Glutamate, the major excitatory neurotransmitter in the central nervous system, transmits signals through several distinct subtypes of receptors. Among others, the N-methyl-d-aspartate (NMDA) subtype of glutamate receptor-associated channels plays a crucial role in neuronal activity-dependent processes including neural development (Contestabile, 2000) and synaptic plasticity (Silva, 2003), owing to their voltage-dependent opening and high Ca²⁺ permeability. On the other hand, overactivation of NMDA receptors and resultant influx of excess Ca²⁺ can be harmful to neurons, which may contribute to various neuropathological consequences associated with neuronal degeneration and death (Hardingham and Bading, 2003).

A remarkable feature of NMDA receptor channels is that, in addition to the agonist glutamate, they require a coagonist glycine for their activation (Johnson and Ascher, 1987; Kleckner and Dingledine, 1988). The idea that glycine site activation is essential for NMDA receptor-mediated physiological processes is supported by the findings that glycine site antagonists block the induction of long-term potentiation of hippocampal synaptic transmission (Izumi et al., 1990) as well as acquisition of spatial memory in rats (Watanabe et al., 1992). Apart from the physiological aspects of NMDA receptor functions, several studies have demonstrated remarkable protective effects of glycine site antagonists against NMDA receptor-mediated neuronal injury in vitro (McNamara and Dingledine, 1990; Patel et al., 1990) and in vivo (Foster et al., 1990), indicating that glycine site stimulation is also crucial for neuropathological events as a consequence of NMDA receptor activation.

Although earlier pharmacological experiments have already shown that d-serine is a potent agonist at glycine site of NMDA receptors (Kleckner and Dingledine, 1988), the compound was not initially regarded as an endogenous ligand because d-amino acid was assumed to be absent in mammalian tissues and organs. Since the discovery of substantial quantities of brain d-serine (Hashimoto et al., 1992; Hashimoto and Oka, 1997), however, emerging evidence supports the hypothesis that d-serine, rather than glycine, may be important as an endogenous “glycine site” ligand at least in several brain regions (Hashimoto et al., 1993; Schell et al., 1995; Snyder and Ferris, 2000). For example, NMDA receptor-mediated neurotransmission in neonatal rat cerebellar slices is markedly attenuated by application of d-amino acid oxidase (DAAOX), an enzyme that efficiently degrades extracellular d-serine (Mothet et al., 2000). DAAOX was also able to attenuate NMDA receptor-mediated synaptic transmission in hippocampal slices and dissociated hippocampal cultures (Mothet et al., 2000), suggesting that endogenous d-serine is essential for maintaining normal synaptic transmission. Recent reports using a similar approach proposed that endogenous d-serine supplied from glial cells may set the sensitivity of NMDA receptors at glutamatergic synapses in the retina (Stevens et al., 2003) and regulate the induction of long-term synaptic plasticity in the hippocampus (Yang et al., 2003).

Biochemical and immunohistochemical observations revealed that d-serine is present in rat cerebral cortex at high levels (Hashimoto et al., 1995; Schell et al., 1995, 1997). The patterns of regional variations and postnatal changes in brain d-serine levels show strong correlations with those of NMDA receptors (Hashimoto et al., 1993). The major cell types containing high levels of d-serine in the telencephalon seem to be astrocytes that are located in close vicinity to NMDA receptor NR2A/B subunits expressed on neuronal dendritic spines (Schell et al., 1995, 1997). Existence of serine racemase, a d-serine synthesizing enzyme, has also been demonstrated in astrocytes of the cerebral cortex (Wolosker et al., 1999). In addition, activation of non-NMDA subtypes of glutamate receptors stimulates efflux of d-serine from cultured astrocytes prepared from rat cerebral cortex (Schnell et al., 1995; Ribeiro et al., 2002). Despite these lines of evidence, however, roles of endogenous d-serine in the cerebral cortex have not been proven. Particularly, the issue concerning whether endogenous d-serine plays a significant role in pathological events such as excitotoxic injury has not been addressed in any brain regions. Using brain slice preparation, we examined potential pathophysiological roles of endogenous d-serine in the cerebral cortex.

Materials and Methods

Materials. MK-801 and 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo-[f]quinoxaline (NBQX) were purchased from Sigma-Aldrich (St. Louis, MO). Homoquinolinic acid and 2,7-dichlorokynurenic acid (DCKA) were obtained from Tocris Cookson Inc. (Bristol, UK). DAAOX was from MP Biomedicals (Irvine, CA) and α-aminoisobutyric acid (AIB) was from Wako Pure Chemicals (Osaka, Japan). Other chemicals and reagents including NMDA, glycine, d-serine, and l-serine were obtained from Nacalai Tesque (Kyoto, Japan).

Slice Preparation. Cerebrocortical slices were prepared from 10- to 16-day-old Wistar rats of both sexes (Nihon SLC, Shizuoka, Japan). Procedures in this study were approved by our institutional animal experimentation committee, and animals were treated in accordance with the Guide for the Care and Use of Laboratory Animals by the National Institutes of Health. Rats under deep anesthesia with diethyl ether were decapitated, and the brains were rapidly removed from the skull and immersed into ice-cold bathing solution saturated with 95% O₂, 5% CO₂. The standard bathing solution consisted of 124 mM NaCl, 3 mM KCl, 1.2 mM NaH₂PO₄, 2.4 mM CaCl₂, 1.3 mM MgSO₄, 26 mM NaHCO₃ and 10 mM glucose. Coronal brain slices (400 μm in thickness) were cut with the use of a vibrating blade microtome (Leica, Nussloch, Germany), and then they were separated into two hemispheres. Slices containing the parietal cortex and the striatum were used for the following experiments, after other brain structures such as the septum and the basal forebrain were removed from each slice. Slices were transferred to a reservoir filled with the standard bathing solution continuously bubbled with 95% O₂, 5% CO₂ and were maintained at 4°C for 1 to 2 h.

Drug Treatment and Simulated Ischemia. Nylon mesh beds (Izumi et al., 1995) of 25 mm in diameter were used for treatment of slices with NMDA or homoquinolinic acid. Slices were transferred into nylon mesh beds (four slices per each bed) that were inserted into a six-well plate, each well containing 6 ml of the standard bathing solution. After preincubation for 1 h, exposure of slices to excitotoxins was achieved by transfer of nylon mesh beds to another six-well plate filled with 6 ml of the bathing solution containing excitotoxins. After 30-min treatment, slices were transferred into yet another plate filled with fresh standard solution and incubated for further 6 h. All solutions were continuously bubbled with 95% O₂,5% CO₂ and maintained at 34°C. In the case of simulated ischemia, slices preincubated for 1 h in the standard oxygenated bathing solution were transferred into glucose-free bathing solution saturated with 95% N₂, 5% CO₂ at 34°C. After oxygen/glucose deprivation for 20 min (unless otherwise indicated), slices were transferred into fresh, glucose-containing standard solution bubbled with 95% O₂, 5% CO₂ and incubated for further 6 h. MK-801, DCKA, and NBQX were applied from the preincubation period of 1 h and were also added in the bathing solution during excitotoxin or ischemic treatment and postincubation for 6 h. d-Serine and glycine were applied only during excitotoxin or ischemic treatment. DAAOX, as well as MK-801 and DCKA in several experiments, was applied from the preincubation period and during excitotoxin or ischemic treatment. Heat inactivation of DAAOX was done by 5-min incubation of DAAOX solution in distilled water at 100°C.

Histological Assessment of Neuronal Damage. At termination of postincubation, slices were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight at 4°C. Then, slices were dehydrated with graded ethanol and xylene, embedded in paraffin, and cut into sections of 5 μm in thickness with a sliding microtome. Nissl staining with toluidine blue was performed on these sections after deparaffinization, and the sections were examined by bright-field optical microscopy. Positively stained round or oval cell bodies showing clear boundary with surrounding parenchyma were judged as viable neurons, and the number of viable neurons in an area of 170 × 245 μm² within the parietal cortex of individual slices was counted. The grid for cell counting was positioned at layer III of the forelimb/hindlimb area, the most dorsome-dial parts of the parietal cortex (Zilles and Wree, 1995). This region was chosen because it contained a high and homogenous density of neurons, and also because it could be easily identified by the presence of a well developed layer IV underneath, a characteristic of the parietal cortex. Data shown in figures are from a representative set of experiments. Reproducibility of the results was confirmed by two or three different sets of experiments.

In several experiments, morphology of cell nuclei was examined by staining with Hoechst 33342. Sections were treated with 2 μg/ml Hoechst 33342 in PBS at room temperature, washed three times in PBS, and viewed under fluorescence microscope with a 4,6-dia-midino-2-phenylindole filter set.

Statistics. Data are expressed as means ± S.E.M (n = 4 for each condition). Statistical significance was evaluated by one-way analysis of variance followed by Student-Newman-Keuls test. Probability values less than 0.05 were considered significant.

Results

NMDA- and Ischemia-Induced Neuronal Injury in Rat Cerebrocoritical Slices. Cerebrocortical slices were prepared from 10- to 16-day-old rats. When these slices were maintained in an oxygenated balanced salt solution for 7.5 h and examined histologically thereafter, the majority of cells inside slices displayed healthy looking, large round or oval cell bodies, which were judged as viable neurons (Fig. 1A). On the other hand, when slices were treated with NMDA for 30 min and incubated for further 6 h in drug-free solution, the number of viable neurons markedly decreased. The cytotoxic effect of NMDA was concentration-dependent within a range of 10 to 300 μM and was abolished by 10 μM MK-801, a noncompetitive NMDA receptor antagonist (Fig. 1B). Simulated ischemia, which was achieved by incubation of slices in a solution devoid of oxygen and glucose followed by postincubation for 6 h, also resulted in a marked decrease in the number of viable neurons (Fig. 1A). The extent of injury was dependent on the duration of oxygen/glucose deprivation (Fig. 1C). The decrease in the number of viable neurons induced by 20 min of simulated ischemia was markedly suppressed by 10 μM MK-801, whereas the compound did not provide a significant protective effect against injury by 30-min ischemia. Although expression levels and subunit compositions of NMDA receptors in the cerebral cortex may undergo developmental changes (Sheng et al., 1994), we did not find obvious changes in NMDA- and ischemia-induced neuronal death within the ages ranging from postnatal 10 to 16 days.

Endogenous Supply of Glycine Site Ligand(s) Is Both Necessary and Sufficient for Neuronal Damage by NMDA and Simulated Ischemia. Because in the present experimental settings the cortical slices were bathed in a balanced salt solution, ligand(s) for glycine site of NMDA receptors should be provided endogenously within the slices. To ascertain whether the presence of endogenous glycine site ligand(s) is critical for NMDA receptor-mediated neuronal damage, we examined the effect of DCKA on NMDA- and ischemia-induced injury. DCKA is a potent antagonist acting at the glycine site of NMDA receptor complex (Baron et al., 1990). DCKA attenuated NMDA cytotoxicity onto cortical neurons in a concentration-dependent manner (Fig. 2A). Significant neuroprotective effect was observed with DCKA at concentrations of 3 μM or higher, and at 100 μM the drug almost abolished the cytotoxicity of 30 μM NMDA. DCKA was also effective in protecting cortical neurons from ischemic injury. At concentrations of 10 μM or higher, DCKA markedly and significantly increased the number of surviving neurons after ischemic insults in a concentration-dependent manner (Fig. 2B). These results confirm that endogenous glycine site ligand(s) are essential for NMDA receptor-mediated injury. MK-801 (10 μM) and DCKA (30-100 μM) were fully protective even if they were omitted from postincubation solution after NMDA or ischemic treatment (Fig. 2C), indicating that NMDA receptor activation during these insults is crucial for neuronal damage.

Several lines of evidence suggest that the glycine site of NMDA receptors is not saturated in in vitro brain slice preparations (Bergeron et al., 1998; Chen et al., 2003). Accordingly, we examined whether exogenous application of glycine site agonists could exacerbate neurotoxic consequences of NMDA and simulated ischemia. As shown in Fig. 3A, addition of glycine (10-30 μM) or d-serine (100-300 μM) did not provide significant influences on the decrease in the number of surviving neurons induced by 30 μM NMDA. In addition, glycine and d-serine showed no significant effect on ischemia-induced neuronal injury (Fig. 3B). Glycine (100 μM) and d-serine (1 mM) alone applied for 30 min did not affect the viability of neurons assessed after 6 h (control, 177.3 ± 3.0 cells/field; glycine, 184.0 ± 11.3 cells/field; d-serine, 182.3 ± 6.0 cells/field). These results suggest that sufficient amount of glycine site ligand(s) is supplied endogenously to allow NMDA receptor overactivation leading to neuronal injury.

DAAOX Inhibits NMDA- and Ischemia-Induced Neuronal Injury. To verify whether d-serine plays an important role in the cerebral cortex as an endogenous glycine site ligand, we used DAAOX. DAAOX is an enzyme that catalyzes oxidative deamination of d-amino acids (D'Aniello et al., 1993). At physiological pH, this enzyme is highly selective for d-serine and does not act on other amino acids, including glycine. Indeed, the specificity of the enzyme has been demonstrated in cerebellar slice preparation, in that DAAOX treatment reduces d-serine levels by 90%, without affecting significantly the levels of other amino acids including glycine, l-serine, and l-glutamate (Mothet et al., 2000). When DAAOX at 1 to 30 U/liter was applied 1 h before and during 30-min treatment of cortical slices with 30 μM NMDA, it showed a marked protective effect against NMDA cytotoxicity in a concentration-dependent manner. At 30 U/liter, DAAOX almost completely inhibited the decrease in the number of surviving neurons induced by NMDA (Fig. 4A).

Although DAAOX is highly selective for neutral d-amino acids and is unlikely to degrade NMDA (Mothet et al., 2000), we further examined the effect of DAAOX against neurotoxicity of homoquinolinic acid, another selective NMDA receptor agonist that does not possess amino acid structure (Patneau and Mayer, 1990). Homoquinolinic acid at 30 μM was cytotoxic to cortical neurons, which was abolished by 10 μM MK-801 (Fig. 4B). Notably, DAAOX afforded a remarkable protective effect against homoquinolinic acid cytotoxicity, being significantly effective at 3 and 10 U/liter. Protective effect of DAAOX was also evident in the case of ischemic neuronal injury. Thus, DAAOX at 1 to 30 U/liter provided concentration-dependent protection against neuronal damage induced by 20 min of simulated ischemia (Fig. 4C).

We also performed Hoechst 33342 staining (Pizzi et al., 2000) to verify the neuroprotective effect of DAAOX against NMDA- and ischemia-induced injury. Nuclear condensation and fragmentation were clearly observed in slices treated with 30 μM NMDA. Changes in nuclear morphology such as shrinkage and distorted appearance were also evident in slices that received 20 min of simulated ischemia. These deleterious effects were markedly prevented by application of 30 U/liter DAAOX before and during NMDA or ischemic treatment (Fig. 5).

Specificity of the effect of DAAOX was examined further. The protective effect of 10 U/liter DAAOX against NMDA neurotoxicity was reversed by exogenous application of 1 mM d-serine or 100 μM glycine (Fig. 6A), which strongly suggest that the effect of DAAOX is attributable to the removal of endogenous d-serine by its enzymatic activity. Moreover, DAAOX protein at equivalent amount with 10 U/liter, when heat-inactivated before application, showed no significant effect on NMDA neurotoxicity (Fig. 6A). We obtained similar results when we did these examinations onto the protective effect of DAAOX against ischemic neuronal injury. That is, the effect of DAAOX was abolished either by concomitant application of excess d-serine and glycine or by prior heat inactivation (Fig. 6B). Moreover, the protective effect of DAAOX against NMDA and ischemic insults was not significantly affected by l-serine (Fig. 7).

A recent report has shown that d-serine is a substrate of a neutral amino acid transporter asc-1 expressed mainly on presynaptic nerve terminals in rodent brain (Helboe et al., 2003). To determine whether inhibition of asc-1-mediated amino acid transport by d-serine contributed to the observed effect of d-serine, we examined the effect of AIB, a well known inhibitor of system A/L transport that was shown to inhibit asc-1 (Helboe et al., 2003). AIB neither reversed the protective effect of DAAOX, nor affected the effect of d-serine (Fig. 7).

Blockade of non-NMDA Receptors Does Not Affect Sensitivity of NMDA- and Ischemia-Induced Neuronal Injury to a Glycine Site Antagonist. Application of non-NMDA receptor agonists has been shown to stimulate efflux of d-serine from cultured astrocytes (Schell et al., 1995; Ribeiro et al., 2002). If this regulatory mechanism of d-serine efflux makes significant contributions to d-serine supply during excitotoxic or ischemic insults, blockade of non-NMDA receptors is expected to decrease the amount of available d-serine and therefore provide neuroprotective effects by compromising NMDA receptor overactivation. According to these assumptions, we examined the influences of a non-NMDA receptor antagonist NBQX on NMDA- and ischemia-induced neuronal damage. NBQX, applied to cerebrocortical slices at 10 μM during the entire course of the incubation periods, did not show significant effect on neuronal damage induced by 30 μM NMDA. In addition, the concentration dependence of the neuroprotective effect of DCKA against NMDA cytotoxicity was not apparently affected by NBQX (Fig. 8A; compare with Fig. 2A). Similar results were obtained in the case of ischemic neuronal injury (Fig. 8B), namely, neuronal injury induced by 20 min of simulated ischemia was not affected by 30 μM NBQX. Moreover, concentration-dependent protective effect of DCKA against ischemic neuronal damage was not altered by NBQX (compare Fig. 8B with Fig. 2B), indicating that non-NMDA receptor-stimulated d-serine efflux does not contribute significantly to the outcome of ischemic neuronal injury under the present experimental conditions.

Discussion

Neuronal damage resulting from overactivation of NMDA receptors contributes to various pathological processes in the central nervous system, including acute disorders such as ischemia and trauma as well as slowly progressing neurodegenerative diseases (Hardingham and Bading, 2003). In the present study, we used cerebrocortical slice preparation maintained in a defined salt solution, to verify unambiguously whether endogenous glycine site ligand, particularly d-serine, plays a key role in determining excitotoxic consequences in the cerebral cortex. DCKA, a potent glycine site antagonist, showed a remarkable protective effect against neuronal injury induced by NMDA or simulated ischemia. On the other hand, addition of d-serine or glycine did not potentiate cytotoxic effects of NMDA and ischemia. These results suggest that endogenous supply of glycine site agonist(s) is both necessary and sufficient for overactivation of NMDA receptors under pathological conditions. In contrast to our findings, a recent electrophysiological study on brain slice preparation suggested that glycine site of NMDA receptors on rat prefrontal cortical neurons was not saturated, because glycine site agonists potentiated synaptically evoked NMDA excitatory postsynaptic currents (Chen et al., 2003). A plausible explanation for the apparent discrepancy is that, unlike in the case of synaptic transmission under physiological conditions, partial occupancy of the glycine site may be sufficient for NMDA receptor overactivation leading to neuronal injury, under pathological conditions where a large amount of glutamate-binding site ligands are present. It is also possible that under pathological conditions such as ischemia, extracellular concentrations of glycine site ligand(s) are elevated, reaching levels sufficient for saturating the glycine site of NMDA receptors (Lo et al., 1998; see below).

Critical roles of glycine site stimulation in induction of NMDA receptor-mediated neuronal damage have been repeatedly demonstrated in several brain regions such as the hippocampus (Foster et al., 1990; Newell et al., 1995) and the striatum (Foster et al., 1990). What we revealed here beyond these previous studies is that d-serine plays an important role as an endogenous glycine site agonist in determining excitotoxic consequences in the cerebral cortex. Potential endogenous agonists at the glycine site of NMDA receptors are d-serine and glycine, and at present, DAAOX is an only available tool to discriminate which compound is actually working physiologically and pathophysiologically. DAAOX is highly selective for neutral d-amino acids (D'Aniello et al., 1993) and is shown to deplete d-serine in cerebellar slices without affecting the levels of other amino acids, including glycine (Mothet et al., 2000). NMDA receptor-mediated synaptic transmission is significantly attenuated by DAAOX in hippocampal slices and immature cerebellar slices (Mothet et al., 2000) as well as in the isolated retina (Stevens et al., 2003), indicating that endogenous d-serine is crucial for supporting synaptic transmission under physiological conditions. Using DAAOX, the present study demonstrated that endogenous d-serine is also important for pathological processes mediated by NMDA receptor overactivation in the cerebral cortex. This conclusion is supported by the findings that neuronal damage induced by NMDA, homoquinolinic acid, or simulated ischemia was markedly inhibited by application of DAAOX to cortical slices. The neuroprotective effect of DAAOX was abolished by prior heat inactivation and also was reversed by exogenous addition of excess d-serine or glycine, confirming that the enzymatic degradation of d-serine by DAAOX is responsible for the observed neuroprotective effects. The fact that AIB did not influence the protective effect of DAAOX suggests that the reversal of neuroprotection by excess d-serine was indeed mediated by glycine site stimulation rather than blockade of neutral amino acid transport (Helboe et al., 2003). The reversal of the effect of DAAOX by glycine is also most likely due to stimulation of glycine site of NMDA receptors, although possible involvement of other actions of glycine, including activation of inhibitory glycine receptors or influences on amino acid transport, cannot be excluded completely. A previous in vivo microdialysis study demonstrating that extracellular concentration of d-serine in the frontal cortex may be high enough to saturate the glycine site of NMDA receptors (Matsui et al., 1995) is also consistent with the critical role of d-serine in this brain structure.

In contrast to d-serine, l-serine showed little effect on the protection afforded by DAAOX against NMDA and ischemic insults. This is not surprising because l-serine itself is a poor agonist at the glycine site of NMDA receptors (McBain et al., 1989). Moreover, although l-serine may be converted to d-serine by serine racemase (Wolosker et al., 1999) after its cellular uptake, or may stimulate d-serine release by direct exchange through neutral amino acid transporters (Ribeiro et al., 2002), a short period of l-serine application is unlikely to elevate extracellular d-serine levels sufficient to overwhelm the rate of degradation of d-serine by DAAOX.

Although the above-mentioned results strongly suggest that d-serine acts as an endogenous glycine site ligand in the cerebral cortex, we do not entirely exclude the possible contribution of glycine as another endogenous ligand, because we have not examined the consequences of removing endogenous glycine. Utilization of glycine-degrading enzymes such as glycine oxidase (Nishiya and Imanaka, 1998) may provide useful information concerning this issue.

d-Serine immunoreactivity is concentrated in astrocytes (Schell et al., 1995, 1997; Wolosker et al., 1999). In hippocampal CA1 region, the distribution of d-serine immunoreactivity roughly parallels that of NR2A/B immunoreactivity (Schell et al., 1997), supporting the view that the glycine site of NMDA receptors is occupied primarily by glia-derived d-serine. Interestingly, efflux of d-serine from cultured astrocytes is stimulated by agonists of non-NMDA subtypes of glutamate receptors (Schell et al., 1995; Ribeiro et al., 2000). In the present study, however, blockade of non-NMDA receptors by NBQX neither affected NMDA- and ischemia-induced injury, nor altered the concentration dependence of the protective effect of DCKA. These results suggest that non-NMDA receptor-mediated enhancement of d-serine efflux, even if it occurs, has little impact on neuronal injury under the present experimental conditions. Our results do not necessarily mean that concentrations of d-serine show no changes during excitotoxic or ischemic insults, because subtle changes in extracellular d-serine levels may have been overlooked by estimation relying on the effect of DCKA. In addition, mechanisms other than those mediated by non-NMDA receptors may also regulate the rate of d-serine efflux from astrocytes. Indeed, extracellular concentrations of d-serine in rabbit cerebral cortex have been shown to increase during transient forebrain ischemia in vivo (Lo et al., 1998). Although the pathways of d-serine efflux across plasma membranes remain to be established, ASCT-like neutral amino acid transporters seem to play an important role (Hayashi et al., 1997; Ribeiro et al., 2002). Transporter-mediated exchange with other neutral amino acids may be among important mechanisms contributing to the efflux of d-serine because l-serine and l-threonine are potent inducers of d-serine efflux from cultured astrocytes, being more effective than kainate (Ribeiro et al., 2002). Also noteworthy are the findings of Hashimoto et al. (2000), who demonstrated that intrastriatal perfusion of NMDA or kainate caused a significant decrease, rather than an increase, in extracellular concentrations of d-serine. Clearly, there remains much to be done for elucidation of the mechanisms regulating extracellular levels of d-serine.

In conclusion, we provided evidence that d-serine acts as an endogenous ligand at the glycine site of NMDA receptors in the cerebral cortex. The present study also demonstrated for the first time that reduction in extracellular d-serine levels affords remarkable protective effect against NMDA receptor-mediated neurotoxic insults. Hence, controlling brain d-serine levels may constitute a novel strategy for neuroprotection against various neuropathological and neurodegenerative conditions (Snyder and Ferris, 2000; Wolosker et al., 2002).