Age-Dependent Mechanisms of White Matter Injury After Stroke

Abstract

White matter is affected in human stroke and treatment strategies derived from experimental work on young adults focused exclusively on gray matter have been ineffective in the more clinically relevant aging population. The risk for stroke increases with age and the white matter also inherently becomes more vulnerable to injury with age. Age-related changes in the molecular architecture of white matter dictate the predominant injury mechanisms and determine the functional outcome. A prominent increase in the main Na⁺-dependent glutamate transporter, GLT-1, together with increased glutamate levels may reflect an increased need for glutamate signaling in the aging white matter to maintain function. Conversely, these adaptive modifications become detrimental under ischemic conditions, leading to increased vulnerability and robust injury. Mitochondria gather and elongate to effectively buffer Ca²⁺ but fail to produce sufficient ATP during an ischemic challenge to sustain aging axon function. Consequently, protective interventions in young white matter become injurious or ineffective to promote recovery in aging white matter after an ischemic episode. An age-specific understanding of the mechanisms of injury processes in white matter is essential to design dynamic therapeutic approaches for stroke victims.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The significance of white matter injury in the clinical manifestations of stroke has been underestimated in experimental animal models. Rodents are the most commonly used animals for the study of stroke but ironically rodent brain constitutes only ~10 % white matter by volume (Fig. 16.1; Zhang and Sejnowski 2000). In addition, the most commonly used stroke model in rodents, the middle cerebral artery occlusion (MCAO) model consistently spares corpus callosum, the white matter tract in rodents (as reviewed in Ginsberg and Busto 1989). This is mainly due to the fact that the middle cerebral is not the main arterial supply to the corpus callosum area. Rodent corpus callosum receives collaterals from deep penetrating pial arteries and striate arteries that arise from the circle of Willis. Conversely, human corpus callosum receives 80 % of its blood supply from the anterior cerebral artery and its branches, the middle cerebral artery being one of them (Wolfram-Gabel et al. 1987; Ture et al. 1996). Predictably, then the injured rodent brain after an ischemic attack is, essentially neuronal injury with minimal or no contribution from white matter. Human brain comprises equal percentages of gray and white matter by volume (Fig. 16.1; Zhang and Sejnowski 2000), which means that injuries sustained after a stroke in humans inevitably involve more white matter than in rodents. In fact, white matter is injured during most strokes, and even small lesions located in a strategic area of white matter can lead to drastic neurological dysfunction and deficits. It is conceivable that all these factors may have contributed to the failure of in clinical trial of drug candidates that selectively conferred protection to neurons in experimental stroke models (Del Zoppo 1995, 1998; Dirnagl et al. 1999; O’Collins et al. 2006). We suggest that failure to protect white matter is one of the principal reasons contributing to the lack of successful stroke therapy.

Fig. 16.1

Brain white matter volume expands as brain size enlarges. Histograms show the percentage of cerebral hemisphere volume composed of white matter in several mammals, ranging from mouse to human. (Figure is from Chap. 8 Molecular Pathophysiology of White Matter Anoxic Ischemic Injury by Ransom B, Goldberg MP and Baltan S; Stroke Edited by Mohr, Wolf, Grotta, Moskowitz, Mayberg, von Kummer, Elsevier 2011. Data calculated from Zhang K, Sejnowski TJ: A universal scaling law between gray matter and white matter of cerebral cortex. Proc Natl Sci U S A 97:5621–5626, 2000)

The risk of ischemic stroke increases drastically with age. A number of pathological changes occur in aging that preferentially influence brain structures and are candidates for causing age-associated neurological impairment, or increased susceptibility to impairment. The majority of studies of aging brain to date have focused on stereological estimates of total neuron numbers in gray matter. Interestingly, aging has global effects on white matter with relatively little effect on gray matter (Vincent et al. 1989; Peters and Sethares 1993). White matter appears to undergo generalized shrinkage in volume relative to gray matter (Albert 1993), due to the reduction in myelin content (Kemper 1994) and number of axons with age (Yamauchi et al. 2002). Myelin provides important insulating properties to axons allowing for propagation of action potentials over large distances at high velocity. Disruption of the myelin sheath could, therefore, contribute to cognitive impairment, as observed during aging (Peters and Sethares 1993; O’Sullivan et al. 2001; Ferro and Madureira 2002). Imaging studies in elderly people demonstrate hypodense lesions in white matter, termed leukoaraiosis (Hachinski et al. 1987). The severity of leukoaraiosis may predict high risk for stroke (Yamauchi et al. 2002) and impaired cognitive function (Schmidt et al. 1991). Other imaging techniques (Basser et al. 1994; Pendlebury et al. 1999) show regional variability in white matter signal characteristics, reflecting age-related degenerative effects on myelination and white matter connectivity (Davatzikos and Resnick 2002). However, the cellular basis of these changes cannot be determined from clinical imaging protocols.

White matter is a target of hypoxic-ischemic injury throughout life, in clinical settings ranging from periventricular leukomalacia in neonates, stroke, and cardiac arrest in adults, to dementia in the aging brain (Goldberg and Ransom 2003). Although there is increasing evidence that the degenerative effects of aging on white matter contribute to cognitive impairment, dementia, and increased risk for stroke (Bonita et al. 1994; Kurtzke 1994), structural and functional changes in aging white matter axons and glial cells have been explored only recently (Baltan et al. 2008; Baltan 2009, 2012). With an increase in the elderly population around the world, there is an urgency to view successful brain aging as the goal of scientists in the broad field of neurobiology. Identification of risk factors, awareness of the relative importance of each factor, and knowledge of their interaction should facilitate maintenance of cognitive, emotive, motor, and sensory functions. The focus of this chapter is to summarize the current information on molecular and functional changes in axons and glial cells as a function of age and how these adaptive mechanisms transform the tissue into a state of increased vulnerability in the face of an ischemic attack.

2 Intact Optic Nerve and Corpus Callosum Slices as Models to Study White Matter

The optic nerve, a purely myelinated central nervous system (CNS) white matter tract, is sensitive to the aging process (Cavallotti et al. 2002, 2003) and offers several advantages to study the mechanisms of white matter injury: (a) tissue isolation does not require extensive surgical interventions so there is minimal preparation injury, (b) the isolated optic nerve is structurally and functionally stable in vitro for at least 18 h, (c) there are no neurons, therefore no synapses or synaptic machinery to indirectly contribute to white matter injury, (d) axon function can be monitored by recording evoked compound action potentials (CAPs) (Fig. 16.2a), (e) the cellular components of optic nerve can be identified using immunohistochemistry, isoform-specific antibodies and confocal imaging (Fig. 16.2b), (f) amino acid release such as glutamate can be quantitatively monitored (Fig. 16.13), (g) protein levels of interest can be quantified by western blots and (h) intravitreal injections enable a route of delivery to axons.

Fig. 16.2

Monitoring white matter function and architecture. (a) Use of suction electrodes allows all axons to be stimulated and a compound action potential (CAP) to be recorded. (b) Using cell-specific antibodies, white matter axons labeled with SMI-31 for neurofilament (green), GFAP for astrocytes (magenta) and APC for mature oligodendrocyte cell bodies (green). Sytox (+) glial nuclei are in blue. Scale bar = 50 μm for SMI-31 and, GFAP, 10 μm for APC. (Reproduced in part from Baltan 2009)

Rat optic nerve (RON) has been widely used to study white matter injury to determine the effects of anoxia (Stys et al. 1990, 1992), aglycemia (Brown et al. 2001) and ischemia (Garthwaite et al. 1999). Mouse optic nerve (MON), however, has important advantages over RON as an in vitro model for studying white matter injury. The diameter of the adult RON is about twice that of the adult MON and metabolism is limited by inadequate glucose diffusion into the RON, especially when glucose utilization is increased as during anoxia (Baltan Tekkök et al. 2002; Tekkok et al. 2003). Taking into account this technical point, and considering the promise of future transgenic mice to analyze specific injury pathways, we use MONs for our experiments.

On the other hand, corpus callosum (CC) slices offer important advantages for the investigation of injury in white matter with neighboring gray matter. Slices preserve the anatomical and structural integrity of the continuity of the axon with the cell body, allowing assessment of axons with or without myelin in the presence of fully differentiated oligodendrocytes. Use of these two in vitro white matter tracts allow exceptional combined function–structure analysis, where glial cells and axons have retained their native relationships to one another within a three-dimensional organization (Fig. 16.2).

An in vivo model that selectively causes ischemic injury to white matter tracts is necessary to test some important parameters of stroke. Several different approaches have been taken to address this issue. For instance, severe applications of the MCAO model cause injury in subcortical white matter in rodents (Pantoni et al. 1996). However, subcortical white matter axons are extensions of cortical neuronal cell bodies (mainly layers 3, 5 and 6). Therefore, subcortical white matter injury in stroke is not a selective white matter injury but is secondary to the injury of cortical neuronal cell bodies. Consequently, neither the injury mechanisms nor the protective approaches can be specified to white matter structures. The modified Rice Vanucci model (carotid artery ligation + recovery under hypoxic conditions) is used to mimic periventricular leukomalacia and/or prenatal hypoxia, and selectively injures immature oligodendrocytes and arrests myelin production in postnatal day 7 rodent brain without any axon damage or motor deficits (Follett et al. 2000; Jensen 2006). Current approaches to corpus callosum injury involve local injections of glutamate analogues (McDonald et al. 1998; Leroux et al. 2010), demyelinating substances (Gadea and Lopez-Colome 2001), or vasoconstrictive substances (Sozmen et al. 2009) to cause glial and/or axon injury. These models do not mimic ischemia and also cause comparable tissue injury due to trauma and volume effects. On the other hand, our earlier efforts to induce ischemic injury in corpus callosum or optic nerve using piglets resulted in inconsistent findings, despite initial promising results (Lee et al. 2006). Primate brains contain 35 % white matter by volume and MCA perfusion territory is closer to human brain, and therefore may be the most appropriate model (Frykholm et al. 2000; Enblad et al. 2001).

3 Architectural Organization of White Matter

Structure and function in white matter are integrated: one is a reflection of the other (Fig. 16.2). The challenge, therefore, is to decipher the meaning of molecular architecture of white matter components. Discovery of the structural design of white matter is an evolving process and has so far proved to be more complex than anticipated. Axons and their myelinating oligodendrocytes, together with nurturing astrocytes (Fig. 16.2b, c), defensive microglia (Fig. 16.4) and progenitor cells forming synapses onto axons form an intricate interactive environment. The morphological variability in these white matter components among different white matter tracts further attest to their sophisticated nature to adapt to their regional function. In particular, glutamate homeostasis of adult white matter and intricate distribution of glutamate receptors and glutamate transporters expressed by astrocytes, oligodendrocytes, axons and microglia in a region-specific manner further illustrate the sophistication of white matter architecture (Figs. 16.7 and 16.8). White matter is particularly vulnerable to excitotoxicity during ischemia (Fig. 16.11) and a detailed knowledge of structural elements is key to prevent or restore function after ischemic injury.

In optic nerve from rat, mouse, or rabbit oligodendrocytes express AMPA receptors that are composed of GluR1, GluR3, and GluR4, but GluR3 and GluR4 in spinal cord (rat) and corpus callosum (mouse). On the other hand, oligodendrocytes from optic nerve are rich in GluR6, GluR7, KA1, and KA2 subunits of kainate receptors while in spinal cord GluR5, GluR6, KA1 form the subunit composition of kainate receptors. Interestingly corpus callosum oligodendrocytes do not express kainate receptors (Baltan and Goldberg, unpublished data). The AMPA and kainate receptors are mainly located on oligodendrocyte cell bodies and are believed to be involved in axon–myelin signaling.

Oligodendrocytes express NMDA receptors that consist of NR1, NR2A-NR2C, and NR3A subunits. Remarkably, NMDAR are expressed in clusters on oligodendrocyte processes, whereas AMPA and kainate receptors are diffusely located on oligodendrocyte soma. In addition, immunogold electron microscopy revealed that NR1, NR2, and NR3 subunits are present in the myelin sheath. Functional aspects of NMDA receptors on oligodendrocytes were discovered by application of several approaches. NMDA-mediated currents were recorded in mature oligodendrocytes from corpus callosum and cerebellar white matter (Karadottir et al. 2005), influx of Ca was identified in myelin formed by oligodendrocytes in the RON (Micu et al. 2006) and the protein levels of NMDA receptor subunits NR1, NR2A-D, and NR3A were detected in different type of oligodendrocytes (Karadottir et al. 2005; Salter and Fern 2005; Micu et al. 2006).

Glutamate is found in ample amount in white matter (Fig. 16.7) and is released during oxygen and glucose deprivation (OGD) (Fig. 16.13), causing overactivation of AMPA/kainate receptors (Tekkok et al. 2007). Potential sources of glutamate include axons (Li et al. 1999), oligodendrocytes (Fern and Moller 2000), and astrocytes via the reversal of Na⁺-dependent glutamate transporters (Li et al. 1999; Tekkok et al. 2007), Ca²⁺ increases (Parpura et al. 1994), or via release from swelling-activated anion channels and hemichannels (Ye et al. 2003). Vesicular glutamate release by vesicular glutamate transporter 1 (VGLUT1), 2 (VGLUT2) and 3 (VGLUT3) is Ca²⁺-dependent and modulated by stimulation frequency in myelinated (Kukley et al. 2007) and in unmyelinated axons (Ziskin et al. 2007) of the corpus callosum. However Na⁺-dependent glutamate transport inhibitors greatly diminish glutamate release in MONs during ischemia (Tekkok et al. 2007). Astrocytes express the greatest density of Na⁺-dependent glutamate transporters in white matter; glutamate transporter-1 (GLT-1) and glutamate aspartate transporter (GLAST) (Figs. 16.7 and 16.8). Energy deprivation causes a gradual dissipation of the transmembrane Na⁺ gradient in astrocytes setting up conditions for reverse exchange and glutamate release (Longuemare et al. 1999).

4 White Matter Components Reorganize with Aging

Across the age groups, there is no change in the proportion of total APC (+) oligodendrocytes, GFAP (+) astrocytes, NG2 (+) precursor cells (Fig. 16.3). Interestingly, microglial processes and the thickness and the length of these processes (ramification) increase three to fivefold in older white matter (Fig. 16.4), in agreement with age-related activation of white matter microglia in monkey (Sloane et al. 1999) and rats (Ogura et al. 1994; Morgan et al. 2004). Additionally, each microglial cell with its processes appears to occupy a distinct domain within the tissue, reminiscent of astrocytic domains (Oberheim et al. 2008) in young white matter. This territorial organization is lost with age as numerous processes reached out towards neighboring domains. Microglial activation is a pathological hallmark of stroke. Curiously, aging seems to unmask this capacity of microglia without any further insult. It is unclear whether the effect of activated microglia on white matter is beneficial, detrimental, or a combination of both.

Fig. 16.3

Aging results in loss of cellularity but relative percentage of glial cells is preserved. (a–c) Pie charts summarize percentile of glial cells (percent of total nuclei) in white matter from 1-, 12-, and 18-month-old optic nerve. (d) Histograms demonstrate ~35 % fewer glial cells by 12 months of age. ***p < 0.001, one-way ANOVA

Fig. 16.4

Aging causes microglial activation. A few Iba (+) microglial cells (red) with their small cell bodies and few thin processes can be detected in 1-month-old MONs (left). Aging causes elaborate changes in their morphology including cell body size, and process ramification (right). Calibration bar = 20 μm

Astrocytes exhibit thicker processes that change their orientation from transverse to a more longitudinal orientation with aging (Baltan et al. 2008, also see Fig. 16.8). However, the GFAP protein levels or the number of astrocytes do not change in older white matter (Baltan et al. 2008). Because astrocyte processes form the end-feet on the vascular wall and contribute to the neurovascular unit, the change in their direction raises the possibility that they may be adapting to vascular changes with aging. In addition, astrocyte processes elongate to the nodes of Ranvier and to sense the metabolic demands of axon function. Age-dependent reorganization of axonal architecture (nodal, paranodal, myelin content-see Figs. 16.5 and 16.6) may need expanded points of surveillance by astrocytes to maintain the metabolic burdens of aging axons.

Fig. 16.5

MBP and MAG are down regulated with aging. Quantification of labeling intensity reveals that MBP and MAG decrease with age compared to 1 month old mice. *p < 0.05, **p < 0.01, ***p < 0.001, one-way ANOVA

Fig. 16.6

Nodal and paranodal structures reorganize with aging. Immunolabeling of Nav1.6 sodium channels at the nodes of Ranvier show that sodium channel clusters (red) become smaller and assume an extranodal localization (line with arrows) in aging MONs. Paranodal CASPR (green) elongates (brackets) and becomes thinner, as summarized in the histograms. *p = 0.0317, **p = 0.0019, two-tailed Student’s t-test. Calibration bar = 2 μM

Myelin provides a unique architecture to allow high-fidelity conduction along axons. Labeling studies for different myelin proteins reveal a significant loss of myelin basic protein (MBP) and myelin associated glycoprotein (MAG) as a function of age. On the other hand, and consistent with preserved oligodendrocyte numbers, CNPase levels show a small drop (Fig. 16.5). Oligodendrocytes and possibly myelin express AMPA/kainate and NMDA receptors (Micu et al. 2006; Baltan et al. 2008) which mediate excitotoxic injury during ischemia. Age-dependent changes in glutamate receptor type and subunit composition is of utmost interest, for they dictate the ensuing injury.

Clustering of Nav1.6, the predominant sodium channel at the nodes of Ranvier (Goldin et al. 2000) is an indicator of proper axo–glial contact dictated by myelin (Boiko et al. 2001) to support saltatory conduction. The contactin-associated protein (CASPR) rich paranodes are flanked by juxtaparanodal Kv1.1 or Kv1.2 potassium channels. Immunohistochemical analysis for CASPR and Nav1.6 indicate that CASPR (Fig. 16.6, green) is longer and thinner, while Nav1.6 clusters (Fig. 16.6, red) become shorter with aging. Although Nav1.6 is also found in unmyelinated axons (Black et al. 1999), it is not detectable in the internodal region of myelinated axons. However, aging causes ectopic regions of Nav1.6 immunoreactivity (Fig. 16.6, white arrows) which might lead to deleterious effects on axons especially if colocalized with the Na⁺–Ca²⁺ exchanger (Craner et al. 2004). A majority of axon profiles with diffuse Na⁺ channel immunoreactivity is associated with reduced MBP immunolabeling in spinal cord experimental allergic encephalomyelitis (Craner et al. 2004) which paralleled structural changes in aging axons in optic nerve. In addition, Kv1.2 (+) potassium channels overlap with CASPR immunostaining indicating loss of the characteristic demarcation between paranodal and juxtaparanodal structures (Dupree et al. 1999; Bhat et al. 2001). Displacement or aberrant localization of nodal, paranodal, juxtaparanodal structures suggest that axonal function at the node may be compromised with age. Similar findings in the aging monkey and RON (e.g., Hinman et al. 2006) implicate age-related molecular reorganization at the nodes of Ranvier as intrinsic to white matter which crosses over species.

White matter glutamate homeostasis and the related machinery also go through a series of age-related restructuring (Figs. 16.7 and 16.8). The dominant glutamate transporter, GLT1, plays an essential role in removing glutamate from the extracellular space and maintaining glutamate below neurotoxic levels under normoxic conditions (Rothstein et al. 1996; Hazell et al. 2001). Although these transporters are predominantly expressed on astrocytes in young white matter, they extend to additional structures with aging (Fig. 16.8), implying that additional white matter constituents may contribute to toxic glutamate accumulation in aging white matter. In addition to GLT-1, the essential members to maintain glutamate homeostasis are GLAST, glutamate and glutamate synthatase (GS). White matter glutamate content increases considerably and in correlation with increased GS levels with age (Fig. 16.8). Together with a twofold increase in GLT-1 levels in older white matter (Baltan et al. 2008), these adjustments may infer an age-related adaptive mechanism in white matter to remove and to convert excessive glutamate to glutamine so as to maintain glutamate homeostasis. As a result glutamate levels under control conditions (Fig. 16.13, Baltan et al. 2008) and axon conduction across aging axons remains impressively stable over time under normoxic conditions. The number of GLT-1 transporters determines the capacity of the tissue to move glutamate between internal and external compartments (Fig. 16.13). However, it is the direction of the pump that acts to save or injure the tissue. During ischemia, these measures act against the tissue due to an accelerated Na⁺ overload as a result of decreased tolerance to energy deprivation in aging white matter. Therefore, numerous GLT-1 transporters reverse and lead to early robust release of glutamate causing enhanced excitotoxicity (Fig. 16.13). Moreover, in young white matter glutamate resumes to baseline levels after the end of OGD, which is a sign of efficient astrocyte uptake of glutamate. However, glutamate levels remain elevated in old white matter, suggesting that aging astrocytes cannot take up excess glutamate and thereby expanding the excitotoxicity duration into the recovery period (Fig. 16.13). Despite the possibility that glutamate may be released from multiple sources in aging white matter, astrocytes are expected to remove and store glutamate efficiently. Therefore, these results suggest a prominent change in aging astrocyte capacity to remove glutamate.

Fig. 16.7

Glutamate and glutamate synthatase (GS) expression increase in 12-month-old MONs. (a) Immunolabeling and (b) quantification of glutamate, GS and GLAST immunolabeling revealed that glutamate and GS labeling intensity increases by 156 ± 11.1 % and 198 ± 21.9 %, respectively, in 12 month old MONs. Note that GLAST labeling intensity decreased to 68.9 ± 11.2 % in 12 month old MONs. *p = 0.0278, **p = 0.004, ***p = 0.0006, two-tailed Student’s t-test

Fig. 16.8

Aging is correlated with upregulation of GLT-1. The overlap of GLT1 (red) and GFAP (green) labeling indicated that GLT-1 was mainly expressed in astrocytes (merged) in 1-month-old MONs. MONs. There was a twofold increase in GLT-1 pixel intensity (188.6 ± 18.5 %) with age. The pattern of GFAP expression in MONs changed with age but without an increase in GFAP pixel intensity. (Reproduced from Baltan et al. 2008)

On the other hand, GLAST expression in white matter falls with age, raising the possibility that, glutamate transporters can functionally substitute for one another with age (Fig. 16.7). It is well-known that the upregulation of GLT-1 participates in the induction of brain ischemic tolerance in gray matter (Romera et al. 2004; Kawahara et al. 2005; Zhang et al. 2007) and in reactive astrocytes in human periventricular leukomalacia (Desilva et al. 2008). Aging seems to saturate these mechanisms (Fig. 16.8), and it is intriguing whether preconditioning is impaired in white matter with aging. Unfortunately, these questions will remain unanswered until an in vivo model of adult white matter stroke is available.

5 Age-Dependent Mechanisms of Ischemic White Matter Injury

White matter axons are dependent on a constant supply of oxygen and glucose to faithfully transmit signals. We reported that CNS white matter function is exceptionally tolerant to a complete lack of oxygen (Tekkok et al. 2003) while there is regional variability in the ability to function and survive anoxia (Baltan 2006). However, young adult white matter is readily susceptible to ischemia induced by combined OGD (Fig. 16.9). Mechanisms underlying ischemic white matter injury prove to be unpredictably complex (Fig. 16.10) (Wrathall et al. 1994; Agrawal and Fehlings 1997; Fern and Ransom 1997; McDonald et al. 1998; Sanchez-Gomez and Matute 1999; Follett et al. 2000; Tekkök and Goldberg 2001; Stys 2004; Tekkok et al. 2007). White matter contains no neuronal soma but instead has myelinated axons, oligodendrocytes, and astrocytes. The cellular elements of white matter are individually under attack during ischemia, while they still remain interactive with each other in intricate mechanisms that are not well understood. Axons are injured (Fig. 16.10) directly by ionic mechanisms, resulting in accumulation of intracellular Na⁺ and Ca²⁺ (Stys et al. 1990; Fern et al. 1995; Wolf et al. 2001; Ouardouz et al. 2003; Underhill and Goldberg 2007) while astrocytes, via the reversal of Na⁺-dependent glutamate transporters (Li et al. 1999; Tekkok et al. 2007) initiate an excitotoxic path resulting in oligodendrocyte death and myelin damage (Fig. 16.11, Tekkök and Goldberg 2001; Tekkok et al. 2007). Glutamate accumulation concomitantly initiates the oxidative pathway attacking white matter constituents due to formation of free radicals mediated by glutamate competing with cysteine at the glutamate-cysteine pump (Oka et al. 1993) and glutamate disrupting mitochondrial function (Chang and Reynolds 2006). Glutamate is necessary but not sufficient to explain ischemic injury in white matter. Consistent with this, exogenous application of glutamate or its agonists fails to initiate injury. Only a short and reversible OGD (15 min) combined with glutamate (or analogues) mimic ischemic injury (Tekkok et al. 2007). In accordance, glutamate levels remain very stable during the initial stage of OGD (25–30 min) and then steadily start to accumulate (Fig. 16.12). These results point to a sequential order of injury pathways converging in order for irreversible injury to ensue, such that an essential first stage of ionic dysfunction primes white matter for glutamate toxicity (Fig. 16.9; Tekkok et al. 2007). The sequential convergence of these pathways also manifests itself as duration-dependent injury (Tekkok et al. 2007; Baltan 2009). Predictably, removal of extracellular Ca²⁺, blockade of AMPA/kainate receptors (Fig. 16.11), blockade of reverse glutamate transport (Tekkok et al. 2007), or prevention of reactive oxygen species (ROS) generation reduces ischemic white matter injury. On the other hand, although developing oligodendrocyte processes (Salter and Fern 2005), mature oligodendrocyte cell bodies (Karadottir et al. 2005; Baltan et al. 2008), and myelin (Micu et al. 2006) express functional NMDARs, blockade of these receptors does not improve axon function after ischemia in young optic nerve (Tekkok et al. 2007) or corpus callosum (Tekkök and Goldberg 2001). These results do not negate the activation of NMDAR during ischemia but imply that their activation does not contribute to axonal damage and raises a caution for clinical implications of NMDAR antagonism during ischemia, particularly in aging white matter (see below). In fact, oligodendroglial NMDAR activation preserves axon function against ischemia by coupling axonal and glial energy metabolisms in developing and young optic nerves (Saab et al. 2012).

Fig. 16.9

Putative mechanisms of ischemic white matter injury. Ionic, excitotoxic, and oxidative stress converge in sequential order to cause irreversible injury in white matter during ischemia. Note that glutamate release, due to reverse Na⁺-dependent transport dictates irreversible nature of the injury. (Reproduced from Baltan 2009)

Fig. 16.10

White matter is susceptible to ischemic injury. Axon function is quantified as the area under the CAP, normalized to control, and plotted against time. Under normal conditions CAP area remains stable over time (brown). A 60 min period of oxygen glucose deprivation (OGD) depresses the CAP gradually until conduction along the axons is completely lost (gray) typically around 30 min. Restoring oxygen and glucose, axon function recovers to ~25 %. Sample traces from control (a), OGD (b) and recovery (c) periods are shown above the plot

Fig. 16.11

Overactivation of AMPA/kainate receptors cause oligodendrocyte death and axon disruption. Under normoxic condition, DAPI (+) glial nuclei are mostly APC (+) oligodendrocytes among SMI-31 (+) labeled axons. A 30 min period of OGD in corpus callosum slices causes wide-spread loss of APC (+) oligodendrocytes and SMI-31 (+) axons with pyknotic bright DAPI (+) dead nuclei. Blockade of AMPA/kainate receptors with NBQX prevents oligodendrocyte death and axon disruption. Scale bar = 10 μm. (Reproduced from Tekkök and Goldberg 2001)

Fig. 16.12

Aging axons are more vulnerable to anoxia. Normalized data show that the CAP area rapidly falls and recovers less after 60 min of anoxia in 12-month-old (red) compared to 1-month-old (black) MONs. The CAP area and SEM are plotted every 3 min. **p < 0.009, one-way ANOVA

Although valid mechanisms, all of these studies were performed entirely on young animals. Because the risk for stroke increases with age, a thorough understanding of white matter ischemic injury in age-appropriate populations is of central importance to meet the challenge of developing effective stroke therapy. Indeed, CNS white matter becomes intrinsically more vulnerable to OGD in older animals and the mechanisms of white matter injury change as a function of age (Baltan et al. 2008). This increased sensitivity to OGD is due, in part, to increased susceptibility of aging axons to a lack of oxygen (Fig. 16.12). Mitochondrial dysfunction and disruption with age presumably underlie the loss of aerobic capacity of aging axons (Fig. 16.13, see mitochondria in insets). On the other hand, whether aging axons are less tolerant to removal of glucose is under investigation.

Fig. 16.13

Enhanced excitotoxicity due to impaired mitochondrial function leads to early and robust glutamate release in aging white matter. The principle Na⁺-dependent glutamate transporter, GLT-1, typically takes up glutamate with co-transport of Na⁺. During ATP depletion, due to increased intracellular Na⁺ levels, the transporter reverses and releases glutamate. Therefore the number of transporters determines the capacity of the system for the amount of glutamate that can be transported, but it is the ATP levels that determine the direction of the transporter (to remove or release glutamate). Mitochondria in aging axons become longer and thicker compared to young axons, which may hinder ATP production and drive GLT-1 in reverse mode. Consistent with this, there is an early and robust glutamate release in aging white matter (blue) compared to optic nerves from young mice (gray). Note, that the glutamate levels return to baseline in young but remain elevated in aging white matter (red arrows). (Reproduced in part from Baltan et al. 2008)

Accumulation of Ca²⁺ is an important step in the development of ischemic injury in young white matter, while Ca²⁺, influx may be more vital to aging axons. The lack of protection by removal of extracellular Ca²⁺ or by blockade of Ca²⁺ entry secondary to reverse operation of the Na/Ca exchanger (NCX), denotes that preventing Ca²⁺ influx is not sufficient to preserve older axon function. Ironically, removal of extracellular Ca²⁺ worsens axon function recovery after OGD, a perplexing observation that implies Ca²⁺ entry during ischemia is a protective measure. It is possible that Ca²⁺ release from intracellular Ca²⁺ stores (ICS), and the interplay between intracellular and extracellular Ca²⁺, is more critical during ischemia in aging axons. Although a role for Ca²⁺ release from endoplasmic reticulum via IP3 and ryanodine receptors is described in young white matter (Thorell et al. 2002), various Ca²⁺-dependent neurophysiological (Landfield and Pitler 1984; Campbell et al. 1996) and signaling pathways (Verkhratsky et al. 1998) remain unexplored in white matter, particularly as these relate to ischemia and aging.

Because blockade of NMDARs cause a similar worsening of axon function recovery, it points to Ca²⁺ entry through NMDARs as an important element to protect aging axon function against ischemia. Alternatively the role of NMDAR as the axon–glia metabolic couplers becomes more critical to support the increased energy burden of aging axons. It is also plausible that removal of Ca²⁺ may lead to glutamate release via hemichannels from aging astrocytes thus exacerbating excitotoxic injury. These mechanisms remain to be explored.

The Ca²⁺-independent nature of ischemic injury in aging white matter raises another important question as to how AMPA/kainate receptor activation mediates injury. Because blockade of AMPA/kainate receptors promotes axon function recovery across all age groups, it suggests that Na⁺ entry, via AMPA/kainate receptors, rather than Ca²⁺, mediates injury during ischemia in older white matter. Overload of Na⁺ exhibits irreversible toxic swelling, even in the absence of extracellular Ca²⁺ (Rothman and Olney 1995) and a rise in intracellular Na⁺ promotes reversal of the Na⁺-dependent glutamate transporter, resulting in glutamate accumulation (Szatkowski et al. 1990). Together with the upregulation in GLT-1 expression, increases in intracellular Na⁺ may be the leading cause of increased and early release of glutamate, overactivating AMPA/kainate receptors and creating a vicious cycle that underlies the vulnerability of aging white matter to ischemia (Fig. 16.13). Furthermore, an increase in Na⁺ concentration interferes with maintenance of the transmembrane ion gradient. This challenges the Na–K ATPase pump, compromises the ability of aging axons to maintain membrane properties and axonal excitability and contributes to an increased vulnerability to ischemia (Fig. 16.17, Scavone et al. 2005).

The finding that blockade of NMDARs worsens outcome in older white matter (Baltan et al. 2008) has important clinical implications as it explains why NMDAR antagonists, that conferred protection to neuronal cell bodies in experimental stroke settings, failed to provide any benefit in clinical stroke trials. In this aspect, these results challenge the existing convention in stroke research that assumes a common mechanism of injury in the brain across the life span. They also question the well-established agreement that glutamate receptor over-activation unequivocally causes injury. It is critical to consider that therapeutic options beneficial for white matter tracts may be ineffective for the gray matter and that therapeutic options proving successful in gray matter may be less useful or even harmful for white matter, especially when mechanisms of injury change as a function of age.

6 Mitochondrial Dysfunction Underlies Vulnerability of Aging Axons to Ischemia

The bioenergetics of mitochondria in neurons and their role in glutamate excitotoxicity are well described in gray matter (Nicholls et al. 2007). Mitochondrial dysfunction and excitotoxicity share common aspects and are believed to act synergistically by potentiating each other (Albin and Greenamyre 1992; Jacquard et al. 2006; Silva-Adaya et al. 2008). Mitochondria are dynamic organelles that travel along microtubules, using axonal transport to reach peripheral locations (Hollenbeck 2005; Hollenbeck and Saxton 2005) (Fig. 16.13). They constantly undergo fission and fusion events (Karbowski et al. 2004), and the relative rates of mitochondrial fusion and fission have been implicated in the regulation of their number, size and shape (Mozdy and Shaw 2003; Scott et al. 2003; Chen et al. 2007). The balanced delivery of mitochondria to cell body, dendrites and axons helps them serve multiple functions, including energy generation, regulation of Ca²⁺ homeostasis, cell death, synaptic transmission and plasticity (Chang and Reynolds 2006). Expectedly, an exciting link between a variety of neurological diseases, as well as aging, and defects in mitochondrial fusion and distribution is emerging (Karbowski and Youle 2003; Chen et al. 2007). Older neurons become more susceptible to glutamate excitotoxicity due to loss of mitochondrial membrane depolarization and increased ROS generation leading to reduced energy supply (Parihar and Brewer 2007). Mitochondria exhibit a cell-specific morphology; neuronal mitochondria are small and round as opposed to the longer tubular mitochondria in white matter axons (Fig. 16.14). Dynamin-related protein 1 (Drp-1) is known to affect the distribution of mitochondria (Otsuga et al. 1998; Smirnova et al. 1998), and Drp-1 protein levels show cell-specific variation, being expressed more in astrocytes compared to neurons (Uo et al. 2009). Mitochondrial function appears to decline in older animals, presumably causing reduced ATP production. This has been demonstrated in cardiac (Lesnefsky et al. 2001), liver (Selzner et al. 2007), and brain (Toescu 2005). Ion transport accounts for about 50 % of all ATP utilization and Na⁺/K⁺ ATPase activity alone is responsible for the majority of this consumption (Erecinska and Silver 1994). A loss of ATP reserve, diminishing the activity of this key enzyme with advanced age, is a plausible contributor to heightened white matter injury susceptibility (Scavone et al. 2005). Consistent with this axon function, when transiently challenged with OGD, was slower to restore normal ion gradients, permitting pathological processes related to ion derangement to operate for longer periods (hence reversing Na⁺-dependent glutamate transporter(s) earlier), and producing more injury in older MONs (Baltan et al. 2008). The disadvantage of compromised ATP levels in older animals was further verified by better recovery of white matter function in older animals when OGD was imposed at lower temperature (Baltan et al. 2008). Axons with high ATP requirements have many more mitochondria per unit length of process (Bristow et al. 2002); therefore, these axons would be preferentially targeted by low ATP conditions.

Fig. 16.14

CFP (+) somal and axonal mitochondria exhibit region-specific morphology. (a) Neuronal mitochondria are observed as CFP (+) structures in Thy-1 mito mice (Misgeld et al. 2007). CFP (+) mitochondria are small and round in neuronal cells bodies (yellow asterisk) labeled with MAP2 (red) but more linear and tubular in primary dendrites (yellow arrows). (b) Numerous long tubular mitochondria (yellow arrows) are evident in young corpus callosum. Note, the smaller round mitochondria in GM. Scale bar = 5 μm

Excitotoxicity and elevated Ca²⁺ induce marked changes in mitochondrial morphology, arresting their motion (Rintoul et al. 2003; Barsoum et al. 2006; Chang and Reynolds 2006) and generating ROS in neurons (Nicholls et al. for review). In young white matter, activation of either AMPA or kainate receptors (Baltan et al. 2008) loads mitochondria with Ca²⁺ and fission is enhanced, associated with loss of fluorescence of mitochondria genetically tagged with CFP (Fig. 16.15a; Misgeld et al. 2007). A Ca²⁺ overload activates n-NOS to produce nitric oxide (NO) and ROS, which are proposed as diffusible second messengers linking oligodendrocyte excitotoxicity to axon injury (Matute et al. 2001; Ouardouz et al. 2006). Axon function directly correlates with tissue energy reserves, since Na⁺–K⁺ ATPase activity is intimately dependent on ATP levels. As a result, OGD causes a significant reduction in ATP levels and CFP (+) mitochondria, which could be prevented by AMPA/kainate receptor blockade (Fig. 16.14).

Fig. 16.15

Blockade of excitotoxicity preserves CFP (+) axonal mitochondria and ATP levels in response to OGD. (a) OGD drastically reduced CFP fluorescence in MONs from mito CFP (+) mice and pretreatment with NBQX (30 μM) protected against this loss. Note the change in mitochondrial morphology from small and tubular to tiny and punctuated form with OGD. Scale bar = 10 μm (insets = 2 μM) (b) Consistent with the preservation of CFP pixel intensity, NBQX pretreatment conserved ATP levels in MONs. ***p < 0.0001, one-way ANOVA. (Reproduced in part from Baltan et al. 2011b)

Aging stimulates mitochondrial fusion (Fig. 16.13) and this may be accompanied by a reduction in Drp-1 levels. The regulated process of mitochondrial fusion and fission controls the spatiotemporal properties of mitochondrial Ca²⁺ responses and the physiological and pathophysiological consequences of Ca²⁺ signals (Szabadkai and Rizzuto 2004; Szabadkai et al. 2004). By enhancing fusion or inhibiting fission, elongated mitochondria possibly absorbs Ca²⁺ efficiently preventing n-NOS activation and subsequent ROS production (Cheung et al. 2007, Fig. 16.17). However, this age-related adaptive reorganization of mitochondria becomes detrimental under ischemic conditions. Ischemia, in aging white matter, further enforces mitochondrial fusion as a result of the age-dependent drop in Drp-1 and age-dependent loss of mitochondrial motility with exposure to glutamate (Chang and Reynolds; Parihar and Brewer). Mitochondria fuse to collectively counteract the already increased excitotoxicity and Ca²⁺ load with aging, and this age-related change in mitochondrial dynamics could hinder ATP production. Because basal ROS generation is already elevated with aging, further increases in ROS accumulation under ischemic conditions result in increased vulnerability to ischemia (Fig. 16.17).

These results were further verified in a series of experiments investigating the protective effects of Class I HDAC inhibitors in young white matter (Baltan et al. 2011a, b). These HDAC inhibitors promoted functional recovery of axons and preserved white matter cellular architecture. This protection correlated with the upregulation of an astrocyte glutamate transporter, delayed and reduced glutamate accumulation during OGD, preservation of axonal mitochondria and oligodendrocytes, and maintained ATP levels in young optic nerves (Fig. 16.16) verifying the proof of principle that excitotoxic injury leads to mitochondrial dysfunction in white matter axons.

Fig. 16.16

The HDAC inhibitor MS-275 preserves CFP (+) axonal mitochondria and ATP levels in response to OGD. (a) OGD (blue) drastically reduced CFP fluorescence in MONs from mitoCFP (+) mice and pretreatment with MS-275 (1 μM) protected against this loss. (b) Consistent with the preservation of CFP pixel intensity, MS-275 pretreatment promoted axon function recovery (green) and conserved ATP levels in MONs. ***p < 0.0001, one-way ANOVA. Scale bar = 10 μm. (Reproduced in part from Baltan et al. 2011b)

7 Region-Specific Mechanisms of Ischemic WM Injury

There is a growing sense that the mechanisms of white matter injury vary from one area of the brain to another (Tekkok et al. 2007). The explanations for regional differences in white matter injury are not yet understood at a cellular level. Regional differences in white matter oligodendrocytes and/or axons are logical possibilities but differences in astrocytes, cells rich in glutamate, should also be considered. Moreover, there may be regional differences in the glutamate receptors that participate in the injury process (Gallo and Russell 1995; Brand-Schieber and Werner 2003a, b; Tekkok et al. 2007). These differences in glutamate receptor pharmacology of white matter injury may reflect regional differences in the receptors themselves (e.g., degree of expression or subunit composition), regional variability in the amount or onset of glutamate release during ischemia, or the existence of regionally diverse oligodendrocytes. The latter point is particularly noteworthy, given that the ratio of myelinating to non-myelinating oligodendrocytes vary between areas where all axons are myelinated, like the optic nerve (Foster et al. 1982), compared with areas containing many non-myelinated axons such as the corpus callosum (Olivares et al. 2001). These molecular properties indeed determine the mechanism of injury and the set of receptors that mediate that injury. For instance, activation of either AMPA or kainate receptors in MONs is sufficient to cause injury (Tekkok et al. 2007) while activation of Ca²⁺-permeable AMPA receptors exclusively mediate the ischemic injury in corpus callosum slices (Tekkök and Goldberg 2001). Moreover, axon function is significantly more resilient to ischemia in optic nerve compared to corpus callosum while corpus callosum axons are more tolerant to anoxia compared to optic nerve. These findings point to a curious divergence between anoxic and ischemic injury mechanisms in white matter. Ischemia causes progressive injury as a function of glutamate accumulation. Anoxia does not cause glutamate release, therefore no region-specific glutamate receptors are involved in the injury process, implying a rather stagnant course for anoxic injury. It is conceivable that axons use energy from glycolysis to prevent Na+/K + pump failure during anoxia to suppress subsequent membrane depolarization and limit the rise in intracellular Ca²⁺ levels. Since more of the smaller diameter unmyelinated axons of corpus callosum survive anoxia, myelin may be the structural element underlying the vulnerability to anoxia, inferring that unmyelinated axons and/or the smallest axons with the thinnest myelin sheath are the resistant group (Baltan Tekkök and Ransom 2004). These results reveal that ischemia and anoxia are not interchangeable forms of injury in white matter and that CNS white matter is remarkably tolerant of anoxia although there is regional variability in their ability to function or survive (Baltan 2006).

8 Conclusions

The main goal of this review is to establish the proof of principle that CNS white matter becomes inherently more susceptible to an ischemic attack with age and that the molecular and cellular mechanisms of ischemic injury change as a function of age (Fig. 16.17). Predictably, age-related changes in the molecular architecture of white matter dictate the predominant injury mechanisms and determine the functional outcome. Consequently, protective interventions in young white matter such as removal of extracellular Ca²⁺ (Fig. 16.17a, black arrow), reduce functional recovery in aging axons (Fig. 16.17a, black dotted arrows). Together with the observation that blockade of reverse NCX fails to protect function in older mice (Fig. 16.17b, Baltan et al. 2008), these results propose a diminished role for the ionic pathway with aging. On the other hand, aging causes a prominent increase in the expression pattern of glutamate, GS and GLT-1 levels that extend to additional structures in white matter. These modifications may imply an age-related adaptive mechanism to maintain glutamate signaling and homeostasis. However, during an ischemic episode these adaptive changes act against the tissue and expedite and aggravate glutamate release (Fig. 16.17a, red dotted arrow) and expand the excitotoxic injury into recovery period. Interestingly, AMPA/kainate receptors (Fig. 16.17b, white triangles) mediate the ischemic injury across age groups (Fig. 16.17a, purple arrows) indicating that certain steps of injury are preserved irrespective of age. Surprisingly, activation of NMDARs turns into an essential mode of protection for aging axons such that blockade of NMDARs impedes functional recovery after ischemia (Fig. 16.17a, green dotted arrow). NMDARs are expressed in oligodendrocyte cell bodies in young white matter but expand to myelin and to myelin processes with age (Fig. 16.17b, yellow squares; Baltan et al. 2008). Whether this modification of NMDA receptor expression enhances axon–glia metabolic coupling or mediates efficient Ca²⁺ influx to myelin (Micu et al. 2006) to promote aging axon function recovery remains to be explored.

Fig. 16.17

Putative molecular and cellular mechanisms responsible for ischemic white matter injury. (a) Protective intervention in young white matter may become injurious or remain the same or become enhanced in aging white matter. (b) Age-related cellular reorganization of white matter components determines the injury mechanisms and functional outcome

Based on these findings, we propose that age-related upregulation of glutamate, GS and GLT-1 (Fig. 16.17b, blue circles) is not limited to astrocytes but extends to other white matter components (Baltan et al. 2008). Na⁺-dependent and Ca²⁺-dependent mechanisms involving astrocytes, oligodendrocytes expressing EAAC1 (Arranz et al. 2008), microglia expressing GLT-1 or axons with VGLUTs (Kukley et al. 2007; Ziskin et al. 2007) become additional sites of glutamate release with aging and contribute to increased excitotoxicity. In young white matter, activation of either AMPA or kainate receptors loads mitochondria with Ca²⁺ and fission is enhanced due to abundant Drp-1 levels. Ca²⁺ overload activates n-NOS to produce NO and ROS which are proposed as diffusible second messengers to link oligodendrocyte excitotoxicity to axon injury (Matute et al. 2001; Ouardouz et al. 2003, 2006). Aging leads to mitochondrial fusion (Fig. 16.17b, elongated mitochondria) due to a reduction in Drp-1 levels. The regulated process of mitochondrial fusion and fission controls the spatiotemporal properties of mitochondrial Ca²⁺ responses and the physiological and pathophysiological consequences of Ca²⁺ signals (Szabadkai and Rizzuto 2004; Szabadkai et al. 2004). By enhancing fusion or inhibiting fission, elongated mitochondria efficiently absorbs Ca²⁺ preventing n-NOS activation and subsequent ROS production (Cheung et al. 2007). However this age-related adaptive reorganization of mitochondria becomes detrimental under ischemic conditions. Ischemia, in aging white matter, further enforces mitochondrial fusion as a result of an age-dependent drop in Drp-1 and an age-dependent loss of mitochondrial motility with exposure to glutamate (Chang and Reynolds; Parihar and Brewer). Mitochondria fuse to collectively counteract the already increased excitotoxicity and Ca²⁺ load with aging, and this age-related change in mitochondrial dynamics could hinder ATP production. This challenges the Na⁺/K⁺ ATP pump to maintain axon excitability and the associated rise in Na⁺ levels challenges the GLT-1 to function in forward a direction to take up glutamate.

An age-specific understanding of the mechanisms of injury processes in white matter is essential to design dynamic therapeutic approaches for stroke victims.

Therefore an age-dependent reduction in mitochondrial bioenergetics may underlie the increased vulnerability of aging axons to ischemia.