Address correspondence and reprint requests to Kevin St P. McNaught, Department of Neurology, Mount Sinai School of Medicine, Annenberg 14–73, One Gustave L. Levy Place, New York, NY 10029, USA. E-mail: email@example.com
Mutations in α-synuclein, parkin and ubiquitin C-terminal hydrolase L1, and defects in 26/20S proteasomes, cause or are associated with the development of familial and sporadic Parkinson's disease (PD). This suggests that failure of the ubiquitin–proteasome system (UPS) to degrade abnormal proteins may underlie nigral degeneration and Lewy body formation that occur in PD. To explore this concept, we studied the effects of lactacystin-mediated inhibition of 26/20S proteasomal function and ubiquitin aldehyde (UbA)-induced impairment of ubiquitin C-terminal hydrolase (UCH) activity in fetal rat ventral mesencephalic cultures. We demonstrate that both lactacystin and UbA caused concentration-dependent and preferential degeneration of dopaminergic neurons. Inhibition of 26/20S proteasomal function was accompanied by the accumulation of α-synuclein and ubiquitin, and the formation of inclusions that were immunoreactive for these proteins, in the cytoplasm of VM neurons. Inhibition of UCH was associated with a loss of ubiquitin immunoreactivity in the cytoplasm of VM neurons, but there was a marked and localized increase in α-synuclein staining which may represent the formation of inclusions bodies in VM neurons. These findings provide direct evidence that impaired protein clearance can induce dopaminergic cell death and the formation of proteinaceous inclusion bodies in VM neurons. This study supports the concept that defects in the UPS may underlie nigral pathology in familial and sporadic forms of PD.
Degeneration of the dopamine-containing neurons of the substantia nigra pars compacta (SNc) is the primary pathology occurring in Parkinson's disease (PD; Forno 1996). Neuronal death is accompanied by the appearance of cytoplasmic Lewy body inclusions which accumulate a wide range of proteins including α-synuclein, ubiquitin, neurofilaments, and oxidized/nitrated proteins (Pollanen et al. 1993; Good et al. 1998; Spillantini et al. 1998; Giasson et al. 2000). Nigral pathology in PD is associated with oxidative stress, mitochondrial dysfunction, and excitotoxicity but it is not clear how these events contribute to the neurodegenerative process (Jenner and Olanow 1998).
Recent evidence suggests that failure of the ubiquitin–proteasome system (UPS) leading to protein accumulation contributes to degeneration of dopaminergic neurons and Lewy body formation in the SNc in both familial and sporadic forms of PD (McNaught et al. 2001). Various deletions and point mutations in the parkin gene (6q15.2–27), which codes for a ubiquitin ligase, lead to a loss of enzyme activity and destruction of the SNc in cases of autosomal recessive juvenile parkinsonism (AR-JP; Shimura et al. 2001). In rare occurrences of autosomal dominant PD, missense mutations in the gene (4q21–q23) encoding α-synuclein have been shown to produce proteins that are prone to misfold and aggregate (Polymeropoulos et al. 1997; Goedert 2001). Mutant α-synucleins resist and inhibit proteolysis, and increase the sensitivity of cells to a variety of toxic insults (Bennett et al. 1999; Lee et al. 2001; Stefanis et al. 2001; Tanaka et al. 2001; Tofaris et al. 2001). These factors are thought to contribute to nigral dopaminergic cell death and Lewy body formation in α-synuclein-linked familial forms of PD (Goedert 2001; Spira et al. 2001). Additionally, a missense mutation in the gene (4p14) encoding for the de-ubiquitinating enzyme, ubiquitin C-terminal hydrolase L1 (UCH-L1), is associated with rare cases of autosomal dominant PD (Leroy et al. 1998). Further, we recently showed that all three proteolytic activities of 26/20S proteasomes are impaired in the SNc of patients with sporadic PD (McNaught and Jenner 2001; McNaught et al. 2002). The occurrence of impaired clearance of proteins in the SNc of PD patients is supported by evidence for an elevation in the levels of protein oxidation products and the aggregation of a variety of proteins (Yoritaka et al. 1996; Alam et al. 1997; Lopiano et al. 2000).
Thus, alterations in the UPS are associated with the development of nigral pathology in the various etiological forms of PD that have been discovered to date. Indeed, proteasomal inhibition can lead to degeneration of PC12 cells with inclusion body formation (Rideout et al. 2001). However, it has not been demonstrated that defects in the UPS can lead to degeneration and the formation of proteinaceous Lewy body-like inclusions in mesencephalic dopaminergic neurons. We therefore examined the effects of inhibition of 26/20S proteasomal function and UCH activity in fetal rat ventral mesencephalic (VM) cultures.
Materials and methods
Synthetic lactacystin and ubiquitin aldehyde (UbA) were obtained from Calbiochem (Beeston, Nottingham, UK). Cell culture plastics, media and reagents were obtained from Life Technologies (Paisley, UK) and Sigma–Aldrich (Poole, UK). Monoclonal mouse antibodies to tyrosine hydroxylase (TH) and glutamic acid decarboxylase (GAD) were obtained from Boehringer Mannheim (East Sussex, UK). Polyclonal rabbit antibodies to α-synuclein (raised against an internal peptide of the human protein) and monoclonal mouse antibodies to ubiquitin were obtained from Chemicon International Ltd (Harrow, UK). Biotinylated goat anti-rabbit IgG, biotinylated goat anti-mouse IgG, avidin and biotin were components of the Vectastain ABC kit (Vectastain Laboratories Inc., CA, USA). [3H]dopamine (32.6 Ci/mmol) and [14C]γ-aminobutyric acid (GABA, 240 mCi/mmol) were obtained from NEN (Boston, MA, USA). All other reagents/materials were of analytical/cell culture grade and obtained from Sigma–Aldrich and other commercial sources.
Enriched fetal rat ventral mesencephalic neuronal cultures
Neuronal-enriched cultures containing dopaminergic neurons were prepared from the VM of fetuses (14–15 days gestation) obtained from Sprague–Dawley rats (Charles River, Kent, UK) as described elsewhere (Pardo et al. 1997). Cells were suspended in Dulbecco's modified Eagle medium (DMEM; supplemented with 15% fetal bovine serum, 1 mm sodium pyruvate and 4 mm l-glutamine) and plated at a density of 104 cells/cm2 on 13-mm poly d-lysine-coated plastic coverslips in 24-well culture plates. Cells were grown in a humidified atmosphere of 5% CO2/95% air at 37°C for 1 day, after which the culture medium was changed to defined serum-free DMEM–F12 medium (supplemented with 25 µg/mL insulin, 100 µg/mL transferrin, 60 µm putrescine, 20 nm progesterone and 30 nm sodium selenite), and cells grown for a further 6 days before use. Immunocytochemical analyses showed that these VM cultures contained approximately 2% astrocytes and > 95% neurons.
Coverslips containing VM neurons were washed with 0.1 m phosphate-buffered slaine (PBS; pH 7.4) then fixed in 4% paraformaldehyde for at least 1 h before being immunostained using the Vector ABC method (Pardo et al. 1997). Primary antibody preparations for TH, GAD, α-synuclein, and β-synuclein were used at a dilution of 1 : 200, and the ubiquitin antibody preparation was used at a 1 : 500 dilution. Control experiments in which primary antibodies were omitted confirmed the specificity of these immunoreactions (data not shown). Stained coverslips were washed with 0.1 m PBS, dehydrated through ethanol, then mounted onto slides. Slides were examined under a Zeiss Axioskop (Carl Zeiss Inc., Thornwood, NY, USA) microscope equipped with a grid-containing eyepiece for determining neuronal count (approximately 10% of coverslip area counted) and morphology at × 200 magnification as previously described (McNaught and Jenner 1999).
[3H]Dopamine and [14C]GABA uptake
Determination of dopamine and GABA uptake into dopaminergic and GABAergic neurons, respectively, was conducted as described previously (Casper et al. 1991). VM cultures were rinsed with Kreb's phosphate buffer (pH 7.4) and incubated for 30 min at 37°C with the same buffer containing 0.2 mg/mL ascorbic acid, 10 µm GABA, 0.5 µCi/mL [3H]dopamine and 0.05 µCi/mL [14C]GABA. After rinsing, the radioactivity was extracted with 1 mL 95% ethanol, which was added to vials containing scintillation cocktail and the radioactivity measured in a scintillation spectrometer (Packard Tri-Carb 2100; Packard Bioscience, Meriden, CT, USA). VM cultures treated with the neuronal dopamine uptake blocker mazindol (10 µm) and the neuronal GABA uptake blocker diaminobutyric acid (1.0 mm) were used as blanks. For uptake studies, VM cultures were grown in minimum essential medium (MEM) containing serum.
Determination of the effects of inhibition of the UPS on dopaminergic cell viability, protein accumulation and inclusion body formation
The growth medium of VM cultures in 24-well culture plates was replaced with fresh filter-sterilized DMEM-F12 containing up to 10 µm lactacystin (a selective proteasome inhibitor; Fenteany and Schreiber 1998) or up to 10 μm UbA (an inhibitor of UCH; Melandri et al. 1996; Schaeffer and Cohen 1996) for a maximum of 48 h. To determine the extent of neuronal death in these cultures, cell counts were conducted following immunostaining. [3H]Dopamine/[14C]GABA uptake into the remaining plated neurons was also measured as an indicator of cell viability. To determine alterations in morphology, protein accumulation and inclusion body formation in lactacystin/UbA-treated VM cultures, coverslips containing neurons were immunostained as described above.
Inhibition of the UPS causes toxicity to dopaminergic neurons
Treatment of VM cultures with inhibitors of the UPS for up to 48 h produced a concentration-dependent decrease in the number of dopaminergic neurons as shown by a reduction in the number of TH-positive cells. After 24 h in culture, there was a loss of approximately 7% of TH-positive cells in control cultures compared to 40% and 55% reductions following exposure to 5 and 10 µm lactacystin, respectively (Fig. 1a). Similarly, in comparison to controls, the addition of 5 and 10 µm lactacystin inhibited [3H]dopamine uptake into dopaminergic neurons by 55% and 58%, respectively (Fig. 1b). In contrast, 10 µm lactacystin had no significant effect on [14C]GABA uptake after 24 h of exposure (Fig. 1b), and only a 22.9% (p < 0.01) reduction after 48 h of exposure. Treatment of VM cultures with 5 and 10 μm UbA caused approximately 70% and 90% loss of TH-positive cells, respectively (Fig. 1a).
Exposure of VM cultures to 10 µm lactacystin or 10 μm UbA for 24 h was associated with alterations in the morphology of the remaining TH-positive neurons. Both treatments caused destruction of neuronal processes and disintegration of dopaminergic neuronal membranes with ghosting (Fig. 2). Lactacystin caused enlargement of neuronal perikaryon while cell body size remained unchanged or only slightly increased following UbA treatment (Fig. 2).
Protein accumulation in VM neurons following inhibition of the UPS
Control VM cultures were immunoreactive for α-synuclein in axons, dendrites, and terminals but perikarya were poorly stained (Fig. 3a). Exposure for 24 h to 5 µm lactacystin (Fig. 3b) or 5 μm UbA (Fig. 3c), which produced preferential alteration in dopamine cell viability (Figs 1 and 2), caused an increase in the intensity of α-synuclein staining in perikaryon and processes. Immunostaining of control VM cultures for ubiquitin revealed staining in neuronal perikarya and processes (Fig. 3d). The intensity of staining in neuronal perikaryon and processes was increased following exposure to 5 µm lactacystin for 24 h (Fig. 3e) but was reduced following treatment with 5 μm UbA (Fig. 3f).
Inhibition of the UPS resulted in the formation of proteinaceous inclusions in the cytoplasm of VM neurons
VM cultures treated with low concentrations of lactacystin (5 µm) or UbA (5 µm) for 24 h, parameters that produced degeneration of dopaminergic neurons with little/no effect on [14C]GABA uptake (Figs 1 and 2), caused cytoplasmic inclusions that were intensely immunoreactive for α-synuclein (Figs 4a and b) and lightly stained for ubiquitin (Fig. 4c). The α-synuclein/ubiquitin-immunoreactive inclusions varied in number and size from a single large inclusion (Fig. 4a) to several small cytoplasmic pebble-like structures (Fig. 4b) which tended to appear at an earlier time point. Inclusions were either round or oval and located in the middle of the cell body or displaced along the periphery of the cell (Figs 4a–c). The intense cytoplasmic staining for α-synuclein made it difficult to clearly delineate inclusion bodies in VM neurons following treatment with UbA although there was localized and concentrated immunoreactivity in perikarya (Fig. 4d).
We demonstrate in primary VM cultures that inhibition of normal proteasomal function with lactacystin induces a concentration-dependent degeneration of dopaminergic neurons and the formation of inclusion bodies that stain positively for α-synuclein and ubiquitin. Inhibition of UPS activity with UbA similarly induces a dose-dependent degeneration of dopamine neurons with an intense accumulation of α-synuclein in focal deposits that resemble inclusion bodies. These findings are consistent with the notion that failure of the UPS plays a role in the etiopathogenesis of PD (McNaught et al. 2001).
Lactacystin, an antibiotic derived from Streptomyces, selectively inhibits the proteolytic activities of 26/20S proteasomes (Craiu et al. 1997; Fenteany and Schreiber 1998; McNaught and Jenner 2001). Lactacystin-mediated impairment of 26/20S proteasomal activity has been shown to cause the accumulation of ubiquitinated proteins and to induce apoptosis in cerebellar granule cells (Canu et al. 2000). It has also been reported that lactacystin induces degeneration with inclusion body formation in cultured PC12 cells (Rideout et al. 2001). In the present study, we demonstrate that lactacystin causes dopaminergic cell death, alterations in protein handling and the formation of proteinaceous inclusion bodies in VM cultures. Interestingly, lactacystin caused swelling of axonal processes of VM neurons with increased staining for ubiquitin and α-synuclein as are found in nigral dopaminergic neurons in PD (Mezey et al. 1998). Taken together, these observations suggest that impairment of 26/20S proteasomal function in SNc neurons could play an important role in the neurodegenerative process occurring in patients with sporadic PD (McNaught and Jenner 2001; McNaught et al. 2002).
A mutation in the gene that encodes for UCH-L1, a de-ubiquitinating enzyme, has been reported in a small number of patients with familial PD, and expression of the mutant protein in Escherichia coli results in a 50% loss of enzymatic activity in comparison to wild-type controls (Leroy et al. 1998). UbA blocks UCH thereby reducing the availability of ubiquitin monomers for degradation of proteins via the UPS (Melandri et al. 1996; Schaeffer and Cohen 1996). We show for the first time that inhibition of UCH with UbA leads to a concentration-dependent degeneration of dopamine neurons in VM cultures. This was associated with focal cytoplasmic deposits which stain intensely for α-synuclein and appear to represent inclusion bodies. Thus, one may speculate that reduced availability of ubiquitin could impair activity of the ubiquitination cycle and inhibit ubiquitin-dependent protein degradation where UCH-L1 defects occur. Indeed, we found a marked reduction in ubiquitin staining in cultures exposed to UbA in this study. Post-mortem brain tissue from PD patients with the UCH-L1 mutation is not yet available for analysis and so it is unknown whether Lewy body formation occurs in this illness. However, in gracile axonal dystrophy in mice, mutations in UCH-L1 cause neuronal death and the presence of inclusion bodies that are immunoreactive for ubiquitin and other proteins (Saigoh et al. 1999). Thus, impairment of UCH-L1 could directly lead to protein handling dysfunction and nigral pathology in this familial form of PD.
Our study suggests that dopaminergic neurons are more susceptible than GABAergic neurons to alterations in viability following inhibition of the UPS with lactacystin, based on the significant differences in dopamine versus GABA uptake. This may relate to the highly oxidative environment within dopaminergic neurons which requires the UPS to clear a high level of damaged proteins that have the potential to be cytotoxic (Hirsch et al. 1997; Davies 2001; Sherman and Goldberg 2001). This may explain why nigral dopaminergic neurons preferentially degenerate in familial forms of PD despite alterations in protein expression and/or UPS function elsewhere in the brain (Shimura et al. 1999).
There is increasing evidence which points to interference with normal UPS function as an important etiopathogenic factor in PD. Gene defects in parkin and UCH-L1 that impair the ubiquitination of proteins necessary for 26/20S proteasomal degradation have been found in a small number of families with PD-like syndromes (Leroy et al. 1998; Shimura et al. 2001). Further, sporadic PD has been shown to be associated with impaired 26/20S proteasomal function (McNaught and Jenner 2001; McNaught et al. 2002). α-Synuclein, which is implicated in both sporadic and familial forms of PD, is degraded by 26/20S proteasomes (Bennett et al. 1999; Goedert 2001; Tofaris et al. 2001). Further, a glycosylated form of α-synuclein has been shown to be a substrate for parkin in SNc dopaminergic neurons in normal human subjects and its ubiquitination and proteasomal degradation are inhibited when parkin activity is impaired in AR-JP (Shimura et al. 2001). Thus, our findings support the concept that defects in the UPS with altered protein handling could play a significant role in the degeneration of dopaminergic neurons, protein accumulation and Lewy body formation that occur in patients with sporadic and familial PD.
KPM and PJ are supported by grants from the Parkinson's Disease Society (UK) and the National Parkinson Foundation (Miami, FL, USA). CM, RJB and JY are supported by a US Army grant (DAMD17-9919557). KPM, CWO and PS are supported by grants from the NIH (NS3377) and the Bachmann–Strauss Dystonia & Parkinson Foundation Inc.