Proteasome mediates dopaminergic neuronal degeneration, and its inhibition causes alpha-synuclein inclusions.

Parkinson's disease is characterized by dopaminergic neuronal death and the presence of Lewy bodies. alpha-Synuclein is a major component of Lewy bodies, but the process of its accumulation and its relationship to dopaminergic neuronal death has not been resolved. Although the pathogenesis has not been clarified, mitochondrial complex I is suppressed, and caspase-3 is activated in the affected midbrain. Here we report that a combination of 1-methyl-4-phenylpyridinium ion (MPP(+)) or rotenone and proteasome inhibition causes the appearance of alpha-synuclein-positive inclusion bodies. Unexpectedly, however, proteasome inhibition blocked MPP(+)- or rotenone-induced dopaminergic neuronal death. MPP(+) elevated proteasome activity, dephosphorylated mitogen-activating protein kinase (MAPK), and activated caspase-3. Proteasome inhibition reversed the MAPK dephosphorylation and blocked caspase-3 activation; the neuroprotection was blocked by a p42 and p44 MAPK kinase inhibitor. Thus, the proteasome plays an important role in both inclusion body formation and dopaminergic neuronal death but these processes form opposite sides on the proteasome regulation in this model.

Parkinson's disease (PD) 1 is characterized by selective cell death of the mesencephalic dopaminergic neurons and by the presence of Lewy bodies. However, the relationship between dopaminergic neuronal death and inclusion body formation has not been elucidated. Autosomal recessive juvenile parkinsonism (ARJP) is caused by mutations in the gene encoding Parkin (1), a ubiquitin ligase E3 of the ubiquitin-proteasome system (2), and the mutant gene products lose E3 ligase activity (3).
Furthermore, in patients with sporadic PD, proteasome activity is well preserved in the striatum (4) but reduced in the midbrain (5). Therefore, it has been assumed that insufficient function of the ubiquitin-proteasome system plays an important role in the pathogenesis of PD including ARJP. ␣-Synuclein, mutation of which causes familial PD (6), is one of the major components of Lewy bodies (7,8). It is degraded by the proteasome (9), and proteasome inhibition leads to ␣-synuclein-positive inclusion formation in vitro (10,11) and ␣-synuclein accumulation in dopaminergic neurons in vivo (12). Therefore, impairment of the ubiquitin-proteasome system is thought to be associated with inclusion body formation. The substrates of ubiquitin ligase E3 Parkin are thought to be accumulated in dopaminergic neurons in ARJP; however, Lewy bodies have not been detected in patients with ARJP (13,14) except for one case (15). Despite the absence of Lewy bodies dopaminergic neuronal degeneration starts earlier in patients with ARJP than in those with sporadic PD. In this context, inclusion body formation is not always required for dopaminergic neuronal death, and the role of Lewy body formation in the pathogenesis of PD has not been fully elucidated.
Although the pathogenesis of sporadic cases has not been clarified, mitochondrial complex I is suppressed in sporadic PD patients (16 -18). Exogenous or endogenous neurotoxins such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (19,20) or MPTP analogues (21), which suppress mitochondrial complex I activity, may be involved. MPTP is unique in causing chronic progressive dopaminergic neuronal degeneration in the human brain. It is converted to MPP ϩ by monoamine oxidase and is taken up into dopaminergic neurons through the dopamine transporter. It inhibits the activity of mitochondrial complex I and causes selective dopaminergic neuronal death (22)(23)(24). Recently, it was shown that chronic exposure to rotenone, a selective inhibitor of mitochondrial complex I, causes selective neuronal death of mesencephalic dopaminergic neurons with Lewy-like ␣-synuclein inclusion body formation (25,26). Here we report that dopaminergic neuronal death induced by mitochondria complex I inhibition using MPP ϩ or rotenone was accompanied by an elevation in proteasome activity and that proteasome inhibition caused ␣-synuclein inclusion body formation but blocked dopaminergic neuronal death.

EXPERIMENTAL PROCEDURES
Primary Neuronal Culture of the Ventral Mesencephalon-Cultures of the rat mesencephalon were established according to methods described previously (27). The ventral two-thirds of the mesencephalon were dissected from rat embryos on the 16th day of gestation. The dissected regions included dopaminergic neurons from the substantia nigra and the ventral tegmental area but not noradrenergic neurons from the locus ceruleus. Neurons were dissociated mechanically and plated out onto 0.1% polyethyleneimine-coated plastic coverslips at a density of 1.1 ϫ 10 5 cells/cm 2 . The culture medium consisted of Eagle's minimum essential medium containing 10% fetal calf serum for the first 1-4 days in culture and horse serum from the 5th day onwards. The animals were treated in accordance with guidelines published in the NIH Guide for the Care and Use of Laboratory Animals.
Treatment of Cultures-To investigate MPP ϩ -induced neurotoxicity, cultured neurons were exposed to 1-100 M MPP ϩ for 24 h or 10 M MPP ϩ for 24 -72 h on the 7th day of culture and then fixed. To determine the effects of proteasome inhibitors (lactacystin, proteasome inhibitor I (PSI), MG-132) on MPP ϩ -induced neurotoxicity, the cultures were incubated simultaneously with these inhibitors and 10 M MPP ϩ for 24 or 48 h on the 7th day of culture. Control experiments were sham operations, similar to the treatment but using minimum essential medium with Earle's salts containing no drugs. MPP ϩ was dissolved in water. Rotenone was dissolved in Me 2 SO. Because incubation with less than 0.1% ethanol for 24 h was found to have no effect on neuronal survival rates, lactacystin, PSI, and MG-132 were dissolved in ethanol. In pilot studies concentrations of lactacystin greater than 10 M, PSI greater than 100 M, and MG-132 greater than 1 M were toxic to cultured neurons. Therefore, the drug concentrations used were 10 nM to 1 M for lactacystin, 10 -100 nM for MG-132, and 0.01-10 M for PSI.
Immunocytochemical Investigation-The numbers of surviving neurons were determined using immunostaining as described in our previous study (28). Briefly, after fixation, cultured cells were incubated with anti-tyrosine hydroxylase (TH) antibody (diluted at 1:1000, PA-152, Chemicon) for 24 h, with the secondary biotinylated antibody for 1 h, and with avidin-biotin complex solution (Vectastain) for 1 h. Finally, the cultures were reacted with diaminobenzidine solution for 6 min. The number of cells stained with anti-TH antibody in the 15 randomly selected fields (ϫ200) was counted as the number of surviving dopaminergic neurons by investigators who were blind to the experimental treatments. Neurotoxicity was evaluated by the reduction in the neuronal survival rate in each experiment. Statistical analysis was performed by ANOVA and post-hoc multiple comparison using Newman-Keul's method. Statistical significance was defined as p Ͻ 0.05.
To investigate the accumulation of ubiquitin and ␣-synuclein, cultured cells were incubated with anti-ubiquitin (diluted 1:300, Zymed Laboratories Inc., San Francisco, CA) and anti-␣-synuclein (diluted 1:4000, Chemicon International, Temecula, CA) antibodies for 24 h. Anti-cleaved caspase-3 antibody (Cell Signaling Technology, Beverly, MA) was used to detect caspase-3 activation. Cultured cells were exposed to 30 M MPP ϩ for 18 h and then immunocytochemically stained with anti-cleaved caspase-3 antibody (1:50) for 24 h. A specific inhibitor FIG. 1. Ubiquitin and ␣-synuclein deposition during MPP ؉ -induced neuronal death. A, B, and C, cultured neurons were exposed to 10 M MPP ϩ for 24, 48, and 72 h and fixed for immunocytochemical analysis of TH, ubiquitin, and ␣-synuclein (A). The number of surviving TH-positive neurons, representing dopaminergic neurons, was reduced by MPP ϩ in a manner dependent on exposure time (left column). After 24 h of exposure, small deposits of ubiquitin were seen in the cytosol (higher magnification seen in B), which diminished after exposure to MPP ϩ for 48 or 72 h (middle column). In contrast, ␣-synuclein-positive deposits were not seen after 24 or 48 h of MPP ϩ exposure but were seen after 72 h exposure (right column, higher magnification in C). Scale bars represent 100 m (tyrosine hydroxylase (A)), 50 m (ubiquitin (A)), 50 m  Immunocytochemical Analysis by Double Labeling-A double-immunostaining method was used to investigate the co-localization of ␣-synuclein and TH, a marker of dopaminergic neurons. On the 8th day in culture cells were fixed and incubated with mouse monoclonal anti-TH antibody (1:600, MB318, Chemicon International) for 2 h at room temperature. Cells were then incubated with biotinylated anti-mouse Ig G antibody for 1 h at room temperature, with peroxidase-conjugated avidin-biotin-complex solution (Vectastain) for 1 h and finally with diaminobenzidine solution containing hydrogen peroxide. Cells were then incubated with rabbit polyclonal anti-␣-synuclein antibody (overnight incubation at 4°C, 1:4000) and with biotinylated anti-rabbit Ig G antibody and reacted with avidin-biotinylated alkaline phosphatase complex (Vectastain). Finally, they were visualized using an anti-alkaline phosphatase kit solution (Vectastain) for 15 min.
Subsequently, membrane-bound horseradish peroxidase-labeled antibodies were detected by an enhanced chemiluminescence detection system (ECL-plus, Amersham Biosciences) and exposed to Fuji x-ray film.
Lactate Dehydrogenase (LDH) Release Assay-The release of LDH was measured from culture medium using an LDH assay kit (Kyokuto MTX-LDH assay kit, Tokyo, Japan). The culture medium samples were collected 48 h after the onset of drug exposures and incubated with a substrate solution containing nitro blue tetrazolium, diaphorase, and NAD at 37°C for 45 min. Then the reaction was terminated with a stop solution (0.5 M HCl). The absorbance measurements at 560 nm were taken as LDH release. LDH contained in the 10% horse serum minimum essential medium with Earle's salts was subtracted from each experiment.
Proteasome Activity Assay-Proteasome activity was quantified by measurement of the release of 7-amino-4-methylcoumarin from the fluorogenic peptide Suc-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin using an assay kit, 20 S proteasome assay kit (Affinity Research Product, Exeter, UK). Cultured cells were washed twice with Tris-buffered saline, harvested on ice, and resuspended into a buffer containing 25 mM Hepes and 0.5 mM EDTA. Cells were centrifuged and lysed by brief sonication, added to the fluorogenic substrate Suc-Leu-Leu-Val-Tyr-7amino-4-methylcoumarin, and incubated at 37°C for 30 min. Proteasome activity was detected by changes in fluorescence intensity at 355 nm of excitation and 460 nm of emission using an automatic multi-well plate reader. The relative activity was standardized by protein concentration, which was determined using the Bio-Rad protein assay kit (Bio-Rad). Co-administration of 0.1 M lactacystin at the beginning of MPP ϩ exposure (Lacta (0 -48H)) provided significant neuroprotection. Administration of lactacystin at 12 h after the beginning of MPP ϩ exposure (Lacta (12-48H)) also provided a neuroprotective effect. In contrast, administration more than 24 h after the beginning of MPP ϩ exposure (Lacta (24 -48H) and Lacta (36 -48H)) did not have any neuroprotective effect. *, p Ͻ 0.001 compared with control by ANOVA. #, p Ͻ 0.001 compared with MPP ϩ by ANOVA and post-hoc multiple comparison using Newman-Keul's test. n ϭ 4 coverslips/experiment. NS, not significant. Intranigral Microinjection of MPP ϩ and Proteasome Inhibitors-Male Wistar rats weighing ϳ180 g were used. The rats were fasted overnight with free access to water. For stereotaxic microinjection rats were anesthetized (sodium pentobarbital, 50 mg/kg, intraperitoneal) and immobilized in a Kopf stereotaxic frame. Subsequently, the rats were injected with MPP ϩ (3 g) and lactacystin (0.12 g) or MG-132 (0.19 g) in a final volume of 4 l of sterilized physiological saline (containing 1% Me 2 SO) (n ϭ 4). This was injected into the substantia nigra via a motor-driven 10-l Hamilton syringe. After 7 days treated rats were perfused through the aorta with 150 ml of 10 mM phosphatebuffered saline followed by 300 ml of a cold fixative consisting of 4% paraformaldehyde, 0.35% glutaraldehyde, and 0.2% picric acid in 100 mM phosphate buffer under deep anesthesia with pentobarbital (100 mg/kg, intraperitoneal). After perfusion the brain was quickly removed and post-fixed for 2 days with paraformaldehyde in 100 mM phosphate buffer and then transferred to a 15% sucrose solution in 100 mM phosphate buffer containing 0.1% sodium azide at 4°C. 20-m-thick nigral sections were cut in a cryostat and collected into 100 mM phosphatebuffered saline containing 0.3% Triton X-100. After several washes, the nigral slices were incubated with rabbit polyclonal TH antibody (1: 20000, AB-151, Chemicon International) for 3 days at 4°C. The antibody was detected by an ABC Elite kit (Vector Laboratories) using diaminobenzidine with nickel enhancement. Subsequently, the number of TH-positive neurons in the nigral sections was counted blind to the experiments.

RESULTS
Protein Ubiquitination in MPP ϩ -induced Dopaminergic Neuronal Degeneration-Using rat mesencephalic cultured neurons, we investigated dopaminergic neuronal death, protein ubiquitination, and ␣-synuclein accumulation in MPP ϩ -induced neurotoxicity by immunostaining using anti-TH (a marker for dopaminergic neurons), anti-ubiquitin, and anti-␣synuclein antibodies, respectively. Exposure to MPP ϩ (1-100 M) for 24 h caused dopaminergic neuronal death in a dose-dependent manner but did not cause ␣-synuclein aggregation (data not shown). Long term exposure to 10 M MPP ϩ (for 24, 48, and 72 h) caused dopaminergic neuronal death, which increased with exposure time. A 24-h exposure caused fine cytoplasmic depositions of a ubiquitin-positive substance but did not cause ␣-synuclein aggregation (Fig. 1, A and B). Cytoplasmic depositions of ubiquitin were not detected after MPP ϩ exposure for 48 h or longer (Fig. 1A). In contrast to ubiquitin depositions, ␣-synuclein-positive fine granules were seen in the cytosol after 72 h exposure (Fig. 1, A and C). Quantitative analysis revealed transient ubiquitination followed by ␣-synuclein deposition (Fig. 1D). Double-labeled immunocyto- chemical analysis using anti-ubiquitin (dark blue) and anti-TH (brown) antibodies revealed cytoplasmic protein ubiquitination in dopaminergic neurons (Fig. 1E). About half of the ubiquitinpositive cells were TH-positive neurons (Fig. 1F). After a 72-h exposure to MPP ϩ , double-labeled immunocytochemical analysis using anti-␣-synuclein (dark blue) and anti-TH (brown) antibodies revealed that more than 80% of the ␣-synucleinpositive neurons were dopaminergic neurons (Fig. 1, G and H).
MPP ϩ -induced Dopaminergic Neuronal Death and the Proteasome-MPP ϩ -induced dopaminergic neuronal death was accompanied by cytoplasmic protein ubiquitination, which can be a target of the proteasome. Therefore, we investigated proteasome activity after exposure to MPP ϩ . Although MPP ϩ did not directly elevate the proteasome activity in a cell-free assay (Supplemental Fig. 1), exposure of cultured mesencephalic cells to MPP ϩ for 2 h significantly elevated proteasome activity ( Fig.  2A). The proteasome activity was significantly elevated at 2-48 h after exposure to MPP ϩ (Fig. 2B). The elevated proteasome activity was suppressed by 0.1-1.0 M lactacystin (A. G. Scientific Inc. San Diego, California), a cell-permeable and irreversible inhibitor of both the 20 S and 26 S proteasomes (Fig. 2C). Therefore, 0.1-1.0 M lactacystin was used to investigate the effect of proteasome inhibition on MPP ϩ -induced neurotoxicity. Simultaneous administration of lactacystin significantly blocked MPP ϩ -induced dopaminergic neuronal death at doses of 0.1-1.0 M (Fig. 2D). A dose-response curve of neuroprotection by lactacystin was similar to that of proteasome inhibition (Fig. 2E). A reversible proteasome inhibitor, benzyloxycarbonyl-Leu-Leu-leucinal (MG-132, Calbiochem, 100 nM) also suppressed the proteasome activity (Fig. 2F) and provided significant neuroprotection against MPP ϩ -induced toxicity (Fig.  2G). 48 h of exposure to 1.0 M lactacystin or 100 nM MG-132 alone had no effect on the number of surviving dopaminergic neurons (Supplemental Fig. 2). Immunocytochemical staining revealed that 10 M MPP ϩ alone reduced the number of surviving dopaminergic neurons and shortened the neurites and that co-administration with lactacystin or MG-132 blocked the MPP ϩ -induced neurotoxic effect (Fig. 2H). A cell-permeable irreversible proteasome inhibitor, benzyloxycarbonyl-Ile-Glu(O-t-butyl)-Ala-leucinal (PSI, Carbiochem) significantly blocked proteasome activity (Fig. 2I) and also blocked dopaminergic neuronal death (Fig. 2, J and K). LDH release assay revealed that exposure to 30 M MPP ϩ for 48 h elevated LDH release, and the LDH release elevation was blocked by proteasome inhibitors (Fig. 3). Furthermore, dopaminergic neuronal death by 72 h of exposure to MPP ϩ was also blocked by lactacystin or MG-132 (Supplemental Fig. 3). The time course of the protective effect of a proteasome inhibitor against MPP ϩ -induced dopaminergic neuronal death was investigated. Lactacystin was administered at 0, 12, 24, and 36 h after the beginning of a 48-h exposure to 10 M MPP ϩ . Administration at 0 or 12 h after the beginning of MPP ϩ exposure provided significant neuroprotection, but administration at 24 h or later did not provide any protection (Fig. 4, A and B).
The Proteasome and Dopaminergic Neuronal Death in Vivo -We investigated the effect of proteasome inhibitors on dopaminergic neuronal death in vivo. MPP ϩ was injected stereotaxically into the unilateral mesencephalon, and the rats were sacrificed 7 days after microinjection to investigate dopaminergic neuronal loss. MPP ϩ caused dopaminergic neuronal loss in the ipsilateral substantia nigra, and this was blocked by the co-administration of lactacystin or MG-132 with MPP ϩ (Fig. 5).
Inclusion Body Formation in Dopaminergic Neurons after Combined Treatment with MPP ϩ and Proteasome Inhibitors-The effect of the simultaneous addition of proteasome inhibi- tors to 24 h of MPP ϩ treatment was investigated to determine the role of the proteasome system in ␣-synuclein aggregation. Combination of MPP ϩ and lactacystin caused the formation of ␣-synuclein-positive inclusions in the cytoplasm (Fig. 6A, ␣-synuclein). Co-administration of PSI or MG-132 with MPP ϩ also caused ␣-synuclein-positive inclusion formation. In contrast to treatment with MPP ϩ alone, massive round-shaped inclusions (5-8 m in diameter) were seen in the cytoplasm. We then investigated the localization of the ␣-synuclein inclusions by double-labeled immunostaining using anti-TH (brown) and anti-␣-synuclein (dark blue) antibodies (Fig. 6A, TH/␣synuclein). In the control experiment, TH was stained in the cytoplasm and neurites, and 24 h of MPP ϩ treatment caused shortening of the neurites. However, ␣-synuclein was not detected in either experiment. Combined treatment with MPP ϩ and proteasome inhibitors (lactacystin, PSI, or MG-132) caused ␣-synuclein inclusion formation (dark blue) in the cytoplasm of dopaminergic neurons (brown). These ␣-synuclein-positive inclusions were seen in about 3-8 neurons/cm 2 , corresponding to about 1% of dopaminergic neurons. The inclusion bodies caused by a combination of MPP ϩ and proteasome inhibitors were stained with anti-ubiquitin antibody (Fig. 6A, ubiquitin). Hematoxylin and eosin staining showed that the inclusions were stained with eosin but not with hematoxylin (data not shown).
Treatment with proteasome inhibitors alone did not cause the appearance of ␣-synuclein inclusions (Fig. 6B). Western blotting analysis revealed that MPP ϩ treatment with or without lactacystin caused an elevation in the ␣-synuclein protein level (Fig. 6C).
Rotenone-induced Dopaminergic Neuronal Death and the Proteasome-Furthermore, we investigated dopaminergic neuronal death induced by rotenone, a pharmacological inhibitor of mitochondrial complex I in rat mesencephalon primary culture. Rotenone caused dopaminergic neuronal death in a dose-and time-dependent manner (Fig. 7, A and B). Exposure to rotenone caused a slight but significant elevation in proteasome activity (Fig. 7, C and D). Although rotenone-induced dopaminergic neuronal death (0.1-10 M for 24 h or 0.1 M for 24 -72 h) was not accompanied by ␣-synuclein aggregation (data not shown), at 0.1 M for 48 h it was significantly blocked by the coadministration of the proteasome inhibitors, lactacystin, or MG-132 (Fig. 7, E and F). Combined treatment with rotenone and either lactacystin or MG-132 caused the appearance of massive round inclusions stained with ␣-synuclein (Fig. 7E).
Neuroprotective Mechanism of Proteasome Inhibition-To elucidate the mechanism of neuroprotection by proteasome inhibitors, we investigated the effects of lactacystin on proteins mediating survival signals. Among them, MAPK (29, 30) and Bcl-2 family proteins (31) are reported to be involved in Lewy bodies, and therefore, we investigated these proteins. MPP ϩ treatment slightly reduced the protein level of total and phosphorylated p42 and p44 MAPK and significantly reduced the relative ratio of phosphorylated MAPK to total MAPK. The protein level of phosphorylated MAPK and the relative level of phosphorylated MAPK to total MAPK were significantly elevated by the addition of 1.0 M lactacystin (Fig. 8A). The addition of MG-132 or PSI to MPP ϩ also elevated the protein level of phosphorylated MAPK, and this was blocked by PD098059, an inhibitor of MEK1/2, the kinase of p42 and p44 MAPK (Fig. 8B). Furthermore, the neuroprotective effect of proteasome inhibitors was significantly blocked by PD098059 (Fig. 8C). Western blotting analysis revealed that the protein level of Bcl-X, but not that of Bcl-2, was elevated by lactacystin, but the elevation of Bcl-X was not affected by PD098059 (Supplemental Fig. 4). These data suggest that neuroprotection by proteasome inhibitors was not linked directly with up-regulation of Bcl-X because the neuroprotection was completely blocked by PD098059. Then we investigated the relationship between proteasome activation and caspase-3 cleavage. MPP ϩinduced neuronal death was significantly blocked by a specific caspase-3 inhibitor (Ac-DMQD-CHO) (Fig. 9, A and B). Immunocytochemical analysis using an anti-cleaved caspase-3 antibody revealed that MPP ϩ caused caspase-3 cleavage, which was blocked by 0.1 M lactacystin (Fig. 9, C and D). DISCUSSION Exposure to MPP ϩ for 24 h caused dopaminergic neuronal death accompanied by the cytosolic deposition of an ubiquitinated substance that was not detected after MPP ϩ exposure for 48 h or longer. These data suggest that MPP ϩ exposure caused the ubiquitination of proteins within 24 h that was eliminated within the following 24 h. Furthermore, we found that MPP ϩ exposure elevated the proteasome activity. Protein ubiquitination is a marker targeting proteolysis by the proteasome system, and ubiquitinated proteins might be proteolyzed by activated proteasome and eliminated within the subsequent 24 h. FIG. 8. Effect of lactacystin on MAP kinase. A, a typical demonstration of immunoblotting and semi-quantitative analysis of the ratio of phosphorylated MAP kinase to total MAP kinase. The protein level of MAP kinases (p44 and p42) was slightly reduced by 24 h of treatment with 10 M MPP ϩ . The protein level of phosphorylated MAP kinases was also slightly reduced by MPP ϩ treatment but was elevated by combined treatment with MPP ϩ and 1.0 M lactacystin. The protein levels of phosphorylated and total MAP kinases were determined using a densitometer. The ratio of phosphorylated MAP kinases to total MAP kinases was significantly reduced by MPP ϩ exposure (*). The reduced ratio was significantly elevated by the co-administration of 1.0 M lactacystin (#). *, p Ͻ 0.001 compared with control; #, p Ͻ 0.001 compared with MPP ϩ alone by ANOVA, n ϭ 3. B, immunoblotting analysis of total and phosphorylated MAPK. Exposure to 10 M MPP ϩ for 24 h slightly reduced the protein level of phosphorylated MAPK, which was elevated by 1.0 M lactacystin, 100 nM MG-132, and 100 nM PSI; and this elevation was blocked by 30 M PD098059. C, effect of PD098059 on neuroprotection provided by lactacystin and PSI. Neuroprotection provided by lactacystin (0.1-1.0 M) or PSI (0.01-0.1 M) was significantly blocked by PD098059, an inhibitor of MEK1/2, a MAP kinase kinase. 30 M PD098059 alone had no effect on the survival rate of the dopaminergic neurons (data not shown). *, p Ͻ 0.01 compared with MPP ϩ alone; #, p Ͻ 0.01 compared with MPP ϩ ϩ lactacystin (Lacta) or MPP ϩ ϩ PSI by ANOVA, n ϭ 4 coverslips/experiment. The proteasome inhibitors, lactacystin, PSI, and MG-132, blocked MPP ϩ -induced dopaminergic neuronal death in vitro and in vivo. A temporal profile of neuroprotection by lactacystin revealed that its administration within 12 h after the beginning of MPP ϩ exposure provided neuroprotection. It did not provide neuroprotection when administered after 24 h when cytoplasmic proteins were ubiquitinated. Taken together, these results show that the proteasome, which catalyzes and eliminates ubiquitinated proteins within 24 -48 h after exposure to MPP ϩ , and the protein degradation process by the proteasome could play an important role in dopaminergic neuronal death.
The proteasome inhibitors caused the appearance of roundshaped inclusions in dopaminergic neurons. These data are partly consistent with previous studies (10, 11), in which proteasome inhibitors caused ␣-synuclein inclusions. However, in the present study, neither MPP ϩ exposure alone nor proteasome inhibitors alone led to ␣-synuclein-positive inclusion body formation. One of the reasons for this discrepancy is concentrations of proteasome inhibitors because sub-lethal doses of proteasome inhibitors were used in the present study. In contrast to proteasome inhibitors alone, combination treatment of rotenone and proteasome inhibitors caused the formation of ␣-synuclein-positive inclusions. Therefore, both proteasome suppression and mitochondrial complex I inhibition were required for inclusion body formation in dopaminergic neurons.
Previous studies show that proteasome inhibitors block cell death in some paradigms (32)(33)(34) and enhance it in other paradigms (35)(36)(37). One of the possible reasons for this discrepancy may be related to the mitotic ability of the cells. The proteasome plays an important role in the regulation of pro- FIG. 9. Effect of lactacystin on caspase-3. A and B, immunostaining with anti-TH antibody revealed that MPP ϩ -induced dopaminergic neuronal death was blocked by a caspase-3 inhibitor, Ac-DMQD-CHO. In the control experiment dopaminergic neurons had long neurites. Exposure to MPP ϩ reduced the number of surviving dopaminergic neurons and shortened the neurites. Co-administration of 10 M Ac-DMQD-CHO with MPP ϩ blocked the neurotoxicity (A). Quantitative analysis revealed that MPP ϩ caused significant neuronal death (*), which was blocked by Ac-DMQD-CHO (#) (B). *, p Ͻ 0.001 compared with control. #, p Ͻ 0.001 compared with MPP ϩ alone. n ϭ 4 coverslips/experiment; C and D, immunostaining with anti-cleaved caspase-3 antibody revealed that lactacystin blocked caspase-3 activation by MPP ϩ . Cleaved caspase-3 was not detected in the control experiment (C, boxed area in the lower column at a higher magnification). After exposure to MPP ϩ for 16 h cells were positively stained with anti-cleaved caspase-3 antibody. Lactacystin (0.1 M) blocked the cleavage of caspase-3 by MPP ϩ . Quantitative analysis showed that lactacystin significantly blocked the cleavage of caspase-3 (D). *, p Ͻ 0.001 compared with control. #, p Ͻ 0.001 compared with MPP ϩ alone. n ϭ 6 coverslips/experiment. Bars represent 100 m (A) or 10 m (C). teins that are related to cell cycle progression such as cyclindependent kinases. Therefore, the blockade of proteasome activity could be toxic to cells with mitotic ability but would not be toxic to primary cultured neurons or in vivo brain neurons, which are post-mitotic. Recently Höglinger et al. (38) demonstrated that MPP ϩ or rotenone neurotoxicity is enhanced by 100 nM epoximicin, a proteasome inhibitor which suppresses the proteasome activity to less than 10%. Therefore, overall inhibition of proteasome activity may enhance toxicity, and neuroprotection may require partial suppression of the activity.
The ubiquitin/proteasome system degrades short-lived proteins including proteins mediating signal transduction (39). Among them MAPK is of particular interest because MKK6, the MAPK kinase that activates c-Jun N-terminal kinase, is activated by the polyubiquitination of lysine 63 of MKK6 kinase (40), and p42 MAPK is expressed in Lewy bodies (29,30). We found that the protein level of total and phosphorylated MAPK was slightly reduced, and the relative ratio of phosphorylated MAPK to total MAPK was significantly reduced by MPP ϩ treatment, and that this was reversed by proteasome inhibitors. These data are supported by a previous report demonstrating that phosphorylation of p42 and p44 MAPK is stimulated by lactacystin (41). Stimulation of MAPK phosphorylation by proteasome inhibitors was blocked by PD098059, a selective inhibitor of MEK1/2, p42 and p44 MAPK kinase, and the neuroprotection provided by proteasome inhibitors was also blocked by PD098059 in the present study. MAPK can be regulated by the degradation of MAPK itself directly by the proteasome (42), but these data suggest that MAPK phosphorylation is mediated by MEK1/2 activation. Because Raf, a kinase of MEK1/2, is degraded by the proteasome (43,44), MEK1/2 might be activated by the stabilization of Raf by proteasome inhibitors. Activation of p42 and p44 MAPK provides an antiapoptotic effect (45)(46)(47)(48)(49)(50), and therefore, the neuroprotection resulting from proteasome inhibition was thought to be mediated by the reversion of p42 and p44 MAPK dephosphorylation.
Previous studies on autopsied brains showed that caspase-3 is elevated and activated in the substantia nigra of PD patients compared with that in control patients (51), indicating that caspase-3 activation is involved in the pathogenesis of PD (52). We revealed that MPP ϩ caused caspase-3 activation, which mediated dopaminergic neuronal death, and that lactacystin blocked the caspase-3 activation. The results of the present study indicate that the proteasome plays an important role both in inclusion body formation and dopaminergic neuronal death. In this PD model, inclusion body formation and neuronal death may form opposite sides on the proteasome activity.