Parkinson's disease is characterized by the progressive and selective loss of dopaminergic neurons in the substantia nigra. It has been postulated that endogenously formed CysDA (5-S-cysteinyldopamine) and its metabolites may be, in part, responsible for this selective neuronal loss, although the mechanisms by which they contribute to such neurotoxicity are not understood. Exposure of neurons in culture to CysDA caused cell injury, apparent 12–48 h post-exposure. A portion of the neuronal death induced by CysDA was preceded by a rapid uptake and intracellular oxidation of CysDA, leading to an acute and transient activation of ERK2 (extracellular-signal-regulated kinase 2) and caspase 8. The oxidation of CysDA also induced the activation of apoptosis signal-regulating kinase 1 via its de-phosphorylation at Ser967, the phosphorylation of JNK (c-Jun N-terminal kinase) and c-Jun (Ser73) as well as the activation of p38, caspase 3, caspase 8, caspase 7 and caspase 9. Concurrently, the inhibition of complex I by the dihydrobenzothiazine DHBT-1 [7-(2-aminoethyl)-3,4-dihydro-5-hydroxy-2H-1,4-benzothiazine-3-carboxylic acid], formed from the intracellular oxidation of CysDA, induces complex I inhibition and the subsequent release of cytochrome c which further potentiates pro-apoptotic mechanisms. Our data suggest a novel comprehensive mechanism for CysDA that may hold relevance for the selective neuronal loss observed in Parkinson's disease.
The underlying pathology in PD (Parkinson's disease) is characterized by a significant degeneration of dopaminergic neurons in the substantia nigra pars compacta of the basal ganglia [1,2]. Although the precise mechanism contributing to the vulnerability of these neurons to degeneration remains unclear, previous evidence suggests that oxidation of DA (dopamine) to its respective o-quinone, and the latter metabolite with cellular thiols to form CysDA (5-S-cysteinyldopamine), may partly mediate this highly specific pathology [3–6]. CysDA formation occurs via the facile nucleophilic addition of cysteine to o-quinone intermediates [7,9], or by an initial glutathionylation and subsequent hydrolysis by γ-glutamyltransferase and cysteinylglycinase respectively to the cysteinyl conjugate [9,10]. Both glutathionyl and cysteinyl derivatives have been detected in mammalian brain tissue [11–13] and are elevated in post-mortem brain tissue from patients with PD . The neurotoxic potential of CysDA has been described in neurons , where increases in oxidative DNA base modification parallel activation of caspase 3 activity . However, the further intracellular oxidation of CysDA may potentiate neuronal injury through the production of oxidants  and through the formation of DHBT-1 [7-(2-aminoethyl)-3,4-dihydro-5-hydroxy-2H-1,4-benzothiazine-3-carboxylic acid], which has been shown to evoke neurobehavioural responses when centrally administered to mice, despite CysDA having no effects [7–9]. These observations raise the prospect that DHBT-1, and perhaps other oxidative metabolites of CysDA, might represent the cadre of endotoxins formed in neuronal cytoplasm that contribute to neuronal damage through interference with mitochondrial complex I.
Previous observations have indicated that both CysDA and DHBT-1 may accumulate in isolated intact rat brain mitochondria where they selectively inhibit complex I of the electron transport chain [15,16], as well as the KGDHC (ketoglutarate dehydrogenase complex)  and the PDHC (pyruvate dehydrogenase complex) , thereby influencing ATP production [19,20] and promoting oxidative stress [21,22]. An increase in the basal level of intracellular oxidative stress is hypothesized to influence pro-apoptotic signalling [23–24], although limited data exist to link either CysDA or DHBT-1 to such an effect. For example, intracellular increases in oxidants such as superoxide, reported to occur previously in association with CysDA [4,16], may activate JNK (c-Jun N-terminal kinase) [25–28] and downstream apoptotic mechanisms within neurons , as well as effect mitochondria through the mitochondrial transition pore and/or release of cytochrome c [30,31]. There is strong evidence linking the activation of JNK to neuronal injury in response to a wide array of pro-apoptotic stimuli in both developmental and degenerative death signalling [25,32]. In particular, JNK has been reported to be activated by a number of oxidative challenges, such as DA , 4-HNE (4-hydroxy-2,3-nonenal) , reduced expression of SOD1 (superoxide dismutase 1) , hydrogen peroxide [36,37] and by oxidized LDL (low-density lipoprotein) .
Steady-state striatal levels of CysDA have been estimated in the nanomolar range if one assumes uniform tissue distribution. However, CysDA is unlikely to be uniformly distributed within striatal tissues and would accumulate within the dopaminergic microenvironment of neuronal synaptosomes to levels within the micromolar range . Although PD symptoms are primarily due to the loss of dopaminergic neurons in the substantia nigra pars compacta, previous studies have reported abnormal mitochondria content/function  and increased oxidative stress  in the cerebral cortex of PD patients. Such results demonstrate the early involvement of the cerebral cortex in PD due to the convergence of multiple metabolic defects . In the present study, we elucidated the potential mechanism of CysDA toxicity by examining the extent to which pro-apoptotic signalling contributes to the neuronal injury. Our data suggest two related, but independent, pathways by which CysDA may induce neuronal injury. In one pathway, the rapid uptake and intracellular oxid-ation of CysDA yields intracellular oxidants which account for most of its neurotoxic potential via the acute and transient activation ERK1/2 (extracellular-signal-regulated kinase 1/2) and caspase 8. The intracellular conversion of CysDA into DHBT-1 exacerbates oxidative stress and contributes to the subsequent activation of ASK1 (apoptosis signal-regulating kinase 1)/JNK and p38 pro-apoptotic signalling. In the other pathway, the inhibition of complex I by DHBT-1, formed during CysDA oxidation, potentiates apoptosis via cytochrome c release and caspase activation.
Reagents were from Sigma–Aldrich and BDH Chemicals unless otherwise stated. CysDA and DHBT-1 were synthesized as reported previously [5,6,13]. Molecular structures and purity of the CysDA and DHBT-1 standards were verified using ESI–MS equipped with a Bruker micrOTOF™ high-resolution time-of-flight mass spectrometer . Stock solutions (10 mM in water) were stored at −20°C. Antibodies used were anti-ACTIVE MAPK (mitogen-activated protein kinase) (ERK1/2) and anti-ACTIVE JNK (JNK1/2); ERK1/2, pASK1 (Ser967), pASK1 (Thr845), ASK1, phospho-c-Jun (Ser73), cleaved caspase 7 (Asp198), caspase 9, U0126 and PD98059 [MEK (MAPK/ERK kinase) inhibitors] were all from New England Biolabs. LY294002 and Wortmannin [PI3K/Akt inhibitors] were from Tocris Biosciences. Horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Sigma–Aldrich), ECL reagent and Hyperfilm ECL were purchased from GE Healthcare. Ultrapure water (18.2 MΩ·cm) passed through a purification system (Purite) was used for all purposes.
Neuronal culture, treatments and assessment of damage
All experimental procedures and protocols used in the present study were conducted according to the specifications of the United Kingdom Animals (Scientific Proceduces) Act, 1986. Primary cultures of mouse cortical neurons were prepared from 14–16-day-old Swiss mouse embryos (NIH, Harlan) as reported previously [37,42]. DIV6 (6 days in vitro) primary neuronal cells were plated on to 24-well plates at a density of 5×105 cells/well and the neurotoxic effects of 5-S-cysteinyl conjugates of DA (0–500 μM) were assessed at 6, 12, 24 and 48 h post-exposure. For the mitogen-activated protein inhibition experiments, primary neurons were treated with U0126 (10 μM) and PD98059 (50 μM) for 30 min before the addition of CysDA (100 μM). Neuronal damage elicited by treatments was evaluated by the Alamar Blue assay and by morphological examination as detailed previously . Morphological assessment of neuronal damage was made by analysis of phenotypic markers under light microscopy using a Nikon eclipse TS100 at ×40 magnification. Images were captured using a Nikon Coolpix 5400 digital camera fitted with a Coolpix MDC lens adaptor. For recovery experiments, the caspase 3/7 inhibitor, Ac-DEVD-CHO (N-acetyl-Asp-Glu-Val-Asp aldehyde) (10 μM) was co-administered with CysDA for 6 h, after which cells were lysed for caspase assessment. In a separate experiment, Ac-DEVD-CHO (10 μM) was removed after 6 h and replaced with fresh medium for 18 h before assessment of neuronal damage. To assess the influence of MEK inhibition on CysDA-induced neuronal injury, U0126 (10 μM) and PD98059 (50 μM) were co-administered with CysDA (100 μM) for 30 min before the addition of CysDA and incubation for 24 h.
Neuronal accumulation and metabolism of CysDA
Assessment of the uptake of CysDA was performed in primary cortical neurons. DIV6 primary neurons were incubated with a sub-toxic concentration of CysDA (30 μM, final concentration) for 0–12 h. Following exposure, cells were washed three times with ice-cold PBS and rapidly lysed on ice using aqueous methanol (50%, v/v) containing HCl (0.1%) as reported previously . Lysed cells were scraped and left on ice to solubilize for 45 min, and then centrifuged at 1000 g for 5 min at 4°C to remove unbroken cell debris and nuclei. To 50 μl aliquots of sample, 12.5 μl of 25% (w/v) metaphosphoric acid was added and samples were centrifuged at 9400 g for 10 min at 4°C. Samples were then snap-frozen and kept at −80°C until analysis. A Rheodyne injection valve with a 5 μl sample loop was used to manually introduce supernatant fractions on to a Bio-Sil ODS-5S, 5 μm particle size, 4.0 mm×250 mm, C18 column (Bio-Rad Laboratories). The column was eluted with a mobile phase consisting of 50 mM NaH2PO4, 3% (v/v) acetonitrile (pH 2.6) at a flow rate of 1 ml/min and detection of eluents was achieved using an eight-channel CoulArray detector (ESA) set at 175, 250, 325, 400, 475, 550, 700 and 750 mV. Further experiments were carried out to address CysDA oxidation into its benzothiazine derivative (DHBT-1) following a 12 h incubation with CysDA. HPLC analyses were performed using an Agilent 1100 Series linked to diode array and fluorescence detectors and connected to a Bruker micrOTOF™ high-resolution time-of-flight mass spectrometer as reported previously .
Measurements of intracellular oxidative stress and total glutathione
The effects of CysDA on the intracellular redox status was performed using DCDHF-DA (2′,7′-dichlorodihydrofluorescein diacetate) as reported previously . Changes in intracellular ROS (reactive oxygen species) levels were expressed as percentage of increase over unexposed cells. Total glutathione (i.e. glutathione plus glutathione disulfide; GSH+GSSG) levels were determined using the Glutathione Assay Kit from Sigma–Aldrich as reported previously .
For the confocal microscopy experiment, cells were cultured on 13 mm glass coverslips coated with poly-D-lysine. Treatments were performed by adding CysDA (0–500 μM) directly to the medium 6 h before fixing. Following exposure, the medium was removed and neurons were washed twice with HBM (Hepes-buffered medium; 140 mM NaCl, 5 mM KCl, 5 mM NaHCO3, 1.2 mM Na2HPO4, 1.2 mM CaCl2, 5.5 mM glucose and 20 mM Hepes, pH 7.4). Then, 100 μM DCDHF-DA in HBM was placed on the neurons for 30 min in 5% CO2/95% air at 37°C. After washing with PBS, cells were fixed in PBS containing 4% PFA (paraformaldehyde; Sigma–Aldrich) for 10 min. Samples were finally mounted on to glass slides using VECTASHIELD® Hard+Set™ Mounting Medium (Vector Laboratories) and visualized using a Leica TCS SP2 laser-scanning confocal microscope (excitation, 488 nm; emission, 530 nm).
Assessment of complex 1 activity
Complex I (NADH-ubiquinone oxidoreductase) activity was determined in cellular homogenates as described previously . Briefly, primary neurons treated with 100 μM CysDA (0–12 h) were harvested by scraping and lysed by incubating with 0.005% digitonin in HBSS/EDTA (Hanks balanced salt solution plus 5 mM EDTA) for 20 s at room temperature. The solubilization was stopped by the addition of 50 volumes of ice-cold HBSS/EDTA. The lysed cells were then centrifuged at 14000 g for 10 min at 4°C and the pellet was diluted to approximately 1 mg/ml protein in HBSS/EDTA containing 1 μM leupeptin, 1 μM pepstatin and 100 μM PMSF (phenylmethylsulfonyl fluoride). Immediately before performing complex I assays, a 200 μl aliquot of protein suspension was sonicated. The complex I assay was performed by sequentially adding 10 μl of 4.2 mM coenzyme Q0 dissolved in methanol, 10 μl of 10 mM NADH and 30–100 μg of protein to a cuvette containing assay buffer (0.025 M potassium phosphate buffer, pH 8.0, 0.25 mM EDTA and 1.5 mM KCN) pre-incubated at 30°C for 2 min. The total reaction volume was 1 ml. The change in absorbance at 340 nm was measured for 2 min, then 5 μl of 500 μM rotenone was added and the absorbance change recorded for another 2 min to determine the rotenone sensitive activity. The complex I activity was calculated by subtracting the rotenone-sensitive rate from the total rate and using 6.81 mM−1·cm−1 as the combined NADH-ubiquinone molar absorptivity at 340 nm. The results are recorded as the recorded of control experiments.
Measurement of cytochrome c release
A solid-phase ELISA kit (Quantikine®M, R&D Systems) was used to measure cytochrome c release into the cytosol of cultured cortical neurons. Cortical neurons (5×106 cells) were incubated for 0–12 h with 100 μM CysDA. Following exposures, neurons were washed three times with PBS and lysed on ice for 10 min using 50 mM Tris/HCl, 0.5% Triton X-100, 150 mM NaCl, 2 nM EGTA and EDTA, and mammalian protease inhibitor cocktail (1:100 dilution; Sigma–Aldrich). Samples were homogenized further by successive passage through a 26-gauge hypodermic needle. Lysates were then centrifuged at 16000 g for 10 min and 100 μl of diluted supernatant fraction (1:5, v/v) was added to the 96-well microplates coated with monoclonal antibody specific for rat/mouse cytochrome c conjugate. Absorbance of samples was measured at 450 nm in a microplate reader. A standard curve was constructed by adding diluted solutions of cytochrome c standards. The results are recorded as the percentage of control experiments.
Caspase 3/8 activity
Following treatment, primary cortical neurons (5×106 neurons/plate) were washed twice with ice-cold PBS and lysed on ice and caspase 3/8-like protease activity was assessed based on the hydrolysis of the peptide substrate Ac-DEVD-p-nitroanilide (N-acetyl-Asp-Glu-Val-Asp-p-nitroanilide) by caspase 3 or Ac-IETD-p-nitroanilide by caspase 8 resulting in the release of the p-nitroaniline moiety, which absorbs at 405 nm (ε405=10.5). Quantification of the absorbance at 405 nm was carried out using a Versa Max UV microplate reader (Molecular Devices). Suitable control experiments were performed using the caspase 3 (Ac-DEVD-CHO) and the caspase 8 (Ac-IETD-CHO) inhibitors at 200 μM (Sigma–Aldrich). To assess the influence of MEK inhibition on caspase 3 activity, U0126 (10 μM) and PD98059 (50 μM) were co-administered with CysDA (100 μM) for 30 min, before the addition of CysDA and eventual caspase 3 activity measurement at 6 h.
Immunoblotting was performed as reported previously . Blots were incubated with either anti-ACTIVE MAPK polyclonal antibody (1:1000 dilution), anti-ERK1/2 polyclonal antibody (1:1000 dilution), anti-phospho ASK1 (Ser967, Thr845) polyclonal antibody (1:1000 dilution), anti-ASK1 (1:1000 dilution), anti-caspase 7 or anti-caspase 9 polyclonal antibody. The molecular mass of the bands were calculated from comparisons with pre-stained molecular-mass markers (MW 27000–180000 and MW 6500–45000; Sigma–Aldrich) that were run in parallel with the samples. The equal loading and efficient transfer of proteins were confirmed by staining the nitrocellulose with Ponceau Red (Sigma–Aldrich).
All results are shown as means±S.D. for three separate experiments unless otherwise stated. The statistical evaluation of the results was performed by one-way ANOVA followed by a Bonferroni's multiple comparison test using GraphPad InStat version 3.05, with P<0.05 considered as significant.
Neurotoxicity of CysDA, uptake and oxidative stress
CysDA (0–500 μM) was observed to generate significant dose-dependent neurotoxicity at 12, 24 and 48 h, with no toxicity observed at the 6 h time point (Figure 1A). Morphological analyses confirmed neuronal injury in response to CysDA treatment (Figure 1B). Vehicle-treated cortical neurons exhibited a large intact cell body and a finely developed dendritic network (Figure 1B, panel i), whereas following exposure to CysDA (100 and 500 μM, 24 h), neuronal cell bodies appeared shrunken and there was a significant loss of dendritic projections (Figure 1B, panels ii and iii).
Neurotoxicity of CysDA
Exposure of neurons to CysDA (30 μM) for 0–12 h resulted in the intracellular accumulation of CysDA [peak at RT (retention time)=10 min] up until 6 h (750±23 nM; P<0.001) (Figure 2A). A non-significant drop in intracellular CysDA concentration (~15%; P>0.05) was observed after 3 h of incubation, which coincided with the appearance of a new peak at RT=6 min (arrows), occurring at the same redox potential as CysDA (Figure 2A). Injection of chlorpromazine, which contains a phenothiazine ring system, presented an electrochemical potential in the range of 500–700 mV, comparable with that of the unknown compound (results not shown), suggesting that the newly formed derivative may likely be DHBT-1, formed from the oxidation of CysDA. Evidence for the formation of DHBT-1 was further substantiated by the fact that after 12 h of incubation there was a large decrease in CysDA medium concentrations (peak at RT=8.9 min; m/z 273) (72±1.6% decrease by 12 h; P<0.001) that coincided with the appearance of a new peak at RT=22.3 min (Figure 2B). MS analysis indicated a compound with m/z 253 and product ion scans yielded four prominent peaks consistent with the fragmentation pattern of DHBT-1  (Figure 2B). Taken together, these results suggest that CysDA is able to accumulate in neuronal cells, where it is oxidized to yield DHBT-1.
Uptake and metabolism of CysDA in cortical neurons
Exposure of neurons to CysDA (100 or 500 μM; 6 h) also led to significant increases in DCDHF-DA-detectable ROS (Figure 3A; 6 h), at all time points (P<0.05), with levels increasing over time (Figure 3B, black bars). These increases in ROS were accompanied by significant concomitant reductions in total glutathione levels over the same time frame (Figure 3B, white bars).
Effect of CysDA on oxidative stress and glutathione levels
MEK/ERK pathway modulation
To investigate whether CysDA-induced ROS formation led to an alteration in neuronal signalling, we assessed the phosphorylation status of ERK1/2 using an antibody that recognizes the dually phosphorylated Thr202 and Tyr204 of ERK1 and Thr185 and Tyr187 of ERK2, and the singly phosphorylated Tyr204 within the catalytic core of the active form of ERK1 (44 kDa) and ERK2 (42 kDa). Treatment with CysDA (100 μM; 0–12 h) resulted in a significant, but transient, activation of ERK2, which peaked at 0.5 h (3.6-fold; P<0.01) and steadily declined over the next 12 h (Figure 4A). ERK1 phosphorylation in response to CysDA followed a similar pattern, although the increases observed (1.3-fold at 0.5 h) did not achieve significance. At 12 h post-exposure, both ERK1 and ERK2 were decreased below basal levels, although this did not reach statistical significance (P>0.05). By contrast, CysDA exposure did not alter the levels of total ERK1/2 at any time point (Figure 4A).
Modulation of MAPKs by CysDA
Selectively blocking the upstream dual-specific kinases, MKK 1 and 2 (MAPK kinase 1 and 2) with U0126 (0.5 h, 10 μM) or PD98059 (0.5 h, 50 μM) was observed to significantly attenuate CysDA-induced ERK1/2 phosphorylation below basal levels (P<0.01; Figure 4B). However, pre-treating cortical neurons with the PI3K (phosphoinositide 3-kinase) inhibitors LY294002 (50 μM) or wortmannin (100 nM) had no significant impact on ERK1/2 phosphorylation levels (Figure 4B). With regard to neurotoxicity, pre-incubation of cortical neurons (30 min) with U0126 (10 μM) or PD98059 (50 μM), before the addition of CysDA (100 μM), partially restored neuronal viability by 29±4% (P<0.01; n=4) and 19±6% (P<0.05; n=4) respectively, in comparison with that observed in CysDA-treated neurons (Figure 4C), whereas LY294002 (50 μM) or wortmannin (100 nM) had no effect on neuronal viability (P>0.05) (Figure 4C). Neither PD98059 nor U0126 at the concentration used in these assays displayed any neurotoxicity (results not shown).
ASK1, JNK and p38 modulation
In contrast with CysDA-induced ERK1/2 activation, exposure of cortical neurons to CysDA (100 μM, 0–12 h) led to a time-dependent activation of ASK1, via the dual de-phosphorylation of ASK1 at Ser967 and phosphorylation of ASK1 at Thr845, relative to vehicle-treated levels (Figure 5A). CysDA exposure also led to significant time-dependent increases in the phosphorylation/activation of JNK1/2 (p46 and p54), c-Jun (Ser73) and dually phosphorylated p38 (Thr180 and Tyr182) which all peaked between 1 and 3 h (Figure 5B). Parallel blots probed with antibodies that detect total levels of JNK and p38, demonstrated that no modification of the total protein levels were induced by CysDA exposure (Figure 5B). Pre-treatment of cortical neurons with either U0126 (10 μM) or PD98059 (50 μM) had no detectable effect on CysDA-induced activation of JNK1/2 (P>0.05), indicating that JNK modulation was not mediated upstream by alterations in ERK activation (results not shown).
Effect of CysDA on pro-apoptotic signalling in neurons
Mitochondrial complex I inhibition and cytochrome c release
To investigate the potential basis of JNK1/2 activation on exposure to CysDA, we investigated the impact of the CysDA on mitochondrial function . CysDA (100 μM; 0–12 h) exposure led to a time-dependent release of cytochrome c into the cytosol, which was significant 1 h post-exposure (85±23% increase; P<0.01), and reached a maximum after 6 h (160±22% increase; P<0.001) (Figure 6). Assessment of the mitochondrial respiratory chain also revealed that CysDA exposure (100 μM, 0–12 h) led to a time-dependent inhibition of mitochondrial complex I activity, which achieved significance 3 h after incubation (27±5% decrease; P<0.01) (Figure 6).
Effect of CysDA on cytochrome c release and mitochondrial complex I inhibition
Exposure of neurons to CysDA (100 μM) led to an enhanced caspase 3 activity in neuronal lysates, as indicated by increased cleavage of the caspase 3-specific substrate Ac-DEVD-p-nitroanilide (absorption 405 nm), which peaked at 6 h before returning to basal levels (Figure 7A). CysDA (100 μM) also induced the early activation of caspase 8 after 0.5 h (P<0.001) and a second peak of caspase 8 activation later at 6 h post-initial exposure (P<0.001) (Figure 7B). Control experiments in which cells were co-incubated with the caspase 3 and caspase 8 inhibitors (Ac-DEVD-CHO and Ac-IETD-CHO respectively) demonstrated no activity (Figure 7). Co-administration of the caspase 3 inhibitor Ac-DEVD-CHO (10 μM) with CysDA (100 μM) led to a significant reduction in CysDA-induced caspase 3 activation and associated neuronal injury (Figure 8). Furthermore, the co-administration of the MEK inhibitors, UO126 (10 μM) and PD89059 (50 μM) with CysDA induced significant reductions in the level of caspase 3 activation (P<0.01) and neuronal damage (P<0.01). Moreover, when both inhibitors were used in combination, there were further significant reductions in both caspase activity and neuronal injury (Figure 8). CysDA also induced the fragmentation of pro-caspase 9 (49 kDa) into a large fragment resulting from the cleavage at Asp368 (39 kDa), which peaked at 6 h. Probing the same lysate with an antibody that recognizes caspase 7 resulted in the appearance of a large fragment of cleaved caspase 7 following cleavage at Asp198 (20 kDa) (Figure 9)
Increased activity of caspase 3- and caspase 8-like proteases induced by CysDA
Involvement of caspase 3 and MEK signalling on CysDA-induced neurotoxicity
Increased cleavage of caspase 9 and caspase 7 induced by CysDA
A hallmark of PD is the progressive and selective loss of dopaminergic neurons in the substantia nigra pars compacta. Although the precise mechanism of neurodegeneration is unclear, it has been hypothesized that DA-mediated neurotoxicity is partly responsible for the loss of neurons in this select region of the basal ganglia [45–48]. DA-mediated neurotoxicity may proceed via DA auto-oxidation to dopamine-o-quinone [49,50] and its reaction directly with a cysteine residue (or with GSH followed by enzymatic processing of the resulting glutathione S-conjugate) to yield neurotoxic CysDA [4,14–16]. We have shown previously that CysDA may induce apoptosis in cortical neurons via a mechanism linked to its induction of intracellular oxidative stress , although the precise mechanism for neurotoxicity was unresolved. In the present investigation, we show that CysDA may induce neuronal apoptosis via two routes, both dependent on the initial intracellular oxidation of CysDA. Our data suggest that the uptake and intracellular oxidation of CysDA, followed by a rapid activation of ERK2 and the sustained activation of caspase 8/3 underpins part of CysDA's neurotoxic potential. Recent studies have shown abnormal mitochondria content and function, and increased oxidative stress and oxidative responses in the cerebral cortex . More recently, intrastriatal injection of CysDA in mice caused extensive oxidative stress, as evidenced by protein carbonylation and glutathione depletion, and affected the expression of α-syn (α-synuclein) in the pre-frontal cortex . Since cortical neurons show low DAT (dopamine transporter) density , it is therefore unlikely that CysDA may penetrate neurons through such mechanism. Rather, CysDA may cross neuronal membranes via NETs [noradrenaline (norepinephrine) transporters], which are known to transport DA as well as noradrenaline in brain regions with low levels of DAT . Data emanating from our studies also suggest a second pathway of neurotoxicity involving the activation of ASK1/JNK and p38, an event also dependent on CysDA oxidation and potentiated by the inhibition of complex I and cytochrome c release by DHBT-1, a neurotoxin formed as a result of CysDA oxidation. Previous data support such a mechanism, where MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) was associated with toxicity in dopaminergic neuronal culture and was found to temporally involve cytochrome c release, activation of caspase 9, caspase 3 and caspase 8 .
The time frame associated with CysDA neurotoxicity and the observations that demonstrate DHBT-1 formation and complex I inhibition suggest that CysDA neurotoxicity is dependent on its initial oxidation. Oxidative stress is rudimentary to the pathogenesis of PD, involving increased nigral levels of oxidized DNA and RNA coupled with decreases in glutathione and increases in unbound iron [21,47,55], suggesting that oxidation reactions are highly likely in this area of the Parkinsonian brain. Our data are in agreement with earlier observations that the oxidation of CysDA leads to an elevation of neuronal oxidants [4,14,16]. In the present study, we suggest that this induction of intracellular oxidative stress triggers pro-apoptotic signalling, which may be chronically enhanced by the inhibition of complex I by the oxidative metabolite of CysDA, DHBT-1 and its subsequent disruption of the electron transport chain [5,15,56]. Such results indicate a common mechanism for a variety of oxidative insults that lead to neuronal death. For example oxygen-induced death of hippocampal neurons , MPP+ (1-methyl-4-phenylpyridinium)-induced death of neurons  and H2O2-induced apoptosis in cortical neurons  all involve oxidative stress and caspase activation. ROS, such as superoxide and peroxide, generated during CysDA oxidation , may initiate the rapid and transient activation of ERK1/2 and is, in part, responsible for the neurotoxicity of CysDA, with specific MEK inhibitors ameliorating a significant fraction of its neurotoxic potential (39.6±2.1% with U0126; P<0.01). The MEK/ERK signalling pathway is generally regarded as being neuroprotective, although its activation during focal cerebral ischaemia  or following neuronal exposure to peroxynitrite  or glutamate  has been associated with induction of neurotoxicity. Previous studies have suggested that ERK1/2 activation is required for mitochondrial membrane depolarization, cytochrome c release and caspase 3 activation in cellular models [63–65], apoptotic events that we observe in our model and within the expected time frames of activation: ERK phosphorylation (0.5 h), cytochrome c release (1–6 h) and caspase 3 activation (0.5–6 h). In addition, co-administration of the MEK inhibitors, UO126 and PD98059 with CysDA induced significant reductions in the level of caspase 3 activation and neuronal damage, therefore confirming our hypothesis regarding the mechanism of action. However, recent evidence suggests that ERK activation may induce caspase 8 and that caspase 8 may directly activate pro-caspase 3, leading to apoptosis , thereby potentiating Fas signalling and cell death . In the present study, caspase 8 was activated at 0.5 h post-CysDA incubation, a time frame that coincides with the observed ERK1/2 activation. As such, we suggest that CysDA and its oxidation products recruit both intrinsic and extrinsic apoptotic pathways through ERK1/2 modulation and caspase 8 activation. Alternatively, ERK activation may also lead to the inappropriate expression of cell-cycle-related genes, such as cyclin D1 , the expression of which has been observed to be increased in post-mortem substantia nigra pars compacta from PD patients [69,70].
Our data also indicate that the intracellular formation of DHBT-1 is an important mediating step in the neurotoxicity of CysDA. Its ability to inhibit complex I activity, reported previously in in vitro studies [15,17] and after direct stereotactic injection in the nigra [7–9], leads to prolonged intracellular oxidative stress capable of activating pro-apoptotic signalling. Under the in vitro conditions employed in the present investigation, 100 μM CysDA evoked inhibition of the mitochondrial complex I by 3 h, which is consistent with the intracellular oxidation of CysDA to DHBT-1. One of the striking features of PD is a profound decrease in the level of mitochondrial complex I activity in the substantia nigra . Although the factors responsible for mitochondrial impairment in PD are unclear, it has been hypothesized that oxidation of CysDA and the formation of DHBT-1  are contributory. Indeed, the cysteinyl S-conjugate of DA and several of its metabolites have been detected in mammalian brain tissue [11–13] and are elevated in post-mortem brain tissue from patients with PD . The inhibition of complex I has been shown previously to lead to an increase in superoxide generation [71–73], which may induce additional complex I disruption  and/or trigger apoptosis.
Consistent with the inhibition of complex I and the generation of intracellular oxidants, we also observed activation of ASK1, a MKK kinase-5 that plays an essential role in stress-induced apoptosis , and the phosphorylation of JNK, c-Jun and p38 above basal levels, peaking at later time points. ASK1 activation followed a time-dependent increase, suggesting that it may derive from the initial oxidative stress induced by CysDA oxidation, whereas JNK, c-Jun and p38 only peaked at 3 h, possibly suggesting that additional oxidative stress brought about by DHBT-1-dependent complex I inhibition may be involved. During the course of our experiments, we considered that the anti-ACTIVE JNK antibody might be cross-reacting with the dually phosphorylated ERK2 as reported previously [36,75]. However, the pre-treatment of cortical neurons with either U0126 (10 μM) or PD98059 (50 μM) had no measurable effect on CysDA-induced activation of JNK1/2, demonstrating that JNK1/2 activation in response to CysDA in our model was not dependent on ERK activation. Rather, in agreement with previous data, it appears that ASK1 activation (via its de-phosphorylation at Ser967), the phosphorylation of JNK and c-Jun (Ser73) are linked to oxidative stress induced by CysDA and DHBT-1 . Furthermore, DA-induced apoptosis has been reported previously to involve the activation of JNK and c-Jun in striatal neuronal cultures , and activation of p38 in SH-SY5Y neuroblastoma cells .
In summary, we detail a comprehensive mechanism for the neurotoxicity of this endogenous neurotoxin, indicating that it is capable of promoting neuronal death through oxidative mechanisms, involving the rapid activation of ERK1/2 and the subsequent activation of apoptotic signalling and DHBT-1 induced complex I inhibition. Such mechanisms are particularly relevant to PD pathophysiology, since both complex I activity and increased oxidative stress are hallmarks of post-mortem pathology in the brains of patients with PD . Thus our data suggest that the endogenous formation of CysDA in the brain of PD patients may be casually related to neurodegeneration.
apoptosis signal-regulating kinase 1
days in vitro
Hanks balanced salt solution plus 5 mM EDTA
c-Jun N-terminal kinase
ketoglutarate dehydrogenase complex
mitogen-activated protein kinase
mitogen-activated protein kinase kinase
reactive oxygen species
David Vauzour and Jeremy Spencer conceived and designed the experiments. David Vauzour prepared the neuronal culture and performed the neuronal toxicity experiments and immunoblotting. John Pinto and Arthur Cooper performed the uptake experiments and analysis. David Vauzour wrote the paper and Jeremy Spencer, John Pinto and Arthur Cooper revised the paper. All authors gave final approval of the version to be published.
This research was supported by the Medical Research Council [grant number G0400278/N102]. Work from the A.J.L.C. and J.T.P. laboratory was supported in part by the National Institutes of Health [grant numbers ES008421 and CA11842].
Present address: Norwich Medical School, Faculty of Medicine and Health Sciences, University of East Anglia, Norwich NR4 7TJ, U.K.