Ischemic preconditioning acts upstream of GluR2 down-regulation to afford neuroprotection in the hippocampal CA1

Abstract

Animals subjected to sublethal transient global ischemia (ischemic preconditioning) exhibit neuroprotection against subsequent global ischemia-induced neuronal death in the hippocampal CA1 (ischemic tolerance). The molecular mechanisms underlying ischemic tolerance are unclear. Here we report that ischemic preconditioning induced a small, transient down-regulation of GluR2 mRNA expression and greatly attenuated subsequent ischemia-induced GluR2 mRNA and protein down-regulation and neuronal death. Ischemic preconditioning and GluR2 antisense knockdown acted synergistically to increase cell death. Sublethal antisense knockdown did not protect against subsequent ischemic insults or antisense knockdown. These findings indicate that ischemic preconditioning acts at step(s) upstream from suppression of GluR2 gene expression to afford neuroprotection and implicate transcriptional regulation of GluR2 expression in the adaptive mechanisms associated with ischemic tolerance.

Ischemic tolerance is a well-known phenomenon in which brief ischemic episodes (ischemic preconditioning) afford robust protection to hippocampal CA1 pyramidal neurons against subsequent more severe insults that are otherwise lethal for these neurons (1–6). Tolerance also can be induced in vivo by spreading depression (7), hypoxia (8, 9), seizures (10), adenosine (11), and inhibitors of oxidative phosphorylation (12) and in vitro by exposure to cytokines (13) and excitotoxins (14). The molecular mechanisms underlying ischemic tolerance are not yet fully delineated.

Ischemic preconditioning is thought to require activation of N-methyl-d-aspartate receptors and enhanced Ca²⁺ influx (15–17), as well as opening of ATP-sensitive K⁺ channels by means of activation of adenosine A₁ receptors (11, 18–20). The considerable delay (6–24 h) from the preconditioning stimulus until onset of ischemic tolerance and the long duration (up to 7 days) is consistent with a role for transcriptional changes in adaptation (21). Ischemic preconditioning is known to activate c-fos and c-jun (22) and a number of survival factors including heat shock protein HSP70 (23–29), superoxide dismutase (30), nitric oxide (31), brain-derived neurotrophic factor (32), and the antiapoptotic factor Bcl-2/Bclx_L (33). Preconditioning also prevents ischemia-induced down-regulation of the α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor (AMPAR) “Ca²⁺ ion gatekeeper” subunit GluR2 (11). The presence of edited GluR2 subunits in recombinant AMPARs greatly reduces Ca²⁺ permeability (34, 35), voltage-dependent block by polyamines (36, 37), and single-channel conductance (38) and may govern AMPAR recycling and targeting to the synapse (39). Because preconditioning by sublethal ischemia largely prevents GluR2 down-regulation induced by more severe ischemic insults, the molecular machinery involved in GluR2 suppression is a candidate mediator of ischemic tolerance in neurons.

Global ischemia modifies AMPAR subunit composition in a manner that promotes Ca²⁺ permeability at CA1 synapses (46) and induces delayed death primarily of hippocampal CA1 neurons. The present study was undertaken to examine a possible role for GluR2 transcriptonal machinery in ischemic tolerance. Here we report that ischemic preconditioning induced a small, transient suppression of GluR2 mRNA expression specifically in CA1 pyramidal neurons with no change in GluR2 protein expression. Preconditioning greatly attenuated ischemia-induced suppression of GluR2 mRNA and protein. Preconditioning and GluR2 antisense knockdown acted synergistically to increase cell death (4). Sublethal antisense knockdown did not protect against subsequent ischemic insults or antisense knockdown. These findings suggest that ischemic preconditioning acts at a step upstream from suppression of GluR2 gene expression to afford neuroprotection and implicate regulation of GluR2 expression in ischemic tolerance.

Methods

Global Ischemia and Ischemic Preconditioning in Gerbils.

All experiments except for those involving antisense injections were performed on gerbils. Gerbils offer an advantage compared with rats in that forebrain ischemia can be produced by the relatively simple two-vessel occlusion model. To induce ischemia in gerbils, age-matched, adult male Mongolian gerbils (Charles River Breeding Laboratories), weighing 60–80 g, were maintained in a temperature- and light-controlled environment with a 14-h light/10-h dark cycle and treated in accord with the principles and procedures of the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. All protocols were approved by the Institutional Animal Care and Use Committee of the Albert Einstein College of Medicine. Gerbils were fasted overnight and (on the day of surgery) anesthetized with halothane (induction, 4%; maintenance, 1%) and subjected to global ischemia by temporary bilateral occlusion of the carotid arteries (5 min for global ischemia, 2 min for sublethal ischemia) or to sham occlusion, followed by reperfusion as described (40). Arteries were visually inspected to ensure adequate reflow, and then anesthesia was discontinued. Treatment groups were as follows: control (sham operation), sublethal ischemia (2 min), lethal ischemia (5 min, i.e., lethal for CA1 pyramidal neurons), or sublethal ischemia (ischemic preconditioning), followed by lethal ischemia at 72 h (to test for ischemic tolerance).

Sublethal Ischemia and Lethal Ischemia in Rats.

Experiments involving a combination of antisense injection and ischemia were performed on rats. Rats were preferred for these experiments, because the GluR2 cDNA sequence is known, and antisense oligonucleotides could be directed against the mRNA (4). To induce global ischemia in rats, age-matched male Sprague–Dawley rats (Charles River Breeding Laboratories), weighing 125–200 g, were fasted overnight. On the day of surgery, rats were anesthetized with halothane as above. Rats were subjected to sublethal (4 min) or lethal (10 min) ischemia by four-vessel occlusion or to sham operation according to Pulsinelli and Brierley (41).

Antisense Oligodeoxynucleotides and Administration of GluR2 Antisense in Rats.

The specific antisense oligonucleotide AS (5′-AACCATTTTATCCACTTCACT-3′) targets positions 1168–1188 of the GluR2 cDNA; “scrambled” antisense SC (5′-CTAACCTCCAATCTTATTCTA-3′) contained the identical base composition in a randomized sequence (4). For administration of oligonucleotides, rats were anesthetized with halothane as above. Body temperature and anesthesia level were monitored. For intracerebroventricular injection, a cannula (28 gauge; stainless steel; inner diameter, 0.18 mm; outer diameter, 0.36 mm) was lowered stereotaxically into the right lateral ventricle, and antisense (10 nmol in 5 μl saline) or saline (5 μl) was delivered at a flow rate of 1 μl/min. A single injection (sublethal antisense) or four injections at 12-h intervals (lethal antisense) were given. Four injections were required to kill CA1 pyramidal neurons. Animals were not observed to exhibit seizures.

Histological Analysis.

Neuronal cell loss was evaluated by histological examination of brain sections at the level of dorsal hippocampus from animals killed at 7 days after ischemia or sham operation or for GluR2 knockdown, after the last injection. Animals were placed under deep anesthesia and fixed by transcardiac perfusion with ice-cold 4% paraformaldehyde. Brains were removed and immersed in fixative (4°C, overnight). Coronal sections (15 μm) were cut and stained with toluidine blue. To determine neuronal counts, the number of surviving neurons per 250 μm length of the CA1 region was counted for four animals/treatment group (four sections/animal) and analyzed by ANOVA with Scheffé's post hoc tests.

In Situ Hybridization.

To examine the effect of ischemic preconditioning on GluR1 and GluR2 mRNA expression in the hippocampus, gerbils were subjected to sham operation, sublethal ischemia, lethal ischemia, or sublethal ischemia followed by lethal ischemia, anesthetized with pentobarbital, and decapitated at 24, 48, or 72 h after the last ischemia or 24 h after sham operation. mRNA expression was assessed by in situ hybridization on coronal sections of brain at the level of the hippocampus as described (42). In brief, frozen brain sections from experimental and control animals were hybridized (overnight, 49°C) with ³⁵S-labeled GluR1 or GluR2 RNA probes (10⁶ cpm/section; 1 ng/μl). For quantitation of mRNA expression, autoradiograms were analyzed with a Scan Jet 4-C computing densitometer using NIH IMAGE 1.61 image-analysis software as described (4). Mean OD values for sections from experimental animals were normalized to the corresponding values for control sections apposed to the same film to enable comparisons among different films. Statistical significance was assessed by ANOVA with Scheffé's post hoc test.

Western Blotting.

To examine effects of preconditioning on GluR2 protein abundance, Western blot analysis was performed on samples isolated from hippocampal subfields from experimental and control gerbils at 24, 48, and 72 h after reperfusion as described (43). The CA1 and dentate gyrus containing the hilus and CA3c (hereafter termed dentate gyrus) were rapidly microdissected and stored at −70°C until use. Protein samples (10 μg) from the CA1 and dentate gyrus were subjected to gel electrophoresis (10% polyacrylamide minigels; Bio-Rad), transferred to nitrocellulose membranes, and labeled with a mAb directed to a sequence within the N-terminal domain of the GluR2 subunit (1:1,000, Pharmigen). Mean band densities for protein samples from the CA1 and dentate gyrus of experimental animals were normalized to the corresponding values of controls on the same membrane to enable comparisons of bands on immunoblots apposed to different films. Statistical significance was by ANOVA with Scheffé's post hoc tests.

Results

Ischemic Preconditioning Protects Against Ischemia-Induced Neuronal Death in CA1.

To examine the neuroprotective effect of ischemic preconditioning, we subjected gerbils to four protocols—sham operation, sublethal ischemia (2 min), lethal ischemia (5 min), and sublethal ischemia—followed after 3 days by lethal ischemia. We assessed neuronal death histologically at 7 days after the last ischemia. Sublethal ischemia did not induce neuronal damage (Fig. 1 C and D). Lethal ischemia induced extensive death of pyramidal cells in CA1; the few remaining pyramidal neurons were severely damaged and appeared pyknotic (Fig. 1 E and F). Little or no cell loss occurred in CA3 or dentate gyrus. Ischemic preconditioning by sublethal ischemia afforded robust and nearly complete protection of CA1 neurons against lethal ischemia induced 3 days later (Fig. 1 G and H). These findings were validated by neuronal counts (Fig. 1I). These data are in confirmation of Kitagawa et al. (44) and Kirino et al. (23).

Figure 1.

Ischemic preconditioning affords protection against death of CA1 pyramidal neurons induced by lethal ischemia. Toluidine blue-stained coronal brain sections at the level of the dorsal hippocampus from sham operated (n = 5; A and B) and experimental gerbils subjected to sublethal ischemia (2 min; n = 5; C and D), lethal ischemia (5 min; n = 5; E and F), or sublethal ischemia, followed after 72 h by lethal ischemia (n = 5; G and H); animals were killed 7 days after the last surgery. Sublethal ischemia did not induce neuronal death (C and D). Lethal ischemia induced loss of pyramidal neurons in CA1; cell loss was not detected in CA3 or dentate gyrus (E and F). Sublethal ischemia afforded nearly complete protection against subsequent lethal ischemia (G and H). Quantitation of cell counts per unit length of CA1 (I). so: stratum oriens; sp: stratum pyramidale; sr: stratum radiatum. (Scale bar: 400 μm in A, C, E, and G; 50 μm in B, D, F, and H.)

Ischemic Preconditioning Induces a Small, Transient Down-Regulation of GluR2 mRNA in CA1.

To examine a possible role for Ca²⁺-permeable AMPARs in ischemic preconditioning, we evaluated the effect on GluR1 and GluR2 expression in the hippocampus by in situ hybridization. Brain sections were cut at the level of the dorsal hippocampus from experimental and control gerbils and hybridized with RNA probes directed to GluR1 and GluR2 mRNAs (Fig. 2). In control hippocampus, GluR1 and GluR2 mRNA expression was intense in the pyramidal cell layers of CA1 and CA3 and in the granule cell layer of the dentate gyrus. Sublethal ischemia induced a modest, but significant, down-regulation of GluR2 mRNA expression in the pyramidal cell layer of CA1, evident at 48 h (reduction to 75.6% ± 3.3% of control at 48 h; P < 0.05; n = 7; Fig. 2A). By 72 h after sublethal ischemia, GluR2 mRNA had recovered to near control levels (97.1 ± 3.2%; n = 6). The reduction in GluR2 mRNA expression was cell-specific, in that its expression was unchanged in the pyramidal cell layer of the CA3 and the granule cell layer of the dentate gyrus. The change in mRNA expression was subunit-specific, in that GluR1 mRNA expression was not altered in any subfield (determined only at 48 h).

Figure 2.

Sublethal ischemia induces a small, transient down-regulation of GluR2 mRNA in CA1 and attenuates GluR2 mRNA down-regulation induced by subsequent lethal ischemia. (Upper) Film autoradiograms of GluR2 mRNA and GluR1 mRNA expression in the hippocampus of control and experimental gerbils at 24, 48, and 72 h (GluR2) or 48 h (GluR1) after sublethal ischemia (A), lethal ischemia (B), and sublethal ischemia, followed 72 h later by lethal ischemia (C). Control data are the same in each case. (Lower) Mean OD values for the pyramidal cell layer of CA1 (GluR2, ●; GluR1, triangles), for the pyramidal cell layer of CA3 (GluR2, shaded circles) and the granule layer of dentate gyrus (GluR2, ○).

Ischemic Preconditioning Attenuates Ischemia-Induced Suppression of GluR2 mRNA in CA1.

We next examined the effects of ischemic preconditioning on ischemia-induced alterations in AMPAR mRNA expression. In unconditioned gerbils, lethal ischemia markedly reduced GluR2 mRNA expression in the pyramidal cell layer of CA1 (Fig. 2B); GluR2 mRNA was reduced to 67.3 ± 8.1% of control at 24 h (P < 0.01; n = 5); 45.7 ± 7.7% of control at 48 h (P < 0.01; n = 6); and 20.2 ± 3.9% of control at 72 h, a time before the onset of neuronal death (P < 0.01; n = 7; see Fig. 1 and ref. 45). The reduction in GluR2 mRNA expression was cell-specific in that its expression was not altered in dentate gyrus granule neurons (Fig. 2B). The ischemia-induced alteration in mRNA expression was subunit-specific, in that GluR1 mRNA was unchanged at 48 h after ischemia, the only time point at which it was evaluated. These data are in confirmation of Pellegrini-Giampietro et al. (42) and Gorter et al. (45).

In contrast, in preconditioned gerbils, lethal ischemia induced a small, transient down-regulation of GluR2 mRNA in the CA1; GluR2 was reduced to 65.4 ± 10.0% of control at 24 h (P < 0.05; n = 5) and 68.6 ± 5.4% of control at 48 h (P < 0.05; n = 4; Fig. 2C). Thus, at 24 h after lethal ischemia, the degree of GluR2 down-regulation was virtually the same in preconditioned vs. naive animals, but at 48 h was greater in the naive (P < 0.05). By 72 h after lethal ischemia in preconditioned animals, GluR2 mRNA had recovered to near control levels (93.6 ± 3.4% of control; n = 6; Fig. 2C). The transient GluR2 down-regulation observed after lethal ischemia in preconditioned animals was cell-specific in that it was unchanged in CA3 and dentate gyrus. The GluR2 down-regulation was subunit-specific, in that GluR1 expression was unchanged in all hippocampal subfields at 48 h. In rat, ischemic preconditioning prevented ischemia-induced GluR2 mRNA down-regulation at 24 h, the only time tested and one when GluR2 mRNA is greatly reduced in naive animals (11).

Ischemic Preconditioning Prevents Ischemia-Induced Suppression of GluR2 Protein in CA1.

Because mRNA changes are not always followed by changes in protein abundance, we examined the effects of ischemic preconditioning on GluR2 protein in CA1 by quantitative Western blot analysis. Protein samples from the CA1 and dentate gyrus of experimental and control gerbils were subjected to electrophoresis and probed with a mAb directed to a sequence within the N-terminal domain of the GluR2 subunit. Analysis of band densities indicated that sublethal ischemia did not significantly alter GluR2 subunit abundance in CA1 or dentate gyrus as late as 72 h (Fig. 3A). In contrast, lethal ischemia reduced GluR2 subunit abundance in CA1, evident at 48 and 72 h (reduction to 60% of the control value at 72 h; Fig. 3B); the reduction was subfield-specific, in that ischemia did not alter GluR2 subunit abundance in dentate gyrus as late as 72 h. Ischemic preconditioning markedly attenuated the ischemia-induced down-regulation of GluR2 protein in CA1 and did not alter GluR2 abundance in dentate gyrus (73.0 ± 3.0% of control; n = 4; Fig. 3C). These findings indicate that ischemic preconditioning acts upstream from GluR2 suppression, which could account for the resistance against ischemia-induced neuronal death.

Figure 3.

Sublethal ischemia prevents suppression of GluR2 protein abundance induced in CA1 by lethal ischemia. Representative Western blots probed with a mAb against the GluR2 subunit (Upper) and relative GluR2 subunit abundance (Lower) for protein samples isolated from the CA1 and dentate gyrus (DG) of gerbils subjected to sublethal ischemia (A), lethal ischemia (B), or sublethal ischemia followed by lethal ischemia (C).

Ischemic Preconditioning Enhances GluR2 Antisense-Induced Neuronal Death.

Administration of specific (but not scrambled) GluR2 antisense (four injections at 12-h intervals, GluR2 knockdown or lethal antisense) induces marked neuronal death of hippocampal pyramidal neurons (Fig. 4 C, D, and M, see ref. 4). To investigate whether ischemic preconditioning also acts downstream from GluR2 expression, we examined the ability of ischemic preconditioning to protect against GluR2 antisense-induced neuronal death. Ischemic preconditioning did not protect, but rather acted synergistically with subsequent lethal antisense (administered 48 h later) to induce neuronal death (≈75% loss of neurons for preconditioning followed by GluR2 antisense vs. ≈40% loss of neurons for antisense alone; Fig. 4 E, F, and M). These findings indicate that preconditioning acts upstream and not downstream from regulation of GluR2 expression. Moreover, the finding that sublethal ischemia and subsequent sublethal GluR2 antisense act synergistically to induce neuronal death indicates that the brief suppression of GluR2 mRNA induced by either treatment can summate to trigger neuronal death.

Figure 4.

Ischemic preconditioning enhances GluR2 antisense-induced neuronal death. GluR2 antisense at a sublethal dose does not protect against antisense- or ischemia-induced neuronal death. Toluidine blue-stained coronal brain sections at the level of the dorsal hippocampus from control (saline injected; n = 4; A and B) and experimental rats subjected to four intracerebroventricular injections of antisense (lethal antisense; n = 4; C and D), sublethal ischemia followed by lethal antisense 48 h later (n = 4; E and F), a single antisense intracerebroventricular injection (sublethal antisense; n = 4; G and H), sublethal antisense followed by lethal antisense 48 h later (n = 4; I and J), sublethal antisense followed by lethal ischemia 48 h later (n = 4; K and L). Animals were killed 7 days after the last injection or induction of ischemia. (M and N) Quantitation of cell counts from sections illustrated in A-L. (Scale bar: 400 μm in A, C, E, G, I, and K; 50 μm in B, D, F, H, J, and L.)

Sublethal GluR2 Antisense Does Not Prevent Antisense or Ischemia-Induced Neuronal Death.

A single, sublethal injection of GluR2 antisense induces a transient reduction in GluR2 mRNA expression and little or no neuronal death (Fig. 4 G, H, and N, see ref. 4). To determine whether that brief suppression of GluR2 mRNA expression is capable of inducing ischemic tolerance, we administered sublethal GluR2 antisense to rats followed after 48 h by lethal ischemia or lethal antisense, i.e., four injections at 12 h intervals. Sublethal antisense did not prevent the nearly total loss of neurons in CA1 induced by subsequent lethal ischemia (Fig. 4 K, L, and N). Moreover, sublethal antisense had no effect on the mortality of CA1 pyramidal neurons induced by subsequent lethal antisense (≈40% neuronal loss for sublethal antisense followed by lethal antisense vs. ≈40% neuronal loss for lethal antisense alone; Fig. 4 I, J, and N).

Discussion

The present study was undertaken to examine molecular mechanisms underlying ischemic tolerance. The principal findings of our study are as follows: (i) Ischemic preconditioning induces a small, transient suppression of GluR2 mRNA expression specifically in CA1 pyramidal neurons with no change in GluR2 protein expression. (ii) Ischemic preconditioning greatly attenuates lethal ischemia-induced down-regulation of GluR2 mRNA and protein. (iii) Ischemic preconditioning does not protect against antisense induced knockdown of GluR2. (iv) GluR2 antisense knockdown does not protect against lethal ischemia or lethal antisense. These findings suggest a molecular mechanism for induction of ischemic tolerance in CA1. Namely, ischemic preconditioning renders refractory the transcriptional machinery that leads to GluR2 down-regulation and thereby prevents expression of Ca²⁺-permeable AMPARs at CA1 synapses and AMPAR-mediated Ca²⁺ influx into CA1 neurons. Interestingly, estrogen at concentrations that afford protection against global ischemia-induced hippocampal injury does not suppress GluR2 down-regulation but blocks induction of the pro-apoptotic molecules, caspase-3 and p75 (50).

The finding that ischemic preconditioning affords robust neuroprotection and prevents GluR2 down-regulation is consistent with a role for glutamate receptor-dependent mechanisms in ischemia-induced death. Global ischemia triggers suppression of GluR2 gene expression and subunit abundance and enhances AMPAR-mediated Ca²⁺ influx in vulnerable CA1 pyramidal neurons before cell death (for review, see ref. 46). Ischemia induces prolonged, Ca²⁺-dependent AMPA excitatory postsynaptic currents at CA1 synapses, which are sensitive to Joro spider toxin and 1-naphthyl acetyl spermine, channel blockers selective for Ca²⁺-permeable AMPARs (47–49). These findings indicate expression of GluR2-lacking, Ca²⁺-permeable receptors at CA1 synapses and predict enhanced vulnerability of CA1 neurons to ambient glutamate (46, 51). Consistent with this model, knockdown of the GluR2 gene by administration of antisense oligonucleotides, even in the absence of an ischemic insult, causes death of pyramidal neurons (4). Antisense-induced cell death is fully prevented by 1-naphthyl acetyl spermine, indicating a role for Ca²⁺-permeable AMPARs.

GluR2 gene expression is under control of REST/NRSF, a gene silencing transcriptional factor, which acts at the RE1 (restrictive element 1) site in the proximal promoter of the GluR2 gene to suppress GluR2 expression (52, 53). Lethal ischemia initiates a signaling cascade that up-regulates REST and suppresses GluR2 transcription in CA1 neurons (A.C., S.Y.G., T.J., H.T., M.V.L.B., and R.S.Z., unpublished observations). These findings are consistent with a model whereby the early rise in Ca²⁺ triggered by ischemia activates a Ca²⁺-dependent CREB regulatory element in the REST promoter (54) to induce REST expression. Findings in the present study raise the possibility that preconditioning renders REST-mediated signaling refractory to a subsequent ischemic insult, thus preventing neuronal death.

Our findings of sublethal ischemia-induced changes in GluR2 mRNA and protein expression are consistent with the findings of others (11, 55). However, one group, which uses a different method to induce ischemia and measure mRNAs, reports nonspecific changes in AMPAR subtypes and rejects the hypothesis that GluR2 down-regulation is involved in delayed cell death (56–58).

In summary, the present study implicates changes in glutamate receptor expression in ischemic preconditioning and tolerance. An understanding of the molecular mechanisms underlying ischemic tolerance may help in the design of novel neuroprotective strategies for intervention in the neuronal death associated with stroke and other neurological disorders.

Abbreviations

AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid

AMPAR

AMPA receptor