ERK signaling leads to mitochondrial dysfunction in extracellular zinc-induced neurotoxicity

Abstract

A zinc-induced signaling pathway leading to ERK1/2 activation and subsequent neuronal death has been investigated. We find that an extracellular zinc application stimulates biphasic phosphorylation of ERK1/2 and p38 MAPK in rat cultured neurons. The activation of ERK1/2, but not p38, is responsible for zinc neurotoxicity as only U0126, a MEK inhibitor that blocks ERK1/2 phosphorylation, significantly protects cortical neurons from zinc exposure. Overexpression of a dominant negative Ras mutant blocks zinc-induced Elk1-dependent gene expression in neurons, indicating the involvement of Ras activation in the zinc pathway leading to ERK phosphorylation and Elk1 signaling. We also find that zinc treatment results in neuronal mitochondrial hyperpolarization. Importantly, both U0126 and bongkrekic acid, an inhibitor of the mitochondrial adenine nucleotide translocase (ANT), effectively reduce zinc-triggered mitochondrial changes. As bongkrekic acid also prevents zinc-triggered neuronal death but not ERK1/2 phosphorylation, activation of MAPK signaling precedes and is required for mitochondrial dysfunction and cell death. These results provide new insight on the mechanism of extracellular zinc-induced toxicity in which the regulation of mitochondrial function by the Ras/MEK/ERK pathway is closely associated with neuronal viability.

Introduction

Zinc is the second most abundant transition metal in the body, and it exists in high concentrations within the brain (Sensi et al. 2009). Although the total concentration of cellular zinc is 100–500 μM, the level of intracellular freely available zinc in the picomolar range (Bozym et al. 2006). Besides participating in cell growth and development as an important structural and functional component in a large number of proteins, zinc can function as a neuromodulator to affect synaptic signaling (Besser et al. 2009; Sensi et al., 2009).

Unregulated increases in intracellular free zinc, which can severely impair zinc homeostasis, are toxic to neurons, and have been observed to occur following transient global ischemia, traumatic brain injury, and epileptic seizures, among other disorders (Sensi et al., 2009). In these situations, high concentrations of free zinc can translocate from presynaptic zinc-containing vesicles into postsynaptic neurons. However, mice lacking vesicular zinc still show zinc accumulation in degenerating neurons following excitotoxic stimuli, suggesting that zinc may be released from cellular locations other than synaptic vesicles (Lee et al. 2000). Indeed, under oxidative or nitrosative stress, cysteines in zinc thiolate complexes in intracellular proteins can be modified. As a result, zinc is liberated, also leading to increases in cytoplasmic free zinc and subsequent neuronal injury (Aizenman et al. 2000; Hershfinkel et al., 2009).

Various molecular mechanisms have been proposed for zinc neurotoxicity. Zinc has been shown to directly inhibit the mitochondrial respiratory chain, cause the opening of the mitochondrial permeability transition pore (mPTP), and initiate the release of pro-apoptotic proteins (Jiang et al. 2001; Jonas 2009). In addition, rising intracellular zinc also promotes the generation of reactive oxygen species (ROS) in cells by activating NADPH oxidase, and by triggering mitochondria superoxide production (Sensi et al. 1999; Noh and Koh 2000; Zhang et al. 2004). Importantly, many neuronal signaling molecules are activated during zinc neurotoxicity, including extracellular signal-regulated kinase 1/2 (ERK1/2) (Park and Koh 1999; Du et al. 2002).

ERK1/2 activation is achieved via its phosphorylation at both tyrosine and threonine residues, primarily by the action of the upstream kinases MEK1/2, themselves targets of Ras activation. Phosphorylateded ERK1/2 can translocate to the nucleus, where it regulates the function of transcription factors such as Elk1. To sustain proper ERK1/2 activity during normal cellular functions, cells have to maintain a critical balance between ERK1/2 phosphorylation and dephosphorylation. Increasing evidence now suggests that activation of ERK1/2 can induce cell death under oxidative stress in a zinc-dependent fashion (Ho et al. 2008).

Ras is an important upstream regulator of ERK1/2 phosphorylation, and the Ras/MEK/ERK signaling pathway is well characterized as a major kinase cascade in many cell types. Studies suggest that Ras activation in PC12 cells is involved in intracellular processes of cell death initiated by different stimuli, via either non-apoptotic (Yan and Greene 1998) or apoptotic signaling pathways (Ferrari and Greene 1994). Importantly, zinc is able to induce Ras-dependent ERK1/2 phosphorylation in PC12 cells (Seo et al. 2001). Other studies have shown that Ras-Raf-MEK-dependent oxidative cell death is associated with the activation of mitochondrial voltage-dependent anion channels (VDACs) in non-neuronal cells (Yagoda et al. 2007), suggesting that mitochondrial membrane potential alterations may play an important role in MAPK-dependent cell death.

In spite of this vast amount of information, it has never been determined whether ERK-dependent zinc neurotoxicity has a mitochondrial component. More importantly, however, it is unclear whether zinc-dependent alterations in mitochondrial function are always the sole result of direct actions of the metal on this organelle, or whether intermediate signaling events contribute to this process. Here, we provide evidence that exogenous zinc stimulates the Ras/MEK/ERK1/2/Elk1 signaling cascade in primary cortical cultures. Furthermore, we show that ERK1/2 activation is required for zinc-triggered mitochondrial hyperpolarization, leading to neuronal cell death.

MATERIALS AND METHODS

Materials

Rabbit anti-phospho-ERK1/2, anti-phospho-p38 MAPK, and anti-p38 MAPK were obtained from Cell Signaling (Beverly, MA); mouse anti-ERK1/2 monoclonal antibody were from BD Biosciences/Transduction Laboratories (San Diego, CA); a horseradish peroxidase (HRP)-conjugated goat secondary antibody against either mouse or rabbit IgG was from Jackson ImmunoResearch (West Grove, PA). SB239063 and U0126 were from Tocris (Ellisville, MO). FTI-2628 was from Calbiochem (La Jolla, CA). Bicinchoninic acid (BCA) based protein assay reagents were from Pierce (Rockford, IL). Both wild-type Ras (RasWT) and dominant negative Ras mutant (RasN17) constructs were kindly provided by Dr. Guillermo Romero (University of Pittsburgh School of Medicine, USA). Unless specified, all other chemicals used were of analytical grade quality or better and were obtained from Sigma-Aldrich (St. Louis, MO).

Cell culture

Mixed neuronal-gial cell cultures were prepared from the cerebral cortex of embryonic day 16–17 Sprague-Dawley rat fetuses (Hartnett et al. 1997). Mixed cultures were used in the studies requiring transfection, normally conducted after 3 weeks in vitro. For neuron-enriched primary cortical cultures, cells were seeded onto either PLO-coated culture plates or dishes at a density of 1.5 × 10⁵ cells/cm². ARA-C was applied to cell cultures 3 days after plating. Neuron-enriched cultures were used for cellular toxicity assay and protein immunoblotting experiments at 12–14 days in vitro. The viability of these cultures was very poor after transfection procedures.

Cell viability assay

Neuronal cultures were treated with zinc and/or inhibitors for designated times, rinsed and returned to the incubator. The incubation medium (phenol red-free minimum essential medium, 10 mM HEPES) contains amino acids as well as low concentration of added bovine serum albumin (0.1%; used to protect cells from extensive media changes). As these molecules can bind zinc, the exact concentrations of free zinc in the medium are not known, but were likely substantially lower than the added amounts. Neuronal viability was determined overnight following zinc treatment, by measuring the activity of lactate dehydrogenase (LDH). LDH values were always normalized to those obtained in sister cultures exposed overnight to 200 μM kainic acid.

ELK1 luciferase reporter assay

The PathDetect Elk1 trans-Reporting System (Stratagene, La Jolla, CA) was used to detect zinc-regulated gene expression in cortical neurons. Briefly, cells were co-transfected with plasmids of pFR-luc reporter, pFA2-Elk1 fusion trans-activator, Renilla luciferase reporter (pRL-TK; Promega, Madison, WI), and one of the following constructors: Ras N17, Ras WT or parent vector using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Renilla luciferase reporter plasmid served as an internal, non-inducible reporter standard to account forvariations in transfection efficiency. Forty-eight hours later, transfected neurons were exposed to zinc (100 μM) for 30 min. Elk1-related luciferase activity was measured ~18 h later using the Dual-Glo luciferase assay system (Promega, Madison, WI). Results were expressed as a ratio of firefly luciferase activity to Renilla luciferase activity.

Western blot analysis

Neurons were briefly rinsed twice with ice-cold PBS and were then gently scraped off the dishes on ice after being exposed to cell lysis buffer supplemented with a protease inhibitor cocktail (Biosource, Camarillo, CA). Debris was pelleted by centrifugation for 10 min, and the remaining lysates were used immediately or stored at −20 °C. The protein concentrations of the samples were measured with a BCA assay. Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) was carried out using the Mini Protean 3 System (Bio-Rad, Hercules, CA). Prior to electrophoresis, protein samples were treated with a reducing sample buffer and boiled at 100 °C for 5 min. Samples with equal amounts of protein were run on 7.5% SDS-PAGE gel. For immunoblotting, separated protein bands were transferred onto a 0.2 μm nitrocellulose membrane (Bio-Rad). The membranes were then blocked with 1% BSA in PBS with 0.05% Tween 20 (PBST) at room temperature for 1 h, and probed with appropriated primary antibodies diluted in PBST. After washing (3×) in PBST, blots were incubated with goat secondary antibody conjugated to horseradish peroxidase at room temperature for 1 h. Blots were then visualized using SuperSignal West Pico Chemiluminescent Substrate Kit (Pierce, Rockford, IL). Quantification was performed with Scion Image software (Scion, Frederick, MD).

Mitochondrial Membrane Potential (ΔΨ_m)

Changes in the mitochondrial membrane potential in cortical neurons were assessed using 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazole-carbocyanine iodide (JC-1; Molecular Probes, Eugene, OR). JC-1, a cationic dye, accumulates in mitochondria in a membrane potential-dependent fashion and forms orange-red fluorescent aggregates. JC-1 outside the mitochondria exists as a green fluorescent monomer. As such, this dye is useful for monitoring relative changes in ΔΨ_m in neurons (Buckman and Reynolds 2001). When excited at 488 nm, the dye emits with peaks at 530 (green) and 590 nm (red). Fours hours following zinc treatment neuron-enriched cultures were incubated with 1.5 μM JC-1 for 20 min at 37 °C. Neurons were then examined using an inverted epifluorescence microscope. Dual emission images (green and red) were obtained for every condition. In addition, a phase-contrast imagine from the same field was always acquired. All digital images were stored in memory for later analysis. NIH ImageJ software was used to quantify the fluorescence signal density from for both red and green images. Background-subtracted fluorescence values were collected from 10–15 fields from a single culture for both red and green channels. The mean red to green ratio for each culture was then calculated. Data were collected and averaged for three independent experiments.

Results

ERK1/2 mediates extracellular zinc neurotoxicity

Neurons were exposed to increasing nominal concentrations of extracellular ZnCl₂ (30 to 300 μM) for 30 min (see Materials and Methods). Zinc treatment resulted in a concentration-dependent decrease in cell viability (Fig. 1a), which began to be statistically significant at a concentration of 100 μM. At 300 μM, zinc killed most neurons, consistent with prior work (e.g. Kim et al. 1999; Park and Koh, 1999). We next determined whether exposure to toxic concentrations of zinc triggered phosphorylation of ERK1/2 and p38 MAPK. We observed that the activation of ERK1/2 and p38 MAPK in neurons was also concentration-dependent (Fig. 1b). A time course analysis of the phosphorylation of these MAPKs showed that both ERK1/2 and p38 MAPK were rapidly activated following zinc exposure (Fig. 1c). ERK1/2 phosphorylation first peaked after 30 min of zinc treatment (i.e. measured immediately after removing zinc; 0 min of post-zinc period in Figs. 1c and d), quickly decreased, and returned to the basal levels within 1 h. Three hours later, however, a second wave of ERK1/2 phosphorylation re-appeared, and it persisted up to 6h following zinc exposure (the longest time period we tested in the experiments). The activation of ERK1/2 by toxic zinc exposure was not mediated by stimulation of the recently described metabotropic neuronal zinc-sensing receptor ZnR/GPR39 (Besser et al., 2009) as phopholipase C inhibitors (e.g. U-73122), which prevent the downstream signaling of this receptor, were without effect (data not shown). Finally, the time course of the activation of p38 MAPK showed a similar pattern to ERK1/2 phosphorylation, except it first peaked at 15 min following zinc removal.

Fig. 1.

Zinc induces neuronal cell death and MAPK activation in rat primary cortical cultures. (a) Neuron-enriched cultures seeded on 24-well plates were exposed to zinc at various concentrations for 30 min and assayed for cell death 24 hr later. Toxicity is expressed as percent of maximal cell death (LDH activity) induced by overnight exposure to kainic acid (200 μM). Values represent the mean ± SEM from three independent experiments performed in quadruplicate (** p< 0.01, compared with control; ANOVA/Dunnett). (b) Concentration-dependent activation of ERK1/2 and p38 by zinc in neurons. Cortical neurons were exposed to 100 and 200 μM zinc for 30 min. Cells were then rinsed and returned to the incubator. Three hours later, cells were harvested and samples were used for western blot analysis. The phosphorylation of ERK1/2 (pERK1/2) and p38 MAPK (pp38) was detected using rabbit antibodies specific against phosphorylated ERK1/2 and p38. Non-phosphorylated ERK1/2 and p38 were also probed in identical membranes and served as their respective loading controls. Results shown are representative of three independent experiments. (c) Time courses of pERK1/2 and pp38 activation in neurons following zinc treatment. Cell cultures were treated with 100 μM zinc for 30 min. After zinc was removed, neurons were incubated for various times as indicated. Samples were collected, and analyzed as described above. Results shown are representative of three independent experiments. (d) Quantification of the time course of ERK1/2 phosphorylation induced by zinc, calculated from the ratio of pERK1/2 to total ERK1/2. The band densities of pERK1/2 and ERK1/2 were measured using Scion Image software (Scion, Frederick, MD) and normalized to controls (ctrl). The data are expressed as mean ± SEM from three independent experiments (*p< 0.05, compared with control; ANOVA/post-hoc. one sample test).

To investigate the roles of ERK1/2 and p38 MAPK activation in zinc-triggered neuronal death, we used chemical inhibitors to specifically block these signaling pathways. We first tested the effect of the selective MEK1/2 inhibitor U0126 (3–10 μM), and observed it significantly reduced zinc-induced cell death when added before and during zinc exposure (Fig. 2a). However, when U0126 was added to the cortical cultures 1 h after the removal of zinc, before the second wave of ERK1/2 phosphorylation began, its protective effect against zinc toxicity was lost (Fig. 2b). This suggests that the initial stage of MAPK activation initiates a series of events that culminate with the demise of the cells. We previously reported that a selective p38 MAP kinase inhibitor, SB239063 (20 μM) could block intracellular liberated zinc toxicity in cortical neurons (McLaughlin et al. 2001). However, we did not observe a similarly protective effect of this drug on extracellular zinc-induced toxicity (Fig. 2c). Our results thus suggest that ERK1/2 and not p38 MAPK, plays a critical role in the neurotoxicity caused by extracellular zinc exposure, as reported earlier (Park and Koh, 1999).

Fig. 2.

Zinc neurotoxicity is ERK1/2 dependent. Neuronal cultures were treated with vehicle or zinc (100 μM) for 30 min. The MEK inhibitor U0126 was present either before, during and after zinc exposure (a), or added only 1 hr. following metal exposure (b). (c) The p38 inhibitor SB239063 was present before, during and after zinc exposure. Toxicity was expressed as a percent of maximal cell death induced by overnight exposure to 200 μM kainic acid. Data represent the means ± SEM of at least three independent experiments performed in quadruplicate. ^Δp< 0.05, ^ΔΔp< 0.01, and ^ΔΔΔp< 0.001, compared with untreated controls (vehicle, open bar), whereas *p< 0.05, compared to zinc-treated controls in the absence of inhibitors (filled bar). The experimental data were analyzed by paired, two-tailed student t tests. (d) Effects of MAPK inhibitors on ERK1/2 phosphorylation in zinc-treated cortical neurons. Neuronal cultures were treated with vehicle or zinc (100 μM) for 30 min in the presence of either U0126 (10 μM) or SB239063 (20 μM) before, during and after zinc exposure. Three hours later, samples were harvested. The levels of pERK1/2 and ERK1/2 were detected by western blotting by using specific antibodies against their respective phosphorylated and non-phosphorylated proteins. Data shown are representative of three independent experiments with similar results.

The effects of these MAPK inhibitors on ERK1/2 phosphorylation were also examined. As expected, zinc-induced increases in ERK1/2 phosphorylation were suppressed by U0126 (10 μM) (Fig. 2d). Interestingly, SB239063 (20 μM) strongly stimulated the phosphorylation of Erk1/2 in both control and zinc-treated neurons (Fig. 2d). Because SB239063 alone had no effect on the cell viability (Fig. 2c), the induction of ERK1/2 phosphorylation by this drug is probably mediated by a molecular mechanism not related to zinc toxicity. The concentration range of the MAPK inhibitors utilized is within the range used by other investigators. For U0126, we chose a concentration range that was similar to those used by our group in an earlier study (Du et al., 2002), which is similar to the effective concentration used in the initial description of this drug (Favata et al., 1998). This drug has very little or no effect on the activities of other kinases such as protein kinase C, Abelson tyrosine kinase, Raf, etc. Similarly, SB239063 is a selective p38 inhibitor with 220-fold less activity at ERK, JNK and other kinases (Tocris Biosciences product information).

Ras mediates zinc-induced activation of ERK1/2

We utilized the PathDetect Elk1 trans-reporting system to measure zinc-mediated activation of the Ras/MEK/ERK pathway in neurons. This system contains a fusion trans-activator Elk1 plasmid (pFA2-Elk1) and a firefly luciferase reporter plasmid (pFR-Luc). Once expressed in the cells, the fusion Elk1 protein can be phosphorylated by activated ERK1/2, subsequently activating transcription of the luciferase gene from the pFR-Luc plasmid. Forty-eight hours following transfection with pFA2-Elk1 and pFR-Luc, neurons were treated with zinc (100 μM) for 30 min and assayed for luciferase expression the next day. We found that zinc stimulation increased Elk1-directed luciferase activity nearly two-fold, when compared to vehicle-treated controls (Fig. 3). To determine whether Ras is involved in zinc-induced Elk1 activation, we pre-treated transfected neurons with FTI-2628 (1 μM), a cell-permeable inhibitor of protein farnesyltransferase that has been shown to block Ras activation (Kowluru et al. 2010). We observed, however, that FTI-2628 only slightly decreased (<10%) zinc-induced Elk1 activation (data not shown). Unfortunately, we could not test higher concentrations of this compound as the vehicle itself, DMSO, was observed to potently activate ERK1/2 phosphorylation and Elk1 regulated gene expression. As such, we used a dominant negative mutant Ras (RasN17) (Rizzo et al. 2000) or wild type Ras construct (RasWT, as positive control), transfecting them together with the PathDetect components. Overexpression of RasN17 mutant suppressed zinc-induced Elk1 activation (Fig. 3). In contrast, RasWT expression in neurons had little effect on zinc-induced increase of Elk1-regulated gene expression, primarily because it had profoundly elevated its overall basal level of activation, thereby mostly occluding any potential actions of the metal. These results demonstrate that Ras is a key upstream activator responsible for zinc-mediated ERK1/2 and Elk1 activation.

Fig 3.

Ras-mediated activation of Elk1 following neuronal zinc exposure. Cortical neurons co-transfected with either a blank parent vector, a RasN17 dominant negative vector, or a Ras wild-type (WT) construct were exposed to vehicle or zinc for 30 min and Elk1-related firefly luciferase activities were measured 18 hours later. The ratios of firefly luciferase activity to renilla luciferase activity, representing Elk1 expression in neurons, were normalized to vehicle-treated control (open bar). Data represent the mean ± SEM from five independent experiments performed in quadruplicate. ^ΔΔΔp< 0.001, compared with the untreated control (vehicle, open bar) and analyzed by one sample, two-tailed t test; whereas *p< 0.05, compared to the control treated with zinc (filled bar) and analyzed by paired, two-tailed student t test.

Zinc-triggered mitochondrial dysfunction is ERK1/2-dependent

Mitochondrial dysfunction has been closely linked to pathological processes, including zinc-mediated neuronal injury (Weiss et al. 2000; Sensi et al. 2009). Of interest, a study demonstrated a physical association between Elk1 and the mitochondrial adenine nucleotide transporter (ANT), as well as other proteins thought to be associated with the mPTP complex (Barrett et al. 2006a). In addition, that study showed that overexpression of Elk1 can increase neuronal cell death, and that Elk1-mediated toxicity could be inhibited with bongkrekic acid, a mitochondrial ANT inhibitor (Barrett et al. 2006a). We thus evaluated whether bongkrekic acid could also block zinc toxicity in our system. Cortical neurons were incubated with 100 μM zinc (30 min) in the presence of bongkrekic acid (3 and 10 μM). As shown in Fig. 4a, neuronal cell death was significantly decreased by bongkrekic acid in a concentration-dependent manner. Bongkrekic acid (10 μM), however, had no effect on zinc-induced ERK1/2 phosphorylation, either immediately following zinc treatment (Fig. 4b), or 3 h later (data not shown), indicating that ERK1/2 activation must occur upstream from mitochondrial dysfunction following metal exposure.

Fig. 4.

Bongkrekic acid attenuates zinc-induced neuronal death, but not ERK1/2 activation. (a) Neuron-enriched cultures were treated with zinc (100 μM) for 30 min. Bongkrekic acid (10 μM) was added into the cultures 20 min before, during and after zinc treatment. LDH activity in the media was measured 24 hr later, and normalized to 200 μM kainic acid-treated sister cultures. Data represent the mean ± SEM from four independent experiments performed in quadruplicate. ^ΔΔp< 0.01, compared with untreated controls (vehicle, open bar), whereas **p< 0.01, compared to zinc-treated controls in the absence of bongkrekic acid (filled bar). The experimental data are analyzed by paired, two-tailed student t test. (b) Bongkrekic acid has no effect on zinc-mediated ERK1/2 activation in neurons. Neuronal cultures were treated with zinc (100 μM) for 30 min in the absence or presence of bongkrekic acid (10 μM) before and during zinc exposure. Samples were then harvested at the end of zinc exposure. Phosphorylated ERK1/2 and total ERK1/2 proteins were detected as described in Fig. 2(d). Data shown are representative of three independent experiments with similar results.

Mitochondrial changes leading to cell death are closely related to alterations in the ΔΨ_m. We therefore used the dye JC-1 to investigate changes ΔΨ_m in cortical neurons after zinc exposure. To simplify this experiment we chose to focus on a single time point, 4 h after zinc removal. As shown in Fig. 5, a pronounced increase in red JC-1 fluorescence was observed in zinc-treated cells, when compared to controls, suggesting a profound level of mitochondrial hyperpolarization. To investigate whether the observed mitochondrial changes were the result of ERK activation we exposed cells to zinc in the presence of U0126 (10 μM). We found this drug to strongly suppress zinc-induced changes in ΔΨ_m, indicating once more that mitochondrial dysfunction was dependent on MAPK activation. Finally, we confirmed that bongkrekic acid, at a neuroprotective concentration (10 μM), could also effectively prevent changes to ΔΨ_m following zinc exposure (Fig. 5), which was indeed the case. To validate the ability of JC-1 to monitor changes in ΔΨ_m, zinc-treated neurons loaded with JC-1 were also exposed to oligomycin (4 μg/ml), an ATPase synthase inhibitor that induces a pronounced mitochondrial hyperpolarization. The red to green JC-1 fluorescence intensity ratio was measured in these cells, also at the 4-hour time point. As expected, naïve cells treated with oligomycin alone showed a pronounced hyperpolarization (Fig. 6a), demonstrated by a significant, nearly two-fold increase in red to green JC-1 fluorescence ratio (Fig. 6b). Zinc (100 μM) alone produced a very similar change in the red/green fluorescence ratio, essentially occluding any additional actions of oligomycin in these cells (Figs. 6a and b).

Fig. 5.

Mitochondrial hyperpolarization induced by zinc exposure in neurons is attenuated by U0126 and bongkrekic acid (BKA). Neuron-enriched cortical cultures were exposed to either vehicle or zinc (30 min; 100 μM) in the presence or absence of either U0126 (10 μM) or BKA (10 μM) before, during and after zinc exposure. Four hours following zinc treatment, cultures were incubated with JC-1 (1.5 μM) for 20 min at 37 °C and visualized under a fluorescent microscope (bottom panels). Phase-contrast micrographs of the same field were also taken (top panels). Note that zinc-induced changes in JC-1 fluorescence, denoting mitochondrial hyperpolarization, are absent in cultures exposed to both U0125 and BKA. Images are representative of three independent experiments.

Fig. 6.

Zinc occludes oligomycin-induced mitochondrial hyperpolarization. Naïve or zinc-treated cortical neuronal cultures were loaded with JC-1 as described in Fig. 5, and then exposed to oligomycin (4 μg/ml). (a) Representative images of JC-1 fluorescence in show pronounced hyperpolarized mitochondria after oligomycin, zinc, and zinc plus oligomycin treatments. (b) Quantification of JC-1 red to green fluorescence ratios. Data are expressed as means ± SEM from three independent experiments and analyzed by paired, two-tailed student t test. We noted a pronounced increase in red to green fluorescence ratio in both zinc and oligomycin-treated cultures (**p< 0.01 and ^Δp< 0.05, compared with vehicle control (open bar), whereas no significant difference (N.S.) in fluorescence ratio was observed in zinc-treated neurons in the absence or presence of oligomycin.

Discussion

In this study, we investigated the role of the Ras/MEK/ERK1/2/Elk1 signaling pathway in zinc-induced toxicity in primary rat neuronal cultures. Zinc treatment stimulated biphasic phosphorylation of ERK1/2 and p38 MAPK. Suppression of ERK1/2 phosphorylation, but not p38, significantly protected neurons from zinc toxicity, suggesting that ERK1/2 activation is a key event associated with zinc-induced neuronal death. Moreover, overexpression of a dominant negative Ras mutant completely blocked activation of the ERK1/2-dependent transcription factor Elk1, demonstrating that Ras is the upstream kinase responsible for zinc-mediated activation of the MAPK. Furthermore, bongkrekic acid, a mitochondrial ANT inhibitor, also effectively reduced neuronal damage in this zinc toxicity model. Both U0126 and bongkrekic acid prevented zinc-mediated mitochondrial hyperpolarization in cortical neurons, suggesting that zinc neurotoxicity triggers a injurious signaling cascade linking ERK1/2 activation to mitochondrial dysfunction.

The ERK1/2 signaling cascade is an important intracellular pathway responding to extracellular stimuli such as growth factors, normally transmitting signals from cell surface receptors to nuclear effectors. In addition to its role in the regulation of cell proliferation, differentiation, migration, and survival, ERK1/2 has emerged as a key mediator of neuronal death (Cheung and Slack 2004). Extracellular zinc can enter neurons through different routes, including NMDA receptors, voltage-gated calcium channels, and calcium-permeable AMPA receptors (Sensi et al., 2009). As cytosolic zinc rises, excess zinc is thought to be gradually taken up by mitochondria, inducing ROS generation and eventually dysfunction (Sensi et al. 1999; Sensi et al. 2000; Dineley et al. 2005). It has been shown that mitochondria-produced ROS are able to activate ERK1/2 in non-neuronal cells, albeit by a yet undefined process (Samavati et al. 2002). In addition, zinc released from intracellular stores caused by glutathione depletion selectively blocks ERK1/2 phosphatases, contributing to the persistent activation of ERK1/2 and consequent neuronal death in HT22 neuroblastoma cell line and immature cortical neurons (Ho et al. 2008). However, as a single independent factor, neither ERK1/2 activation (Luo and DeFranco 2006) nor ERK phosphatase inhibition (Levinthal and Defranco 2005) is sufficient to produce toxicity in neurons. This evidence suggests that both zinc-dependent ERK1/2 activation and phosphatase inhibition may be simultaneously required for toxicity to sometimes ensue. The Zn-mediated activation of the Ras pathway reported here combined the previously described inhibition of phosphatases by intracellular zinc (Ho et al., 2008) may therefore account for the delayed and sustained activation of the MAPK pathway observed in our study. Regardless, the late ERK activation, in and of itself, is not sufficient for extracellular zinc toxicity, as only inhibition of the early phase of the MAPK activation can abrogate the lethal actions of the metal.

The mechanism by which MEK1/2 typically activates ERK1/2 is initiated by Ras. Moreover, Elk1, a transcription factor, is a primary downstream target of activated ERK1/2. The contribution of the Ras/MEK/ERK1/2 pathway in the mediation of zinc-induced neuronal Elk1 activation shown in our study is in agreement with zinc-dependent ERK1/2 phosphorylation in differentiated PC12 cells and IIC9 embryonic fibroblasts (Seo et al. 2001; Klein et al. 2006). How an increase in intracellular zinc leads to Ras activation in neurons remains undefined, although it could be the result of the aforementioned zinc-induced actions on the mitochondria. Consistent with this hypothesis, Ras is a reported target of different oxidants and serve as a sensor of cellular redox stress (Lander et al. 1995). Superoxide produced as a result of mitochondrial dysfunction can be readily converted to long-lived hydrogen peroxide (Bao et al. 2009), likely contributing to oxidative signaling processes. In addition, mitochondrial ROS have been suggested as a trigger for NADPH oxidase (Lee et al. 2006), and inhibitors of this enzyme, as well as antioxidants, reduce zinc-induced cellular ROS generation and neurotoxicity (Kim et al. 1999; Noh and Koh 2000).

In non-neuronal cells Ras can translocate to the mitochondria following its activation by cell death signals (Bivona et al. 2006). Moreover, Ras, MEK, ERK1/2, and Elk1 all have been found to exist in mitochondria in a variety of tissues, including the brain (Nowak 2002; Alonso et al. 2004; Barrett et al. 2006a). It is therefore possible that these kinases may phosphorylate and activate their substrates entirely within this organelle. The expression of Elk1 in the brain in both nuclear and cytoplasmic compartments is exclusively neuronal, and is coincident with active ERK1/2, suggesting phosphorylated Elk1 could also function outside the nucleus in neurons (Sgambato et al. 1998). Indeed, the association of cytoplasmic Elk1 with the ANT and other mPTP-related proteins has been shown to be intimately involved in hippocampal cell death induced by DNA-damaging agents (Barrett et al. 2006a), with bongkrekic acid found to be neuroprotective in that model. Likewise, local Elk1 transcription and translation in dendrites has been shown to induce cell death (Barrett et al. 2006b). Therefore, the involvement of Elk1 may provide a critical link between ERK1/2 activation, mitochondrial dysfunction and cell death following zinc exposure.

Of course, it is entirely possible that zinc-mediated ERK1/2 activation, either directly or indirectly, modulates mitochondrial proteins that then affect mitochondrial function. The mPTP complex has three major components: cyclophilin D, the mitochondrial voltage-dependent anion channel (VDAC), and ANT (Galluzzi et al. 2009). Mitochondrial VDAC is physically associated to ANT in the regulation of the mPTP (Halestrap and Brennerb 2003), and activated Ras-Raf-MEK has been recently found to lead to oxidative cell death through VDAC (Yagoda et al. 2007). In ischemic brain, zinc induces VDAC activity in the mitochondrial outer membrane (Bonanni et al. 2006). Therefore, bongkrekic acid could also prevent zinc-induced conformational changes in the ANT via direct binding to and/or inhibition of its target. Regardless of the mechanism, the neuroprotective effects of bongkrekic acid shown here are consistent with previous observations in depolarized cortical neurons exposed to zinc (Jiang et al. 2001). Our results, however, finally provide experimental evidence for an association between ERK signaling to a bongkrekic-acid sensitive mitochondrial injurious process in zinc neurotoxicity.

ΔΨ_m is an important marker of mitochondrial function, and its alteration plays a key role in the regulation of cell death in many cell types, including neurons. In rat hippocampal neurons and human D283 medulloblastoma cells, staurosporine-induced mitochondrial hyperpolarization is an early event associated with cell death (Poppe et al. 2001). Mitochondral hyperpolarization is sometimes reversible and could reflect an attempt by the injured cells to restore their bioenergetics (Perl et al., 2004). Here, we found zinc stimulated mitochondrial hyperpolarizaton in cortical neurons within a few hours after metal exposure. This phenomenon, prevented by both U0126 and bongkrekic acid, correlated well with decreased neuronal survival, and thus we believe that the hyperpolarization event preceded eventual mitochondrial dysfunction, leading to cell death (Skulachev, 2006). Since bongkrekic acid had no effect on zinc-induced ERK1/2 activation, MAPK activation must occur upstream form mitochondrial dysfunction.

Alternatively, the ANT or other bongkrekic acid-sensitive proteins may function independent of ERK1/2, but must cross-talk with ERK1/2-directed proteins, such as egr1 (Park and Koh, 1999), in the mitochondria to jointly regulate ΔΨ_m. MEK and ANT inhibitors have been shown to protect brain tissues from brain injury (Namura et al. 2001; Weaver et al. 2005), and both these types of injury have been separately linked to zinc neurotoxicity (Sensi et al. 2009). Therefore, targeting the Ras/MEK/ERK signaling cascade, as well as mPTP regulators may provide a mechanistically justified approach to treat brain disorders associated with zinc dysregulation.

Abbreviations

ERK1/2

extracellular signal-regulated kinase 1/2

MAPK

mitogen-activated protein kinase

LDH

lactate dehydrogenase

SDS-PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis

JC-1

5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazole-carbocyanine iodide

ROS

reactive oxygen species

ΔΨ_m

mitochondrial membrane potential

mPTP

mitochondrial permeability transition pore

ANT

adenine nucleotide translocator