Mitochondrial Inhibitor Models of Huntington’s Disease and Parkinson’s Disease Induce Zinc Accumulation and Are Attenuated by Inhibition of Zinc Neurotoxicity in vitro or in vivo

An impairment of energy metabolism underlies neuronal death in neurodegenerative diseases (reviewed in [1]). Huntington’s disease (HD) is a progressive neurodegenerative disorder characterized by chorea, psychiatric disturbances and dementia. The striatum is the primary site of neuronal loss in HD. The mechanism of neurodegeneration and the underlying cause of the selectivity for the striatum are thought to involve inhibition of mitochondrial electron transport at complex II and the selective vulnerability of the striatum to energy impairment [2,3]. Chronic systemic injection of 3-nitropropionic acid (3-NPA, complex II inhibitor) into mice or rats induces bilateral striatal lesions with many neuropathological features of HD and, therefore, is widely used as a model of HD [4,5]. Reactive oxygen species (ROS) have been suggested to play a substantial role in the neurotoxicity of 3-NPA-induced neuronal injury; effective antioxidants partially attenuate (up to 50%) 3-NPA lesions [6,7]. These radicals cause intracellular and extracellular Zn²⁺ accumulation in slices exposed to 3-NPA as determined by Zn²⁺ fluorescence [8].

Oxidative stress is also thought to contribute to dopaminergic cell death in Parkinson’s disease (PD) as evidenced by biochemical changes occurring in the substantia nigra (such as increased iron levels, inhibition of complex I activity and decreased glutathione levels) [9,10]. The catecholaminergic neurotoxin 6-hydroxydopamine (6-OHDA) has been shown to form endogenously in patients suffering from PD and its injection is used as an animal model of PD [11,12].

The involvement of ROS in N-methyl-4-phenylpyridinium (MPP⁺) toxicity of dopamine-secreting neurons is unclear. Free-radical-sensitive fluorophores, as well as measurements of lipid peroxidation and vital dyes, were used to demonstrate increased production of ROS in dopaminergic neurons treated with 6-OHDA compared to MPP⁺ [13,14]. 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) is the precursor to MPP⁺, and both have been shown to induce Zn²⁺ release and an increase in intracellular zinc ([Zn²⁺]_i) in degenerating striatal neurons in vivo and in situ [15,16]. We recently showed that ROS induces an increase in [Zn²⁺]_i and a loss of nicotinamide adenine dinucleotide (NAD⁺), causing inhibition of glycolysis and neuronal death; accordingly, preventing this increase or loss prevents this outcome [17].

In this study, we investigated the role of Zn²⁺ neurotoxicity in 3 neurotoxins (3-NPA, 6-OHDA and MPP⁺) commonly used as models for HD and PD. We examined [Zn²⁺]_i, NAD⁺/NADH levels, glycolytic inhibition and the efficacy of compounds and genotypes, which restore NAD⁺ or reduce Zn²⁺ levels, to attenuate mitochondrial inhibition-mediated injury in vitro, and against 3-NPA injections in vivo. These compounds and genotypes were used due to their efficacy against Zn²⁺ neurotoxicity.

Near-pure cortical neuronal cell cultures were prepared from E15 Swiss-Webster, C57/Bl6J, ZnT3–/–, MT3–/–, or Wallerian degeneration-slow (Wld^s) mice as previously described [18]. Cultures from striatum/midbrain of Swiss-Webster mice were prepared likewise (except at 12 hemispheres/plate), and the growth of glia was inhibited by feeding with Neurobasal medium A + B27 after 3 days [19].

Slow toxicity was initiated after 8 days in vitro, by exposure to 100–1,000 µm 3-NPA, 50–200 µm 6-OHDA or 5–400 µm MPP⁺ in MEM-defined media + 26 mm sodium bicarbonate and 21 mm glucose + 1 µm MK-801 and 50 ng/ml of brain-derived neurotrophic factor (BDNF, Cellsciences, Canton, Mass., USA). Due to the lack of glia, BDNF was included to prevent trophic deprivation, and MK-801 was included to prevent wash-induced glutamate release and excitotoxicity in near-pure neuronal cultures. The efficacy of pyruvate, lactate, thiamine or nicotinamide (10 mm), 6 mm NAD⁺, 10 µm sirtinol or 10 µm N,N,N’N’-tetrakis(2-pyridylmethyl) ethylenediamine (TPEN, specific [Zn²⁺]_i chelator) were tested after dose-finding experiments determined these concentrations to maximally reduce toxicities. Cell death was assayed 24–48 h later by lactate dehydrogenase efflux to the bathing medium, or by staining with 3-(4,5-dimethylthiazol-2-Yl)-2,5-diphenyltetrazolium bromide (MTT) or by staining with propidium iodide (PI), with an n = 9–12 for each condition [20]. Toxicity mediated by 3-NPA takes more time to occur than MPP⁺ or 6-OHDA toxicities (death starts at 18 ± 2 h vs. 4 ± 1 h; n = 9–12).

Near-pure cortical neuronal cultures were loaded with 5 µm Fluo-Zin3-AM for 30 min at 37°C. Cultures were exposed to 800 µm 3-NPA and 200 µm 6-OHDA, 400 µm MPP⁺, 25 µm ethacrynic acid and 80 µm H₂O₂ (ROS generators, positive control). Increases in Fluo-Zin3 fluorescence began at 6 h for 3-NPA and peaked at 8–12 h, began at 2 h for the other toxins and peaked at 3 h. Increases in Fluo-Zin3 fluorescence were seen in striatal/midbrain cultures at lower levels of toxins (5 µm MPP⁺), but with similar kinetics. Striatal/midbrain cultures were coloaded with 5 µm Fluo-Zin3 (Invitrogen/Life Technologies, Carlsbad, Calif., USA) and 10 µm 5–7,dihydroxytryptamine (weakly identifies dopaminergic neurons) [19], and then exposed to 5 µm MPP⁺ for 3 h. Epifluorescence photomicrographs were taken at equivalent exposures at 485/530 for Fluo-Zin3 and 360/450 for 5–7,dihydroxytryptamine. For Zn²⁺ fluorescence in striatal sections, mice were sacrificed after 7× 3-NPA injections (50 mg/kg i.p. 2×/day) and fresh frozen sections of striatum were made, dried and immediately stained in 5 µm Zinpyr-1 (ZP1, TefLabs, Galveston, Tex., USA) for 2 min, and washed 3× in phosphate-buffered saline (PBS). Epifluorescence photomicrographs were then taken at equivalent exposures at 485/530.

Near-pure striatal/midbrain neuronal cultures were fixed, permeabilized in 0.4% Triton X-100, blocked in 10% normal goat serum, incubated with the primary antibody (anti-tyrosine hydroxylase (1:1,000), Chemicon/Millipore, Billerica, Mass., USA) and developed using goat anti-mouse-CY3 secondary antibodies (1:200). Tyrosine hydroxylase positive neurons (TH⁺, >300/pit for control) that remained in each pit were counted.

Near-pure cortical neuronal cultures (7–9 DIV), were used for the NAD⁺ and NADH measurements; cultures were exposed to 800 µm 3-NPA, or to 400 µm MPP⁺ or 200 µm 6-OHDA. Cultures were then washed and lysed by addition of 75% ethanol-0.05 m K₂HPO₄ after toxin exposure. NAD⁺ and NADH were measured as described [21]. The NAD⁺/NADH ratio decreased starting at 12–14 h after 3-NPA exposure and was maximally reduced at 18 h, whereas it decreased starting at 3 h and was maximally reduced at 4 h for MPP⁺ and 6-OHDA which is just prior to the onset of cell death (data not shown).

Dihydroxyacetone phosphate (DHAP) and fructose bisphosphate (FBP) measurements were made on neuronal cell lysates prepared at the same timepoints and conditions, but were lysed by addition of 6% perchloric acid and protein precipitated by addition of potassium carbonate to pH 3.5. DHAP and FBP were measured as described [21].

Long-Evans male rats (body weight: 250–300 g, Charles Rivers, Wilmington, Del., USA), C57/BL6J mice (Jackson Laboratories, Bar Harbor, Me., USA), or Wld^s mice (Harlan, Loughborough, UK) were used in this study. Six rats or mice were used for each condition studied. Housing and anesthesia concurred with guidelines established by the Washington University and LSUHSC institutional Animal Studies Committee, and were in accordance with the PHS Guide for the Care and Use of Laboratory Animals, USDA regulations and the AVMA Panel on Euthanasia guidelines. Two groups of 6 rats were anesthetized with 40 mg/kg pentobarbital intraperitoneally (i.p.), placed into a stereotax and injected with 750 nmol of 3-NPA in the presence or absence of 400 nmol of pyruvate by direct stereotaxic microiontophoretic injection into the striatum at 2 µl/min (0.5 mm anterior bregma, 2.6 mm lateral bregma, 5 mm ventral bregma; the syringe was left in for 5 min after injection). Animals were monitored for 24 h and perfused with cold PBS on day 3 for 2,3,5-triphenyl tetrazolium chloride (TTC) staining. Subacute systemic intraperitoneal injections of 3-NPA in 16-week-old C57/Bl6 mice were performed as described [4]. Six Wld^s male mice aged 16 weeks were similarly exposed to 3-NPA. The surviving, nonakinetic animals from each of the 4 groups of 6 mice were behavior-tested and perfused for TTC staining on the evening of their last 3-NPA injection.

Infarct volumes were measured by morphometric analysis of infarct areas that were defined by a lack of staining in 2% TTC of 1- or 2-mm coronal sections using mouse or rat brain matrices, respectively (Harvard Bioscience, South Natick, Mass., USA). A cross-sectional area of the nonstained region for each brain slice was determined using an image analyzer (DUMAS, Drexel University, Philadelphia, Pa., USA) as described [22].

Mice were tested just prior to euthanasia for clasping, grasping strength and gait, using the tail inversion, inverted-screen and footprint-mapping tests. Animals were scored 0–5 by a rater blinded to the experimental conditions; scores from all 3 tests were added for each mouse, and results are presented as mean ± SEM (n = 6). For tail inversion: 0 = dead, 1 = akinetic, 2 = curled into a ball, 3 = major clasping, 4 = minor clasping and 5 = no clasping. For inverted screen: 0 = dead, 1 = nongrasping/akinetic, 2 = holding on for less than 2 s, 3 = holding on for 2–6 s, 4 = holding on for 6–12 s, and 5 = holding on for longer than 12 s or ‘climb over’. For footprint mapping: 0 = dead, 1 = akinetic, 2 = hindlimb dragging, 3 = footprint nonoverlap >1.5 cm, 4 = footprint nonoverlap 0–1.5 cm, and 5 = complete footprint overlap. Out of six 3-NPA- or 3-NPA + lactate-exposed mice, 3 or 2 died respectively, whereas only 1 Wld^s mouse and no 3-NPA + pyruvate-exposed mice died.

The changes in neuronal death and intracellular NAD⁺ ([NAD⁺]_i) were determined in cultures under the conditions and genotypes stated. The mean ± SEM were plotted and the n given for each experiment in the figure legends. Results were compared to sham-wash or saline-injection controls, as well as to toxin or injury exposure alone. One-way ANOVA was used to assess variance in each set of experiments, followed by a Bonferroni test. Significance was achieved by a p value of less than 0.05.

Unless otherwise stated, all reagents were from Sigma Chemical Co. (St. Louis, Mo., USA).

The toxins studied here (MPP⁺, 6-OHDA and 3-NPA) have been reported to induce ROS, and ROS have been shown to induce [Zn²⁺]_i release [17]. We show that 3 or 8 h after exposure of striatal/midbrain or cortical neuronal cultures to toxic levels of these compounds (fig. 1a, b), [Zn²⁺]_i levels are optimally increased as measured using Fluo-Zin3 fluorescence dye microscopy. The increase observed in cortical neurons is equal to, or for 3-NPA greater than, the strong ROS generators, H₂O₂, or ethacrynic acid (quantified in fig. 1c). 5–7,Dihydroxy-tryptamine is at best a weak fluorescent marker for TH⁺ neurons [19], and Fluo-Zin3 fluorescence was also shown to increase after MPP⁺ exposure in these neurons. The increase in [Zn²⁺]_i was oxidant mediated and specific for zinc because trolox (antioxidant) and TPEN partially and fully attenuated the increase in [Zn²⁺]_i after 3-NPA, respectively (fig. 1c). The partial efficacy of antioxidants against each of these neurotoxins has been reported previously [19,23,24].

The concentrations and time courses of the neurotoxins were chosen to result in approximately 50% neuronal death. The concentrations of compounds used to attenuate toxicity were optimized (data not shown). The therapeutic compounds pyruvate, nicotinamide, NAD⁺, sirtinol and TPEN were demonstrated to attenuate death induced by 3-NPA and MPP⁺ in near-pure cortical neuronal cultures to varying extents (fig. 2). Sirtinol is an inhibitor of the sirtuin pathway, which is an NAD⁺ catabolic deacetylase enzyme involved in transcriptional regulation. Neurotoxicity induced by 6-OHDA was not responsive to Zn²⁺ chelation (TPEN) or Zn²⁺ excess, but was responsive to the restoration of NAD⁺ levels (pyruvate, nicotinamide, NAD⁺ and sirtinol). In contrast, equivalent concentrations of lactate were less effective than pyruvate for all toxins; lactate maintained partial efficacy for 3-NPA and 6-OHDA, but not for MPP+. Thiamine, a pyruvate dehydrogenase and α-ketoglutarate dehydrogenase cofactor, was ineffective for all toxins. Pyruvate, nicotinamide, lactate, thiamine and NAD⁺ do not chelate Zn²⁺ [17,25]. Other than lactate, the beneficial compounds increase the intracellular concentration of NAD⁺ (table 1 and [17]). The number of TH⁺ neurons was reduced by lower toxin exposures and restored by pyruvate, NAD⁺ and TPEN in striatal/midbrain cultures, confirming that similar mechanisms are involved in TH⁺ neurons (fig. 3).

Near-pure cortical neuronal cultures were exposed to 3-NPA for 18 h and to MPP⁺ and 6-OHDA for 4 h (prior to cell death), which were optimal timepoints for the loss of metabolic intermediates (data not shown). Cultures were assayed for total NAD⁺ and NADH content as stated in the Experimental Procedures section. These neurotoxins all induced a substantial decrease in NAD⁺ levels, and the effect was greatest for 6-OHDA. However, only MPP⁺ (and not 6-OHDA) induced a substantial increase in NADH levels, as expected for a mitochondrial complex I inhibitor. Extracellular pyruvate, nicotinamide and NAD⁺ prevented the decrease in [NAD⁺]_i levels for each neurotoxin; pyruvate did so at the expense of NADH levels, thereby restoring the NAD⁺/NADH ratio. Lactate did not significantly prevent the decrease in the NAD⁺/NADH ratio (table 1).

Cultures were similarly exposed and assayed for DHAP and FBP content as already stated. These neurotoxins (3-NPA, 6-OHDA and MPP⁺) all induced a substantial increase in these glycolytic intermediates indicative of glycolytic inhibition at GAPDH. The increase in DHAP and FBP was greatest for 3-NPA. Extracellular pyruvate, nicotinamide and NAD⁺ restored the intracellular levels of DHAP and FBP, demonstrating that the glycolytic inhibition was indirect and mediated through the NAD⁺/NADH ratio. Lactate partially reduced the increase in DHAP and FBP levels after 3-NPA, but not to the extent of pyruvate or NAD⁺ (table 2).

Long-Evans male rats were injected with high doses of 3-NPA into the striatum (750 nmol) in the presence or absence of pyruvate; lesion volume was determined after 3 days. We also determined lesion volumes induced by 9 low-dose intraperitoneal injections of 3-NPA into mice. Pyruvate attenuated the lesions in both the rat intrastriatal injections (fig. 4) and in the mouse intraperitoneal injections (fig. 5). Nicotinamide, but not lactate, was also effective at attenuating lesions induced both by intrastriatal or intraperitoneal injections of 3-NPA (data not shown, fig. 4) [26]. Furthermore, Wld^s mice also had smaller lesions after 3-NPA administration. Cortical neuronal cultures derived from Wld^s mice were also resistant to each of the 3 toxins, and the therapeutic compounds further attenuated neurotoxicity (online suppl. fig. 1b; for all online supplementary material, see www.karger.com?doi=10.1159/000336558). Cortical neuronal cultures derived from mice with a genetic deletion of the Zn²⁺ homeostatic proteins, zinc transporter 3 (ZnT3 KO, no synaptic zinc), or metallothionein 3 (MT3 KO, reduced [Zn²⁺]_i) were not significantly different in their sensitivities to the toxins (online suppl. fig. 1a).

Figure 6 shows the mean ± SEM of the summed scores from 3 behavioral tests of each mouse: inverted screen, tail inversion and footprint mapping. The results from each individual test were similar to the combined results. In the animals receiving 3-NPA alone (that scored 0, 0, 0, 10, 9 and 5) or 3-NPA + lactate (that scored 0, 0, 3, 10, 10 and 7), 50% were dead or completely moribund, and the gait and strength of the remainder were very abnormal at the time of sacrifice. In those mice receiving pyruvate (that scored 15, 15, 15, 11, 10 and 10), all were active and had a nearly normal gait, normal strength and little clasping behavior. In Wld^s mice (that scored 0, 10, 12, 10, 9 and 7), one mouse died and the behavioral deficits in the remainder were intermediate (fig. 6).

We have previously demonstrated that exogenous Zn²⁺ [21,27] or oxidant-induced [Zn²⁺]_i release [17] causes selective loss of NAD⁺ levels, glycolytic inhibition (causing a buildup of the intermediates DHAP and FBP), ATP depletion and neuronal death. Manipulations that prevent Zn²⁺ accumulation (chelation) or restore NAD⁺ levels (exposure to pyruvate, nicotinamide, NAD⁺ and sirtuin pathway inhibitors or the use of Wld^s neurons) allow for glycolytic flux and reduce Zn²⁺-mediated neurotoxicities in vitro and in vivo. Zn²⁺ chelation attenuates 3-NPA and MPP⁺ toxicity and pyruvate, nicotinamide and the direct application of NAD⁺ attenuate 3-NPA, MPP⁺ and 6-OHDA toxicities. These data strongly implicate a reduction in the NAD⁺/NADH ratio as an important mechanism, with Zn²⁺-mediation for 3-NPA and MPP⁺ (see: this paper and [26]). Zn²⁺ accumulation has now been demonstrated for each of these neurotoxins in vitro and for 3-NPA and MPTP in vivo (online suppl. fig. 2 and [8,15,16]). Because subacute 3-NPA lesions form over 5–7 days, Zn²⁺ chelator studies are difficult to perform, requiring multiple intracerebral ventricular injections. ROS have been implicated in both the toxicities of these compounds and in the diseases they model, and antioxidants were shown to be partially effective (reviewed in [28,29]). Studies of these toxins suggest that their mechanisms of injury rely on ROS generation induced by mitochondrial inhibition (reviewed in [30,31]). ROS scavengers protect against each of these neurotoxins [19,23,24]. In addition, a combination of an antioxidant and a mitochondrial activator improves motor performance and extends survival in the R6/2-expanded polyglutamine repeat mouse model of HD [32]. Mitochondrial dysfunction has also been demonstrated in patients and in genetic models of PD (e.g. parkin and synuclein), and in patients and genetic models of HD (CAG repeat models) (reviewed in [33,34,35]).

However, 6-OHDA toxicity, when compared to MPP⁺ or 3-NPA toxicity was relatively unresponsive to sirtuin inhibition, Zn²⁺ chelation or Zn²⁺ loading (fig. 2, 3). This may be because 6-OHDA had already induced maximal NAD⁺ loss through an alternate mechanism (table 1), such that Zn²⁺-mediated loss of NAD⁺ is less important. We suggest that the ROS generated by these toxins cause the Zn²⁺ release and accumulation. The differential cell selectivity of these inhibitors likely results from a combination of their differential uptake or activation and the differential susceptibility of the vulnerable regions to oxidative stress and mitochondrial dysfunction [3,30]. For 3-NPA and MPP+, the increase in [Zn²⁺]_i could cause the loss of NAD⁺, or it could exacerbate the NAD⁺ loss induced by the toxins.

MPP⁺ (but not 6-OHDA) induced a reduction in NAD⁺ levels with a proportionate increase in NADH levels, which resulted in glycolytic inhibition. This is consistent with the known inhibition of mitochondrial NADH dehydrogenase by MPP⁺. 3-NPA, and particularly 6-OHDA, induced a reduction in NAD⁺ levels without an increase in NADH levels. This suggests a mechanism involving catabolism of NAD⁺ that results in a block of glycolysis (tables 1, 2). Restoration of NAD⁺ levels by pyruvate, nicotinamide or NAD⁺ prevents glycolytic inhibition and death, irrespective of the source of the dyshomeostasis in the NAD⁺/NADH ratio. Because 6-OHDA is unresponsive to Zn²⁺ chelation, it is also possible that it acts through a non-Zn²⁺-mediated mechanism to induce NAD⁺ loss, perhaps by activating poly-ADP ribosyl polymerase (PARP) (online suppl. fig. 3). PARP is activated by 3-NPA and MPTP, causing a loss of NAD⁺ levels; PARP inhibition is effective against these neurotoxicities in vivo [26,36,37,38]. A similar PARP-mediated mechanism has not been reported for 6-OHDA neurotoxicity, though the dramatic loss of NAD⁺ levels and the efficacy of nicotinamide suggest that PARP- 1 knockout animals should be partially resistant to 3- NPA and 6-OHDA, as has been shown for MPTP [39]. Wld^s mice overexpressed the NAD⁺ synthetic enzyme NMNAT1, allowing better NAD⁺ restoration [40], and they had reduced susceptibility to MPTP-induced nigrostriatal degeneration [41]. They also displayed a dramatically reduced cell loss and Zn²⁺ fluorescence after target deprivation [17], as well as a modestly reduced 3-NPA lesion volume (fig. 5). Measurements of wild-type and Wld^s striatal NAD⁺ levels after 3-NPA exposure are ongoing. Neurons derived from Wld^s mice were resistant to each of these mitochondrial toxins, and were synergistic with nicotinamide (online suppl. fig. 1), which mimics its effect on neuronal death mediated by another Complex I inhibitor, rotenone [42]. Neurons from ZnT3 KO or MT3 KO were not resistant to toxicity from 3-NPA, 6-OHDA or MPP⁺. ZnT3 is the Zn²⁺ transporter which loads synaptic zinc [43]. MT3 is the metallothionein present in neurons, and is a zinc homeostatic protein and an antioxidant [44]. The ZnT3 KO data is not surprising because synaptic release of zinc is unlikely to be involved in vitro due to the large culture volumes. However, the lack of effect of MT3 KO on neuronal toxicity is somewhat surprising, and suggests that the role of MT3 as an antioxidant protein is equally important as its role as an [Zn²⁺]_i homeostasis protein in these in vitro toxicities.

We demonstrate here that when the [NAD⁺]_i levels were maintained by pyruvate conversion to lactate, by nicotinamide precursor supplementation (and NAD⁺ catabolic blocker), or by direct extracellular NAD⁺ supplementation, then glycolytic inhibition and death induced by mitochondrial toxins were prevented. Pyruvate was previously shown to attenuate MPP⁺-induced and 6-OHDA-induced neurotoxicities. However, the mechanism for this efficacy was suggested to be due to its antioxidant capabilities, whereas here we demonstrate that pyruvate acts by restoring the NAD⁺/NADH ratio [45,46,47]. Powerful ROS scavengers have only been shown to moderately attenuate (<50%) these mitochondrial inhibitors, whereas pyruvate was more efficacious both in vitro and in vivo. Also, pyruvate is not an antioxidant in the usual sense of absorbing excess electrons. Rather, its antioxidant effect results from the ability of alpha-ketoacids to undergo nonenzymatic decarboxylation in the presence of H₂O₂ causing their mutual destruction [48]. By withdrawing H₂O₂, this mildly reduces levels of other oxidative species via Fenton chemistry reactions. Pyruvate was ineffective against redox active metal toxicity (Cu²⁺, Cd²⁺) and much less effective against ethacrynic acid or menadione toxicities (ROS-mediated), suggesting its relative weakness as a normal antioxidant [22,49,50]. We suggest that pyruvate (and nicotinamide or NAD⁺) restores the NAD⁺/NADH ratio, whether the cause of the decrease is complex 1 inhibition (MPP⁺), PARP-1 activation (MPP⁺ or 6-OHDA), or Zn²⁺-mediated NAD⁺ loss (3-NPA or MPP⁺). This restoration of the NAD⁺/NADH ratio allows glycolytic flux. Exogenous lactate is unable to restore the NAD⁺/NADH ratio, but may be beneficial to remaining astrocytes or as an additional energy substrate.

Thiamine had no beneficial effect on 3-NPA, 6-OHDA or MPP⁺ neurotoxicity. This suggests that in contrast to redox active metals or exogenous oxidative agents, direct ROS-mediated pyruvate dehydrogenase or α-ketoglutarate dehydrogenase inhibition might not be involved in the neurotoxicity of these neurotoxins. These data suggest that 3-NPA and MPP⁺ cause a ROS-mediated insult leading to Zn²⁺ accumulation and NAD⁺ loss resulting in glycolytic blockade, which are all necessary for neuronal death. Non-Zn²⁺-mediated NAD⁺ loss, especially for 6-OHDA, also contributes to neuronal death. It also suggests that these compounds should be investigated for their efficacy against HD and PD.

We would like to thank Darlene Guillot for assistance with graphics and Ryan Labadens for scientific editing. This work was funded by NIH grants NS030337 and DK073446 (C.T.S.).

J.Z. and W.Z contributed equally to this work.An abstract of this study has been presented at the Society for Neurosciences in 2003, New Orleans, La., USA