CHAPTER 327 Neurochemical Pathomechanisms in Traumatic Brain Injury
Traditionally, TBI has been divided into primary and secondary forms of injury. Primary injury is due to the unavoidable direct mechanical forces occurring at the time of the traumatic insult.1 Secondary injury is derived from complications initiated by the primary injury and includes potentially avoidable entities such as hypoxic-ischemic injury, cerebral edema, metabolic dysfunction, alterations in vascular permeability, diminished blood flow, inflammation, diffuse axonal injury, and the consequences of intracranial hypertension. Almost all clinical treatments are aimed at modulating these secondary injury mechanisms.
Both primary and secondary brain injury can be further classified as focal or diffuse. The distinction between focal and diffuse injuries is historically derived from the presence or absence of radiographic mass lesions on computed tomography.2,3 This distinction has now evolved to also consider the distinct pathobiologic mechanisms imparted by the trauma in regions local to and remote from the point of impact. However, any attempt to conclusively classify brain injury remains a difficult task because most TBIs consist of a heterogeneous admixture of focal and diffuse damage. Research efforts and clinical trials have sought to tip the balance of these secondary events toward facilitating neuroprotection rather than autodestruction. Central to research efforts is the neuronal response to brain injury. It is thought that the mortality and morbidity associated with TBI can be greatly reduced by a better understanding of the mechanisms that cause neuronal injury, dysfunction, and death. This information is critical within the clinical realm—not only to minimize neuronal death after TBI but also to potentially facilitate augmentation of neurological reorganization and repair. Accordingly, there is new ongoing research aimed at the use of stem cells for neuroregeneration (see Chapter 328) and for the development of deep brain stimulation techniques that may modulate the sequelae of TBI.4,5
This chapter covers the complex pathophysiology of TBI. It begins by briefly detailing the biomechanical consequences of trauma on the central nervous system (CNS) (also see Chapter 325) and continues with a comprehensive discussion of primary and secondary injury mechanisms.
Primary Injury: Molecular And Microscopic Aspects
Focal versus Diffuse Primary Brain Injury
Diffuse Primary Brain Injury
The best example of a pure diffuse primary brain injury is “mild TBI” or concussion (see Chapter 332). Concussion is a broadly applied term for the clinical manifestations of blunt head trauma that result in rapid-onset, functional disturbance of the CNS (rather than a structural injury) secondary to the inertial forces of TBI. Concussion usually resolves spontaneously and, contrary to popular belief, may or may not be associated with loss of consciousness (LOC). When LOC is present, it is thought that either the magnitude or the biomechanical directionality of the traumatic forces is sufficient to transiently perturb the brainstem reticular activating system and result in LOC.
Focal Primary Brain Injury
Primary focal brain damage is a direct result of the physical forces delivered at the time of injury. These “impact” injuries are manifested clinically as cortical contusions, brain parenchymal lacerations, and vascular lesions with resultant hemorrhage and hematoma formation.6 Focal contusions most commonly result from contact of the brain with more rigid structures such as the skull, dural edges, or physical objects used in assaults. These impact injuries may occur in association with a more significant diffuse injury, such as when the rapidly decelerating cranium strikes a windshield during a motor vehicle collision. The inertial forces of such a diffuse injury are thought to generate widespread damage and are implicated in prolonged unconsciousness. However, focal injuries do not cause LOC but instead may cause permanent discrete neurological deficits because of the immediate effects of the penetrating/focal injury. The forces of impact imparted by primary injuries are largely responsible for rapid necrosis via physical destruction of cellular elements.7 Focal injuries include skull fractures, penetrating TBI, and vascular injuries and are covered additionally in Chapter 324 and Chapter 325.
Relationship between Brain Movement, Mechanical Forces of Injury, and Histologic Effects
The destructive energy imparted by any given trauma is transmitted to the skull—and thereby its contents—in relatively unpredictable patterns. At one end of the spectrum are pure inertial injuries, where rapid deceleration and rotational forces cause devastating diffuse injuries. The cranium may never contact a solid object, yet the brain is irreversibly damaged. Such injuries maximally damage axons and most frequently occur as a result of motor vehicle accidents (although almost always compounded by an additional impact component). At the opposite extreme are unusual impact injuries in which the stationary head (e.g., of a machine operator) is slowly crushed by slow-moving machinery. Such injuries classically produce massive fractures, extra-axial hematomas, and contusions, but these patients do not usually lose consciousness because axonal injury is absent and the reticular activating system/projection fibers are not disturbed. Thibault and Gennarelli used a primate impact acceleration injury model to characterize the relationship between the magnitudes of acceleration/deceleration force, the duration over which it is applied, and the consequences for the intracranial contents.9 A brief, high-intensity deceleration force will tear parasagittal bridging veins and cause an acute subdural hematoma. When the deceleration force is of higher magnitude and longer in duration, as in motor vehicle accidents, diffuse axonal injury may occur. When both the magnitude and the duration of the deceleration force are less, transient unconsciousness (concussion) results but few structural effects are seen when the brain is examined either ultrastructurally or by light microscopy (see Chapter 332).
Brain Movement during Impact
Laboratory models of brain injury have demonstrated that the brain moves substantially within the cranial cavity in response to deceleration forces.10,11 The brain is anchored within the cranial cavity by only the parasagittal bridging veins, parasinusoidal granulations, cranial nerves, and tentorium. Movement of the brain forward toward the anterior cranial-basal structures, particularly the sphenoidal ridges, concentrates force at the bases of the frontal lobes and the tips of the temporal lobes.11 Surface contusions therefore occur very much more frequently at these sites than elsewhere. There is evidence in the human brain that shearing force also concentrates in deep white matter structures such as the corona radiata, thus explaining the frequent finding of parasagittal gliding contusions.12 Finally, it is generally believed that shearing forces transmitted through the brainstem and the reticular activating system are responsible for the immediate LOC.
Damage to Cells/Tissue
Astrocytes
Astrocytes are increasingly becoming understood to have a significant role in the deleterious effects of TBI. Although the surface area and complexity of membranes may be less for astrocytes, there is now clear evidence that astrocytes are excitable, possess ion channels, and may be depolarized (although to a much lesser extent than neurons).13–15 Astrocytic membranes also constitute an important component of the blood-brain barrier (BBB), and there are now extensive data showing that this barrier function is transiently disturbed by mechanical trauma. Furthermore, disturbed ionic and neurotransmitter homeostasis is recognized as one of the most important mechanisms contributing to the secondary brain swelling after TBI, and astrocytes play a role in maintaining this homeostasis. One example is maintenance of potassium homeostasis after TBI. Astrocytes are known to function as potassium uptake buffers in that they have the capacity to rapidly take up potassium from the extracellular space.14,16–18 Kimelberg and Norenberg hypothesized that astrocytes function to conduct potassium away from neurons, particularly in injured brain tissue, and thereby aid in the establishment of ionic homeostasis.14 Thus, there is a net loss of potassium from injured tissue into the microvasculature that begins hours after onset. In addition, stretch-injured astrocytes express a dysfunctional cation current as opposed to an osmoregulatory anion current. This mechanism may contribute to the cytotoxic swelling seen after TBI.19 Such astrocyte swelling is the ultrastructural hallmark of both acute cerebral ischemia and focal cerebral contusion and is almost always seen in animal models of trauma and in humans after trauma.16,20
Axons
About 50 years ago, neuropathologic studies first demonstrated an accumulation of axoplasmic “retraction balls” at sites of axonal discontinuity.12,21 They were chiefly found on large myelinated fibers in patients who were unconscious from the time of injury and subsequently died. These retraction balls were found in high density in white matter tracts in approximately 25% of severely head-injured patients and were thought to occur immediately as a result of tearing.22 Traumatic axonal injury (the experimental correlate to diffuse axonal injury) is now known to be a progressive process involving transient mechanoporation of the axolemma that allows unregulated calcium entry.23 The mechanical insult induces a sequence of events culminating in failure of axoplasmic transport, pooling of intra-axonal contents, and pinching off of the axon from its distal segment (Fig. 327-1). This disconnection occurs within 24 to 72 hours after the traumatic event and is termed delayed or secondary axotomy (because the primary mechanical insult provokes secondary biochemical processes that result in axotomy). This suggests that axons that subsequently show the changes associated with diffuse axonal injury may be functioning, in some capacity, immediately after the injury before eventually degrading. It also suggests that other, less affected axon tracts may not progress to diffuse axonal injury. Thus, diffuse axonal injury is amenable to therapeutic intervention.
At the molecular and microscopic level, calcium influx initiates activation of calpain24,25 and mitochondrial swelling,26 with release of cytochrome c and activation of caspase27 leading to further axonal injury, apoptosis, and detachment over time. These changes have far-reaching consequences for neuronal function. Interruption of the axon causes proximal wallerian degeneration of the affected neuron. Distally, the axon degenerates, fragments, and disappears, thereby resulting in deafferentation of the affected neuronal fields. The functional consequences of this process may include seizures because of lack of inhibitory effects, spasticity, intellectual decline, and unmodulated behavior patterns. When this process is widespread and wallerian degeneration destroys many neurons, the whole brain becomes atrophic, with ventriculomegaly and, in the worst cases, a persistent vegetative state.22,28
Clinical Implications
The benefits of attenuating secondary axotomy may be enormous. These hopes are bolstered by recent ultrastructural studies revealing that neuron somata show evidence of the potential for reorganization and repair for up to 7 days after traumatically induced axonal injury.29 Cyclosporine is a widely investigated immunosuppressive drug that has been shown to blunt traumatically induced axotomy in experimental models of TBI.30 Its neuroprotective properties are thought to be derived from inhibitory actions on the protein phosphatase calcineurin, as well as from its modulatory effect on mitochondria and the mitochondrial permeability transition pore (see discussion later in this chapter). There has been one clinical trial aimed at taking advantage of these properties via the postinjury administration of cyclosporine in the hope of limiting mitochondrial damage and axonal injury.31 Although the results of this small trial demonstrated that administration of the drug was safe, it did not reveal any benefit in terms of neurological outcome. Future phase III clinical trials are still being considered.
Shear Effect on the Microvasculature
It has been estimated that the magnitude of shear required to damage the pial vasculature may be 5 times greater than that needed to damage axons.9,32 Although this gradation of force suggests that vascular structures should be damaged less frequently than axons and membranes, in reality this may not be the case because traumatic forces are more readily translated to surface vessels than to deeper axons. In the majority of significant head injuries, focal concentrations of force develop at the tips of the frontal and temporal poles and are sufficient to disrupt these pial vessels and cause a focal contusion. Ultrastructural studies in both head-injured humans and appropriate animal models have demonstrated major anatomic changes in the injured microvasculature. Such changes include the following16,17: (1) swelling of perivascular astrocytic end-feet; (2) increased endothelial microvacuolation and micropseudopodial activity; (3) perivascular hemorrhage and transvascular diapedesis of red cells, which may coalesce to form a frank intracerebral hematoma or hemorrhagic contusion; and (4) increased intravascular leukocyte adherence. Frank vascular disruption has been found to be unexpectedly atypical in human pericontusional biopsy material, thus suggesting that small vessels “stretch and leak” much more frequently than they “tear or burst.” These microvascular changes have profound functional consequences, chiefly a reduction in local cerebral blood flow (CBF) and the development of vasogenic and cytotoxic edema with increased intracranial pressure (ICP).20
Ion Channels
Although dendritic spines, synapses, gap junctions, and myelinated axons constitute specialized regions of neurons, ion channels are by far the most frequent structures embedded in neuronal membranes. Using patch clamp techniques and in vitro tissue culture of neurons growing on flexible plastic membranes, studies have shown that ion channel function may be radically altered by mechanical deformation—in this case by delivery of a brief jet air impulse to the flexible membrane.13,33,34 Specific classes of “mechanotransducing” ion channels have been identified by using similar techniques in both neurons and glia.13,33 Some of these ion channels remain perturbed for several hours after mechanical deformation.34 Still other experiments have shown rapid entry of calcium and subsequent neuronal death, along with efflux of lactate and potassium into the culture medium, after mechanical deformation.34 Further implicating channelopathy as a significant factor in TBI are data from in vivo trauma models such as fluid percussion injury and contusion impact models. These models show a massive, approximately threefold to fourfold rapid transient efflux of potassium into the extracellular fluid (ECF) associated with a fall in the sodium content of the ECF.35–38 About a third of the potassium release could be blocked with tetrodotoxin (a selective blocker of voltage-gated sodium channels), thus suggesting that two thirds of the potassium release was occurring through agonist-operated channels. Another investigation similarly revealed that preinjury blockade of voltage-operated ion channels failed to ameliorate the negative neurological and behavioral effects of the trauma and produced only a modest effect on K+ flux in the ECF. This suggests that agonist-operated ion channels are more important in mediating ionic events after TBI.36 Finally, channelopathy has been implicated in the etiology of the calcium influx seen after TBI. Work by Wolf and colleagues suggests that calcium influx occurs, in part, via mechanical alteration of tetrodotoxin-sensitive sodium channels after traumatic axonal injury.39 This implies that excessive calcium influx may not be fully attributable to direct axolemmnal poration but instead is also related to sodium channelopathy. The resulting sodium influx then triggers depolarization-induced calcium influx through voltage-gated calcium channels and reversal of the sodium-calcium exchanger, both acting to increase influx of calcium. Work is ongoing to further explore this phenomenon.
Synapses
Direct investigation of synaptic function is difficult in the acute stages of trauma. Microdialysis studies have investigated the time course of changes in neurotransmitters within the extracellular space after fluid percussion injury, and the results demonstrated brief transient surges in the release of excitatory amino acids (EAAs) and acetylcholine. From experimental models of injury, this posttraumatic excitotoxicity is marked by increases in glutamate, which leads to an increase in extracellular potassium as a result of channel activation. Potassium levels determined by microdialysis techniques were increased in 20% of patients after severe TBI and were also noted to correlate directly with reduced CBF.40 Data from microdialysis studies of patients who have sustained severe head injury and patients with ischemic events superimposed on their primary trauma show that ECF EAAs rise to levels 50 to 60 times higher than normal values when a secondary ischemic event is superimposed on the trauma.35,41 The excitatory neurotransmitters released from damaged cells and neuron processes may be responsible for these increases through a positive feedback loop. EAAs may also come from the intravascular compartment. This conclusion is supported by the finding that serum levels of structural amino acids in these patients were also raised and appeared to fluctuate in parallel with EAAs.42
The behavioral changes that persist up to weeks or months after TBI, even in animals without any evidence of structural damage, have been taken as evidence to support functional changes at the synaptic level or in relation to second messenger systems. Neurochemical studies have shown evidence of synaptic alterations, as well as G protein–coupling variations, within the cell membrane that are manifested as prolonged amplification of protein synthesis in response to activation of muscarinic cholinergic and certain catecholaminergic receptors.43–45 These changes may translate into effects on long-term potentiation in the hippocampus (which have been demonstrated in the absence of structural changes after trauma and may be an important mechanism underlying the traumatic effects on learning and memory). In addition, matrix metalloproteinases (MMPs) are known to modulate molecules forming the extracellular matrix (ECM). MMP proteolysis of ECM molecules may perform a permissive or inductive role in the fiber remodeling or synaptogenesis initiated by deafferentation.46 The significance of this interaction in TBI was explored in animals by intraventricular infusion of the MMP inhibitor FN-439 after unilateral lesions of the entorhinal cortex.46 The lesioned rats receiving the MMP inhibitor failed to develop the capacity for long-term potentiation and showed persistent cellular debris. These results underscore the importance of continuing to improve our understanding of MMPs, the ECM, and other mechanisms involved in remodeling after trauma.
Secondary Injury Processes
The concept of delayed secondary neurological damage after head injury is supported by “lucid interval” statistics. Between 30% and 40% of severely head-injured patients who die will, at some time, have demonstrated a period of lucidity sufficient to obey commands or speak.22,47 This implies that the primary impact events were not sufficiently severe to damage the brain beyond the capacity for function, thus emphasizing the importance of the secondary damage.48 The principal mechanisms to consider are hypoxia-ischemia, edema, excitotoxicity, calcium dysregulation, apoptosis, cytoskeletal proteolysis, metabolic and mitochondrial derangements, oxidative stress, and inflammation.49,50 The deleterious mechanisms at work are diverse and interrelated—often with both sequential and parallel cascades of neuronal reaction and cell death.51
Finally, it is useful to emphasize that many clinical trials, using either physiologic or pharmacologic interventions, view these secondary injury mechanisms as the main therapeutic targets (Table 327-1).52–72 Many of these trials will be discussed briefly throughout this chapter under the most relevant heading. That said, there is no questioning the heterogeneity of TBI and, unfortunately, the lack of favorable results from previous clinical trials. Some of these failures may be due partially to flawed classification systems in terms of optimally reflecting the target population for the drug in question. These classifications have relied on several variables, including clinical severity, pathophysiology, pathoanatomic and prognostic indicators, etiology, and symptomatology. Today, the most commonly used system is the Glasgow Coma Scale, which is based solely on the severity of the neurological injury. As a result, there is growing support to establish a more clinically derived classification system for TBI based chiefly, but not exclusively, on pathoanatomic features. It is hoped that this would improve the application of appropriate treatment strategies targeting the various causes of TBI.73
Hypoxia/Ischemia
A central factor involved in secondary damage after TBI is the onset of hypoxic-ischemic damage. The incidence of ischemic brain damage seen at autopsy in patients who sustained severe TBI is extremely high, with estimates ranging between 60% and 90%.47 During life, most of these patients do not manifest the long periods of low cerebral perfusion pressure (CPP) that are known to be necessary for the generation of ischemic damage. Likewise, in animal models of impact-type head injuries, widespread ischemic damage is not seen other than around the periphery of focal contusions. Thus, there is a fundamental paradox, and the high incidence of ischemic brain damage is not easily explained, although necrosis of neurons, secondary to release of EAAs, may be a factor exacerbating cell death.
The Genesis of Ischemic Brain Damage after Severe Human Traumatic Brain Injury
On a global scale, CBF can be decreased by as much as 50% during the first 48 hours after injury and lead to ischemic changes.74 Impaired CBF has well-described cellular consequences, and a time-dependent hierarchy of neuronal events is summarized in Table 327-2.75–77 Both hemorrhage and contusion can lead to local ischemia through compromise of the microcirculation by thrombotic occlusion of blood vessels.1,78–80 Additionally, at the local level, blood flow within focal areas of contusion is dramatically reduced.16 Focal ischemia can also result from the formation of a mass lesion, which will raise ICP and impede blood flow to the damaged region in accordance with the modified Monro-Kellie Doctrine.81 Finally, ischemia secondary to occlusive or hemorrhagic stroke develops as a result of damage to or interruption of the blood supply from a parent vessel to an area of vulnerable parenchyma—such as traumatic carotid dissections.49
TABLE 327-2 Cellular Consequences of Impaired Cerebral Blood Flow
From Jones TH, Morawetz RB, CrowelI RM, et al. Threshold of focal cerebral ischemia in awake monkeys. J Neurosurg. 1981;54:773-782.
| CEREBRAL BLOOD FLOW (mL/100 g/min) | CONSEQUENCES |
|---|---|
| 40-60 | Normal |
| 20-30 | Start of neurological symptoms |
| 16-20 | Isoelectric electroencephalogram, loss of evoked potentials |
| 10-12 | Na+ and K+ pump failure Cytotoxic edema |
| <10 | Complete metabolic failure with gross disturbance of cellular energy homeostasis (infarction) |
 |
|
Infarction versus Selective Neuronal Loss
When flow is profoundly reduced (i.e., <5 to 10 mL/100 g per minute) within the distribution of one cerebral end artery for more than 60 to 90 minutes, infarction ensues (immediate necrosis of all cell types within a zone of the brain). However, when the reduction in flow is to levels of approximately 15 to 18 mL/100 g per minute for a period longer than 30 minutes, selective neuronal loss may occur—especially in the hippocampal neurons (in the molecular layer, CA1 and CA3 sectors), cerebellar granular cells, and cortical neurons (particularly the larger cells in areas such as the cuneate visual cortex).82,83 Within the context of head injury, this type of neuronal loss is especially important in patients with raised ICP, in whom CPP may be marginal (≈30 to 40 mm Hg) for many hours or even days. In such patients, a high frequency of ischemic neuronal loss is seen in the hippocampus.47 This may explain the high frequency of memory disorders and coordination difficulty noted in the majority of severely head-injured survivors. This concept is also in accord with the almost universal finding of marked cerebral atrophy in patients who survive severe head injuries.
Clinical Implications
Clearly, reductions in CBF can have devastating consequences and will directly affect metabolic profiles. Noting that historical strategies for managing severe TBI followed ICP-directed protocols (and therefore would indirectly augment CBF), a phase III clinical trial with CPP-directed therapy was performed.65 To ensure adequate CBF and therefore oxygen delivery, the primary goal in this trial was to maintain higher CPP (versus lowering ICP). The results suggested that CPP-directed therapy improves several physiologic parameters, such as brain perfusion; however, it failed to show any incremental benefit in outcomes when CPP was targeted to levels greater than 70 mm Hg (versus 60 mm Hg). This was predominately due to an increased incidence of acute respiratory distress syndrome, which had a negative impact on mortality measures.65 Still other trials have targeted increased oxygenation by using hemoglobin substitutes, or hyperbaric oxygen, to augment the oxygen-carrying capacity of the microcirculation to damaged tissue.42 One such agent is Oxycyte, a third-generation perfluorocarbon (PFC) that improves the oxygen-carrying capacity of blood. In animals, PFCs have been shown to improve cerebral oxygenation and mitochondrial function after TBI.84 However, increased free radical formation with higher doses was also seen in these same studies. The authors suggested the need for further studies combining PFCs with free radical scavengers. Trials of PFCs and hyperbaric oxygen are ongoing—including the upcoming Brain Tissue Oxygen Monitoring in Traumatic Brain Injury trial, which will aim to implement therapy directed at increasing the partial pressure of oxygen in brain tissue (PbtO2) to further evaluate direct measurements of cerebral oxygenation on outcome.
Ischemia and Associated Acidosis/Hydrogen
Although hydrogen ions in the extracellular space are powerful cerebral vasodilators, high concentrations of hydrogen ions within cells are harmful because they alter the function of intracellular enzymes.85 For example, low pH causes conformational changes in the N-methyl-D-aspartate (NMDA) ion channel that prevent further ingress of sodium and calcium and egress of potassium during cellular acidosis. The potential benefits of mild acidosis therefore include inactivation of glutamate receptors, decreased free radical generation,86,87 inhibition of phospholipase A2 (which generates free radicals), decreased energy demand because of hyperpolarization, and inhibition of the Na+/H+ exchange transporter, which prevents intracellular entry of Na+ and Ca2+.88
Clinical Implications
The acidosis and elevated lactate levels so often seen accompanying TBI became a therapeutic target in a clinical study in which tris-(hydromethyl)-aminomethane (THAM) was administered to victims of TBI. THAM is an alkalizing agent that can buffer CO2 and acids. Its use in animal models resulted in reduced edema, lower ICP, and more favorable energy kinetics. Unfortunately, this did not translate as well to humans, and no advantage in outcome was observed.61,89
Edema/Increased Intracranial Pressure
Brain swelling occurs in almost all patients with severe brain injury and in 5% to 10% of those with moderate injuries (also see Chapter 322 and Chapter 324).90,91 The morbidity associated with brain injury was once thought to largely be correlated with the extent of posttraumatic edema. Currently, the significance of edema is acknowledged, but we now appreciate the multifaceted pathobiology constituting TBI.
Without question, posttraumatic edema contributes heavily to intracranial hypertension. The hypertension may then lead to decreased CPP (CPP = MAP − ICP, where MAP is mean arterial pressure) and to ischemia, thereby inciting continued cytotoxic edema and progressive increases in ICP. This vicious cycle is depicted in Figure 327-2. Historically, breakdown of the BBB with protein extravasation (termed vasogenic edema) was believed to be the primary component of the edema seen after TBI.92 This misconception led some practitioners to use corticosteroids for the management of TBI. This has now clearly been shown to be harmful, with the Corticosteroid Randomization After Significant Head Injury trial (>9000 patients enrolled, the largest TBI trial to date) revealing that corticosteroids were associated with worse outcomes when used in patients with severe TBI.93 Although vasogenic edema does contribute to the overall edema seen in TBI, early work by Marmarou and colleagues used mathematical modeling techniques to reveal that the vascular engorgement component of brain swelling after severe brain injury probably represents only about 25% of the overall increase in brain bulk, with the remainder being due predominantly to cytotoxic edema.94 Vasogenic edema probably becomes important around focal contusions on the 2nd through the 10th to 15th days and is most likely associated with transient opening of the BBB to medium-molecular-weight markers (50 to 70 kD).95 In humans studied with both gadolinium-enhanced magnetic resonance imaging (MRI) and pertechnetate-enhanced single-photon emission computed tomography, vasogenic edema is seen only at later time points around contusions and not at all in patients with diffuse nonfocal injuries.96–98 There is now supporting evidence that the majority of early brain edema, both global and focal, is cytotoxic rather than vasogenic. This was further substantiated recently by Marmarou and colleagues, who used diffusion-weighted MRI to evaluate patients with severe TBI. This study revealed that cytotoxic (intracellular) edema is indeed the predominant form of edema present after TBI.99 The mechanisms responsible for generating this cytotoxic edema are a continuing topic of great interest. Identified mechanisms responsible for the generation of cytotoxic edema include hypoxia-ischemia, channelopathy/excitotoxicity, mechanical membrane disruption, and aquaporins.
Hypoxia-Ischemia
Hypoxic-ischemic events are common after TBI and result in failure of ionic pumps, specifically the sodium-potassium pump (see Table 327-2). Pump malfunction causes the cell to lose its innate homeostatic environment via failure of sodium extrusion and potassium uptake. This failure will bring about the accumulation of sodium within the cell and lead to the influx of water because of the altered osmotic gradient.16
Channelopathy/Excitotoxicity
A major source of cytotoxic edema is traumatically induced accumulation of EAA neurotransmitters such as glutamate and glycine.49 These neurotransmitters bind their receptors, activate and open membrane channels, and induce sodium influx. As with ionic pump failure, osmotic pressure dictates that water will follow. In addition to the oncotic effects of sodium entry, sodium influx will also cause membrane depolarization with an influx of chloride—also resulting in swelling.100
Mechanical Membrane Disruption
Trauma induces brief and spontaneously recovering mechanoporation of neuronal and axonal membranes. Membrane disruption causes efflux of potassium from neurons and resultant astrocytic swelling, as confirmed by ultrastructural examination of pericontusional tissue taken as little as 10 minutes after impact.101
Aquaporins
Aquaporins are a relatively new family of at least nine proteins that have been shown to be involved in the formation of cellular water channels under a variety of circumstances.102 At this time, aquaporin expression across injury models has proved inconsistent and is incompletely understood. For example, increased aquaporin expression is present after ischemic injury, whereas it is decreased after a cortical impact model of TBI.103 These conflicting data appear to indicate that aquaporin expression may lead to the accumulation intracellular edema under some circumstances while playing a role in the clearance of edema in others.104 Aquaporin-4 (AQP4) is currently thought to be one of the primary cellular water channels in the brain and is localized to astrocytic foot processes along the basal lamina and brain–cerebrospinal fluid interface. Our understanding of AQP4 has been greatly augmented by the study of AQP4 knockout mice in several models of brain injury.102 In models of cytotoxic edema, AQP4 deletion or alteration has been shown to be protective.104 In contrast, AQP4 deletion in models of vasogenic edema results in decreased clearance of edema and greater progression of disease.104 In addition to AQP4, Tran and associates described a potential role for AQP1 in water homeostasis after experimental TBI.105 AQP1 is another member of the aquaporin family found in the brain and can participate in CO2 transportation across the cellular membrane. Interestingly, AQP1’s promoter contains a glucocorticoid response element. Thus, Tran and colleagues hypothesized that “AQP1 may be involved with edema-related brain injury and might be modulated by external conditions such as the pH and the presence of steroids.” The importance of aquaporins in modulating edema after trauma is thus clear. It further suggests the potential for the development of pharmacologic strategies targeting aquaporin function and expression to dramatically alter our ability to treat cerebral edema, which is currently limited to osmotic agents.
Excitotoxicity
Glutamate excitotoxicity is a self-perpetuating process that triggers numerous injurious intracellular mechanisms. After TBI there is direct release of excessive EAAs, such as glutamate and aspartate, from presynaptic nerve terminals and astrocytes into the extracellular space. This process is depicted in Figure 327-3 and begins when these EAAs bind to the appropriate postsynaptic receptor (NMDA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate [AMPA]). Activation of these ion channels will cause intracellular Ca2+ and Na+ levels to rise, followed passively by movement of Cl− and water. The resultant combination of intracellular volume and Ca2+ overload induces organelle swelling, plasma membrane swelling,106 necrosis,107 and apoptosis and leads to the activation of destructive enzymes108 (phospholipases, calpain, caspase, and nitric oxide synthase [NOS]), as depicted in Figure 327-3. Glutamate-driven excitotoxicity will further depolarize the cell, activate voltage-dependent calcium channels, and thereby propagate a dangerous positive feedback loop.6 The downstream effects of these events are shown in Figure 327-3 and are individually discussed under the appropriate subheading within this chapter.
Clinical Implications
There has been great historical interest in pharmacologic modification of glutamate excitotoxicity. Six phase II-III clinical trials (Eliprodil, Selfotel Int, Selfotel U.S., Cerestat, Saphir/D-CPP-ene, and CP101-606) directed by this strategy have been performed, but only one showed a survival benefit.56 Dexanabinol, a cannabinoid with pluripotent functionality, including noncompetitive NMDA receptor antagonist properties (as well as being a free radical scavenger and inhibitor of tumor necrosis factor-α [TNF-α]), was evaluated in a phase II trial; it was shown to be safe and well tolerated and seemed to decrease ICP. However, the phase III efficacy trial showed no effect on outcome.69
Calcium Dysregulation
As mentioned, TBI always leads to intracellular influx of calcium through numerous routes, including but not limited to (1) opening of voltage-dependent channels induced by mechanical deformation of membrane and ion channels, (2) opening of agonist-dependent channels mediated by neurotransmitter substances released in excess into ECF, and (3) opening of specific calcium channels. It is clear that calcium is strongly implicated in the propagation of several deleterious cascades responsible for the generation of neuronal injury and death after TBI.109,110 It has been linked to many downstream mechanisms of injury, including activation of cysteine proteases and subsequent cytoskeletal proteolysis,111 mitochondrial permeability transition,112 free radical toxicity, and mechanical perturbation of neuronal membranes.113 These mechanisms have been implicated in the initiation of various forms of cell death, including apoptosis and necrosis.114 Regardless of the etiology, unregulated calcium influx can cause calcium to be released from intracellular stores, as well as glutamate-containing exocytotic vesicles, thereby further increasing cytosolic calcium in a vicious cycle of cell destruction. This massive calcium surge can overwhelm cells’ buffering capability, a duty largely handled by mitochondria and ionic pumps under normal circumstances.115 The result is the activation of various enzymes, including the aforementioned cysteine proteases (calpain, caspase, and cathepsin). Induction of these important proteases has been shown in focal116 and diffuse117 models of TBI. Calpain activity has been noted in both necrotic and apoptotic forms of cell death, whereas caspase-3 activity is seen only in apoptosis.118 Finally, mitochondria have a central role in calcium homeostasis, and recent evidence highlights their function in cell death via opening of the mitochondrial permeability transition pore (MPTP), a calcium-dependent process.119
Clinical Implications
The central role of calcium in mediating cell damage and death led to trials of calcium antagonists in head-injured patients. The dihydropyridine calcium antagonist nimodipine has shown little overall benefit in unselected populations with head injury.52,53 Nimodipine has, however, shown a trend for a more favorable outcome in patients with traumatic subarachnoid hemorrhage.120 Poor brain penetration has been implicated as a major factor in the limited effect of nimodipine in TBI.
Cytoskeletal Proteolysis
The cytoskeleton consists of three main protein components: microfilaments, neurofilaments, and microtubules. After TBI, calcium-induced activation of calpain results in proteolysis of the cytoskeleton and may play an integral role in delayed neuronal degeneration—so-called calpain-mediated spectrin proteolysis (CMSP).121 In axonal stretch injury, within minutes there is malalignment and distortion of the cytoskeletal components,122 which leads to loss of microtubules and increased spacing of neurofilaments, especially at the node of Ranvier. In addition to trauma, inhibition of calpains can limit the proteolysis of MAP2 (a type of membrane-associated protein contained in microtubules) after ischemia.123,124 The caspases have also been linked to the breakdown of cytoskeletal proteins such as MAP2, α-spectrin, and neurofilaments.125–127
Clinical Implications
Targeting of CMSP has yet to reach the clinical trial stage. It has been used experimentally with modest success through the use of calpain inhibitors in TBI. In one particular study, a fluid percussion injury model was used to generate TBI in rats.128 The calpain inhibitor MDL-28170 was administered after injury, and axons in the corpus callosum were evaluated with amyloid precursor protein immunohistochemistry. When the drug was given within 30 minutes after injury, there was both structural and functional benefit in terms of axonal injury burden. However, the functional benefits diminished when the drug was given beyond 30 minutes after injury. In short, calpain inhibition has demonstrated potential (i.e., reduction of cytoskeletal breakdown and neuronal degeneration in animal models), yet the results are modest and require further investigation before they can be translated to human trials.
Derangements in Brain Metabolism after Traumatic Brain Injury
Because the brain is dependent on aerobic metabolism for delivery of substrate (oxygen and glucose) and because of the frequent impairment of oxygenation and perfusion that occurs after severe head injury, metabolic derangement is an extremely common and important consequence of TBI. The metabolic changes may be global or focal, with evidence demonstrating metabolic derangement after TBI coming from several methodologies, including (1) a 2-deoxyglucose technique to measure glucose metabolism, (2) positron emission tomography (PET) studies using fluorodeoxyglucose in humans, (3) measurement of the jugular/arterial differences in oxygen and lactate to yield global measures of oxygen consumption and lactate production in humans (AVDO2 and CMRO2), (4) measurements of whole-brain metabolic indices with magnetic resonance spectroscopy, and (5) measurements of focal ECF indices using microdialysis and PbtO2 monitors. The data from these studies allow certain generalized conclusions. TBI induces massive ion fluxes across neuronal membranes, widespread loss of resting membrane potential, and release of neurotransmitters into the extracellular space. Within minutes of these events, the brain attempts to restore ionic homeostasis by ionic pumping and reuptake of neurotransmitters. These processes are intensely energy dependent and result in an abrupt increase in glucose utilization. Studies based on the fluid percussion model in rats have shown that this increase in glucose metabolism, to facilitate the generation of adenosine triphosphate (ATP), is brief and maximally localized to parts of the brain that are maximally deformed by the shearing forces.128 When focal lesions such as subdural hematoma, focal infarction, or cerebral contusion are present, glucose use increases for a period in the penumbral border zone.129,130 Furthermore, in humans, PET studies have shown that these increases in glucose are maximal in the penumbral zone of contusions and in the hemisphere underlying hematomas when this part of the brain is viable.131 This increase may persist for 2 to 4 hours in the rat and 5 to 7 days in humans.129,131 In both animal models and humans, glucose use is depressed when measured days after diffuse injury and remains so for weeks after impact, which is consistent with the reduced metabolic needs of the comatose brain.132
Clinical Implications
Hypothermia was an initially promising intervention intended to improve outcomes after TBI that is best discussed as a targeted therapy for the metabolic derangements seen after TBI. Historically, clinical interest in induced hypothermia is derived from experimental models of head injury in which hypothermia has been shown to (1) reduce energy requirements of the brain, (2) stabilize cell membranes,133 (3) improve posttraumatic CBF-metabolic uncoupling, (4) attenuate axonal injury, (5) reduce inflammation, and (6) reduce ICP. There have been at least 12 clinical trials performed, with approximately 6 ongoing trials, and thus far no outcome benefit has been revealed. As of this writing, it is unclear why no benefit has been observed. Hypothermia does have many associated potential systemic complications, and these may be negating the beneficial effects. Therefore, prophylactic hypothermia has currently lost momentum as a standalone therapy for TBI. However, ongoing research is being performed to see whether it might expand the window of opportunity to introduce other therapeutic strategies, both pharmacologic and physiologic. In addition, work is ongoing to more carefully evaluate the role of therapeutic hypothermia, in particular for intracranial hypertension, in the management of TBI.
Mitochondrial Permeability Transition
Clearly, mitochondrial integrity is pivotal in the maintenance of cellular metabolic homeostasis. The occurrence of a process involving the increased permeability of mitochondrial membranes en route to cellular death has long been speculated. Kroemer and coauthors proposed the term mitochondrial permeability transition (MPT) to describe a calcium-induced process of increased mitochondrial membrane permeability.134 Since that time there has been a tremendous explosion in research on MPT and cell death. MPT has been described as increased permeability of the inner membrane of mitochondria via the opening of channels called MPTPs. Opening of MPTPs is a devastating event with resultant loss of transmembrane potential, mitochondrial swelling, and eventual rupture of the outer mitochondrial membrane. This loss of mitochondrial function produces profound deficiencies in neuronal metabolism and ionic equilibrium. To clarify terminology, mitochondrial permeability transition is the term used to identify the entire process, whereas mitochondrial permeability transition pore is the term used to describe the actual channel (therefore, opening of the MPTP allows the process of MPT to occur).
Clinical Implications
Cyclophilin D (CyD) is a member of the cyclophilin family involved in protein folding that is normally found in the mitochondrial matrix but migrates to the inner membrane during MPT. In experiments using the drug cyclosporine to block CyD, MPT activity was significantly attenuated, thereby establishing CyD as a critical player in MPT. As discussed previously, cyclosporine is a Food and Drug Administration–approved drug that has been shown to confer neuroprotection after TBI in multiple animal models and is being evaluated for potential use in neurocritical care.135 Interestingly, a recent in vitro study using isolated mitochondria revealed that the presence of phosphate was necessary for the inhibition of MPTP opening by cyclosporine administration or CyD knockout.136 Investigations are ongoing to determine the relationship, if any, of the neuroprotective effect of cyclosporine specifically on the blunting of MPT. Studies have investigated the role of MPT in apoptosis. Although cyclosporine-dependent MPT has been tied to the release of cytochrome c137 and thus to apoptosis, other studies have failed to show inhibition of apoptosis by cyclosporine-mediated blockage of MPT.138 Furthermore, one study indicated that CyD-dependent MPT regulates necrotic cell death but not apoptotic cell death.139 Finally, overexpression of CyD has been shown to be somewhat protective of apoptotic death. Overall, it is clear that MPT is an important mediator of necrotic cell death and can be implicated in apoptosis. However, its exact role in apoptosis remains unclear and warrants further investigation.
DNA Damage
Figure 327-3 illustrates several of the pathways that lead to DNA damage, including caspase-independent apoptosis (via apoptosis-inducing factor [AIF]), caspase-dependent apoptosis, and oxidative stress (through the generation of nitric oxide [NO]). DNA damage will then activate poly(adenosine diphosphate ribose) polymerase (PARP), which will induce the release of AIF through sequential activation of calpain and then Bax. It should be pointed out that DNA damage is usually associated with apoptosis, but this is not always the case because apoptosis can occur without DNA fragmentation.140 Furthermore, the presence of DNA fragmentation should not be used to exclusively indicate apoptosis.
Neuronal injury mechanisms involving DNA damage are also often mentioned in concert with induction of the tumor suppressor gene p53, which can serve as a transcription factor for a variety of genes with many actions (e.g., growth arrest, promotion of apoptosis). Increased expression of p53 with an associated increase in DNA fragmentation and neuronal apoptosis has been shown after focal141 and diffuse TBI.142 Increased endonuclease activity resulting in DNA fragmentation has been demonstrated in animal models of head injury.143 These delayed apoptotic mechanisms may be more amenable to pharmacologic blockage because of their long “window of opportunity” in contrast to other neuroprotective mechanisms, and at least two current neuroprotective trials involving TBI (Solvay SL334 and Neuren NZ1366) are in part targeting this mechanism.
Free Radical Formation
Free radicals are highly reactive ionic molecules bearing an unpaired electron in their outer electron shell. This unpaired electron confers high chemical reactivity. Free radicals are the normal by-product of oxidative metabolism within mitochondria, and they fulfill important physiologic roles such as signaling and polymorphonuclear leukocyte–mediated destruction of bacteria.144,145 Reactive oxygen species (ROSs) are inherently injurious, however, and contribute to many diseases affecting the CNS, including Parkinson’s disease, Alzheimer’s disease, multiple sclerosis, and amyotrophic lateral sclerosis.146 ROSs are also important contributors to secondary injury and are produced early after neurotrauma.147,148 Indeed, many secondary injury processes lead to free radical production, and in turn, free radicals feedback positively to increase the activity of many of these harmful processes.
