Introduction

Abstract Introduction Methods Results Discussion References

Excessive activation of glutamate receptors caused by an abnormal release or leak of excitatory amino acids is a pivotal event in ischemia-induced neuron death (Choi, 1992). This "excitotoxic" (Olney, 1989) process can readily be demonstrated in a variety of animal models and has been implicated in the neuropathology associated with both acute (e.g., stroke, traumatic brain injury, hypoglycemia) and chronic (e.g., epilepsy, Parkinson's and Huntington's diseases) conditions (Carter, 1992; Choi, 1988; Robinson and Coyle, 1987). These findings have resulted in a variety of therapeutic strategies aimed at limiting glutamate receptor-mediated neurotoxicity (Palfreyman et al., 1994).

Pharmacological studies indicate that both the ionotropic and metabotropic families of glutamate receptors contribute to the excitotoxic process (Choi, 1992; Pizzi et al., 1993; Sheardown et al., 1993). Nonetheless, among these multiple molecular targets, the development of compounds that block glutamatergic transmission through the ionotropic family of NMDA receptors remains the most intensively investigated at both the preclinical and clinical levels (Cherkofsky, 1995; Maccecchini, 1995; Muir et al., 1994; Sveinbjornsdottir et al., 1993). Among the features that make NMDA receptors attractive targets for drug design are its multiple, allosteric sites that control channel activity (Palfreyman et al., 1994) and the apparent requirement for coordinate occupation of glycine and glutamate recognition sites for channel opening (Carter, 1992; Huettner, 1989; Kleckner and Dingledine, 1988). This latter feature led to the hypothesis that if glycine concentrations are at or near saturating concentration in situ, then glycine partial agonists would exhibit NMDA antagonist properties (Skolnick et al., 1989). Consistent with this hypothesis, ACPC, a high-affinity partial agonist at strychnine-insensitive glycine receptors (Clos et al., 1996; Marvizon et al., 1989; Watson and Lanthorn, 1990; Witkin et al., 1995), attenuates glutamate-induced neurotoxicity in vitro (Boje et al., 1993; Fossom et al., 1995a, b). This effect was concentration dependent and was abolished by addition of exogenous glycine (Boje, et al., 1993). Moreover, when administered at the time of ischemic insult, ACPC was neuroprotective and improved neurological outcome in animal models of global, focal and spinal ischemia (Fossom et al., 1995b; Long and Skolnick, 1994; Zapata et al., 1996).

Chronic treatment with ACPC was also neuroprotective in several of these in vivo models (Long and Skolnick, 1994; Lopez and Lanthorn, submitted; Von Lubitz et al., 1992). However in chronic studies, the final dose of ACPC was administered 24 h before ischemia, and plasma and brain concentrations of ACPC were undetectable at the time of ischemic insult (Von Lubitz et al., 1992). This finding suggests that the neuroprotective effects of acute and chronic administration of ACPC are mediated by different mechanisms (Von Lubitz et al., 1992).

Although the mechanism responsible for the neuroprotective effects of chronic ACPC is unknown, two recent observations suggest that sustained exposure to this glycine partial agonist can produce an alteration in the subunit composition of NMDA receptors. Thus, alterations in radioligand binding to cortical NMDA receptors have been demonstrated after chronic (14-day) administration of ACPC to mice (Nowak et al., 1993). Moreover, in primary cultures of rat cerebellar granule neurons, the levels of mRNA encoding the epsilon-3 subunit homolog were increased after a 24-h exposure to ACPC (Fossom et al., 1995a). The hypothesis that ACPC-induced alterations in subunit composition are responsible for these neuroprotective effects is consistent with the demonstrations that the physiological and pharmacological properties of NMDA receptors are largely defined by their subunit composition (Mori and Mishina, 1995). To further explore this hypothesis, we examined the effects of chronic ACPC administration on the levels of mRNAs that encode NMDA receptor  and epsilon subunits by in situ hybridization. We now report that this chronic regimen of ACPC produces robust region-specific changes in mRNA levels encoding the epsilon family of NMDA receptor proteins.

	Methods

Abstract Introduction Methods Results Discussion References

Animals. Male NIH-Swiss mice (20-25 g) were housed in wire bottom cages (five per cage) with free access to a standard diet and tap water. The vivarium was maintained at 22-25°C on a 12-h light/dark cycle (lights on at 6:00 A.M.). All in vivo procedures were performed in accordance with NIH Animal Care and Use Committee guidelines.

Tissue collection and fixation. Twenty-four hours after the last injection, mice were deeply anesthetized with pentobarbital and perfused transcardially with saline followed by 4% paraformaldehyde in 0.1 M sodium borate. The brains were removed, postfixed in 4% paraformaldehyde in 0.1 M sodium borate for 1 h at 4°C, and subsequently placed in 4% paraformaldehyde-0.1 M sodium borate solution containing 10% sucrose overnight at 4°C. The brains were then frozen in isopentane and stored at 70°C until sectioned. Frozen brains were cut into 15-µm sagittal sections. The slices were collected in a cold cryoprotectant solution (50 mM sodium phosphate buffer, pH 7.3, 30% ethylene glycol, 20% glycerol) and stored at 20°C.

In situ hybridization histochemistry. Hybridization histochemical localization of each transcript was carried out on every sixth brain section with ³⁵S-labeled cRNA probes. General protocols for riboprobe synthesis, probe hybridization and autoradiographic localization of mRNA signal were adapted from Simmons and co-workers (1989). Tissue sections mounted on poly-L-lysine-coated slides were fixed in 4% paraformaldehyde for 20 min. Sections were then: 1) digested with Proteinase-K (10 µg/ml in 100 mM Tris-HCl, pH 8, 50 mM EDTA for 30 min at 37°C); 2) rinsed in diethylpyrocarbonate water; 3) dipped in triethanolamine (0.1 M TEA, pH 8, for 10 min); 4) acetylated by dipping in 0.25% acetic anhydride in 0.1 M TEA, pH 8; and 5) dehydrated through graded concentrations (50, 70, 95 and 100%, 3 min each) of ethanol. The hybridization mixture (70 µl, ~10⁷ cpm/ml) was spotted on each slide, sealed under a coverslip and incubated for 15 to 20 h at 60°C. The coverslips were removed and the slides rinsed (4× SCC: 0.6 M NaCl, 60 mM trisodium citrate buffer, pH 7), digested (RNase-A, 20 µg/ml, 37°C), washed (2× SSC for 10 min, 1× SCC for 10 min, 0.5× SCC for 10 min at 23°C and 0.1× SSC for 30 min at 65°C) and dehydrated by sequential dipping in 50 to 100% ethanol. The slides were exposed to BioMax MR film (Kodak) at 4°C for 24 to 72 h (depending on the probe), then defatted in xylene and dipped in LM1 nuclear emulsion. After 4- to 15-day exposures ( probe, 4 days; epsilon-1 and epsilon-2 probes, 6 days; epsilon-3 probe, 15 days), emulsion-dipped slides were developed in D19 (Kodak) for 3.5 min at 14°C. Thereafter, tissues were rinsed in distilled running water for 1 h, dehydrated through graded concentration of alcohol and embedded with DPX (Aldrich Chemical Co., Milwaukee, WI).

cRNA probe synthesis and preparation. An antisense cRNA probe to detect RNAs encoding  subunits was generated from the 1530-nt PstI fragment (+1330/+2860) of a cDNA encoding the rat homolog of the -1 subunit (denoted NMDAR1 in the rat; Boje et al., 1993), subcloned into pGEM-3zf (Promega Corporation, Madison, WI) and conserved in all splice variants of the  subunit. An epsilon-1 cRNA probe was generated from the 1240-nt PstI fragment (+3110/+4350) of epsilon-1 cDNA, subcloned into pGEM-3zf (Promega Corporation, Madison, WI). An epsilon-2 cRNA probe was generated from the 760-nt ApaI fragment (+3510/+4270) of epsilon-2 cDNA, subcloned into pGEM-11zf (Promega Corporation, Madison, WI). An epsilon-3 cRNA probe was generated from the 1020-nt NotI/HindIII fragment (+2630/+3650) of epsilon-3 cDNA, subcloned into pGEM-11zf (Promega Corporation, Madison, WI). These fragments of cDNA were chosen from regions that lack homology with other subunits. Furthermore, each of these fragments identified a single, appropriately sized RNA species, when used as a probe for Northern blot analysis of mouse brain RNA (data not presented). For the in situ studies reported here, sense-strand cRNA probes were also used to verify the specificity of each antisense-strand probe.

Quantitative analysis. Brain sections from slides dipped in nuclear emulsion were used to obtain hybridization signals. Sections were analyzed with Image I Software (Universal Imaging Corporation, West Chester, PA) with a Nikon optical system coupled to a PC. The optical density of the hybridization signals was measured under dark field illumination at 40× magnification. Brain areas of interest were digitized and subjected to densitometric analysis. The optical densities of each specific region were then corrected for the average background signal, which was determined by sampling cells located outside of the areas of interest. The analysis was performed on the following brain regions: frontal, parietal and occipital cortices (layers II-VI), hippocampal fields (CA1, CA2, CA3 and CA4), dentate gyrus and cerebellum.

Statistical analysis. Data were expressed as optical density (O.D.) values. Data for each subunit in each subfield of cortex were analyzed by two-way ANOVA, with treatment and layer as factors. Data for each subunit in hippocampus were analyzed by two-way ANOVA, with treatment and region (CA1, CA2, CA3, CA4 and dentate gyrus) as factors. Data for each subunit in cerebellum were analyzed by one-way ANOVA. Differences between saline and ACPC treatment for a subunit in a particular layer were determined by the least significant difference (LSD) test.

	Results

Abstract Introduction Methods Results Discussion References

Chronic ACPC differentially alters levels of mRNAs encoding epsilon subunits in cerebral cortex. Chronic treatment with ACPC produced significant changes in the levels of mRNAs encoding the NMDA receptor subunits designated epsilon-1 to epsilon-3 (Kutsuwada et al., 1992) in mice (corresponding to NMDAR2A-C in rats; Monyer et al., 1992; Moriyoshi et al., 1991). These effects depended on both the cortical subfield and layer. ACPC treatment increased levels of mRNA encoding epsilon-1 in frontal (13-20%), parietal (18-39%) and occipital (26-38%) cortex (fig. 1A). Figure 2 shows autoradiograms of epsilon-1 in frontal cortex from representative control (A) and ACPC-treated (B) animals. In contrast, ACPC treatment decreased levels of mRNAs encoding epsilon-2 and epsilon-3 in cortex. mRNA levels for epsilon-2 were decreased in frontal (25-35%) and parietal (10-26%) cortex (fig. 1B). Figure 2 shows autoradiograms of epsilon-2 in the frontal cortex from representative control (C) and ACPC-treated (D) animals. mRNA levels for epsilon-3 were decreased in frontal (18-27%) and occipital (10-42%) cortices and tended to be decreased in parietal cortex (15-21%, P < .075; fig. 1C). Figure 2 shows autoradiograms of epsilon-3 in frontal cortex from representative control (E) and ACPC-treated (F) animals.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Chronic ACPC treatment differentially alters expression of NMDA receptor subunit mRNAs in regions of cerebral cortex. Mice were injected daily for 14 days with saline (open bars) or 200 mg/kg ACPC (solid bars). Twenty-four hours after the final injection animals were sacrificed and RNA levels were quantified for epsilon-1 (panel A), epsilon-2 (panel B) and epsilon-3 (panel C) by in situ hybridization as described under "Methods." Data for each subunit in each region of cortex were analyzed by two-way ANOVA, with treatment and layer as factors. A significant main effect of ACPC treatment is indicated by a horizontal line above the data bars, with significance level noted. A significant effect of ACPC treatment for a subunit in particular layer was determined by two-tailed least significant difference (LSD) test (* P < .05, ** P < .01). Values represent mean ± S.E.M. (n = 5-10 mice).

View larger version (214K): [in this window] [in a new window]   Fig. 2.   Representative autoradiograms of frontal cortex illustrating an ACPC-induced increase in epsilon-1 (panel B compared with control in panel A), and ACPC-induced decreases in epsilon-2 (panel D compared with control in panel C) and epsilon-3 (panel F compared with control in panel E) subunit mRNA levels.

Chronic ACPC differentially alters levels of mRNAs encoding epsilon subunits in hippocampus. Significant elevations in the levels of epsilon-1 mRNA were present in hippocampus (10-19%), particularly CA1 (19%) and CA2 (17%), after chronic ACPC treatment (fig. 3A). Figure 4 shows autoradiograms of epsilon-1 in hippocampus from representative control (A) and ACPC-treated (B) animals. epsilon-2 mRNA levels were not altered in hippocampus (fig. 3B). Figure 4 shows autoradiograms of epsilon-2 in hippocampus from representative control (C) and ACPC-treated (D) animals. Although epsilon-3 mRNA was present in relatively low abundance compared with levels of the other epsilon species, ACPC treatment produced an overall reduction in epsilon-3 mRNA levels, particularly in CA1 (30%) and dentate gyrus (26%) (fig. 3C). This is illustrated in autoradiograms from representative control and ACPC-treated animals in figure 4E, F. 

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Chronic ACPC treatment differentially alters expression of NMDA receptor subunit mRNAs in regions of hippocampus. Mice were injected daily for 14 days with saline (open bars) or 200 mg/kg ACPC (solid bars). RNA levels for epsilon-1, epsilon-2 and epsilon-3 subunits were quantified by in situ hybridization as described under "Methods." Data for each subunit in hippocampus were analyzed by two-way ANOVA, with treatment and area (CA1, CA2 CA3, CA4 and dentate gyrus) as factors. A significant main effect of ACPC treatment is indicated by a horizontal line above the data bars, with significance level noted. A significant effect of ACPC treatment for a subunit in a particular area was determined by a two-tailed least significant difference (LSD) test (* P < .05, ** P < .01). Values represent mean ± S.E.M. (n = 7-10 mice).

View larger version (175K): [in this window] [in a new window]   Fig. 4.   Representative autoradiograms of hippocampus illustrating an ACPC-induced increase in epsilon-1 (panel B compared with control in panel A), no effect of ACPC on epsilon-2 (panel D compared with control in panel C) and an ACPC-induced decrease in epsilon-3 (panel F compared with control in panel E) subunit mRNA levels.

Chronic ACPC does not alter levels of mRNAs encoding either epsilon subunits in cerebellum or  subunits in any brain area studied. ACPC did not alter levels of epsilon-1, epsilon-2 or epsilon-3 mRNAs in cerebellum (fig. 5). Analysis of mRNA encoding  subunit(s), with a probe that detects all splice variants equally, revealed no significant differences between control and ACPC-treated animals in any brain regions analyzed, including cerebral cortex (frontal, parietal or occipital subfields), hippocampus (CA1, CA2 CA3-4 or dentate gyrus), cerebellum, striatum or thalamus (data not presented).

View larger version (6K): [in this window] [in a new window]   Fig. 5.   Chronic ACPC treatment does not alter expression of NMDA receptor subunit mRNAs in cerebellum. Mice were injected daily for 14 days with saline (open bars) or 200 mg/kg ACPC (solid bars). Twenty-four hours after the final injection animals were sacrificed and RNA levels were quantified for epsilon-1, epsilon-2 and epsilon-3 by in situ hybridization as described under "Methods." Values represent mean ± S.E.M. (n = 6-10 mice).

	Discussion

Abstract Introduction Methods Results Discussion References

In this study we demonstrate that chronic administration of ACPC, a glycine partial agonist, alters mRNA levels encoding NMDA epsilon-1, epsilon-2 and epsilon-3 (but not ) subunits in a region-specific manner. This investigation was prompted by the demonstration that chronic administration of ACPC was neuroprotective in animal models of global (Von Lubitz et al., 1992), focal (Lopez and Lanthorn, submitted) and spinal (Long and Skolnick, 1994) ischemia. ACPC was also reported to be neuroprotective when administered at the time of ischemic insult (Long and Skolnick, 1994; Zapata et al., 1996), but it was hypothesized that the neuroprotective effects produced by acute and chronic treatment were not mediated by an identical mechanism (Fossom et al., 1995b). Thus, chronic treatment regimens included a 24-h washout before the induction of ischemia, and at the time of insult, tissue levels of ACPC were below the limits of detection in gerbils that received seven daily injections of drug (Von Lubitz et al., 1992). Although drug levels were not measured in the focal and spinal ischemia studies (Long and Skolnick, 1994; Lopez and Lanthorn, submitted), both the short plasma half-life of ACPC in mice and rats (1.5 and 2.5 h, respectively) (Maccecchini, 1995) and the lack of identified metabolites (Cherkofsky, 1995) argue that no significant accumulation would occur in these species after chronic administration.

Excessive activation of NMDA receptors is one of a complex series of extra- and intracellular events in the "excitotoxic cascade" initiated by an ischemic insult (Choi, 1992; Greene and Greenamyre, 1996; Maccechini, 1995). Thus, it would not be necessary to invoke a direct link between the neuroprotection associated with chronic ACPC treatment and changes in the composition of NMDA receptors. Nonetheless, we examined mRNA levels encoding NMDA receptor subunits because two recent reports indicate that sustained exposure to ACPC can alter the properties of this family of ligand-gated ion channels. Thus, Nowak and co-workers (1993) reported an ~2-fold reduction in the potency of glycine to inhibit [³H]5,7-dichlorokynurenic acid binding (an antagonist at strychnine-insensitive glycine receptors; Baron et al., 1991) to NMDA receptors in cortical membranes after a chronic regimen of ACPC. Furthermore, we observed a ~2.5-fold increase in epsilon-3 mRNA levels after a 24-h exposure of cerebellar granule cell neurons to ACPC in cell culture (Fossom et al., 1995a). The present demonstration of changes in mRNA levels encoding the epsilon family of NMDA receptor subunits is consistent with the hypothesis that sustained exposure to ACPC can alter NMDA receptor function in vivo. Future studies will be needed to determine whether the changes in mRNA levels that follow chronic ACPC administration are reflected in changes in NMDA receptor subunit protein levels. If these region-specific, bidirectional effects on mRNA levels reflect corresponding changes in the expression of NMDA receptor proteins, then these data offer some insight into the mechanism responsible for the neuroprotective effects produced by chronic treatment with ACPC.

The physiological and pharmacological properties of both wild-type and recombinant NMDA receptors are largely determined by subunit composition (Laurie and Seeburg, 1994; Lynch et al., 1994; Mori and Mishina, 1995; Wafford et al., 1993). Wild-type NMDA receptors are likely constituted as heterooligomers, assembled from combinations of  and one or more epsilon subunits (Sheng et al., 1994). ACPC-induced changes in the level of each specific epsilon mRNA were unidirectional among the brain regions examined. For example, ACPC treatment either increased or did not affect epsilon-1 mRNA levels, but in no instance were epsilon-1 levels decreased. Likewise, ACPC decreased (or did not significantly alter) epsilon-2 and epsilon-3 mRNA levels. In recombinant NMDA receptors expressed in Xenopus oocytes, the affinities of glutamate and glycine are both lower (10.5- and 2.4-fold, respectively) in receptors composed of -1 and epsilon-1 subunits than those constituted by -1 and epsilon-3 subunits (Kutsuwada et al., 1992). Similarly, the affinities of glycine and glutamate are 2.1- and 7-fold lower in receptors constituted with -1 and epsilon-1 subunits than in those containing -1 and epsilon-2 subunits. With the corresponding rat cRNAs, Wafford and co-workers (1993) reported that the affinities of glycine and glutamate are 24- and 6.6-fold lower in receptors composed of -1 and epsilon-1 subunit homologs than in receptors constituted by -1 and epsilon-3 homologs. Moreover, the affinities of glycine and glutamate at NMDA receptors containing both epsilon-1 and epsilon-3 with -1 homologs were higher than at receptors containing only -1 and epsilon-1 subunit homologs. If chronic administration of ACPC results in a higher proportion of NMDA receptors containing epsilon-1 and/or a lower proportion of receptors constituted with epsilon-2 and epsilon-3 subunits (as indicated by the changes in mRNA levels presented here), then the affinities of glycine and glutamate will be lower in this new receptor pool. Because ischemia results in sustained elevations of both glycine and glutamate levels (Globus et al., 1991), it would be predicted that receptor populations with a lower affinity for these agents would be less susceptible to excitotoxic damage. Although speculative, there is evidence to support this hypothesis. For example, the potency of glycine to inhibit [³H]5,7-dichlorokynurenic acid binding to strychnine-insensitive glycine sites in cortical membranes is reduced ~2-fold after chronic ACPC treatment (Nowak et al., 1993). Conversely, sustained exposure of primary cultures of cerebellar neurons to ACPC, which results in significantly larger increases in both NMDA-stimulated changes in intracellular calcium levels and glutamate-induced cell death than in control cultures, is accompanied by 2.5-fold increases in levels of epsilon-3 mRNA (Fossom et al., 1995a). Although these data demonstrate a clear contrast in the effects of sustained exposure to ACPC in vivo and in cell culture, changes in the levels of epsilon-3 mRNA apparently parallel the sensitivity to ischemic insult and NMDA (glutamate) exposure, respectively, in these two model systems. Whether this effect is caused by the partial agonist properties of ACPC or would also be observed with other ligands that occupy strychnine-insensitive glycine sites (i.e., without regard to the intrinsic efficacy) merits further investigation.

Because it is not possible to predict the occurrence of the most common causes of brain ischemia (i.e., stroke, heart attack and traumatic brain injury), a therapeutic strategy with use of chronic administration of a glycinergic ligand seems rather limited. However, such a strategy may prove suitable as prophylaxis for procedures such as coronary artery bypass grafting, in which the patient is hypoxic for extended periods. Moreover, excessive activation of NMDA receptors has been linked to several chronic neurodegenerative disorders including Parkinson's disease (Blandini et al., 1996; Ossowska, 1994), in which such a strategy may also have therapeutic value.