Extracellular Space

Figure 7 Diagram summarizing events of the acute phase of CNS white matter injury. White matter is critically dependent on a continuous supply of O2 and either glucose or lactate supplied by astrocytes and then shuttled to axons through monocarboxylate transporters (Wender et al., 2000). An energy deficit and/or excess demand lead to impairment of ATP-dependent pumps such as the Na-K-ATPase (1a) and Ca-ATPase (1b), including those located on the "axo-plasmic reticulum." Axonal Ca accumulation may begin due to release from internal stores, triggered by depolarization via L-type Ca channels (2a), and generation of IP3 (2b). The rise in [Na]i by flux through non-inactivating Na channels (3a) coupled with depolarization caused by K efflux through a variety of K channels (3b) stimulates the Na-Ca exchanger to operate in the reverse, Ca import mode (4). This Ca accumulation (5) triggers a variety of destructive events including mitochondrial Ca overload (especially during reoxygenation) (6), and over-activation of a number of Ca-dependent enzyme systems (7). NO generated by NO-synthase may inhibit mitochondrial respiration and alter other cellular proteins. Some Na influx may occur through Na/K permeable inward rectifier channels (8) (Eng et al., 1990; Stys et al., 1998). Na overload and depolarization also stimulate glutamate release through reversal of Na-dependent glutamate transport (9), leading to glial injury from activation of ionotropic glutamate receptors (10). Very recently, ATP-activated P2X purinergic receptors were shown to cause Ca-dependent oligodendroglial injury (11) (Matute et al., 2003). A component of Ca influx into damaged axons directly through voltage-gated Ca channels is also likely (12). Endogenously released transmitters such as GABA and adenosine may play an "autoprotective" role (13). Anion transporters such as the K-Cl co-transporter participate in volume dysregulation in glia and the myelin sheath, contributing to conduction abnormalities (14). The locations of the various channels and transporters are drawn for clarity and do not necessarily reflect their real distributions in myelinated axons. Modified from reference Stys, 2004a with permission from Bentham Science Publishers Ltd.

Figure 7 Diagram summarizing events of the acute phase of CNS white matter injury. White matter is critically dependent on a continuous supply of O2 and either glucose or lactate supplied by astrocytes and then shuttled to axons through monocarboxylate transporters (Wender et al., 2000). An energy deficit and/or excess demand lead to impairment of ATP-dependent pumps such as the Na-K-ATPase (1a) and Ca-ATPase (1b), including those located on the "axo-plasmic reticulum." Axonal Ca accumulation may begin due to release from internal stores, triggered by depolarization via L-type Ca channels (2a), and generation of IP3 (2b). The rise in [Na]i by flux through non-inactivating Na channels (3a) coupled with depolarization caused by K efflux through a variety of K channels (3b) stimulates the Na-Ca exchanger to operate in the reverse, Ca import mode (4). This Ca accumulation (5) triggers a variety of destructive events including mitochondrial Ca overload (especially during reoxygenation) (6), and over-activation of a number of Ca-dependent enzyme systems (7). NO generated by NO-synthase may inhibit mitochondrial respiration and alter other cellular proteins. Some Na influx may occur through Na/K permeable inward rectifier channels (8) (Eng et al., 1990; Stys et al., 1998). Na overload and depolarization also stimulate glutamate release through reversal of Na-dependent glutamate transport (9), leading to glial injury from activation of ionotropic glutamate receptors (10). Very recently, ATP-activated P2X purinergic receptors were shown to cause Ca-dependent oligodendroglial injury (11) (Matute et al., 2003). A component of Ca influx into damaged axons directly through voltage-gated Ca channels is also likely (12). Endogenously released transmitters such as GABA and adenosine may play an "autoprotective" role (13). Anion transporters such as the K-Cl co-transporter participate in volume dysregulation in glia and the myelin sheath, contributing to conduction abnormalities (14). The locations of the various channels and transporters are drawn for clarity and do not necessarily reflect their real distributions in myelinated axons. Modified from reference Stys, 2004a with permission from Bentham Science Publishers Ltd.

events may allow a more rational and, it is hoped, more effective design of therapeutic intervention. "Energy deficit," whether due to inadequate supply such as ischemia, impaired energy producing machinery such as mitochondria altered by NO, or excess energy utilization as would occur in a demyelinated axon (see Fig. 8), will inhibit ATP-dependent pumps such as the Na-K- and Ca-ATPases, resulting in flux of ions down their electrochemical gradients. Ca leak from internal stores plays a major role in some white matter tracts, whereas Na influx through noninactivat-ing channels is a proximal event, which in turn triggers a variety of deleterious cascades including the following: (a) Na loading; (b) depolarization; (c) reverse Na-Ca exchange and Ca overload; (d) glutamate efflux from axons as well as glia through reversal of electrogenic Na-dependent glutamate transporters, with subsequent activation of ionotropic and metabotropic glutamate receptors on glial cells and possibly on axons; (e) activation of voltage-gated Ca channels, which may flux Ca directly or further exacerbate release from internal stores; and other steps illustrated in Fig. 7.

From such an "injury map," one can now select targets that will most likely be beneficial. On the one hand, blocking one or several of the Ca-dependent enzymes (e.g., cal-pain) may confer modest protection, because the accumulated Ca ions will go on to stimulate other deleterious Ca-sensitive pathways, a prediction borne out experimentally (Jiang and Stys, 2000) (see also Chapter 20). On the other hand, addressing a proximal step on which several downstream events are dependent appears most attractive. One obvious candidate is the voltage-dependent Na channel, and in particular, the noninactivating component of the Nav1.6 channel (Smith et al., 1998; Herzog et al., 2003b). This channel has been shown to be the predominant Na channel subtype at nodes of mature myelinated axons (Caldwell et al., 2000) and along demyelinated axons in EAE (Craner et al., 2004a) and in MS (Craner et al., 2004b). This channel presumably contributes a substantial fraction of the resting axon membrane Na permeability that is known to be present in axons of the optic nerve (Stys et al., 1993). Interestingly, the inactivation properties of the

Nav1.6 channel have been shown to be Ca-dependent. Increased intracellular Ca leads to slower inactivation and thus an increased Na flux through the channel (Herzog et al., 2003a). This raises the possibility that, as a result of increased levels of intracellular Ca, there is an increased Na influx that could, in turn, have a positive feedback effect, further increasing Na-dependent Ca entry mechanisms such as reverse Na-Ca exchange. Such a mechanism is quite plausible because Ca imaging experiments in optic axons have shown that slowing Na channel inactivation pharmacologically with veratridine greatly increases the amount of Ca influx, even after a single action potential (Verbny et al., 2002). Therefore, active fibers may be particularly vulnerable to such a runaway positive feedback cycle of increased Ca, slowed Na channel inactivation, further activity-dependent increase in Na loads leading to greater Ca entry, and so on, finally culminating in degeneration of the fiber.

In fact a selective block of the persistent component of sodium current may be essential to avoid interfering with normal excitability. For example, TTX, a state-independent Na channel blocker (Catterall, 1980), is protective in many white matter injury models (Stys et al., 1992a; Imaizumi et al., 1997; Teng and Wrathall, 1997; Rosenberg et al., 1999b; Tekkok and Goldberg, 2001). However, this agent is unlikely to be clinically useful because it potently blocks normal neural excitability. There are several classes of Na channel blockers (e.g., local anesthetic, antiarrhythmic, anti-convulsant) that display "use-dependence," that is, potency of block that varies with the rate of activation of the channels and are more selective for the open state of the Na channel. These state-dependent Na channel blockers have the potential of allowing normal signaling to proceed unhindered along axons, yet effectively blocking a persistently open state that would leak excessive Na ions during pathological conditions. This proposition has been tested in the in vitro anoxic optic nerve model using analogs of local anesthetics and antiarrhythmic agents known to preferentially block open, noninactivating Na channels (Stys et al., 1992b, 1995). The permanently charged quaternary lido-

Figure 8 Architectural changes occurring in demyelinated axons, and the hypothesis of an unbalanced energy supply-demand equation. Normal axons (left) are only vulnerable when energy supply is interrupted (e.g., during ischemia). In a demyelinated fiber (right), the increased energy demand, particularly in active axons, coupled with a reduced ATP-producing capacity by altered mitochondria, may create a state of "virtual hypoxia" promoting a loss of Ca home-ostasis and ultimately structural failure of the fiber, manifested as spheroid formation and finally transection. Nav, Kv, Cav: voltage-gated ion channels; NCX: Na-Ca exchanger. (Reproduced from Stys, 2004b, with permission.)

Figure 8 Architectural changes occurring in demyelinated axons, and the hypothesis of an unbalanced energy supply-demand equation. Normal axons (left) are only vulnerable when energy supply is interrupted (e.g., during ischemia). In a demyelinated fiber (right), the increased energy demand, particularly in active axons, coupled with a reduced ATP-producing capacity by altered mitochondria, may create a state of "virtual hypoxia" promoting a loss of Ca home-ostasis and ultimately structural failure of the fiber, manifested as spheroid formation and finally transection. Nav, Kv, Cav: voltage-gated ion channels; NCX: Na-Ca exchanger. (Reproduced from Stys, 2004b, with permission.)

Figure 9 Bar graph showing effects of selected Na channel blockers on pre-anoxic and post-anoxic optic nerve compound action potential (CAP). Gray bars show depression of excitability before anoxia is applied, reflecting the drug's anesthetic properties. Black bars show CAP recovery after anoxia and wash of drug. Control recovery without blockers is «20-30% of control. Lidocaine (1 mM) is an effective neuroprotectant but at anesthetic concentrations. In contrast, charged compounds such as QX-314 and prajmaline that are thought to be more selective for the open conformation of the Na channel, are highly neuroprotective with minimal CAP depression. Prajmaline and tocainide are anti-arrhythmic drugs in clinical use. (Reproduced from Stys, P. K., and Waxman, S. G. [2004]. Ischemic white matter damage. In: "Myelin Biology and Disorders," R. Lazzarini, J. Griffin, H. Lassmann, K.-A. Nave, R. J. Miller and B. D. Trapp, eds. Copyright 2004, with permission from Elsevier. Data from Stys, 1995.)

Figure 9 Bar graph showing effects of selected Na channel blockers on pre-anoxic and post-anoxic optic nerve compound action potential (CAP). Gray bars show depression of excitability before anoxia is applied, reflecting the drug's anesthetic properties. Black bars show CAP recovery after anoxia and wash of drug. Control recovery without blockers is «20-30% of control. Lidocaine (1 mM) is an effective neuroprotectant but at anesthetic concentrations. In contrast, charged compounds such as QX-314 and prajmaline that are thought to be more selective for the open conformation of the Na channel, are highly neuroprotective with minimal CAP depression. Prajmaline and tocainide are anti-arrhythmic drugs in clinical use. (Reproduced from Stys, P. K., and Waxman, S. G. [2004]. Ischemic white matter damage. In: "Myelin Biology and Disorders," R. Lazzarini, J. Griffin, H. Lassmann, K.-A. Nave, R. J. Miller and B. D. Trapp, eds. Copyright 2004, with permission from Elsevier. Data from Stys, 1995.)

caine analog QX-314 (known to be more selective for the open conformation of Na channels) (Yeh and Tanguy, 1985; Wang et al., 1987; Khodorov, 1991) was very effective at a concentration that showed little inhibition of normal electro-genesis. A sampling of effective compounds against in vitro white matter anoxic injury is shown in Fig. 9. The prototypical local anesthetic lidocaine was a potent neuroprotectant, but at concentrations that severely impaired conduction. Charged quaternary amines such as QX-314 and prajmaline were also effective, but at concentrations that exhibited little suppression of excitability.

Unfortunately such charged molecules are unlikely to be effective in vivo because of poor penetration of the blood-brain barrier and cell membranes, which is required for access to the local anesthetic binding site on the cytoplasmic face of the Na channel (Hille, 1977; Ragsdale et al., 1996; Salazar et al., 1996). However, certain anticonvulsants and antiarrhythmic agents can penetrate into the brain and have been shown to be effective against brain ischemia (e.g., mexiletine) (Lee et al., 1999; Hewitt et al., 2001). Even more intriguing is the ability of some of these agents to improve outcome in EAE after systemic administration. Lo et al. (2002, 2003) showed a substantial protective effect of the Na-channel blocking anticonvulsant phenytoin, which substantially lessened the incidence of degeneration of optic nerve and spinal cord axons, maintained conduction in these axons, and improved clinical outcome in progressive EAE. Phenytoin also protected against the reduction in conduction velocity seen in untreated EAE, leading these investigators to suggest that it protected against demyelination (Lo et al., 2003). Another study using flecainide (a use-dependent Na channel blocking antiarrhythmic agent) in a chronic relapsing EAE model also showed a substantial protective effect against axonal degeneration, improvement in electrophysio-logical conduction parameters in the cord, clinical scores, and a sparing of myelin loss (Bechtold et al., 2004). Together, these studies unite at least part of the injury model proposed for anoxia/ischemia, with the still-unresolved sequence of events responsible for white matter injury in neuroinflammatory disorders. Indeed, even the myelin-spar-ing effect of phenytoin and flecainide can be explained by a reduction of glutamate release (by reverse transport, or later because of axonal degeneration), which is thought to target glial elements, potentially including myelin.

It is of course possible that the positive results discussed in the previous paragraph may not be achievable clinically in humans. The concentrations of a Na channel blocker required to adequately block Na influx may not be tolerated in vivo, or other pathways (e.g., poorly selective cation channels such as the inward rectifier, or AMPA/kainate receptors activated by glutamate from other sources such as inflammatory cells) may also contribute to Na influx. Although the primary site of action (on glial cells? on axons?) is not known at this time, a number of studies have confirmed the efficacy of AMPA receptor antagonists such as NBQX and GYKI52466 in in vitro anoxic, ischemic, and traumatic white matter injury models (Agrawal and Fehlings, 1997; Li et al., 1999; Stys and Ouardouz, 2002), as well as in in vivo models of ischemia, spinal cord injury, and EAE (Wrathall et al., 1994, 1997; Rosenberg et al., 1999a; Kanellopoulos et al., 2000; Pitt et al., 2000; Smith et al., 2000; McCracken et al., 2002; Groom et al., 2003). Together with the known neuroprotective actions of these agents in gray matter (Akins and Atkinson, 2002), this class of drug represents an attractive option, perhaps for use in combination with Na channel blockers, and deserves further study.

0 0

Post a comment