Involvement of mitochondrial K+ release and cellular efflux in ischemic and apoptotic neuronal death

Authors


Address correspondence and reprint requests to Mark P. Mattson, Laboratory of Neurosciences, National Institute on Aging GRC 4F01, 5600 Nathan Shock Drive, Baltimore, MD 21224, USA. E-mail: mattsonm@grc.nia.nih.gov

Abstract

We measured and manipulated intracellular potassium (K+) fluxes in cultured hippocampal neurons in an effort to understand the involvement of K+ in neuronal death under conditions of ischemia and exposure to apoptotic stimuli. Measurements of the intracellular K+ concentration using the fluorescent probe 1,3-benzenedicarboxylic acid, 4,4′-[1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diylbis(5-methoxy-6,2-benzofurandiyl)]bis-, tetrakis [(acetyloxy) methyl] ester (PBFI) revealed that exposure of neurons to cyanide (chemical hypoxia), glutamate (excitotoxic insult) or staurosporine (apoptotic stimulus) results in efflux of K+ and cell death. Treatment of neurons with 5-hydroxydecanoate (5HD), an inhibitor of mitochondrial K+ channels, reduced K+ fluxes in neurons exposed to each insult and increased the resistance of the cells to death. K+ efflux was attenuated, levels of oxyradicals were decreased, mitochondrial membrane potential was stabilized and release of cytochrome c from mitochondria was attenuated in neurons treated with 5HD. K+ was rapidly released into the cytosol from mitochondria when neurons were exposed to the K+ channel opener, diazoxide, or to the mitochondrial uncoupler, carbonyl cyanide 4(trifluoromethoxy)phenylhydrazone (FCCP), demonstrating that the intramitochondrial K+ concentration is greater than the cytosolic K+ concentration. The release of K+ from mitochondria was followed by efflux through plasma membrane K+ channels. In vivo studies showed that 5HD reduces ischemic brain damage without affecting cerebral blood flow in a mouse model of focal ischemic stroke. These findings suggest that intracellular K+ fluxes play a key role in modulating neuronal oxyradical production and cell survival under ischemic conditions, and that agents that modify K+ fluxes may have therapeutic benefit in stroke and related neurodegenerative conditions.

Abbreviations used
CBF

cerebral blood flow

DCF

2,7-dichlorofluorescin diacetate

FCCP

carbonyl cyanide 4(trifluoromethoxy)phenylhydrazone

HBSS

Hank's buffered saline solution

5HD

5-hydroxydecanoate

Mito-KATP

mitochondrial ATP-sensitive K+ channels

PBFI

1,3-benzenedicarboxylic acid, 4,4′-[1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diylbis(5-methoxy-6,2-benzofurandiyl)]bis-, tetrakis [(acetyloxy) methyl] ester

PBS

phosphate-buffered saline

ROS

reactive oxygen species

TEA

tetraethylammonium

TMRE

tetramethylrhodamine ethyl ester

TTC

2,3,5-triphenyltetrazolium chloride

Potassium (K+) channels are widely expressed in all eukaryotic cells. K+ channels located in the plasma membrane have been heavily studied and shown to play key roles in regulating neuronal excitability and synaptic transmission (Robertson 1997; Sanguinetti and Spector 1997; Aguilar-Bryan et al. 1998). Moreover, several different neurological disorders have been linked to abnormalities in plasma membrane K+ channels, including epilepsy, episodic ataxia and deafness (Lehmann-Horn and Jurkat-Rott 1999; Jentsch 2000). In contrast to K+ channels in the plasma membrane, very little information is available concerning K+ channels in intracellular organelles. Based on studies employing pharmacological agents, and recent electrophysiological data, it has been proposed that mitochondria contain ATP-sensitive K+ (Mito-KATP) channels and charybdotoxin-sensitive K+ channels (Inoue et al. 1991; Siemen et al. 1999; Grover and Garlid 2000; Liu et al. 2001; Xu et al. 2002). However, the protein composition and electrophysiological properties have not been fully established. Moreover, it has not been established whether the intramitochondrial K+ concentration is less or greater than the cytoplasmic K+ concentration, and therefore whether opening of mitochondrial K+ channels results in potassium influx or release. Recent studies have identified candidate Mito-KATP channel proteins, including a 28 kDa protein that was isolated from heart muscle mitochondria based upon its selective binding to the sulfonylurea compound glibenclamide (Szewczyk et al. 1997), a charybdotoxin-sensitive channel (Xu et al. 2002) and a putative brain Mito-KATP channel protein which exhibits ligand-binding properties similar to those of heart Mito-KATP channels (Bajgar et al. 2001; Debska et al. 2001). The latter study provided evidence that the amount of Mito-KATP channels in brain cells is at least sixfold higher than in heart cells, suggesting an important role for these channels in neuronal function. Mito-KATP channels exhibit a distinct pharmacological profile that allows their functions to be studied; they are selectively activated by low concentrations of diazoxide and blocked by 5-hydroxydecanoate (5HD) (Jaburek et al. 1998; Ghosh et al. 2000; Sato et al. 2000; Liu et al. 2001).

Several lines of evidence suggest that K+ fluxes play important roles in the neuronal deaths that occur in various settings. Drugs that activate plasma membrane K+ channels have been reported to protect neurons against excitotoxic and oxidative insults (Abele and Miller 1990; Goodman and Mattson 1996), and K+ channel blockers can induce apoptosis (Choi et al. 1999). On the other hand, plasma membrane K+ channels are activated in neurons undergoing apoptosis in response to glutamate receptor activation or exposure to ceramide (Yu et al. 1999a, b). Studies of heart muscle cells subjected to ischemia have suggested an important role for mitochondrial K+ channels in ischemic pre-conditioning, a process in which exposure of cells to a brief mild ischemia increases their resistance to a subsequent more severe ischemic insult (Matsushima and Hakim 1995; Barone et al. 1998; Stagliano et al. 1999; Yellon and Dana 2000; Shimizu et al. 2001). Opening of Mito-KATP with diazoxide can mimic the cytoprotective effect of pre-conditioning, whereas Mito-KATP blockers such as 5HD can prevent ischemic pre-conditioning (Takashi et al. 1999; Eells et al. 2000; O'Rourke 2000). Recent studies suggest that mitochondrial K+ channels also play a role in ischemic neuronal death (Domoki et al. 1999; Liu et al. 2002). Neuronal death after a stroke involves overactivation of glutamate receptors (excitotoxicity), oxidative stress and cellular calcium overload (Lee et al. 2001), and a form of programmed cell death called apoptosis (Mattson and Culmsee 2000; Bratton and Cohen 2001; Graham and Chen 2001). Alterations in mitochondria, including membrane depolarization and permeability transition and release of cytochrome c, are believed to be pivotal events in neuronal apoptosis (Fiskum 2000; Gorman et al. 2000). In the present study, we therefore sought to determine the role of intracellular K+ fluxes in excitotoxic and ischemic neuronal apoptosis. Surprisingly, we demonstrate that the intramitochondrial K+ concentration is greater than the cytoplasmic K+ concentration in cultured hippocampal neurons. We further show that K+ release from mitochondria plays a role in neuronal apoptosis, and that blockade of this K+ release is neuroprotective both in cell culture and in a mouse model of focal ischemic brain injury.

Materials and methods

Hippocampal cultures and quantification of cell survival

Hippocampi from embryonic day 18 rat embryos were removed, incubated in Hank's buffered saline solution (HBSS) containing 0.2% trypsin, and cells were mechanically dissociated by passage through the narrowed bore of a fire-polished Pasteur pipette. Cells were seeded onto polyethyleneimine-coated plastic dishes or 22 mm2 glass coverslips, and maintained in Neurobasal medium containing B-27 supplements, 2 mm l-glutamine, 0.001% gentamicin sulfate and 1 mm HEPES (pH 7.2). Experiments were performed in 7 to 9-day-old cultures; the neurons in these cultures are vulnerable to chemical hypoxia and apoptosis induced by staurosporine (Mattson et al. 1998; Kruman et al. 1998; Kruman and Mattson 1999). Potassium cyanide was prepared as a 500× stock in water, glutamate was prepared as a 200× stock in HBSS, and 5HD and staurosporine were prepared as 500× stocks in dimethylsulfoxide. Cell survival was quantified by counting undamaged neurons in a pre-marked microscope field prior to, and at indicated time points after, exposure to experimental treatments using methods described previously (Mattson et al. 1997). Neurons that died in the intervals between examination points were usually absent, and the viability of the remaining neurons was assessed by morphological criteria; neurons with intact neurites and soma with a smooth round appearance were considered viable, whereas neurons with fragmented neurites and vacuolated soma were considered non-viable. Analyses were performed without knowledge of the treatment history of the cultures.

Imaging of intracellular K+ concentrations

The intracellular [K+] in cultured hippocampal neurons was assessed with the K+-selective fluorescence indicator dye 1,3-benzenedicarboxylic acid, 4,4′-[1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diylbis(5-methoxy-6,2-benzofurandiyl)]bis-, tetrakis [(acetyloxy) methyl] ester (PBFI-AM) using methods similar to those described previously (Jezek et al. 1990; Meuwis et al. 1995; Cook et al. 2002). Cells were incubated for 45 min in the presence of 5 µm PBFI-AM (Molecular Probes, Eugene, OR, USA). Cells were washed twice with K ± free saline (159 mm NaCl, 3.6 mm NaHCO31 mm MgCl2, 5 mm HEPES, 2.3 mm CaCl2, 10 mm glucose; pH 7.4) and were maintained in K+-free saline during imaging. For some experiments, sodium (Na+) was replaced with an equimolar concentration of N-methyl-d-glucamine (Mattson et al. 1989). Images of PBFI fluorescence in neuronal cell bodies were acquired using a Zeiss Axiovert microscope (Zeiss, Thornwood, NY, USA) with a 40× oil immersion objective lens; the dye was sequentially excited at 340 and 380 nm (510 nm emission). Pairs of images were acquired at 5–7 s intervals using an Attofluor system (Zeiss) which calculated a ratio of the fluorescence images acquired at each excitation wavelength (340/380 nm); data are presented as the ratio for each neuron. Diazoxide, cyanide, staurosporine, glutamate, tetraethylammonium, ouabain, glibenclamide, 5HD, carbonyl cyanide 4(trifluoromethoxy)phenylhydrazone (FCCP) (a mitochondrial uncoupler) and valinomycin (a K+ ionophore) were prepared as 200 to 500× concentrated stocks and were added directly to the culture incubation medium during imaging.

Measurements of levels of reactive oxygen species and mitochondrial membrane potential

Levels of cellular reactive oxygen species (ROS) were measured using the fluorescent probe 2,7-dichlorofluorescin diacetate (DCF) as described previously (Goodman and Mattson 1994). Briefly, cells were incubated for 30 min in the presence of 50 µm DCF followed by washing in HBSS. Cells were imaged using a confocal laser scanning microscope coupled to an inverted microscope; cells were located under bright-field optics and then scanned once with the laser (488 nm excitation and 510 emission). The laser beam intensity and photodetector sensitivity were held constant across cultures to allow quantitative comparisons of relative fluorescence intensity of cells between treatment groups. DCF localizes mainly in the cytoplasm and is oxidized by several ROS, most notably hydrogen peroxide (Page et al. 1993). The fluorescent probe tetramethylrhodamine ethyl ester (TMRE) (Molecular Probes) was employed as an indicator of mitochondrial membrane potential using methods described previously (Krohn et al. 1999; Scaduto and Grotyohann 1999). Following exposure of cells to experimental treatments, cells were incubated for 20 min in the presence of 100 nm TMRE, washed three times in fresh culture medium, and confocal images of cellular TMRE fluorescence acquired (543 nm excitation and 585 nm emission). All images were coded and analyzed without knowledge of experimental treatment history of the cultures. Values of the average fluorescence intensity in the cell body of each neuron imaged were obtained using the software supplied by the manufacturer of the confocal microscope (Zeiss).

Assessment of cytochrome c release from mitochondria

In order to evaluate the subcellular localization of cytochrome c, we employed confocal imaging of cells double-labeled with Mitotracker Red CMX Ros (Molecular Probes) and a cytochrome c antibody using methods similar to those described previously (Cheng et al. 2001). After experimental treatment, cells were incubated with 100 nm Mitotracker Red CMX Ros for 30 min at 37°C (the dye is taken up by mitochondria where it forms thiol conjugates with peptides and is thereby trapped in the mitochondria), washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde in PBS at 37°C for 30 min. Fixed cells were permeabilized with 0.1% Triton X-100 for 5 min at 4°C, followed by a 2 h incubation at room temperature in blocking solution (2% normal goat serum, 0.1% Triton X-100 in PBS, pH 7.4) containing primary monoclonal cytochrome c antibody (10 µg/mL; PharMingen, San Diego, CA, USA). After washing, cells were incubated for 2 h in PBS containing FITC-conjugated goat anti-mouse IgG (1 : 100, ImmunoResearch Laboratory, West Grave, PA, USA). Cells were then imaged in dual-scan mode on a Zeiss CLSM 510 confocal microscope using a 40× water immersion objective (numerical apeture (N.A.) = 1.4). The excitation and emission wavelengths for Mitotracker CMX Ros were 510 and 590 nm, and for FITC were 488 and 510 nm, respectively. In one experiment, immunoblot analysis of cytochrome c was performed on the cytosolic fraction of neurons that had been exposed to staurosporine alone or in combination with 5HD using methods described previously (Cheng et al. 2001). Each lane of the gel was loaded with 50 µg cytosolic protein and equal loading was confirmed by probing the same blot with an actin antibody.

Measurement of mitochondrial permeability transition

Brain mitochondria were isolated from rat cerebral cortex tissue using a discontinuous Percoll gradient according to Sims, method B (Sims 1990), with slight modification, as described previously (Hansson et al. 2003). Activation of mitochondrial permeability transition was monitored by measuring the decrease in right angle light scattering at 520 nm (mitochondrial swelling) using a Perkin-Elmer Spectrometer LS-50B (Emeryville, CA, USA). 5HD was added to mitochondrial suspensions resulting in final concentrations of 1, 10,300 and 1000 μM. Experiments were run under deenergized conditions as described previously (Hansson et al. 2003), and under energized conditions (250 mM Sucrose, 20 mM Mops, 10 mM Tris Baze, 2 mM Pi, 1 mM Mg2+, 1 μM EGTA, 1 μg/ml Oligomycin, 20 μM ADP, 5 mM Malate/Glutamate, pH 7.2, 37°C). Calcium was added to induce swelling, 100 μM and 2 μmoles/mg mitochondrial protein, respectively. Cyclosporin A (1 μM) was used as positive control and experiments were terminated with addition of alamethicin to induce maximal swelling.

Mice, drug administration and focal cerebral ischemia model

Three-month-old male C57BL/6 mice weighing 25–28 g were obtained from the National Cancer Institute and maintained on a 12 h light/12 h dark cycle with continuous access to food and water. Mice were fasted overnight prior to surgery. 5HD was infused intravenously in the femoral vein, at a dose 20 mg/kg, 20–30 min prior to middle cerebral artery occlusion. To induce permanent focal cerebral ischemia, the left middle cerebral artery was exposed and cauterized using methods described previously (Nawashiro et al. 2000). Mice were anesthetized using isofluorane administered as a vapor, and body temperature was maintained at 37°C throughout the surgical procedure and recovery period. A 1 cm vertical incision was made between the left eye and ear; the temporal muscle was split, a portion of the skull at the junction of the zygomatic arch and squamous bone was removed, and the left middle carotid artery was exposed. The left middle cerebral artery was occluded by electrocoagulation. Laser Doppler flowmetry was used to measure cerebral blood flow (CBF) using a flexible 0.5 mm fiberoptic probe (Perimed, Stockholm, Sweden); CBF recordings were begun 5–10 min prior to drug administration and continued for at least 30 min after middle cerebral artery occlusion. Brain temperature was measured using a probe placed directly on the brain surface. Mice were killed 24 h after middle cerebral artery occlusion by isofluorane overdose. All animal procedures were approved by the National Institute on Aging Animal Care and Use Committee and complied with NIH guidelines for the care and use of laboratory animals.

Quantification of infarct volume

These methods were identical to those used in our previous studies (Liu et al. 2002; Yu et al. 1998; Yu and Mattson 1999). Briefly, brains were removed, rinsed in cold PBS, and 2 mm-thick coronal sections cut. Tissue sections were incubated for 30 min at room temperature in a solution of 2% 2,3,5-triphenyltetrazolium chloride (TTC) in PBS, and then fixed for 30 min with 4% paraformaldehyde in PBS. The borders of the infarct in each brain slice were outlined and the area quantified using NIH image 6.1 software. To correct for brain swelling the infarct area was determined by subtracting the area of undamaged tissue in the left hemisphere from that of the intact contralateral hemisphere. Infarct volume was calculated by integration of infarct areas for all slices of each brain (Swanson et al. 1990).

Results

5-Hydroxydecanoate protects cultured neurons against death induced by glutamate and staurosporine

Previous studies had provided evidence that 5-hydroxydecanoate (5HD) is a selective blocker of Mito-KATP channels (O'Rourke 2000). In order to determine whether opening of Mito-KATP channels plays a role in excitotoxic neuronal death and apoptosis, we pre-treated cultured hippocampal neurons with 100 µm 5HD or vehicle (0.2% dimethylsulfoxide) immediately prior to exposure to glutamate and staurosporine at concentrations which we had shown in previous studies induce apoptosis in hippocampal cultures identical to those employed in the present study (Mattson et al. 1998; Glazner et al. 2000). Whereas approximately 80% of the neurons were killed by glutamate in vehicle-treated cultures, only 40% of the neurons were killed in cultures pre-treated with 5HD (Fig. 1). Staurosporine killed approximately 90% of the neurons in vehicle-treated cultures, whereas only 30% of the neurons were killed by staurosporine in cultures that had been treated with 5HD (Fig. 1). These results suggested that flux of K+ across the mitochondrial membrane played an important role in excitotoxicity and apoptosis, and prompted us to examine K+ fluxes in neurons exposed to agents that modulate mitochondrial K+ channels.

Figure 1.

5HD protects cultured hippocampal neurons against excitotoxicity and apoptosis. Hippocampal cultures were pre-treated for 30 min with 5HD (100 µm). Cultures were then exposed for 24 h to 10 µm glutamate or 200 nm staurosporine and neuronal survival was quantified. Values are the mean and SD of determinations made in four to six cultures. **p < 0.01, ***p < 0.001 compared with corresponding control value (anova with Scheffe posthoc tests).

Evidence that activation of Mito-KATP results in the release of K+ from mitochondria

Although it has been presumed that the cytosolic K+ concentration is greater than the intramitochondrial K+ concentration, and that opening of Mito-KATP should therefore result in K+ influx into the mitochondria, recent findings suggest that the intramitochondrial K+ concentration can be greater than the cytoplasmic K+ concentration (Garlid 1996). We therefore determined the effects of diazoxide and 5HD on the cytoplasmic K+ concentration in cultured hippocampal neurons by imaging cells loaded with the fluorescent K+ indicator dye PBFI (Jezek et al. 1990; Meuwis et al. 1995; Cook et al. 2002). Exposure of neurons to a concentration of diazoxide (100 µm) expected to activate primarily Mito-KATP channels resulted in a transient increase in PBFI fluorescence ratio which returned to the baseline level during a period of several minutes (Figs 2 and 3a), consistent with a release of K+ from mitochondria. Exposure of neurons to higher concentrations of diazoxide (400 µm) resulted in a large cytosolic K+ transient followed by a progressive depletion of cytosolic K+ to levels far below basal levels (Fig. 3b). The latter result is consistent with the high concentration of diazoxide activating both Mito-KATP (resulting in the rapid increase of cytosolic K+ concentration) and plasma membrane K+ channels (resulting in depletion of cytosolic K+). To rule out the possibility that changes in PBFI fluorescence were due to Na+ fluxes, we incubated neurons in medium in which Na+ was replaced with N-methyl-d-glucamine. Under the latter conditions, neurons exhibited diazoxide-induced changes in PBFI fluorescence indistinguishable from those observed in neurons incubated in Na+-containing buffer (data not shown).

Figure 2.

Diazoxide induces a transient increase in cytosolic [K+] in cultured hippocampal neurons. Images of PBFI fluorescence indicating relative intracellular [K+] in hippocampal neurons prior to, and at different time points after, exposure to 100 µm diazoxide; cells were incubated in buffer lacking K+. Note that diazoxide induces a transient increase in PBFI fluorescence ratio, suggesting that it induced K+ release from the mitochondria.

Figure 3.

Effects of diazoxide and an uncoupler of oxidative phosphorylation on cytosolic [K+] in hippocampal neurons. Cultured hippocampal neurons were loaded with the fluorescent K+ indicator dye, PBFI, and images of PBFI fluorescence were acquired every 5 s prior to, and after exposure of neurons to the indicated treatments. (a) 100 µm diazoxide; (b) 400 µm diazoxide; (c) 10 µm FCCP and 400 µm diazoxide; (d) 5 mm TEA and 400 µm diazoxide. For each graph, the values are the mean and SD of measurements made in 12–20 neurons. Similar results were obtained in three to six separate experiments.

In order to confirm that the diazoxide-induced increase of the PBFI signal was the result of K+ release from mitochondria, we exposed neurons to the mitochondrial uncoupler FCCP. FCCP induced a rapid and very large increase in the PBFI fluorescence ratio (Fig. 3c), consistent with a massive release of K+ from mitochondria. When neurons were treated with tetraethylammonium (TEA) to block plasma membrane K+ channels, exposure to 400 µm diazoxide resulted in a transient increase in the cytosolic K+ which was not followed by the depletion of cytosolic K+ that occurred in the absence of TEA (Fig. 3d). To determine whether mitochondria were capable of releasing K+ under conditions of ongoing cellular efflux of K+, we first exposed neurons to 400 µm diazoxide and then applied a second dose of diazoxide at a time point when the cytosolic K+ was decreasing. The second dose of diazoxide elicited a cytosolic K+ transient, indicating that mitochondria still contained considerable K+, even as K+ was exiting the cell (Fig. 4a). 5HD completely blocked the K+ transient induced by 100 µm diazoxide (Fig. 4b) and markedly attenuated the K+ response to 400 µm diazoxide (Fig. 4c). 5HD did not prevent K+ efflux from the cell (Figs 4c and d), consistent with a relatively selective ability of 5HD to block mitochondrial K+ channels. Glibenclamide attenuated K+ release from mitochondria and blocked K+ efflux from the cells (Fig. 4e), suggesting that both mitochondrial and plasma membrane K+ channels are sensitive to this agent. When neurons were treated with ouabain, an inhibitor of the Na+/K+-ATPase diazoxide induced K+ release and efflux from the cell (Fig. 4f), suggesting that a sodium gradient was not required for K+ release or efflux.

Figure 4.

Evidence that diazoxide induces and 5HD inhibits K+ release from mitochondria in hippocampal neurons. Cultured hippocampal neurons were loaded with the fluorescent K+ indicator dye, PBFI, and images of PBFI fluorescence were acquired every 5 s prior to, and after exposure of neurons to the indicated treatments. (a) 400 µm diazoxide; (b) 100 µm 5HD and 100 µm diazoxide; (c) 100 µm 5HD and 400 µm diazoxide; (d) 400 µm diazoxide and 200 µm 5HD; (e) 10 µm glibenclamide and 400 µm diazoxide; (f) 1 mm ouabain and 400 µm diazoxide. For each graph, the values are the mean and SD of measurements made in at least 10 neurons. Similar results were obtained in three to six separate experiments.

Evidence for mitochondrial release and cellular extrusion of K+ in neurons subjected to hypoxic and apoptotic insults

The death of neurons that occurs as the result of ischemia is believed to involve alterations in mitochondria, including increased oxyradical production, membrane potential changes and release of apoptotic factors such as cytochrome c (Mattson and Kroemer 2003). Although pre-treatment of mice with diazoxide can protect neurons against focal ischemic brain injury by a pre-conditioning mechanism (Liu et al. 2002), fluxes of K+ in neurons under ischemic conditions have not been measured previously. When neurons were subjected to hypoxia (3 h exposure to potassium cyanide) and then challenged with diazoxide, no rapid K+ transient was observed (Fig. 5a), indicating that mitochondria were depleted of K+ during the course of the hypoxia. However, when neurons were treated with 5HD during the period of chemical hypoxia, the mitochondria retained sufficient K+ to be released upon exposure to diazoxide (Fig. 5b).

Figure 5.

Mitochondrial K+ is depleted in neurons subjected to hypoxia. Cultured hippocampal neurons were loaded with the fluorescent K+ indicator dye, PBFI, and images of PBFI fluorescence were acquired every 5 s prior to, and after exposure of neurons to the indicated treatments. (a) 1 mm KCN and 400 µm diazoxide; (b) 100 µm 5HD, 1 mm KCN and 400 µm diazoxide. For each graph, the values are the mean and SD of measurements made in at least 12 neurons. Similar results were obtained in four separate experiments.

We next measured the cytosolic K+ in neurons exposed to either cyanide or staurosporine in order to determine if and how K+ fluxes occur in neurons subjected to hypoxic and apoptotic insults. Exposure of neurons to cyanide resulted in a progressive decrease in the PBFI fluorescence ratio during a 10 min exposure period, suggesting that cyanide induced K+ efflux from the cell (Fig. 6a). This relatively slow apparent decrease of cytosolic K+ concentration contrasted with the rapid and transient increase of the cytosolic K+ concentration (followed by a depletion of K+) in neurons exposed to a high concentration of diazoxide (Fig. 3b). However, we found that when neurons were treated with 5HD, the cyanide-induced depletion of cytosolic K+ was attenuated (Fig. 6b). Staurosporine induced a progressive decrease in cytosolic K+ which was attenuated by 5HD (Figs 6c and d); a similar depletion of cytosolic K+ occurred in neurons exposed to glutamate and this was also attenuated by 5HD (data not shown). 5HD had no effect on mitochondrial permeability pore opening in isolated mitochondria (data now shown), providing further evidence for a specific effect on mitochondrial K+ channels. Collectively, these data suggest that K+ efflux from the cell occurs in neurons subjected to hypoxic, excitotoxic and apoptotic insults, and that this efflux is preceded by K+ release from mitochondria.

Figure 6.

Hypoxic and apoptotic insults induce cellular efflux of K+ which requires, in part, release of K+ from mitochondria. Cultured hippocampal neurons were loaded with the fluorescent K+ indicator dye, PBFI, and images of PBFI fluorescence were acquired prior to, and after exposure of neurons to the indicated treatments (1 mm cyanide, 300 nm staurosporine, 100 µm 5HD). For each graph, the values are the mean and SD of measurements made in 14–24 neurons. Similar results were obtained in three to five separate experiments.

5-Hydroxydecanoate suppresses oxidative stress and cytochrome c release in neurons exposed to hypoxic and apoptotic insults

Oxidative stress and an increase in mitochondrial membrane permeability resulting in the release of cytochrome c are believed to be pivotal events in neuronal apoptosis (Budd et al. 2000), including the deaths of some neurons that occur after a stroke (Mattson et al. 2000). To determine the role of Mito-KATP in the generation of oxyradicals and in apoptosis-related events in mitochondria, we first determined the effect of 5HD on levels of ROS in cultured hippocampal neurons. 5HD significantly decreased levels of DCF fluorescence, suggesting that Mito-KATP play a role in generating ROS under basal culture conditions (Figs 7a and b). As expected from previous studies (Pong et al. 2001), staurosporine induced an increase in levels of DCF fluorescence; however, the magnitude of the increase was significantly reduced in neurons treated with 5HD (Fig. 8a). We next measured mitochondrial membrane potential in neurons using the fluorescent probe TMRE (see Methods). 5HD had no significant effect on mitochondrial membrane potential under basal culture conditions (Figs 7a and c). Exposure of neurons to staurosporine resulted in a decrease in mitochondrial membrane potential, and 5HD significantly attenuated this effect of staurosporine (Fig. 8b).

Figure 7.

Effects of 5HD on ROS production and mitochondrial membrane potential. ROS production was measured using the fluorescent probe, DCF, and mitochondrial membrane potential was assessed using the fluorescent probe, TMRE, in hippocampal neurons. (a) Representative images of DCF fluorescence and TMRE fluorescence in neurons that had been exposed to either vehicle (control) or 100 µm 5HD for 30 min. Note that DCF fluorescence was decreased in the 5HD-treated cultures, whereas TMRE fluorescence was not decreased. (b) and (c) Values are the mean and SD of determinations made in four separate cultures (20–30 neurons assessed per culture). **p < 0.01 compared with the control value.

Figure 8.

5HD suppresses ROS production and stabilizes mitochondrial membrane potential in neurons exposed to an apoptotic insult. Cells were pre-treated for 30 min with vehicle or 200 µm 5HD, and were then exposed to vehicle or 200 nm staurosporine (STS) for 30 min. DCF fluorescence (a) and TMRE fluorescence (b) were quantified. Values are the mean and SD of determinations made in four to six separate cultures (20–30 neurons assessed per culture). **p < 0.01 (anova with Scheffe posthoc tests).

To determine the effects of 5HD on release of cytochrome c from mitochondria of neurons exposed to hypoxic and apoptotic insults, hippocampal cultures were pre-treated with 5HD and then exposed to either staurosporine or cyanide. Exposure of neurons to staurosporine resulted in release of cytochrome c from mitochondria in nearly all neurons during a 6 h exposure period (Fig. 9). 5HD treatment greatly decreased the number of neurons that released cytochrome c (Figs 9 and 10). Cyanide induced the release of cytochrome c from the mitochondria of the majority of neurons during a 6 h exposure period, and 5HD decreased the number of neurons in which cytochrome c was released from mitochondria in response to cyanide (Fig. 10a). The inhibitory effect of 5HD on cytochrome c release was confirmed by immunoblot analysis of the cytosolic and mitochondrial fractions of neurons that had been exposed to staurosporine alone or in combination with 5HD (Fig. 10c).

Figure 9.

5HD prevents cytochrome c release from mitochondria in hippocampal neurons subjected to an apoptotic insult. Cultures were pre-treated for 30 min with vehicle (control) or 100 µm 5HD. Cultures were then exposed to 200 nm staurosporine (STS) for 6 h. Cells were then double-labeled with Mitotracker Red and a cytochrome c antibody. Confocal images of Mitotracker Red fluorescence (red), cytochrome c immunoreactivity (green) and the merged image (yellow indicates sites of co-localization) are shown. Note that STS induced release of cytochrome c from mitochondria, and that 5HD prevented cytochrome c release from many neurons.

Figure 10.

5HD prevents cytochrome c release from mitochondria in hippocampal neurons exposed to hypoxia and staurosporine. (a) Confocal images of hippocampal neurons double-labeled with Mitotracker Red and a cytochrome c antibody in cultures that had been exposed for 6 h to 30 µm cyanide (CN) alone or in combination with 100 µm 5HD. Note that Mitotracker red and cytochrome c are not co-localized in CN-treated neurons, but are co-localized in most neurons co-treated with 5HD and CN, indicating that 5HD prevents cytochrome c release from mitochondria. (b) Results of determinations of the numbers of neurons that released cytochrome c in cultures exposed to staurosporine (STS) alone (200 nm) or in combination with 100 µm 5HD. Values are the mean and SD of determinations made in four to six cultures per condition (40–60 neurons assessed per culture). **p < 0.01. (c) Cytochrome c immunoblot of the cytosolic and mitochondrial fractions of neurons in an untreated control culture, and cultures exposed for 4 h to 200 nm STS alone or in combination with 100 µm 5HD. The graph shows the ratio of the relative amounts of cytochrome c in the cytosol compared with that in the mitochondria.

Collectively, these findings suggest that blockade of Mito-KATP protects neurons against apoptosis by suppressing oxyradical production, stabilizing mitochondrial membranes and preventing cytochrome c release.

5-Hydroxydecanoate decreases focal ischemic brain injury in mice

To determine the role of Mito-KATP in ischemic neuronal injury in vivo, we administered 5HD (20 mg/kg, i.v.) to mice either 30 min before or 30 min after permanent occlusion of the middle cerebral artery occlusion. The dose of 5HD was chosen based upon previous studies documenting the efficacy of this dose in blocking the ability of diazoxide to induce ischemic pre-conditioning in dogs (Shake et al. 2001). Mice were killed 24 h after middle cerebral artery occlusion, and infarct volume was quantified by analysis of brain sections stained with TTC. Infarct volume was significantly decreased in mice treated with 5HD 30 min prior to middle cerebral artery occlusion compared with control mice that received vehicle (Fig. 11). 5HD had no significant effect on infarct size when administered 30 min after middle cerebral artery occlusion (Fig. 11b). 5HD had no significant effect on basal cerebral blood flow (Fig. 12) and did not affect the reduction in cerebral blood flow following permanent occlusion of the middle cerebral artery (80–90% decreases in cerebral blood flow in all eight control and 10 5HD-treated mice analyzed), suggesting that the neuroprotective effect of 5HD was not the result of a vascular action.

Figure 11.

5HD reduces infarct volume in mice subjected to permanent focal cerebral ischemia. Mice were administered either saline (vehicle control) or 5HD (20 mg/kg) intravenously 30 min prior to, or 1 h after permanent middle cerebral artery occlusion. Mice were killed at 24 h after the onset of ischemia, and brains were removed, sectioned and stained with TTC. (a) Representative brain sections from a control mouse and a 5HD-pre-treated mouse are shown. Note decreased size of the cortical infarction in the mouse that received 5HD. (b) Infarct volumes were quantified. Values are the mean and SD (n = 6–10 mice per group). **p < 0.01 compared with the control value (anova with Scheffe posthoc test).

Figure 12.

5HD does not affect cerebral blood flow. Laser Doppler flowometry was used to measure cerebral blood flow (CBF) using a flexible 0.5 mm fiberoptic probe. Regional CBF was determined in the territory of the middle cerebral artery. (a) Representative recording in a mouse prior to, and after intravenous administration of 5HD (20 mg/kg). (b) Summary data. Values are the mean and SD (n = 5 mice).

Discussion

Data in the present study demonstrate that K+ release from mitochondria, and subsequent efflux from the cell, occurs in neurons subjected to several insults that induce apoptosis, including chemical hypoxia, glutamate and staurosporine. A surprising finding is that mitochondria apparently release K+ upon exposure to the Mito-KATP opener, diazoxide, or when exposed to the uncoupling agent, FCCP or the mitochondrial toxin, cyanide. 5HD prevented the release of K+ from mitochondria, consistent with efflux through Mito-KATP being a major route of K+ efflux. The cytosolic K+ concentration is known to be approximately 150 mm, but the mitochondrial matrix K+ concentration has not been established due to methodological difficulties; thus, conflicting data have arisen from different studies (Kowaltowski et al. 2002). Although the assumption has been that the intramitochondrial K+ concentration is lower than the cytosolic concentration (Debska et al. 2001), the most direct measurements suggest an intramitochondrial K+ concentration of 180 mm or greater (Garlid 1996).

Previous studies have provided evidence that efflux of K+ through plasma membrane K+ channels occurs in different cell types undergoing apoptosis in response to a range of apoptotic stimuli. For example, activation of plasma membrane K+ channels contributes to nitric oxide-induced apoptosis of vascular smooth muscle cells because K+ channel blockers (iberiotoxin and tetraethylamonium) inhibited the cell death (Krick et al. 2002). Electrophysiological analyses showed that K+ channels are activated in cultured cortical neurons during apoptosis induced by staurosporine and trophic factor deprivation (Yu et al. 1999a), and ceramide-induced apoptosis (Yu et al. 1999b). In the latter studies, K+ channel antagonists tetraethylammonium and clofilium attenuated the apoptosis, suggesting a key role for K+ efflux in the cell death process. The specific plasma membrane K+ channels activated during apoptosis are not fully established, although recent studies suggest that two-pore domain K+ channels mediate rapid K+ efflux during apoptotic volume decreases in embryonic cells (Trimarchi et al. 2002). Our data support the increasing evidence for the involvement of plasma membrane K+ channels in neuronal apoptosis. However, our K+ imaging data, and results obtained using mitochondrial K+ channel activators and blockers, suggest that K+ release from mitochondria is an early event that precedes and accompanies K+ efflux through plasma membrane channels.

The ability of 5HD to reduce ischemic brain injury in vivo and apoptosis of cultured neurons suggests a role for modulation of Mito-KATP in neuronal death cascades. The protective effect of 5HD against focal ishcemic brain injury was only observed when 5HD was administered prior to the onset of ischemia, suggesting that opening of Mito-KATP soon after the onset of ischemia plays an important role in the cell death cascade. 5HD had no effect on cerebral blood flow, suggesting that an effect of 5HD on the vasculature is unlikely to account for its neuroprotective action. During ischemia, Mito-KATP channels may be open and contribute to the increased ROS production and ATP depletion. By blocking Mito-KATP channels, 5HD would maintain mitochondrial membrane potential, decrease ROS production and promote maintenance of ATP levels, resulting in neuroprotection (Fig. 13). While our data suggest that inhibition of Mito-KATP is the major mechanism responsible for the neuroprotective actions of 5HD, it was recently reported that 5HD can be converted to 5HD-CoA by acyl-CoA synthetase (an enzyme present in mitochondria); 5HD-CoA has the potential to be an inhibitor of β-oxidation, which could contribute to the ability of 5HD to suppress ROS production (Hanley et al. 2002; Lim et al. 2002). High concentrations of 5HD (≥ 300 µm) have been reported to suppress mitochondrial respiration, an additional cytoprotective action (Jaburek et al. 1998; Lim et al. 2002).

Figure 13.

Model for the mechanisms whereby inhibition and activation of mitochondrial potassium channels modulate neuronal vulnerability to apoptotic cell death. Apoptotic stimuli, such as ischemia, glutamate and trophic factor deprivation, induce the activation of mitochondrial K+ channels which, by mechanisms that remain to be established, causes production of ROS, increased membrane permeability and release of cytochrome c from mitochondria, resulting in activation of caspases and apoptotic cell death. Opening of plasma membrane K+ channels results in efflux of cytosolic K+ and cell shrinkage during apoptosis. Subtoxic levels of activation of mitochondrial K+ channels (with diazoxide or mild ischemia, for example) results in K+ release and production of ROS, which induces a cellular stress response (pre-conditioning) and thereby increases the resistance of the neurons to various insults, including ischemic injury. Blockade of mitochondrial K+ channels (with 5HD, for example) suppresses K+ release and decreases ROS production, thereby stabilizing mitochondrial membranes and preventing apoptosis.

Previous studies have documented several changes in the mitochondria that occur in response to ischemia, including an increase in oxyradical production (Keller et al. 1998; Murakami et al. 1998), membrane depolarization or hyperpolarization (Budd et al. 2000) and release of cytochrome c (Ouyang et al. 1999; Sugawara et al. 1999; Noshita et al. 2001). In many cases of apoptosis, changes in membrane potential precede, and may be required for, release of cytochrome c (Kruman and Mattson 1999; Budd et al. 2000). We found that 5HD stabilized mitochondrial membrane potential in neurons exposed to apoptotic insults, consistent with a role for opening of Mito-KATP in the apoptotic process. We found that 5HD did indeed stabilize mitochondrial membrane potential and prevented cytochrome c release in neurons exposed to staurosporine or chemical hypoxia (cyanide exposure). Cells undergoing apoptosis typically exhibit increased production of ROS that may result from mitochondrial alterations, but may also play an important role in triggering mitochondrial membrane permeability changes. The ability of 5HD to reduce levels of ROS in cultured hippocampal neurons suggests a role for Mito-KATP activation in oxyradical production in neurons undergoing apoptosis. The latter interpretation is consistent with data showing that depolarization of mitochondria induces oxyradical production, as does exposure of cells to diazoxide, an opener of Mito-KATP (Samavati et al. 2002).

The neuroprotective effect of 5HD against focal ischemic brain injury in vivo and neuronal apoptosis in vitro was unexpected in the light of recent findings showing that activation of Mito-KATP with diazoxide protects cardiac myocytes (Sato et al. 2000) and neurons (Liu et al. 2002) against ischemic injury. Moreover, the latter studies showed that 5HD could block the cytoprotective effects of diazoxide. Emerging evidence suggests that diazoxide protects cells by inducing a mild stress response (Fig. 13). It was recently suggested that the pre-conditioning effect of diazoxide may itself require induction of oxidative stress (Samavati et al. 2002). Pre-conditioning by diazoxide may also involve activation of protein kinase C and mitogen-activated protein (MAP) kinases (Baines et al. 1999; Takashi et al. 1999). Activation of protein kinase C and MAP kinases can protect neurons against conditions relevant to ischemic brain injury, including oxidative stress, excitotoxins and hypoxia (Maiese et al. 1996; Han and Holtzman 2000; Skaper et al. 2001). In addition, it was recently reported that diazoxide can induce the expression of Bcl− 2 and its association with mitochondria in cultured hippocampal neurons, while preventing Bax translocation to mitochondria (Liu et al. 2002). Therefore, paradoxically, both activators and inhibitors of Mito-KATP can be neuroprotective.

A partial purification and reconstitution of brain Mito-KATP was recently reported, and a comparison of relative levels of the putative Mito-KATP proteins in heart and brain suggest that their levels are approximately sixfold higher in brain (Bajgar et al. 2001). We found that changes in activity of Mito-KATP can modulate ROS production and mitochondrial membrane potential in neurons. Collectively, the emerging data therefore suggest that Mito-KATP may play particularly important roles in neuronal physiology as well as in pathophysiological conditions where mitochondrial alterations are prominent. Importantly, the present findings suggest that drugs that inhibit mitochondrial K+ channels may useful in the treatment of stroke and related neurodegenerative conditions.

Ancillary