Mitochondria have been identified as targets of the neurotoxic actions of zinc, possibly through decreased mitochondrial energy production and increased reactive oxygen species accumulation. It has been hypothesized that impairment of mitochondrial trafficking may be a mechanism of neuronal injury. Here, we report that elevated intraneuronal zinc impairs mitochondrial trafficking. At concentrations just sufficient to cause injury, zinc rapidly inhibited mitochondrial movement without altering morphology. Zinc chelation initially restored movement, but the actions of zinc became insensitive to chelator in <10 min. A search for downstream signaling events revealed that inhibitors of phosphatidylinositol (PI) 3-kinase prevented this zinc effect on movement. Moreover, transient inhibition of PI 3-kinase afforded neuroprotection against zinc-mediated toxicity. These data illustrate a novel mechanism that regulates mitochondrial trafficking in neurons and also suggest that mitochondrial trafficking may be closely coupled to neuronal viability.
- green fluorescent protein
- organelle transport
- signal transduction
- mitochondrial membrane potential
- oxidative stress
Neural zinc is a highly regulated ion, much of which is tightly coordinated to proteins, leaving a small pool of zinc that remains free and mobile (Frederickson, 1989). Elevated intracellular free zinc ([Zn2+]i) results from entry through several Ca2+-permeable pathways (Weiss et al., 1993; Koh and Choi, 1994; Yin and Weiss, 1995; Sensi et al., 1997; Cheng and Reynolds, 1998) and also from oxidant-mediated release from intracellular stores (Aizenman et al., 2000). Elevated [Zn2+]i results in neuronal injury in vitro (Choi et al., 1988; Koh and Choi, 1994), and it has been suggested that its accumulation may contribute to neurodegeneration associated with ischemia (Koh et al., 1996; Lee et al., 2002), epileptic seizures (Frederickson et al., 1988), and head trauma (Suh et al., 2000).
Although an unambiguous mechanism for Zn2+-mediated neurotoxicity has not been identified, several lines of evidence suggest mitochondria and energy metabolism as subcellular targets for the toxic actions of [Zn2+]i (Dineley et al., 2003; Sensi and Jeng, 2004). Zinc can inhibit glycolysis (Sheline et al., 2000), the tricarboxylic acid cycle (Brown et al., 2000), and complexes in the electron transport chain (Skulachev et al., 1967; Kleiner, 1972, 1974; Link and von Jagow, 1995; Mills et al., 2002). It has also been shown that Zn2+ dissipates mitochondrial membrane potential (ΔΨm), decreases oxygen consumption, and enhances reactive oxygen species (ROS) accumulation (Dineley et al., 2005). However, it remains unclear which, if any, of these processes represent the critical target for the neurotoxic actions of zinc.
It has long been appreciated that mitochondria constantly move, divide, and fuse throughout the cell (Bereiter-Hahn and Voth, 1994). It is reasonable to suggest that mitochondrial trafficking in neurons serves to position mitochondria to deliver ATP to regions of high energy demand (van Blerkom, 1991) and to aid in the regulation of local Ca2+ homeostasis (Spira et al., 2001; Yi et al., 2004), although direct evidence that mitochondrial trafficking is driven by energy demand remains sparse. Nevertheless, recent studies suggest that interruption of trafficking is one consequence of excessive activation of NMDA receptors and the subsequent calcium entry (Rintoul et al., 2003). Other potential neurotoxins, such as nitric oxide (Rintoul et al., 2004) and ATP depletion (Rintoul et al., 2003), also inhibit mitochondrial movement, suggesting that impaired delivery of mitochondria could be an important contributor to neuronal injury.
In this report, we investigated the effects of [Zn2+]i on mitochondrial trafficking. Using a mitochondrially targeted enhanced yellow fluorescent protein (mt-eYFP), we observed that movement is substantially diminished with ionophore-induced elevations of [Zn2+]i that are just sufficient to cause injury. We further show that these actions of zinc are mediated by a signaling cascade that involves activation of phosphatidylinositol (PI) 3-kinase, and that transient inhibition of PI 3-kinase prevents both the impairment of trafficking and zinc-induced neuronal injury.
Materials and Methods
Materials. All reagents were purchased from Sigma (St. Louis, MO) unless otherwise specified.
Cell culture. All procedures using animals were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. Primary cultures of rat forebrain neurons were prepared from embryonic Sprague Dawley rat pups (17 d in gestation) and grown in a 37°C incubator containing 5% CO2. Cortices were removed, trypsinized at 37°C for 30 min, and plated on poly-d-lysine-coated 31 mm glass coverslips. Cells were plated in medium containing DMEM with 10% FBS and 1% penicillin/streptomycin. Cultures contain <5% glia in the total cell population. Five hours after initial plating, plating medium was completely removed and replaced with N2-NB media (Neurobasal media with 0.5% penicillin/streptomycin and 1% N2 supplement). Cells were refed with N2-NB media 4 d after plating, whereby approximately one-third of the volume was removed and replaced with an equal volume of fresh N2-NB media. Cells were fed 8 and 11 d after plating with B27-NB media (Neurobasal media with 0.5% penicillin/streptomycin and 2% B27 supplement minus antioxidants) using the same feeding protocol described above. For preparation of the insulin-depleted cells, cultures were fed only with B27-NB media beginning at 5 h after initial plating. All experiments were performed on neurons after 13-15 d in culture.
Transfection using DNA constructs. The mt-eYFP construct (Llopis et al., 1998) consists of eYFP fused to the mitochondrial targeting sequence from subunit IV of human cytochrome c oxidase. The luciferase reporter plasmid consists of the luciferase gene placed in a mammalian expression vector under the control of a cytomegalovirus promoter (Gossen and Bujard, 1992). After 10-12 d in vitro, neurons were transiently transfected using a modified calcium phosphate transfection technique (Xia et al., 1996). This procedure typically generates 1-2% efficiency, with maximal protein expression 24-72 h after transfection.
Fluorescence microscopy. All experiments were performed using a HEPES-buffered salt solution (HEBSS) containing the following (in mm): 137 NaCl, 5 KCl, 10 NaHCO3, 0.6 KH2PO4, 0.6 Na2HPO4, 0.9 MgSO4, 1.4 CaCl2, 20 HEPES, and 5.5 glucose, pH adjusted to 7.4 with NaOH. Neurons were perfused at 5 ml/min during the course of each experiment, and the chamber temperature was maintained at 37°C with heated buffer. The personal-computer-based system for data acquisition is as described previously (Dineley et al., 2002) using SimplePCI imaging software (Compix, Cranberry, PA). Using a 40× water immersion objective, mitochondrial movement in a field containing a single cell was observed by illuminating at 490 nm and acquiring images every 6 s. After a 3-5 min rinse in HEBSS buffer, cells were perfused with the appropriate stimulus (diluted in buffer from stock concentrations).
For [Zn 2+]i measurements, we used the Zn 2+-sensitive fluorophore mag-fura-2 AM (Invitrogen, Eugene, OR). Coverslips were incubated with 5 μm mag-fura-2 for 20 min at 37°C. With constant perfusion at 10 ml/min, fluorescence was measured by illuminating the indicator alternatively at 335 and 375 nm. For ΔΨm experiments, neurons were incubated with 5 μm rhodamine 123 (Invitrogen) for 10 min at 37°C (Vergun et al., 2003). Cells were excited at 490 nm and fluorescence was measured in 10-20 individual neurons for each coverslip. For nucleic acid staining, coverslips were incubated in HEBSS containing 1.5 μm Hoechst-33342 for 15 min at room temperature (Dineley et al., 2000). Cells were illuminated at 340 nm and fluorescent images were obtained randomly from three to six cell-containing regions per coverslip.
Quantification of mitochondrial movement. Mitochondrial movement was analyzed using a macro-based analysis program as described previously (Rintoul et al., 2003). Briefly, a 255 × 255 pixel subfield containing mitochondria was selected, in which pixel images in successive images are subtracted. A movement event for each pixel was registered if the change in fluorescence between successive fields exceeded 20 fluorescence units and a quantitative measurement of movement was obtained by combining pixel events in a field over 2 min. Movement was represented as average event counts/pixel, which was normalized to prestimulus movement values.
Toxicity. For the simplicity of assaying only cells that express mt-eYFP, we transfected neurons with the luciferase construct to measure neuronal injury. For toxicity experiments, coverslips of luciferase-transfected neurons were washed three times with HEBSS. Reagents were diluted in HBSS at the desired concentration and 1 ml of solution was applied to the coverslips for the desired time at room temperature. Stimuli were terminated by washing at least three times with excess HEBSS. The wash buffer was then replaced with the original media and returned to the incubator. Twenty-four hours later, cells were rinsed and lysed in PBS. Lysates were used to measure luciferase enzyme activity using the BriteLite Ultra-High Sensitivity Luminescence Reporter Gene Assay System (PerkinElmer Life Sciences, Boston, MA) according to the manufacturer's instructions. Arbitrary luminescence units were normalized to control, nontreated cells.
Western blotting. Cells were lysed in lysis buffer containing the following: 50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1.5 mm MgCl2, 1% Triton X-100, 5 mm EGTA, 20 μm leupeptin, 1 mm AEBSF [4-(2-aminoethyl) benzenesulfonylfluoride], 1 mm NaVO3, 10 mm NaF, and 1 tablet of protease inhibitor. Protein concentration was determined by a microprotein assay using the BCA protein assay kit according to the manufacturer's instructions (Pierce, Rockford, IL). Approximately 30 μg of lysates were mixed with equal volumes of 2× SDS sample loading buffer (60 mm Tris-HCl, pH 7.5, 2 mm EDTA, 10 mm 2-mercaptoethanol, 20% glycerol, and 2% SDS) and size-fractionated by electrophoresis on 4-15% Tris-HCl Ready Gels (Bio-Rad, Hercules, CA) at 30 mA for 1 h followed by electrotransfer onto a nitrocellulose membrane at 80 V for 2 h. The membrane was preblotted with 5% dry milk in PBS-Tween (1× PBS, 0.1% Tween 20) at room temperature for 1 h. The blots were probed with rabbit antiserum raised against Akt, and phosphosphorylated-S473 Akt (Cell Signaling Technology, Beverly, MA) antibodies. Goat anti-rabbit IgG coupled to horseradish peroxidase (1:1000; Pierce) was used as the secondary antibody. Bands were visualized by the SuperSignal West Dura Extended Duration Substrate (Pierce).
Statistical analyses. Statistical analysis was performed using Prism 4.01 (GraphPad Software, San Diego, CA). All data are presented as mean ± SE consisting of four to six experiments for each condition from at least three different cell culture preparations. Comparisons were made using Student's t test and one-way ANOVA with Bonferroni's multiple comparison and Dunnett's post tests to compare all conditions to control, with p values <0.05 regarded as significant.
Concentrations of Zn2+ that inhibit mitochondrial movement parallel Zn2+ concentrations that are neurotoxic
We fluorescently labeled neuronal mitochondria using an mt-eYFP. The ionophore sodium pyrithione (Pyr) was used to increase [Zn2+]i (Dineley et al., 2000). As shown in Figure 1a, buffer, sodium pyrithione alone (20 μm), or 0.3 μm Zn2+/Pyr did not affect mitochondrial movement. However, higher concentrations of Zn2+ significantly inhibited mitochondrial movement, with concentrations as low as 1 μm resulting in a significant decrease in movement (supplemental video 1, available at www.jneurosci.org as supplemental material). Notably, this stimulus consistently stopped movement of all mitochondria in a given field and did not preferentially alter the movement of retrograde-versus anterograde-moving mitochondria. This Zn2+-mediated effect was irreversible up to 20 min after the stimulus was washed out. Interestingly, this concentration-dependent movement inhibition corresponded to the concentration dependence of Zn2+ toxicity (Fig. 1a, inset).
To establish that this was not a Ca2+-induced phenomenon, we perfused cells with Zn2+/Pyr under Ca2+-free conditions to confirm that the stimulus was still effectively inhibiting movement (Fig. 1b). There is evidence that, under depolarizing conditions, cellular Zn2+ entry occurs through plasma membrane Ca2+-permeable pathways (Sensi et al., 1999). As a more physiological approach to elevate [Zn2+]i, we observed that mitochondrial movement was significantly inhibited with elevated Zn2+ (300 μm) in the presence of high K+ buffer (Fig. 1b). Evidence that exposure to strong oxidizing agents can generate [Zn2+]i fluxes in neurons is indicative of Zn2+ release from cytosolic reservoirs, including metallothionein (Aizenman et al., 2000). Application of the sulfhydryl oxidizing agent 2,2′-dithiodipyridine (DTDP) (10 μm) significantly inhibited mitochondrial movement, suggesting that [Zn2+]i release in addition to Zn2+ entry is sufficient to induce the effect (Fig. 1b). Our previous study showed that Ca2+-induced cessation of movement was also associated with a substantial change in mitochondrial morphology (Rintoul et al., 2003). However, the representative fluorescent images taken before (Fig. 1c) and after (Fig. 1d) Zn2+ exposure demonstrated that there are no gross morphological changes that occur with Zn2+ exposure.
Zn2+ inhibition of mitochondrial movement is partially sensitive to chelation
To determine whether chelation can restore Zn2+-mediated movement inhibition, we applied the membrane-permeant heavy metal chelator N,N,N′,N′-tetrakis-(2-pyridylmethyl)-ethylenediamine (TPEN) (50 μm) immediately after Zn2+/Pyr exposure (Fig. 2a). Notably, TPEN partially restored movement when applied after 1 μm Zn2+ but was ineffective at restoring movement when applied 10 min after 3 μm Zn2+. Additionally, Figure 2b shows that TPEN restored movement when applied after shorter exposures to 3 μm Zn2+ but became ineffective after treatments longer than 6 min. We also observed that TPEN rescued neurons from injury when immediately applied after a 5 min Zn2+/Pyr exposure (Fig. 2c) but did not protect against toxicity after a 10 min (Fig. 2d) Zn2+ treatment. Together, these results indicate that the TPEN effects are concentration and time dependent. The insensitivity to TPEN after longer or larger Zn2+ exposures suggested the initiation of a signaling cascade, which is irreversible after simple Zn2+ chelation and therefore becomes independent of Zn2+.
As described above, oxidant exposure diminished mitochondrial movement when 2,2′-DTDP (30 μm) was applied at concentrations that inhibit mitochondrial movement to the same extent as 3 μm Zn2+/Pyr. If TPEN was applied after 2,2′-DTDP, movement was not restored (supplemental Fig. 1a, available at www.jneurosci.org as supplemental material). However, if 2,2′-DTDP and TPEN were applied simultaneously, movement was inhibited to a lesser degree (supplemental Fig. 1c, available at www.jneurosci.org as supplemental material). Likewise, a similar trend was observed when 100 μm H2O2 was applied to mimic ROS exposure (supplemental Fig. 1b,d, available at www.jneurosci.org as supplemental material). Interestingly, the reducing agent dithiothreitol also did not reverse movement inhibition (supplemental Fig. 1e, available at www.jneurosci.org as supplemental material), suggesting that although this phenomenon is potentially ROS mediated, movement is not restored when shifted to a more reduced environment. These findings are consistent with ROS-mediated Zn2+ release contributing to the inhibition of trafficking but also suggest that there may be an additional ROS-induced Zn2+-independent effect on mitochondrial movement.
Inhibitors of PI 3-kinase prevent Zn2+-induced movement inhibition
To identify the Zn2+-activated signaling cascade involved in the movement effect, we investigated several agents that selectively inhibit protein kinases (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). Of all of the compounds that we tested in our movement paradigm, only the PI 3-kinase inhibitors, wortmannin (1 μm) and 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002) (30 μm), were effective in that (1) they did not affect mitochondrial movement by themselves, and (2) they blocked the Zn2+-mediated effect at the concentration (3 μm Zn2+/Pyr) and time (10 min) that was TPEN insensitive (Fig. 3a). One concern was that wortmannin was preventing ionophore-induced [Zn2+]i accumulation by blocking its cellular entry or chelating the [Zn2+]i before it could mediate any effects on movement. We excluded these possibilities by monitoring [Zn2+]i in mag-fura-2-loaded neurons that were perfused with 3 μm Zn2+/Pyr, either with or without wortmannin (1 μm) pretreatment (Fig. 3b). Both conditions produced equal mag-fura-2 responses that were TPEN-reversible, confirming that wortmannin was not preventing entry or binding [Zn2+]i.
Preventing Akt phosphorylation does not prevent Zn2+-induced movement inhibition
As a possible downstream effector of PI 3-kinase, we investigated the role of Akt phosphorylation in Zn2+-induced movement inhibition (Fig. 4). We monitored phosphorylated Akt levels in neurons treated with 3 μm Zn2+/Pyr for 10 min, either in the presence or absence of wortmannin (1 μm) or a putative Akt inhibitor [1L-6-hydroxymethyl-chiro-inositol-2-(R)-2-O-methyl-3-O-octadecykcarbonate] (10 μm; Calbiochem, La Jolla, CA) (Fig. 4a). This phosphatidylinositol ether analog has been reported to selectively inhibit Akt phosphorylation and prevent growth of HT-29, MCF, HeLa, and PC-3 cancer cell lines. This novel class of Akt inhibitors selectively blocks Akt activation and downstream substrates without affecting upstream kinases or other kinase pathways. We used insulin (100 nm) as a positive control in these experiments. As demonstrated in Figure 4a, both insulin and Zn2+ rapidly induced Akt phosphorylation within 10 min in a wortmannin-sensitive manner. Interestingly, treatment with Akt inhibitor alone increased phosphorylation from control levels, and this compound did not significantly reduce phosphorylation in neurons compared with either insulin or Zn2+ alone.
We also monitored the effects of these treatments on mitochondrial movement. As shown in Figure 4b, perfusion with 100 nm insulin for 10 min did not significantly decrease mitochondrial movement compared with untreated cells, although a 10 min stimulus of 100 nm insulin is sufficient for Akt phosphorylation in primary neurons (Schubert et al., 2004). Treatment with the Akt inhibitor itself slightly decreased mitochondrial movement and also did not prevent the Zn2+-induced movement effect. Although previous studies have used this inhibitor to prevent Akt phosphorylation in cell lines (Kozikowski et al., 2003), it appears to be less effective in neurons.
As an alternative to the Akt inhibitor, we transiently cotransfected a hemagglutinin-tagged dominant-negative (DN)-Akt (K179M) plasmid along with the mt-eYFP plasmid. Overexpression of DN-Akt did not prevent the Zn2+-mediated movement inhibition (supplemental Fig. 3, available at www.jneurosci.org as supplemental material), arguing against Akt activation by PI 3-kinase as a downstream event. We additionally tested other targets of PI 3-kinase that are activated in Akt-independent pathways (supplemental Fig. 3, available at www.jneurosci.org as supplemental material). Blocking ADP-ribosylation factor 6 (ARF6) phosphorylation using a dominant-negative construct (Langille et al., 1999) or inhibiting target of rapamycin (TOR) phosphorylation using rapamycin (Lynch et al., 2001) did not prevent the Zn2+-mediated inhibition, indicating that these protein kinases are not directly involved in the movement paradigm. Thus, although PI 3-kinase activation appears to be an essential component of the effect of zinc, it alone is apparently not sufficient to account for the actions of zinc on mitochondrial movement.
Zn2+ effects on mitochondrial movement do not result from compromised mitochondrial function
We have shown previously that acute application of the mitochondrial uncoupler carbonyl cyanide p-trifluoromethoxy-phenylhydrazone (FCCP), which rapidly dissipates ΔΨm, decreases mitochondrial movement in primary neurons (Rintoul et al., 2003). Zinc, at some concentrations, can also impair mitochondrial function (Sensi et al., 1999; Dineley et al., 2000). However, an examination of the effects of these zinc treatments on mitochondrial membrane potential show that zinc only produced a significant depolarization at relatively high concentrations (Fig. 5a,b). In addition, pretreating neurons with wortmannin (1 μm) did not prevent the inhibition of mitochondrial movement produced by FCCP (750 nm) exposure (Fig. 5c). Together, these results argue against mitochondrial dysfunction as the key mechanism in Zn2+-mediated movement inhibition.
Wortmannin protects against Zn2+-induced neurotoxicity
As shown in Figure 3a, preperfusion with wortmannin prevented Zn2+-mediated inhibition of mitochondrial movement. However, the effectiveness of wortmannin was decreased when Zn2+ was increased from 3 to 10 μm (Fig. 6a). Interestingly, wortmannin protected against Zn2+-induced neurotoxicity with concentrations at or less than 3 μm Zn2+/Pyr, but not with 10 μm (Fig. 6b,c), suggesting a correlation between PI 3-kinase activation in Zn2+-mediated movement inhibition and kinase activity during neurotoxicity.
Sustained as opposed to transient activation of PI 3-kinase does not protect against Zn2+-induced toxicity
Although PI 3-kinase is generally considered to be a prosurvival signaling mechanism, a recent study demonstrated that transient PI 3-kinase inhibition protects primary neurons from oxidative stress (Levinthal and DeFranco, 2004). As described above, our data suggest that wortmannin protected against Zn2+-induced neurotoxicity at concentrations that inhibited mitochondrial movement. Although wortmannin is a fast-acting inhibitor (half-life of 90 min), it is also relatively unstable in an aqueous solution. The other selective PI 3-kinase inhibitor, LY294002, has a longer half-life (3.5 h) and produces a more sustained kinase inhibition (Jones et al., 1999). Because we were unable to prevent Zn2+-induced neurotoxicity with LY294002 (Fig. 7a), we speculated that the neuroprotection that we observed with wortmannin could be explained by a transient inhibition of the kinase. To explore this possibility, neurons were repeatedly exposed to wortmannin during the course of Zn2+ toxicity. Sustained kinase inhibition with wortmannin no longer prevented neurons against elevated [Zn2+]i (Fig. 7b), in agreement with the hypothesis that only transient inhibition of PI 3-kinase affords protection against Zn2+ toxicity.
The main finding from this study is that elevated [Zn2+]i modulates mitochondrial movement through a signaling mechanism that is most commonly portrayed as a prosurvival pathway. In our experiments, acute induction of pathophysiological [Zn2+]i consistently stopped mitochondrial movement in cortical neurons. This effect was reversible only with early Zn2+ chelation, indicating that a Zn2+-induced signaling cascade had been initiated, which subsequently mediated movement inhibition. Interestingly, both Zn2+-induced movement changes and neurotoxicity were prevented with inhibitors of the PI 3-kinase pathway. Thus, we demonstrated a signaling mechanism that regulates both mitochondrial trafficking and neuronal viability in parallel.
Recent studies have identified several modulators of mitochondrial trafficking in central neurons. In particular, elevated intracellular Ca2+ is an effective inhibitor of movement, an effect that is likely to be mediated by inhibition of ATP synthesis as well as potential disruption of the cytoskeleton (Rintoul et al., 2003). Nitric oxide also inhibits movement (Rintoul et al., 2004), probably as a result of inhibition of the electron transport chain. However, the mechanism by which Zn2+ inhibits movement is distinct in that it appears to occur without clear disruption of mitochondrial function. Our findings of a role for PI 3-kinase are consistent with a recent report by Chada and Hollenbeck (2003), who reported that focal stimulation of NGF receptors in sympathetic neurons causes mitochondria to “dock” or stop in the region of NGF application via a mechanism sensitive to inhibition by wortmannin, and likely by undergoing docking interactions with the actin cytoskeleton (Chada and Hollenbeck, 2004). Interestingly, this phenomenon did not occur with global application of NGF, which would more closely parallel the conditions used here. Nevertheless, the basic concept of PI 3-kinase-mediated control of mitochondrial trafficking is clearly similar between the two model systems. From our experiments, it appears that PI 3-kinase is involved in the actions of zinc described here, but the downstream target of the signaling cascade is less obvious. Although the conditions used here result in Akt phosphorylation, agents like insulin that mimic this action of zinc do not recapitulate the effects on trafficking. Also, neither Akt inhibitors nor dominant-negative Akt prevents the effects of zinc. Likewise, neither TOR nor ARF6 inhibition interrupted the actions of zinc. Thus, the step(s) between PI 3-kinase activation and the mitochondrial target remains unidentified. Moreover, recent studies have highlighted the role of other signaling cascades in the regulation of mitochondrial trafficking, including interactions with phosphatidylinositol 4,5-bisphosphate species (De Vos et al., 2003), activation of protein kinase A (Okada et al., 1995), and tumor necrosis factor-induced p38 activation (De Vos et al., 2000). In addition to mitochondria, others have implicated the regulation of synaptic vesicular transport by kinesin motor proteins (Okada et al., 1995; Zhao et al., 2001). With regard to the present study, we also demonstrated that mitochondrial trafficking is regulated by a signaling mechanism critical for cell survival.
A number of different mechanisms for zinc-mediated toxicity have been suggested. The data shown here illustrate the point that there are multiple effects of Zn2+, which occur over a range of concentrations. We observed that low concentrations of Zn2+ that inhibit movement and produce injury have no effect on ΔΨm. Higher concentrations of zinc produce a form of injury that is not prevented by wortmannin and is also associated with mitochondrial depolarization. Zinc-mediated mitochondrial depolarization has been reported in a number of studies, both in intact cells (Sensi et al., 1999) as well as in isolated mitochondria (Dineley et al., 2005). However, these results suggest that, although mitochondria are clearly a target for zinc action, the most sensitive mechanisms associated with cytotoxicity do not involve gross disruption of mitochondrial function. Evidently, this phenomenon is not strictly dependent on the source of zinc. To mimic oxidant-labile [Zn2+]i, we applied 2,2′-DTDP to neurons and observed that mitochondrial movement significantly decreased with concentrations as low as 10 μm. This was surprisingly low, given that 100 μm 2,2′-DTDP was applied to cells to see mag-fura-2 responses in neurons (Aizenman et al., 2000). In addition, we also observed movement inhibition with 100 μm H2O2, which was used to mimic ROS exposure. Interestingly, TPEN did not restore mitochondrial movement when applied after either 2,2′-DTDP or H2O2. Together, these results raise the possibility that, apart from a direct Zn2+ effect, a secondary ROS-induced mechanism may also modulate mitochondrial movement.
Indirect PI 3-kinase activation by Zn2+ has been reported in other studies. Eom et al. (2001) found that Zn2+ induced stimulation of the JNK (c-Jun N-terminal protein kinase) through the PI 3-kinase pathway in primary mouse cortical neurons in culture, and Kim et al. (2000) demonstrated that extracellular Zn2+ activates p70 S6 kinase through PI 3-kinase signaling in Swiss 3T3 cells. These findings differ from our own in that PI 3-kinase activation is dependent on the activation of other pathways, which may be attributable to the exposure of 10 times higher concentrations of Zn2+ (except in the absence of ionophore), longer Zn2+ exposures necessary to achieve Akt phosphorylation, or varying sensitivities among cell types. We did not observe that the activation of other protein kinases plays a direct role in the Zn2+-mediated movement effect, because inhibitors of these pathways did not protect against movement inhibition (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). We cannot entirely exclude the possibility that these kinases are activated either downstream of or in parallel with PI 3-kinase; however, their activation did not affect Zn2+-mediated changes in mitochondrial trafficking or neuronal injury. Zn2+ activates a number of other protein kinases involved both in pro-cell and anti-cell survival mechanisms. For example, studies have demonstrated that [Zn2+]i can activate members of the MAP (mitogen-activated protein) kinase family. Zn2+ activation of the stress-activated p38 (McLaughlin et al., 2001), as well as its activation of the traditionally prosurvival ERK (extracellular signal-regulated kinase) (Du et al., 2002), is hypothesized to contribute to neuronal injury and subsequent cell death. In addition, activation of PKC (protein kinase C) in neurons plays a role in Zn2+-induced oxidative neuronal injury (Noh, 1999). Because signal transduction pathways are rarely activated in a linear manner, it is more than likely that there is cross talk between these pathways with respect to Zn2+-mediated neuronal injury.
PI 3-kinase activity has traditionally been associated with many key cellular processes, including cell growth and survival, membrane trafficking, neurite outgrowth, and cytoskeletal reorganization (Foster et al., 2003). Indeed, extensive literature demonstrates that growth factor-mediated activation of the PI 3-kinase/Akt pathway by insulin (Patel et al., 1993) or IGF-1 (insulin-like growth factor 1) (Dudek et al., 1997; Schubert et al., 2004), NGF (Chada and Hollenbeck, 2003), and BDNF (Zheng and Quirion, 2004), is neuroprotective against a range of cellular stresses. In addition, Luo et al. (2003) have demonstrated that constitutive activation of Akt is neuroprotective, whereas Akt deactivation promotes multiple models of neuronal cell death, including NMDA excitotoxicity, or NO- and H2O2-elicited injury. Including the present study, however, only one other report by Levinthal and DeFranco (2004) recently provided an example of PI 3-kinase activity associated with cell death. The authors demonstrated that glutamate-induced oxidative neuronal injury is mediated by activation of PI 3-kinase. Moreover, a window of transient PI 3-kinase inhibition afforded protection against oxidative glutamate toxicity, which is consistent with our results involving Zn2+ toxicity. We also observed that transient PI 3-kinase inhibition was neuroprotective, but this protection was overcome with sustained kinase inhibition. This provides insight about the bipolar nature of PI 3-kinase: short-term activation is detrimental to cell health, whereas long-term activation is essential for cell survival.
This work was supported by National Institutes of Health Grant NS34138 (I.J.R.), American Heart Association (AHA) Postdoctoral Fellowship 0325734U (G.L.R.), and AHA Predoctoral Fellowship 0315338U (L.M.M.). We thank Dr. Roger Y. Tsien for providing the mt-eYFP construct and Dr. Elias Aizenman for providing the luciferase reporter construct. We thank David J. Levinthal, Drs. Donald B. DeFranco, Jes K. Klarlund, Yu Jiang, and Daniel Altschuler for providing critical reagents and helpful suggestions, and our colleagues in the lab for assistance in preparing this manuscript. We also thank Anthony J. Filiano for preparation of cell cultures, Xiaoping Hu, Yue Luo, and Ganwei Lu for assistance with Western blots, and Groundskeeper Willie for technical assistance in fluorescence microscopy.
Correspondence should be addressed to Dr. Ian J. Reynolds, Merck Research Laboratories, 770 Sumneytown Pike, Mail Stop WP42-229, West Point, PA 19486. E-mail:.
Copyright © 2005 Society for Neuroscience 0270-6474/05/259507-08$15.00/0