Address correspondence and reprint requests to P. Shashidharan, Department of Neurology, Box 1137, Mount Sinai School of Medicine, New York, NY 10029, USA. E-mail: email@example.com
Childhood-onset dystonia is an autosomal dominant movement disorder associated with a three base pair (GAG) deletion mutation in the DYT1 gene. This gene encodes a novel ATP-binding protein called torsinA, which in the central nervous system is expressed exclusively in neurons. Neither the function of torsinA nor its role in the pathophysiology of DYT1 dystonia is known. In order to better understand the cellular functions of torsinA, we established PC12 cell lines overexpressing wild-type or mutant torsinA and subjected them to various conditions deleterious to cell survival. Treatment of control PC12 cells with an inhibitor of proteasomal activity, an oxidizing agent, or trophic withdrawal, resulted in cell death, whereas PC12 cells that overexpressed torsinA were significantly protected against each of these treatments. Overexpression of mutant torsinA failed to protect cells against trophic withdrawal. These results suggest that torsinA may play a protective role in neurons against a variety of cellular insults.
One form of autosomal dominant dystonia is due to a mutation of the DYT1 gene, which is localized on chromosome 9q34 (Kramer et al. 1990; Ozelius et al. 1992), and encodes for the protein torsinA. In most pedigrees with the clinical syndrome of childhood-onset dystonia, a GAG deletion, corresponding to loss of a single glutamic acid in torsinA at residue 302/303, has been found (Ozelius et al. 1997). This alteration in amino acid sequence of the protein seems necessary, but alone is not sufficient for the development of dystonic symptoms, as not all mutant gene carriers develop dystonia. The reason for the lack of complete penetrance remains to be elucidated. Within the central nervous system, torsinA expression is restricted to neurons (Shashidharan et al. 2000a; Konakova and Pulst 2001; Konakova et al. 2001; Walker et al. 2001), but its neuronal function remains elusive. However, based on its homology to heat shock proteins (HSPs) and to the AAA+ superfamily of chaperone-like proteins (Neuwald et al. 1999), an HSP-like function has been proposed (Ozelius et al. 1997). HSPs have been shown to have protective effects against neurodegeneration in a number of in vivo models (Auluck et al. 2002; Dedeoglu et al. 2002), and a similar neuroprotective function has been suggested for torsinA, as overexpression has been shown to reduce protein aggregation (McLean et al. 2002; Caldwell et al. 2003). Indeed the C-terminus of torsinA, which includes the glutamic acid deletion in DYT1 dystonia, is predicted to be involved in formation of a homo-oligomeric ring structure and in interactions with other proteins (Neuwald et al. 1999). Therefore, it is reasonable to consider that DYT1 mutation may interfere with its function.
In order to investigate the potential neuroprotective functions of torsinA, we tested the effect of overexpression of torsinA and mutant torsinA in PC12 cells on survival against a variety of cellular insults.
Materials and methods
The fluorescein-based in situ cell death detection kit was from Roche (Indianapolis, IN, USA). Cell culture media and biochemicals were obtained from Gibco-Life Technologies (Gaithersburg, MD, USA) and Sigma-Aldrich Corp (St Louis, MO, USA). Horse serum was obtained from Gemini (Calabasas, CA, USA) and fetal bovine serum (FBS) from Collaborative Biomedical Products (Bedford, MA, USA). Dulbecco's modified Eagle's medium (DMEM) and trypsin were from Mediatech (Herndon, VA, USA). ECL western blot chemiluminescence kit was from Amersham-Pharmacia (Piscataway NJ, USA), MG-132 was from Calbiochem (San Diego, CA, USA). Polyvinylidene difluoride (PVDF)-membrane was from Bio-Rad (Hercules, CA, USA). The phosphorimager used was Model Storm 860 (Molecular dynamics, Sunnyvale, CA, USA).
Cell cultures. PC12 cells were propagated in DMEM with 10% horse serum, 5% FBS, 2 mm l-glutamine, 50 U/mL penicillin, and 50 µg/mL streptomycin (PC12 media). Two hundred thousand cells were plated in 24-well plates and used for performing various treatments and analyses. For TdT-mediated dUTP nick end labeling (TUNEL) studies, cells were plated on coverslips coated with polylysine.
Establishment of PC12 cells stably overexpressing torsinA. The coding regions of the cloned human wild-type torsinA and ΔE-torsinA cDNAs were subcloned into expression vector pcDNA3. PC12 cells were plated at a density of 3 × 105 cells/35 mm dish in DMEM containing 5% FBS and 10% horse serum. Cells were transfected with the plasmid DNA as described previously (Shashidharan et al. 1999). Stable cell lines expressing green fluorescent protein (GFP) were established as described previously (Shashidharan et al. 1999) and served as controls. Both GFP and torsinA expressing PC12 cells were derived from the same clonal background. The transfected cells were treated with G418, and cells resistant to the drug were isolated and passaged to establish stable PC12 cell lines. Several clonal cell lines were isolated, and levels of wild-type and ΔE-torsinA expression determined by western blot analysis.
Treatment protocol. Treatment of cultures was performed with a complete change of the feeding medium. For proteasomal inhibition,10 µm MG-132 was added to the medium. To induce oxidative stress, hydrogen peroxide (H2O2, 0.2 mm) was added to the medium, cells were incubated for 6 h and then in fresh complete medium as described previously (Shashidharan et al. 1999). To induce trophic withdrawal, cultures were grown in serum-free medium for 6 or 24 h. Eight wells of cultures were used per each concentration and/or treatment and the experiments were repeated at least three times.
MTT assay. Cell viability was determined by MTT reduction assay, as described previously (Han et al. 1996). In brief, 50 µL MTT (5 mg/mL) was added to each cell culture well containing 0.5 mL medium. After 1-h incubation at 37°C, the medium was carefully removed and the formazan crystals dissolved in 1 mL isopropyl alcohol by gentle shaking of the plate. Absorbance was determined at 570 nm in a microplate reader (Spectramax 250, Molecular Devices Corporation, Sunnyvale CA, USA).
TUNEL reaction. The fluorescein-based in situ cell death detection kit was used to perform the TUNEL reaction according to the manufacturer's instructions. Briefly, after the various treatments, the cells on coverslips were fixed with 4% paraformaldehyde for 1 h, washed with phosphate-buffered saline (PBS), permeabilized with 0.1% triton X-100 on ice for 2 min, rinsed with PBS, and coverslips incubated with 50 µL TUNEL reaction mixture (enzyme + nucleotides + fluorescein-dUTP) in a humidified atmosphere at 37°C for 1 h. To demonstrate nuclear morphology, cells were counterstained with Hoechst stain (33258). Following these incubations, coverslips were rinsed with PBS three times and analyzed by fluorescence microscopy. The fluorescent photomicrographs were processed using Adobe Photoshop. Cells undergoing cell death were quantified by counting TUNEL-positive nuclei by an investigator (blinded to the treatment). Values shown are means of eight fields using a 0.25-mm2 grid and data are represented per cm2.
Western blotting. Western blotting was performed using cell homogenates obtained from control cells and torsinA-transfected PC12 cells using rabbit polyclonal antibody to human torsinA generated against a peptide derived from the C-terminus of torsinA as described previously (Shashidharan et al. 2000a). The antibody has been previously characterized and shown to immunoreact with human and rodent torsinA, and human ΔE-torsinA (Shashidharan et al. 2000a; Walker et al. 2001). Protein concentrations were determined using a protein assay kit (Sigma Diagnostics, St. Louis, MO, USA). Twenty micrograms of protein homogenate was resolved on SDS-PAGE and electroblotted onto PVDF membrane using 50 mm Tris-HCl buffer (pH 8.4). The blot was incubated with antibody to torsinA for 1 h and detected using ECL western blot chemiluminescence kit. Relative levels of protein were quantified using Phosphorimager Model #Storm 860 (Molecular Dynamics, Sunnyvale, CA, USA).
Values were expressed as means ± SEM. Significance of differences between two groups was determined by two-tailed Student's t-test. For multiple comparisons, one-way anova followed by the Tukey–Kramer test was used.
Establishment of PC12 cells stably overexpressing torsinA and ΔE-torsinA
Stable PC12 cells were established by transfection with torsinA-pcDNA3 or ΔE-torsinA-pcDNA3 constructs. Several stable cell lines were established expressing varying levels of torsinA and ΔE-torsinA (Fig. 1). The expression of endogenous torsinA in PC12 cells is very low and is visible as a faint band (Fig. 1, lane C).
PC12 cell lines expressing ΔE-torsinA were difficult to maintain, as expression of the mutant protein was not found after a few passages, therefore the studies performed with ΔE-torsinA cells were limited to the effects of serum withdrawal.
Survival of torsinA-expressing clones
Using the MTT assay, we studied the effect of torsinA expression upon cell survival under various toxic conditions. Clonal cell lines expressing wild-type torsinA were treated with the proteasomal inhibitor MG-132 (10 µm). Clone 4 showed significant protection against proteasomal inhibition as compared with GFP-transfected cells, which were used as controls (Fig. 2a). When torsinA-transfected cells were exposed to H2O2 a more marked protective effect was seen, statistically significant in all three clonal lines (Fig. 2b). A similar effect was seen in the setting of serum deprivation (Fig. 2c). As clone 4 demonstrated the most robust survival under various toxic conditions, this line was used for further studies.
Effect of ΔE-torsinA expression in the setting of serum deprivation
We compared the survival of control PC12 cells with cells overexpressing torsinA (clone 4) or ΔE-torsinA (clone 7) following serum withdrawal. After 24 h of trophic withdrawal, 76% of control PC12 cells died, compared with cells overexpressing torsinA, of which only 25% died (Fig. 3). In contrast, 64% of cells overexpressing ΔE-torsinA died. The protection afforded by wild-type torsinA to cells upon trophic withdrawal was highly significant (p < 0.001) and showed a significant difference from controls and cells overexpressing ΔE-torsinA (p < 0.001). Similar results were obtained by performing TUNEL experiments, showing an absence of TUNEL-positive nuclei in cell lines overexpressing torsinA, in contrast to numerous TUNEL positive nuclei in control cells or cells overexpressing ΔE-torsinA (Fig. 4).
Quantitative determination of cell death using TUNEL
Representative fluorescent photomicrographs of TUNEL studies of control PC12 cells and those overexpressing torsinA or ΔE-torsinA, under various conditions of cellular insults, are shown in Figs 4 and 5. The number of TUNEL-positive cells in cultures subjected to proteasome inhibition, oxidative stress, and serum deprivation, respectively, were counted by an investigator blinded to the treatment (Fig. 6). The number of TUNEL-positive nuclei was significantly reduced in cells overexpressing torsinA, demonstrating that torsinA renders PC12 cells resistant to these cellular insults as compared with control cells. Cells overexpressing ΔE-torsinA were not protected against the effects of serum deprivation.
A number of recent studies have suggested an HSP-like function of torsinA. McLean et al. (2002) showed that overexpression of torsinA suppressed α-synuclein aggregation in a human H4 neuroglioma cell culture model. Similarly, in a Caenorhabditis elegans model utilizing polyglutamine repeat-induced protein aggregation, Caldwell et al. (2003) found that overexpression of both human and C. elegans torsin proteins resulted in a dramatic decrease in protein aggregations. In PC12 cultures, Hewett et al. (2003) demonstrated that endogenous torsinA is sensitive to H2O2, with immunoreactivity shifting from a codistribution with the ER marker PDI throughout the cytoplasm to increased intensity in the perinuclear region and into protrusions from the cell surface. They also observed a slight change in mobility after H2O2 treatment, indicating a rapid covalent modification of torsinA in response to oxidative stress (Hewett et al. 2003). In addition, torsinA is seen in Lewy bodies in Parkinson's disease (Shashidharan et al. 2000b; Sharma et al. 2001) and in inclusion bodies in trinucleotide repeat diseases (Walker et al. 2003), as are a number of HSPs, chaperones and other proteins involved in protein degradation (Sherman and Goldberg 2001; McNaught et al. 2002; Schmidt et al. 2002).
We demonstrate that overexpression of torsinA significantly protects PC12 cells from cell death induced by the reversible proteasomal inhibitor MG-132, oxidant stress, and trophic withdrawal. When cells overexpressed mutant torsinA in the setting of trophic withdrawal, this protective effect was absent. As protein expression was lost after a few passages, studies of ΔE-torsinA were limited to testing in this condition only.
Overexpression of other HSPs has been shown to protect cells against similar stresses (Sherman and Goldberg 2001). Overexpression of HDJ-1, a member of the HSP 40 family, renders SH-SY5Y cells resistant to the cytotoxicity associated with oxidative stress and proteasomal inhibitors, with preservation of mitochondrial function and proteasomal activity following oxidative injury (Ding and Keller 2001). Based on these results the authors suggest a possible role for proteasome inhibition in the toxicity of oxidative stress, and further that HSPs may confer resistance to oxidative stress by attenuating the toxicity of proteasome inhibition. TorsinA, when overexpressed, plays a protective function in cells against the cellular insults described in this paper, although the exact mechanism by which it achieves this effect remains to be elucidated.
It is reasonable to consider that the loss of protection with mutant torsin A may in some way relate to the development of DYT1 dystonia. In humans with DYT1 dystonia, ΔE-torsinA has not yet been demonstrated to cause neurodegeneration or protein aggregation (Walker et al. 2002; Rostasy et al. 2003), but our findings suggest that this should be more carefully explored particularly as only a small number of brains have been studied. Dopamine neurons in the substantia nigra pars compacta, which have been implicated in the pathophysiology of dystonia (Augood et al. 2002), contain a significant amount of torsinA (Augood et al. 1998) and in DYT1 dystonia have normal intracellular distribution (Walker et al. 2002; Rostasy et al. 2003) and increased size (Rostasy et al. 2003). We postulate that the ΔE-torsinA mutation interferes with the function of normal torsinA, possibly in response to environmental stress (Jankovic and Van der Linden 1988; Fletcher et al. 1991; Frucht et al. 1999; Schrag et al. 1999; Frucht et al. 2000). It is possible, however, that use of other techniques such as electron microscopy or novel antibodies, for example, to ubiquitin–protein conjugates as have been used in Parkinson's disease (McNaught et al. 2002), may yet reveal structural markers of neuronal dysfunction and protein aggregation in dystonia, as they have in Parkinson's disease. Indeed over expression of ΔE-torsinA has been shown to cause membrane inclusions and aggregation of mutant protein in neural cell cultures (Hewett et al. 2000). To date attention has been focused on dopamine neurons, and there is evidence suggesting that torsinA plays a role in dopaminergic neurotransmission (Augood et al. 2002). However it is also possible that changes primarily occur in interneurons of extranigral regions which play a role in modulating dopaminergic activity such as the striatum or centromedian nucleus of the thalamus (Saka et al. 2002). The different neurochemical classes of interneurons each make up only a small percentage of these neuronal populations (Yelnik et al. 1991) and might not be detected on routine neuropathological examination.
Our studies support a protective role for torsinA, possibly related to its HSP-like properties. Whether this is related to the pathogenesis of dystonia requires further investigation.
The work was supported by the Bachmann-Strauss Dystonia & Parkinson Foundation, Inc. and NIH grant NS43038 (PS). We thank Dr Catherine Mytilineou for her critical reading and many suggestions in preparation of this manuscript.