Pharmacological induction of heat-shock proteins alleviates polyglutamine-mediated motor neuron disease

Abstract

Spinal and bulbar muscular atrophy (SBMA) is an adult-onset motor neuron disease caused by the expansion of a trinucleotide CAG repeat encoding the polyglutamine tract in the first exon of the androgen receptor gene (AR). The pathogenic, polyglutamine-expanded AR protein accumulates in the cell nucleus in a ligand-dependent manner and inhibits transcription by interfering with transcriptional factors and coactivators. Heat-shock proteins (HSPs) are stress-induced chaperones that facilitate the refolding and, thus, the degradation of abnormal proteins. Geranylgeranylacetone (GGA), a nontoxic antiulcer drug, has been shown to potently induce HSP expression in various tissues, including the central nervous system. In a cell model of SBMA, GGA increased the levels of Hsp70, Hsp90, and Hsp105 and inhibited cell death and the accumulation of pathogenic AR. Oral administration of GGA also up-regulated the expression of HSPs in the central nervous system of SBMA-transgenic mice and suppressed nuclear accumulation of the pathogenic AR protein, resulting in amelioration of polyglutamine-dependent neuromuscular phenotypes. These observations suggest that, although a high dose appears to be needed for clinical effects, oral GGA administration is a safe and promising therapeutic candidate for polyglutamine-mediated neurodegenerative diseases, including SBMA.

Expansion of a trinucleotide CAG repeat encoding the polyglutamine tract causes inherited neurodegenerative disorders, including spinal and bulbar muscular atrophy (SBMA), Huntington's disease, dentatorubral pallidoluysian atrophy, and several forms of spinocerebellar ataxia (1, 2). All these polyglutamine diseases show progressive and refractory neurological symptoms with selective neuronal cell loss within the susceptive regions of the nervous system. SBMA is a lower motor neuron disease exclusively affecting males and characterized by adult-onset proximal muscle atrophy, weakness, fasciculations, and bulbar involvement (3, 4). The molecular basis of this disease is elongation of a polyglutamine tract in the androgen receptor (AR) protein (5), the toxicity of which is considered a major cause of neurodegeneration in SBMA (6, 7). It has been postulated that pathogenesis in SBMA results from testosterone-dependent accumulation of pathogenic, polyglutamine-expanded AR in the cell nucleus (8, 9). This hypothesis is strongly supported by the observation that intranuclear accumulation of disease-causing protein leads to transcriptional dysregulation, a supposed pathway of neurodegeneration in polyglutamine diseases (10, 11).

Accumulated polyglutamine-containing protein is commonly seen as diffuse nuclear staining or as inclusion bodies, the histopathological hallmarks of polyglutamine diseases. Although inclusion bodies appear to represent a cellular defensive response, diffusely accumulated polyglutamine-containing protein in the nucleus possesses a distinctly toxic property (12-14). Accumulation of pathogenic protein is, thus, a major target of therapeutic strategies for polyglutamine diseases. This view is supported by animal studies showing that hormonal interventions lowering serum testosterone levels successfully prevents nuclear accumulation of pathogenic AR and, thereby, rescue the phenotypes of mouse models of SBMA (8, 15, 16).

Heat-shock proteins (HSPs) are stress-induced molecular chaperones that play a crucial role in maintaining correct folding, assembly, and intracellular transport of proteins (17, 18). Under toxic conditions, HSP synthesis is rapidly up-regulated and nonnative proteins are refolded. There is increasing evidence that HSPs abrogate polyglutamine toxicity by refolding and solubilizing pathogenic proteins (19-21). Overexpression of Hsp70, together with Hsp40, inhibits toxic accumulation of abnormal polyglutamine protein and suppresses cell death in a variety of cellular models of polyglutamine diseases including SBMA (22-24). Hsp70 has also been shown to facilitate proteasomal degradation of abnormal AR protein in a cell culture model of SBMA (25). The salutary effects of Hsp70 have been verified in studies by using mouse models of polyglutamine diseases (26, 27). However, clinical applications based on these studies have certain limitations because they used genetic overexpression of Hsp70.

Geranylgeranylacetone (GGA) is an acyclic isoprenoid compound with a retinoid skeleton that induces HSP synthesis in various tissues including the gastric mucosa, intestine, liver, myocardium, retina, and central nervous system (28-32). Oral administration of GGA rapidly up-regulates HSP expression in response to a variety of stresses, although this effect is weak under nonstress conditions (29). With an extremely low toxicity, this compound has been widely used as an oral antiulcer drug. The aim of the present study is to investigate whether oral GGA induces HSP expression and thereby suppresses polyglutamine toxicity in cell culture and mouse models of SBMA.

Materials and Methods

Adenovirus Vector. Adenovirus vectors were constructed with the BD Adeno-X Expression system according to the manufacturer's protocol (Invitrogen). Briefly, truncated AR constructs containing GFP (24 CAG repeats, 215 N-terminal amino acids of AR or 97 CAG repeats, and 442 N-terminal amino acids of AR) (23) were cloned into the pShuttle vector between the NheI and XbaI sites. pShuttle vectors with truncated AR24 or AR97 were digested with I-CeuI and PI-SceI. After in vitro ligation, recombinant adenovirus vector constructs containing the respective transgenic fragments were used to transfect HEK293 cells, and the vectors were isolated by using the freeze-thaw method. Finally, virus titer was determined by using the BD Adeno-X Rapid Titer kit (Invitrogen).

Cell Culture. The human neuroblastoma cell line (SH-SY5Y, American Type Culture Collection No. CRL-2266) was maintained with DMEM/F12 (Invitrogen) supplemented with 10% FCS. After neural differentiation in differentiation medium (DMEM/F12 supplemented with 5% FCS and 10 μM retinoic acid) for 4 days, SH-SY5Y cells were infected with the recombinant adenovirus vectors containing truncated AR24 or AR97 at a multiplicity of infection of 20 for 1 h and then treated with GGA. At each time point (0, 2, 4, and 6 days) after infection, cells were fixed with 4% paraformaldehyde for 10 min at room temperature, counterstained with propidium iodide (Molecular Probes), and mounted in Gelvatol. A confocal laser scanning microscope (MRC1024, Bio-Rad) and a conventional fluorescent microscope were used to determine the degree of neuronal cell death and the presence of GFP-labeled AR24 or AR97 protein in diffuse nuclear aggregates or in inclusion bodies. Quantitative analyses were made from triplicate determinations. Duplicate slides were graded blindly in two independent trials as described in ref. 23.

Immunocytochemistry. Cells were fixed with 4% paraformaldehyde and incubated with an anti-HSF-1 (HSF-1, heat-shock factor-1) antibody (1:5,000, Stressgen, Victoria, Canada) and anti-rabbit Alexa Fluor 568 antibody (1:1,000, Molecular Probes), then counterstained with Hoechst 33342 (Molecular Probes).

Animals. AR-24Q and AR-97Q mice were generated by using the pCAGGS vector as described in 8 and 33. The mouse rotarod task was performed with an Economex rotarod (Colombus Instruments, Columbus, OH), and cage activity was measured with the AB system (Neuroscience, Tokyo) as described in ref. 34. Each cage contained three mice, which were subjected to a 12-h light/dark cycle. All animal experiments were approved by the Animal Care Committee of Nagoya University Graduate School of Medicine.

GGA Treatment. GGA was a gift from Eisai, Inc. (Tokyo). For treating cultured SH-SY5Y cells, GGA was dissolved in absolute ethanol supplemented with 0.2% α-tocopherol, and ethanol with α-tocopherol alone was used as vehicle. For oral administration to mice, GGA granules were mixed with powdered rodent chow at concentrations of 0.25%, 0.5%, 1%, and 2%. GGA was administrated to mice from 6 weeks of age until the end of the analysis without withdrawal or dose reduction. All mice had unlimited access to food and water. Net consumption of GGA was determined based on the amount of food consumed in each cage.

Western Blotting. SH-SY5Y cells were lysed in CellLytic lysis buffer (Sigma-Aldrich) containing a protease inhibitor mixture (Roche Diagnostics). Mouse tissues were homogenized in buffer containing 50 mM Tris (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, and 1 mM 2-mercaptoethanol with 1 mM PMSF, and 6 μg/ml aprotinine and then centrifuged at 2,500 × g for 15 min. To extract the nuclear and cytoplasmic fractions, mouse tissues were treated with NE-PER nuclear and cytoplasmic extraction reagents (Pierce); cultured cells were lysed in buffer containing 10 mM Tris·HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, and 0.5% Nonidet P-40 and then suspended in buffer containing 50 mM Tris·HCl (pH 6.8), 2% SDS, 6% glycerol, and protease inhibitor mixture (Roche Diagnostics). Equal amounts of protein were separated by 5-20% SDS/PAGE and transferred to Hybond-P membranes (Amersham Pharmacia Biotech). Primary antibodies and concentrations were as follows: AR (H-280, 1:1,000, Santa Cruz Biotechnology) Hsp70 (1:1,000, Stressgen Biotechnologies, Victoria, Canada), Hsc70 (1:5,000, Stressgen Biotechnologies), Hsp25 (1:5,000, Stressgen Biotechnologies), Hsp40 (1:5,000, Stressgen Biotechnologies), Hsp60 (1:5,000, Stressgen Biotechnologies), Grp78 (1:5,000, Stressgen Biotechnologies), Hsp90 (1:1,000, Stressgen Biotechnologies), Hsp105 (1:250, Novocastra Laboratories, Newcastle, U.K.), HSF-1 (1:5,000, Stressgen Biotechnologies), and thioredoxin (1:2,000, Redox Bioscience, Kyoto, Japan). Primary antibody binding was probed with horseradish peroxidase-conjugated secondary antibodies at a dilution of 1:5,000, and bands were detected by using immunoreaction enhance solution (Can Get Signal, Toyobo, Japan) and enhanced chemiluminescence (ECL Plus, Amersham Biosciences, which is now GE Healthcare). An LAS-3000 imaging system (Fuji) was used to produce digital images. Signal intensities of three independent blots were quantified with image gauge software version 4.22 (Fuji) and expressed in arbitrary units. Membranes were reprobed with anti-α-tubulin (1:5,000, Santa Cruz Biotechnology), or anti-histone H3 (1:500, Upstate Biotechnology, Lake Placid, NY) antibodies for normalization.

Immunohistochemistry. Mice anesthetized with ketamine-xylazine were perfused with 4% paraformaldehyde fixative in phosphate buffer (pH 7.4). Tissues were dissected, postfixed in 10% phosphate-buffered formalin, and processed for paraffin embedding. Sections to be stained with anti-polyglutamine antibody 1C2 were treated with formic acid for 5 min at room temperature; those to be incubated with anti-HSF-1 antibody were boiled in 10 mM citrate buffer for 15 min. Primary antibodies and dilutions were as follows: polyglutamine (1:20,000, Chemicon, Temecula, CA), Hsp70 (1:500, Stressgen Biotechnologies), and anti-HSF-1 (1:5000, Stressgen Biotechnologies). Primary antibody binding was probed with a labeled polymer of secondary antibody as part of the Envision+ system containing horseradish peroxidase (DakoCytomation, Gostrup, Denmark). The number of 1C2-positive cells in the spinal cord and muscle were determined as described in ref. 27.

Statistical Analyses. We analyzed data by using the Kaplan-Meier and log-rank test for survival rate, ANOVA with Dunnett's post hoc test for multiple comparisons, and an unpaired t test from statview software version 5 (Hulinks, Tokyo).

Results

GGA Suppresses Polyglutamine Toxicity in Cellular Model of SBMA. To test whether GGA suppresses cellular toxicity induced by expanded polyglutamine, we generated a cultured cell model of SBMA. Adenovirus vector-mediated expression of a truncated AR with 97 CAGs (tAR97Q) resulted in the formation of inclusion bodies in the nucleus and cytoplasm as well as eventual cell death in human neuroblastoma cell line SH-SY5Y, whereas expression of AR containing only 24 CAGs (tAR24Q) showed no such toxicity (Fig. 1 A and B). GGA administration reduced neuronal cell death as detected by propidium iodide staining in the cells expressing tAR97Q, the strongest effect occurring at a dose of 10^-9 M (Fig. 1 B and C). Although GGA failed to decrease the number of the cells containing inclusion bodies, Western blot analysis using an anti-AR N terminus antibody demonstrated that GGA significantly diminished the amount of a high-molecular-weight complex, which likely corresponds to oligomers of tAR97Q (Fig. 1 D and E) (35). Thus, GGA treatment suppresses cytotoxicity caused by accumulation of AR with elongated polyglutamine without inhibiting inclusion body formation.

Fig. 1.

Effects of GGA on polyglutamine toxicity in cultured cell. (A) Punctuated aggregates visualized with GFP (arrowhead) are formed in SHSY-5Y cells infected with an adenovirus vector containing truncated AR with 97 CAGs (tAR97Q-GFP) but not in those bearing tAR24Q. (B) Frequency of cell death 6 days after infection as detected by propidium iodine staining (^*, P < 0.05 compared with untreated tAR97Q cells). (C) Suppression of cell death by GGA. (D) Frequency of cells bearing aggregates. (E) Anti-AR analysis of Western blots of extracts from cells infected with tAR97Q. Error bars indicate SD.

GGA Induces HSPs in Cellular Model of SBMA. To determine whether the GGA-mediated mitigation of polyglutamine toxicity is due to HSP expression, we determined HSP levels in the cell culture model of SBMA after GGA treatment. GGA up-regulated expression of Hsp70, Hsp90, and Hsp105 in the cells with tAR97Q but did not in those with tAR24Q (Fig. 2 A and B). Cycloheximide treatment eliminated GGA-mediated HSP induction and suppression of cell death (Fig. 2 C and D). Expression of Hsc70, a constitutively expressed HSP, was not increased by GGA treatment; no GGA-mediated up-regulation was detected for other HSPs tested, such as Hsp40 and Hsp60, or for thioredoxin, a redox-regulating protein (data not shown). Western blotting (Fig. 2 E and F) and immunocytochemistry (Fig. 2G) revealed that GGA increased the nuclear uptake of hyperphosphorylated HSF-1, a transcription factor regulating HSP expression in the nucleus. Given that activated HSF-1 forms a hyperphosphorylated trimer and translocates into the nucleus, these findings suggest that GGA activates HSF-1, leading to HSP up-regulation.

Fig. 2.

GGA-mediated HSP induction in cultured cell. (A) Anti-HSP analysis of Western blots from cells infected with tAR97Q and treated with GGA. (B) Quantification of the levels of HSPs from tAR97Q-infected cells after 2 days of GGA treatment. (C) Anti-Hsp70 analysis of Western blots from tAR97Q cells treated with or without cycloheximide. (D) Frequency of cell death 2 days after infection as detected by propidium iodine staining (^**, P < 0.05 compared with tAR97Q cells treated with GGA but not with cycloheximide). (E and F) Anti-HSF-1 analysis of Western blots of the cellular nuclear fraction (E) and that of total cell lysate (F). Upper bands correspond to the hyperphosphorylated, active form of HSF-1. (G) Immunocytochemistry for HSF-1. Error bars indicate SD.

GGA Ameliorates Symptomatic Phenotypes of SBMA Mouse. To examine whether pharmacological induction of HSPs alleviates polyglutamine-dependent neuronal dysfunction, oral GGA was administrated to transgenic mice bearing human AR with 97 CAGs (AR-97Q). The actual amount of GGA was constant in each treatment group during the treatment period (see Table 1, which is published as supporting information on the PNAS web site). Oral GGA ameliorated muscle atrophy, gait disturbance, rotarod disability, and body weight loss in AR-97Q mice at both doses of 0.5 and 1% of food, which correspond to ≈600 and 1,200 mg·kg^-1·day^-1, respectively (Fig. 3 A-E and Table 1). The life span of AR-97Q mice treated orally with 0.5 or 1% GGA was significantly extended compared with that of untreated AR-97Q mice. (P < 0.001) (Fig. 3F). GGA failed to alleviate motor dysfunction in AR-97Q mice at a dose of 0.25%. A higher dose of GGA, 2% of food, inhibited body growth and had no beneficial effects on the neurological phenotypes of the AR-97Q mice. Although no hepatic or renal toxicity was demonstrated at other doses, this high dose caused liver enlargement and dysfunction in wild-type and transgenic mice (see Table 2, which is published as supporting information on the PNAS web site).

Fig. 3.

Effect of GGA on neurological phenotypes of AR-97Q mice. (A) Muscle atrophy of 13-week AR-97Q mice. (B) Footprints of 13-week AR-97Q mice. Front paws are shown in red, and hind paws are shown in blue. (C) Stride distance of 13-week AR-97Q mice (n = 3 for each group). (D-F) Rotarod task (D), body weight (E), and cumulative survival (F) of male AR-97Q mice treated with GGA (n = 12 for each group) and untreated counterparts (n = 15). Rotarod performance significantly improved after GGA at doses of 0.5% and 1.0% (P < 0.0001 at both doses compared with nontreated mice at 20 weeks), and body weight increased significantly at a dose of 0.5% (P < 0.005 at 0.5% and P < 0.05 at 1.0%, at 14 weeks). Error bars indicate SD.

GGA Induces HSP Expression in SBMA Mice Through HSF-1 Activation. To examine whether the GGA-induced improvement in the phenotypes of AR-97Q mice was due to induction of HSPs, the expression levels of HSPs were determined. Oral GGA increased the expression of Hsp70, Hsp90, and Hsp105 in the central nervous system and in the skeletal muscle of AR-97Q mice at the doses (0.5 and 1% of food) that were shown to improve symptomatic phenotypes of AR-97Q mice (Fig. 4 A-C and Fig. 6 A and B, which is published as supporting information on the PNAS web site). The induction of HSPs was not clearly observed in the central nervous system until 3 weeks after treatment initiation, but it continued for at least 4 weeks thereafter (see Fig. 7A, which is published as supporting information on the PNAS web site). HSP induction by GGA was undetectable at a dose of 0.25% and was not significant at 2%, in agreement with the lack of therapeutic effect on motor function at these doses. Grp78, Hsp25, Hsp40, Hsp60, and thioredoxin were not induced by GGA administration (see Fig. 7B).

Fig. 4.

GGA-mediated HSP induction in AR-97Q mice. (A) Western blotting for various HSPs in the spinal cord of 14-week, wild-type (Wt), AR-24Q and AR-97Q mice. (B) Western blotting for various HSPs in skeletal muscle of 14-week wild-type, AR-24Q, and AR-97Q mice. (C) Immunohistochemistry for Hsp70 in 14-week wild-type, AR-24Q, and AR-97Q mice. (D) Western blotting of nuclear fraction from spinal cord and that from muscle using anti-HSF-1 antibody. Upper bands correspond to the hyperphosphorylated active form of HSF-1.

To examine whether GGA induced HSP expression through HSF-1 activation, the nuclear translocation of HSF-1 was investigated after GGA treatment. In the untreated state, the level of nuclear accumulated hyperphosphorylated HSF-1 in the central nervous systems of AR-97Q mice was lower than in the wild-type mice. However, when AR-97Q mice received 0.5% oral GGA, nuclear translocation of HSF-1 was higher than in the nontreated mice (Fig. 4D and Fig. 8, which is published as supporting information on the PNAS web site). In contrast, nuclear translocation of HSF-1 in skeletal muscle of untreated AR-97Q mice is already much higher than in wild-type mice, thus perhaps explaining the high degree of Hsp70 induction in AR-97Q mice. After GGA treatment, nuclear translocation of HSF-1 in skeletal muscle of AR-97Q mice was even higher than it was in untreated AR-97Q mice, contributing to a higher induction of Hsp70 (Figs. 4D and 8). These experiments suggest that oral GGA restores activation of HSF-1, which is inhibited by expanded polyglutamine in the affected nervous tissues of AR-97Q mice.

GGA Inhibits Accumulation of Pathogenic AR in Nucleus. With the aim of evaluating the effect of GGA on the nuclear accumulation of abnormal AR, immunohistochemistry with anti-polyglutamine antibody 1C2 was performed on tissues from GGA-administrated and untreated AR-97Q mice. Oral 0.5% GGA decreased the number of 1C2-positive cells in nervous tissues and, to a lesser extent, in muscle (Fig. 5 A and B). Western blot analysis using an antibody against AR demonstrated that 0.5% oral GGA reduced the amount of the high-molecular-weight complex of abnormal AR (Fig. 5C). These findings suggest that oral GGA-mediated HSP induction inhibits nuclear accumulation of abnormal AR, leading to mitigation of polyglutamine-dependent pathogenesis.

Fig. 5.

Effect of GGA on accumulation of abnormal AR. (A) Immunohistochemistry of 14-week wild-type, AR-24Q, and AR-97Q mice using 1C2 antibody. (B) Quantification of 1C2-positive cells in spinal cord and muscle of AR-97Q mice treated with or without GGA. (C) Western blotting for AR of 14-week AR-97Q mice and quantification of the high-molecular-weight, abnormal AR complex indicated by a smear from the top of the gel. Error bars indicate SD.

Discussion

GGA Induces HSP Expression. In the present study, GGA induced Hsp70, Hsp90, and Hsp105 in a cultured cell model of SBMA, leading to abrogation of polyglutamine-mediated cytotoxicity. Furthermore, oral GGA alleviated neuronal dysfunction through induction of HSPs in SBMA mice.

GGA was first introduced as a nontoxic inducer of Hsp70 in rat gastric mucosa (28). Oral GGA has also been reported to induce Hsp70 in the central nervous system as well as in the small intestine, liver, heart, and retina of rodents without any adverse effects (29-32, 36, 37). The present study suggests that the required dose for HSP induction in the SBMA mouse model is ≈600 mg·kg^-1·day^-1, whereas 200 mg·kg^-1·day^-1 induces HSP expression in nonneuronal tissues of rodents under stress (28, 36). Several studies have verified that Hsp70 induction is due to GGA-mediated activation of HSF-1, a transcription factor that regulates expression of Hsp70 (28, 37). In the SBMA mice, GGA facilitated nuclear translocation of HSF-1, leading to induction of Hsp70, in the affected tissues.

GGA showed no adverse effects at the salutary doses used in the present study, although hepatic toxicity was detected at a higher dose. Low toxicity of GGA is advantageous, because continuous administration of GGA at a high dose is required for treating slowly progressive neurodegenerative disease (6, 7). Pharmacological induction of HSP by using GGA thus appears to be an applicable therapeutic strategy for SBMA, although careful attention should be paid to adverse effects during long-term treatment.

HSPs as Therapeutics for Polyglutamine Diseases. In the present study, GGA-mediated HSP induction resulted in inhibiting the accumulation of abnormal AR in the cellular and transgenic mouse models of SBMA. Accumulation of abnormal protein has been considered central to the pathogenesis of polyglutamine diseases, including SBMA. It has been postulated that expanded polyglutamine confers a monomeric protein conformational change from random coil to β-sheet, leading to formation of a polyglutamine oligomer (38, 39). The misfolded monomer and oligomer exercise their toxic effects by interacting with normal cellular proteins. Direct inhibition of polyglutamine oligomerization by Congo red has been demonstrated to exert therapeutic effects in a mouse model of Huntington's disease (40). Whereas oligomerization of causative proteins has been implicated in the pathogenic processes of neurodegeneration in polyglutamine diseases, the formation of inclusion bodies or mature amyloid fibrils appears to possess cytoprotective properties (13, 41). Based on this hypothesis, HSPs have been drawing a great deal of attention because they inhibit oligomer assembly and thereby mitigate polyglutamine toxicity (20, 21, 38). This view is supported by the fact that overexpression of Hsp70 attenuates the accumulation of polyglutamine-containing protein, resulting in amelioration of neurodegeneration in animal models of spinocerebellar ataxias or SBMA (26, 27).

GGA treatment significantly suppressed nuclear accumulation of abnormal AR in SBMA mice but did not inhibit inclusion body formation in cultured cells. This inconsistency does not necessarily deny a beneficial effect of GGA on polyglutamine aggregation, because it can be explained by several lines of evidence: (i) HSPs facilitate amyloid fibril formation by stabilizing the conformation of abnormal polyglutamine-expanded protein (42), and (ii) HSPs biochemically alter the structure of inclusion bodies (43, 44).

Hsp70 overexpression, however, fails to alleviate neurodegeneration or aggregate formation in the R6/2 mouse model of Huntington's disease (45, 46). This discord appears to indicate that higher levels of Hsp70 or the concomitant induction of other HSPs is required to alleviate Huntington's disease pathology. In addition to Hsp70, various molecular chaperones that colocalize with aggregates have also been shown to suppress polyglutamine toxicity: Hsp40-associated Hsp70 (23, 43), Hsp90, and Hsp105 (47, 48). Oral GGA induced Hsp90 and Hsp105 in the mouse model of SBMA and such diverse HSP up-regulation might contribute to the beneficial effects of GGA in the SBMA mice.

HSP in Pathogenesis of polyQ Diseases. Not only are HSPs considered potent suppressors of polyglutamine toxicity, but they are also implicated in the pathogenesis of neurodegeneration (20). There are several lines of evidence that polyglutamine elongation weakens the protective responses for coping with cellular stress. Truncated AR with expanded polyglutamine delays the induction of Hsp70 after heat shock (49). In the SBMA mice we examined, the level of Hsp70 in spinal cord was decreased before the onset of motor dysfunction. A similar finding has also been reported in the R6/2 mouse model of Huntington's disease (46). In our SBMA mice, abnormal, polyglutamine-expanded AR seems to inhibit nuclear translocation of HSF-1 in the central nervous system, leading to a decrease in the level of Hsp70. In mammalian cells, induction of Hsp70 requires activation and nuclear localization of HSF-1. In the presence of nonnative protein, HSF-1 is derepressed, forming a trimer that translocates into the nucleus and binds to heat-shock elements within the gene encoding Hsp70 (50). In cellular models, this stress-induced nuclear accumulation of HSF-1 has been designated nuclear granules (51). Aggregates of abnormal ataxin-1, the causative protein in spinocerebellar ataxia 1, have been shown to hinder induction of nuclear granules in response to heat shock (52). Therefore, failure of HSF-1 activation appears to enhance polyglutamine toxicity. In this context, it is intriguing that inhibition of the nuclear accumulation of HSF-1 was detected in spinal cord but not in muscle of SBMA transgenic mice. Given that the threshold for HSP induction is relatively high in motor neurons (53), motor-neuron-specific inactivation of HSP transcription might partially explain why the central nervous system is selectively affected in polyglutamine diseases including SBMA.

HSP-Based Therapy for Neurodegeneration. Both genetic and pharmacological manipulations of HSPs have been demonstrated to mitigate the pathogenesis of neurodegeneration (54-57). These observations suggest that GGA-mediated HSP induction may provide a therapeutic strategy for diverse neurodegenerative disorders, because these diseases share common pathogenic mechanisms such as abnormal protein aggregation, disruption of the ubiquitin-proteasome system and activation of the apoptotic pathway.

In summary, our observations indicate that GGA is a safe and promising therapeutic approach for treating many devastating neurodegenerative diseases, including SBMA.