Abstract
Cultured cortical neurons maintained in 25 mm glucose underwent a widespread neuronal death after exposure to NMDA, AMPA, and kainate. Among these, NMDA toxicity was substantially reduced in neurons maintained in 100 mm glucose. NMDA-induced increase in [Ca2+]i and reactive oxygen species was attenuated in neurons maintained in high glucose that revealed increased mitochondrial membrane and redox potentials as determined using rhodamine 123 and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide.p-trifluoromethoxy-phenylhydrazone, KCN, and rotenone, the selective inhibitors of mitochondrial potential, abrogated neuroprotective effect of high glucose against NMDA. The neuroprotective action of high glucose was extended against oxygen or combined oxygen–glucose deprivation. The present study provides evidence that prolonged exposure of cortical cells to high glucose attenuates NMDA- and free radical-mediated neuronal death via enhanced mitochondrial function.
- glucose
- mitochondria
- membrane potential
- redox potential
- excitotoxicity
- Ca2+
- reactive oxygen species
- ischemia
Glucose in brain supplies energy essential for maintenance of the nervous system. Deficiency in glucose that results from hypoglycemia or ischemic insults can trigger neuronal injuries. Balance in ion homeostasis is disturbed, which in turn results in membrane depolarization and massive release of neurotransmitters, including glutamate (Siesjo, 1988; Erecinska and Silver, 1989). The extracellular accumulation of glutamate results in neuronal death by activating ionotropic glutamate receptors sensitive to NMDA or AMPA–kainate (Choi, 1988). In addition, neurons impaired of energy metabolism appear to be highly sensitive to excitotoxicity (Simon et al., 1984; Wieloch, 1985; Monyer et al., 1989; Cebers et al., 1998). Intracellular free Ca2+ and reactive oxygen species (ROS) have been well documented as causative mediators of excitotoxicity (Choi, 1988; Coyle and Puttfarcken, 1993). In addition, several studies suggest that brain supplied with excess glucose becomes more vulnerable to ischemic injuries, presumably by increasing lactic acid production and generation of ROS (Smith et al., 1986; Nedergaard, 1987; Lundgren et al., 1992; Li et al., 1998). However, administration of high glucose before hypoxic–ischemia has been reported to reduce brain damage (Ginsberg et al., 1987; Zasslow et al., 1989; Kraft et al., 1990; Vannucci et al., 1996). In addition to this beneficial effect of hyperglycemia, maneuvers increasing glucose entry into neurons were shown to protect neurons from glutamate neurotoxicity (Ho et al., 1995), stroke or seizure (Lawrence et al., 1995, 1996), and mitochondrial toxins (Dash et al., 1996). Addition of glucose metabolites, pyruvate and malate, attenuated neuronal death after exposure to glutamate or H2O2 (Desagher et al., 1997; Ruiz et al., 1998). Recognizing the apparent controversial effects of high glucose on neuronal injuries, we set out experiments to examine how prolonged exposure to high glucose modulates excitotoxicity, oxidative stress, and deprivation of oxygen or glucose induced in primary cortical cell cultures.
MATERIALS AND METHODS
Primary cortical cell culture. Neocortices were prepared from brains of fetal ICR mice at 14–15 d gestation and mechanically triturated as previously described (Noh and Gwag, 1997). Dissociated cells were plated on 24 well plates (five hemispheres per plate, ∼105 cells per well) in a plating medium consisting of Eagle’s minimal essential media (MEM; Earle’s salts, 11090–081) supplemented with 5% horse serum, 5% fetal bovine serum, and 2 mm glutamine. Two different groups of cultures were prepared by supplementing plating medium with 21 mmglucose (final glucose concentration, 25 mm) or 96 mm glucose (final glucose concentration, 100 mm). Under both conditions, glia become confluent at 7–8 d in vitro (DIV). Thereafter, overgrowth of glial cells was halted by 2 d exposure to 10 μm cytosine arabinoside. Cultures were then fed twice a week with plating medium lacking fetal serum. Cultures were maintained at 37°C in a humidified 5% CO2 atmosphere.
Neurotoxicity experiments. Cortical cell cultures (DIV 12–14) were washed in MEM (bicarbonate-free, 11700–010) supplemented with 26.6 mm bicarbonate and 21 mm glucose. Cultures were then exposed to excitotoxins (NMDA, AMPA, or kainate) or free radical-inducing agents [Fe2+ or buthionine-S,R-sulfoximine (BSO)] for 24 hr.
For oxygen or glucose deprivation, cortical cell cultures (DIV 15–17) were transferred to an anaerobic chamber containing 5% CO2, 10% H2, and 85% N2 as described before (Gwag et al., 1997). For oxygen deprivation, culture medium was replaced with deoxygenated balanced salt solution containing (in mm): 143.6 NaCl, 5.4 KCl, 1.8 CaCl2, 0.8 MgSO4, 1 NaH2PO4, 26.2 NaHCO3, 5.5 glucose, and 10 mg/l phenol red. For combined oxygen and glucose deprivation, culture medium was replaced with deoxygenated balanced salt solution lacking glucose. To terminate injuries, cultures were taken out of the anaerobic chamber, added with concentrated glucose to the exposure solutions (final glucose concentration, 5.5 mm), and placed to the aerobic CO2 incubator.
Analysis of neuronal death. Overall cell injury was assessed microscopically under phase-contrast optics or by measuring amount of lactate dehydrogenase (LDH) released into the bathing medium 24 hr after neurotoxic insults as previously described (Koh and Choi, 1987). The percent neuronal death was normalized to the mean LDH value released 24 hr after continuous exposure to 500 μm NMDA (= 100) or a sham control (= 0).
Calcium imaging. Measurement of intracellular free calcium concentration ([Ca2+]i) was performed using a Ca2+-sensitive indicator fura-2 under a fluorescence microphotometry (Grynkiewicz et al., 1985). Cortical cell cultures (DIV 12) grown on a glass-bottom dish were loaded with 5 μm fura-2 AM plus 2% Pluronic F-127 for 30 min at room temperature. Cells were washed three times with a salt solution containing (in mm): 120 NaCl, 5 KCl, 2.3 CaCl2, 15 glucose, 20 HEPES, and 10 NaOH, pH 7.4. The fura-2 fluorescent signals (Ex = 340/380 nm; Em = 510 nm) were acquired with a Nikon Diaphot inverted microscope and CCD camera. Fura-2 ratio images were analyzed using a Quanticell 700 system (Applied Imaging).
ROS imaging. Cortical cell cultures (DIV 12) grown on a glass-bottom dish were loaded with 5 μmdichlorodihydrofluorescein diacetate (DCDHF-DA; Molecular Probes, Eugene, OR) plus 2% Pluronic F-127 in HEPES-buffered control salt solution (HCSS) containing (in mm): 120 NaCl, 5 KCl, 1.6 MgCl2, 2.3 CaCl2, 15 glucose, 20 HEPES, and 10 NaOH. Cultures were incubated for 20 min at 37°C, washed three times with HCSS, and the fluorescence signal of DCF (Ex = 490 nm; Em = 510 nm), the oxidation product of DCDHF-DA by free radicals, was analyzed on the stage of a Nikon Diaphot inverted microscope equipped with a 100 W Xenon lamp. To minimize background signal caused by direct oxidation of DCDHF-DA by illumination at 490 nm, intracellular levels of ROS were analyzed within 3 sec after illumination using a Quanticell 700 system (Applied Imaging).
Measurement of mitochondrial transmembrane potential.Cortical cell cultures (DIV 12) grown on a glass-bottom dish were loaded with 5 μm rhodamine 123 (Molecular Probes, Eugene, OR), a fluorescent dye indicating mitochondrial transmembrane potential (Bellomo et al., 1991), in HEPES-buffered control salt solution (HCSS) containing (in mm): 120 NaCl, 5 KCl, 1.6 MgCl2, 2.3 CaCl2, 15 glucose, 20 HEPES, and 10 NaOH. Cultures were incubated for 10 min at 37°C, washed three times with HCSS, and the fluorescence signal (Ex = 480 nm; Em = 520 nm) of rhodamine 123 was analyzed on the stage of a Nikon Diaphot inverted microscope equipped with a 100 W Xenon lamp. Background fluorescence signal of rhodamine 123 was determined on a glass-bottom dish without cells and subtracted from rhodamine 123 signals obtained in cortical neurons maintained in 25 and 100 mm glucose. All images were analyzed using a Quanticell 700 system (Applied Imaging).
MTT assay. The mitochondrial dehydrogenase activity that cleaves 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) was used to determine mitochondrial redox potential in a quantitative colorimetric assay (Mosmann, 1983). Cortical cell cultures (DIV 12) were incubated with 100 μg/ml MTT in PBS for 1 hr at 37°C. The supernatant was then aspirated, and the formazan product was dissolved in dimethylsulfoxide and analyzed at 570 nm.
RESULTS
Attenuation of NMDA-induced neurotoxicity in cortical cell cultures maintained in high glucose
We first examined the possibility that excitotoxicity would be altered in cortical cell cultures grown in high glucose. Cortical cell cultures (DIV 12) maintained in 25 mm glucose showed swelling of neuronal cell body within 6 hr after exposure to 20 μm NMDA (Fig.1A). The NMDA-induced cell swelling was not produced in cortical cell cultures maintained in 100 mm glucose. However, cortical neurons (DIV 12) maintained in 25 mm or 100 mm glucose underwent cell body swelling after exposure to 40 μm kainate or 10 μm AMPA (data not shown). Whereas cortical cell cultures grown in 25 mm glucose showed dose-dependent neuronal death 24 hr after exposure to 10–40 μm NMDA, 3–30 μm AMPA, or 20–80 μm kainate, cultures grown in 100 mm glucose were highly resistant to neuronal death induced by NMDA, but not the other excitotoxins (Fig.1B).
Fura-2 fluorescence microphotometry was performed to determine if cortical neurons grown in high glucose would override rise in [Ca2+]i after exposure to NMDA. In cortical neurons grown in 25 mmglucose, [Ca2+]iwas increased gradually over 120 min after exposure to 20 μm NMDA. Although the baseline levels of [Ca2+]i were not different between cortical neurons maintained in 25 and 100 mm glucose, NMDA-induced rise in [Ca2+]i was markedly reduced in neurons grown in 100 mm glucose (Fig.2). Accumulation of ROS after activation of NMDA receptors has been implicated as a necessary route to neuronal death (Dyken, 1994; Dugan et al., 1995). In cortical neurons grown in 25 mm glucose, addition of 20 μm NMDA increased levels of ROS within the 3 hr that was analyzed using DCDHF-DA, a redox-sensitive dye (Fig. 3). The baseline levels of [ROS]i were similar in cortical neurons maintained in 25 and 100 mm glucose, but [ROS]i produced after activation of NMDA receptors was significantly reduced in cortical cell cultures grown in 100 mm glucose.
The neuroprotective effect of high glucose requires enhanced mitochondrial potential
Activation of NMDA receptor results in selective uptake of Ca2+ into mitochondria and subsequent production of ROS that appears to mediate neuronal death (Dugan et al., 1995; Reynolds and Hastings, 1995; Peng and Greenamyre, 1998). Glucose entered into cells likely enhances mitochondrial potentials that play a central role in regulation of [Ca2+]i and ROS. This possibility was examined using the positively charged and lipophilic rhodamine 123 that permeates into the negatively charged mitochondria and therefore reflects the mitochondrial transmembrane potential (Johnson et al., 1980; Emaus et al., 1986). The fluorescence signal of rhodamine 123 was significantly increased in cortical neurons grown in 100 mm glucose compared to 25 mmglucose (Figs. 4A,B). We also analyzed mitochondrial redox potential by measuring reduction of the tetrazolium salt MTT by α-nicotinamide adenine dinucleotide (NADH) dehydrogenase. Prolonged exposure to high glucose enhanced mitochondrial redox potential as shown by increased MTT reduction (Fig. 4C).
Additional experiments were performed to determine if the enhanced mitochondrial potentials by high glucose would be required for attenuation of NMDA neurotoxicity. Inclusion of 0.1 μmp-trifluoromethoxy-phenylhydrazone (FCCP), a hydrophobic protonophore reducing mitochondrial membrane potential, was not toxic itself but abrogated the neuroprotective action of high glucose against NMDA (Fig. 5). The neuroprotective effect of high glucose against NMDA was also reversed with inclusion of subtoxic doses of rotenone or KCN that reduces mitochondrial redox potential by blocking NADH dehydrogenase or cytochrome oxidase. We performed additional experiments to examine if effect of high glucose reducing NMDA-induced accumulation of [Ca2+]i and [ROS]i would be abrogated in the presence of FCCP, KCN, or rotenone. In cortical neurons maintained in 100 mm glucose, neither [Ca2+]i nor [ROS]i was accumulated after exposure to 100 μm KCN, 0.1 μm FCCP, or 0.1 μm rotenone for 4–12 hr (data not shown). However, concurrent treatment with 100 μm KCN, 0.1 μm FCCP, or 0.1 μmrotenone elevated accumulation of [Ca2+]i and [ROS]i after exposure to 20 μm NMDA (Table1). These results imply that prolonged exposure of cortical neurons to high glucose attenuates NMDA-induced accumulation of [Ca2+]i and [ROS]I through enhanced mitochondrial potentials.
Attenuation of free radical neurotoxicity in cortical cell cultures maintained in high glucose
We next examined the neuroprotective potential of high glucose against oxidative stress. Cortical cell cultures grown in 25 mm glucose underwent widespread neuronal death 24 hr after exposure to 50 μm Fe2+ or 1 mm BSO that produces ROS by a Fenton reaction or depletion of glutathione, respectively. This free radical-mediated neurotoxicity was substantially decreased in cortical cell cultures grown in 100 mm glucose (Fig. 6). Accumulation of ROS after exposure to Fe2+was significantly reduced in cortical neurons grown in 100 mm glucose compared to cortical neurons grown in 25 mm glucose (Fig. 7). The baseline levels of [ROS]i were similar between cortical cells maintained in 25 and 100 mm glucose.
Cortical neurons maintained in high glucose are protected from deprivation of oxygen or combined oxygen and glucose
Cortical cell cultures grown in 25 mm glucose produced swelling of neuronal cell body at 6 hr and subsequent neuronal death at 24 hr after deprivation of oxygen for 6–14 hr (Fig.8A,B). Maintaining cultures in 100 mm glucose markedly protected cortical neurons from the oxygen deprivation for 6–12 hr. Cortical neurons maintained in high glucose were also resistant to insults induced by deprivation of oxygen and glucose for 60–100 min (Fig. 8C).
DISCUSSION
We have demonstrated that cortical neurons maintained in high glucose are spared against excitotoxicity induced by NMDA, but neither AMPA nor kainate. Neurons maintained in high glucose reveal enhanced mitochondrial transmembrane and redox potential that likely counteract toxic accumulation of [Ca2+]i and [ROS]i. The neuroprotective effects of high glucose are extended to injuries by oxygen or glucose deprivation as well as oxidative stress.
Activation of NMDA or AMPA–kainate glutamate receptors results in rapid influx of ions such as Na+, Ca2+, and Cl− that in turn causes cell body swelling and lysis (Choi, 1988). In the present study, prolonged exposure to high glucose attenuated NMDA neurotoxicity without influencing AMPA- and kainate-induced neuronal death. Recognizing the unique property of the NMDA receptor complex as a major route of Ca2+ entry, the preferential neuroprotective effect of high glucose against NMDA may be attributable to reducing toxic actions of Ca2+. In support of this, administration of excess glucose to PC12 cells was reported to reduce the depolarization-induced increase of [Ca2+]i (Chan and Greenberg, 1991). In the present study, NMDA-induced [Ca2+]iaccumulation was not observed in cortical neurons maintained in high glucose. Moreover, NMDA-induced production of [ROS]i, which results from Ca2+-dependent uncoupling of mitochondrial electron transport (Malis and Bonventre, 1986; Lafon-Cazal et al., 1993; Kiedrowski and Costa, 1995; Reynolds and Hastings, 1995), was reduced in cortical neurons maintained in high glucose. The present findings that cortical neurons maintained in high glucose are resistant against neurotoxicity by NMDA, but neither kainate nor AMPA, provide further evidence supporting causative role of Ca2+ and ROS for NMDA-mediated excitotoxicity (Tymianski et al., 1993; Reynolds and Hastings, 1995;Dugan et al., 1995; Stout et al., 1998).
Accumulated [Ca2+]i is extruded or can be redistributed through ATP-dependent reuptake into the endoplasmic reticulum and passive entry into the mitochondria matrix (Wang and Thayer, 1996; Peng et al., 1998). Mitochondrial transmembrane potential was markedly increased in cortical neurons maintained in high glucose. This increased potential can drive ATP production that enhances efflux of Ca2+and therefore counteracts NMDA-induced [Ca2+]iaccumulation. Alternatively, the mitochondrial membrane potential enhanced by high glucose can detoxify excess accumulation of [Ca2+]i by accelerating entry of Ca2+ to intracellular Ca2+ pools irrespective of ATP production. In support of this, hippocampal and cortical neurons treated with 2.5 mm pyruvate and 1.5 mmmalate show enhanced Ca2+ entry into the endoplasmic reticulum or the mitochondria without increasing ATP production (Villalba et al., 1994). With enhanced mitochondrial redox potential by high glucose that results in substantial reduction of [ROS]i, NMDA-induced accumulation of [ROS]i as well as [Ca2+]i can be lowered to subtoxic levels in cortical neurons. Selective inhibitors of mitochondrial membrane and redox potential reversed the neuroprotective effect of high glucose against NMDA, suggesting that the enhanced mitochondrial potentials are necessary for the neuroprotective effect of high glucose.
The mitochondria play a role in generation and processing of [ROS]i. Abnormal changes in mitochondrial DNA and proteins have been observed in brain during aging and neurodegenerative process (Coyle and Puttfarcken, 1993; Beal, 1995). The mitochondrial dysfunction can result in accumulation of [ROS]i and cause neuronal death. Mitochondrial toxins such as N-methyl,4-phenyl-1,2,3,6 tetrahydropyridine and 3-nitropropionic acid appear to produce ROS-mediated neurotoxicity via inhibition of electron transporters (Beal et al., 1993; Smith and Bennett, 1997). In the present study, increasing mitochondrial potentials protected neurons against Fe2+and BSO by interfering with accumulation of [ROS]i. Further study will be needed to determine whether the neuroprotective effect of increased mitochondrial potentials can be extended to other physiological and pathological cell deaths mediated through ROS.
NMDA neurotoxicity and oxidative stress have been well documented as mechanisms underlying hypoxic–ischemic brain injuries. The protective effects of high glucose that attenuated NMDA- and ROS-mediated neuronal death were also observed against deprivation of oxygen or glucose. The neonatal rat administrated with high glucose is resistant to hypoxic–ischemic injuries (Palmer and Vannucci, 1993). In contrast to the beneficial effects of high glucose, systemic administration of glucose before ischemia accentuates brain damage in the adult animals after hypoxic–ischemic injuries (Sieber and Traystman, 1992). Increased accumulation of lactic acid and [ROS]i has been correlated with the worsening effects of glucose. However, neuroprotective effect of high glucose after prolonged exposure does not conflict with the deleterious one of glucose systemically administered 1 hr before onset of hypoxic–ischemic injury. In our study, acute treatment (1–3 d) with high glucose did not protect cortical neurons from exposure to NMDA or oxygen–glucose deprivation (S. Seo and B. Gwag, unpublished data).
Prolonged exposure to high glucose protects selectively against NMDA- and ROS- dependent neurotoxicity. The enhanced mitochondrial membrane and redox potentials by high glucose demonstrate as key mechanisms underlying the neuroprotective effects of high glucose. Maneuvers increasing mitochondrial potentials can be merited to treat acute and chronic neurodegenerative diseases that depend upon excess activation of NMDA receptors or toxic accumulation of [ROS]i.
Footnotes
This work was supported by Korea Science and Engineering Foundation through the Brain Disease Research Center at Ajou University (B.J.G.).
Correspondence should be addressed to Dr. Byoung J. Gwag, Department of Pharmacology, Ajou University School of Medicine, San 5, Wonchon-dong, Paldal-gu, Suwon, Kyungki-do 442–749, Korea.