Androgen receptors containing expanded polyglutamine tracts exhibit progressive toxicity when stably expressed in the neuroblastoma cell line, SH-SY 5Y.

The pathogenesis of X-linked spinal and bulbar muscular atrophy (SBMA) has been traced to an expansion of repeated glutamine (Gln) residues within the amino terminus of the human androgen receptor (AR). To examine the mechanisms by which these expanded repeat ARs (Exp-ARs) are toxic to neurons, we have established and characterized a cell culture model by stably transfecting SH-SY 5Y neuroblastoma cells with cDNAs containing either normal AR (81 series; 23 Glns) or Exp-AR (902 series; 56 Glns). At a low passage number, no differences in cell morphology, growth properties, or susceptibility to toxic insults were observed between clones expressing normal AR or Exp-AR. Initially, both types of cultures were found to express similar levels of specific hormone binding in monolayer binding assays. Immunohistochemical studies demonstrated the vast majority of both the normal AR and Exp-AR were localized to the nucleus in the absence and presence of androgen. As the 902 series of clones were propagated, the Exp-AR content in the cells appeared to decline progressively. However, this decrease actually reflects a gradual disappearance of the Exp-AR cell population. No such selection occurred during the propagation of cells expressing the normal AR. This selection against cells expressing physiological levels of Exp-AR occurs in the absence of intracellular aggregates and suggests that mechanisms other than those involving the formation of aggregates underlie the observed toxicity of Exp-ARs.

Expression of the human androgen receptor in eukaryotic cells using a recombinant adenovirus vector yields high levels of the soluble, functional receptor protein.

The androgen receptor (AR) and closely related members of the steroid receptor family have proven difficult to obtain in native form in large quantities. In the case of the human AR (hAR), high-level expression in prokaryotic or non-mammalian cells leads to the synthesis of a high proportion of non-binding, insoluble, or degraded forms of the receptor protein. To circumvent these difficulties, we have constructed a recombinant adenovirus that directs the expression of hAR under the control of a potent, constitutive promoter. Infection of eukaryotic cells with this recombinant virus leads to the synthesis of large quantities of the intact AR. In contrast to expression methods designed to direct the full-length AR in bacteria, yeast, and insect cells, AR expressed in mammalian cells using this adenoviral vector accumulates at high levels, retains many properties of the native AR, and is not rapidly proteolyzed. This method will prove useful for large-scale preparations of hAR for use in functional and structural studies.

Assessment of androgen receptor function in genital skin fibroblasts using a recombinant adenovirus to deliver an androgen-responsive reporter gene.

Mutations of the androgen receptor (AR) cause defects in virilization and can result in a spectrum of phenotypic abnormalities of male sexual development that includes patients with a completely female phenotype (complete testicular feminization) and individuals with less severe defects of virilization, such as Reifenstein syndrome. These phenotypes are not specific for mutations of the AR gene, however, and defects in other genes can also result in similar abnormalities of male development. For this reason, the diagnosis of an AR defect is laborious and requires data from endocrine studies, the family history, and in vitro binding experiments. To assist in the evaluation of patients with possible AR defects, we previously employed the use of a recombinant adenovirus to deliver an androgen-responsive gene into fibroblast cultures to assay AR function in normal subjects and patients with complete forms of androgen resistance. Although these studies demonstrated measurable differences between these two groups of subjects, we did not assay samples from patients with partial defects of androgen action. In the current study, we have modified this method to examine AR function in three groups of patients with known or suspected defects of AR function: patients with Reifenstein syndrome, patients with spinobulbar muscular atrophy, and patients with severe forms of isolated hypospadias. When assayed using this method, the AR function of patients with Reifenstein syndrome was intermediate between that of normal control subjects and that of patients with complete testicular feminization. Using the parameters established by the aforementioned experiments, we found that defective AR function can be detected in fibroblasts established from patients with spinobulbar muscular atrophy and in some patients with severe forms of isolated hypospadias, including two with a normal AR gene sequence. These results suggest that this method may have some utility in screening samples to detect defects of AR function, particularly when viewed in the context of other AR assays results.

The adenovirus-mediated delivery of a reporter gene permits the assessment of androgen receptor function in genital skin fibroblast cultures. Stimulation of Gs and inhibition of G(o).

Defects in the androgen receptor cause a spectrum of abnormalities in genetic males ranging from phenotypic women with testicular feminization to men with minor defects in fertility and/or virilization. The diagnosis of androgen resistance can be quite cumbersome, including analysis of the family history, karyotyping, endocrine studies, measurement of androgen binding in genital skin fibroblasts, and, in some instances, sequencing of mutant cDNAs. Furthermore, androgen-binding studies may be normal in patients with qualitative receptor abnormalities or mutations in the DNA-binding domain of the receptor. To circumvent these difficulties, we have used a recombinant adenovirus to deliver an androgen-responsive reporter gene (mouse mammary tumor virus-luciferase) to fibroblasts cultured from genital skin from 12 normal controls and from eight individuals with complete testicular feminization. Following incubation with androgen (2 nM mibolerone) for 72 h, luciferase activity in normal fibroblasts increased > 10-fold (range 11-200-fold) in a manner that corresponded with the level of androgen receptor detected in ligand-binding assays. By contrast, luciferase activity increased negligibly in fibroblasts from individuals with testicular feminization (average = 1.2-fold increase). This assay permits a direct assessment of endogenous androgen receptor function in cells and should be a powerful aid in the diagnosis of androgen resistance.

Androgen increases androgen receptor protein while decreasing receptor mRNA in LNCaP cells.

We have examined the effect of androgen treatment on androgen receptor mRNA and protein expression in the LNCaP human prostate carcinoma cell line. Incubation with androgen caused a decrease in cellular androgen receptor mRNA content that was concentration and time dependent. Maximal suppression to approximately 35% of control level was observed after 49 h of exposure to androgen. By contrast, incubation of LNCaP cells with androgen resulted in a 2-fold increase in the cellular content of androgen receptor protein at 24 h. At 49 h androgen receptor protein increased 30% as assayed by immunoblots and 79% as assayed by ligand binding. These results suggest that ligand-induced changes in androgen receptor stability and/or the translational efficiency of androgen receptor mRNA account for the phenomenon of androgen receptor upregulation observed in cultured LNCaP cells. Furthermore, the suppression of androgen mRNA and protein that is caused by prolonged incubation with androgen is incomplete and is reversible upon removal of ligand.

Expression of a mutant androgen receptor in cloned fibroblasts derived from a heterozygous carrier for the syndrome of testicular feminization.

Thermolability of androgen binding was compared in fibroblasts cloned from normal female skin, skin from a subject with testicular feminization whose mutation is known to be associated with a thermolabile androgen receptor, and from the mother of the subject with testicular feminization. Seven of 28 clones studied from the mother exhibited thermolability of binding, indicating that the mutant gene that causes thermolability of binding, like the gene responsible for the normal androgen receptor, is X-linked.

Variation in steroid 5 alpha-reductase activity in cloned human skin fibroblasts. Shift in phenotypic expression from high to low activity upon subcloning.

Steroid 5 alpha-reductase in skin fibroblasts varies in two ways: 1) mean activity in uncloned fibroblasts corresponds to the activity in skin from which fibroblasts are derived (high in genital skin and low in nongenital skin); 2) activity in normal genital skin fibroblasts varies over a wide range (more than 200-fold). Studies of activity in fibroblast clones provides an explanation for this variability. Activity in clones of genital skin fibroblasts varies from the limits of detectability to very high levels, whereas activity in fibroblasts cloned from nongenital skin is uniformly low. The mean value of 5 alpha-reductase in genital skin clones is similar to the value in uncloned fibroblasts from the same explant. This suggests that the numbers of high and low activity cells in a given explant influence the overall enzyme activity. Variability among normal genital skin strains is also due to inherent differences in the range of enzyme activities. Furthermore, high activity genital skin clones give rise to a mixture of both high and low-activity subclones, whereas low activity genital skin and nongenital skin clones give rise to populations with uniformly low activity. Thus, with time genital skin gives rise to high activity fibroblasts that continually undergo shift in phenotypic expression to low activity.