Reduction in Intracellular Calcium Levels Induces Injury in Developing Neurons

Experimental Neurology 178, 21–32 (2002) doi:10.1006/exnr.2002.8027

Reduction in Intracellular Calcium Levels Induces Injury in Developing Neurons Christopher P. Turner*, Diane Pulciani*, and Scott A. Rivkees* ,† *Department of Pediatrics and †Interdepartmental Program in Neuroscience, Yale School of Medicine, YCHRC, 464 Congress Avenue, New Haven, Connecticut 06520 Received April 3, 2002; accepted July 19, 2002

neurons, glutamate action is mediated, in part, by increases in intracellular Ca 2⫹ levels (27). Maintaining intracellular Ca 2⫹ within a normal range is important for cell survival (22), as significantly increased or decreased intracellular Ca 2⫹ levels can lead to cell injury (6, 24, 31). Although altered NMDAR function may influence intracellular Ca 2⫹ concentrations (27), it is not known whether neuronal injury associated with NMDAR blockade in neonates is mediated through altered Ca 2⫹ signaling. Adenosine is a neuromodulator that can also influence neuronal survival and intracellular Ca 2⫹ levels (10, 44). Adenosine acts through a family of G-proteincoupled receptors, which includes the A 1 adenosine receptor (A 1AR) (13, 35). Activation of A 1ARs inhibits accumulation of cyclic adenosine monophosphate within cells and reduces intracellular Ca 2⫹ levels (9, 58). A 1ARs are the most heavily expressed adenosine receptor subtype in the central nervous system (CNS) and are present in nearly all brain regions (41, 47, 48). During development, A 1ARs are expressed in the brain as early as embryonic day 14 in rodents (40). High levels of A 1ARs are present in the CNS at the end of gestation, and A 1AR expression in the brain persists into adulthood (42, 55). An important effect of A 1AR activation is the ability to block the release of glutamate and reduce postsynaptic responses to glutamate receptor activation (7, 9, 39). Because NMDAR blockade induces neuronal cell death during early neonatal life (20), A 1AR activation may influence postnatal neuronal viability by modulating glutamate action. To address the idea of whether adenosine can influence CNS development by modulating the action of other neurochemicals, we have examined the influences of A 1AR activation and NMDAR blockade on neuronal survival in vitro and in vivo. We now report that the combined effects of A 1AR activation coupled with NMDAR blockade lead to profound cell death during the early neonatal period. Our observations also suggest that reductions in intracellular Ca 2⫹ play a role in the observed toxicity.

The neurotransmitter glutamate influences intracellular Ca 2ⴙ levels and plays an essential role in maintaining neuronal viability during early development. Blockade of NMDA receptors induces cell death in the neonatal forebrain via mechanisms that are not understood. Other neuromodulators that can influence intracellular Ca 2ⴙ levels include the nucleoside adenosine, which acts via A 1 adenosine receptors subtypes (A 1ARs). Because A 1AR activation inhibits glutamate release and action, A 1AR activation may also contribute to neonatal brain injury. To examine this possibility, we treated primary neuronal cultures with the A 1AR agonist CPA, the NMDAR antagonist MK801, or CPA ⴙ MK801. Combined MK801 ⴙ CPA treatment resulted in profound cellular injury, exceeding that seen in other groups. In keeping with the hypothesis that altered Ca 2ⴙ signaling mediates CPA ⴙ MK801 injury, reduction of Ca 2ⴙ levels with EGTA, thapsigargin, or BAPTA-AM enhanced CPA ⴙ MK801-induced neuronal damage. In contrast, increasing intracellular Ca 2ⴙ using ionomycin reversed CPA ⴙ MK801 toxicity. Direct visualization of intracellular Ca 2ⴙ by confocal microscopy revealed that CPA ⴙ MK801 inhibited KClevoked increases in intracellular Ca 2ⴙ. Supporting the concept that A 1AR activation and NMDAR blockade results in brain injury, neonatal rats injected with A 1AR agonists ⴙ MK801 showed widespread apoptosis in many brain regions. These observations show that A 1AR activation and NMDAR blockade lead to early postnatal cell injury by mechanisms that involve inhibition of intracellular Ca 2ⴙ signaling. © 2002 Elsevier Science (USA)

Key Words: TUNEL; caspase-3; confocal; neonatal; rat; apoptosis.

INTRODUCTION

Recent data show that the neurotransmitter glutamate plays an essential role in maintaining neuronal viability during early development, as blockade of Nmethyl-D-aspartate receptors (NMDAR) induces cell death throughout the forebrain of neonatal rats (20). In 21

©

0014-4886/02 $35.00 2002 Elsevier Science (USA) All rights reserved.

22

TURNER, PULCIANI, AND RIVKEES

MATERIALS AND METHODS

Neuronal Cultures The use of animals in all experiments was approved by the Yale Animal Care and Use Committee. Embryonic day 18 rat embryos were removed from the uteri of anesthetized dams (halothane, 1–2%). Cortical neurons were prepared from embryonic day 18 rat brains, as described (3). Embryos were kept on ice in Hanks’ balanced salt solution (no calcium or magnesium salts; 10 mM Hepes, 2 mM sodium pyruvate; Gibco/BRL, Rockville, MD). Whole brains were isolated, the meninges, hippocampus, and subcortical tissue were removed, and the remaining cortical tissue was placed into Hibernate E medium (supplemented with B27; Gibco/BRL). The cortical tissue was pooled, minced with a sterile razor blade, triturated gently, and centrifuged (200g, 1 min). The cell pellet was resuspended in Neurobasal medium (Gibco/BRL) containing 25 ␮M L-glutamate (Sigma), 0.5 mM L-glutamine, penicillin– streptomycin, and B27 supplements (Gibco/BRL). Cells were placed in poly-D-lysine-coated, 12-well plates (Sigma, St. Louis, MO; Costar, Corning, Inc., Corning, NY). Medium was changed on the third day after initial dispersion and every 7 days thereafter using glutamate-free Neurobasal medium. Isolated cortical neurons were grown in Neurobasal medium (Gibco/BRL) containing 25 ␮M L-glutamate (Sigma), 0.5 mM L-glutamine, penicillin–streptomycin, and B27 (Gibco/BRL) for the first 3 days, which was then replaced with glutamate-free Neurobasal medium for the remaining time in culture. Experiments were performed at 11 days in vitro. Each set of in vitro experiments was performed at least three times using different sets of neuronal cultures isolated from separate litters. To estimate the percentage of neurons in cultures, NeuN-immunocytochemistry was performed. After 11 days in vitro, cells were fixed with 4% PFA and incubated with a mouse anti-NeuN antibody (1:5000, Chemicon, Temecula, CA) overnight at 4°C. The pattern of primary antibody labeling was established using a biotinylated goat anti-mouse secondary (1:200), ABC Elite, and diaminobenzidine/NiCl staining kits (Vector, Burlingame, CA). NeuN-positive cells were expressed relative to total cell counts. In all cultures studied, we observed that more than 98% of cells were positive for this neuronal marker (56). Fluorescein Diacetate and Propidium Iodide Staining Neurons were exposed to vehicle or drugs for 24 h. Cells were then stained for 5 min with fluorescein diacetate to identify live cells (FDA, 2 ␮g/ml; Sigma) or with propidium iodide to identify dead cells (PI, 2 ␮g/ ml; Sigma). FDA is esterified by living cells, whereas PI

requires a loss of cell membrane integrity to penetrate and label cells (5). After incubation with FDA or PI, cells were washed three times with phosphate-buffered saline (PBS) and photographed at low magnification using UV fluorescence. Four separate fields per well were captured digitally. Each image was analyzed using software that readily distinguishes labeled cells from background (Image Pro software; Image Pro, Inc., Boston, MA). The mean number of FDA- or PI-positive cells was determined in triplicate wells per treatment group. Confocal Microscopy Neurons were plated on poly-D-lysine-coated, 22 ⫻ 40 mm cover slips, grown for 11 days in vitro. Neurons were loaded for 20 min with the Ca 2⫹-sensitive dye, Fluo-3-AM (25 ng/ml; Molecular Probes, Eugene, OR), plus Pluronic F127 (50 ng/ml; Molecular Probes). After coverslips were mounted into confocal chambers, Neurobasal medium was replaced with Hibernate E/B27 medium, which maintains neuronal viability for extended periods (2). Medium was exchanged when needed by suction from one side of a coverslip resting above the cells and fresh medium was added to the opposite side. This was performed rapidly (taking about 5 s) and repeated at least three times for complete washout. To determine direct effects of drugs on intracellular Ca 2⫹ levels, baseline Fluo-3 imaging was performed in the presence of vehicle or drug. Because intracellular Ca 2⫹ concentrations increase following depolarization (12), we examined changes in Fluo-3 intensity in cells stimulated with KCl (25 mM) in the presence or absence of drugs. After each drug addition and KCl stimulation, cells were washed with fresh medium and allowed to recover. KCl stimulation was then repeated to verify recovery of predrug responses before additional studies were performed. Cells were exposed to KCl for the duration of the recording period and recovery from KCl exposure was usually observed after the first wash. After the addition of KCl, changes in Fluo-3 intensity were monitored every 5 s over 90 s. Fluo-3 images were captured as multi-TIFF files and expressed as plots of intensity over time (FluoView; Olympus, Melville, NY). For each cell, changes in intensity over time were normalized to baseline, averaged for all cells, and the mean intensity at each time point was determined. The transformed data were then fitted to sigmoidal curves using GraphPad Prism v3.0 (GraphPad, San Diego, CA). The curves fitted the general equation Y ⫽ bottom ⫹ (top – bottom)/(1 ⫹ exp({V50-X)/slope}). Slopes and maximums were derived from these curves. Confocal imaging was performed on at least three separate fields of cells for each coverslip, in five separate stud-

INTRACELLULAR CALCIUM REDUCTION INJURES DEVELOPING NEURONS

ies, using neurons isolated from different litters for each study. In Vivo Animal Studies Pregnant females (Sprague–Dawley) were used. The day of birth was defined as postnatal day 0 (P0). Litters generally included 10 pups, and individual pups within each litter were randomly assigned to different treatment groups for drug injections. Pups were injected with drugs (subcutaneous) at various postnatal ages. After injections, pups were returned to their dam. At 24 h after injections pups were anesthetized (halothane, 1%) and decapitated, and brains were dissected and frozen on dry ice. Brains were stored at ⫺80°C until they were cut on a cryostat. TUNEL Assay Cryostat-cut, 20-␮m brain sections were thawmounted onto Super-Frost Plus slides (Fisher, Pittsburgh, PA) and stored at ⫺80°C. Terminal transferasedUTP- nick-end-labeling (TUNEL) was performed using the ApoTag kit procedure (Intergen, Purchase, NY). Briefly, sections were air-dried and fixed in 1% phosphate-buffered paraformaldehyde (PFA) for 10 min. Sections were then washed (three times in PBS) and postfixed in precooled ethanol/acetic acid (2:1) for 5 min at ⫺20°C. Following 5 min in equilibration buffer, sections were incubated for 1 h in a humidified chamber at 37°C with terminal transferase enzyme in reaction buffer. Sections were then washed with stop buffer for 10 min followed by 30 min with anti-digoxigenin peroxidase. Finally, the sections were stained with a diaminobenzidine/NiCl staining kit (Vector), washed in PBS, dehydrated in ascending alcohols, and mounted with coverslips. As a negative control, terminal transferase was omitted. TUNEL-positive cells were counted from digitally captured images at low magnification (Image Pro). Mean TUNEL counts for all sections sampled were expressed per square millimeter. TUNEL was also performed on cultured embryonic neurons and TUNELpositive cells were photographed digitally and quantified, as described for FDA and PI staining. Caspase-3 Immunohistochemistry Brain sections adjacent to those used in the TUNEL assay were processed for caspase-3 immunohistochemistry. Sections were fixed in 4% PFA for 10 min, followed by incubation with a rabbit polyclonal antibody to a synthetic amino acid sequence within the cleaved caspase-3 gene (1:50; Cell Signaling Technology, Beverly, MA). Sections were then incubated with a goat anti-rabbit, biotinylated secondary antibody (1:200; Vector) followed by ABC Elite kit (Vector). Antibody labeling was visualized with diaminobenzidine/NiCl

23

(Vector). Between each incubation, sections were washed three times (10 min each) in 0.1 M Tris-buffered saline (TBS). Antibodies were diluted in 0.1% BSA/0.01% Triton X-100/TBS (Sigma) and incubated with sections at room temperature for 1–2 h. As a negative control, primary antibody was omitted. Drugs The A1AR agonist, N 6-cyclopentyladenosine (CPA), the A 1AR-selective antagonist, 8-sulfophenyltheophilline (8SPT), the A1AR antagonist 8-phenyl-1,3-dipropylxanthine (DPCPX), the NMDA-specific antagonist dizocilpine (MK801), ethylene glycol-bis(␤-aminoethyl ether)N,N,N⬘,N⬘-tetraacetic acid (EGTA), thapsigargin, and halothane were obtained from Sigma Chemical Co. Ionomycin was obtained from Calbiochem (San Diego, CA). 1,2-Bis(O-aminophenoxy)ethane-N,N,N⬘,N⬘-tetraacetic acid-AM (BAPTA-AM) was obtained from Molecular Probes (Eugene, OR). The A1AR partial agonist 8-propylamino-N 6-cyclopentyladenosine was a gift from Dr. Adrian Ijzerman (Leiden University, The Netherlands). Most drugs were dissolved in dimethylsulfoxide (DMSO; Sigma) and diluted 1000-fold or more from stock solutions with PBS. DMSO without drugs was diluted 1000fold or more with PBS and served as the vehicle in all experiments. Data Analysis For FDA and PI analysis, data were normalized to vehicle treatments and the means were compared using ANOVA with a Bonferroni posttest comparison of means (GraphPad Prism v3.0). For TUNEL data, means were compared using ANOVA and a Bonferroni posttest comparison of means (GraphPad Prism v3.0). For confocal analysis, curves were derived, and slopes and maximums for each set of data were determined using GraphPad Prism v3.0. RESULTS

Effects of CPA and MK801 Treatment on Isolated Neurons Because adenosine can reduce glutamate action (9, 39), we tested whether A 1AR activation and NMDAR blockade can influence neuronal viability. In these studies, CPA was used to activate A 1ARs and MK801 was used to block NMDARs. Isolated cortical neurons were exposed to drug treatments for 24 h. To assess cell viability, cells were stained with FDA to identify living cells (Fig. 1A). In some experiments PI staining was used to assess numbers of dead cells (Fig. 1A). Quantitative analysis of FDA and PI labeling revealed concentration-dependent toxicity of CPA in the presence of MK801 (Figs. 1B and 1C). When combined with MK801, concentrations of CPA as low as 10 nM

24

TURNER, PULCIANI, AND RIVKEES

FIG. 1. Effects of A 1AR activation plus NMDAR blockade on isolated neurons. Isolated neurons were treated with various drugs for 24 h and stained with either FDA (A1–A4) or PI (A5–A8). Scale bar in A8 is 200 ␮m. Quantitative analysis revealed concentration-dependent decreases in FDA-stained cells (B), and an increase in PI-stained cells (C) when CPA (1–1000 nM) was combined with MK801 (100 ␮M). (D) Neurons were exposed to the A 1AR antagonist, DPCPX (1–100 nM) 30 min before CPA ⫹ MK801 (100 nM and 100 ␮M, respectively). DPCPX produced a concentration-dependent increase in FDA-positive cells (ns, not significant). Data in B, C, and D represent quadruplicate observations for each well performed in triplicate.

decreased FDA labeling and increased PI labeling (P⬍0.001 for FDA; P⬍0.05 for PI, compared to MK801 alone). CPA or MK801 alone produced only modest toxic effects on isolated neurons compared to vehicle (Fig. 1; P⬍0.05, P⬍0.01, respectively, for FDA; P⬎0.05 for PI). To assess the specificity of CPA action, cells were pretreated with different concentrations of the A 1AR antagonist DPCPX (1–100 nM) 30 min before CPA ⫹ MK801 were added. The studies above showed that significant toxicity could be achieved with 100 nM CPA and 100 ␮M MK801 and these concentrations were used in this and subsequent studies. We observed that the toxic effects of CPA ⫹ MK801 were reduced by DPCPX in a concentration-dependent manner (Fig. 1D). In addition to assessing FDA and PI staining, we performed TUNEL on neurons treated in the same manner as described above. CPA ⫹ MK801 resulted in a higher number of TUNEL-positive neurons, exceeding that seen with MK801 or CPA alone (TUNELpositive cells/mm 2, mean ⫾ SEM: vehicle ⫽ 138 ⫾ 17;

CPA (100 nM) ⫽ 198 ⫾ 16; MK801 (100 ␮M) ⫽ 220 ⫾ 17*; CPA (100 nM) ⫹ MK801 (100 ␮M) ⫽ 447 ⫾ 37*; p⬍0.001 compared to MK801). Thus, using three independent measurements of cell injury (FDA, PI, and TUNEL), we observed synergistic toxicity when CPA was combined with MK801. CPA-MK801 Toxicity Is Mediated by Changes in Relative Intracellular Ca 2⫹ Levels Having observed that A 1AR activation plus NMDAR blockade resulted in injury to isolated neurons, we next sought to identify potential mechanisms. Activation of A 1AR lowers intracellular Ca 2⫹ (58), whereas MK801 blocks NMDAR-mediated Ca 2⫹ flux (29, 52). We therefore tested whether changes in intracellular Ca 2⫹ levels mediate CPA–MK801 toxicity. First, we examined CPA–MK801 effects under conditions in which Ca 2⫹ levels were reduced. EGTA was used to sequester extracellular Ca 2⫹ (43), the Ca 2⫹ATPase pump inhibitor, thapsigargin, was used to deplete store-operated, intracellular Ca 2⫹ (21), and the

INTRACELLULAR CALCIUM REDUCTION INJURES DEVELOPING NEURONS

25

FIG. 2. Effects of changes in Ca 2⫹ levels on A 1AR activation and NMDAR blockade-induced toxicity. Cells were treated with EGTA (A), thapsigargin (B), BAPTA-AM (C), or ionomycin (D), with or without CPA ⫹ MK801. Other cells were treated with either CPA ⫹ MK801 or vehicle (Veh). After 24 h, cells were stained with FDA. When CPA ⫹ MK801 was combined with EGTA, thapsigargin, or BAPTA-AM, toxicity was greater than with CPA ⫹ MK801. In contrast, ionomycin reversed CPA ⫹ MK801 toxicity. In C: ⫺B, no BAPTA-AM, ⫹B; with BAPTA-AM (2 ␮g/ml). Data represent quadruplicate observations performed in triplicate.

intracellular Ca 2⫹ chelator, BAPTA-AM, was used to sequester cytosolic Ca 2⫹ (15). Cultured cortical neurons were exposed to vehicle, CPA ⫹ MK801 (100 nM and 100 ␮M, respectively), and EGTA (1 mM), thapsigargin (1 ␮M), or BAPTA-AM (1–10 ␮g/ml), with or without CPA ⫹ MK801. After 24 h, FDA staining was performed We observed that EGTA enhanced CPA ⫹ MK801 toxicity when compared with CPA ⫹ MK801 (P⬍0.001) or EGTA alone (P⬍0.001) (Fig. 2A). When thapsigargin was combined with CPA ⫹ MK801, neuronal survival was greatly reduced compared to CPA ⫹ MK801 (P⬍0.001) or thapsigargin alone (p⬍0.001) (Fig. 2B). When BAPTA-AM (2 ␮g/ml) was combined with CPA (100 nM) ⫹ MK801 (100 ␮M), toxicity was greater than that seen with CPA ⫹ MK801 (p⬍0.001) or BAPTA-AM alone at 2 ␮g/ml (P⬍0.001) (Fig. 2C). Compared to vehicle, both EGTA and BAPTA (2 and 10 ␮g/ml) lowered FDA labeling (P⬍0.01), whereas thapsigargin had little effect (P⬎0.05). EGTA reduces extracellular Ca2⫹ and thapsigargin and BAPTA reduce intracellular Ca2⫹, which may explain why EGTA effects appear additive whereas those of thapsigargin and BAPTA appear synergistic.

We then tested whether increasing intracellular Ca 2⫹ levels affected the observed toxicity. Cells were treated with vehicle, CPA ⫹ MK801 (100 nM and 100 ␮M, respectively), ionomycin (0.1 ␮M for 5 min), or CPA ⫹ MK801 ⫹ ionomycin. After 24 h, cells were stained with FDA. When ionomycin was combined with CPA ⫹ MK801, there was increased cell survival compared to CPA ⫹ MK801 treatment in the absence of ionomycin (P⬍0.01) (Fig. 2D). Ionomycin alone had little effect on cell viability compared to vehicle (P⬎0.05). CPA and MK801 Influences on Relative Intracellular Ca 2⫹ Concentrations The above studies suggest that CPA ⫹ MK801 toxicity is dependent upon relative intracellular Ca 2⫹ levels. Thus, to directly visualize the effects of CPA and MK801 on Ca 2⫹ signaling, confocal imaging was performed on isolated neurons using the Ca 2⫹-sensitive dye, Fluo-3-AM. Cells were preloaded with Fluo-3-AM for 30 min. Individual neurons were identified within a microscopic field (Fig. 3A) and continuously observed. First,

FIG. 3. Changes in intracellular Ca 2⫹ levels following drug treatment. KCl-stimulated, Fluo-3 signaling in the presence of vehicle at beginning of recording (A) or at the maximum of Fluo-3 intensity (B). In vehicle-treated neurons, KCl stimulation produced a sigmoidal increase in Fluo-3 intensity (C). Reductions in KCl-stimulated signaling were seen following CPA or MK801 treatments (C). However, no signal was obtained when cells were treated with CPA ⫹ MK801 (C). Curves represent single cell observations normalized to baseline and averaged across all cells within each time point (N ⫽ 7 cells in this study). Scale bar in B, 20 ␮m. A relative scale of Fluo-3 intensity in arbitrary units (0, lowest, to 100, highest) is shown (A).

26

TURNER, PULCIANI, AND RIVKEES

FIG. 4. Distribution of TUNEL in the striatum and thalamic regions 24 h following drug treatment. Coronal cut hemi-sections from brains of P7 neonates following vehicle (A and C) or MK801 (B and D) treatments. Sections taken at the level of the striatum (A and B) or thalamus (C and D). Note intense staining in MK801-treated animals in the striatum (arrow in B), cerebral cortex (arrowheads in D), dorsal thalamus (arrow in D), and hypothalamic regions. High-power images of TUNEL-positive cells are shown (E1 from thalamus, E2 from striatum). (F) Negative control section (terminal transferase omitted) from an MK801-treated P7 neonate. (G) Quantitative analysis of TUNEL in the dorsal thalamus from vehicle- (Veh), 8PCPA- (PA), MK801- (MK), or MK801 ⫹ 8PCPA-treated animals at different ages. Scale bars: 1mm (A–D, F); 20 ␮m (E1); 10 ␮m (E2). N ⫽ 4 for vehicle-injected pups, N ⫽ 6 for all other treatment groups.

cells were exposed to KCl (25 mM) to induce depolarization in the presence of vehicle (Fig. 3B). After a washout period, CPA, MK801, or CPA ⫹ MK801 was added for 5 min before KCl stimulation. Between each drug treatment, cells were washed with fresh medium and predrug responses to KCl in the presence of vehicle were verified. In vehicle-treated cells, KCl induced a rapid increase in Fluo-3 intensity (slope ⫽ 4.3 ⫾ 0.2), which reached a maximum (310.9 ⫾ 1.2) at the plateau phase (Fig. 3C). Following CPA treatment, KCl-stimulated neurons displayed a less pronounced rapid phase (slope ⫽ 6.8 ⫾ 0.4) and a reduction in the maximum (275.1 ⫾ 1.4) at the plateau (Fig. 3C). Following MK801 (100 ␮M)

treatment, there was a reduction in the KCl-induced rapid phase (slope ⫽ 10.3 ⫾ 1.1) and the maximum (158.9 ⫾ 2.1) at the plateau of Fluo-3 signaling (Fig. 3C). However, the combination of CPA ⫹ MK801 completely blocked KCl-evoked, Fluo-3 signaling (slope ⫽ -0.1 ⫾ 0.0, maximum ⫽ 98.4 ⫾ 0.4; Fig. 3C). To test for possible order effects, drug treatments were repeated in a random order, and the same results were consistently observed. In addition, following KCl stimulation, Fluo-3 intensity rapidly returned to resting levels when neurons were placed back in normal medium. Finally, CPA and MK801 both reduced baseline Fluo-3 intensity in the absence of KCl (data not shown).

INTRACELLULAR CALCIUM REDUCTION INJURES DEVELOPING NEURONS

A 1AR Activation Enhances MK801-Induced Apoptosis in the Neonatal Brain Having established that A 1AR activation and NMDAR receptor blockade is toxic to isolated neurons, we next assessed whether these treatments influence cell survival in vivo. We used two different approaches to test whether A 1AR activation in whole animals could produce toxicity or enhance MK801-induced toxicity. In the first set of studies, pups of different ages were injected with the A 1AR partial agonist 8PCPA, MK801, or 8PCPA ⫹ MK801. Vehicle-injected animals served as controls (N ⫽ 4 for vehicle-injected pups, N ⫽ 6 for all other treatments). 8PCPA has minimal cardiovascular effects, and 2mg/kg is sufficient to activate A 1ARs (53). MK801 was used at a dose of 1 mg/kg, which results in neuronal injury in neonates (20). Pups were injected at postnatal day 4 (P4), P7, or P21. Twentyfour hours later, brains were examined for patterns of TUNEL. At 24 h after injections, intense TUNEL was observed in the cortical, thalamic, and hypothalamic regions (Figs. 4A– 4F). Cell injury was consistently found in the thalamus, particularly the dorsal nuclei. Therefore, we focused on this brain region for quantitative analysis. Every third section was analyzed to determine numbers of TUNEL-positive cells throughout the thalamus, approximating ⫺1.60 to ⫺4.52 relative to Bregma (37). At all ages (P4, P7, and P21), vehicle-injected rats displayed the least number of TUNEL-positive neurons (Fig. 4G). At P4 and P7, 8PCPA treatment increased TUNEL above that seen with vehicle treatment (Fig. 4G; P⬍0.05 at P4 and P7). At P21, TUNEL following 8PCPA treatment was the same as that seen with vehicle treatment. In contrast, MK801 induced robust TUNEL at P4 and P7 compared to vehicle (P⬍0.001 at P4 and P7), whereas by P21 levels of TUNEL were the same as in vehicle-treated animals (Fig. 4G). However, we found that 8PCPA ⫹ MK801 produced TUNEL significantly higher than that seen with MK801 alone (Fig. 4G), at either P4 (P⬍0.001) or P7 (P⬍0.001). By P21, 8PCPA ⫹ MK801 treatment displayed similar levels of TUNEL as that seen with vehicle injections (Fig. 4G). In our in vitro studies we had used CPA to activate A 1ARs. Thus, in our second set of studies, we used this drug to activate A 1ARs in vivo. CPA effects on the CNS are well documented (10) and it readily reaches the brain after peripheral injection (50). Because CPA can influence hemodynamic status (45, 49), 8SPT was given with CPA to block peripheral A 1AR activation. 8SPT is a charged xanthine that is excluded from the brain (54). As above, MK801 was used to block NMDARs. Because P4 and P7 animals were sensitive to A 1AR activation and NMDAR blockade, we focused on these ages. In these additional studies we examined both

27

TUNEL and caspase-3 immunoreactivity (caspase-3-ir) to confirm that cell injury was apoptotic in nature (57). Animals were given single subcutaneous injections of vehicle, CPA (0.1 mg/kg), or 8SPT (20 mg/kg) (N ⫽ 3 for each group). Other animals received MK801 (1 mg/ kg), CPA ⫹ MK801, or CPA ⫹ MK801 ⫹ 8SPT (N ⫽ 6 for each group). The CPA dose used has been shown to effectively activate A 1ARs in vivo (8). The 8SPT dose has been shown to effectively block peripheral A 1ARs in vivo (54). Using similar approaches to those describe here, caspase-3 expression reaches a broad peak between 6 and 16 h, but returns to basal levels by 24 h (31, 36). In contrast, TUNEL is lower at 8 h but peaks 24 h following MK801 treatment in neonates (20). Thus, brain sections were analyzed at 8 or 24 h after injections for TUNEL or caspase-3-ir. At 8 h, modest TUNEL and robust caspase-3-ir was observed in MK801-, CPA ⫹ MK801-, or CPA ⫹ MK801 ⫹ 8SPT-treated animals (Table 1; see Figs. 5 and 6), whereas only light to modest TUNEL and light caspase-3-ir were observed in vehicle-, CPA-, or 8SPTtreated animals (Table 1; Figs. 5 and 6). At 24 h after injection, robust TUNEL but light caspase-3-ir was observed in MK801-, CPA ⫹ MK801-, or CPA ⫹ MK801 ⫹ 8SPT-treated animals (Table 1; Figs. 5 and 6). Vehicle-, CPA-, or 8SPT-treated animals had only light to modest TUNEL and light caspase-3-ir (Table 1; Figs. 5 and 6). Importantly, 8SPT did not block either caspase-3-ir (at 8 h) or TUNEL (at 24 h) following CPA ⫹ MK801 injections. DISCUSSION

Based on in vitro and in vivo observations, we find that A 1AR activation coupled with NMDAR blockade results in profound injury to the developing brain. Our findings suggest that the effects of A 1AR activation and NMDAR blockade appear to be mediated by reductions in intracellular Ca 2⫹ levels. Furthermore, we identify the early neonatal period as a time when the developing brain is especially susceptible to this form of neuronal injury. It has previously been shown that exposure of isolated neurons to MK801 results in cell death (19). Similarly, we also observed MK801 to be toxic to neurons. However, the adverse effects of CPA ⫹ MK801 treatment on isolated neurons were much more pronounced than that observed with CPA or MK801 treatment alone, indicating synergistic effects. Demonstrating that the observed effects are A 1AR-mediated, the A 1AR antagonist DPCPX reversed CPA–MK801 toxicity in a concentration-dependent manner. We also observed that CPA contributed to MK801-induced cell death at concentrations as low as 10 nM, which is consistent with concentrations that activate A 1ARs (13, 35). At the concentration used in this study we

28

TURNER, PULCIANI, AND RIVKEES

TABLE 1 Distribution of TUNEL and Caspase-3 Staining in Forebrain Structures Caspase-3 (8 h)

Cerebral cortex Striatum Hippocampus Thalamus

Vehicle

CPA

8SPT

MK801

CPA–MK801

CPA–MK801–8SPT

⫹ ⫹ ⫹ ⫹

⫹ ⫹⫹ ⫹ ⫹⫹

⫹ ⫹ ⫹ ⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

TUNEL (24 h)

Cerebral cortex Striatum Hippocampus Thalamus

Vehicle

CPA

8SPT

MK801

CPA–MK801

CPA–MK801–8SPT

⫹ ⫹ ⫹ ⫹

⫹ ⫹⫹ ⫹ ⫹⫹

⫹ ⫹ ⫹ ⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹

Note. TUNEL assessment: ⫹ light (0 –5 TUNEL-positive cells per section); ⫹⫹, modest (6 –10 TUNEL-positive cells per section); ⫹⫹⫹, robust (⬎10 TUNEL-positive cells per section). Caspase-3 immunoreactivity: ⫹, light (0 –10 positive cells per section); ⫹⫹, modest (11–20 cells per section); ⫹⫹⫹, robust (⬎20 cells per section). Data shown are taken from brains of P4 neonates; N ⫽ 3 for vehicle-, CPA-, and 8SPT-injected animals; N ⫽ 6 for all other groups.

acknowledge that MK801 may have actions not related to blockade of NMDARs. Our observations suggest that CPA-MK801 toxicity is mediated by a perturbation of intracellular Ca2⫹ signaling. First, reducing extracellular Ca2⫹ with EGTA enhanced CPA ⫹ MK801 toxicity. Second, disruption of intracellular Ca2⫹ stores with the Ca 2⫹- ATPase pump inhibitor thapsigargin markedly enhanced CPA–MK801induced toxicity. Third, chelation of cytosolic Ca2⫹ with BAPTA-AM enhanced CPA–MK801-induced toxicity. In contrast, raising intracellular Ca2⫹, by briefly exposing neurons to the Ca2⫹-ionophore ionomycin, reversed CPA– MK801-induced toxicity. We observed that ionomycin did not completely reverse CPA–MK801 toxicity, raising the possibility that ionomycin did not normalize Cai2⫹ or that Ca 2⫹-independent events may also play a role in CPA– MK801 injury. These approaches provide indirect evidence that changes in Ca i2⫹ levels mediate neuronal injury but are consistent with the hypothesis that combined A 1AR activation and NMDAR blockade lower intracellular calcium to produce toxicity. Direct support of this hypothesis for altered Ca 2⫹ signaling came from confocal studies, as KCl-evoked changes in Fluo-3 intensity were completely abolished by the combination of CPA ⫹ MK801. We were surprised by the degree to which combined CPA ⫹ MK801 inhibited KCl-evoked Fluo-3 changes, as depolarization is thought to lead to entry of Ca 2⫹ via voltage-gated calcium channels. It is possible that the changes in Fluo-3 intensity we observed are, at least in part, determined by changes in subcellular pools of Ca 2⫹ (such as those found in mitochondria) that may be sensitive to the drug treatments used in this study (32).

It has been hypothesized that, for continued cell viability, there is a critical level, or “set point,” at which intracellular Ca 2⫹ concentrations must be maintained (22). Necrotic cell death is thought to occur when Ca 2⫹ levels exceed this set point, whereas apoptotic cell death is induced when levels fall below the critical level (59). Following excessive glutamate release, which triggers Ca 2⫹ overloading of cells, necrotic cell death predominates (6, 23). In contrast, lowering intracellular Ca 2⫹ can induce neuronal cell death (24) and is associated with increased expression of caspase-3, a pro-apoptotic protein (31). Our data are consistent with the calcium set point hypothesis, as the loss in cell viability induced by CPA ⫹ MK801 is enhanced when Ca 2⫹ levels are reduced. In contrast, raising intracellular Ca 2⫹ prevented CPA–MK801-induced cell death. In neonates, MK801-induced injury is highest in the first week of neonatal life (20), which we also observed. However, using two different approaches to activate central A 1ARs in vivo, we find that A 1AR activation enhances MK801-induced cell damage in neonates, demonstrating a direct effect of A 1AR agonists on the CNS, as suggested by others (26, 54). Supporting the idea that cell injury was apoptotic in nature, we observed expression of caspase-3, a marker of programmed cell death, associated with increased signs of DNA fragmentation (TUNEL). TUNEL-positive cells were consistently found in the dorsal thalamic nuclei, suggesting that this region is particularly sensitive to brain injury. In neonates, thalamic injury has been observed following hypoxia and ischemia (34), supporting the notion that this is a developmentally sensitive brain area. The thalamus is a critical brain region that is involved in processing ascending sensory information and organizing complex

INTRACELLULAR CALCIUM REDUCTION INJURES DEVELOPING NEURONS

29

FIG. 5. TUNEL induced by A 1AR activation and NMDAR blockade in the thalamus. Coronal sections from the thalamus showing TUNEL staining at 8 and 24 h after injecting P4 or P7 neonates with vehicle (A, D, G, J), MK801 (B, E, H, K), or CPA ⫹ MK801 ⫹ 8SPT (C, F, I, L). Robust TUNEL was observed 24 h after CPA–MK801– 8SPT injections at both P4 (F) and P7 (L). Quantitative analysis (see Fig. 4G) revealed that the A 1AR activation enhanced MK801-induced cell damage. However, 8SPT did not block the action of CPA ⫹ MK801 (compare E with F and K with L). Scale bar in L is 250 ␮m. N ⫽ 3 for vehicle-, CPA-, and 8SPT-injected animals, N ⫽ 6 for all other groups.

motor activity (28) and may play a role in the pathogenesis of psychiatric disorders (4). Thus, determining whether altered A 1AR and NMDAR activity during critical developmental periods leads to long-term changes in sensory processing, motor control, or behavior will be important in future studies. Interestingly, whereas MK801 protects neurons from injury in the mature brain (46), NMDAR antagonism is toxic to the developing brain (20). Similarly, whereas A 1AR activation reduces or prevents brain injury in mature animals (38), we show that A 1AR activation promotes or enhances brain injury in neonates. It has been suggested that A 1AR activation is not neuroprotective in the young due to a lack of A 1AR– G-protein coupling in neonates (1). However, as suggested by others (17, 33), we find that functional A 1AR–G protein coupling is indeed present in neonates at the same ages studied here (51). During development, neuronal cell numbers are naturally reduced by apoptotic mechanisms, which are active in the first postnatal week in the rat (16). It is therefore possible

that during early postnatal life, this “cell-pruning” process can be enhanced by altered neuronal transmission or by traumatic injury (18). We recognize that activation of A 1ARs and blockade of NMDARs is unlikely to occur under normal physiological conditions. However, during hypoxia, hypoglycemia, or brain trauma, adenosine levels rise dramatically (11, 14, 25). Acting through A 1ARs, adenosine can inhibit glutamate action by blocking glutamate release or attenuating NMDAR-mediated action (7, 9, 39), leading to a potential cascade of events seen when NMDAR action is blocked in neonates. Furthermore, efforts are now directed at developing drugs for treating neonatal brain injury, some of which include NMDAR antagonists (30). Our data suggest that if NMDAR antagonists are administered when adenosine levels are high, rather than being protective, there is actually a heightened risk of neonatal brain injury. Overall, our data demonstrate that increased A 1AR activation and decreased NMDAR action can lead to robust cell death during early postnatal periods. Our

30

TURNER, PULCIANI, AND RIVKEES

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8. 9.

FIG. 6. Caspase-3 immunoreactivity in the thalamus following A 1AR activation and NMDAR blockade. Coronal sections from the thalamus showing caspase-3 immunoreactivity at 8 (A, B, E, F) and 24 h (C, D, G, H) following injection of P4 or P7 neonates with vehicle or CPA ⫹ MK801 ⫹ 8SPT. Robust caspase-3 immunoreactivity was observed at 8 h (B and F) but not 24 h (D and H) following CPA– MK801– 8SPT. Vehicle-injected animals showed little or no caspase-3 immunoreactivity at all times studied (A, C, E, G). As with TUNEL analysis (Fig. 5), cell injury was still observed when A 1AR activation was combined with NMDAR blockade while in the presence of the A 1AR antagonist 8SPT (B and F). Inset in G: negative assay control (primary antibody omitted). Scale bar in H is 250 ␮m. N ⫽ 3 for Vehicle-, CPA-, and 8SPT-injected animals, N ⫽ 6 for all other groups.

10.

11.

12.

13.

14.

data suggest that the mechanisms of CPA-MK801 toxicity involve reduction in relative intracellular Ca 2⫹ concentrations. These observations demonstrate that the brain may be vulnerable to this form of injury during the early neonatal period.

ACKNOWLEDGMENTS This work was supported by NIH Grant RO1-NS33439, the Donaghue Medical Research Foundation, and the Fannie F. Ripple Foundation Cell Imaging Facility.

15.

16.

17.

Aden, U., K. Lindstrom, E. Bona, H. Hagberg, and B. B. Fredholm. 1994. Changes in adenosine receptors in the neonatal rat brain following hypoxic ischemia. Brain Res. Mol. Brain Res. 23: 354 –358. Brewer, G. J., and P. J. Price. 1996. Viable cultured neurons in ambient carbon dioxide and hibernation storage for a month. Neuroreport 7: 1509 –1512. Brewer, G. J., J. R. Torricelli, E. K. Evege, and P. J. Price. 1993. Optimized survival of hippocampal neurons in B27-supplemented Neurobasal, a new serum-free medium combination. J. Neurosci. Res. 35: 567–576. Byne, W., M. S. Buchsbaum, L. A. Mattiace, E. A. Hazlett, E. Kemether, S. L. Elhakem, D. P. Purohit, V. Haroutunian, and L. Jones. 2002. Postmortem assessment of thalamic nuclear volumes in subjects with schizophrenia. Am. J. Psychiatry 159: 59 – 65. Chen, K., and M. D. Maines. 2000. Nitric oxide induces heme oxygenase-1 via mitogen-activated protein kinases ERK and p38. Cell. Mol. Biol. (Noisy-le-grand) 46: 609 – 617. Choi, D. W., M. Maulucci-Gedde, and A. R. Kriegstein. 1987. Glutamate neurotoxicity in cortical cell culture. J .Neurosci. 7: 357–368. de Mendonca, A., and J. A. Ribeiro. 1993. Adenosine inhibits the NMDA receptor-mediated excitatory postsynaptic potential in the hippocampus. Brain Res. 606: 351–356. Dragunow, M. 1991. Adenosine and seizure termination. Ann. Neurol. 29: 575. Dunwiddie, T. V., and B. B. Fredholm. 1989. Adenosine A1 receptors inhibit adenylate cyclase activity and neurotransmitter release and hyperpolarize pyramidal neurons in rat hippocampus. J. Pharmacol. Exp. Ther. 249: 31–37. Dunwiddie, T. V., and S. A. Masino. 2001. The role and regulation of adenosine in the central nervous system. Annu. Rev. Neurosci. 24: 31–55. Fowler, J. C. 1993. Purine release and inhibition of synaptic transmission during hypoxia and hypoglycemia in rat hippocampal slices. Neurosci. Lett. 157: 83– 86. Franklin, J. L., and E. M. Johnson, Jr. 1992. Suppression of programmed neuronal death by sustained elevation of cytoplasmic calcium. Trends Neurosci. 15: 501–508. Fredholm, B. B., M. P. Abbracchio, G. Burnstock, J. W. Daly, T. K. Harden, K. A. Jacobson, P. Leff, and M. Williams. 1994. Nomenclature and classification of purinoceptors. Pharmacol. Rev. 46: 143–156. Hagberg, H., P. Andersson, J. Lacarewicz, I. Jacobson, S. Butcher, and M. Sandberg. 1987. Extracellular adenosine, inosine, hypoxanthine, and xanthine in relation to tissue nucleotides and purines in rat striatum during transient ischemia. J. Neurochem. 49: 227–231. Han, K. S., H. J. Kang, E. Y. Kim, W. J. Yoon, S. Sohn, H. J. Kwon, and B. J. Gwag. 2001. 1,2-bis(2-Aminophenoxy)ethaneN,N,N⬘,N⬘-tetraacetic acid induces caspase-mediated apoptosis and reactive oxygen species-mediated necrosis in cultured cortical neurons. J. Neurochem. 78: 230 –239. Haydar, T. F., C. Y. Kuan, R. A. Flavell, and P. Rakic. 1999. The role of cell death in regulating the size and shape of the mammalian forebrain. Cereb. Cortex 9: 621– 626. Hettinger, B. D., M. Leid, and T. F. Murray. 1998. Cyclopentyladenosine-induced homologous down-regulation of A1 adenosine receptors (A1AR) in intact neurons is accompanied by

INTRACELLULAR CALCIUM REDUCTION INJURES DEVELOPING NEURONS

18.

19.

20.

21. 22.

23.

24.

25.

26.

27.

28. 29.

30.

31.

32. 33.

34.

35.

receptor sequestration but not a reduction in A1AR mRNA expression or G protein alpha-subunit content. J. Neurochem. 71: 221–230. Hu, B. R., C. L. Liu, Y. Ouyang, K. Blomgren, and B. K. Siesjo. 2000. Involvement of caspase-3 in cell death after hypoxia– ischemia declines during brain maturation. J. Cereb. Blood Flow Metab. 20: 1294 –1300. Hwang, J. Y., Y. H. Kim, Y. H. Ahn, M. B. Wie, and J. Y. Koh. Y. 1999. N-Methyl-D-aspartate receptor blockade induces neuronal apoptosis in cortical culture. Exp. Neurol. 159: 124 –130. Ikonomidou, C., F. Bosch, M. Miksa, P. Bittigau, J. Vockler, K. Dikranian, T. I. Tenkova, V. Stefovska, L. Turski, and J. W. Olney. 1999. Blockade of NMDA receptors and apoptotic neurodegeneration in the developing brain. Science 283: 70 –74. Inesi, G., and Y. Sagara. 1994. Specific inhibitors of intracellular Ca2⫹ transport ATPases. J. Membr. Biol. 141: 1– 6. Johnson, E. M., Jr., T. Koike, and J. Franklin. 1992. A “calcium set-point hypothesis” of neuronal dependence on neurotrophic factor. Exp. Neurol. 115: 163–166. Ko, H. W., K. Y. Park, H. Kim, P. L. Han, Y. U. Kim, B. J. Gwag, and E. J. Choi. 1998. Ca2⫹-mediated activation of c-Jun Nterminal kinase and nuclear factor kappa B by NMDA in cortical cell cultures. J. Neurochem. 71: 1390 –1395. Koh, J. Y., and C. W. Cotman. 1992. Programmed cell death: its possible contribution to neurotoxicity mediated by calcium channel antagonists. Brain Res. 587: 233–240. Koos, B. J., L. Kruger, and T. F. Murray. 1997. Source of extracellular brain adenosine during hypoxia in fetal sheep. Brain Res. 778: 439 – 442. Lorenzen, A., A. M. Sebastiao, A. Sellink, H. Vogt, U. Schwabe, J. A. Ribeiro, and I. J. Ap. 1997. Biological activities of N6,C8disubstituted adenosine derivatives as partial agonists at rat brain adenosine A1 receptors. Eur. J. Pharmacol. 334: 299 – 307. MacDermott, A. B., M. L. Mayer, G. L. Westbrook, S. J. Smith, and J. L. Barker. 1986. NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones. Nature 321: 519 –522. McCormick, D. A., T. and Bal. 1994. Sensory gating mechanisms of the thalamus. Curr. Opin. Neurobiol. 4: 550 –556. Melena, J., and N. N. Osborne. 2001. Voltage-dependent calcium channels in the rat retina: involvement in NMDA-stimulated influx of calcium. Exp. Eye Res. 72: 393– 401. Mishra, O. P., K. I. Fritz, and M. Delivoria-Papadopoulos. 2001. NMDA receptor and neonatal hypoxic brain injury. Ment. Retard. Dev. Disabil. Res. Rev. 7: 249 –253. Moran, J., T. Itoh, U. R. Reddy, M. Chen, E. S. Alnemri, and D. Pleasure. 1999. Caspase-3 expression by cerebellar granule neurons is regulated by calcium and cyclic AMP. J. Neurochem. 73: 568 –577. Nicholls, D. G., and S. L. Budd. 1998. Neuronal excitotoxicity: the role of mitochondria. Biofactors 8: 287–299. Nicolas, F., J. Oillet, V. Koziel, and J. L. Daval. 1994. Characterization of adenosine receptors in a model of cultured neurons from rat forebrain. Neurochem. Res. 19: 507–515. Northington, F. J., D. M. Ferriero, D. L. Flock, and L. J. Martin. 2001. Delayed neurodegeneration in neonatal rat thalamus after hypoxia- ischemia is apoptosis. J. Neurosci. 21: 1931–1938. Olah, M. E., and G. L. Stiles. 1995. Adenosine receptor subtypes: Characterization and therapeutic regulation. Annu. Rev. Pharmacol. Toxicol. 35: 581– 606.

36.

37. 38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49. 50.

51.

52.

53.

54.

31

Olney, J. W., T. I. Tenkova, D. M. Holtzman, and K. A. Roth. 2001. Ethanol triggers widespread caspase-3 activation and apototic neurodegeneration in wild type but not BAX-deficient mice. Soc. Neurosci. Abs. 27, 655.8. Paxinos, G., and C. Watson. 1998. The Rat Brain in Stereotaxic Coordinates, 4th ed. Academic Press, San Diego. Phillis, J. W., and H. G. Goshgarian. 2001. Adenosine and neurotrauma: Therapeutic perspectives. Neurol. Res. 23: 183– 189. Ribeiro, J. A. 1995. Purinergic inhibition of neurotransmitter release in the central nervous system. Pharmacol. Toxicol. 77: 299 –305. Rivkees, S. A. 1995. The ontogeny of cardiac and neural A1 adenosine receptor expression in rats. Brain Res. Dev. Brain Res. 89: 202–213. Rivkees, S. A., S. L. Price, and F. C. Zhou. 1995. Immunohistochemical detection of A1 adenosine receptors in rat brain with emphasis on localization in the hippocampal formation, cerebral cortex, cerebellum, and basal ganglia. Brain Res. 677: 193–203. Rivkees, S. A., Z. Zhao, G. Porter, and C. Turner. 2001. Influences of adenosine on the fetus and newborn. Mol. Genet. Metab. 74: 160 –171. Rogawski, M. A. 1986. Single voltage-dependent potassium channels in cultured rat hippocampal neurons. J. Neurophysiol. 56: 481– 493. Schubert, P., T. Ogata, C. Marchini, S. Ferroni, and K. Rudolphi. 1997. Protective mechanisms of adenosine in neurons and glial cells. Ann. NY Acad. Sci. 825: 1–10. Shryock, J. C., and L. Belardinelli. 1997. Adenosine and adenosine receptors in the cardiovascular system: Biochemistry, physiology, and pharmacology. Am. J. Cardiol. 79: 2–10. Simon, R. P., J. H. Swan, T. Griffiths, and B. S. Meldrum. 1984. Blockade of N-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Science 226: 850 – 852. Swanson, T. H., J. A. Drazba, and S. A. Rivkees. 1995. Adenosine A1 receptors are located predominantly on axons in the rat hippocampal formation. J. Comp. Neurol. 363: 517–531. Swanson, T. H., S. E. Krahl, Y. Z. Liu, J. A. Drazba, and S. A. Rivkees. 1998. Evidence for physiologically active axonal adenosine receptors in the rat corpus callosum. Brain Res. 784: 188 –198. Tabrizchi, R., and S. Bedi. 2001. Pharmacology of adenosine receptors in the vasculature. Pharmacol. Ther. 91: 133–147. Trivedi, B. K., C. J. Blankley, J. A. Bristol, H. W. Hamilton, W. C. Patt, W. J. Kramer, S. A. Johnson, R. F. Bruns, D. M. Cohen, and M. J. Ryan. 1991. N6-substituted adenosine receptor agonists: potential antihypertensive agents. J. Med. Chem. 34: 1043–1049. Turner, C. P., H. Yan, T. Othman, M. Schwartz, T. Othman, and S. A. Rivkees. 2002. A1 adenosine receptor activation induces ventriculomegaly and white matter loss. Neuroreport 13(9): 1199 –1204. Uematsu, D., J. H. Greenberg, N. Araki, and M. Reivich. 1991. Mechanism underlying protective effect of MK-801 against NMDA-induced neuronal injury in vivo. J. Cereb. Blood Flow Metab. 11: 779 –785. Van Schaick, E. A., R. A. Math-ot, J. M. Gubbens-Stibbe, M. W. Langemeijer, H. C. Roelen, A. P. Ijzerman, and M. Danhof. 1997. 8-Alkylamino-substituted analogs of N6-cyclopentyladenosine are partial agonists for the cardiovascular adenosine A1 receptors in vivo. J. Pharmacol. Exp. Ther. 283: 800 – 808. von Lubitz, D. K., and P. J. Marangos. 1990. Cerebral ischemia in gerbils: Postischemic administration of cyclohexyl adenosine and 8-sulfophenyl-theophylline. J. Mol. Neurosci. 2: 53–59.

32 55.

TURNER, PULCIANI, AND RIVKEES

Weaver, D. R. 1996. A1-adenosine receptor gene expression in fetal rat brain. Brain Res. Dev. Brain Res. 94: 205–223. 56. Wolf, H. K., R. Buslei, R. Schmidt-Kastner, P. K. SchmidtKastner, T. Pietsch, O. D. Wiestler, and I. Bluhmke. 1996. NeuN: A useful neuronal marker for diagnostic histopathology. J. Histochem. Cytochem. 44: 1167–1171. 57. Woo, M., R. Hakem, M. S. Soengas, G. S. Duncan, A. Shahinian, D. Kagi, A. Hakem, M. McCurrach, W. Khoo, S. A. Kaufman, G. Senaldi, T. Howard, S. W. Lowe, and T. W. Mak. 1998. Essen-

tial contribution of caspase 3/CPP32 to apoptosis and its associated nuclear changes. Genes Dev. 12: 806 – 819. 58. Wu, L. G., and P. Saggau 1994. Adenosine inhibits evoked synaptic transmission primarily by reducing presynaptic calcium influx in area CA1 of hippocampus. Neuron 12: 1139 – 1148. 59. Zipfel, G. J., D. J. Babcock, J. M. Lee, and D. W. Choi. 2000. Neuronal apoptosis after CNS injury: the roles of glutamate and calcium. J. Neurotrauma 17: 857– 869.