Therapeutic approaches to spinal and bulbar muscular atrophy

Abstract

Spinal and bulbar muscular atrophy is a hereditary motor neuron disease caused by trinucleotide repeat expansion in the androgen receptor gene. The disease mechanism probably involves a toxic gain of function in the mutant protein, because other mutations that cause a loss of androgen receptor function result in a different phenotype, and the mutant protein is toxic in mouse models. In these models, the toxicity is ligand-dependent and associated with protein aggregation, as well as altered transcriptional regulation, axonal transport, and mitochondrial function. Various therapeutic approaches have shown efficacy in the mouse models, including androgen reduction, heat shock protein 90 (HSP90) inhibition, and insulin-like growth factor (IGF)-1 over-expression. Clinical trials of androgen reducing agents have had mixed results, with indications of efficacy but no proof of clinically meaningful benefit to date. These clinical studies have established outcome measures for future trials of other agents that have been beneficial in the animal studies.

Spinal and bulbar muscular atrophy (SBMA), or Kennedy’s disease, was the first neurodegenerative disease found to be caused by expansion of a trinucleotide (CAG) repeat encoding a polyglutamine tract in the mutant protein. The genetic defect places SBMA in a family of polyglutamine expansion diseases that also includes Huntington’s disease and at least six spinocerebellar ataxias. SBMA is an X-linked, adult-onset, lower motor neuron disease, and the causative mutation is in the androgen receptor (AR) gene [1]. The repeat length is polymorphic, with a range of 11–35 CAGs in normal individuals and expansion to 37–65 CAGs in SBMA. The severity of the disease increases and the age of onset decreases with increasing repeat length [1]. Pathologically, SBMA patients have a loss of lower (primary) motor neurons in the brainstem and spinal cord and also a subclinical loss of sensory neurons in the dorsal root ganglia [2]. Muscle cramps are often the initial complaint; other common symptoms include arm and leg weakness, tremor, and difficulty with speech and swallowing [3]. SBMA patients also often have signs of androgen insensitivity, including breast enlargement and reduced fertility [3].

Since the identification of the causative mutation, much has been learned about the molecular mechanisms of SBMA through studies of transgenic animal models [4–10]. Preclinical therapeutic studies have indicated opportunities for intervention to prevent or slow the disease progression. In this review, we discuss the proposed mechanisms and treatment strategies based on the evidence from these studies.

AR and toxic gain of function

Among the polyglutamine diseases, SBMA is unique in that ligand binding alters subcellular localization and toxicity of the mutant protein. The AR is a nuclear receptor that binds to testosterone and dihydrotestosterone (DHT). It is a heat shock protein 90 (HSP90) client, and it is normally sequestered in the cytoplasm in a complex with HSP90 and other chaperones. Upon ligand binding, the AR is actively translocated to the nucleus and functions as a transcription factor [11]. The polyglutamine repeat expansion is in the amino terminal transactivation domain of the AR. The mutant protein has a propensity to aggregate and form nuclear inclusions in motor neurons and other cells. Neuronal toxicity may result from aberrant interactions with crucial nuclear proteins, resulting in transcriptional dysregulation [11–13].

Anti-androgen treatment

Studies in cell culture and animal models have provided evidence for ligand-dependent toxicity in SBMA [5, 13–15]. Exogenous androgens are required to elicit disease manifestations in transgenic flies, and in transgenic mice, only males (which have higher androgen levels than females) fully manifest the degenerative phenotype. Androgen administration elicits the phenotype in the female mice, and anti-androgen treatment alleviates the disease manifestations in males. The effects of anti-androgens such as flutamide and leuprorelin, a lutenizing hormone-releasing hormone (LHRH) agonist that reduces testosterone levels, have been studied in transgenic mice. Leuprorelin treatment and surgical castration mitigate the disease phenotype in the mice, while flutamide does not [5, 13]. A report that two women homozygous for the SBMA mutation have milder manifestations than their affected male relatives [16] indicates that the toxicity of the mutant AR is ligand-dependent and might be mitigated by anti-androgen treatment in humans, as well as in mice.

In 2009, Gen Sobue’s research team published the results of an interventional trial with the androgen-reducing agent leuprorelin [17]. In the 48-week placebo-controlled portion of this trial, there were signs that the treatment had an effect on the nuclear accumulation of the mutant protein and slightly improved swallow function (cricopharyngeal opening duration), although it was not clear whether the effects were clinically meaningful. A larger, multi-center trial of leuprorelin is currently in progress in Japan.

At the U.S. National Institute of Neurological Disorders and Stroke (NINDS), a clinical trial was undertaken with dutasteride, a 5-α-reductase inhibitor that blocks the conversion of testosterone to DHT. The study was based in part on the rationale that dutasteride might reduce the toxicity of the mutant androgen receptor in motor neurons but maintain the anabolic benefits of testosterone in the muscle. This study was a two-year, double blind, placebo-controlled trial. Physical, neuropsychological, quality of life, and biochemical measures were assessed in 50 SBMA subjects randomized to receive either dutasteride or placebo. There was no significant difference in the primary outcome measure, the change in strength as indicated by quantitative muscle assessment (QMA). Of the secondary outcome measures, physical quality of life (as measured by the physical component summary of the Short-Form Health Survey SF-36v2 questionnaire) showed a benefit with dutasteride treatment, but this was balanced by a negative effect on the SF-36v2 mental component summary. Subjects taking dutasteride also had fewer falls. In the placebo group, the QMA showed an average decrease in muscle strength of about 2% per year, indicating that SBMA patients might need to be studied over a longer period to show a significant effect on disease progression (see http://clinicaltrials.gov/ct2/show/NCT00303446?term=kennedy+disease&rank=1). A cross-sectional study of 57 SBMA subjects recruited for the dutasteride study showed that higher rather than lower levels of testosterone are associated with increased muscle strength and function [3]. This indicates that there might be a fine balance between the anabolic benefits of androgens and the deleterious ligand-activated toxicity of the mutant AR. At present, neither leuprorelin nor dutasteride has been proven to be effective treatment for SBMA, although the recent studies show indications of benefit, and further prospective, randomized clinical trials are needed. The relative lack of efficacy of anti-androgen treatment points to the need to identify other therapeutic targets for this disease.

Modulators of chaperones

One of the pathological hallmarks of SBMA is the presence of intranuclear inclusions in motor neurons of brainstem and spinal cord in SBMA patients [6, 18]. These inclusions are likely protective rather than toxic, but the tendency of the mutant protein to aggregate correlates with its toxicity and might be involved in the disease mechanism [11, 19]. The toxicity of the mutant protein is mitigated by heat shock proteins (HSPs), which act as chaperones in regulating protein folding and targeting misfolded proteins for degradation. Cell culture and animal studies have addressed the role of chaperones, primarily HSPs 40, 70 and 90 in mutant protein aggregation and cell death in SBMA [20–22]. The carboxy-terminus of AR has affinity for HSP90, a chaperone that is important for the function and stability of AR. HSP90 inhibitors reduce ligand binding and induce degradation of AR. Two geldanamycin derivatives, 17-AAG and 17-DMAG, which are HSP90 inhibitors in development for malignancy, have been studied in SBMA animal models. Research has shown that 17-AAG inhibits intranuclear aggregate formation by mutant AR and leads to improved motor performance in the AR97Q transgenic mouse model of SBMA [22].

Impairment of the ubiquitin proteasome system (UPS) by aggregated mutant protein has been suggested by in vitro and in vivo studies [20, 23, 24]. Normal function of this protein degradation pathway is important for the action of HSP90 inhibitors. However, experiments with ubiquitin reporters have shown no global impairment of the UPS in Huntington’s disease [25]. Using GFP-based ubiquitin reporter mice, Sobue and colleagues tested whether the UPS in the AR97Q transgenics is preserved and whether 17-DMAG is effective in this mouse model of SBMA [9]. They found that the UPS is indeed functional and showed an increase in its activity as the mice developed severe manifestations. The UPS activity was well-preserved in the spinal cord and even increased in the skeletal muscle of the AR97Q SBMA mice. In both cell culture and transgenic mice, exposure to 17-DMAG led to increased degradation of mutant AR, with an induction of the chaperones HSP70 and HSP40. 17-DMAG improved motor function in the transgenics, increased survival in a dose dependent manner, reduced mutant polyglutamine staining in affected tissues, and improved the skeletal muscle pathology.

These findings point to 17-DMAG as a candidate therapy. Both 17-AAG and 17-DMAG have been used in cancer trials, but neither has been given for more than 6 months to date [26]. Other HSP90 inhibitors are currently under development as cancer treatment. If any of these turns out to be well tolerated and safe for long-term use, then it might be worth pursuing in preclinical and clinical studies of SBMA.

Neurotrophic factors

A lack of target-derived trophic support might contribute to motor neuron dysfunction in SBMA. This could arise through decreased retrograde axonal transport. Murine studies have suggested that the mutant AR causes accumulation of neurofilament proteins and synaptic proteins, and reversibly disrupts the p150 subunit of dynactin, an important motor protein for retrograde transport [27]. In addition, fast axonal transport is altered in a kinase-dependent manner in the presence of mutant AR [28]. A recent study of insulin-like growth factor 1 (IGF-1) provided evidence that it might have therapeutic potential, whether or not its trophic activity is altered in SBMA [29]. In this study, IGF-1 over-expression reduced muscle atrophy and increased survival in the SBMA transgenic mice. AR97Q transgenic mice were crossed with mice over-expressing a non-circulating, muscle specific isoform of IGF-1 (mIGF-1). The double transgenics had delayed disease onset, improved muscle strength and motor activity, and prolonged lifespan. IGF-1 attenuated the pathological signs of myopathy and denervation in the muscle. In addition, the mice over-expressing IGF-1 in muscle had better motor neuron survival in the spinal cord. These findings are interesting when considering evidence for myopathic involvement in patients and the knock-in mice [7]. This raises a question regarding the relative contributions of motor neurons and muscle in SBMA pathogenesis. Both in vitro and in vivo studies suggest that IGF-1 acts via activation of the pro-survival Akt enzyme, reducing ligand binding and transactivation by the AR [29, 30]. This provides a potential target for therapeutic development. Although IGF-1 has been tested in clinical trials for amyotrophic lateral sclerosis and myotonic muscular dystrophy without benefit, SBMA might be a more appropriate disease for this approach to treatment. Further mouse studies with injected IGF-1 are in progress, and if these confirm the benefit, then a clinical trial would likely follow. In addition, an earlier study suggested that expression of the murine variant of vascular endothelial growth factor (VEGF-164) is altered in transgenic mice expressing mutant AR with 100 CAGs (YAC AR100Q) [31]. Motor neuron toxicity was rescued with over-expression of VEGF in cell culture, indicating that VEGF might be an important neurotrophic survival factor in SBMA. In certain cell types, IGF-1 is considered an effective inducer of VEGF. In light of the myopathic component of SBMA, it might be appropriate to use intramuscular delivery of growth factors to produce neuroprotection.

Other candidates: evidence from cell culture and animal studies

Transcriptional dysregulation probably plays a role in the pathogenesis of SBMA, and this might be mediated in part by the depletion of histone acetyltransferases [32, 33]. In a cellular model, over-expression of CREB-binding protein and treatment with histone deacetylase (HDAC) inhibitors, such as sodium butyrate, trichostatin A, and vorinostat, mitigated cell death [34]. Furthermore, pre-symptomatic administration of sodium butyrate in AR97Q transgenic mice resulted in increased survival and motor performance. Although it also delayed neuropathological features, the HDAC inhibition did not have an effect on nuclear localization and inclusion formation by mutant AR protein [33]. HDAC inhibitors remain worthy of consideration as therapeutic candidates for SBMA, although the benefit might be limited by the toxicity of these compounds. Better efficacy might be achievable with selective HDAC inhibition. A case in point is regulation of the gene that encodes C terminus of Hsc70-interacting protein (CHIP). Over-expression of CHIP, a U-box type E3 ubiquitin ligase, decreased nuclear accumulation of the mutant AR protein and ameliorated the motor impairment in SBMA transgenic mice [24]. Recent studies in models of amyotrophic lateral sclerosis (ALS) and in striated muscle suggest that CHIP can selectively activate the autophagic removal of aggregated proteins [35, 36]. Although this mechanism has not yet been observed in SBMA models, the possible involvement of CHIP in autophagy in SBMA is worth investigating. It might be useful to identify a small molecule that increases CHIP expression as a potential therapeutic approach.

In light of the mixed effects of androgen reduction in recent clinical trials, other approaches to treatment are worth considering. In 2007, Chawnshang Chang’s research team reported benefit with the synthetic curcumin-related compound, ASC-J9, in the AR97Q transgenic mice [10]. They showed that this compound promoted AR degradation and delayed the onset and progression of motor impairment. This treatment had negligible influence on circulating testosterone, and the mice had normal sexual activity and improved fertility. ASC-J9 is also less toxic in animal studies than 17-AAG or DMAG. A biotech company, AndroScience, is currently developing the drug for use in humans, and safety testing and confirmatory mouse efficacy studies are planned. If ASC-J9 is confirmed to be safe and effective in mice, then a clinical trial might follow.

The toxicity of mutant AR is greater with nuclear localization. Experiments in neuronal cells and animal models showed that deleting residues in the nuclear localization domain reduced inclusion formation and toxicity even in the presence of the ligand [37]. Furthermore, pharmacological activation of autophagy completely rescued the effects of mutant AR in the nucleus. Although further research is needed to define the role of autophagy in SBMA, the results suggest that enhancing cytoplasmic degradation might be useful.

Although the AR is a nuclear transcription factor, the mutant protein might have effects on cytoplasmic organelles. Mitochondrial dysfunction has been described in SBMA, as in other polyglutamine diseases, and might lead to cell death through oxidative damage and caspase activation. Our recent findings in a motor neuron-derived cell line and affected tissues from the knock-in mouse model show that mutant AR causes mitochondrial abnormalities [38]. We found a decrease in mitochondrial number, loss of mitochondrial membrane potential, and increase in reactive oxygen species. The mutant AR also altered peroxisome proliferator-activated receptor γ coactivator-1 (PGC-1β) mRNA, a key regulator of mitochondrial biogenesis and function, and its target genes, including mitochondrial transcription factor A (TFAM). The mutant protein also decreased the transcript levels of mitochondrial and cellular antioxidant genes in both cell culture and mouse models. Interestingly, we found that mutant AR localizes to the mitochondria in cultured cells in addition to down-regulating gene expression. These results then suggest that mutant AR might act directly on mitochondria or indirectly through transcriptional dysregulation of genes encoding mitochondrial proteins. The mitochondrial abnormalities and cell toxicity can be attenuated with mitochondrial modulators such as cyclosporine A (which blocks the opening of the mitochondrial membrane transition pore) and the antioxidants coenzyme Q and idebenone. Idebenone has been used with mixed results in clinical trials in Friedreich’s ataxia, a disorder caused by mutation of frataxin, a mitochondrial protein [39–41]. Nevertheless, the results from the in vitro study suggest that mitochondrial modulators and antioxidants such as these might be worth evaluating as potential therapy for SBMA.

Finally, exercise has been found to have beneficial effects in studies of ALS and SBMA subjects. A larger scale clinical trial in ALS patients is planned. A small study published recently from a group in Copenhagen had mixed results in SBMA patients who did regular cycling exercise for 12 weeks. A longer-term trial or a trial with a different type of exercise or different outcome measures might show clearer evidence of benefit [42]. A systematic study of endurance and/or resistance exercise, modeled after the ALS clinical trial, is currently under consideration at NINDS.

Conclusions

In summary, there are currently several approaches to treatment for SBMA that have shown efficacy in transgenic mouse models and have a reasonable likelihood of benefit in clinical trials. Because these approaches have different targets, any successful treatments might be synergistic.

The ligand-dependent toxicity of the mutant AR is most directly targeted by reducing androgen levels, but because androgens probably have beneficial as well as deleterious effects on neuromuscular function in SBMA, the benefits of this approach are likely to be limited. Targeting the heat shock response through HSP90 inhibition is currently too toxic for such a slowly progressive and generally non-fatal disease. The mechanism of action of ASC-J9 warrants further investigation, as evidence from animal studies indicates that it might offer a less toxic means to mitigate mutant AR toxicity. Symptomatic treatments with anti-oxidants and exercise are worth evaluating further, but the effects of these downstream interventions are likely to be modest at best.

Perhaps the best current approaches to SBMA treatment are (1) finding selective androgen receptor modulators (SARMs) that block mutant AR toxicity with little or no anti-anabolic activity, and (2) developing IGF-1 derivatives or agonists with a better therapeutic window, i.e., less toxicity or greater efficacy. The IGF-1 pathway is particularly attractive because of its protective effects on neurons and muscle as well as the changes it induces in the post-translational modification of mutant AR that diminish toxicity in cell culture and animal models.

SBMA is a model neuromuscular disease with a well defined cause and pharmacological handles on treatment. Successfully mitigating the manifestations of this disease could point the way to effective treatments for other, more common causes of skeletal muscle weakness.

Figure 1. Therapeutic targets for SBMA.

(1) Decreased ligand availability mitigates the disease phenotype in transgenic animals and shows signs of efficacy in SBMA clinical trials. (2) HSP90 inhibition and increased expression of heat shock proteins (HSPs) have been beneficial in transgenic mice. (3) Axonal dysfunction and loss of neurotrophic support are implicated in SBMA. Overexpressing VEGF and IGF1 in transgenic mice has shown promise in mitigating clinical disease manifestations, as well as improving spinal cord and muscle pathology. (4) ASC-J9, a curcumin-related compound, reduces intranuclear accumulation of mutant protein (shown in black) through increased degradation. (5) Transcriptional dysregulation through altered histone acetylation is a possible mechanism. Correcting this with histone deacetylase inhibitors has been beneficial in cell culture and animal models. (6) Recent studies indicate that retaining AR in the cytoplasm by blocking its nuclear translocation and inducing autophagy (arrow indicating a membrane-bound autophagosome) might be effective. (7) Cell culture studies suggest that modulating the function of mitochondria (shown in red) and reducing oxidative stress with antioxidants might be beneficial.