Human Molecular Genetics

Human Molecular Genetics Advance Access originally published online on June 13, 2006
Human Molecular Genetics 2006 15(14):2225-2238; doi:10.1093/hmg/ddl148

This Article Abstract FREE Full Text (PDF) Supplementary Data A corrigendum has been published OA All Versions of this Article: 15/14/2225    most recent ddl148v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (7) Google Scholar Articles by Thomas, P. S. Articles by La Spada, A. R. Search for Related Content PubMed PubMed Citation Articles by Thomas, P. S., Jr Articles by La Spada, A. R. Social Bookmarking          What's this?

© 2006 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Loss of endogenous androgen receptor protein accelerates motor neuron degeneration and accentuates androgen insensitivity in a mouse model of X-linked spinal and bulbar muscular atrophy

Patrick S. Thomas, Jr¹, Gregory S. Fraley⁵, Vincent Damien¹, Lillie B. Woodke², Francisco Zapata¹, Bryce L. Sopher¹, Stephen R. Plymate² and Albert R. La Spada^1,2,3,4,*

¹ Department of Laboratory Medicine, ² Department of Medicine, ³ Department of Neurology, ⁴ Center for Neurogenetics and Neurotherapeutics, University of Washington Medical Center, Seattle, WA, USA and ⁵ Department of Biology, Hope College, Holland, MI, USA

^* To whom correspondence should be addressed at: Department of Laboratory Medicine, University of Washington Medical Center, PO Box 357110, Room NW 120, Seattle, WA 98195-7110, USA. Tel: +1 2065982138; Fax: +1 2065986189; Email: laspada{at}u.washington.edu

Received April 9, 2006; Accepted June 2, 2006

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
X-linked spinal and bulbar muscular atrophy (SBMA; Kennedy's disease) is a polyglutamine (polyQ) disease in which the affected males suffer progressive motor neuron degeneration accompanied by signs of androgen insensitivity, such as gynecomastia and reduced fertility. SBMA is caused by CAG repeat expansions in the androgen receptor (AR) gene resulting in the production of AR protein with an extended glutamine tract. SBMA is one of nine polyQ diseases in which polyQ expansion is believed to impart a toxic gain-of-function effect upon the mutant protein, and initiate a cascade of events that culminate in neurodegeneration. However, whether loss of a disease protein's normal function concomitantly contributes to the neurodegeneration remains unanswered. To address this, we examined the role of normal AR function in SBMA by crossing a highly representative AR YAC transgenic mouse model with 100 glutamines (AR100) and a corresponding control (AR20) onto an AR null (testicular feminization; Tfm) background. Absence of endogenous AR protein in AR100Tfm mice had profound effects upon neuromuscular and endocrine-reproductive features of this SBMA mouse model, as AR100Tfm mice displayed accelerated neurodegeneration and severe androgen insensitivity in comparison to AR100 littermates. Reduction in size and number of androgen-sensitive motor neurons in the spinal cord of AR100Tfm mice underscored the importance of AR action for neuronal health and survival. Promoter–reporter assays confirmed that AR transactivation competence diminishes in a polyQ length-dependent fashion. Our studies indicate that SBMA disease pathogenesis, both in the nervous system and the periphery, involves two simultaneous pathways: gain-of-function misfolded protein toxicity and loss of normal protein function.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
X-linked spinal and bulbar muscular atrophy (SBMA; Kennedy's disease) is an inherited neurodegenerative disorder that occurs exclusively in men, and typically presents in the 4th to 5th decade of life, although disease onset in juveniles may occasionally occur (1,2). Neurodegeneration in SBMA is largely restricted to lower motor neurons in the spinal cord and to bulbar nuclei of the brain stem; upper motor neurons are spared (3). However, neuropathology is not restricted to motor neurons in SBMA. With disease progression, dorsal root ganglion neurons may degenerate, thereby causing a mild distal sensory deficit (4,5). SBMA patients also develop varying degrees of testicular atrophy, gynecomastia, impaired fertility and elevated androgen levels—all signs of mild-to-moderate androgen receptor (AR) insensitivity (6,7).

The AR gene was evaluated as a candidate gene for SBMA due to the androgen insensitivity signs and X-linked pattern of inheritance. Linkage mapping studies independently confirmed the AR gene as a valid candidate (8). This work culminated in the discovery of a novel mutation, expansion of a CAG trinucleotide repeat, in the first exon of the AR gene as the cause of SBMA (9). Whereas CAG repeats range in size from 5 to 35 triplets in normal individuals, patients with SBMA always display 37 or more CAG repeats in the first exon of the AR gene. As CAG encodes the amino acid glutamine and the AR CAG repeat begins at codon 58, SBMA is a polyglutamine (polyQ) repeat expansion disease. Subsequent studies of several other dominantly inherited neurodegenerative diseases have yielded causal CAG repeats encoding polyQ tract expansions. At present, SBMA is one of nine such polyQ repeat diseases in a family of disorders that includes Huntington's disease (HD), dentatorubral-pallidoluysian atrophy and six forms of spinocerebellar ataxia: SCA 1, 2, 3, 6, 7 and 17 (10,11).

Unlike most other polyQ repeat diseases, the function of the protein mutated in SBMA is well established. The AR protein is a steroid hormone transcription factor that contains an amino-terminal transactivation domain, a DNA-binding domain with two zinc fingers, a hinge region and a ligand-binding domain (12,13). AR acts as a typical transcription factor by interacting with co-activators, co-repressors and other components of the transcriptional machinery to drive the expression of its target genes. AR-driven transcription regulates muscle growth, bone growth, spermatogenesis and the development of secondary sexual characteristics (14). Humans or mice lacking any functional AR protein fail to develop male external genitalia, appear to be female and are completely infertile despite a 46, XY karyotype—a disorder known as testicular feminization (15,16). In humans, a continuum of phenotypes due to varying degrees of impaired AR function have also been described and are collectively known as the androgen insensitivity syndrome (AIS) (17). Interestingly, although SBMA patients always appear male, they display obvious signs of incomplete virilization, such as reduced fertility and reduced testicular size (6,7). This observation suggests that the polyQ expansion mutation in SBMA imparts a partial loss-of-function in tissues important for the expression of male secondary sexual characteristics. However, the SBMA polyQ expansion mutation is dominant in its action, with female SBMA carriers protected due to decreased levels of circulating ligand (18). The dominant inheritance pattern of all other polyQ repeat diseases, together with an extensive literature on the production and generation of cell culture and animal models of these diseases, strongly supports a toxic gain-of-function effect of the polyQ expansion tract (11,19,20).

Although polyQ expansion mutations produce a dominant gain-of-function toxicity, gain-of-function and loss-of-function mechanisms are not mutually exclusive in these diseases. Indeed, there is considerable evidence for a pathogenic role of diminished normal function of disease proteins containing polyQ tract expansions. In one study, post-natal elimination of huntingtin protein expression yielded striatal degeneration in HD conditional knock-out mice (21). In another murine model, the HD yeast artificial chromosome (YAC) 128 mouse, absence of endogenous huntingtin expression accentuated HD neuropathology (22,23). Various in vitro studies have also found that cells with depressed levels of huntingtin expression are more susceptible to polyQ toxicity (24,25). As most polyQ disease proteins are widely expressed throughout the neuraxis but only produce degeneration in select regions of the central nervous system, partial loss of disease protein function could render particular neurons exquisitely sensitive to the toxic gain-of-function effects that a polyQ expansion imparts to the disease protein. Although a mechanistic explanation for how loss-of-function of a polyQ protein contributes to polyQ disease pathogenesis is lacking, a number of recent studies have further underscored the importance of polyQ disease protein normal function (26–31). In the case of SBMA, whether partial loss of AR function contributes to the motor neuron degeneration remains unclear.

To investigate whether normal AR function plays a role in SBMA neurodegeneration, we employed a highly representative SBMA mouse model (i.e. the AR100 mouse), in which the human AR gene carrying 100 CAG repeats is expressed under proper temporal and spatial regulation within a 450 kb DNA fragment contained in a YAC (32). The AR100 mouse displays a late-onset, neuromuscular, gender-specific phenotype characterized by neurogenic atrophy. To evaluate the contribution of normal AR function to SBMA neurodegeneration, we compared male mice carrying the human AR gene with 100 glutamine repeats (Ar^+/Y;Tg::AR^Q100 or AR100) to mice carrying the same transgene, but lacking the endogenous AR gene (AR^Tfm/Y;Tg::AR^Q100 or AR100Tfm). Both AR100 and AR100Tfm mice express the same amount of polyQ-expanded AR; however, AR100Tfm mice exhibit earlier onset of weight loss, kyphosis, hindlimb atrophy, weakness and neurodegeneration compared with AR100 mice. Interestingly, the AR null background not only accelerated the neuromuscular phenotype of AR100 transgenic mice, but also uncovered the androgen insensitivity in sexual organs and in the spinal cord. These results support a role for the loss of wild-type AR protein function in SBMA disease pathogenesis.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
AR100Tfm mice develop an accelerated neuromuscular phenotype
To study the role of normal AR function in the AR100 YAC transgenic mouse model of SBMA (32), we obtained female mice heterozygous for the testicular feminization (Tfm) mutation and crossed these female Tfm carriers with male AR100 and AR20 transgenic mice. The Tfm mutation is a spontaneous single base-pair frameshift mutation in the first exon of the murine AR gene that eliminates the expression of endogenous AR protein (33). Tfm mice are viable, but are effectively AR null and, therefore, recapitulate the phenotype of human Tfm syndrome (15,33). As the Tfm mice are maintained on a C57BL/6J background, and our AR20 and AR100 lines had been backcrossed onto the C57BL/6J background for >12 generations, a confounding effect of strain background differences did not complicate this breeding scheme. All offspring obtained from female Tfmxmale AR100 (or AR20) matings were genotyped for the human AR YAC transgene, the Tfm mutation and the Y chromosome.

To determine the effect of endogenous AR function upon the SBMA phenotype, we compared the behavioral phenotypes of AR100, AR100Tfm, AR20, AR20Tfm, non-transgenic (NT) and Tfm male mice obtained from crossing female Tfm mice with male AR100 or AR20 transgenic mice. By 8 months of age, AR100Tfm mice were already smaller and lighter than their AR100 littermates (Fig. 1A). During the course of their lives, both AR100 and AR100Tfm mice displayed a failure to gain weight and even lost weight compared with controls; however, the AR100Tfm weight loss was significantly worse than the AR100 weight loss. A distinguishing feature of the AR100 SBMA mouse model is kyphosis (i.e. curvature of the spine) due to proximal muscle weakness (32). With disease progression at 11 months of age, AR100 mice are visibly kyphotic and smaller than normal (Fig. 1B). AR100Tfm mice at 11 months of age, however, always appeared the smallest and most kyphotic of the TfmxAR100 progeny (Fig. 1B). To compare the phenotypic severity of littermate progeny obtained from these crosses, we developed a phenotype rating scale (Table 1). When a blinded observer rated the phenotypic severity of TfmxAR100 and TfmxAR20 progeny at 7 and 10 months of age using this scale, a worsening of the SBMA disease phenotype in AR100Tfm mice became apparent (Fig. 1C and D). To further validate and quantify the differences in phenotypic severity between AR100 and AR100Tfm mice, we performed a grip strength analysis on age-matched littermates. Although AR100 mice did not display significant impairments in grip strength at 10 months of age in comparison to AR20 mice, AR100Tfm mice did exhibit significant weakness (Fig. 1E). By 17 months of age, AR100 mice were significantly weaker than age-matched AR20 mice in the grip strength paradigm; however, AR100Tfm were, by far, the weakest cohort tested (Fig. 1F).

View larger version (59K): [in this window] [in a new window]   Figure 1. Loss of the endogenous AR accelerates the SBMA phenotype in AR100 YAC transgenic mice. (A) AR100Tfm mice lose more weight than AR100 mice. We followed the weights of male control and AR transgenic mice (n=4/group) on a bi-monthly basis. Here, we see a chart of weight as a function of age for the different genotypes. Both AR100 and AR100Tfm display significant progressive weight loss over time compared with Tfm and AR20Tfm mice (P<0.05 by ANOVA). However, the weight loss in the AR100Tfm cohort is significantly greater than the weight loss in their AR100 littermates (P<0.05 by ANOVA). To evaluate the individual contribution of endogenous AR status and of CAG repeat length, we performed multiple regression analysis and found that each of these independent factors has a significant effect upon weight (P<0.005). (B) AR100Tfm mice are smaller and more kyphotic than AR100 mice. Here, we see representative photographs of 11-month-old male NT, Tfm, AR100 and AR100Tfm mice. Note the greater spine curvature and smaller size of the AR100Tfm individual in comparison to his AR100 littermate. (C) AR100Tfm mice display a more severe phenotype at 7 months of age. Sets of control and AR transgenic mice (n=5 to 8/group) were assigned a phenotype severity score (Table 1) by a blinded observer. Although AR100 mice do not display signs of disease, 80% of the AR100Tfm cohort were scored as abnormal, indicating more severe disease (P<0.05 by Mann–Whitney U test). (D) AR100Tfm mice display a more severe phenotype at 10 months of age. Sets of control and AR transgenic mice (n=5 to 8/group) were assigned a phenotype severity score (Table 1) by a blinded observer. All but one of the AR100Tfm mice were rated as abnormal, and the severity of disease in the AR100Tfm group significantly exceeded phenotypic severity in the other cohorts including the AR100 group (P<0.05 by Mann-Whitney U test). (E) Grip strength analysis at 10 months of age. Cohorts of control and AR transgenic mice (n=5 to 8/group) underwent combined forepaw—hindpaw grip strength testing to derive mean grip strength force measurements. By this time point, AR100Tfm mice are already significantly weaker than their AR100 littermates (P<0.05 by ANOVA). Differences among AR100 and the controls were not significant. Multiple regression analysis indicated that endogenous AR status and glutamine tract length are independent significant factors in grip strength performance (P<0.001). (F) Grip strength analysis at 17 months of age. Cohorts of control and AR transgenic mice (n=5 to 8/group) underwent combined forepaw–hindpaw grip strength testing to derive mean grip strength force measurements. Both AR100 and AR100Tfm are significantly weaker than age-matched controls (P<0.05 by ANOVA); however, AR100Tfm mice were the weakest, displaying a significant difference in grip strength in comparison to their AR100 littermates (P<0.05 by ANOVA). Differences among the control animals (i.e. AR20, AR20Tfm, NT and Tfm) were not significant. Multiple regression analysis indicated that endogenous AR status and glutamine tract length are independent significant factors in grip strength performance (P<0.001).

 

View this table: [in this window] [in a new window]   Table 1. Phenotype severity scale

 
Although AR100Tfm mice display muscle atrophy reminiscent of SBMA, confirmation of neurogenic muscle atrophy requires the use of special histopathology stains. To confirm the existence of a motor neuronopathy and compare its severity among the TfmxAR100 progeny, we combined hematoxylin and eosin (H&E) staining of quadriceps muscle sections with immunostaining for neural cell adhesion molecule (NCAM), a protein known to be up-regulated during muscle renervation (34). At 6.5 months of age, AR100 mice display only modest muscle pathology with limited fiber size variation and some replacement of muscle by fibrous tissue and fat, in comparison to NT and Tfm controls (Fig. 2A–C). Muscle sections from AR100Tfm mice, however, display significant variation in fiber size with considerable connective tissue replacement and fat infiltration (Fig. 2D). NCAM staining correspondingly revealed only occasional positive muscle fibers in AR100 mice, but detected frequent positive muscle fibers in AR100Tfm mice (Fig. 2E–H). These findings suggest that the absence of endogenous AR protein results in more rapidly progressive muscle atrophy due to denervation in SBMA mice. Given these obvious differences in muscle histopathology, basic dye staining of lumbar cord sections from NT, Tfm, AR100 and AR100Tfm indicated a trend toward increased motor neuron drop-out in the AR100Tfm sections (Fig. 2I–L); however, counting of neurons from cord sections did not yield a significant difference in motor neuron number (data not shown). This was not surprising as the production of a neurological phenotype in mouse models for the polyQ repeat diseases is not clearly associated with neuron cell death (35).

View larger version (129K): [in this window] [in a new window]   Figure 2. AR100Tfm mice develop more rapidly progressive neurogenic muscle atrophy. (A–H) Muscle histopathology in AR transgenic mice. Quadriceps muscle sections from sets of 6.5-month-old male NT (A, E), Tfm (B, F), AR100 (C, G), and AR100Tfm (D, H) littermate mice (n=3 to 4/group) were stained with H&E (A–D) and an anti-NCAM antibody (E–H). While H&E staining of control NT and Tfm sections reveals muscle fibers of similar size and shape with peripheral nuclei (A, B), AR100 and AR100Tfm sections display signs of muscle pathology, including infiltration of fat cells and connective tissue (C, D). Comparison of AR100 and AR100Tfm sections, however, reveals that AR100 muscle is only mildly abnormal, whereas AR100Tfm muscle pathology is much more advanced. NCAM staining does not detect any evidence of up-regulation in control NT and Tfm muscle sections (E, F), but does highlight occasional NCAM-positive fibers in AR100 muscle sections (G). NCAM staining of AR100Tfm muscle sections detects many more positive muscle fibers (H). AR20Tfm muscle sections appeared identical to age-matched NT mice (data not shown). (I–L) Motor neuron histopathology in AR transgenic mice. Sets of 20 µm lumbar spinal cord sections from 6.5-month-old male NT (A, E), Tfm (B, F), AR100 (C, G) and AR100Tfm (D, H) littermate mice (n=3 to 4/group) were stained with a basic dye stain (Richardson's) and the ventrolateral columns serially evaluated. Here, we see a representative set of images. NT, Tfm and AR100 mice have numerous motor neurons of large size and with prominent nuclei surrounded by pronounced cytosolic basophilia (I–K). AR100Tfm lumbar spinal cord sections tend to display slightly reduced numbers of motor neurons, and the surviving motor neurons are denser-staining and smaller in size (L), suggesting that degenerative changes occur earlier in AR100Tfm mice.

 
Transgenic expression of AR100 protein does not rescue the Tfm phenotype in SBMA mice
SBMA patients display signs of androgen insensitivity (6,7). The AR100 YAC mouse model of SBMA does not display androgen insensitivity signs as the mutant human AR transgene is expressed in the presence of the normal endogenous mouse AR gene (32). After crossing male AR20 and AR100 transgenic mice with female Tfm carriers, we derived AR20Tfm and AR100Tfm mice, and evaluated them for signs of any androgen insensitivity phenotype. We first examined the distance between the anus and genitals of mice [i.e. the anogenital distance (AGD)], as AGD is a reliable measure of external masculinization in mice (36,37). Normal male mice have an AGD at least 50% greater than that of female mice or androgen-insensitive Tfm mice (Fig. 3A and B). Visual inspection of external genitalia of AR20Tfm and AR100Tfm mice revealed that neither human AR20 protein nor AR100 protein completely restored the AGD to normal male values (Fig. 3C and D). Although the AR20 (line C9B) transgene and AR100 (line C25) transgene are expressed equivalently at roughly endogenous levels (32), AGD in the AR20Tfm mice was roughly twice that of AGD in the AR100Tfm mice, indicating that the AR20 protein was better able to rescue the female-like anogenital appearance produced by the Tfm mutation in comparison to AR100 protein (Fig. 3E). Indeed, AR100Tfm mice displayed AGDs indistinguishable from female mice, indicating that the AR100 protein did not affect any appreciable change upon the appearance of the external genitalia. Thus, rescue of the Tfm mutation's effect upon the appearance of the external genitalia inversely correlated with polyQ tract length, suggesting that presence of a polyQ expansion in the human AR protein impairs its function.

View larger version (72K): [in this window] [in a new window]   Figure 3. AR100Tfm mice display a severe androgen insensitivity phenotype. (A) Representative photograph of the genitalia of a 5-month-old male NT mouse. Note prominent distance between the anus and fully descended testes, known as the AGD. (B) Representative photograph of the genitalia of a 5-month-old female mouse. Note difference in AGD and appearance. (C) Representative photograph of the genitalia of a 5-month-old AR20Tfm XY mouse. Note that AGD is intermediate between that of a normal male and normal female mouse. (D) Representative photograph of the genitalia of a 5-month-old AR100Tfm XT mouse. Note that AGD is indistinguishable from that of a female mouse. (E) Comparison of AGD in AR transgenic mice. AGD lengths were measured by ruler by a blinded examiner for cohorts (n=7/group) of NT male, NT female, Tfm, AR20Tfm and AR100Tfm mice. AR100Tfm mice displayed a mean AGD comparable with Tfm and female mice, while AR20Tfm mice had a significantly longer AGD in comparison to mice with these genotypes (P<0.05 by ANOVA with Tukey HSD post hoc test). Although AR20Tfm mice displayed partial rescue of AGD length, AR20Tfm mean AGD length was still significantly shorter than normal male mice (P<0.05 by ANOVA with Tukey HSD post hoc test). Error bars correspond to SEM. Comparison of total testosterone (F), FT (G) and LH (H) levels for sets of NT, Tfm, AR20, AR20Tfm, AR100 and AR100Tfm mice (n=6/group). AR100Tfm mice have significantly higher total testosterone, FT and LH levels compared with AR20Tfm mice (P<0.05 by ANOVA with a Games–Howell post hoc test). Differences between AR20 and AR100 animals, however, were not significant. Error bars correspond to SEM. AR20, AR20Tfm and NT mice have comparable total testosterone and LH levels (P>0.05 by ANOVA).

 
Luteinizing hormone (LH) is released from the pituitary and binds to receptors on Leydig cells of the testes to stimulate testosterone biosynthesis. As most of the circulating total testosterone is bound to carrier proteins in the bloodstream, measurement of unbound free testosterone (FT) reflects the bioactive testosterone fraction. In AIS patients, testosterone and LH levels are often elevated in proportion to the degree of insensitivity (17). We therefore obtained serum samples from adult NT, Tfm, AR20, AR20Tfm, AR100 and AR100Tfm mice, and measured their levels of LH, total testosterone and FT. Expression of the AR100 protein was associated with increased total testosterone and FT levels, and comparison of AR100Tfm mice with AR20Tfm mice revealed that total testosterone and FT levels were approximately three to five times greater in the AR100Tfm mice in comparison with AR20Tfm mice (Fig. 3F and G). This pronounced difference was also reflected in LH, as AR100Tfm mice displayed LH levels that were approximately eight times greater than LH levels in AR20Tfm mice (Fig. 3H). Levels of total testosterone and LH in AR20Tfm mice were comparable with those measured in AR20 and NT mice, whereas Tfm mice displayed very low testosterone levels and elevated LH (Fig. 3F–H). These findings indicated that the AR100 protein was unable to adequately perform its expected endocrine end organ functions in the AR100Tfm mice. As a final measure of AR function, we perfused AR transgenic mice of various genetic constitutions, dissected out their testes and compared testes weights, as testicular atrophy is also a hallmark of androgen insensitivity in SBMA. Surprisingly, we noted a significant reduction in testes weight in AR100 mice when compared with AR20 mice (Supplementary Material, Fig. S1), which, when taken together with the increased FT levels in AR100 mice (Fig. 3F), suggests a possible dominant negative effect of AR100 protein upon normal endogenous AR protein function or polyQ gain-of-function toxicity. Although AR20Tfm and AR100Tfm testes only weighed 25% of testes from control male mice, AR20Tfm and AR100Tfm testes each weighed twice as much as testes obtained from Tfm mice, indicating a modest degree of rescue.

To further evaluate the ability of AR protein with different length polyQ tracts to rescue the Tfm phenotype, we compared testes sections from AR transgenic and control mice after H&E staining. Testes sections from normal male mice display a stereotypical anatomical organization with well-stratified seminiferous tubules and large numbers of spermatids being extruded into the central lumen (Fig. 4A). Testes sections from Tfm mice, however, show poorly organized seminiferous tubules devoid of properly developing germ cells or spermatids (Fig. 4B). When we assessed testes sections from AR20Tfm and AR100Tfm mice, we noted severe abnormalities. Both AR20Tfm and AR100Tfm testes lacked spermatids and displayed failure of germ cell progression in their remnant seminiferous tubules (Fig. 4C and D). Although both AR20Tfm and AR100Tfm testes were markedly abnormal, AR20Tfm testes better approximated the expected circumferential layering of germ cells, whereas AR100Tfm testes were utterly disorganized and contained degenerating spermatocytes marked by regions of dense chromatin and frequent vacuolization. Thus, although neither the AR20 nor the AR100 protein could rescue the testes abnormalities in mice lacking functional endogenous AR, AR100Tfm testes always appeared worse than AR20Tfm testes. We also evaluated testes sections from AR100 transgenic mice, and noted that their seminiferous tubules were relatively normal in appearance and contained numerous extruded spermatids, consistent with their fertile status (Fig. 4E).

View larger version (138K): [in this window] [in a new window]   Figure 4. PolyQ length-dependent testicular pathology in AR transgenic mice on the Ar null background. Testes sections from sets of 10 month-old male control (A), Tfm (B), AR20Tfm (C), AR100Tfm (D), and AR100 (E) transgenic mice (n=3–4/group) were evaluated by H&E staining. Note normal stratified, circumferential organization of seminiferous tubulues with numerous spermatids extruded into central lumen in control testes (A). Tfm testes lack stratified seminiferous tubules and fail to produce spermatids (B). AR20Tfm (C) and AR100Tfm (D) testes are also markedly abnormal, but AR20Tfm testes have partially stratified seminiferous tubules, whereas AR100Tfm testes are utterly disorganized. Note presence of degenerating spermatocytes (red arrows) in AR100Tfm testes, indicative of extremely severe pathology. (E) AR100 testes closely resemble normal testes. All micrographs are of identical magnification, and a scale bar corresponding to 50 µm is included in panel D.

 
How does impaired AR100 protein function worsen SBMA neurodegeneration?
As studies of endocrine and reproductive tissues from AR100Tfm mice revealed evidence for androgen insensitivity due to impaired normal AR function, we wondered if and how impaired normal AR function could account for the more severe neuromuscular phenotype documented in the AR100Tfm mice vis-à-vis the AR100 mice. As AR100Tfm mice have higher levels of ligand in their bloodstream (Fig. 3E), and nuclear localization of AR requires ligand binding (38,39), we performed immunohistochemistry to assess the subcellular localization of AR in motor neurons of the lumbar spinal cord of transgenic animals. Using a human-specific AR antibody, we noted that AR100Tfm mice display a significantly higher proportion of motor neurons with AR immunostaining in their nuclei in comparison to AR100 mice (Fig. 5). NT, AR20 and AR20Tfm mice do not show appreciable accumulation of human AR in motor neuron nuclei (Fig. 5; data not shown), because only polyQ-expanded proteins have impaired turnover (40,41). To exclude increased nuclear AR accumulation in AR100Tfm motor neurons due to the increased AR protein expression, we measured human AR mRNA expression by quantitative PCR and protein expression by quantitative western blot analysis and we found that AR100 and AR100Tfm animals express the human AR transgene in the spinal cord at comparable levels (Supplementary Material, Fig. S2). As previous studies have linked nuclear localization of mutant AR protein with AR polyQ toxicity (42,43), enhanced nuclear accumulation of mutant AR protein in motor neurons of AR100Tfm mice likely contributes to their accelerated neuromuscular phenotype.

View larger version (47K): [in this window] [in a new window]   Figure 5. Increased nuclear localization of AR protein in AR100Tfm motor neurons. (A–C) Confocal immunostaining analysis of spinal cord sections. A set of 8-month-old NT, Tfm, AR100 and AR100Tfm male mice (n=3/group) were studied by obtaining 40 µm thick lumbar spinal cord sections and immunostaining with an anti-AR antibody (red) in combination with DAPI (blue). Whereas nuclear accumulation of human AR protein is not observed in motor neurons from NT mice (A), many motor neurons from AR100 (B) and from AR100Tfm (C) mice display intense nuclear staining. Greater numbers of motor neurons from AR100Tfm mice displayed such nuclear staining. (D) We counted the numbers of motor neurons showing nuclear staining with an AR-specific antibody, as shown in (A–C), for sets of section (n=4/section/three mice/genotype). Comparison of AR-labeled motor neuron nuclei revealed that AR100Tfm mice display 2.5xmore AR-positive motor neuron nuclei than AR100 mice (P<0.0001 by t-test).

 
Although alteration in ligand levels due to impaired AR function represents an indirect pathway by which loss of normal AR function could contribute to SBMA disease pathogenesis, we sought to determine if impaired AR function could directly underlie accelerated motor neuron degeneration in AR100Tfm mice. To do this, we evaluated the spinal nucleus of the bulbocavernosus (SNB), the dorsolateral nucleus (DLN) and the ventral medial pool (VMP), as motor neurons in the SNB and DLN of male mice require the action of androgen and reflect the function of the AR protein (44), whereas VMP motor neurons do not (45,46). Spinal cord sections from cohorts of NT, Tfm, AR20, AR20Tfm, AR100 and AR100Tfm male mice (n=6/group) were thus stained with cresyl violet to permit analysis of SNB, DLN and VMP motor neurons (Fig. 6A–F). While AR20, AR20Tfm and AR100 mice displayed comparable numbers of SNB motor neurons of roughly equivalent size, SNB motor neurons in AR100Tfm mice were significantly fewer in number and smaller in size (Fig. 6G and H). Although the sizes of SNB motor neurons in AR100Tfm mice and Tfm mice were similar, the presence of AR100 protein did partially rescue SNB motor neuron number in AR100Tfm mice (Fig. 6G and H). Analysis of VMP motor neuron number and size revealed no significant differences between NT, Tfm, AR20, AR20Tfm, AR100 and AR100Tfm male mice (Supplementary Material, Fig. S3).

View larger version (107K): [in this window] [in a new window]   Figure 6. Androgen-responsive motor neurons are depleted in AR100Tfm mice. (A–F) Assessment of lumbar spinal cord motor neuron regions by cresyl violet staining. Here we see representative images of lumbar cord sections from cohorts of 10 month-old AR transgenic mice and controls (n=6/group). In NT controls (A), numerous motor neurons in the SNB (marked by arrows) and in the DLN (marked by arrowheads) are apparent. Note that ‘cc’ corresponds to ‘central canal’. SNB motor neurons are noticeably absent in Tfm mice (B), and DLN motor neurons are diminished in size. (C) AR100Tfm sections also display a profound loss of SNB motor neurons and reduction in DLN motor neurons. Red arrows indicate possible SNB motor neurons that failed to migrate to the appropriate anatomical location. (D) AR100 sections reveal much greater numbers of SNB motor neurons (arrows) and DLN motor neurons (arrowheads), though migratory mislocalization is common (red arrows). AR20Tfm mice exhibit reductions in SNB and DLN motor neurons with some mislocalized neurons (E), but in AR20 mice (F), numerous healthy appearing SNB motor neurons (arrows) and DLN motor neurons (arrowheads) are apparent. (G) Comparison of SNB motor neuron size. The soma of surviving SNB motor neurons in AR100Tfm and Tfm mice are similar in size, but significantly smaller than motor neuron cell bodies from AR20Tfm or AR100 mice (P<0.05 by ANOVA). Error bars correspond to SEM. Multiple regression analysis indicated that glutamine tract length has a significant effect upon SNB motor neuron size (P<0.01). (H) Comparison of SNB motor neuron numbers. AR100Tfm mice have significantly more SNB motor neurons than Tfm mice (P<0.05 by ANOVA), but significantly fewer SNB motor neurons than AR100 or AR20Tfm mice (P<0.05 by ANOVA). Error bars correspond to SEM. Multiple regression analysis revealed that both endogenous AR status and glutamine tract length are significant independent factors in determining SNB motor neuron number (P<0.01).

 
Effect of polyQ length expansion upon AR transactivation competence
Endocrine, reproductive and neuroanatomical characterization of the AR100Tfm mice strongly supports the view that polyQ tract expansion in AR decreases its normal function of ligand-dependent transcription activation. To test this hypothesis, we obtained M12 cells, a line of cells derived from a primary prostate carcinoma that does not express AR (47). We then co-expressed human AR cDNA's of different polyQ lengths (n=20, 66, 112 and 250), a triple probasin AR-response element (ARE) promoter–reporter construct and Renilla luciferase, and measured the AR transactivation competence. PolyQ tract expansion produced significant repression of ARE-driven transactivation in a polyQ length-dependent manner (Fig. 7). This reduction could not be attributed to differences in AR protein translation, as we documented comparable expression of AR20, AR66, AR112 and AR250 protein on western blots (data not shown), and also normalized promoter activity to protein levels (Fig. 7).

View larger version (13K): [in this window] [in a new window]   Figure 7. PolyQ expansion impairs AR transactivation in vitro. AR transcription activation was assayed by co-transfecting the AR-null human prostate epithelial M12 cell line with a triple AR probasin firefly luciferase promotor–reporter construct (AAR3) and AR CMV expression constructs with polyQ tracts of either 20, 66, 112 or 250. Luciferase activity was normalized to protein expression by calculating the ratio of luciferase activity to the intensity of AR protein measured from an aliquot of transfected cells by western blot analysis. Ratio values were then normalized to the AR20 mean result that was arbitrarily set to 1.0. Promoter activation by AR20 was significantly greater than promoter activation by AR66, whose transactivation significantly exceeded that of AR112 or AR250 (P<0.05 by ANOVA; post hoc testing significant at P<0.05 for all comparisons, except AR112 versus AR250). All transfections were performed in duplicate, and measurements were done in triplicate. Error bars correspond to SEM. Sequestration of AR protein into aggregates was excluded by examining the stacking portions and wells of immunoblotted protein lysates from transfected M12 cell lines analyzed by western blot, and noting the absence of immunoreactivity (not shown).

 

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Although the paradigm of polyQ expansion as a toxic gain-of-function mutation is supported by numerous investigations (11,19,20), the question of whether loss or alteration of a polyQ disease protein's normal function might concomitantly contribute to polyQ neurodegeneration has arisen. In HD, there is strong evidence supporting a role for loss or alteration of normal huntingtin protein function in disease pathogenesis—post-natal elimination of huntingtin gene expression in the forebrain of mice appears to result in striatal degeneration (21). How the diminished huntingtin function contributes to HD striatal degeneration remains unclear, however, as some studies point to huntingtin's anti-apoptotic actions (24,25), other studies implicate impaired huntingtin regulation of transcription (48) and still others argue that huntingtin normally participates in axonal transport (30,49). In spinocerebellar ataxia type 7, recent studies have found that the ataxin-7 protein is a core component of the STAGA co-activator complex (27,50,51), with one study demonstrating that polyQ-expanded ataxin-7 diminishes STAGA complex co-activator function to produce SCA7 retinal degeneration in mice (27). As the normal functions of other polyQ disease proteins, such as ataxin-2, ataxin-3 and TATA-binding protein, appear to impinge upon crucial pathways of translation regulation, protein degradation and transcription regulation (26,28,29,31,52), it seems plausible that expansion of the polyQ tract could also contribute to the neurodegenerative process in these diseases by adversely affecting the normal protein function.

SBMA is unique among the CAG/polyQ repeat diseases for a number of reasons. Concomitant loss of a well-established normal protein function is a defining feature of SBMA, as affected male patients display signs of androgen insensitivity (2,4,6). Indeed, clinicians have extensively characterized the androgen insensitivity phenotype of SBMA patients (1,7). However, whether impaired AR function in SBMA contributes to the motor neuron degeneration is unknown. To address this question and thereby determine if loss of normal protein function is a feature of polyQ-mediated neurodegeneration in this disorder, we humanized a highly representative AR YAC CAG100 transgenic mouse model (32), by crossing AR100 (and control AR20) male mice with female Tfm carriers. AR100 male mice with the endogenous mouse AR gene (=AR100) and AR100 male littermate mice lacking the endogenous mouse AR gene (AR100Tfm) were compared with one another, and to male (XY) NT, Tfm, AR20 and AR20Tfm control mice. Absence of endogenous AR protein function in AR100Tfm mice had profound effects upon both the neurodegenerative phenotype and the endocrine-reproductive features of this SBMA mouse model. AR100Tfm mice displayed an accelerated neuromuscular disease phenotype in comparison with AR100 mice, based upon physical examination, behavioral testing and histopathology analysis. AR100Tfm mice appeared female, failed to develop functional testes, and exhibited elevations in FT and LH. The worsening of the SBMA neurodegenerative phenotype and endocrine-reproductive functions could not be attributed to strain background effects, as the female Tfm mice were on a pure C57BL/6J background and the AR transgenic mice were incipient C57BL/6J congenic mice at the time of the cross-breeding.

In SBMA, affected males typically do not present with neuromuscular disease until the 4th decade of life; however, it is not uncommon for SBMA patients to be recognized for breast enlargement after puberty or to present to endocrinologists due to concerns over reduced fertility prior to the onset of their neurological disease. Although the androgen insensitivity phenotype in SBMA is highly variable, SBMA patients always appear male from birth and there has never been any report of SBMA patients who are incompletely masculinized or have ambiguous genitalia. In the AR100Tfm mice, we observed a dramatic AIS phenotype, as male mice displayed AGD sizes that were more consistent with the appearance of females, had incompletely descended testes and were infertile. This severe androgen insensitivity consequently yielded marked elevations in FT and LH at levels well beyond what would typically be seen in human SBMA males. The reason for the profound androgen insensitivity may reflect the use of an exceptionally long CAG100 repeat tract in the AR YAC used to model SBMA. In SBMA, patients seldom have CAG repeats >65 triplets, well short of the CAG100 repeat used to derive the SBMA mice. Previous studies of SBMA have suggested different possible explanations for the mechanistic basis of the androgen insensitivity, with some groups favoring decreased AR protein expression and other groups proposing impaired AR transactivation competence (53–57). In this study, we analyzed AR transactivation competence in AR null M12 cells and observed a progressive reduction in AR transactivation as a ratio of AR protein expression with increasing polyQ tract length. Our results support impaired AR transactivation function as the basis for the androgen insensitivity phenotype in SBMA, and our observation of a dramatic AIS phenotype in the AR100Tfm mice likely reflects the greatly diminished transactivation function of an AR protein with an exceptionally long 100 glutamine repeat tract. Comparison of the endocrine-reproductive phenotypes of AR100Tfm and AR20Tfm are consistent with this interpretation, as AR100Tfm mice had significantly shorter AGD's, higher FT and LH levels and more atypical testes than AR20Tfm mice. However, despite achieving levels of protein expression approximating endogenous Ar, the expression of human normal AR20 protein was not capable of completely rescuing the Tfm endocrine-reproductive phenotype, as AR20Tfm mice were not fertile and displayed incompletely masculinized genitalia. As the AR20 transgene was carried on a YAC, additional regulatory information for proper developmental expression of AR may reside outside of the 5' and 3' regions contained in the AR YAC construct. Consistent with this explanation is the observation that AR knock-in mice whose murine AR genes have been replaced with human exon 1 sequences, containing either 21 or 41 CAG repeats, develop fully masculinized genitalia, are fertile, have normal testis morphology and display normal testosterone and LH levels (58).

In addition to characterizing the endocrine-reproductive features of the AR100Tfm mice, we decided to consider the effects of impaired AR action upon motor neuron degeneration in this mouse model. Numerous studies have demonstrated that AR action is crucial for neuron survival and recovery from injury in rodents (59–61). These studies suggest that AR-mediated gene transcription helps to maintain the health of motor neurons, a function that may be particularly important in the face of stress and injury. As it turns out, certain populations of motor neurons (i.e. SNB and DLN) in the spinal cord of rodents depend upon androgen action for their growth and survival. We therefore performed a stereological analysis of the SNB motor neurons, and observed a significant reduction in SNB size and number in AR100Tfm mice in comparison to both AR100 mice and AR20Tfm mice. We also noted a decrease in DLN motor neurons in AR100Tfm mice, but did not detect any differences in motor neuron size or number in the VMP, a non-androgen-dependent region of the spinal cord. These findings indicate that impaired AR function in the AR100Tfm mice accounts for reduced numbers of SNB and DLN motor neurons, and suggests that AR may serve as a trophic factor in motor neurons in the spinal cord.

For many years, it has been known that androgen accumulates in the spinal cord (62). The distribution of androgen-responsive motor neurons in the spinal cord may reflect differential susceptibility to a variety of toxic insults. In humans, Onuf's nucleus is a set of androgen-sensitive motor neurons that is analogous to the SNB in rodents (63). Interestingly, Onuf's nucleus is resistant to injury and degeneration in amyotrophic lateral sclerosis and autosomal recessive spinal muscular atrophy (64–66). While the mechanism of this protective effect is unknown, the basis of AR's trophic action may stem from AR-regulated gene expression. Consistent with this hypothesis, one group has documented a ligand-dependent, polyQ length-dependent effect of AR upon neuron survival in motor neuron-like cells in culture (67). Although AR's trophic action likely depends upon the transactivation of target genes by binding at promoters and regulatory regions containing AREs, AR may also mediate signaling events and trigger cellular responses through a recently identified non-genomic pathway (68). In either case, the presence of functional AR in motor neurons may be crucial for supporting motor neuron survival in the face of various insults.

In addition to finding evidence for impaired AR function as a potential contributor to spinal cord motor neuron degeneration in AR100Tfm mice, we noted that spinal cord motor neurons in AR100Tfm male mice more frequently accumulate the mutant AR protein in their nuclei in comparison to AR100 male mice. As AR100Tfm mice display markedly increased levels of androgen, we attributed this increased nuclear localization of mutant AR in motor neuron nuclei to the presence of higher circulating levels of androgen in their bloodstreams. Indeed, both human SBMA females and female SBMA transgenic mice do not develop a pronounced motor neuronopathy, as they do not produce sufficiently high levels of testosterone (18,32,42). Studies of SBMA transgenic mice corroborate this view, as castration or pharmacological ablation of testis function with leuprolide can prevent or ameliorate mutant AR toxicity, whereas delivery of testosterone to female mice can produce a neurological phenotype (42,69,70). The deleterious effect of ligand upon mutant AR has been linked to the ligand's ability to favor nuclear translocation of the mutant AR protein (43), although other actions of ligand have not been excluded. The molecular basis of the mutant AR gain-of-function effect in the nucleus may involve transcription interference, possibly upon CREB-binding protein's co-activation of vascular endothelial growth factor gene expression (32). Thus, in the SBMA AR100Tfm mice, our studies indicate that their accelerated motor neuron degeneration may involve two independent pathways: enhanced gain-of-function mutant AR nuclear toxicity and loss-of-function reduced AR trophic action. In agreement with our results, AR knock-in mice carrying 113 CAG repeats display testes abnormalities that reflect both a toxic gain-of-function and loss of normal AR protein function (58). Our findings thus support an emerging view of polyQ repeat disease pathogenesis stemming from two independent processes: a gain-of-function misfolded protein toxicity and a concomitant loss or alteration of the protein's normal function. Further studies will be needed to dissect the interplay of these pathways, but resolution of this problem will be crucial for therapy development, as a number of promising therapeutic options rely upon RNAi knock-down of mutant gene expression (71). Unless such approaches can discriminate between the mutant allele and the normal allele, non-specific knock-down could actually worsen polyQ disease degeneration and produce untoward side effects due to an accentuated loss of the protein's normal function.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Mouse genetics and expression analysis
Production and characterization of the AR YAC CAG20 (AR20) and AR6 YAC CAG100 (AR100) transgenic mice have been described (32). Male AR transgenic mice that were on an incipient C57BL/6J congenic strain background (after >12 generations of backcrossing) were mated with C57BL/6J female Tfm carrier mice obtained from Jackson Laboratories (Bar Harbor, ME, USA). Three PCR amplifications were performed to genotype resulting progeny: (i) To assess transgene status, PCR amplification with a forward primer (5'-catctgagtccaggggaacagc-3') and a reverse primer (5'-gcccaggcgctgccgtagtcc-3') was followed by digestion of resultant PCR products with Bgl I. Agarose gel electrophoresis of digested products yielded a 450 bp murine band for NT mice or 450, 350 and 100 bp bands for mice carrying the human transgene. (ii) To identify the Tfm mutation, genomic DNA was PCR amplified using a forward primer (5'-tctggcagcagtgaagca-3') and a reverse primer (5'-cagcactggatcgcccccagc-3'). The product was digested with Mwo 1. Mice with the AR Tfm mutation display a 300 bp band whereas mice with a non-mutant AR gene possess a 200 bp band. (iii) To determine chromosomal sex, genomic DNA was PCR amplified with Zfxy primers [5'-gagtgtggtaaaggttttcgtc-3' (forward) and 5'-gattcgcatgtgctttttga-3' (reverse)]. XY mice display a 100 bp band. Specific cycling conditions are available upon request. Real-time RT-PCR was performed on spinal cord RNA samples isolated and analyzed using the TaqMan approach on an ABI 7500 as previously described (32,72,73). Quantitative western blot analysis was performing by immunoblotting of spinal cord lysates with antibody H-280 as previously described (32), and determining signal intensity as a ratio of AR—beta-actin using the NIH Image-J program.

Behavioral analysis and physical examination
A blinded examiner performed the following phenotype assessment. Mice were weighed, rated for kyphosis, analyzed for quadriceps muscle thickness by palpation, observed during ambulation and judged for activity levels. Using these metrics, a phenotype severity score was assigned using a pre-determined rating scale (Table 1). We designed a grip strength testing apparatus by attaching an Accuforce dynamometer (Ametek, Largo, FL, USA) to an 18'' high plastic stand. Each animal was lowered by the tail onto a wire grid to measure the combined forepaw–hindlimb strength force. Each animal was tested five times in succession, and the highest grip strength was recorded as the grip strength for that animal. For determination of AGD, the distance between the border of the anus and genitals was measured to the nearest centimeter with calipers by a blinded examiner. Cages of mice with the genotypes of interest were selected randomly for all behavioral assays and physical examinations.

Histology and immunohistochemistry
After anesthetization, the mice were perfused transcardially with heparinized saline followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (pH 7.4). In some cases, the spinal cord was immediately removed and processed as described previously (32). In other cases, spinal columns were removed and immersed in 4% PFA and then coded for blind analysis. Such spinal cords were subsequently removed, static-fixed in 4% PFA for 2 h, and then placed into 0.1 M phosphate buffer with 20% sucrose O/N for cryoprotection. Spinal cords were then embedded in 8% gelatin and stored at –80°C until processed. Three sets of 40 µm transverse sections were direct mounted onto Superfrost-Plus slides (Fisher Scientific, Chicago, IL, USA). Sections were processed for Nissl substance using cresyl violet staining. Slides were then dehydrated in graded ethanol, cleared in citra-solv and cover-slipped. For double-blind analysis, the slides were coded, randomized and examined under light microscopy by an individual unaware of coding. Motor neurons of the SNB, DLN and VMP were identified based upon size and location, and analyzed under light microscopy as previously described (74,75). Motor neurons from each pool were counted and 50% of neurons (all unilaterally) were digitized under a 40xobjective using a Leica DMB500 5.0 Mpix digital camera connected to a Leica DM5000b microscope. Digitized images of motor neurons were analyzed for somal area using the NIH Image-J program. For analysis of testes, mice were perfused as described above, and then testes were dissected out and weighed. Each testis was embedded in a paraffin block, serially sectioned (5 µm thickness), stained with H&E and examined. Muscle histopathology studies were performed as previously described (32). Confocal microscopy of frozen spinal cord transverse sections was performed as previously described (76). For counting AR-positive motor neurons, a blinded examiner scored 200 motor neurons in different quadrants of the ventrolateral horn for at least four sections/individual for three mice/group to derive percentages of AR-positive motor neurons/genotype.

Measurement of hormone levels
Blood was collected from 10-month-old male mice by retro-orbital collection with heparinized capillary tubes (Chase Scientific, Rockwood, TN, USA). Whole blood was centrifuged at 2000gx10 min in a 4°C microfuge (Eppendorf Centrifuge 5417R) to collect serum for analysis. Each sample was stored at –20°C and sent to the Center for Research in Reproduction Ligand Assay and Analysis Core at the University of Virginia for analysis of FT and LH. Total testosterone levels were measured in the Endocrinology clinical laboratory at the University of Washington Medical Center.

Promoter–reporter assays
Transient transfections of M12 cells with the triple AR probasin luciferase promoter (AAR₃) (generous gift from Dr Matusik) and CMV-AR expression vectors were performed using Lipofectamine Plus (Life Technologies) according to the manufacturer's protocol. The AAR₃ construct is an artificial reporter containing three repeats of the rat probasin ARE1 and ARE2 regions upstream of the thymidine kinase promoter (77). A total of 8.6x10⁴cells/well were seeded in 12-well plates in RPMI medium containing 5% FBS. After 24 h, each well received 1.2 µg AAR₃ construct, 1.2 µg CMV-AR full-length expression vector (either AR-20-FL, AR-45-FL, AR-66-FL, AR-112-FL or AR-250-FL) in 6 µl Plus reagent (Invitrogen) and 3 µl of Lipofectamine reagent (Invitrogen) in serum-free media. After a 3 h exposure to the Lipofectamine/DNA/plus mixture, the medium was supplemented with 1% charcoal-stripped FBS and incubated for 24 h, after which the transfection medium was removed and DHT at a concentration of 10 nM was added for 24 h. Luciferase activity was determined using the Luciferase Assay System (Promega) according to the manufacturer's protocol.

Statistical analysis
All statistical analysis was performed using SPSS 12.02 for Windows (SPSS, Chicago, IL, USA), GB-Stat for the Mac (General Dynamics, Inc., Bethesda, MD, USA), or the VassarStats website http://faculty.vassar.edu/lowry/VassarStats.html. For parametric data, the Univariate General Linear Model (GLM) of ANOVA and an appropriate post hoc test were used. For non-parametric analysis, raw data was transformed by natural log, and GLM ANOVA with an appropriate post hoc test was employed. The level of significance (alpha) was always set at 0.05.

	SUPPLEMENTARY MATERIAL

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Supplementary Material is available at HMG Online.

	ACKNOWLEDGEMENTS

 
We would like to thank A. Smith, J. Huang and D.E. Possin for excellent technical assistance. This work is supported by research grants from the Muscular Dystrophy Association (A.R.L.) and from the NIH (R01 NS41648 to A.R.L., K01 DK66238 to G.F.S., T32 GM07266 to P.S.T. and T32 GM07108 to P.S.T.). P.S.T. is supported by the Cora May Poncin Scholarship Fund and A.R.L. was the recipient of a Paul Beeson Faculty Scholar in Aging Research award from the American Foundation for Aging Research (AFAR). This study utilized the Center for Research in Reproduction Ligand Assay and Analysis Core at the University of Virginia that is funded by the NIH (U54-HD28934).

Conflict of Interest statement. The authors declare that they have no competing financial interests or other conflicts of interest.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 

1. Echaniz-Laguna A., Rousso E., Anheim M., Fleury M., Cossee M., Tranchant C. (2005) A clinical, neurophysiological and molecular study of 12 patients from 4 families with spinal and bulbar muscular atrophy. Rev. Neurol. (Paris) 161:437–444.[Medline]

2. Kennedy W.R., Alter M., Sung J.H. (1968) Progressive proximal spinal and bulbar muscular atrophy of late onset. A sex-linked recessive trait. Neurology 18:671–680.[Free Full Text]

3. Li M., Miwa S., Kobayashi Y., Merry D.E., Yamamoto M., Tanaka F., Doyu M., Hashizume Y., Fischbeck K.H., Sobue G. (1998) Nuclear inclusions of the androgen receptor protein in spinal and bulbar muscular atrophy. Ann. Neurol. 44:249–254.[CrossRef][ISI][Medline]

4. Barkhaus P.E., Kennedy W.R., Stern L.Z., Harrington R.B. (1982) Hereditary proximal spinal and bulbar motor neuron disease of late onset. A report of six cases. Arch. Neurol. 39:112–116.[Abstract]

5. Olney R.K., Aminoff M.J., So Y.T. (1991) Clinical and electrodiagnostic features of X-linked recessive bulbospinal neuronopathy. Neurology 41:823–828.[Abstract/Free Full Text]

6. Arbizu T., Santamaria J., Gomez J.M., Quilez A., Serra J.P. (1983) A family with adult spinal and bulbar muscular atrophy, X-linked inheritance and associated testicular failure. J. Neurol. Sci. 59:371–382.[CrossRef][ISI][Medline]

7. Dejager S., Bry-Gauillard H., Bruckert E., Eymard B., Salachas F., LeGuern E., Tardieu S., Chadarevian R., Giral P., Turpin G. (2002) A comprehensive endocrine description of Kennedy's disease revealing androgen insensitivity linked to CAG repeat length. J. Clin. Endocrinol. Metab. 87:3893–3901.[Abstract/Free Full Text]

8. Fischbeck K.H., Ionasescu V., Ritter A.W., Ionasescu R., Davies K., Ball S., Bosch P., Burns T., Hausmanowa-Petrusewicz I., Borkowska J., et al. (1986) Localization of the gene for X-linked spinal muscular atrophy. Neurology 36:1595–1598.[Abstract/Free Full Text]

9. La Spada A.R., Wilson E.M., Lubahn D.B., Harding A.E., Fischbeck K.H. (1991) Androgen receptor gene mutations in X-linked spinal and bulbar muscular atrophy. Nature 352:77–79.[CrossRef][Medline]

10. Ross C.A. (2002) Polyglutamine pathogenesis: emergence of unifying mechanisms for Huntington's disease and related disorders. Neuron 35:819–822.[CrossRef][ISI][Medline]

11. Zoghbi H.Y. and Orr H.T. (2000) Glutamine repeats and neurodegeneration. Annu. Rev. Neurosci. 23:217–247.[CrossRef][ISI][Medline]

12. O'Malley B. (1990) The steroid receptor superfamily: more excitement predicted for the future. Mol. Endocrinol. 4:363–369.[ISI][Medline]

13. Gelmann E.P. (2002) Molecular biology of the androgen receptor. J. Clin. Oncol. 20:3001–3015.[Abstract/Free Full Text]

14. Mooradian A.D., Morley J.E., Korenman S.G. (1987) Biological actions of androgens. Endocr. Rev. 8:1–28.[Medline]

15. Lyon M.F. and Hawkes S.G. (1970) X-linked gene for testicular feminization in the mouse. Nature 227:1217–1219.[CrossRef][Medline]

16. Pinsky L., Trifiro M., Kaufman M., Beitel L.K., Mhatre A., Kazemi-Esfarjani P., Sabbaghian N., Lumbroso R., Alvarado C., Vasiliou M., et al. (1992) Androgen resistance due to mutation of the androgen receptor. Clin. Invest. Med. 15:456–472.[ISI][Medline]

17. Brinkmann A.O. (2001) Molecular basis of androgen insensitivity. Mol. Cell. Endocrinol. 179:105–109.[CrossRef][ISI][Medline]

18. Schmidt B.J., Greenberg C.R., Allingham-Hawkins D.J., Spriggs E.L. (2002) Expression of X-linked bulbospinal muscular atrophy (Kennedy disease) in two homozygous women. Neurology 59:770–772.[Abstract/Free Full Text]

19. Ordway J.M., Tallaksen-Greene S., Gutekunst C.A., Bernstein E.M., Cearley J.A., Wiener H.W., Dure L.S.T., Lindsey R., Hersch S.M., Jope R.S., et al. (1997) Ectopically expressed CAG repeats cause intranuclear inclusions and a progressive late onset neurological phenotype in the mouse. Cell 91:753–763.[CrossRef][ISI][Medline]

20. Paulson H.L., Bonini N.M., Roth K.A. (2000) Polyglutamine disease and neuronal cell death. Proc. Natl. Acad. Sci. USA 97:12957–12958.[Free Full Text]

21. Dragatsis I., Levine M.S., Zeitlin S. (2000) Inactivation of Hdh in the brain and testis results in progressive neurodegeneration and sterility in mice. Nat. Genet. 26:300–306.[CrossRef][ISI][Medline]

22. Leavitt B.R., Guttman J.A., Hodgson J.G., Kimel G.H., Singaraja R., Vogl A.W., Hayden M.R. (2001) Wild-type huntingtin reduces the cellular toxicity of mutant huntingtin in vivo. Am. J. Hum. Genet. 68:313–324.[CrossRef][ISI][Medline]

23. Van Raamsdonk J.M., Pearson J., Rogers D.A., Bissada N., Vogl A.W., Hayden M.R., Leavitt B.R. (2005) Loss of wild-type huntingtin influences motor dysfunction and survival in the YAC128 mouse model of Huntington disease. Hum. Mol. Genet. 14:1379–1392.[Abstract/Free Full Text]

24. Leavitt B.R., Raamsdonk J.M., Shehadeh J., Fernandes H., Murphy Z., Graham R.K., Wellingtson C.L., Raymond L.A., Hayden M.R. (2006) Wild-type huntingtin protects neurons from excitotoxicity. J. Neurochem. 96:1121–1129.[CrossRef][ISI][Medline]

25. Rigamonti D., Bauer J.H., De-Fraja C., Conti L., Sipione S., Sciorati C., Clementi E., Hackam A., Hayden M.R., Li Y., et al. (2000) Wild-type huntingtin protects from apoptosis upstream of caspase-3. J. Neurosci. 20:3705–3713.[Abstract/Free Full Text]

26. Li F., Macfarlan T., Pittman R.N., Chakravarti D. (2002) Ataxin-3 is a histone-binding protein with two independent transcriptional corepressor activities. J. Biol. Chem. 277:45004–45012.[Abstract/Free Full Text]

27. Palhan V.B., Chen S., Peng G.H., Tjernberg A., Gamper A.M., Fan Y., Chait B.T., La Spada A.R., Roeder R.G. (2005) Polyglutamine-expanded ataxin-7 inhibits STAGA histone acetyltransferase activity to produce retinal degeneration. Proc. Natl. Acad. Sci. USA 102:8472–8477.[Abstract/Free Full Text]

28. Chai Y., Berke S.S., Cohen R.E., Paulson H.L. (2004) Poly-ubiquitin binding by the polyglutamine disease protein ataxin-3 links its normal function to protein surveillance pathways. J. Biol. Chem. 279:3605–3611.[Abstract/Free Full Text]

29. Donaldson K.M., Li W., Ching K.A., Batalov S., Tsai C.C., Joazeiro C.A. (2003) Ubiquitin-mediated sequestration of normal cellular proteins into polyglutamine aggregates. Proc. Natl. Acad. Sci. USA 100:8892–8897.[Abstract/Free Full Text]