Percussion Concussion

The fluid percussion (FP) injury model produces a brief (22 msec) deformation of the exposed cortex. A craniotomy is performed, and a plastic modified Luer-Lok injury tube is attached to the exposed dura and sealed. A Plexiglas cylinder filled with sterile saline is attached to the injury tube via a metal fitting. A pendulum is released on the other end and at impact produces a pressure wave that results in a bolus injection of saline onto the exposed dura. By varying the distance of the pendulum swing, injury severity—mild (1.1-1.3 atm), moderate (2.0-2.3 atm), and severe (2.4-2.6 atm)—can be studied in a reproducible fashion. This is an important model characteristic because of the heterogeneous nature of human TBI and the possibility that treatment strategies may vary with injury severity. The central, lateral, and parasagittal FP models are characterized by brief behavioral responsiveness (e.g., coma), metabolic alterations, changes in local cerebral blood flow and blood-brain barrier (BBB) permeability, and behavioral deficits. The central FP model tends to have variable and small contusions in the vicinity of the fluid pulse, and scattered axonal damage mostly limited to the brainstem. In contrast, lateral and parasagittal FP is characterized by a lateral cortical contusion that is remote from the impact site. Evidence of axonal damage is seen throughout the white matter tracts in the ipsilateral cerebral hemisphere, and tissue tears are seen at gray matter/white matter interfaces. Hippocampal damage is pronounced in the lateral and parasagittal FP injury models, with little brainstem damage. The FP injury model thus produces a range of disorders including contusion, widespread axonal injury, and selective neuronal necrosis.

Another commonly used rodent TBI model is controlled cortical impact injury. In this model, a bone flap is removed, and the impact device is vertically driven into the cerebral cortex to produce tissue displacement. This model produces a well-demarcated cortical contusion with variable degrees of hippocampal involvement dependent on velocity and deformation depth. An advantage of the controlled cortical impact model is that it can be used in mice and allows the testing of genetically altered mice to help determine the cause-and-effect relationships between gene expression and cell injury.

Acceleration Concussion

Inertial acceleration models can produce pure acute subdural hematomas and diffuse axonal injury. These models accelerate the head of the animal in one plane, followed by a rapid deceleration, resulting in movement of the brain within the cranial cavity. The inertial acceleration models are designed to mimic motor vehicle accidents. Tissue tear hemorrhages occur in the central white matter, and gliding contusions occur in the parasagittal gray matter/white matter junctions. These models are characterized by a variable period of coma and axonal damage in the upper brainstem and cerebellum. For the impact acceleration model, the animal is placed on a foam bed with a metal disk attached to the exposed skull, and a weight is dropped from a specified height to produce brain injury. This model also produces prolonged coma and widespread axonal damage. However, these models are characterized by variable and somewhat uncontrolled skull fractures.

Human head injury is never as pure as an experimental model, and the total human injury condition may not be addressed adequately with a single animal model. Therefore, the use of complicated models in assessing pathomechanisms after TBI can provide valuable data (Statler et al., 2001). Many of these studies have added secondary insults such as hypotension (shock), hypoxia, or hyperthermia after the primary injury to more adequately mimic human head trauma. However, this added component to the model can result in complex data interpretation. Thus, once a particular feature of human brain injury is produced in an experimental model, the pathogenesis of the injury can be critically investigated.

Military Models

The incidence of brain injury is currently estimated to be as high as 40% to 70% of combat casualties. Advancements in protective body armor have limited the number of high-velocity impact injuries to the brain, but the use of improvised explosive devices has caused blast injury to become more prevalent. Military-relevant models have only recently been developed, and much work is needed to fully explore pathomechanisms induced by these models. Two models of significance are the penetrating ballistic brain injury model and a blast injury model using a metal tube with a small amount of plastic explosives at one end. The penetrating brain injury (PBI) model mimics a bullet wound by expanding and contracting a balloon that has been inserted into the brain (Williams et al., 2005). Injury severity and placement of the probe can be varied so that different types of bullet injuries can be studied. PBI produces robust histopathological damage and increases in intracranial pressure (ICP) and hemorrhage. In addition, sensorimotor deficits are observed in this model, along with seizure activity. Blast injury models have been developed that can mimic whole-body blast or local blast injury (Cernak and Noble-Haeusslein, 2010). This type of trauma produces cognitive deficits, with ultrastructural and oxidative damage to the hippocampus. Axonal damage is also present, as indicated by increases in phosphorylated neurofilament proteins. These military-relevant injury models have the potential to provide valuable information for managing and treating combat casualties.

In Vitro Models

A shortcoming of animal models is that they preclude a critical assessment of individual cell responses to trauma. In animal experiments, for example, the cellular response to injury may be a consequence of both primary and secondary events initiated by a complex cascade of cellular interactions. To critically investigate the consequences of injury on a specific cell type in the absence of confounding cellular and systemic factors, several in vitro cell culture models have been developed. Models range from scratching the culture with a pipette tip to inducing cellular deformation by stretching cultured cells.

Using these approaches, investigators have found that trauma induces a wide range of primary cellular alterations. Astrocytic responses include hyperplasia, hypertrophy, and increased glial fibrillary acidic protein content. Increases in intracellular calcium occur, which are blocked by specific receptor antagonists. Traumatized astrocytes also produce interleukins and neurotrophic factors. Neonatal cortical neurons that are stretched undergo delayed depolarization that depends on the activation of specific receptor populations. Combined mechanical trauma and metabolic impairment in vivo also induces N-methyl-d-aspartate receptor–dependent neuronal cell death and caspase-3–dependent apoptosis. In vitro experimental approaches provide novel data concerning intracellular signaling cascades, mechanisms underlying cellular responses to trauma, and the role of specific cell types in the pathophysiology of brain trauma (Chen et al., 2009; LaPlaca et al., 2009).

Temporal Patterns of Neuronal Death

The neuropathological sequelae of experimental and human TBI have been well described (Bramlett and Dietrich, 2004; Povlishock and Katz, 2005). In experimental TBI, temporal patterns of neuronal damage have also been characterized. As early as 6 hours after cortical contusion injury, the contused tissue appears edematous, and pyknotic neurons are apparent at the injury site. By 8 days, a cortical cavity has developed that is surrounded by a border containing necrotic tissue, a glial scar, or both. The temporal profile of neuronal damage after parasagittal FP brain injury has also been assessed with light and electron microscopy. As early as 1 hour after impact, dark shrunken neurons indicative of irreversible damage are seen in cortical layers overlying the gliding contusion that displays BBB breakdown to protein tracers. Ultrastructural studies demonstrate that early BBB dysfunction results from mechanical damage of small venules in vulnerable regions including the external capsule. In some brain regions, focal sites of acute neuronal damage are associated with extravasated protein, whereas neuronal damage in other regions appears to occur without overt BBB breakdown. Astrocytic swelling is observed early after injury, with increased glial fibrillary acidic protein immunoreactivity apparent at later times in areas, demonstrating histopathological damage (Fig. 50A.1). In terms of neuroprotection, the acuteness of this damage limits the potential for therapeutic interventions directed against the early neuronal and glial response to TBI.

Fig. 50A.1 Three days after fluid percussion injury, glial fibrillary acidic protein–positive astrocytes and processes are prominent in the hippocampus (×1200).

More subacute patterns of neuronal injury have been documented in various TBI models. At 3 days after moderate parasagittal FP brain injury, scattered necrotic neurons are present throughout the frontoparietal cerebral cortex remote from the impact site (Fig. 50A.2). In addition, selective neuronal damage is seen in the CA3 and CA4 hippocampal subsectors, the dentate hilus, and lateral thalamus ipsilateral to the trauma. These patterns of selective neuronal damage are associated with a well-demarcated contusion overlying the lateral external capsule. Ultrastructural changes consistent with apoptosis have been described after TBI, so delayed patterns of neuronal cell death may involve necrotic and programmed cell death processes.

Fig. 50A.2 Photomicrographs of rat brain 3 days after fluid percussion injury. A, Focal contusion at gray/white interface (×120). B, Damaged cortical neurons overlying the contusion (×1200). C, Damaged and eosinophilic neurons in CA3 hippocampus (×1200).

Selective Neuronal Vulnerability

Damage to the hippocampus is commonly reported in autopsy studies of head-injured patients. Clinical and experimental studies describe cognitive abnormalities thought to be associated with hippocampal dysfunction. In an acceleration model of brain injury in nonhuman primates, CA1 hippocampal histopathological damage was reported in the majority of animals. CA1 damage was not produced by secondary global ischemia, elevated ICP, or seizure activity alone. In this regard, CA1 hippocampal damage is not routinely reported in other TBI models, including cortical contusion and moderate FP injury. However, midline FP injury followed by a delayed sublethal global ischemic insult leads to CA1 neuronal damage. The studies indicate the importance of injury severity on outcome and the vulnerability of the posttraumatic brain to secondary insults. Finally, the dentate gyrus, bilateral dentate hilus, and CA3 hippocampus have also been reported to be selectively damaged in FP models.

Thalamic damage after brain injury is described in clinical and experimental studies. In human brain injury, the loss of inhibitory thalamic reticular neurons is proposed to underlie some forms of attention deficits. In radiographic studies of patients with TBI using magnetic resonance imaging, relationships between injury severity, lesion volume, ventricle-to-brain ratio, and thalamic volume have been reported. Patients with moderate to severe injuries have smaller thalamic volumes and greater ventricle-to-brain ratios than patients with mild to moderate injuries. Decreased thalamic volumes suggest that subcortical brain structures may be susceptible to transneuronal degeneration after cortical damage. Focal damage to thalamic nuclei seen after long-term FP injury may result from progressive circuit degeneration after axonal damage, neuronal cell death, or lack of neurotrophic delivery.

Progressive Damage

Only recently has the progressive nature of the histopathological consequences of TBI been appreciated (Bramlett and Dietrich, 2007). At 2 months after moderate parasagittal FP injury, significant atrophy of the cerebral cortex, hippocampus, and thalamus is apparent in histological sections (Fig. 50A.3, A). Progressive tissue loss in the cortex and hippocampus at various times up to 1 year after lateral FP injury have been reported. Similar findings have been found in a model of controlled cortical impact injury. At 3 weeks and 1 year after injury, analysis demonstrated a significant hemispheric volume loss and expansion of the ipsilateral lateral ventricle. Thus atrophy of gray matter structures is associated with significant enlargement of the lateral ventricle.

Fig. 50A.3 Photomicrographs of rat brain 1 year after fluid percussion injury (×30). A, Brain shows gross atrophy with marked expansion of lateral cortex. B, Severe loss of white matter in external capsule (arrows) (×300).

Ventricular expansion not associated with hydrocephalus or increased ICP is felt to be a sensitive indicator of structural damage and an indirect measure of white matter atrophy. Indeed, ventricular size has been correlated with memory disturbances; patients with the highest ventricular volumes demonstrated significantly lower memory scores. Recent findings provide direct evidence of progressive white matter damage after FP brain injury (Bramlett and Dietrich, 2007). At 1 year after TBI, severe atrophy of specific tracts including the external capsule and cerebral peduncle was documented (see Fig. 50A.3, B). The importance of a progressive injury cascade after TBI in terms of other neurological conditions merits consideration. For example, if mild head trauma leads to a progressive reduction in neuronal reserve, would the person who sustained such trauma be more susceptible than others to neurodegenerative processes associated with aging? Some epidemiologic studies have indeed indicated that a history of brain trauma is a risk factor for Alzheimer disease.

Secondary and Repetitive Damage

A challenging problem faced by medical personnel responsible for the health care of amateur and professional athletes is recognizing and managing mild head injury. Returning an injured athlete to competition when the brain needs time to recover is an obvious concern. Also, a basic understanding of posttraumatic consequences that affect the vulnerability of the brain to secondary or repeated head injury remains unknown. This is a clinically important issue because TBI often is associated with respiratory suppression, resulting in secondary hypoxic insults (McHugh et al., 2007). Experimental studies have documented the detrimental consequences of secondary insults after mild to moderate TBI. The effects of secondary hypoxia on histopathological and behavioral outcome were investigated. Secondary hypoxia induced immediately after moderate FP injury resulted in significantly greater cortical and hippocampal CA1 damage and sensorimotor and cognitive deficits than were found in normoxic animals. Mild hypotension (shock) after TBI was reported to also worsen traumatic outcome. Taken together, these findings indicate the enhanced vulnerability of the posttraumatic brain to mild secondary insults.

Clinical studies indicate that patients with mild head injury may be at risk if they have a subsequent head injury (secondary impact syndrome). Recently an animal model was developed to investigate the behavioral and pathological changes associated with repetitive head injury (Weber, 2007). The consequences of single and repetitive injury induced 24 hours apart were assessed. Repetitive head injury led to greater functional impairment and structural damage than was found in the single-injury group. At the cellular level, repeated mild injury in an in vitro hippocampal cell culture model has shown elevations of neuron-specific enolase and S-100β compared to a single insult. Repetitive injury in transgenic mice that expressed mutant human Aβ precursor protein produced elevated Aβ levels and increased Aβ deposition, thus linking TBI to the mechanisms of Alzheimer disease. The known risk of developing neurodegenerative disease later in life is greater after repetitive brain trauma and makes this type of investigation extremely important.

Axonal and Dendritic Injury

Traumatic axonal injury exists as a spectrum involving widespread areas of the brain in experimental models of TBI (Buki and Povlishock, 2006). This pattern of white matter pathology can evolve from focal axonal alternations to complete transection of the axon. Reactive axonal changes using monoclonal antibodies targeted at neurofilament subunits or β-amyloid precursor protein have been characterized after FP brain injury and controlled cortical impact injury (Fig. 50A.4). Within 1 to 2 hours of injury, reactive axonal change is most conspicuous in brainstem regions, including the pontomedullary junction. This pattern of axonal damage seen in experimental models is in contrast to the human condition, in which callosal and subcortical white matter axonal damage predominates. In contrast, moderate parasagittal brain injury leads to widespread axonal damage in forebrain regions that represent reversible, irreversible, and delayed axonal perturbations. These findings may explain some of the transient and delayed functional consequences of TBI.

Fig. 50A.4 β-Amyloid precursor protein immunoreactivity after traumatic brain injury. A, Large numbers of immunoreactive profiles in external capsule (×120). B, High magnification of axonal swellings (×1200). C, Dense reactive profiles within the fimbria of hippocampus, consistent with primary axotomy (axonal shearing) (×120).

In addition to axonal damage, studies indicate that TBI also leads to significant changes in neuronal dendrites. A common feature of damage of injured neurons is loss of microtubule-associated protein 2 (MAP2) antigenicity. MAP2 is an important microtubule cross-linking protein that is found predominantly in somatodendritic environments, so changes in MAP2 may reflect damage to dendrites. Using this strategy, evidence of dendritic damage not necessarily associated with neuronal death has been obtained that may participate in some of the functional consequences of TBI.

Recent evidence for neurodegeneration following severe controlled cortical impact injury in the mouse has been shown using silver staining (Hall et al., 2005). This marker identifies degenerating neurons and processes. Both hippocampal and cortical neurodegeneration is observed as late as 72 hours and continues to 7 days in select brain regions. This degeneration is associated with calpain-mediated proteolysis of cytoskeletal proteins and decreases in growth-associated protein 43 (GAP-43) expression, which may inhibit the brain’s response to elicit an attempt at plasticity to restore function.

Importance of Gender

Recent clinical and experimental data have emphasized the importance of gender on the consequences of TBI (Stein, 2007). In a study of 334 patients with TBI, female patients had a better predicted outcome at the time of discharge from an inpatient rehabilitation program. However, a meta-analysis of eight previous studies in which outcome was reported separately for men and women reported a worse outcome in women than in men. The impact of gender on TBI is an understudied area of clinical neurotrauma and should be emphasized in future trials.

In experimental models of TBI, gender also appears to influence traumatic outcome (Stein, 2007; Bramlett and Dietrich, 2001). The hemodynamic consequences of brain injury and contusion volume were significantly less in female than in male rats (Bramlett and Dietrich, 2001). In this study, ovariectomy 10 days before TBI removed the volume differences between male and female rats. Thus intact females appear to have an endogenous neuroprotective mechanism that reduces the detrimental consequences of TBI (Stein, 2007). In terms of testing neuroprotective strategies in models of TBI, gender differences must be assessed.

In this regard, in a recent randomized double-blind, placebo-controlled phase II clinical trial known as Progesterone for Traumatic Brain Injury Experimental Clinical Treatment (ProTECT), intravenous progesterone reduced the overall death rate by 50% compared to placebo. Also, there was significant improvement in the functional outcome and level of disability among patients with moderate brain injury. Based on these encouraging results, a 17-center phase III trial clinical trial is proposed (ProTECT III) that will enroll approximately 1140 subjects over a 3- to 6-year period (Wright et al., 2007). The findings from this study may provide a viable treatment option for moderate to severe TBI patients.

Primary Injury Mechanisms

Two major types of forces are responsible for brain injury: one localized at the impact site and a second characterized by rotational forces. Depending on the force and location of the primary impact, head trauma can produce acute damage to blood vessels and axonal projections. Contact phenomena generate superficial or contusional hemorrhages through coup and countercoup (or contrecoup) mechanisms. Direct injury is commonly superficial, and the coup-countercoup hemorrhages may be adjacent or central.

Axonal shearing is a common lesion of the cerebral white matter that occurs particularly in acceleration-deceleration injury. Only recently has morphological evidence of axonal shearing (primary axotomy) become available in FP models. Ultrastructural evidence demonstrates the tearing or shearing of axons in nonhuman primates exposed to lateral acceleration. Perturbations of the axolemma leading to the accumulation of cytoskeletal components and organelles or activation of intracellular mediators of injury such as calpain activation may represent secondary injury processes that can be treated.

Shearing strains may also damage blood vessels and cause petechial hemorrhages, deep intracerebral hematomas, and brain swelling. Mechanical damage to small venules, resulting in focal BBB breakdown and platelet accumulation, is reported immediately after FP injury. Vascular damage leads to the formation of hemorrhagic contusions. Early vascular damage may increase neuronal vulnerability by causing posttraumatic perfusion deficits and extravasation of potentially neurotoxic bloodborne substances.

Secondary Injury Mechanisms

In head-injured patients, the extent of neurological recovery depends on the contribution of posttraumatic secondary insults (Statler et al., 2001). In the clinical setting, secondary insults include hypotension, hypoxia, hyperglycemia, anemia, sepsis, and hyperthermia. Experimental evidence indicates an increased susceptibility of the posttraumatic brain to secondary insults. For example, after midline FP brain injury, CA1 hippocampal vulnerability is enhanced with superimposed secondary ischemia. An important area of research regarding the treatment of brain injury involves the characterization of secondary injury processes, which may be targeted for intensive care management or pharmacotherapy.

A high frequency of hypoxic or ischemic brain damage occurs in patients who die as a result of nonmissile head injury. Hypoxic damage in the form of hemorrhagic infarction and diffuse neuronal necrosis is most common in arterial boundary zones between the major cerebral arteries. Hypoxic damage is also common in patients who have experienced an episode of intracranial hypertension. A significant correlation between hypoxic brain damage and arterial spasm in patients with nonmissile TBI has been reported. Posttraumatic hypoxia aggravates the BBB consequences of FP brain injury.

Posttraumatic hemodynamic impairments represent another injury mechanism. Clinical and experimental investigations report moderate reductions in local cerebral blood flow (LCBF) after TBI. After moderate parasagittal FP brain injury, widespread reductions in LCBF range from 40% to 80% of control. In contrast, severe FP injury leads to LCBF reductions that reach ischemic levels. Focal reductions in LCBF are associated with subarachnoid and intracerebral hemorrhage and local platelet accumulation. Reductions in LCBF result from the mechanical occlusion of cerebral vessels or the release of vasoactive substances, or possibly as a secondary consequence of reductions in neuronal activity or metabolism. Injury severity is therefore a critical factor in determining the hemodynamic and histopathological consequences of experimental TBI.

Cortical spreading depression (CSD) is caused by ionic changes within tissue resulting in waves of depressed electrical activity. CSDs are not necessarily harmful to tissue unless there is damage present. It was previously thought that CSDs could only be induced experimentally; however, spontaneous self-propagating waves within damaged tissue, known as periinfarct depolarizations and similar to CSDs, have been reported (Fabricius et al., 2006). The incidence of CSD has been well described in the experimental focal ischemia literature and associated with increasing amounts of neuronal damage within the cerebral cortex. However, the presence of CSDs in the human brain-injured population had remained controversial until recently because of the inability to adequately measure the phenomenon; Strong and colleagues (2002) clearly showed the presence of CSDs following clinical TBI. Although it does not occur in all acutely injured individuals, the incidence is higher in younger patients. Such findings have resulted in a new initiative to document the occurrence of CSDs after TBI and study their relationship to outcome.

Hypotension and shock are present in a significant number of patients with TBI. In severely injured patients, outcome is correlated with reduced mean arterial blood pressure. Hemorrhagic hypotension after FP injury results in more severe histopathological outcome than TBI alone. The increased sensitivity of the posttraumatic brain to moderate levels of hypotension may result from deficits in autoregulation, which have been reported in patient and experimental studies. Hypotensive periods that may occur during surgical procedures and anesthesia may produce secondary insults and be hazardous to the head-injured patient.

Many patients experience fever after head injury, and clinical data indicate that brain temperature may be higher than core or bladder temperature (Hayashi et al., 2004). Experimentally, posttraumatic brain hyperthermia induced artificially 24 hours after trauma increases mortality rate and aggravates histopathological outcome, including contusion size and axonal pathology. In the clinical setting, posttraumatic hyperthermia may represent a secondary injury mechanism that might negate the beneficial effects of a therapeutic agent.

A significant debilitating consequence of TBI is the development of seizures (Vespa, 2005). Some 40% to 50% of moderate to severe TBI patients develop epilepsy, and brain injuries account for 20% of symptomatic epilepsy cases (Garga and Lowenstein, 2006). Trauma-induced mechanisms of reduced seizure threshold include damage to specific neuronal populations and circuits, increased BBB damage, and aberrant mossy-fiber sprouting of dentate granule cells. Importantly, posttraumatic epilepsy has been shown to worsen traumatic outcome and aggravate cognitive problems in some clinical studies. Unfortunately, chronic seizures after brain injury are poorly controlled by available antiepileptic drugs (Temkin, 2009).

Glutamate Antagonists

Excitatory amino acid neurotransmitters have been implicated in the pathophysiology of TBI. Microdialysis techniques have documented elevated levels of extracellular excitatory amino acids, whereas N-methyl-d-aspartate (NMDA) receptor antagonists, including MK-801 (dizocilpine), provide behavioral and histopathological protection against brain trauma. The role of glutamate antagonists in treating central nervous system injury has been reviewed, and clinical stroke trials of the competitive NMDA antagonist, MK-80I, were withdrawn because of harmful side effects. However, the effects of topiramate on inhibiting glutamate release have been evaluated using microdialysis in TBI patients (Alves et al., 2003). This treatment strategy shows promise in attenuating glutamate levels, but further evaluation is necessary to demonstrate clinical improvement.

Free Radical Scavengers

The genesis of oxygen free radicals has also been implicated in the pathophysiology of TBI. Experimental FP injury increases the production of oxygen radicals, and free radical scavenger inhibitors have been reported to be protective in brain injury models. Sensitive indicators of hydroxy radical production and microdialysis have detected elevations in extracellular hydroxy radical production after FP injury. The free radical scavenger, superoxide dismutase (SOD), reduces BBB opening and the genesis of brain edema after TBI. Transgenic mice that overexpress human copper-zinc SOD1 are protected against brain trauma. Clinical trials using polyethylene glycol SOD and free radical scavengers have been initiated in brain trauma, with little success.

The reason for these negative results is not known, although questions regarding brain penetration and limited therapeutic window have been discussed. In one experimental study, treatment with a potent inhibitor of lipid peroxidation and an antioxidant provided significant neuroprotection in an FP model when treatment was delayed 30 minutes but not 3 hours after TBI. Based on neurochemical and treatment studies, the therapeutic window for treatments that target free radical reactions and lipid peroxidation after moderate TBI seems to be limited. Whether these treatments may protect against secondary insults remains to be investigated.

Neurotrophic Factors

A unique problem of brain trauma is diffuse axonal injury. Axonal injury leading to circuit disruption may not only produce immediate functional consequences but also affect trophic signaling between neuronal populations. Addition of trophic factors after TBI may help maintain neuronal survival and promote circuit reorganization and functional recovery. Neurotrophins have been shown to be neuroprotective by in vitro and in vivo models of neuronal injury. After experimental TBI, delayed treatment with basic fibroblast growth factor significantly reduced histopathological damage and improved cognitive function. Several neurotrophic factors have also been shown to be protective when administered exogenously to animals after experimental TBI. For example, posttraumatic infusion of nerve growth factor (NGF) into the injured cortex or lateral ventricle was reported to improve learning and memory and decrease apoptotic neuronal loss in the septum of rats. Systemic administration of insulin-like growth factor 1 (IGF-1) after TBI improved learning and neuromotor function. In contrast, brain-derived neurotrophic factor (BDNF) administration was not protective against behavioral or histopathological defects caused by TBI. Immortalized neural stem cells retrovirally transduced to produce NGF when transplanted into the injured brain improved cognitive and neuromotor function and rescued hippocampal CA3 neurons. Therefore, an important direction of future research will be to use engineered cell lines to produce neurotrophins that could synthesize and locally release factors that enhance plasticity and circuit reorganization. If experimental studies continue to show a benefit of neuroprotection on neuronal injury and behavioral outcome in TBI models, this may be an important direction for future clinical trials in brain trauma.

Protection by Nitric Oxide–Related Species

In the nervous system, nitric oxide (NO) may serve as a neurotransmitter, a signal between cells, and an autocrine signal within a given cell. After brain injury, an increase of neuronal calcium triggers constitutive NO synthase (cNOS) activity, leading to release of NO that may enhance excitotoxicity. Studies of mutant mice deficient in neuronal (nNOS) or endothelial NOS (eNOS) activity have demonstrated that whereas nNOS exacerbates ischemic injury, eNOS protects against it.

Therapeutic strategies directed at the NO pathway have been reported in models of brain injury. Data indicate that FP injury leads to the acute activation of cNOS, and that the selective inhibition of nNOS by 3-bromo-nitroindazole protects histopathologically and behaviorally. In addition, inhibition of inducible NOS with aminoguanidine (selective iNOS inhibitor) also improved histopathological outcome. However, recent work using iNOS knockout mice has shown a beneficial role of iNOS in brain injury. Thus whereas early iNOS activity after brain trauma may contribute to secondary injury mechanisms, later or chronic activation may participate in reparative processes. These points regarding what processes should be targeted and when they should be inhibited are critical as novel strategies are developed to target NO-mediated cell injury. Additional studies using mutant mice deficient in nNOS, eNOS, or iNOS activity will be important to advance this area of investigation.

Antiapoptotic Agents

Apoptosis is a mode of cell death in both physiological and pathological processes. Evidence of apoptotic cell death has also been observed after TBI. After FP injury, apoptotic cells have been identified in the ipsilateral cortex, hippocampus, and thalamus as soon as 4 hours after injury, and immunocytochemical markers for caspase-3, -8, and -9 activation have been reported (Fig. 50A.6). The effects of FP injury on the expression of the bc1-2 protein, which regulates developmental programmed cell death, have also been investigated. Evidence of apoptosis of oligodendroglia in long tracts undergoing wallerian degeneration has been reported after spinal cord injury. Therefore, demyelination of tracts after brain or spinal cord trauma may result from apoptotic death of oligodendrocytes. Importantly, indicators of apoptotic cell death have also been observed in human tissues.

Fig. 50A.6 Histopathological and immunohistochemical evidence of apoptotic cell change 1 day after traumatic brain injury. A, Hematoxylin and eosin–stained section showing apoptotic bodies (×1200). B, Caspase-8 immunoreactivity (×1200). C, Caspase-9 immunoreactivity in the thalamus (×1200). D, Immunohistochemistry of active caspase-3 in dentate gyrus (×1200).

Although the molecular events leading to apoptosis are not fully understood, the family of cysteine proteases (caspases) play an active role in its pathogenesis. In reference to neuroprotection, in vitro studies have demonstrated that protease inhibitors specific to caspase-3 inhibit apoptosis. Whereas some experimental data indicate that inhibition of caspase-3 improves outcome after TBI, other studies suggest that this approach has limitations. Current emphasis is on more upstream apoptotic processes, in contrast to targeting caspase activation to attenuate damage. Calpain inhibition may target not only apoptotic cell death by regulating caspase activation but also cytoskeletal protein degradation (Raynaud and Marcilhac, 2006). More research is needed to clarify the various pathways involved in apoptotic cell death and determine which pathways may be most sensitive to therapeutic interventions. Using agents specific to apoptosis or in combination with agents that target necrosis is a potential research direction.

Therapeutic Hypothermia

Numerous studies have demonstrated that although mild to moderate hypothermia is neuroprotective in models of TBI, mild hyperthermia worsens outcome (Dietrich and Bramlett, 2010). After brain trauma, hypothermia has been shown to improve histopathological and behavioral outcomes and to influence a wide range of injury processes. Microdialysis studies report that posttraumatic hypothermia reduces the acute surge in levels of extracellular glutamate and hydroxy free radicals after injury. Posttraumatic hypothermia protects against BBB dysfunction. Hypothermia attenuates progressive cortical atrophy and subsequent ventricular enlargement. The ability of any therapeutic intervention to provide long-term protection is an important requirement for the advancement of any therapeutic strategy to the clinical setting.

As previously discussed, a significant number of patients with TBI sustain a secondary insult that may include hypotension, hypoxia, or hyperthermia. This fact has led to the use of complicated models to test novel neuroprotective agents before clinical trials. This point is important because experimental therapeutic strategies are commonly tested in simple models of brain injury as proof of concept (McIntosh et al., 1998). In this regard, posttraumatic hypothermia followed by a controlled rewarming period has been evaluated in TBI models complicated by secondary hypoxia. Taken together, these studies showed that hypothermia was protective in complicated models, but the degree of protection depended on injury severity, duration of hypothermic period, and the rewarming procedure.

The use of moderate levels of hypothermia (>32°C) also improves outcome in patient studies. Systemic hypothermia (32°-33°C) begun within 6 hours of injury (Glasgow Coma Score 4-7) resulted in no cardiac or coagulopathy-related complications, a lower seizure frequency, and more patients in the good recovery to moderate disability category. In other head trauma studies, therapeutic hypothermia attenuated intracranial hypertension but did not affect the frequency of delayed intracerebral hemorrhage. Results from a recent U.S. multicenter TBI trial failed to demonstrate a protective effect of hypothermia on traumatic outcome, but a subgroup analysis showed that patients younger than age 45 who came into the emergency room hypothermic demonstrated improvement with hypothermic treatment (Clifton et al., 2001). Obviously, more experimental and clinical studies are needed to determine what factors are most important in providing protection when using hypothermic strategies. Temperature is known to affect many pathophysiological processes after TBI, and this characteristic may be advantageous because of the multifactorial nature of trauma pathomechanisms. The cooling and rewarming periods are also important variables in determining the extent of neuroprotection. In TBI studies, prolonged periods of hypothermia (i.e., >24 hours) may therefore be necessary to protect the brain from primary and secondary injury processes. In a multicenter TBI trial from China, long-term cooling (5 days) was significantly better than short-term cooling (2 days) in terms of improved outcome in patients (Jiang et al., 2006). Because brain temperature can be elevated compared with bladder temperature in head-injured patients, normothermia or mild hypothermia should be maintained during critical postinjury periods.

Environmental Enrichment

The effect of environmental enrichment (EE) that exposes animals to a complex, highly stimulatory, and social environment has been studied in a number of TBI models. Using a midline FP injury model that produced no noticeable histopathology, EE was reported to improve cognitive function. In addition, EE has been shown to decrease overall contusion volume and improve performance in the Morris Water Maze task. The effects of EE have been suggested to be reflected in changes in dendritic arborization. Recent experimental studies have reported that animals raised with EE have larger contusion injuries and greater sensorimotor deficits after mild TBI, compared to standard laboratory cage-reared animals (Kozlowski et al., 2004). However, the sensorimotor deficit improved at a much more rapid pace in the EE animals compared to age-matched controls. Future studies combining EE with neuroprotective and reparative strategies probably will be needed to maximize improvement in outcome.

Reparative and Transplantation Strategies

After a variety of acute central nervous system injuries, there is a massive proliferation of stem or progenitor cells (Richardson et al., 2010). The identification and origin of the fate of these cells is an area of intensive investigation. In a model of FP injury, the total number of proliferating cells as identified with 5-bromo-deoxyuridine, a marker of mitotic activity, was shown to significantly increase in areas of the subventricular zone and hippocampus. In that study, proliferating cells did not express cell markers and therefore appeared not to have begun to differentiate. Targeting this endogenous proliferative response to injury may be one way to enhance recovery following TBI.

In contrast to attempts to enhance endogenous reparative events, providing new cells from exogenous sources is an alternative approach and may be necessary when neuronal loss and axonal injury are severe. Neural transplantation has been explored in TBI models (Lu et al., 2003). Fetal cortical tissue transplanted into the injury cavity improved motor function and transiently attenuated cognitive dysfunction alone and in combination with NGF infusion. Although the reestablishment of normal adult neural circuitry has not been demonstrated with fetal tissue grafts, one mechanism for improved function may be neuroprotection by release of trophic factors from the grafts (Gao et al., 2006). Recent studies have also attempted to provide cellular replacement and host-graft integration using self-renewing cell lines (Wallenquist et al., 2009). Using transplanted immortalized neural progenitor cells transduced with the mouse NGF gene to secrete NGF, improved neuromotor and cognitive function and reduced hippocampal CA3 cell death have been reported. Continued study in this exciting field may establish transplantation procedures relevant to clinical strategies to promote recovery after TBI.