Glutamate Release from Axons as a Trigger of Ca Overload

Another consequence of axonal Na loading and depolarization, regardless of initial insult, is the aberrant ion-coupled transport of small organic molecules, such as neurotransmitters, in particular, the excitatory transmitter glutamate. Under normal conditions in the CNS, glutamate is taken back up into cells by a series of Na-dependent glutamate transporters. These molecules support electrogenic transport and are also coupled to K (and protons) (Zerangue and Kavanaugh, 1996; Levy et al., 1998), therefore depolarization, Na influx, and/or K loss (all of which occur in damaged axons) will drive these transporters in the Na and glutamate efflux mode. In support of this contention, exper iments on anoxic dorsal columns have shown that reverse Na-dependent glutamate transport contributes to a significant degree to anoxic glutamate release in this tissue (Fig. 4B), with most of the transmitter released from axon trunks (Li, 1999). This does not preclude additional non-vesicular modes of glutamate release of which there are several, including flux through volume-sensitive anion channels (Kimelberg et al., 1990; Rutledge et al., 1998), exocytosis from astrocytes (Pasti et al., 2001), or release from astrocytes through gap junction hemichannels (Ye et al., 2003). Volume changes and early cytosolic Ca dys-regulation promote these modes of glutamate release, so it is quite plausible that these systems contribute as well; however no firm evidence has yet been provided in injured CNS white matter tracts.

The search for glutamate-release mechanisms in damaged white matter was prompted by prior studies from several laboratories indicating a prominent role of ionotropic glutamate receptors in the pathogenesis of white matter ischemia and trauma (Agrawal and Fehlings, 1997; Wrathall et al., 1997; Li et al., 1999; Rosenberg et al., 1999a; Tekkok and Goldberg, 2001). Some studies have even shown that excitotoxic mechanisms play a role in immune-mediated demyelinating disease in rodents (Pitt et al., 2000; Smith et al., 2000), again emphasizing a stereotyped response of CNS white matter to injury (see previously). Glial cells express a variety of glutamate receptors and thus represent a major target of glutamate-triggered Ca overload. AMPA and kainate receptors are found on mature oligodendrocytes and astrocytes (Jensen and Chiu, 1993; Garcia-Barcina and Matute, 1996; Agrawal and Fehlings, 1997; Matute et al.,

1997) (for reviews see Steinhauser and Gallo, 1996; Matute et al., 2002; and Dewar et al., 2003), whereas NMDA receptor-mediated currents only appear transiently on both glial cell types until approximately 2 weeks of age (Ziak et al.,

1998). White matter astrocytes express all AMPA and kainate receptor subunits except GluR4, whereas oligodendrocytes express only GluR3 and GluR4 AMPA receptor subunits (notably lacking GluR2), as well as all kainate sub-units except GluR5 (Garcia-Barcina and Matute, 1996, 1998; Matute et al., 2002). Persistent activation of these non-NMDA ionotropic receptors injures oligodendrocytes, both in cell culture and in vivo (Yoshioka et al., 1995, 1996; Matute et al., 1997; Matute, 1998; McDonald et al., 1998; Liu et al., 2002). The absence of GluR2 imparts significant Ca permeability and may render oligodendrocytes particularly susceptible to Ca flux through these receptors. AMPA and kainate receptor expression is particularly robust in oligodendrocyte progenitors (Barres et al., 1990; Kastritsis and McCarthy, 1993; Holzwarth et al., 1994), which may contribute to the exquisite sensitivity of these cells to ischemia (Fern and Moller, 2000; Follett et al., 2000). Astrocytes can also be injured when exposed to AMPA receptor agonists (Li and Stys, 2000), particularly when desensitization is blocked (David et al., 1996).

The question of whether myelinated axons per se possess glutamate receptor subunits has not been resolved. Axonal protection by the AMPA/kainate receptor blocker NBQX has been demonstrated in an animal model of EAE (Pitt et al., 2000; Smith et al., 2000). More recently, Tekkök and Goldberg (2001) presented evidence of axonal protection by NBQX in an in vitro model of central white matter ischemic injury. However, more recent data from this group indicate that the axonal protection by glutamate antagonists may be secondary to sparing of oligodendroglia, thereby reducing the generation of free radicals that could be toxic to neighboring axons (Underhill and Goldberg, 2002). Irrespective of whether they are present on axons or glial cells, AMPA/kainate receptors appear to play a prominent role in various modes of white mat ter injury, including hypoxia/ischemia (Li et al., 1999; Follett et al., 2000; Kanellopoulos et al., 2000; Tekkok and Goldberg, 2001; McCracken et al., 2002), trauma (Agrawal and Fehlings, 1997; Wrathall et al., 1997; Rosenberg et al., 1999a), and inflammatory demyelination (Pitt et al., 2000; Smith et al., 2000; Werner et al., 2000; Groom et al., 2003). Representative results from anoxic dorsal columns in the adult rat are shown in Fig. 4. Under normoxic conditions, application of either glutamate or kainate (the latter to activate AMPA and kainate receptors) causes functional injury to dorsal columns, whereas NMDA has no effect, as measured by the propagated compound action potential. Moreover, the broad-spectrum ionotropic glutamate receptor antagonist kynurenic acid, or the more selective AMPA receptor blocker GYKI52466, display robust neuroprotective activity against dorsal column anoxia. Similarly, blocking glutamate release via Na-depend-ent transporters with dihydrokainate or L-trans-pyrrolidine-2,4-dicarboxylic acid (Arriza et al., 1994; Griffiths et al., 1994) is also protective against anoxic damage in this tissue. These results do not distinguish between axonal and glial targets of glutamate-triggered injury, but taken together, they indicate that not only is dorsal column white matter injured by glutamate exposure, but that endogenous glutamate is released in sufficient quantities during anoxia to trigger substantial excitotoxicity.

The mechanisms of excitotoxic damage in white matter may be more complex than simple activation of Ca-perme-able receptors. For instance, preliminary data indicate that selective AMPA receptor blockade (using GYKI52466 or SYM2206) is far more protective against white matter anoxia/ischemia using the optic nerve model (in contrast to spinal cord white matter) than combined inhibition of AMPA and kainate receptors using a less selective agent such as NBQX (Jiang and Stys, 2003) or kynurenic acid. Although these observations have yet to be confirmed and mechanisms elucidated, they raise the intriguing, if not counterintuitive, possibility that kainite-receptor activation may in fact be partially protective in some central white matter tracts. Complicating matters further is the possibility that metabotropic glutamate receptors may also contribute to the injury cascade in ischemia and trauma, by coupling back to the potentially important mechanism of release from internal Ca stores, via a phospholipase C-dependent mechanism acting on IP3 receptors (Agrawal et al., 1998; Stys and Ouardouz, 2002). On the basis of immunocytochemical findings, one study suggests the presence of metabotropic glutamate receptors along axons in MS (Geurts et al., 2003).

It is generally accepted that NMDA receptors do not play a significant role in the pathophysiology of mature white matter injury, as these receptors appear absent in adult tissue (Wyllie et al., 1991; Ziak et al., 1998), although they are transiently expressed on glia of immature spinal cord (Ziak et al., 1998), and neither NMDA receptor activation nor antagonism

Figure 4 (A) Effect of glutamate (Glu), kainate (KA), or NMDA on in vitro dorsal columns. Representative CAP tracings after 180 minutes of exposure and bar graph show controls to be stable for 180 minutes, whereas glutamate (1 mM) or kainate (500 mM) irreversibly reduced CAP amplitude to approximately 40% of control. In contrast, NMDA (500 mM, with 20 mM glycine and in the absence of Mg2+) had no effect on propagated CAP (B) Protective effects of ionotropic glutamate receptor antagonists or glutamate transport blockers against dorsal column anoxic injury. The broad spectrum (NMDA, AMPA, kainate receptor) blocker kynurenic acid (1 mM) and the more selective AMPA receptor antagonist GYKI52466 (30 mM) are both protective against anoxic injury. Similarly, the Na-dependent glutamate transport blockers dihydrokainate (DHK) or L-trans-pyrrolidine-2,4-dicarboxylic acid (PDC; 1 mM each) were also protective. Taken together, these data indicate that endogenous glutamate is released from dorsal columns during anoxia/ischemia via reverse transport due to Na influx and depolarization, causing damage by activation of AMPA-preferring ionotropic receptors. (A: Modified from Li and Stys, 2000, with permission. Copyright 2000 by the Society for Neuroscience; B: Modified from Li et al., 1999, with permission. Copyright 1999 by the Society for Neuroscience.)

Figure 4 (A) Effect of glutamate (Glu), kainate (KA), or NMDA on in vitro dorsal columns. Representative CAP tracings after 180 minutes of exposure and bar graph show controls to be stable for 180 minutes, whereas glutamate (1 mM) or kainate (500 mM) irreversibly reduced CAP amplitude to approximately 40% of control. In contrast, NMDA (500 mM, with 20 mM glycine and in the absence of Mg2+) had no effect on propagated CAP (B) Protective effects of ionotropic glutamate receptor antagonists or glutamate transport blockers against dorsal column anoxic injury. The broad spectrum (NMDA, AMPA, kainate receptor) blocker kynurenic acid (1 mM) and the more selective AMPA receptor antagonist GYKI52466 (30 mM) are both protective against anoxic injury. Similarly, the Na-dependent glutamate transport blockers dihydrokainate (DHK) or L-trans-pyrrolidine-2,4-dicarboxylic acid (PDC; 1 mM each) were also protective. Taken together, these data indicate that endogenous glutamate is released from dorsal columns during anoxia/ischemia via reverse transport due to Na influx and depolarization, causing damage by activation of AMPA-preferring ionotropic receptors. (A: Modified from Li and Stys, 2000, with permission. Copyright 2000 by the Society for Neuroscience; B: Modified from Li et al., 1999, with permission. Copyright 1999 by the Society for Neuroscience.)

exerts any obvious deleterious or protective effects (respectively) in central white matter tracts (Fig. 4) (Agrawal and Fehlings, 1997; Li and Stys, 2000; Yam et al., 2000). This receptor has not been studied in great detail in this context, so its role in white matter pathophysiology cannot be ruled out.

VII. Adenosine and GABA as Modulators of Ca Influx

Experiments on rat optic nerve in vitro have provided evidence suggesting that release of two other neurotransmit-ters/neuromodulators, GABA and adenosine, from endogenous stores within white matter may provide a mechanism that modulates calcium influx into axons (Fig. 5). Fern et al. (1994, 1995b, 1996) showed that anoxia triggers the release of GABA and adenosine from endogenous stores within the optic nerve. These act on GABA-B and adenosine receptors, respectively, and increase resistance to anoxia, thus playing an autoprotective role. This autoprotection appears to involve a G-protein/protein kinase C (PKC) pathway that is activated as a result of binding of GABA or adenosine to their receptors. Although the target of PKC activity is not known, these experiments indicate that there are endogenous mechanisms that can modulate at least some of the pathways for calcium influx into axons. Thus, the response of CNS white matter to injury displays a surprisingly rich dependence on a spectrum of traditional neurotransmitters.

VIII. Nitric Oxide and Free Radicals

Nitric oxide (NO) is synthesized from L-arginine by three isoforms of NO synthase, the constitutively expressed

Figure 5 Proposed model for protective effect of adenosine and GABA in the anoxic optic nerve. According to this model, anoxia triggers K efflux and Na influx which lead to reverse operation of the Na-Ca exchanger and influx of calcium into axons. Anoxia also induces release of GABA and adenosine from endogenous stores within white matter with a resultant activation of GABA-B and adenosine receptors, initiating a G-protein/PKC cascade, which increases the tolerance of axons to anoxia. The target of PKC phos-phorylation is not known, but could be the Na-Ca exchanger or Na channels. (From Fern et al., 1996, with permission.)

eNOS and nNOS, and a third (iNOS), which is induced during inflammation (for a review see Brown and Borutaite, 2002). The biological chemistry of NO and its reaction pathways are complex (for reviews see Brown and Borutaite, 2002 and Chapter 18). However, this molecule may be converted to a number of more active derivatives, known collectively as reactive nitrogen species, which in turn can react with and chemically modify a large number of key cellular proteins, lipids, and nucleic acids. In the present context, perhaps one of the most important structures adversely influenced by NO are mitochondria. This gas inhibits cytochrome oxidase (complex IV) competitively with O2, and certain NO derivatives such as perox-ynitrite (ONOO-) may inactivate all respiratory complexes (I-IV) and the ATP synthase. To make matters worse, per-oxynitrite stimulates proton leak across the inner mitochondrial membrane (Gadelha et al., 1997), which, together with impaired complex activity, can effectively render a cell hypoxic or at a minimum exacerbate an existing injury.

The effects of NO may have profound effects on central white matter tracts. Smith and colleagues (Redford et al., 1997) (see also Chapter 18) showed that exposure of demyelinated spinal axons to NO induced reversible conduction block, whereas myelinated axons in the same tract were affected only at higher concentrations (Fig. 6). Subsequent studies confirmed and extended this observation, suggesting that active axons are particularly sensitive to NO exposure (Smith et al., 2001). Although the precise mechanisms of NO toxicity are not fully understood, direct nitration of key proteins, such as Na channels (Hammar-strom and Gage, 1999; Renganathan et al., 2002) and mito-

Figure 6 CAPs recorded from a demyelinated lesion induced in the rat dorsal column by a previous intraspinal injection of ethidium bromide. Records are plotted at 2-minute intervals. Application of the NO donor spermine NONOate (NO concentration aapproximately 3 jM) resulted in conduction block in many of the demyelinated axons, whereas normal fibers remained largely unaffected. This study underscores the increased susceptibility of demyelinated fibers to NO that may exert its effects by partial mitochondrial inhibition and/or Na channel nitrosylation (see text). (Modified from Redford et al., 1997, with permission.)

Figure 6 CAPs recorded from a demyelinated lesion induced in the rat dorsal column by a previous intraspinal injection of ethidium bromide. Records are plotted at 2-minute intervals. Application of the NO donor spermine NONOate (NO concentration aapproximately 3 jM) resulted in conduction block in many of the demyelinated axons, whereas normal fibers remained largely unaffected. This study underscores the increased susceptibility of demyelinated fibers to NO that may exert its effects by partial mitochondrial inhibition and/or Na channel nitrosylation (see text). (Modified from Redford et al., 1997, with permission.)

ANOXIA

ADENOSINE

GABA STORE

ENOS

ADENOSINE

GABA

ADENOSINELf G H GABA-B RECEPTOR I \ y| RECEPTOR I

AXON

chondrial impairment, are highly likely. Indeed, it appears that the steps set in motion by NO may overlap to a significant extent with those during hypoxia/ischemia. Blockade of voltage-gated Na channels or reduction of axonal Ca load with pharmacological block of the Na-Ca exchanger protect against NO exposure (Garthwaite et al., 2002; Kapoor et al., 2003). Thus, the major effect of NO production under pathological conditions such as inflammatory demyelination may be to further exacerbate the cellular hypoxic state, driving the anoxic/ischemic injury cascade (Fig. 7) even more strongly. If so, neuroprotective strategies (see next section)

may simultaneously mitigate at least some of the damaging effects of NO in CNS white matter.

0 0

Post a comment