Na(+)-Ca2+ exchange in anoxic/ischemic injury of CNS myelinated axons

The back and forth of axonal injury and repair after stroke

PURPOSE OF REVIEW

The axon plays a central role in both the injury and repair phases after stroke. This review highlights emerging principles in the study of axonal injury in stroke and the role of the axon in neural repair after stroke.

RECENT FINDINGS

Ischemic stroke produces a rapid and significant loss of axons in the acute phase. This early loss of axons results from a primary ischemic injury that triggers a wave of calcium signaling, activating proteolytic mechanisms and downstream signaling cascades. A second progressive phase of axonal injury occurs during the subacute period and damages axons that survive the initial ischemic insult but go on to experience a delayed axonal degeneration driven in part by changes in axoglial contact and axonal energy metabolism. Recovery from stroke is dependent on axonal sprouting and reconnection that occurs during a third degenerative/regenerative phase. Despite this central role played by the axon, comparatively little is understood about the molecular pathways that contribute to early and subacute axonal degeneration after stroke. Recent advances in axonal neurobiology and signaling suggest new targets that hold promise as potential molecular therapeutics including axonal calcium signaling, axoglial energy metabolism and cell adhesion as well as retrograde axonal mitogen-activated protein kinase pathways. These novel pathways must be modeled appropriately as the type and severity of axonal injury vary by stroke subtype.

SUMMARY

Stroke-induced injury to axons occurs in three distinct phases each with a unique molecular underpinning. A wealth of new data about the molecular organization and molecular signaling within axons is available but not yet robustly applied to the study of axonal injury after stroke. Identifying the spatiotemporal patterning of molecular pathways within the axon that contribute to injury and repair may offer new therapeutic strategies for the treatment of stroke.

Primary paranode demyelination modulates slowly developing axonal depolarization in a model of axonal injury

Neurological sequelae of mild traumatic brain injury are associated with the damage to white matter myelinated axons. In vitro models of axonal injury suggest that the progression to pathological ruin is initiated by the mechanical damage to tetrodotoxin-sensitive voltage-gated sodium channels that breaches the ion balance through alteration in kinetic properties of these channels. In myelinated axons, sodium channels are concentrated at nodes of Ranvier, making these sites vulnerable to mechanical injury. Nodal damage can also be inflicted by injury-induced partial demyelination of paranode/juxtaparanode compartments that flank the nodes and contain high density of voltage-gated potassium channels. Demyelination-induced potassium deregulation can further aggravate axonal damage; however, the role of paranode/juxtaparanode demyelination in immediate impairment of axonal function, and its contribution to the development of axonal depolarization remain elusive. A biophysically realistic computational model of myelinated axon that incorporates ion exchange mechanisms and nodal/paranodal/juxtaparanodal organization was developed and used to study the impact of injury-induced demyelination on axonal signal transmission. Injured axons showed alterations in signal propagation that were consistent with the experimental findings and with the notion of reduced axonal excitability immediately post trauma. Injury-induced demyelination strongly modulated the rate of axonal depolarization, suggesting that trauma-induced damage to paranode myelin can affect axonal transition to degradation. Results of these studies clarify the contribution of paranode demyelination to immediate post trauma alterations in axonal function and suggest that partial paranode demyelination should be considered as another "injury parameter" that is likely to determine the stability of axonal function.

Pathophysiology of HNPP explored using axonal excitability

OBJECTIVE

Hereditary liability to pressure palsies (HNPP) is an autosomal dominant disorder of myelination resulting in susceptibility to pressure palsies from compression or stretching of peripheral nerves.

PATIENTS AND METHODS

This study examined axonal excitability at two sites (one distal and one proximal) in five patients with biopsy and genetically proven HNPP to understand the pathophysiology of the disease. Comparisons were made with age-matched control subjects as well as five Charcot-Marie-Tooth type 1A patients to contrast the findings and explain the different phenotypes of diseases affecting the same gene.

RESULTS

Changes in axonal excitability were found in HNPP subjects, but these were not uniform along the nerve: at the wrist there were prominent alterations in threshold electrotonus, whereas at the elbow there were only subtle alterations in the recovery cycle and the response to strong long-lasting hyperpolarisation. Threshold was raised at both sites, but the nerves were probably not hyperpolarised. Not unexpectedly, changes in CMT1A subjects were more marked than those in HNPP subjects and were uniform along the nerve.

CONCLUSIONS

Structural abnormalities at the node of Ranvier are sufficient to explain the changes in axonal excitability in HNPP, and these abnormalities would predispose the nerves to conduction block when subjected to pressure or stretch.

Region-specific sensitivity of the spinal cord to ischemia/reperfusion: the dynamic of changes in catalytic NOS activity

This study was designed in order to consider whether the release of neuronally derived nitric oxide (NO) in the lumbosacral spinal cord during ischemia/reperfusion is region-specific and whether changes in Ca(2+)-dependent NO synthase (cNOS) activity paralell with functional outcome. The cNOS activity was measured in the spinal cord regions after 13-, 15- and 17-min ischemia alone and that followed by 24 h of reperfusion. In addition, the Tarlov's criteria were applied to define the neurological consequences of ischemia/reperfusion in experimental animals. Based on the results, it is evident that only the 17-min ischemia alone was quite sufficient to cause changes in cNOS activity, however, without alterations in functional outcomes. On the other hand, the ischemic episodes followed by reperfusion caused dynamic, region-specific alterations in cNOS activity and consequently led to deterioration of motor function of hindlimbs in affected animals. Our results indicate that the motoneurons in the ventral horns respond more sensitively to ischemia/reperfusion than do neurons localized in the other spinal cord regions and that changes in cNOS activity may also influence the axonal conductance in the white matter and account for the impairment of motoneuronal activity in affected animals.

Traumatic injury of the spinal cord and nitric oxide

In the current report, we summarize our findings related to the involvement of nitric oxide (NO) in the pathology of spinal cord trauma. We initially studied the distribution of nitric oxide synthase (NOS)-immunolabeled and/or nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd; which is highly colocalized with NOS)-stained somata and fibers in the spinal cord of the rabbit. Segmental and laminar distribution of NADPHd-stained neurons in the rabbit revealed a large number of NADPHd-stained neurons in the spinal cord falling into six categories, N1-N6, while others could not be classified. Large numbers of NADPHd-stained neurons were identified in the superficial dorsal horn and around the central canal. Four morphologically distinct kinds of NADPHd-stained axons 2.5-3.5 microm in diameter were identified throughout the white matter in the spinal cord. Moreover, a massive occurrence of axonal NADPHd-staining was detected in the juxtagriseal layer of the ventral funiculus along the rostrocaudal axis. The prominent NADPHd-stained fiber bundles were identified in the mediobasal and central portion of the ventral funiculus. The sulcomarginal fasciculus was found in the basal and medial portion of the ventral funiculus in all cervical and thoracic segments. Since the discovery that NO may act as a neuronal transmitter, an increasing interest has focused on its ability to modulate synaptic function. NO passes through cell membranes without specific release or uptake mechanisms inducing changes in signal-related functions by several means. In particular, the activation of the soluble guanylyl cyclases (sGC), the formation of cyclic guanosine 3',5'-monophosphate (cGMP) and the action of cGMP-dependent protein kinases has been identified as the main signal transduction pathways of NO in the nervous system including spinal cord. It is known that the intracellular level of cGMP is strictly controlled by its rate of synthesis via guanylyl cyclases (GC) and/or by the rate of its degradation via 3',5'-cyclic nucleotide phosphodiesterases (PDE). GC can be divided into two main groups, i.e., the membrane-bound or particular guanylyl cyclase (pGC) and the cytosolic or sGC. In the spinal cord, the activation of pGC has only been demonstrated for natriuretic peptides, which stimulate cGMP accumulation in GABA-ergic structures in laminae I-III of the rat cervical spinal cord. These neurons are involved in controlling the action of the locomotor circuit. In view of the abundance of NO-responsive structures in the brain, it is proposed that NO-cGMP signaling will be part of neuronal information processing at many levels. In relation to this, we found that surgically induced Th7 constriction of 24 h duration stimulated both the constitutive NOS activity and cGMP level by 120 and 131%, respectively, in non-compartmentalized white matter of Th8-Th9 segments, located just caudally to the site of injury. NO-mediated cGMP formation was only slightly increased in the dorsal funiculus of Th5-Th9 segments. There are some other sources that may influence the NO-mediated cGMP formation in spinal cord. A high level of glutamate produced at the site of the lesion and an excessive accumulation of intracellular Ca2+ may stimulate NOS activity and create suitable conditions for NO synthesis and its adverse effect on white matter. An increased interest has focused on the role of NO at the site of injury and in areas located close to the epicenter of the impact site and, in these connections an upregulation of NOS was noted in neurons and interneurons. However, the upregulation of NOS expression was also seen in interneurons located just rostrally and caudally to the lesion. A quantitative analysis of laminar distribution of multiple cauda equina constriction (MCEC) induced NADPHd-stained neurons revealed a considerable increase in these neurons in laminae VIII-IX 8h postconstriction, and a highly statistically significant increase of such neurons in laminae VII-X 5 days postconstriction in the lumbosacral segments. Concurrently, the number of NADPHd-stained neurons on laminae I-II in LS segments was greatly reduced. It is concluded that a greater understanding of NO changes after spinal cord trauma is essential for the possibility of targeting this pathway therapeutically.

Glaucoma: thinking in new ways-a rôle for autonomous axonal self-destruction and other compartmentalised processes?

Glaucoma is a common neurodegenerative disease that affects retinal ganglion cells (RGCs). Substantial effort is being expended to determine how RGCs die in glaucoma. As in other neurodegenerative diseases, the majority of effort focuses on characterising apoptotic self-destruct pathways. However, apoptosis is not the only self-destruct mechanism that may be activated in neurons. It is now known that neurons have distinct classes of self-destruct programme that are spatially compartmentalised. In addition to the well-described intracellular suicide machinery in the neuronal soma, responsible for apoptosis, there is another, molecularly distinct, self-destruct programme localised in the axon. Evidence also supports the existence of compartmentalised degeneration programmes in synapses and dendrites. RGCs are no exception to this. Recent data, from in vitro studies and from an inherited mouse model of glaucoma, suggest that molecularly distinct degenerative pathways underlie the destruction of RGC somata and RGC axons. In various neurodegenerative diseases, axons, dendrites and synapses often degenerate well before the cells die, and there is increasing evidence that this is important for the production of clinical symptoms and signs. We hypothesise that such compartmentalised and autonomous programmes are of critical importance in the pathophysiology of glaucoma, and we suggest that studies of these processes are essential for a complete understanding of this complex disease.

Neurotrophin expression in the hippocampus and cerebellum is affected by chronic placental insufficiency in the late gestational ovine fetus

Our aim was to determine the effects of chronic placental insufficiency (CPI) during late gestation on the expression of neurotrophic factors and their receptors in the hippocampus and cerebellum in the near-term fetus. Structural alterations were also assessed in these brain regions. CPI was induced in eight fetal sheep by umbilicoplacental embolization (UPE) from 120 to 140 days of gestation (term approximately 147d) such that fetal arterial O2 saturation (SaO2) was maintained at approximately 50% of pre-UPE values. Five non-UPE fetuses served as controls. UPE resulted in fetal hypoxemia, hypoglycaemia, and growth restriction. In hippocampi from UPE fetuses, there were reductions in the optical density (OD) of the immunoreactivity (IR) of brain-derived neurotrophic factor (BDNF) protein within the mossy fibre collaterals of the polymorphic layer and in stratum lucidum (p<0.05); there was no consistent effect on tyrosine-related kinase (Trk) B receptor or neurotrophin-3 (NT-3) expression. Within the cerebellum, there was an increase in BDNF-IR (p<0.05) in the molecular layer; however, Trk B-IR and NT-3-IR were unaltered. There were no significant alterations to the structural parameters measured in the hippocampus. We conclude that CPI in late gestation affects the expression of BDNF in the fetal hippocampus and cerebellum, but these changes do not have a well-defined relationship to structural outcome.

Effect of hypoxia on calcium channels depends on extracellular calcium in CA1 hippocampal neurons

Our previous studies have shown that short lasting hypoxia induces an increase of Ca(2+) influx into the cell through high voltage-activated Ca(2+) channels in hippocampal neurons. This effect was abolished by removing of free Ca(2+) from intracellular solution. The aim of this study was to compare hypoxic responses at different extracellular Ca(2+) concentrations ([Ca(2+)](e)) in hippocampal neurons to ascertain whether the hypoxic sensitivity is restricted to Ca(2+) ions. Whole-cell patch-clamp recordings were made from acutely dissociated CA1 hippocampal neurons of rats. Polarographic method for measurements of O(2) partial pressure was used. Here we found that at 2 mM [Ca(2+)](e) the hypoxic effect was significant (up to approximately 94%), whereas [Ca(2+)](e) elevations to 5 and 15 mM resulted in gradual decreasing of the effect. We found, that total Ca(2+) charge carried into the cell under the hypoxia was similar at all [Ca(2+)](e), whereas Ca(2+) charge carried at normoxia was different for different [Ca(2+)](e), being larger at higher [Ca(2+)](e). These data indicated a saturation of the hypoxic effect due to limitation in the channel conductance. Therefore, we suggested that the hypoxic effect can be connected with increase of channel conductance, and the level of channel conductance at normoxia can determine the amplitude of hypoxic effect.

Resistance of isolated mammalian spinal cord white matter to oxygen-glucose deprivation

We found that isolated guinea pig spinal cord white matter is resistant to acute oxygen-glucose deprivation. Sixty minutes of oxygen-glucose deprivation resulted in a 60% reduction of compound action potential (CAP) conductance, and there was a near complete recovery after 60 min reperfusion. Corresponding horseradish peroxidase-exclusion assay showed little axonal membrane damage. To further deprive the axons of metabolic substrate, we added 2 mM sodium cyanide or 2 mM sodium azide, both mitochondrial suppressors, to the ischemic medium, which completely abolished CAP and resulted in a 15 to approximately 30% recovery postreperfusion. Both compounds preferentially reduced the conductance of large diameter axons. We suggest the residual ATP in our ischemic model can protect anatomic integrity and physiological functioning of spinal axons following ischemic insult. This further suggests that oxygen-glucose deprivation alone cannot be solely responsible for short-term functional and anatomic damage. The damaging effects of ischemia in vivo may be mediated by factors originating from the gray matter of the cord or other systemic factors; both were largely eliminated in our in vitro white matter preparation.

Spinal cord regeneration: from gene to transplants

The past 20 years has seen the emergence of many exciting and promising experimental therapeutic strategies to promote regeneration of the injured spinal cord in laboratory animals. A greater understanding of the pathophysiologic mechanisms that contribute to the initial and secondary cord injury may facilitate the development of neuroprotective strategies that preserve axonal function and prevent apoptotic cell death, thus optimizing neurologic function. Neurotrophic factors have been used to augment the poor intrinsic regenerative capacity of central nervous system neurons, and the need for sophisticated delivery of such trophic agents has stimulated the application of gene therapy in this context. In addition to augmenting the neuronal capacity to regenerate axons, many researchers are developing strategies to overcome the inhibitory environment into which these axons must grow. Characterizing the inhibitory elements of the glial scar at the site of injury and of myelin in the distal tracts is therefore a focus of intense scientific interest. To this effect, a number of strategies have also been developed to bridge the injury site and facilitate axonal growth across the lesion with a variety of cellular substrates. These include fetal tissue transplants, stem cells, Schwann cells, and olfactory ensheathing cells. With the collaboration of basic scientists and clinicians, it is hoped that these experimental strategies coupled with a greater understanding of the neurobiology of spinal cord injury will be translatable to the clinical setting in the near future.

Time course studies on the effectiveness of tetrodotoxin in reducing consequences of spinal cord contusion

Focal injection of the sodium channel blocker tetrodotoxin (TTX) into the injury site at either 5 or 15 min after a standardized thoracic contusion spinal cord injury (SCI) reduces white matter pathology and loss of axons in the first 24 hr after injury. Focal injection of TTX at 15 min after SCI also reduces chronic white matter loss and hindlimb functional deficits. We have now tested the hypothesis that the reduction in chronic deficits with TTX treatment is associated with long-term preservation of axons after SCI and compared both acute (24 hr) and chronic (6 weeks) effects of TTX administered at 15 min prior to and 5 min or 4 hr after SCI. Our results indicate a significant reduction of acute white matter pathology in rats treated with TTX at 15 min before and 5 min after injury but no effect when treatment was delayed until 4 hr after contusion. Compared with injury controls, groups treated with TTX at 5 min and 4 hr after injury did not show a significant deficit reduction, nor was there a significant sparing of white matter at 6 weeks compared with injury controls. In contrast, the group treated with TTX at 15 min before SCI demonstrated significantly reduced hindlimb functional deficits beginning at 1 week after injury and throughout the 6 weeks of the study. This was associated with a significantly higher axon density in the ventromedial white matter at 6 weeks. The results demonstrate that blockade of sodium channels preserves axons from loss after SCI and points to the importance of time of administration of such drugs for therapeutic effectiveness.

Platelet hyperactivity and abnormal Ca(2+) homeostasis in diabetes mellitus

We sought to determine the mechanisms for hyperactivity and abnormal platelet Ca(2+) homeostasis in diabetes. The glycosylated Hb (HbA(1c)) level was used as an index of glycemic control. Human platelets were loaded with Ca- green-fura red, and cytosolic Ca(2+) ([Ca(2+)](i)) and aggregation were simultaneously measured. In the first series of experiments, the platelets from diabetic and normal subjects were compared for the ability to release Ca(2+) or to promote Ca(2+) influx. A potent and relatively specific inhibitor of Na(+)/Ca(2+) exchange, 5-(4-chlorobenzyl)-2',4'-dimethylbenzamil (CB-DMB), increased the second phase of thrombin-induced Ca(2+) response, suggesting that the Na(+)/Ca(2+) exchanger works in the forward mode to mediate Ca(2+) efflux. In contrast, in the platelets from diabetics, CB-DMB decreased the Ca(2+) response, indicating that the Na(+)/Ca(2+) exchanger works in the reverse mode to mediate Ca(2+) influx. In the second series of experiments we evaluated the direct effect of hyperglycemia on platelets in vitro. We found that thrombin- and collagen-induced increases in [Ca(2+)](i) and aggregation were not acutely affected by high glucose concentrations of 45 mM. However, when the platelet-rich plasma was incubated with a high glucose concentration at 37 degrees C for 24 h, the second phase after thrombin activation was inhibited by CB-DMB. In addition, collagen-stimulated [Ca(2+)](i) response and aggregation were also increased. Thus in diabetes the direction and activity of the Na(+)/Ca(2+) exchanger is changed, which may be one of the mechanisms for the increased platelet [Ca(2+)](i) and hyperactivity. Prolonged hyperglycemia in vitro can induce similar changes, suggesting hyperglycemia per se may be the factor responsible for the platelet hyperactivity in diabetes.

Calcium imaging in live rat optic nerve myelinated axons in vitro using confocal laser microscopy

Intracellular Ca(2+) plays a major role in the physiological responses of excitable cells, and excessive accumulation of internal Ca(2+) is a key determinant of cell injury and death. Many studies have been carried out on the internal Ca(2+) dynamics in neurons. In constrast, there is virtually no such information for mammalian central myelinated axons, due in large part to technical difficulty with dye loading and imaging such fine myelinated structures. We developed a technique to allow imaging of ionized Ca(2+) in live rat optic nerve axons with simultaneous electrophysiological recording in vitro at 37 degrees C using confocal microscopy. The K(+) salt of the Ca(2+)-sensitive indicator Oregon Green 488 BAPTA-2 and the Ca(2+)-insensitive reference dye Sulforhodamine 101 were loaded together into rat optic nerves using a low-Ca(2+)/low-Na(+) solution. Axonal profiles, confirmed immunohistochemically by double staining with neurofilament-160 antibodies, were clearly visualized by S101 fluorescence up to 800 microm from the cut ends. The Ca(2+) signal was very low at rest, just above the background fluorescence intensity, indicating healthy tissue, and increased significantly after caffeine (20 mM) exposure designed to release internal Ca(2+) stores. The health of imaged regions was further confirmed by a virtual absence of spectrin breakdown, which is induced by calpain activation in damaged CNS tissue. Red and green fluorescence decayed to no less than 70% of control after 60 min of recording at 37 degrees C, with the green:red fluorescence ratio increasing slightly by 21% after 60 min. Electrophysiological responses recorded simultaneously with confocal images remained largely stable as well.

Involvement of the glutamate transporter and the sodium-calcium exchanger in the hypoxia-induced increase in intracellular Ca2+ in rat hippocampal slices

Hippocampal slices prepared from adult rats were loaded with fura-2 and the intracellular free Ca2+ concentration ([Ca2+]i) in the CA1 pyramidal cell layer was measured. Hypoxia (oxygen-glucose deprivation) elicited a gradual increase in [Ca2+]i in normal Krebs solution. At high extracellular sodium concentrations ([Na+]o), the hypoxia-induced response was attenuated. In contrast, hypoxia in low [Na+]o elicited a significantly enhanced response. This exaggerated response to hypoxia at a low [Na+]o was reversed by pre-incubation of the slice at a low [Na+]o prior to the hypoxic insult. The attenuation of the response to hypoxia by high [Na+]o was no longer observed in the presence of antagonist to glutamate transporter. However, antagonist to Na+-Ca2+ exchanger only slightly influenced the effects of high [Na+]o. These observations suggest that disturbance of the transmembrane gradient of Na+ concentrations is an important factor in hypoxia-induced neuronal damage and corroborates the participation of the glutamate transporter in hypoxia-induced neuronal injury. In addition, the excess release of glutamate during hypoxia is due to a reversal of Na+-dependent glutamate transporter rather than an exocytotic process.

Effects of chronic placental insufficiency on brain development in fetal sheep

Clinical evidence has linked intrauterine compromise such as fetal hypoxemia to poor neurologic outcome in the newborn. In this study we examined the effects of inducing chronic fetal hypoxemia by impairment of placental function on brain development in fetal sheep. Placental insufficiency was induced from 120 to 140 d of gestation (term = 145-148 d) by injection of microspheres into the umbilical circulation in five fetal sheep. Fetal partial pressure of oxygen, PaO2, was reduced from 24.1 +/- 0.5 mm Hg before embolization to 14.8 +/- 0.4 mm Hg after embolization (p < 0.05). In another three fetuses a similar level of hypoxemia (PaO2, 13.8 +/- 0.4 mm Hg) occurred spontaneously. At 140 d of gestation the fetal brains were perfused with fixatives and compared with five control fetuses for the assessment of structural and immunohistochemical alterations. Hypoxemic fetuses demonstrated severe gliosis in the cerebral cortex and reduced myelination of subcortical white matter as visualized by glial fibrillary acidic protein and myelin basic protein staining, respectively (p < 0.05). White matter lesions were observed in two fetuses. The diameter of cerebral capillaries was increased in hypoxemic fetuses (p < 0.05), but there was no change in the number of nitric oxide synthase immunoreactive cells. Growth of neuronal processes was affected in the cerebellum, where there was also a reduction in the number of Purkinje neurons (p < 0.05). These results show that a prolonged period of placental insufficiency, resulting in moderate fetal hypoxemia during the last third of gestation, can affect neurodevelopmental processes that occur late in gestation such as myelination and growth of the cerebellum. This prenatal damage could affect neural connectivity and have functional consequences after birth.