Mitochondrial complex I dysfunction is regarded as underlying dopamine neuron death in Parkinson’s disease models. However, inactivation of the Ndufs4 gene, which compromises complex I activity, does not affect the survival of dopamine neurons in culture or in the substantia nigra pars compacta of 5-wk-old mice. Treatment with piericidin A, a complex I inhibitor, does not induce selective dopamine neuron death in either Ndufs4+/+ or Ndufs4−/− mesencephalic cultures. In contrast, rotenone, another complex I inhibitor, causes selective toxicity to dopamine neurons, and Ndufs4 inactivation potentiates this toxicity. We identify microtubule depolymerization and the accumulation of cytosolic dopamine and reactive oxygen species as alternative mechanisms underlying rotenone-induced dopamine neuron death. Enhanced rotenone toxicity to dopamine neurons from Ndufs4 knockout mice may involve enhanced dopamine synthesis caused by the accumulation of nicotinamide adenine dinucleotide reduced. Our results suggest that the combination of disrupting microtubule dynamics and inhibiting complex I, either by mutations or exposure to toxicants, may be a risk factor for Parkinson’s disease.
Parkinson’s disease is a common aging-related neurodegenerative disorder, which is characterized by the selective loss of dopamine neurons in the substantia nigra pars compacta (SNpc) of the brain. Despite intense research, mechanisms underlying selective dopamine neuron death are not well defined. Inhibition of mitochondrial complex I has long been one of the leading theories (Abou-Sleiman et al., 2006). The observation that drug abusers accidentally exposed to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) developed Parkinsonism provided the first evidence for this hypothesis because 1-methyl-4-phenylpyridinium (MPP+), the toxic metabolite of MPTP, is a mitochondrial complex I inhibitor (Langston et al., 1983; Dauer and Przedborski, 2003). Furthermore, complex I activity is decreased in the substantia nigra, skeletal muscle, and platelets of patients with Parkinson’s disease (Mizuno et al., 1989; Parker et al., 1989; Schapira et al., 1989). A recent study suggests that some of the subunits of complex I in human Parkinson’s disease brains are oxidatively damaged, resulting in the misassembling and functional impairment of complex I (Keeney et al., 2006).
Chronic treatment of rats and mice with rotenone, a well-established complex I inhibitor, induces many key features of Parkinson’s disease (Betarbet et al., 2000; Sherer et al., 2003b; Inden et al., 2007; Pan-Montojo et al., 2010). These findings provide further support for the mitochondrial complex I inhibition hypothesis. Ectopic expression of the Ndi1 gene, a rotenone- and MPP+-insensitive single-subunit NADH dehydrogenase from Saccharomyces cerevisiae, protects dopamine neurons against MPTP and rotenone toxicities in vitro and in vivo, leading to the conclusion that complex I inhibition is obligatory for dopamine neuron death in the rotenone and MPTP models of Parkinson’s disease (Sherer et al., 2003a, 2007; Seo et al., 2006; Richardson et al., 2007; Marella et al., 2008).
To examine the concept that complex I inhibition underlies dopamine neuron degeneration, we studied a mouse strain lacking a functional Ndufs4 gene that encodes one of the 46 subunits comprising mitochondrial complex I and is required for complete assembly and function of complex I (van den Heuvel et al., 1998; Budde et al., 2000; Petruzzella and Papa, 2002; Scacco et al., 2003; Vogel et al., 2007). We confirmed that deletion of the Ndufs4 gene abolished complex I activity in midbrain mesencephalic neurons cultured from embryonic day (E) 14 mice (Choi et al., 2008). Surprisingly, dopamine neurons in Ndufs4−/− cultures appeared normal and survived as well as neurons from wild-type mice (Choi et al., 2008). The absence of complex I activity did not protect dopamine neurons against MPP+ or rotenone toxicity as would be expected if these compounds act by inhibiting complex I, and Ndufs4−/− dopamine neurons were even more sensitive than Ndufs4+/+ neurons to rotenone toxicity (Choi et al., 2008). These data question the long-held complex I inhibition hypothesis and suggest that there is a complex I–independent mechanism that renders dopamine neurons more susceptible than other neurons to rotenone and MPP+. In this study, we provide further evidence to support our prior finding and elucidate complex I–independent mechanisms responsible for rotenone-induced dopamine neuron death.
Complex I inhibition is insufficient to induce dopamine neuron death in culture and in the substantia nigra of Ndufs4−/− mice
We previously reported a lack of correlation between complex I inhibition and dopamine neuron death upon rotenone treatment or Ndufs4 deletion (Choi et al., 2008). Piericidin A is another well-characterized mitochondrial complex I inhibitor (Gutman et al., 1970; Murai et al., 2006). It is at least as potent as rotenone in inhibiting complex I activity in primary mesencephalic cells (IC50 = 20 or 10 nM for rotenone or piericidin A, respectively; Fig. 1, A and B). We used antibodies against tyrosine hydroxylase (TH), the rate-limiting enzyme in dopamine biosynthesis, as a marker for dopamine neurons. Although 5 nM rotenone had very little effect on complex I activity, it selectively killed 50% of the TH+ dopamine neurons (Fig. 1 C). In contrast, 20 nM piericidin A, which inhibited 65–70% of complex I activity, did not induce selective dopamine neuron death (Fig. 1 D).
We next quantified the number of TH+ dopamine neurons in the SNpc of 5-wk-old Ndufs4+/+ and Ndufs4−/− mice by stereological analysis. Although Ndufs4−/− mice die at postnatal week 7, they appear healthy until ∼5 wk of age (Kruse et al., 2008). The TH-staining pattern and cell distribution were comparable in both genotypes (Fig. 2 A). There was no difference in the total number of TH+ dopamine neurons in the SNpc (Fig. 2 B). Similar results were obtained with 6–7-wk-old mice (Fig. S1). The number of NeuN+ (neuronal nuclei) total neurons in the SNpc was also similar in both Ndufs4+/+ and Ndufs4−/− mice (unpublished data).
One might argue that the Ndufs4 deletion in Ndufs4−/− mice causes developmental compensation by other genes or other developmental defects that minimize the effect of complex I inhibition on dopamine neuron death. An in vitro gene knockout system was used to exclude this possibility. Primary mesencephalic neurons were cultured from wild-type or Ndufs4lox/lox embryos and infected with an adeno-associated virus (AAV) expressing Cre-GFP. Western analysis confirmed that AAV-Cre indeed knocked down Ndufs4 protein expression in Ndufs4lox/lox cultures (Fig. S2). For both cultures, ∼50% of total TH+ neurons were AAV1-Cre-GFP+, indicating that Cre deletion of the Ndufs4 gene in culture does not decrease dopamine neuron survival (Fig. 2 C). Together, these data further support our prior finding that complex I inhibition is insufficient to induce selective dopamine neuron death.
Rotenone-induced dopamine neuron death may involve microtubule depolymerization, accumulation of dopamine, and reactive oxygen species (ROS)
If rotenone is toxic for cultured dopamine neurons lacking complex I activity, what is the alternative mechanism for rotenone toxicity? It has been reported that rotenone induces microtubule depolymerization, which disrupts vesicular transport and causes dopamine accumulation and oxidative stress (Brinkley et al., 1974; Marshall and Himes, 1978; Ren et al., 2005). To evaluate microtubule depolymerization, we performed immunostaining for α-tubulin to visualize microtubules (Fig. 3 A) and Western blot analysis to detect free (depolymerized) versus complex (polymerized) tubulin (Fig. 3 B). Interestingly, rotenone, but not piericidin A, caused microtubule depolymerization in primary mesencephalic cultures, correlating with their selective toxicity to dopamine neurons. In addition, the level of free tubulin was increased 3–12 h after rotenone treatment (Fig. 3 C), whereas the detyrosinated, relatively “old” or stable form of tubulin (Glu-tubulin; Ertürk et al., 2007) was decreased 9–12 h after rotenone treatment (Fig. 3 C). Inhibiting microtubule depolymerization by cotreatment with taxol, a microtubule-stabilizing agent, attenuated rotenone-induced TH+ neuron death as well as cytochrome c release (Fig. 3, D and E). Furthermore, treatment with colchicine, a known microtubule-depolymerizing agent, induced TH+ neuron death to the same extent as rotenone (Fig. 3 D). These data implicate microtubule depolymerization as an initial cell death signal upstream from cytochrome c release and activation of the mitochondrial apoptotic pathway.
Treatment of cell cultures with rotenone or colchicine increased intracellular dopamine (Fig. 4 A). Rotenone-induced dopamine accumulation was blocked by cotreatment with taxol, linking microtubule depolymerization to dopamine accumulation. Rotenone-induced dopamine neuron death was attenuated by pretreatment with α-methyl-ρ-tyrosine, an inhibitor of dopamine biosynthesis (Fig. 4 B). α-Methyl-ρ-tyrosine also prevented rotenone-induced dopamine accumulation (Fig. 4 C). These data suggest that rotenone and colchicine increased intracellular dopamine to a toxic level, which contributes to their toxicity to dopamine neurons.
Consistent with dopamine accumulation, rotenone treatment increased intracellular ROS content (Fig. 5 A). Importantly, taxol attenuated rotenone induction of ROS. Finally, treatment with N-acetylcysteine, a free-radical scavenger and ROS inhibitor, protected TH+ neurons from rotenone toxicity (Fig. 5 B). Collectively, these data suggest that the accumulation of dopamine and ROS acts downstream from rotenone-induced microtubule depolymerization and underlies rotenone toxicity to dopamine neurons.
Role of NADH, microtubule dysfunction, dopamine, and ROS in rotenone toxicity and the hypersensitivity of Ndufs4−/− dopamine neurons
To investigate the mechanism underlying the enhanced sensitivity of Ndufs4−/− dopamine neurons to rotenone, we cultured Ndufs4+/+ and Ndufs4−/− neurons from mesencephalons of individual E14 embryos obtained by breeding Ndufs4+/− mice. As with wild-type cultures, piericidin A did not induce dopamine neuron death in Ndufs4−/− cultures (Fig. 6 A). Furthermore, rotenone caused microtubule depolymerization (Fig. 6 B), and either rotenone or colchicine induced dopamine neuron death (Fig. 6, C and D) in Ndufs4−/− cultures lacking complex I activity. These data suggest that microtubule depolymerization may be responsible for toxicity independent of complex I inhibition.
Significantly, Ndufs4−/− dopamine neurons were even more sensitive to either rotenone or colchicine (Fig. 6, C and D). These data suggest that although complex I inhibition by itself is insufficient to kill dopamine neurons, it can potentiate dopamine neuron toxicity induced by microtubule dysfunction. What mechanisms underlie this hypersensitivity? Because rotenone induced similar levels of microtubule depolymerization in both Ndufs4+/+ and Ndufs4−/− cultures (Fig. 6 B), loss of complex I activity must act outside of microtubule stability to make dopamine neurons more vulnerable to microtubule dysfunction. We hypothesize that loss of complex I activity elevates the complex I substrate NADH, which stimulates endogenous dopamine biosynthesis by enhancing the recycling of tetrahydrobiopterin, a cofactor for TH (Vrecko et al., 1993, 1997). However, under resting conditions, the increase in dopamine may be modest and insufficient to cause oxidative stress because vesicular transport and release are sufficient to maintain low cytoplasmic dopamine concentrations. Upon rotenone treatment, the normal vesicular dopamine transport and release may be blocked as a result of microtubule destabilization. Thus, the combination of Ndufs4 deletion and microtubule dysfunction causes more dopamine accumulation and ROS production than microtubule dysfunction alone, resulting in the hypersensitivity of Ndufs4−/− dopamine neurons to rotenone or colchicine.
To test this hypothesis, we measured NADH content in mesencephalic neurons cultured from Ndufs4+/+ and Ndufs4−/− littermates by two-photon microscopy (Kuksa et al., 2003). Indeed, there was a statistically significant increase in NADH levels in cultures from Ndufs4−/− mice (Fig. 7 A). Rotenone treatment increased intracellular dopamine content in TH+ neurons from Ndufs4+/+ cultures, and the increase was much greater in TH+ neurons prepared from Ndufs4−/− mice (Fig. 7 B). Cotreatment with taxol significantly blocked rotenone induction of dopamine accumulation in both cultures (Fig. 7 C). Furthermore, α-methyl-ρ-tyrosine improved the survival of TH+ neurons cultured from Ndufs4−/− mice to the same level as those from Ndufs4+/+ mice.
Consistent with the dopamine increase, rotenone-induced ROS increase was greater in TH+ neurons prepared from Ndufs4−/− mice than in those from Ndufs4+/+ mice (Fig. 8 A). Taxol attenuated rotenone induction of ROS to the same level in both cultures (Fig. 8 B). Finally, treatment with N-acetylcysteine protected TH+ neurons of both genotypes from rotenone toxicity to a similar degree (Fig. 8 C). Collectively, these data suggest that the increase in basal NADH and the greater accumulation of dopamine and ROS after rotenone treatment in Ndufs4−/− cultures may explain why dopamine neurons lacking complex I activity are more sensitive to rotenone.
Rotenone toxicity to dopamine neurons may involve VMAT2 inhibition and the accumulation of cytoplasmic dopamine
Cytosolic, but not vesicular, dopamine is subject to oxidation, which promotes oxidative stress (Hastings et al., 1996). Thus, we determined whether rotenone increases cytosolic dopamine. Cultures treated with reserpine, an inhibitor of VMAT (vesicular monoamine transporter), were used as a positive control. As expected, reserpine increased cytosolic dopamine while decreasing vesicular dopamine (Fig. 9, A and B). Rotenone did not change the level of vesicular dopamine but significantly increased total (Fig. 4 A) as well as cytosolic (Fig. 9 B) dopamine. To further test this hypothesis, we enhanced VMAT2 function by transducing cultures with an adenovirus expressing VMAT2 (Mosharov et al., 2009); this treatment inhibited rotenone-induced TH+ neuron death (Fig. 9 C). These data support the hypothesis that rotenone disrupts VMAT2 function and thereby enhances accumulation of cytoplasmic dopamine, which is toxic.
Although mitochondrial complex I dysfunction has long been implicated in dopamine neuron death and the pathogenesis of Parkinson’s disease, a causal relationship has not been firmly established. Dopamine neurons cultured for 1 wk from the mesencephalon of E14 Ndufs4−/− mouse embryos have no significant complex I activity, yet they appear normal without apparent defects in survival (Choi et al., 2008). The objective of this study was to provide additional evidence to support our prior finding that complex I inhibition is not sufficient to induce dopamine neuron death and to elucidate alternative mechanisms underlying rotenone-induced, complex I–independent dopamine neuron toxicity. We demonstrate in this study that piericidin A, another potent complex I inhibitor, does not induce selective dopamine neuron death, that in-culture deletion of the Ndufs4 gene does not kill dopamine neurons, and that Ndufs4 deletion does not change the number of TH+ neurons in the SNpc in vivo in 5-wk-old mice. These data support our prior finding that complex I inhibition is not a direct cause of dopamine neuron death.
What is the mechanism by which rotenone selectively induces the death of dopamine neurons lacking complex I activity? We illustrate that rotenone causes microtubule depolymerization and the accumulation of dopamine and ROS in mesencephalic cultures. Inhibition of microtubule depolymerization by cotreatment with taxol almost completely inhibited these processes and blocked rotenone-induced loss of TH+ neurons cultured from both Ndufs4+/+ and Ndufs4−/− littermates. Furthermore, piericidin A did not cause microtubule depolymerization. Treatment with colchicine, a microtubule-depolymerizing agent, induced dopamine accumulation and dopamine neuron death. These data implicate microtubule depolymerization and the accumulation of dopamine and ROS as alternative mechanisms underlying rotenone-induced, complex I–independent dopamine neuron death. Our data are consistent with previous work from other laboratories (Ren et al., 2005; Feng, 2006; Ogburn and Figueiredo-Pereira, 2006).
We discovered that loss of complex I activity elevates baseline NADH levels. The NADH increase and microtubule destabilization associated with rotenone treatment cause greater dopamine and ROS accumulation in Ndufs4−/− dopamine neurons than in Ndufs4+/+ dopamine neurons. This may explain why rotenone-induced TH+ neuron loss is greater in Ndufs4−/− neurons compared with Ndufs4+/+ neurons. This hypothesis agrees with the report that overproduction of dopamine is sufficient to induce dopamine neuron death (Park et al., 2007b).
Interestingly, several Parkinson’s disease–related genes, including parkin and α-synuclein, have been implicated in the regulation of microtubule dynamics. Overexpression of α-synuclein impairs microtubule-dependent trafficking (Lee et al., 2006). Parkin, a protein–ubiquitin E3 ligase, ubiquitinates α- and β-tubulin and accelerates their degradation (Ren et al., 2003). Parkinson’s disease–linked mutations abolish these effects of parkin, which may result in the accumulation of misfolded or damaged tubulin. Parkin also protects dopamine neurons against colchicine (Ren et al., 2009). Transgenic mice expressing a mutant microtubule-associated protein tau found in frontotemporal dementia develop early onset tremor, bradykinesia, abnormal gait, and postural instability, which are reversible by l-3,4-dihydroxyphenylalanine (Ittner et al., 2008). Our data illustrate that treatment with the microtubule-depolymerizing agent colchicine is sufficient to cause dopamine neuron death in primary cultures. Together with our data on the differential effects of rotenone and piericidin A on microtubule function and dopamine neuron death, these results suggest that damage and malfunction of microtubules may underlie dopamine neuron death in Parkinson’s disease. Microtubule dysfunction could interfere with the axonal transport of essential molecules from the cell body to the axon terminals as well as removing molecules destined for degradation away from the axon. Thus, the microtubule dysfunction mechanism could explain, at least in part, the “dying back” phenomenon of dopamine neurons in Parkinson’s disease, a process in which neurites degenerate before cell death.
Microtubule function is required for vesicular transport and release. Our data showed that rotenone-induced microtubule depolymerization led to an increase in total dopamine. However, only cytosolic, but not vesicular, dopamine is subject to oxidation, which elevates ROS and can cause neuron death (Hastings et al., 1996). Indeed, rotenone-induced increase in total dopamine was coupled with an increase in cytosolic dopamine. Because VMAT2 is the primary transporter that packages dopamine into presynaptic vesicles in dopamine neurons of the SNpc, our data implicate rotenone inhibition of VMAT2 function in primary dopamine neurons. The combined effect of microtubule depolymerization and VMAT2 inhibition increases total and cytosolic dopamine after rotenone treatment. Consistent with our findings, transgenic mice with low VMAT2 activity have reduced vesicular dopamine content and develop progressive dopamine neuron degeneration (Caudle et al., 2007). Rotenone inhibits and redistributes VMAT2 and increases cytosolic dopamine in SHSY5Y cells (Watabe and Nakaki, 2007, 2008).
Our results do not seem to agree with the findings derived from ectopic expression of Ndi1, a yeast single-subunit NADH dehydrogenase gene. Ectopic expression of Ndi1 rescues dopamine neurons from rotenone toxicity in vitro and in vivo, leading to the conclusion that complex I inhibition is obligatory for dopamine neuron death in the rotenone model of Parkinson’s disease (Sherer et al., 2003a, 2007; Marella et al., 2008). However, alternative explanations exist for these results. For example, per our hypothesis, ectopic overexpression of Ndi1 could reduce NADH concentration and subsequent dopamine biosynthesis in dopamine neurons, making them less vulnerable to rotenone. Furthermore, because intracellular ATP and other NTPs can bind to cytochrome c, inhibit apoptosome formation, and are critical prosurvival factors (Chandra et al., 2006), it is conceivable that overexpressing Ndi1 may increase ATP production and promote the survival of dopamine cells against a host of cell death signals that are unrelated to complex I inhibition, such as oxidative stress. Indeed, overexpression of Ndi1 completely blocked cell death induced by several pesticides that are not known to be complex I inhibitors, including pyridaben, fenpyroximate, fenazaquin, and tebufenpyrad (Sherer et al., 2007). Ndi1 expression also prevented cell death induced by paraquat (Park et al., 2007a), whose cytotoxicity is largely caused by oxidative stress but is independent of complex I inhibition (Day et al., 1999; Jones and Vale, 2000; Dauer and Przedborski, 2003; Bonneh-Barkay et al., 2005; McCormack et al., 2005; Richardson et al., 2005).
Intriguingly, Ndufs4−/− dopamine neurons were even more sensitive to rotenone or colchicine. The loss of complex I activity as a result of Ndufs4 gene inactivation coupled with microtubule dysfunction exacerbated the accumulation of dopamine and ROS. These data suggest that although complex I inhibition is insufficient to kill dopamine neurons, it can potentiate dopamine neuron death caused by microtubule dysfunction. Our results suggest that the combination of inhibiting complex I and disrupting microtubule dynamics, either by genetic mutations or exposure to toxicants, may present an increased risk for Parkinson’s disease.
Materials and methods
Generation of Ndufs4-null (Ndufs4−/−) mice or embryos
The generation and characterization of Ndufs4−/− mice or embryos have been previously described (Kruse et al., 2008). The Ndufs4 heterozygotes (Ndufs4+/−) were bred to generate littermates of Ndufs4+/+, Ndufs4+/−, and Ndufs4−/− mice or embryos. PCR genotyping results were matched to each embryo at the end of the experiment. For the in vitro knockout experiments, mesencephalic primary neurons were cultured from E14 embryos taken from Ndufs4lox/lox mice (Kruse et al., 2008); embryos from wild-type C57BL/6 mice were used as controls.
Primary mesencephalic neuron cultures and drug treatments
Primary cultured dopamine neurons were prepared from E14 mouse embryos as previously described (Choi et al., 2008), either as single embryo cultures (for Ndufs4+/+ and Ndufs4−/− cultures) or as pooled cultures from one to two pregnant C57BL/6 or Ndufs4lox/lox mice. The plug day was defined as E0.5. Drugs were obtained from Sigma-Aldrich unless otherwise specified. Drug treatments were performed in defined serum-free N2 medium (Invitrogen). Half of the media was replaced with N2 medium on the day before drug treatment and then again at the time of drug treatment. Cells treated with vehicle alone were used as controls. Cultures were treated on day in vitro (DIV) 5–6 with various drugs or on DIV 4 with AAV1-Cre-GFP. VMAT2-expressing adenovirus or control GFP virus (gift from E.V. Mosharov and D. Sulzer, Columbia University, New York, NY) was infected 2 d before rotenone treatment.
Monitoring of mitochondrial oxygen consumption was performed in cells (not purified mitochondria) as previously described (Choi et al., 2008). In brief, healthy intact E14 cultured neurons grown on coverslips were placed in respiration buffer, pH 7.3, consisting of 225 mM mannitol, 75 mM sucrose, 10 mM KCl, 5 mM Hepes, 5 mM K2HPO4, and 1 mg/ml of freshly added defatted BSA in an oxygen-monitoring apparatus (5300A system; YSI Life Sciences) at 37°C. Mitochondrial complex I activity was measured by monitoring a 1.25-µM rotenone-sensitive oxygen consumption rate in the presence of complex I–specific substrates (10 mM pyruvate/2 mM malate/1 mM ADP). The amount of oxygen consumed was calculated, assuming the initial oxygen concentration in the buffer was 0.223 µmol O2/ml.
Immunocytochemistry and quantification of TH+ or γ-aminobutyric acid (GABA)–releasing neurons
These procedures were performed as previously described (Choi et al., 2008). In brief, cells were fixed with 4% paraformaldehyde/4% sucrose at room temperature for 30 min and blocked for 1 h in blocking buffer (PBS containing 5% BSA, 5% normal goat serum, and 0.1% Triton X-100). Cells were then incubated with primary antibodies in blocking buffer at 4°C overnight. Primary antibodies included mouse monoclonal antibody against TH (TH, 1:500; Sigma-Aldrich), rabbit polyclonal antibody against TH (1:50,000; Pel-Freez), rabbit polyclonal antibody against GABA (1:5,000; Sigma-Aldrich), and α-tubulin (1:2,000; Sigma-Aldrich). After three washes with PBS, cells were incubated at room temperature for 1 h with appropriate secondary antibodies. Secondary antibodies were Alexa Fluor 488 (or 568) goat anti–rabbit IgG and Alexa Fluor 568 (or 488) goat anti–mouse IgG (1:200; Invitrogen). Images for α-tubulin staining (Fig. 3 A) were taken at room temperature using an imaging system (Marianas; Intelligent Imaging Innovations, Inc.), which incorporates a fluorescence microscope (Axiovert 200M; Carl Zeiss, Inc.) with a motorized stage, a shuttered 175-W xenon lamp, a digital camera (CoolSNAP HQ; Roper Industries), and a 40× objective lens (Axiovert). Cells that were immunostained positive for TH or GABA antibodies and had neurites twice the length of the soma were scored as TH+ or GABA+ cells. All TH+ cells on a 9-mm diameter ACLAR embedding film were scored. The number of TH+ neurons in one ACLAR embedding film was ∼100–200 in control cultures (vehicle control). Cells that were immunostained positive for the GABA antibody were counted from five representative fields from each ACLAR embedding film and scored as the GABAergic population.
Immunohistochemistry and quantification of TH+ neurons in the SNpc
5-wk-old Ndufs4+/+ and Ndufs4−/− littermate mice were perfused with heparinized saline followed by 4% paraformaldehyde as previously described (Choi et al., 2010). Harvested brains were postfixed in 4% paraformaldehyde overnight and then cryoprotected in PBS with 30% sucrose for ≥2 d. Fixed brains were cut into 40-µm sections and collected in cryoprotectant. Sections were incubated with an antibody against TH (1:5,000; Pel-Freez) in PBS containing 0.1% Triton X-100, 5% BSA, and 5% goat serum. Sections were washed and incubated with biotin-labeled secondary antibody and avidin-biotin solution using the VECTASTAIN Elite kit (Vector Laboratories) for 1 h each at room temperature. Sections were developed in DAB for 2–5 min, rinsed several times in tap water, and mounted. Images for TH staining (Fig. 2 A) were collected at room temperature with an inverted fluorescence microscope (E400; Nikon) using a 4× objective lens (Nikon) equipped with a digital microscope camera (MagnaFire; Optronics, Inc.). MagnaFire software was used for system control and image processing. TH+ neurons were counted stereologically as previously described (Mandir et al., 1999). In brief, low magnification (2.5×) images were taken under a microscope (Axiovert 200M) equipped with motorized stage in three axes. The boundaries of SNpc were outlined using the set of anatomical landmarks according to the mouse brain atlas (Hof et al., 2000). The number of TH+ neurons was counted at high power (65×) at automatically selected random fields of the SNpc, and the number of total TH+ neurons on each slide was calculated following optical dissector rules using a cell counter (StereoInvestigator System; MicroBrightField, Inc.). We began cell counting at the first slide of the SNpc section when TH+ neurons were visible and then on every fourth slide through the entire SNpc. The estimate of the total number of TH+ neurons in each brain was calculated using the optical dissector method (Mandir et al., 1999). Five brains from each genotype (Ndufs4+/+ and Ndufs4−/−) were analyzed for stereological cell counting and presented as the total number of TH+ neurons in the SNpc of each brain.
Protein lysates were prepared from cells and analyzed by SDS-PAGE gel electrophoresis and Western blotting as previously described (Choi et al., 2010). Mouse monoclonal anti–cytochrome c antibody (1:3,000) was purchased from Santa Cruz Biotechnology, Inc. Antibodies against α-tubulin and Glu-tubulin were purchased from Sigma-Aldrich.
Preparation of the cytosolic and mitochondrial fraction for cytochrome c release assay
Cytochrome c release was assayed as previously described (Gross et al., 1998; Choi et al., 2004). In brief, cells were washed with PBS, resuspended in isotonic buffer A (250 mM sucrose, 1 mM EGTA, 10 mM Hepes, pH 8.0, 1 mM DTT, and protease inhibitors), and homogenized using an A-type Dounce homogenizer (Kontes) 40 times. Nuclei and unbroken cells were removed by centrifugation at 120 g for 5 min as the pellet. The supernatant was centrifuged again at 10,000 g for 10 min to collect the heavy membrane pellet (mitochondria), whereas the supernatant was designated the cytosolic fraction (cytosol). The heavy membrane pellet was lysed by vortexing and resuspending it in resuspension buffer (10 mM Hepes, 5 mM MgCl2, 42 mM KCl, and 1% Triton X-100) for 5 min on ice.
Preparation of free and complex tubulin
Free or polymerized tubulin was extracted from cultured cells as previously described (Jiang et al., 2006). In brief, midbrain neuronal cultures maintained in 24-well plates were washed twice at 37°C with 1 ml of buffer A containing 0.1 M MES, pH 6.75, 1 mM MgSO4, 2 mM EGTA, 0.1 mM EDTA, and 4 M glycerol. The cultures were then incubated at 37°C for 5 min in 300 µl of free tubulin extraction buffer (buffer A plus 0.1% [vol/vol] Triton X-100 and protease inhibitors). The extracts were centrifuged at 37°C for 2 min at 16,000 g. The supernatant fractions contained free cytosolic tubulin. The pellet fraction and lysed cells remaining in the culture dish retained polymerized tubulin in microtubules and were dissolved in 600 µl of 25 mM Tris, pH 6.8, plus 0.5% SDS. Equal amounts of total protein from free or polymerized tubulin fractions were analyzed by Western blotting with an anti–α-tubulin antibody.
ROS labeling and quantification
Cells were incubated at 37°C for 15 min with 5 µM CM-DCFDA (5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate acetyl ester; Invitrogen) and washed with and then incubated in PBS at 37°C for another 15 min. The images of eight fields from each well were captured with an inverted fluorescence microscope (E400) equipped with a digital camera (MagnaFire) and a temperature-controlled stage at 37°C. Cells were then fixed with 4% paraformaldehyde and processed for anti-TH immunostaining. After staining, images from the same eight fields were captured and compared. The staining intensity of the CM-DCFDA dye (ROS) in the TH+ neuron population was quantified using the ImageJ program (National Institutes of Health). All TH+ neurons in captured images were scored, and the mean ROS level per TH+ neuron was calculated.
Measurement of intracellular NADH content
Two-photon excitation microscopy was used to measure the intracellular NADH content as previously described (Kuksa et al., 2003). Cells were grown for 5 d on 25-mm glass coverslips (Corning). The coverslips were mounted on a temperature-controlled microscopic stage (Heating Insert P; Carl Zeiss, Inc.) installed on the microscope to maintain the medium temperature at 37°C. The objective lens was preheated up to 37°C by an air stream incubator (ASI 400; Nevtek). NADH autofluorescence (435–485 nm) from the sample was collected through the objective lens, separated from the excitation light by a dichroic mirror, filtered by BG39 to remove scattered UV light, and directed to a photomultiplier tube detector (LSM 510; Carl Zeiss, Inc.). NADH intensity was imaged using a 20× 1.3 NA Plan Neofluar objective lens (Carl Zeiss, Inc.) and 76-MHz, 100-fs pulses of a 730-nm light from a mode-locked Ti:Sapphire laser (Mira; Coherent, Inc.). Images were analyzed with the AIM Image Examiner program (Carl Zeiss, Inc.) and the ImageJ program.
E14 mesencephalic primary neurons (3–4 × 105 cells) were plated on each well of the 24-well plates. Duplicate sets of cells were treated equally with drugs or vehicle control. One set was fixed and immunostained for TH to quantify the total number of TH+ neurons. The other set (2–4 wells for each treatment) of cells was frozen on dry ice, stored at −80°C, and sent to the Neurochemistry Core Laboratory at Vanderbilt University’s Center for Molecular Neuroscience Research to analyze for dopamine content, which was done without prior knowledge of genotypes or drug treatments. In brief, cells were homogenized with 0.1 M trichloroacetic acid containing 10 mM sodium acetate/0.1 mM EDTA/1 mM isoproterenol (internal standard)/10.5% methanol, pH 3.8, and centrifuged at 10,000 g for 20 min. Total dopamine content in the supernatant was quantified by HPLC coupled with electrochemical detection (0.7 V). The HPLC system (Antec Leyden) consisted of a 515 HPLC pump, a 717 plus autosampler, an electrochemical detector (Decade II; Antec Leyden), and an HPLC column (150 × 4.6 mm; Nucleosil C18; Phenomenex). The homogenization buffer was used as the mobile phase (0.7 ml/min), and 20 µl of the sample was injected onto an HPLC column (3.9 × 300 mm; Nova-Pak C18; Waters). Dopamine content was normalized to the number of TH+ neurons in each treatment group. For the cytosolic or vesicular dopamine measurement, synaptosome and cytosol fractions were prepared using a synaptic vesicle preparation kit (SV0100; Sigma-Aldrich) and sent to the Neurochemistry Core Laboratory for dopamine measurement.
Data were from at least three independent experiments, each with at least duplicate or triplicate determinations. Statistical analyses of data were performed using two-way analysis of variance and posthoc Student’s t test.
Online supplemental material
Fig. S1 shows the quantification of TH immunostaining of SNpc from 6–7-wk-old Ndufs4+/+ and Ndufs4−/− mouse brains. Fig. S2 shows that AAV-Cre knocks down Ndufs4 protein expression in Ndufs4lox/lox cultures prepared from E14 mesencephalon.
We thank Dr. Eugene V. Mosharov (Columbia University) for the adenovirus expressing VMAT2 and Dr. Linda Wordeman for scientific discussions.
This work was supported by grants ES012215 and ES013696 from the National Institutes of Health (Z. Xia) and the Postdoctoral Fellowship Program of the Korea Science and Engineering Foundation (W.-S. Choi) and facilitated by grant P30 HD02274 from the National Institute of Child Health and Human Development and by grant NIEHS P30ES07033 from the University of Washington National Institute of Environmental Health Sciences–sponsored Center for Ecogenetics and Environmental Health.