Mitochondrial Fragmentation Is Involved in Methamphetamine-Induced Cell Death in Rat Hippocampal Neural Progenitor Cells

Methamphetamine (METH) induces neurodegeneration through damage and apoptosis of dopaminergic nerve terminals and striatal cells, presumably via cross-talk between the endoplasmic reticulum and mitochondria-dependent death cascades. However, the effects of METH on neural progenitor cells (NPC), an important reservoir for replacing neurons and glia during development and injury, remain elusive. Using a rat hippocampal NPC (rhNPC) culture, we characterized the METH-induced mitochondrial fragmentation, apoptosis, and its related signaling mechanism through immunocytochemistry, flow cytometry, and Western blotting. We observed that METH induced rhNPC mitochondrial fragmentation, apoptosis, and inhibited cell proliferation. The mitochondrial fission protein dynamin-related protein 1 (Drp1) and reactive oxygen species (ROS), but not calcium (Ca2+) influx, were involved in the regulation of METH-induced mitochondrial fragmentation. Furthermore, our results indicated that dysregulation of ROS contributed to the oligomerization and translocation of Drp1, resulting in mitochondrial fragmentation in rhNPC. Taken together, our data demonstrate that METH-mediated ROS generation results in the dysregulation of Drp1, which leads to mitochondrial fragmentation and subsequent apoptosis in rhNPC. This provides a potential mechanism for METH-related neurodegenerative disorders, and also provides insight into therapeutic strategies for the neurodegenerative effects of METH.


Introduction
Methamphetamine (METH) is abused by over 35 million people worldwide and is an illicit and potent psychostimulant with strong action on the CNS [1,2]. A serious neuropsychological consequence of METH abuse is cognitive impairment, with working memory deficits remaining long after withdrawal [3]. It is well documented that METH increases glutamate (Glu) levels in the mammalian brain. The high levels of Glu can activate ionotropic receptors, such as N-methyl-D-aspartate (NMDA) and AMPA receptors, resulting in increased intracellular Ca 2+ levels and formation of reactive nitrogen species (RNS) [4][5][6]. All of these factors contribute to METH-mediated neurotoxicity. In addition, in rodents it has been suggested that METH induces apoptosis of striatal glutamic acid decarboxylase-containing neurons due to the interactions of ER stress and mitochondrial death pathways [7][8][9][10][11]. Astroglial activation was also found in METH-induced toxicity [12,13]. During development and following brain injury, NPC are the source of new neurons and astrocytes in the brain. However, the effects of METH on NPC are not well understood.
The mammalian hippocampus retains its ability to generate neurons throughout life [14][15][16]. Granule neurons are generated from a population of continuously dividing progenitor cells residing in the subgranular zone of the dentate gyrus in the rodent brain [17]. Newborn neurons generated from these progenitor cells migrate into the granule cell layer, differentiate, extend axons and express neuronal marker proteins [18][19][20][21]. It is also known that the hippocampus is particularly vulnerable to METH. A single challenge of METH suppresses granule cell proliferation in adult gerbils and initiates rewiring of neuronal networks in the prefrontal cortex (PFC), which occurs concurrently with development of severe deficits in PFC-related behaviors. Developmental dysfunction during hippocampal formation is proposed to play a major role in the pathogenesis of neurodegenerative disorders [22]. Defects such as a reduction in hippocampal volume, shape deformations, abnormalities in the granule cell layer, modifications in the mossy fiber pathway, changes in hippocampal cell density and orientation and changes in several cellular markers have been reported [23].
Mitochondria serve as the powerhouse in most eukaryotic cells and play crucial roles in energy metabolism, thermogenesis, maintenance of Ca 2+ homeostasis and apoptosis [24]. Mitochondria are also dynamic organelles which undergo continuous fission and fusion to form a reticulum structure. Increasing evidence has demonstrated that the changes in mitochondrial morphology depend on the physiological requirements of a cell and are an important determinant of mitochondrial function [25,26]. In mammalian cells, mitochondrial fission and fusion rely on multiple proteins, such as the dynamin superfamily, which mediate the remodeling of the outer and inner mitochondrial membranes [27]. Among of them, Drp1 [28,29], Fission 1 (Fis1) [30,31] and Endophilin B1 (Bif-1/SH3GLB1) [32] control mitochondrial fission. Defects in either mitochondrial fusion or mitochondrial fission may cause severe neurodegenerative diseases [33][34][35].
Recent evidence indicates that mitochondrial dynamics play important roles in the apoptotic process. Many apoptotic stimuli can elicit mitochondrial morphologic changes during the early apoptotic stage, resulting in small, round and more numerous organelles [36][37][38]. Inhibition of mitochondrial fragmentation can not only preserve the mitochondrial architecture but also prevent the release of cytochrome c and subsequent apoptotic steps [39,40]. How METH regulates the mitochondrial dynamics in rhNPC is not clear.
In this study, we utilized rhNPC as a model and observed that METH induces cell injury of rhNPC, which is accompanied by changes in mitochondrial morphology. Our results further demonstrate that Drp1 not only regulates mitochondrial fragmentation in rhNPC in the same way as described for apoptosis-related fission, but also its oligomerization and translocation from the cytosol to the mitochondria induced by METH contribute to mitochondrial fragmentation. These findings suggest that mitochondrial fragmentation plays an important role in METHinduced rhNPC damage, and they also provide insight into the pathological mechanism of neurodegenerative disorders related to METH abuse.

METH inhibits the proliferation of rhNPC
We examined the effects of METH on rhNPC derived from the hippocampus of embryonic day 17 (E17) fetus. These rhNPC form neurospheres (Fig. 1A), which are more than 90% nestin positive (a marker for progenitor cells) (Fig. 1B). After changing to specific differentiation medium for astrocytes or neurons and culturing the cells for 7 days, rhNPC appeared to be completely differentiated into astrocytes (Fig. 1C, GFAP as a marker) or neurons (Fig. 1D, b-tubulin as a marker) which was confirmed by immunoblotting ( Fig. 1E, 1F).
Next, we investigated whether METH affected the neurogenesis of rhNPC. We treated rhNPC with different concentrations of METH for 24 hours, and then performed cell-cycle analysis. The number of proliferating cells decreased in a dose-dependent manner after METH treatment (,60% decrease, Fig. 2A; p,0.05). In addition, pre-treatment of rhNPC with N-acetylcysteine (NAC), a non-specific reductant, reversed this effect (Fig. 2B). These results suggest that METH inhibits the proliferation of rhNPC in an oxidative stress-dependent manner.

The inhibition of rhNPC by METH is associated with cell apoptosis and accompanied by mitochondrial fragmentation
To investigate the mechanism by which METH inhibits rhNPC proliferation, we examined rhNPC apoptosis induced by different concentrations of METH by nuclear staining. The number of condensed nuclei demonstrating apoptosis was significantly increased with the increasing METH concentrations (Fig. 3A, b2 and c2, indicated by arrow; Fig. 3B). Increasing evidence indicates that mitochondrial dynamics play important roles in the apoptotic process. Many apoptotic stimuli elicit mitochondrial morphologic changes during early apoptosis, resulting in small, round and more numerous organelles [36][37][38]. To examine if the mitochondrial morphological changes are also found in METH-induced rhNPC apoptosis, we stained mitochondria with MitoTrackerHRed in addition to DAPI staining. The MitoTracker staining suggests that the mitochondrial network is severely damaged after METH treatment (Fig. 3A, b1 and c1) and the number of cells with fragmented mitochondria increases with increasing concentrations of METH (Fig. 3C).
To investigate whether the METH-induced mitochondrial fragmentation is related to the duration of METH exposure, we treated rhNPC with 300 mM METH for different time periods and then stained mitochondria with MitoTrackerHRed and rhNPC with nestin monoclonal antibody as indicated in Fig. 4. The results demonstrate that prolonged treatment with METH increases the extent of mitochondrial fragmentation (Fig. 4). Taken together, these results suggest that mitochondrial fragmentation is involved in METH-induced NPC apoptosis in a dose-and timedependent manner.

Drp1 is involved in the regulation of METH-mediated mitochondrial fragmentation in rhNPC
Mitochondrial fission and fusion rely on the function of multiple proteins that mediate the remodeling of the outer and inner membranes [27]. In mammalian cells, at least two proteins, Drp1 and Fis1, are required for mitochondrial fission [41,42]. Recent evidence indicates that Drp1 is an essential mediator of apoptosisrelated mitochondrial fission, and during apoptosis translocates from the cytoplasm to mitochondria, where it forms clusters at prospective sites of mitochondrial fission [39,43,44]. Elimination of Drp1 activity during apoptosis results in a block of mitochondrial fission, which leads to long interconnected organelles and a delay in cell death. To evaluate the contribution of Drp1 in regulating mitochondrial morphology during METH treatment, we examined the expression and localization of Drp1 in rhNPC. The results show that under normal conditions, Drp1 has a punctate distribution in the cytoplasm around mitochondrial networks (Fig. 5A, upper panel); however, after treatment with 300 mM METH for 24 hours, the distribution of Drp1 in the cytoplasm was mainly localized at both ends of the mitochondrial fragments ( Fig. 5A lower panel: indicated by arrows in b4). These data suggest that Drp1 is involved in METH-induced mitochondrial fragmentation. To further confirm this possibility, we downregulated Drp1 expression in rhNPC by siRNA duplexes targeted against rat Drp1 mRNA, and then examined the expression of Drp1 by western blotting (Fig. 5C) and mitochondrial morphology after METH treatments (Fig. 5B). The results demonstrate that mitochondrial structure in the Drp1 siRNA-transfected cells still remains connected with mitochondrial networks despite METH treatment ( Fig. 5B-c1 and d1); however, mitochondria undergo fragmentation in siRNA control-transfected cells after METH treatment ( Fig. 5B-b1). These results further support the notion that Drp1 plays an important role in METH-mediated mitochondrial fragmentation.

Oxidative stress, but not Ca 2+ influx, is involved in METHmediated mitochondrial fragmentation in rhNPC
It is well documented that METH increases Glu levels in the mammalian brain, which can activate ionotropic receptors, such as NMDA and AMPA receptors, resulting in increased intracellular Ca 2+ levels and ROS formation [4][5][6]. All of those factors contribute to METH-mediated neurotoxicity. To investigate the effects of METH on intracellular Ca 2+ levels and ROS, we first examined Ca 2+ influx in rhNPC by loading them with Fura-2 and monitoring Ca 2+ influx by live-fluorescent imaging. Using Ca 2+ imaging on groups of 20 rhNPC, we found that METH had no effects on Ca 2+ influx, even after treatment with 5 mM METH (Fig. 6A). Moreover, we observed that 100 mM Glu stimulation caused Ca 2+ influx, and the observed Ca 2+ influx was greatly increased using the AMPA receptor-specific potentiator, cyclothiazide, but it was almost completely blocked by AMPA/kainate receptor antagonist, CNQX [45]. However, METH had no effects on Glu-mediated Ca 2+ influx through AMPA receptors (Fig. 6B).
Next, we examined the ROS production in rhNPC that were pre-treated with 10 mM NAC, and then treated with 300 mM METH or treated with 300 mM METH alone for 24 hours. Cells were subjected to staining with MitoTrackerHRed CM-H2XROS which fluoresces upon oxidation and is indicative of ROS generation. We observed that cells with or without NAC treatment had no significant ROS generation (Fig. 7, a2, b2), whereas, ROS were induced by METH treatment (Fig. 7, c2), and this was blocked by pre-treatment with NAC ( Fig. 7 d2). H 2 O 2 treatment mimicked the METH action in an NAC-sensitive manner (Fig. 7, e2, f2). The results suggest that METH can induce ROS generation, but has no effects on Ca 2+ influx in rhNPC.
METH-mediated mitochondrial fragmentation is associated with the oligomerization and translocation of Drp1 from the cytoplasm to the mitochondria Drp1 has been proposed to encircle mitochondria and function as a drawstring to mediate constriction and scission resulting in the fission of mitochondria [46]. Purified Drp1 has the ability to selfassemble into ring-like multimers on the mitochondrial outer membrane in vitro [28,29,47]. To explore whether METHmediated mitochondrial fragmentation is through Drp1 oligomerization induced by oxidative stress, we treated rhNPC with different concentrations of METH for 24 hours and then analyzed Figure 1. Isolation and differentiation of rhNPC. rhNPC were dissociated from rat hippocampi and cultured with NPBM neural progenitor basal medium containing EGF, bFGF and NSF-1. The neurospheres were formed after culture for 7 days (Fig. 1A). Neurospheres were digested and plated on poly-D-lysine-coated cover slips and cultured with astrocyte differentiation medium or neuron differentiation medium for 7 days and then subjected to immunostaining with nestin ( Fig. 1B), GFAP (Fig. 1C) and b-tubulin III (Fig. 1D) antibodies. rhNPC were cultured in six-well plates with differentiation medium for 3, 5 and 7 days, respectively, then subjected to western blotting with GFAP and b-tubulin III antibodies (Fig. 1E, 1F). doi:10.1371/journal.pone.0005546.g001 the cytosolic Drp1 by non-reducing SDS-PAGE. Our results demonstrated that Drp1 oligomerization was induced by METH treatment in a dose-dependent manner ( Fig. 8A; Fig. 8B). Importantly, the Drp1 oligomerization was reversed by NAC pre-incubation (Fig. 8C, left panel) similar to NAC effects on H 2 O 2 treatment (Fig. 8C, right panel). This is consistent with the METH-induced ROS generation in Fig. 7 and strengthens the association between Drp1 oligomerization and mitochondrial damage.
To further investigate whether METH-mediated mitochondrial fragmentation is caused by increased recruitment of oligomeric Drp1 from the cytoplasm to the mitochondrial outer membrane, we performed in vitro subcellular fractionations on untreated or METH-treated rhNPC. Our results showed that METH induced an increase of oligomeric Drp1 in the mitochondrial fraction, whereas NAC prevented the association of METH-induced oligomeric Drp1 with mitochondria (Fig. 8D). These data suggest that METH-mediated mitochondrial fragmentation is not only associated with the formation of Drp1 oligomers but is also determined by the translocation of Drp1 oligomers from the cytosol to the mitochondrial outer membrane. Furthermore, our data demonstrate that METH-induced ROS may be an important mediator for mitochondrial fragmentation.

Discussion
In this study we observed that METH induced a decrease in proliferation, an increase in apoptosis and a disruption of mitochondrial networks in rhNPC, resulting in mitochondrial fragmentation. Further investigation also suggested that Drp1 was involved in regulating mitochondrial morphology, and that the oligomerization and subsequent translocation of Drp1 contributed to the regulation of METH-mediated mitochondrial fragmentation in which ROS acted as an important mediator.
It has been suggested that the subgranular zone (SGZ) of the dentate gyrus in the hippocampus of the rodent brain is an important source of dividing progenitor cells, and newborn neurons generated from these progenitor cells migrate into the granule cell layer, differentiate, extend axons and express neuronal marker proteins [18][19][20][21]. The hippocampus is an area of particular interest as it is central to many aspects of the addictive process, including drug use relapse. Moreover, the hippocampus is particularly vulnerable to METH [48,49]. METH induces apoptosis of striatal glutamic acid decarboxylase-containing neurons due to the interactions between ER stress and mitochondrial death pathways [7][8][9][10][11]. In addition to neuronal death, astroglial activation was also found in METH-induced toxicity [12,13]. Decreased neurogenesis in the SGZ of the hippocampus is another product of opiate and psychostimulant drugs [50]. Neurogenesis in the dentate gyrus is also decreased after repeated administration of psychostimulants such as phencyclidine (PCP), dizocilpine (MK-801), and METH [51]. This inhibition of neurogenesis affects the replacement and repair of damaged cells and may finally contribute to neurodegenerative disease phenotypes. Thus, is necessary to have a better understanding of the contribution of METH abuse to neurodegenerative disorders via its effects on the regeneration of the brain. We utilized rhNPC as a model to examine the effects of METH on cell proliferation. Our results demonstrate that METH inhibits rhNPC proliferation (Fig. 2), which may contribute to decreased neurogenesis. It is well known that cell death of NPCs, alteration of the proliferative environment of the subventricular zone (SVG) and direct action of drugs of abuse on progenitor cells contribute to decreased neurogenesis [50,52]. Similarly, when treating rhNPC with METH, we found that the frequency of apoptosis was increased (Fig. 3). These results suggest that METH-induced inhibition of rhNPC proliferation is associated with METHinduced apoptosis. Furthermore, the inhibition of proliferation precedes the METH-induced apoptosis. Z-VAD, a pan-caspase inhibitor, has no significant effects on METH-induced inhibition of proliferation but prevents the METH-induced apoptosis (data not shown). These data suggest that the apoptotic cascade may be triggered after rhNPC proliferation is inhibited. Interestingly, we also observed that METH-induced apoptosis was accompanied by the changes in mitochondrial morphology.
Mitochondria are dynamic organelles, frequently dividing and fusing with one another, and the mitochondrial fission and fusion processes participate in apoptosis [46,[53][54][55]. Mitochondrial fission is normally recognized as an early event during apoptosis. Moreover, excessive mitochondrial fragmentation appears to be a prerequisite step in intrinsic apoptosis pathways, as several components of the mitochondrial fission machinery, including Drp1, Fis1, and Endophilin B1, have been implicated in programmed cell death progression [56]. Our results demonstrate that METH induces mitochondrial fragmentation during the early stages of apoptosis ( Fig. 3; Fig. 4), which is consistent with reports that this process of fragmentation occurs early in the cell-death pathway [39,57,58]. Moreover, the fragmentation of mitochondrial networks induced by METH was found to be mediated by Drp1, a dynamin-related protein which regulates apoptosis-related fission, and inhibition of Drp1 activity by Drp1 siRNA also blocked METH-induced mitochondrial fragmentation (Fig. 5). These results suggest that mitochondrial fragmentation is involved in METH-induced apoptosis and Drp1 plays a crucial role in regulating METH-induced mitochondrial fragmentation in rhNPC. It is well known that METH produces dopaminergic deficits, in part, by affecting neuronal dopamine transport and inducing reactive species, such as ROS and/or RNS in dopaminergic neurons [59,60]. Moreover, METH increases Glu levels in the mammalian brain, which can activate ionotropic receptors, such as NMDA and AMPA receptors. This increase in intracellular Ca 2+ levels and RNS formation [4][5][6] contributes to METHmediated neurotoxicity. However, our results suggest that METH has no effects on Ca 2+ influx or AMPA receptor-mediated Ca 2+ influx induced by glutamate in rhNPC, whereas it does produce an increase in ROS generation in rhNPC ( Fig. 6; Fig. 7). These results indicate that oxidative stress elicited by METH plays an important role in METH-induced mitochondrial fragmentation in rhNPC, although some other factors may be involved in this process.
Mitochondria serve as a main source of ROS [61] and also as the target of ROS. Endogenous formation of ROS and alteration of the cells' redox status have been suggested to facilitate cell death [62,63]. These ROS may target some proteins which are associated with mitochondrial fragmentation and change their structure to direct mitochondrial networks towards fragmentation. Recent studies have suggested that mitochondrial fragmentation may be regulated by a variety of cellular events, including cell division, metabolic flux, cell death and differentiation [64]. Most of these processes seem to alter the post-translational modification of Drp1, including its phosphorylation [65], ubiquitinylation [66] and sumoylation [67,68]. In this study, we found that the oligomerization and localization of Drp1 in cells was also associated with METH-induced mitochondrial fragmentation (Fig. 8). These findings suggest that METH-induced mitochondrial fragmentation is mediated by ROS, which indirectly regulates the oligomerization and translocation of Drp1 from the cytosol to the mitochondria.
A common theme throughout the diverse dynamin superfamily is that self-assembly and oligomerization play important roles in the function of these proteins [69][70][71]. Although Drp1 exists primarily as a cytosolic homo-tetramer, it can also self-assemble into higher order structures on the mitochondrial outer membrane, where it is required for proper mitochondrial division. How Drp1 forms oligomers and subsequently mediates mitochondrial fragmentation remains to be elucidated. It has been reported that C-terminal GTPase effector (GED) domains are important for stimulation of GTPase activity, formation and stability of higher order complexes and efficient mitochondrial division [47]. We screened the primary structure of Drp1 and found that it contains several cysteine residues that may be sensitive to ROS. The thiol crosslinking agent diamide also induces Drp1 oligomerization [72]. These findings support the idea that oxidative stress plays an important role in the process of Drp1 oligomerization.
Drp1 oligomerization was a prerequisite for mitochondrial fragmentation, and these oligomers cannot function until they translocate from the cytosol to mitochondrial outer membrane. Our results demonstrate that oxidative stress increases the translocation of oligomeric Drp1 to the mitochondrial membrane, and this was reversed by the reducing agent, NAC (Fig. 8). Taken together, our results indicate METH-induced oxidative stress plays an important role in mitochondrial fragmentation, particularly in the process of Drp1 oligomerization and translocation from the cytosol to the mitochondria. These findings will provide insights into understanding the pathogenesis of neurodegenerative disorders related to METH abuse.

Materials and Methods
Isolation and Culture of rhNPCs rhNPC cultures were prepared as reported previously [73], with some modifications. Briefly, brain hippocampi were dissociated from embryonic day 17 (E17) Sprague-Dawley rat embryos following protocols approved previously by the University of Nebraska Medical Center Institutional Animal Care and Use Committee and utilizing National Institutes of Health (NIH)  for 20 min at 37uC. After washing with HBSS twice, cells were titrated to a single-cell suspension and cultured on poly-D-lysine pre-coated 6-well plates, 24-well plates or plates containing precoated cover slips. They were then used for immunostaining assay, Ca 2+ flux analysis, western blotting, cell cycle analysis and ROS detection.

Cell cycle assay
Cell cycle was examined by propidium iodide (PI, Sigma) staining [74]. 5610 6 cells were harvested and washed twice with Ca 2+ /Mg 2+ free phosphate-buffered saline (PBS), fixed overnight in 70% cold ethanol, digested with RNase A (Sigma) and stained with PI (100 mg/ml). Data were obtained and analyzed by flow cytometry with the CellQuest software on a FACScan (Becton-Dickinson, New Jersey, USA) using a cell population from which debris were gated out.

Mitochondrial staining
Cells grown on cover slips were stained with MitoTrackerHRed CM-XRos (Invitrogen, Carlsbad, CA) at 37uC in a humidified 5% CO 2 atmosphere for 15 min and fixed with 3.7% formaldehyde in culture medium for another 15 min at 37uC. After washing with PBS twice, cells were subjected to immunostaining and/or mounted with SlowFade light antifade reagent (Molecular Probes) and analyzed by Zeiss Axiovert microscope.

Immunofluorescence
Cells grown on cover slips were fixed with 3.7% formaldehyde in culture medium for 15 min at 37uC. After washing with PBS twice, cells were permeabilized with 0.2% Triton X-100 in PBS at 4uC for 10 min, then blocked with 2% bovine serum albumin (BSA) at room temperature for 2 hours. Cells were further incubated for 2 hours at room temperature with antibodies to nestin

Intracellular calcium measurements
Cells cultured on glass cover slips were loaded with 7.1 mM fura II-AM for 30 min at 37uC in Ringer's solution [145 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 10 mM HEPES, 2 mM CaCl 2 and 10 mM D-glucose, pH 7.4]. Cells were then washed twice and incubated again for 20 min in Ringer's solution to allow for intracellular dye cleavage. The cover slips were inserted into the chamber of a PTI Deltascan System, and fura II was excited at wavelengths of 340 and 380 nm as previously described [75,76].

ROS production determination
rhNPC grown on cover slips were stained with the reduced mitochondrial dye, MitoTrackerHRed CM-H2XROS (Invitrogen), which fluoresces upon oxidation, and were cultured at 37uC in humidified 5% CO 2 atmosphere for 15 min, and fixed with 3.7% formaldehyde in culture medium for another 15 min at 37uC. After washing with PBS twice, cells were subjected to immunostaining and/or mounted with SlowFade light antifade reagent (Molecular Probes) and analyzed by Zeiss Axiovert microscope.

Non-reducing SDS-PAGE and Western Blot assay
Immunoblotting was performed as described [78]. Briefly, the cells were resuspended in lysis buffer [10 mM HEPES (pH 7.4), 2 mM EGTA, 0.5% NP-40, 1 mM NaF, 1 mM NaVO 4 , 1 mM PMSF, 50 mg/ml trypsin inhibitor, 10 mg/ml aprotinin and leupeptin] and placed on ice for 30 min. The lysates were centrifuged at 12,0006g for 12 min at 4uC, and the protein concentration was determined with BSA as a standard. Equivalent samples (30 mg protein) were subjected to non-reducing SDS-PAGE (without b-mercaptoethanol or DTT in loading buffer) on 6% acrylamide gel. The proteins were then transferred onto nitrocellulose membranes, and probed with primary antibodies (mouse monoclonal anti-DLP1/Drp1, BD Transduction Laboratories, Lexington, KY; b-actin, Sigma-Aldrich, St. Louis, MO, as a loading control) followed by the appropriate secondary antibodies conjugated to horseradish peroxidase (KPL, Gaithersburg, MD, USA). Immunoreactive bands were visualized using enhanced chemiluminescence (Pierce, Rockford, IL). The molecular sizes of the proteins were determined by comparison with pre-stained protein markers (Invitrogen). rhNPC were treated with different concentrations of METH as indicated, then subjected to non-reducing SDS-PAGE and western blot assay with Drp1 monoclonal antibody and b-actin as a loading control. The data were quantified by ImageJ software (B) ## denotes p,0.05 as compared to control. (C) Cells were treated with 300 mM METH or 200 mM H 2 O 2 for 24 hours or were pre-incubated with 10 mM NAC for 30 min then with 300 mM METH or 200 mM H 2 O 2 treatment for 24 hours. They were then subjected to non-reducing SDS-PAGE and western blot assay with Drp1 monoclonal antibody and b-actin as a loading control. (D) Cells were treated with 300 mM METH alone or pre-incubated with 10 mM NAC for 30 min, then with 300 mM METH treatment for 24 hours, and subsequently subjected to cell fractionation and western blotting with Drp1 monoclonal antibody as described in materials and methods. VDAC1 was used as a mitochondrial marker and loading control. doi:10.1371/journal.pone.0005546.g008

Subcellular fractionation
Cells were fractionated by differential centrifugation as previously described [79][80][81]. Briefly, cells were harvested through Trypsin-EDTA solution digestion, and then centrifuged and resuspended in three volumes of hypotonic buffer [210 mM sucrose, 70 mM mannitol, 10 mM HEPES (pH 7.4), 1 mM EDTA] containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 50 mg/ml trypsin inhibitor, 10 mg/ml leupeptin, 5 mg/ml aprotinin and 10 mg/ml pepstatin. After gentle homogenization with a Dounce homogenizer, cell lysates were centrifuged at 10006g for 5 min to remove unbroken cells and nuclei, and then cytosolic fractions and the mitochondrial pellets were obtained by further centrifugation at 10,0006g for 30 min.

Statistical analysis
Data were analyzed as means6SD. The data were evaluated statistically by the analysis of variance (ANOVA). Significance was determined as p,0.05 and p,0.01.