Chapter 9 Neural Injury at the Molecular Level
The histopathologic appearance of chronic cervical spondylotic myelopathy has been well described and includes the characteristic features of regional demyelination extending axially from the site of compression, preferential lateral column axonal loss, and anterior horn neuron dropout.1–5 Ongoing research projects are creating a better understanding of myelopathy on a molecular level, and recent studies indicate that a significant portion of cell loss appears to be caused by the process of programmed cell death, also known as apoptosis. Although the molecular pathways regulating apoptosis are extremely complex, programmed cell death affects restricted populations of spinal cord cells—including oligodendrocytes and some neuronal and astrocytic subpopulations—suggesting the possibility that targeted antiapoptotic therapy may be a reasonable goal for the treatment or prevention of myelopathy.
Microbiology of the Oligodendrocyte
The oligodendrocyte has been shown to play a pivotal role in several complex biologic processes, including development, injury repair, disease process modulation, and the formation and maintenance of myelin.6,7 During the early stages of human development, a large oligodendroglial population is generated, and an estimated 50% of these cells eventually disappear by the process of apoptosis.8 As the central nervous system matures, the oligodendroglia become responsible for the creation and maintenance of myelin sheaths. These sheaths, although formed directly from oligodendroglial cell membrane, demonstrate key biochemical differences from the parent cell membrane in terms of both chemical and protein composition.9 The biochemical and physiologic characteristics of the relatively small protein constituent are especially important, and absence or alteration of the major protein components (i.e., proteolipid protein or myelin basic protein) can lead to the advent of severe demyelinating disease.10 Another unique feature of the oligodendrocyte is the high concentration of microtubules, which contribute to formation of an elaborate cytoskeletal framework, allowing myelin sheath formation at remote distances from the cell karyon.6
Considerable progress has been made in understanding the response of oligodendroglial cells to injury, and a more complete understanding of this complex process may lead to a greater appreciation of the mechanism of injury in such processes as cervical spondylotic myelopathy. Studies suggest that the oligodendrocyte is particularly sensitive to a wide range of oxidative, chemical, radiation-induced, and mechanical injuries. High iron content and relatively inefficient antioxidant defense mechanisms appear to render the oligodendrocyte vulnerable to oxidative stress.11–14 Injury-related release of intracellular iron may contribute to the generation of damaging hydroxyl radicals through the Fenton reaction.15 In addition, in vitro exposure of mature oligodendrocytes to hydrogen peroxide has been shown to induce apoptotic cell death, but preincubation of these cells with an iron chelator, such as deferoxamine, appears to confer some protection from oxidative cytotoxicity and apoptosis.16,17
Toxins that impair mitochondrial respiration, such as cuprizone and ethidium bromide, have also been shown to trigger apoptosis in oligodendroglial cells. Subsequently, these chemicals have been used to develop experimental models of demyelinating disease and injury. It has been established that radiation exposure directly damages DNA and has been shown to lead to apoptotic cell death in many cell types. However, several studies of delayed neurologic injury after radiation therapy have revealed that oligodendrocytes are the most radiation-sensitive cell population in spinal cord tissue.18,19
In addition to the previously mentioned sources of oligodendrocyte injury, mechanical stress has been repeatedly shown to trigger oligodendrocyte apoptosis. Mechanical injury appears capable of triggering a specific immune response with formation of antibodies and subsequent cytotoxicity directed against oligodendrocyte antigens.20 This immune-mediated injury may be caused by macrophage activity and appears to involve several different cytokines, such as tumor necrosis factor, lymphotoxin, and gamma-interferon.21–24 These activated macrophages also generate free radicals and nitric oxide, which have been shown to lead to apoptosis.25,26 Formation of the membrane attack complex through activation of the complement cascade is another consequence of macrophage activation and has been implicated in oligodendrocyte injury.
In addition to the macrophage, at least two specific subpopulations of T cells may also be involved in oligodendroglial apoptosis. CD4+ T cells adhere to target cells through the Fas receptor identified on oligodendrocyte cells, thereby triggering apoptosis. Gamma-delta T cells have been found to co-localize with oligodendrocytes (expressing heat-shock protein 65), and may trigger cell death through production of gamma-interferon.27
Apoptosis
Apoptosis, also known as “programmed cell death,” may be the primary cellular process underlying the disappearance of oligodendrocytes in the earliest histologic stages of traumatic spinal cord injury (SCI) and other processes such as cervical spondylotic myelopathy. The process of apoptosis is distinct from necrosis and involves a sequence of intracellular events that includes chromatin aggregation and internucleosomal DNA fragmentation, nuclear pyknosis, and subsequent cell shrinkage.28,29 Apoptosis ultimately results in phagocytic engulfment of cells without extracellular discharge of cytosolic contents, and without generation of a local inflammatory response.30
In contrast to necrotic cell death, apoptosis is a much more abbreviated process that has made its study relatively difficult. Apoptotic cells initially shrink and lose contact with adjacent cells, forming membrane blebs and expressing prophagocytic cell surface signals. The cell chromatin then condenses and fragments, and the process ends in compartmentalization of the entire cell into small, membrane-bound vesicles that are quickly phagocytized. By comparison, cell necrosis is a relatively prolonged affair that is characterized by cell membrane disruption, mitochondrial swelling, random DNA cleavage, and the generation of a local inflammatory reaction.31
Several molecular biology assays have been developed for identification of apoptosis in various settings. A marker of DNA cleavage, such as the terminal deoxynucleotidyltransferase (TdT)-mediated nick-end labeling (TUNEL) technique, is a popular assay. Interpretation of studies relying solely on TUNEL staining has been criticized as possibly being limited by the observation that this method has been found to label cells undergoing necrosis as well and may not be as specific for apoptosis as once thought.31 Internucleosomal DNA cleavage, a hallmark of apoptosis, is demonstrated by a characteristic “laddering” pattern on gel electrophoresis, and this finding can reinforce the results of TUNEL staining. The most specific method for identifying apoptotic cells, however, remains direct histologic examination and the identification of chromatin condensation along the nuclear periphery, condensation of the cytoplasm with intact organelles, and membrane blebbing.32 A newly developed commercial assay is also available that uses monoclonal antibody to single-stranded DNA (Apostain; eBioscience, San Diego, CA). This method is purported to detect the earliest stages of apoptosis occurring before DNA fragmentation and supposedly has no cross-reactivity for necrotic cells.33
Molecular Mechanisms of Apoptosis
The molecular pathways involved in apoptosis have been extensively examined, but were initially studied in the roundworm, Caenorhabditis elegans. These studies led to the discovery of one of the first genes associated with apoptosis, which was appropriately named CED 3 in honor of this worm.34 Subsequently, a homologous family of apoptosis-related protein products has been identified in mammals and termed the CED 3/ICE (interleukin-1β-converting enzyme) family.35–37 These proteins, also known as caspases, serve as functional cysteine proteases.38 At least 10 distinct members of this gene family have been identified thus far, and at least 2 of these proteins, caspase-3 and caspase-9, have been strongly associated with apoptosis in human cells.39,40 The intracellular cascade involving caspase-3 ends in activation of specific endonucleases that cleave DNA strands into the characteristic internucleosomal fragments.41 Production of these 185 base-pair fragments results in the DNA laddering that is one of the histologic hallmarks of apoptosis. Activation of caspase-9 appears specifically to induce mitochondrial release of cytochrome c, which is one of the earliest intracellular events in apoptosis.42 Targeted inhibition of caspase-1 (ICE) and caspase-3 (CPP-32) in oligodendrocytes has been shown to prevent apoptotic death of these cells.40
As previously described, numerous chemical and biologic triggers for apoptosis have been identified. Mature oligodendrocytes are particularly sensitive to oxidative stress.11 Experimental exposure of oligodendroglial cells to hydrogen peroxide leads to increased expression and nuclear translocation of transcription factors nuclear factor-κB (NF-κB) and activator protein-1 (AP-1), both implicated as critical elements in the apoptotic pathway.17
One of the most important biologic triggers of oligodendrocyte apoptosis in SCI may be tumor necrosis factor-α (TNF-α). TNF-α has been shown to induce apoptosis in oligodendrocytes, both in vitro and in vivo.43–45 Designated death domains located on the intracellular side of the type I receptor for TNF-α (TNFR1) and related receptors have been associated with activation of caspase-3 and caspase-8, which subsequently leads to apoptosis.40 Gamma-interferon may further enhance susceptibility of oligodendrocytes to TNF-α-triggered apoptosis through up-regulation of the so-called death receptor, Fas.46 It has also been reported that the p38 and Jun N-terminal kinase (JNK) pathways play a role in the transmission of apoptosis signals following SCI. Further findings indicate that activation of JNK by TNF-α promotes expression of apoptosis signal-regulating kinase 1 (ASK1).47
The oligodendrocyte apoptotic signal transduction pathway appears to begin with ligand binding to either Fas (CD95 or Apo1) or p75 (low-affinity neurotrophin receptors) cell surface receptors. These proteins are members of the TNFR family and have been shown to co-localize with cells undergoing apoptosis in a rat model of cervical SCI.32 Binding of Fas ligand (FasL) to the extracellular cysteine-rich domain of Fas results in formation of oligomers, which allows interaction of the intracellular death domain with Fas-associated death domain protein (FADD).48 Once the association is made, the death domain of FADD then interacts with procaspases 8 and 10 and triggers a caspase activation cascade that ultimately ends in activation of at least three different effector enzymes, caspases 3, 6, and 7.32 These effector molecules presumably interact with additional downstream targets, ultimately leading to cell apoptosis.49 FLICE (FADD-like interleukin-1β-converting enzyme) proteins are proteins demonstrating sequence homology with the caspases, but acting as inhibitors of the apoptosis-triggering pathway.50
Another important apoptosis pathway involves the p53 tumor suppressor protein, as well as the proteins p21, Bcl-2, and Bax.51 In a rat model of SCI, p53 protein appeared within 30 minutes of injury, co-localizing with apoptotic glial cells and spreading in distribution over the course of 2 days.51 Cellular studies have further demonstrated that exposure of oligodendroglial cells to hydrogen peroxide leads to rapid translocation of p53 from the cytosol to the nucleus and cell death by apoptosis.52
Apoptosis in Traumatic Spinal Cord Injury
It has been well established that cell loss in traumatic SCI occurs both at the time of injury and secondarily over a period of days to weeks after the event. At the epicenter of injury, the majority of cell death occurs through necrosis, with macrophages and microglia becoming actively engaged in phagocytosis of necrotic cell debris.53 However, cell loss in spinal cord white matter continues throughout a much more extensive axial section of the cord for up to several weeks in a process referred to as secondary injury. Although it has become apparent that this continued cell loss significantly worsens neurologic outcome in SCI, the underlying biologic mechanisms remain poorly understood. Several studies have suggested, however, that the primary process involves oligodendrocyte apoptosis.54–59
Initial evidence that apoptosis contributes to ongoing cell death after acute SCI came from animal studies involving the rat.60 It was demonstrated that acute compressive cord injury leads to preferential apoptosis of oligodendrocytes along degenerating longitudinal white matter tracts.55 These initial findings were subsequently supported by similar results in other animal models, including primates.56 In most of these animal studies, visible signs of oligodendrocyte apoptosis appear within 24 hours and continue for at least 3 weeks after injury.54–57,60–63
A histopathologic study of human SCI indicates that oligodendrocyte cell death by apoptosis can continue from 3 hours to at least 8 weeks after injury.64 In this study, oligodendrocyte apoptosis appeared to correlate with specific patterns of wallerian degeneration and was associated with intracellular activation of caspase-3. Apoptosis was more pronounced in ascending white matter tracts, and the authors speculated that this finding may reflect the histopathologic observation that wallerian degeneration affects ascending tracts before descending ones.65 The extent of oligodendrocyte apoptosis was shown to correlate with the severity of neurologic injury, being significantly less extensive in patients with incomplete neurologic deficits. This correlation of apoptosis and neurologic impairment is in agreement with previous findings from animal studies.57 Of note, neuronal apoptosis was not seen, suggesting that neuronal loss occurs through the process of necrosis.
The biochemical trigger for oligodendrocyte apoptosis related to traumatic SCI is currently unknown but is likely to be multifactorial. It has been observed that SCI is characterized by significant intracellular Ca2+ shifts, and several apoptotic processes are Ca2+ dependent, including DNA fragmentation and proteolysis.66,67 Similarly, acute SCI has been associated with hypoxia and free radical formation, which are also established triggers of apoptosis.68,69 Glutamate excitotoxicity has also been implicated in secondary SCI and appears to lead to apoptotic cell death.70
Animal models have provided most of the information regarding biochemical responses to SCI. A rat model of SCI has demonstrated increased local TNF-α expression within 1 hour of injury, followed by increased nitric oxide levels at 4 hours.71 This model used a neutralizing antibody against TNF-α, and significantly reduced nitric oxide levels as well as the extent of apoptosis. Similarly, addition of a nitric oxide synthase inhibitor, N-monomethyl-l-arginine acetate (L-NMMA) also reduced the number of apoptotic cells. These findings suggest that TNF-α signaling triggers apoptotic cell death after SCI, and that this effect is at least partly mediated by nitric oxide. Of note, the amount of decrease in apoptosis after administration of L-NMMA (42%) was less than half that observed after TNF-α antibody administration (89%), implying the existence of multiple parallel apoptotic pathways. A recent study by Genovese et al. demonstrated the neuroprotective effects of selective adenosine A2A receptor agonists, which act by decreasing the overall expression of myeloperoxidase, NF-κB, and inducible nitric oxide synthase (iNOS), and decreasing the activation of JNK mitogen-activated protein kinase (MAPK) in oligodendrocytes.72 In addition, another recent study found that mice with SCI, when treated with ethyl pyruvate, showed no increase of TNF-α expression and a decrease in oligodendrocyte apoptosis.73
Several studies of development suggest that specific trophic factors are produced by axons and that absence of these factors results in oligodendrocyte apoptosis.74–76 Members of the neuregulin ligand family, in particular the glial growth factor (GGF), bind to the HER4 receptor on the surface of oligodendrocytes and appear to play an important role in cell differentiation and survival.77 Alternatively, the traumatic event may result in direct release of proapoptotic factors into spinal cord tissue. It is well established that activated microglia release several factors that may cause apoptosis, including TNF-α, reactive oxygen intermediates, and nitric oxide.78,79 Administration of exogenous thyroid hormone (triiodothyronine [T3]) during the early period after acute SCI has also been found to increase the population of apoptotic cells.80
Apoptosis in Chronic Spinal Cord Compression
Several studies have suggested an important role for ischemic tissue injury in the pathogenesis of myelopathy in the setting of cervical spondylosis. On the cellular level, the sensitivity of oligodendrocytes to hypoxic injury is well established and appears to support the possibility of an ischemic cause.81 However, neurons are relatively more vulnerable to ischemic injury, and their sparing in early myelopathy makes a purely ischemic cause for cervical spondylotic myelopathy somewhat unlikely.
Although necrosis and apoptosis often occur simultaneously, distinguishing the two processes provides important information regarding the causes of specific disease processes. Although ischemia has been associated with apoptotic cell death, severe ischemia is characteristically thought to result in cell necrosis. Because oligodendrocyte disappearance in both trauma and chronic spondylotic myelopathy is apoptotic in nature, it is thought that mechanisms other than pure ischemia are involved.55
Animal models strongly support a role for apoptotic cell death in the tissue degeneration seen in chronic, compression-related cervical myelopathy. The tiptoe-walking Yoshimura (twy) mouse is a specific strain of inbred mouse that has been useful as a model for chronic spinal cord compression.82 Twy mice become quadriparetic 4 to 8 months after birth because of the development of local hyperostosis along the dorsolateral margins of the C1 and C2 vertebrae, which results in severe cord compression at this level.83 Histologic examination of spinal cord tissue from these mice has revealed a characteristic pattern of descending degeneration affecting the anterior and lateral columns and ascending degeneration along the posterior columns. These findings are in addition to severe tissue damage at the level of compression.84 Cavity formation and myelin ovoids (myelin debris) were observed extending from the zone of compression into adjacent levels without gross deformation of the spinal cord. Detection of apoptotic cells using the TUNEL assay revealed a distribution of glial apoptosis that appeared to mirror the pattern of degeneration, whereas cell-specific staining confirmed that apoptotic cells were oligodendrocytes. The investigators included an autopsy study of a human patient dying with cervical myelopathy resulting from ossification of the posterior longitudinal ligament, in which a pattern of neuronal loss, demyelination, and apoptosis was observed that was similar to the findings in the twy mouse. Further studies of the twy mouse showed increased expression of TNFR1 and TNFR2 in chronically compressed spinal cord tissue, which further elucidates the effect of chronic compression on apoptosis and demyelination.85
Oligodendrocyte survival depends on the presence of specific so-called survival factors produced by neighboring axons, leading to the possibility that oligodendroglial cell loss merely reflects prior neuronal injury. However, oligodendrocyte apoptosis likely precedes axonal degeneration in chronic myelopathy, as evidenced by both human and animal studies of spinal cord compression demonstrating apoptotic oligodendrocytes in the setting of intact demyelinated axons.65,86,87
Prevention of Apoptosis
Oxidative stress has been shown to be a potent trigger for apoptotic death of oligodendrocytes.16 Conversely, antioxidant therapy with pyrrolidine dithiocarbamate (PDTC) and vitamin E appears to moderate this effect considerably.17 The asymmetric distribution of phospholipid polar-head groups across the plasma membrane bilayer may play a role in determining vulnerability to oxidative stress.88 Normally, there is an over-representation of choline phosphoglyceride and sphingomyelin in the outer leaflet, whereas the aminophospholipids, ethanolamine phosphoglyceride (EPG) and serine phosphoglyceride (SPG), are over-represented in the inner leaflet. Apoptosis has been associated with redistribution of SPG and EPG and loss of aminophospholipid asymmetry.89
