciummediated damage may be more complex than in other experimental conditions and perhaps J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury Figure 2 (a) This plastic thick section demonstrates that, with mild traumatic brain injury, reactive axonal change (RA)can occur without any evidence of altered axolemmal permeability to intrathecally applied horseradish peroxidase (arrowheads). XI ,000. (b) In this schematic illustration, w e attempt to portray those axonal events ongoing with traumatic brain injury of varying severity that lead to the formation of reactive axonal swellings. In the first panel, one can see a normal axon profile (A) and an enlargement of its basic cytoskeletal detail showing aligned neurofilaments (NF) with prominent side arms and microtubules (MT). With mild to moderate injury (B&C), one can see that the traumatic episode focally deforms the axon and translates into primary cytoskeletal (neurofilament and microtubule) misalignment, which then translates into focal impairment of axoplasmic transport. Continued anterograde transport of axoplasm and organelles leads to accumulation of these components a t the site of cytoskeletal injury, with progressive convergence of the axolemma (not shown), and disconnection of the proximal axonal segment (D) from its distal appendage. In this process, it is important to note that, although the traumatic event initially can be associated with infolding of the axolemma, there is no evidence of overt axolemmal perturbation. With more severe injury (B&E), focal intra-axonal abnormalities again occur; however, in this case, they involve concomitant changes within the axolemma and the related cytoskeleton. As illustrated, the traumatic episode is associated with altered axolemmal permeability (arrows), as well as rapid neurofilament side arm loss and subsequent neurofilament compaction (enlargement). Further, in the same field, microtubules are lost, and the local mitochondria swell. It is envisioned that all of these axonal changes are a direct reflection of consequences triggered by the altered axolemmal permeability which is assumed to initiate a generalized cytoskeletal collapse leading to a focal impairment of axoplasmic transport, localized swelling of the axon, with pinching off of the axon at the swollen site, and subsequent disconnection (F). In this illustration, the related myelin sheaths are not portrayed. J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury constitutes a manifestation not only of the presence of calcium, but also its concentration and duration of delivery via the altered axolemma. Recently, we have shown that some traumatically injured axons can remain permeable to extracellular tracers for several hours postinjury, suggesting that there may be a continuous and, perhaps, slow leak of the membrane, allowing for a continuing influx of calcium (51, 5 3 ) . While, obviously, more remains to be done in relation to the calcium story in traumatic brain injury, other recent studies serve to further illustrate the complexity of the issues at hand. Specifically, Maxwell et al. (42), have shown, in the stretched optic nerve, that injury results in the loss of the calcium-ATPase activity at the nodal axolemma and the loss of the ecto-Ca-ATPase in the outer aspect of the myelin sheath. Collectively, these events were associated 4-6 hours later with a delayed dissolution of the axonal cytoskeleton. Although traumatically induced injury is seen here to disrupt the cytoskeleton, it still constitutes a delayed event in contrast to other experimental conditions. Further, since it involves the perturbation of multiple mechanisms involved in the of flux and compartmentalization of calcium, the potential complexity of traumatically induced damage becomes even more apparent. thereby translate into impaired axoplasmic transport. Therefore, it appears that there are at least two initiating mechanisms in the pathobiology of reactive axonal change, one involving axolemmal/ cytoskeletal interaction and the other involving direct cytoskeletal perturbationhisalignment (Fig. 2b). The validity of this in vivo interpretation is indirectly substantiated by in vitru studies which show that differential tensile loading of isolated axons can result in a repertoire of change comparable to that described above (20). Specifically, with low tensile loading, axons, assessed in vitru, show axoplasmic disruption independent of any change in axolemmal integrity, entirely consistent with those changes described with mild traumatic brain injury. With more severe injuries, on the other hand, these same axons now reveal axolemmal change that correlates with dramatic axoplasmic failure (20). As these differential in vitru findings are consistent with our observations made in vivo, they tend to confirm our belief that the initiating pathobiology of traumatically induced reactive axonal change is varied and may be a function of the severity of tensile loading to which the axon is initially subjected. Possible diversity of the axonal response to injury. Although, in the case of moderate to severe traumatic brain injury, the evidence is mounting in support of the above-described scenario of axolemmal perturbation and cytoskeletal collapse, it is important to note that these same initiating factors may not be operant across the full spectrum of traumatic brain injury. In this regard, it is important to note that, in the same experiment described above using extracellular peroxidase, we also observed, with milder traumatic brain injuries, a population of axons that demonstrated reactive change without any evidence of axolemmal perturbation reflected in peroxidase passage from the extracellular to axoplasmic compartment (49) (Fig. 2a). Further in these axons, no evidence of neurofilament compaction or microtubular loss were observed. Rather, it appeared that the traumatic episode caused a direct focal misalignment of the neurofilament network which, over time, translated into impaired axoplasmic transport, swelling and disconnection. In these axons, the neurofilament misalignment and disorganization could be seen within 12 minutes of injury, and these sites were soon associated with a focal accumulation of organelles long linked to impaired axoplasmic transport (49). Precisely how this neurofilament disorganization translates into impaired axoplasmic transport is unknown. However, as the neurofilaments and microtubules are structurally aligned and integrated, it would seem logical that misalignment of one cytoskeletal element would cause comparable misalignment of others and Temporal course of reactive axonal change. One of the most puzzling aspects of reactive axonal change in both animals and man, centers on the temporal progression of this change from the moment of initiation to its termination as a fully formed reactive axonal swelling. While typically we have described the progression of reactive axonal change over a 3-6 hour time period, with fully mature reactive swellings seen within 12-24 hours of injury, we must admit that this temporal framework is not cast in stone, and indeed shows variation. In fact, Maxwell and colleagues have shown disconnection and the formation of mature reactive swellings in the optic nerve of the guinea pig with 6 hours of injury (41). Unfortunately, at present, little additional experimental evidence exists to shed light on to this issue. To add to this problem, it is our impression that the temporal sequence of reactive axonal change may reflect the species used (56). Although not supported by statistical studies, we have the strong impression that the temporal progression of reactive axonal change is more accelerated in rats than in micropigs and humans (56). It is unclear if this is an issue of species specificity, or alternatively a function of the distance between the sustaining soma and the reactive axon, which would naturally be much shorter in the rat than in the micropig or human. For example, reactive axonal change in the corticospinal tract of a rat will involve a pathway with a projection distance a fraction of that seen in higher order animals or man. Also related to this issue of the temporal progression is the possibility that this progression may be influenced by the severity of the injury. That is to say that the more severely injured J.T. Povlishock and L W. Jenkins: Pathobiology of Traumatic Brain Injury axons may undergo a more rapid progression of change than less severely injured axons. To date, little evidence exists to support either of these assumptions, and obviously this is an area which mandates continued investigation. Site of axonal failure. No consensus has been reached on precisely what axonal site is most vulnerable to the forces of injury. Many advocate that the nodal region is exquisitely sensitive due to the forces of injury, and thus, constitutes the most characteristic site of axolemmal and cytoskeletal failure (5,23). Our observations, on the other hand, do not support this assumption, as we have consistently found reactive axonal change, reflected in axolemmal and/or cytoskeletal abnormality within nodal, paranodal and internodal domains (11,19,49,53,54,55,56). In our estimation, axons fail at the site of maximal tensile loading and not at any preordained site. It should be noted that most of the fibers undergoing reactive axonal change are large caliber, long-track axons that are most likely subjected to maximal strain at one point along their length. Typically, in these fibers, reactive axonal swellings are seen at points where the axons decussate, cross blood vessels, or change their intraaxial course, suggesting that these sites are most prone to failure under maximal tensile loading. This implies, therefore, that the actual site of maximal tensile loading is more important than the presence or absence of a node of Ranvier. While we cannot preclude the fact that the node may be the most sensitive site for potential failure, we would argue that large caliber axons which are separated by large internodal distances most commonly do not fail at the nodal domains. Evidence to support.such an argument is also found in in vitro studies wherein mechanical stretching of peripheral axons has been shown to result in their collapse at internodal and paranodal segments rather than at nodal sites (Saatman, personal communication). Evidence for the Traumatized Brain's Hypersensitivity t o Delayed Secondary Cerebral Ischemia the posttraumatic secondary insults exacerbate ongoing pathologies found after traumatic brain injury, our studies have taken a slightly different focus, arguing that traumatic brain injury can also alter the neurons in such a fashion to display an increased sensitivity to a secondary ischemic insult which would have been well tolerated in the nontraumatized state. In the introduction, it was noted that the pathobiological sequelae of traumatic brain injury are not always immediate and irreversible and in this context, the delayed and progressive pathology associated with traumatically induced axonal injury was considered. Now, we turn our attention to the issue of the brain parenchyma's sensitivity to secondary ischemic insult. Many have recognized that secondary insults contribute substaqtially to the morbidity associated with human traumatic brain injury (9,26,37,44,26,53), and moreover, laboratory evidence has also shown that secondary insults such as hypertension, hypoxia, and global ischemia worsen the bioenergetic, electrophysiologic and behavioral status of animals subjected to traumatic brain injury (16,30,31,32). While most advocate that Evidence for the traumatically injured brain's hypersensitivity to delayed cerebral ischemia. Although this issue was first considered in cats (39, it was most critically assessed in rodent models of traumatic brain injury in which mild traumatic brain injury was complicated by imposed forebrain ischemia (33,34,36). In this approach, fasted rats were subjected to mild fluid percussion brain injury (17) of sufficient severity to evoke behavioral suppression while eliciting no overt change within the brain parenchyma, as defined by traditional histopathological studies of both the neuronal somata and their axonal appendages. These mildly injured animals were then subjected, one hour later, to the onset of 6 min of transient forebrain ischemia, accomplished by bilateral carotid clamping and systemic hypotension. In addition to assessing each of these insults in combination, each insult was individually assessed to fully appreciate the respective pathobiology of each insult. In this context, we observed, as anticipated, that mild traumatic brain injury was associated with no histological change within the brain parenchyma. Similarly, when the 6 min of transient forebrain ischemia was studied in isolation, it, too, was observed to result in no morphological change. However, when the traumatic insult was combined with an ischemic insult one hour postinjury, extensive bilateral CA1 hippocampal cell loss was found at 7 days postinjury, but not within the first day of survival (33,34,36). Through this paradigm, we recognized that mild brain injury, in some fashion, enhanced selectively vulnerable and regionally restricted ischemic delayed neuronal cell death over 7 days in the absence of any initial overt pathological change within these loci. Mechanisms involved in the induction of the brain's hypersensitivity to secondary cerebral ischemia. Following upon our observation that mild traumatic brain injury can enhance the potential for delayed neuronal cell death following ischemic insult, the obvious question arose as to what could be causing the brain's hypersensitivity to this secondary cerebral ischemia. At that time, it was becoming well established that a mild traumatic brain injury, not capable of evoking overt structural cell change, was able to elicite a net phase of neuroexcitation in which multiple transmitters were released to participate in receptor- mediated pathology throughout the brain (29). Against this backdrop, it J.T. Povlishock and L.W. Jenkins: Pathobiology of Traumatic Brain Injury was postulated that the increased regional, delayed neuronal cell death found in our studies was consistent with a sublethal excitotoxic process which occurred with mild traumatic brain injury, and which was then potentiated to lethality by a second, normally sublethal-ischemic initiated excitotoxic process. Additionally, since the observed selected neuronal vulnerability was similar to the pattern of delayed ischemic neuronal necrosis found in pure forebrain ischemia in fasted rats (14) in which there is substantial evidence for excitotoxicity and calcium mediated mechanisms (15,62), there appeared to be compelling evidence for the role of excitotoxic processes in our observed pathobiological responses. This issue was critically addressed by the repetition of the above-described experiments, now using multiple receptor antagonists (34). For reasons, which will be detailed below, we first used lmg/kg of scopolamine and 4 mg/kg of phencyclidine which were administered 15 mins prior to the induction of mild traumatic brain injury, followed one hour later by secondary cerebral ischemia as described previously. Through this approach, we found the use of the competitive muscarinic (scopolamine) and noncompetitive NMDA (phencyclidine) antagonist provided substantial protection against the previously observed delayed CA1 neuronal death typically seen at 7 days after a traumatic injury complicated by cerebral ischemia (34). Interestingly, in this approach, it was also found that the drug action was effective whether it was given prior to or during both trauma and the imposed forebrain ischemia. The rationale for antagonizing both receptor populations, i.e. the muscarinic and NMDA receptors, was based upon evidence that the ion channel controlled by the NMDA receptor is voltage sensitive and that maximum calcium influx via this channel occurs when the membrane potential is decreased or depolarized. The magnesium block is removed from the NMDA channel in proportion to the membrane voltage decrease. Obviously, then by blocking the action of either acetylcholine acting on the muscarinic receptor or other excitatory transmitters which contribute to CA1 depolarization, one may inhibit the removal of the NMDA channel magnesium block and thereby blunt calcium entry via voltage sensitive calcium channels. Paralleling these traumatically induced changes, it is also well established that during ischemia, there are extensive ionic shifts which include the loss of calcium homeostasis. This is related to energy failure which inactivates energy dependent ion membrane pumps abolishing neuronal membrane potential and resulting in calcium influx via voltage operated calcium channels (VSCC) (62). Further, the ischemiamediated massive release of neurotransmitters also results in calcium influx via receptor operated calcium channels (ROCC). All these events result in loss of neuronal calcium homeostasis. In models of pure neuronal ischemia, this calcium dysregulation may not be of sufficient magnitude and duration to result in immediate death during the ischemic insult but may require further calcium entry during the postischemic period before lethal, levels of calcium toxicity are obtained and neuronal death ensues. It has been postulated that delayed neuronal death following cerebral ischemia is due to subsequent calcium entry via ROCCs and possibly VSCCs due to increased neuronal excitability to excitatory neurotransmitters (12,15,62). In contrast to ischemia, the immediate ionic events during traumatic brain injury are less well documented (29). It is our hypothesis that mechanical impact results in sufficient physical brain deformation to alter ion channel permeability and result in the transient loss of neuronal ion homeostasis. This, in turn, results in the excessive release of neurotransmitters which causes additional supra-physiological neuronal calcium entry possibly via both ROCCs and VSCCs and net excitation of certain neuronal populations. These normal populations may have long lasting changes in excitability thresholds and abnormal neuronal function as a consequence of aberrant stimulation of key regulatory calcium response elements. Such change may, in turn, render some neuronal populations more vulnerable to subsequent challenges that produce further loss of ion homeostasis and calcium regulation. In the traumatic or ischemic insults employed in our double insult paradigm neither insult alone is sufficient to produce over immediate or delayed neuronal death. Yet, when these insults are combined within an, as yet undefined, period of time, delayed neuronal death occurs in fasted rats. It is our hypothesis that this is the result of a cumulative calcium signal that reaches a lethal threshold during the postischemic period that can at least be partially attenuated by blocking muscarinic and NMDA receptors and reducing receptor operated and possibly voltage operated calcium entry as well. Duration of the posttraumatic increased ischemic sensitivity. Having determined that the traumatized brain is hypersensitive to a secondary ischemic insult, the question naturally arose regarding the temporal window during which the brain retained this increased posttraumatic sensitivity to secondary cerebral ischemia. To address this issue, the previously described studies of traumatic brain injury followed by cerebral ischemia were repeated in fasted rats, with the caveat that the induction of sublethal ischemia was now delayed for 24 hrs following traumatic brain injury in contrast to the one hour interval previously described. In such an approach, it was again recognized that CA1 cell loss occurred within the hippocampus (34). In this new paradigm which employed the induction of secondary ischemia 24 hrs postinjury, the use of J .T. Povlishock and L W Jenkins Pathobiology of Traumatic Brain Injury receptor antagonists also proved efficacious. Once again, scopolamine and phencyclidine administered 15 min pretrauma resulted in a significant reduction in the delayed CA1 neuronal cell death, although the degree of protection was not as great as seen in those animals subjected to traumatic brain injury followed by ischemia one hour later. Collectively, these studies utilizing cerebral ischemic insult initiated 24 hrs posttrauma plus receptor antagonism confirm that the traumatically induced hypersensitivity t o secondary cerebral ischemia persists for at least up to 24 hrs postinjury. Moreover, once again, it appears to implicate a receptor-mediated pathobiology, most likely triggered by the traumatic event. As the pharmacokinetics of the receptor antagonists utilized indicate that they have a relatively short half life, this also argues that their protective effect was exerted primarily on the traumatic event itself rather than some long-term effect on the delayed ischemic insult administered 24 hrs posttraumatic brain injury. What is unknown in this issue of delayed posttraumatic ischemia, centers on t h e brain's response to insults occurring beyond the 24 hrs time window. This is not an inconsequential issue as a number of laboratories evaluating postischemic sensitivity t o imposed secondary ischemia have shown that such strategies may evoke either hypersensitivity or induced tolerance dependent upon the temporal window between the primary and secondary ischemic insult (38,39). Altered postischemic sensitivity t o secondary imposed ischemia has been shown to have a receptormediated component, and interestingly, this response varies with time. When the secondary insult is imposed within 24 hrs of the primary ischemic injury, the brain appears hypersensitive to this insult. While, if more than 24 hrs elapse between insults, the brain becomes hyposensitive (38,39). Although there are considerable differences between the receptor-mediated injury that occurs with cerebral ischemia compared t o traumatic brain injury (29), such changes in posttraumatic ischemic sensitivity over time may also occur with traumatic brain injury and obviously warrant continued investigation. Summary and Conclusions Regarding Traumatically Induced Axonal Injury and the Brain's Hypersentivity to Secondary Ischemic Insult remain t o be clearly identified, so that therapies aimed at blunting these can be administered postinjury to block their damaging consequences. In the case of the brain's delayed hypersensitivity to cerebral ischemia, the window of vulnerability spans to 24 hours postinjury. As traumatically brain-injured patients are one of the few clinical populations in which ischemic events can be anticipated due to the development of secondary insults such as intracranial hypertension, it is obvious that more efforts must be directed at controlling and preventing these and other related secondary insults. In sum, against the backdrop of these two examples of traumatically induced pathobiology, it would seem most likely that many other events occuring after traumatic brain injury are also associated with evolving neuronal change and thus may be amenable t o therapeutic intervention. Therefore, we could argue that there is now legitimate hope for the better care and management of traumatically brain-injured humans, whose long-range outcome can no longer be envisioned t o be cast in stone at the moment of impact. Acknowledgements This work was supported by NIH grants NS20193 and R01-NS19550 and the Commonwealth Center for the Study of Brain Injury.

Journal

Brain PathologyWiley

Published: Oct 1, 1995

There are no references for this article.