Stability of an expanded trinucleotide repeat in the androgen receptor gene in transgenic mice.

The expansion of trinucleotide repeat sequences underlies a number of hereditary neurological disorders. To study the stability of a trinucleotide repeat and to develop an animal model of one of these disorders, spinal and bulbar muscular atrophy (SBMA), we have generated transgenic mice carrying either the normal or expanded repeat human androgen receptor (AR) gene. Unlike the disease allele in humans, the AR cDNA containing the expanded repeat in transgenic mice showed no change in repeat length with transmission. Expression of the SBMA AR was found in transgenic mice, but at a lower level than normal endogenous expression. The lack of a physiological pattern of expression may explain why no phenotypic effects of the transgene were observed.

Amino acid substitutions in the hormone-binding domain of the human androgen receptor alter the stability of the hormone receptor complex.

We have investigated the basis of androgen resistance in seven unrelated individuals with complete testicular feminization or Reifenstein syndrome caused by single amino acid substitutions in the hormone-binding domain of the androgen receptor. Monolayer-binding assays of cultured genital skin fibroblasts demonstrated absent ligand binding, qualitative abnormalities of androgen binding, or a decreased amount of qualitatively normal receptor. The consequences of these mutations were examined by introducing the mutations by site-directed mutagenesis into the androgen receptor cDNA sequence and expressing the mutant cDNAs in mammalian cells. The effects of the amino acid substitutions on the binding of different androgens and on the capacity of the ligand-bound receptors to activate a reporter gene were investigated. Substantial differences were found in the responses of the mutant androgen receptors to incubation with testosterone, 5 alpha-dihydrotestosterone, and mibolerone. In several instances, increased doses of hormone or increased frequency of hormone addition to the incubation medium resulted in normal or near normal activation of a reporter gene by cells expressing the mutant androgen receptors. These studies suggest that the stability of the hormone receptor complex is a major determinant of receptor function in vivo.

Complete testicular feminization caused by an amino-terminal truncation of the androgen receptor with downstream initiation.

We have characterized the molecular defect causing androgen resistance in two 46,XY siblings with complete testicular feminization. Although binding studies in genital skin fibroblasts showed a reduced Bmax, an increased dissociation rate of ligand, and an 8S peak of dihydrotestosterone binding on sucrose density gradient centrifugation, no immunoreactive androgen receptor (AR) was detected in immunoblots using anti-NH2-terminal antibodies, suggesting an abnormal amino terminus. Sequence analysis of the AR gene revealed a point mutation CAG-->TAG (Gln-->Stop) at nucleotide 340. In vitro mutagenesis studies suggest the synthesis of the mutant AR is initiated downstream of the termination codon at reduced levels and that each molecule is functionally impaired. These results define a novel mechanism causing androgen resistance: the combination of decreased amount and functional impairment of AR caused by an abnormality within the amino terminus of the receptor. These findings suggest that domains important to the in vivo function of the receptor reside within the amino terminus and that disruption of these domains can occur with only subtle effects on receptor binding. Identification of this mutation made it possible to identify the mutant allele within the family and to ascertain antenatally that it was not present in a 46,XY fetal sibling of the proband at 9 wk gestation.

A mutation in the DNA-binding domain of the androgen receptor gene causes complete testicular feminization in a patient with receptor-positive androgen resistance.

Androgen resistance is associated with a wide range of quantitative and qualitative defects in the androgen receptor. However, fibroblast cultures from approximately 10% of patients with the clinical, endocrine, and genetic features characteristic of androgen resistance express normal quantities of apparently normal androgen receptor in cultured genital skin fibroblasts (receptor-positive androgen resistance). We have analyzed the androgen receptor gene of one patient (P321) with receptor-positive, complete testicular feminization and detected a single nucleotide substitution at nucleotide 2006 (G----C) within the second "zinc finger" of the DNA-binding domain that results in the conversion of the arginine residue at position 615 into a proline residue. Introduction of this mutation into the androgen receptor cDNA and transfection of the expression plasmid into eukaryotic cells lead to the synthesis of a receptor protein that displays normal binding kinetics but is inactive in functional assays of receptor activity. We conclude that substitution mutations in the DNA-binding domain of the androgen receptor are one cause of "receptor-positive" androgen resistance.

Molecular basis of androgen resistance in a family with a qualitative abnormality of the androgen receptor and responsive to high-dose androgen therapy.

We have examined the nature of the mutant androgen receptor in a family with a severe defect in virilization associated with a qualitative defect in receptor function. The androgen receptor gene in this family contains two structural alterations: a single nucleotide substitution at position 2444 in exon 5 (adenosine----guanosine) that converts tyrosine 761 to a cysteine residue and a shortened glutamine homopolymeric segment in exon 1 that encodes 12 rather than the usual 20-22 glutamines. A family study was performed using polymerase chain reaction amplification of the glutamine-rich segment, and it was shown that the sister of the proband does not carry the mutant allele. The effects of these two mutations on the function of the androgen receptor were studied by introducing the changes, individually and in combination, into cDNAs encoding the normal human androgen receptor and analyzing the receptor protein produced after transfection of the cDNAs into eukaryotic cells. The presence of a cysteine residue at position 761 causes rapid dissociation of dihydrotestosterone from the receptor protein. Marked thermolability of the transfected receptor protein, however, was demonstrable only upon introduction of an androgen receptor cDNA containing both the partial deletion of the glutamine homopolymeric segment and a cysteine residue at position 761. Likewise, the ability of the receptor to stimulate a reporter gene is strikingly diminished only when both alterations are present, suggesting that the shortened glutamine homopolymeric segment amplifies the impairment of receptor function caused by the tyrosine to cysteine substitution.

A single nucleotide substitution introduces a premature termination codon into the androgen receptor gene of a patient with receptor-negative androgen resistance.

Mutations of the androgen receptor that impair the action of 5 alpha-dihydrotestosterone and testosterone result in abnormal male sexual development. The definition of the organization of the androgen receptor gene has permitted us to examine its structure in nine patients with androgen resistance that exhibit absent 5 alpha-dihydrotestosterone binding in cultured fibroblasts (receptor-negative androgen resistance). Using labeled probes specific for each individual coding exon, we find no gross rearrangements, insertions, or deletions of the androgen receptor gene in these patients. To analyze the genetic defect in these receptor-negative patients, we used the polymerase chain reaction to amplify each individual exon of the androgen receptor gene in nine affected patients. In all patients, the size of each amplified exon segment was identical to that in normal individuals. The nucleotide sequence of the entire coding region of the androgen receptor was determined in one of these patients. A single nucleotide substitution was identified that results in a premature termination codon in exon 6 at amino acid 794. S1 nuclease protection assays demonstrated that normal levels of androgen receptor mRNA are present in skin fibroblasts of this patient. Transfection of a mutated androgen receptor cDNA containing a termination codon at position 794 into eukaryotic cells resulted in formation of a normal amount of receptor protein, as indicated by immunoblotting, but the expressed protein does not bind 5 alpha-dihydrotestosterone. These findings suggest that the presence of a premature termination codon at amino acid 794 of the androgen receptor is the cause of androgen resistance in this patient.

Mutations in the ligand-binding domain of the androgen receptor gene cluster in two regions of the gene.

We have analyzed the nucleotide sequence of the androgen receptor from 22 unrelated subjects with substitution mutations of the hormone-binding domain. Eleven had the phenotype of complete testicular feminization, four had incomplete testicular feminization, and seven had Reifenstein syndrome. The underlying functional defect in cultured skin fibroblasts included individuals with absent, qualitative, or quantitative defects in ligand binding. 19 of the 21 substitution mutations (90%) cluster in two regions that account for approximately 35% of the hormone-binding domain, namely, between amino acids 726 and 772 and between amino acids 826 and 864. The fact that one of these regions is homologous to a region of the human thyroid hormone receptor (hTR-beta) which is a known cluster site for mutations that cause thyroid hormone resistance implies that this localization of mutations is not a coincidence. These regions of the androgen receptor may be of particular importance for the formation and function of the hormone-receptor complex.

Use of the probasin promoter ARR2PB to express Bax in androgen receptor-positive prostate cancer cells.

BACKGROUND: Adenovirus-mediated overexpression of the apoptosis-inducing protein Bax can induce apoptosis in prostate cancer cell lines. Constitutive overexpression of Bax could result in unwanted apoptosis in every site of accidental Bax accumulation in vivo. Therefore, we developed an adenoviral construct (Av-ARR2PB-Bax) in which the probasin promoter, modified to contain two androgen response elements, drives Bax expression. This promoter would be expected to limit expression of Bax to cells expressing the androgen receptor. METHODS: A variety of androgen receptor (AR)-positive and -negative cell lines of prostatic or nonprostatic origin were infected with Av-ARR2PB-Bax or a control virus, Av-ARR2PB-CAT, in which the same promoter drives expression of the chloramphenicol acetyl transferase-reporter gene. Bax expression and apoptosis in vitro were assessed by western blot analysis. Tumor size and apoptosis in vivo were assessed after four weekly injections of Av-ARR2PB-Bax or Av-ARR2PB-CAT into subcutaneous LNCaP xenografts growing in uncastrated male mice. All statistical tests were two-sided. RESULTS: Bax was overexpressed in an androgen-dependent way in AR-positive cell lines of prostatic origin but not in AR-positive cells of nonprostatic origin or in AR-negative cell lines of either prostatic or nonprostatic origin. The androgen dihydrotestosterone activated apoptosis in LNCaP cells infected with Av-ARR2PB-Bax but not in those infected with Av-ARR2PB-CAT. Av-ARR2PB-Bax-injected LNCaP xenograft tumors decreased in tumor size from 34.1 mm3 (95% confidence interval [CI] = 25.1 mm3 to 43.1 mm3) to 24.6 mm3 (95% CI = -2.5 mm3 to 51.7 mm3), but the difference was not statistically significant (P =.5). Tumors injected with Av-ARR2PB-CAT increased in size, from 28.9 mm3 (95% CI = 12.7 mm3 to 45.1 mm3) to 206 mm3 (95% CI = 122 mm3 to 290 mm3) (P =.002) and contained statistically significant more apoptotic cells (23.3% [95% CI = 21.1% to 25.6%] versus 9.5% [95% CI = 8.0% to 11.1]) (P<.001). CONCLUSIONS: Av-ARR2PB-Bax induces androgen-dependent therapeutic apoptosis in vitro and in vivo by activating apoptosis in AR-positive cells derived specifically from prostatic epithelium and does not affect nonprostatic cells.

Androgen receptor gene expression in human prostate carcinoma cell lines.

Responses to androgen vary widely among prostate cancers and prostatic carcinoma cell lines. We have explored the basis for this heterogeneity by examining the levels of androgen receptor expression in a prostate carcinoma cell line (LNCaP) that expresses the androgen receptor and two prostate carcinoma cell lines that do not contain detectable androgen receptor. We find that while the LNCaP cell line contains high levels of both the androgen receptor protein and mRNA, the receptor-negative cell lines DU-145 and PC-3 do not express androgen receptor protein as detected by immunoblotting or mRNA as detected by Northern analysis or S1 nuclease protection. These results indicate that the absence of androgen receptor expression in the androgen receptor-negative cell lines is caused by diminished androgen receptor mRNA levels. Genomic Southern analysis indicates that the differences in androgen receptor expression in each of these cell lines is not associated with detectable alterations in the structure of the androgen receptor gene.

Functional activities of the A and B forms of the human androgen receptor in response to androgen receptor agonists and antagonists.

The androgen receptor (AR) is present in many cells in two forms. The B form migrates with an apparent mass of 110 kDa and constitutes more than 80% of the immunoreactive receptor in most cell types. The A form of the AR migrates with an apparent mass of 87 kDa, appears to derive from internal translation initiation at methionine-188 in the AR open-reading frame, and usually constitutes 20% or less of the immunoreactive AR present. Previous experiments designed to examine the functional capacity of the A and B forms of the AR have been hampered by marked differences in the expression levels of the two isoforms, as the nucleotide sequence surrounding the codon encoding methionine-188 causes it to be used inefficiently as a translation initiation site. To circumvent this, we altered the nucleotide sequence surrounding methionine-188 to render it more similar to that surrounding the codon encoding methionine-1. Transfection of a cDNA containing these changes resulted in similar levels of expression of A and B forms of the AR as assessed by immunoblot assays using antibodies directed at an epitope preserved in both. Functional activities of these cDNAs were assessed using cotransfection assays that employed two model androgen-responsive genes (MMTV-luciferase and PRE2-tk-luciferase) in response to mibolerone, a potent androgen agonist, in three different cell lines. These studies demonstrated subtle differences in the activities of the A and B isoforms, which depended on the promoter and cell context. Additional studies failed to reveal any major differences in the responses of the AR-A and AR-B isoforms to a variety of androgen agonists and antagonists, suggesting that the previously reported functional defect of the AR-A is due principally to its level of expression. When assays of AR function are performed under conditions in which levels of expression of the two isoforms are equivalent, the AR-A and AR-B possess similar functional activities.

Sebright and Campine chickens express aromatase P-450 messenger RNA inappropriately in extraglandular tissues and in skin fibroblasts.

As the result of a single gene mutation, Sebright and Campine chickens have increased activity of cytochrome P-450 aromatase and increased formation of estrogen in extragonadal tissues. Affected roosters develop a feminizing state characterized by a feathering pattern typical of hens. In this paper we demonstrate that the expression of extraglandular aromatase in these chickens is due to the accumulation of aromatase mRNA similar to that expressed in the ovary of Leghorn and Campine hens. Furthermore, in fibroblasts cultured from Sebright skin, but not in normal Leghorn fibroblasts, aromatase mRNA levels are enhanced in response to 5-azacytidine and sodium butyrate, and aromatase mRNA levels in these fibroblasts correlate with enzymatic activity. We conclude that the accumulation of aromatase mRNA is a critical step in the expression of this mutation.

RFG (ARA70, ELE1) interacts with the human androgen receptor in a ligand-dependent fashion, but functions only weakly as a coactivator in cotransfection assays.

Abnormalities of the human androgen receptor (hAR) cause a range of clinical defects in male development. A large proportion of these mutations are single amino acid substitutions in the hormone-binding domain (HBD) that alter AR function by interfering with the capacity of the AR to bind androgen or to form stable hormone-receptor complexes. Prior studies have suggested that the formation of such stable, active hormone-receptor complexes is a crucial step in the modulation of genes by the AR. It is presumed that these hormone-receptor complexes interact with other proteins to participate in the formation of active transcription complexes at the initiation sites of androgen-responsive genes. Using a yeast two-hybrid screening method, we isolated a partial cDNA encoding the carboxy terminus of a protein that interacts with the hAR-HBD (amino acid residues 623-917) in a ligand-dependent fashion in a yeast two-hybrid assay. Sequence analysis of this clone revealed that it encoded a portion of a protein that had been previously characterized as RFG (RET Fused Gene). Using glutathione-S-transferase (GST) fusions of the hAR HBD and immunoprecipitation of the in vitro translated proteins, we have demonstrated that this interaction can be reproduced in vitro. To determine the capacity of this protein to modulate the activity of the AR in transfection assays, we expressed full-length RFG in the CV1 and DU145 cell lines, in combination with an AR expression vector and model androgen-responsive genes [mouse mammary tumor virus (MMTV) and PRE2-tk luciferase]. Our results demonstrate that RFG alters the induction of these reporter genes very weakly (no greater than 2-fold compared with transfections without the RFG expression plasmid). Thus, while our findings are in agreement with published reports which indicate that RFG interacts with AR-HBD in a ligand-dependent fashion, in our assays RFG does not exert major effects on the activity of the hAR in response to androgen or to other steroid hormones.

Amino acid substitutions in the DNA-binding domain of the human androgen receptor are a frequent cause of receptor-binding positive androgen resistance.

In some subjects with genetic and endocrine evidence of androgen resistance, no defect is demonstrable in the binding of androgen to its receptor in cultured genital skin fibroblasts. We have defined the molecular defect in the androgen receptor in four unrelated subjects in this category (termed receptor positive) with the phenotype of compete or incomplete testicular feminization. In these patients we detected amino acid substitutions in either exon 2 or exon 3, which encodes the DNA-binding domain of the androgen receptor. In one patient with incomplete testicular feminization, two separate mutations were present in exon 3. Introduction of these amino acid substitutions into the androgen receptor-coding segment leads to the expression of receptor proteins that bind ligand in a normal fashion but do not activate the transcription of the androgen-responsive mouse mammary tumor virus promoter. Mobility shift assays using androgen receptor fusion proteins produced in E. coli indicate that these mutations impair binding of the receptor to specific DNA sequences. In the subject with incomplete testicular feminization, a Ser-Gly substitution at amino acid residue 595 is able to partially restore DNA-binding activity to a mutant receptor protein that carries an Arg-Pro substitution at position 615. These findings indicate that mutations in amino acid residues crucial to the binding of the androgen receptor to target DNA sequences are a common cause of receptor-binding positive androgen resistance and that variable impairment of DNA binding can lead to distinctive phenotypes.

Cyclin D1 binds the androgen receptor and regulates hormone-dependent signaling in a p300/CBP-associated factor (P/CAF)-dependent manner.

The androgen receptor (AR) is a ligand-regulated member of the nuclear receptor superfamily. The cyclin D1 gene product, which encodes the regulatory subunit of holoenzymes that phosphorylate the retinoblastoma protein (pRB), promotes cellular proliferation and inhibits cellular differentiation in several different cell types. Herein the cyclin D1 gene product inhibited ligand-induced AR- enhancer function through a pRB-independent mechanism requiring the cyclin D1 carboxyl terminus. The histone acetyltransferase activity of P/CAF (p300/CBP associated factor) rescued cyclin D1-mediated AR trans-repression. Cyclin D1 and the AR both bound to similar domains of P/CAF, and cyclin D1 displaced binding of the AR to P/CAF in vitro. These studies suggest cyclin D1 binding to the AR may repress ligand-dependent AR activity by directly competing for P/CAF binding.

Definition of the human androgen receptor gene structure permits the identification of mutations that cause androgen resistance: premature termination of the receptor protein at amino acid residue 588 causes complete androgen resistance.

We have isolated and characterized the gene encoding the human androgen receptor. The coding sequence is divided into eight coding exons and spans a minimum of 54 kilobases. The positions of the exon boundaries are highly conserved when compared to the location of the exon boundaries of the chicken progesterone and human estrogen receptor genes. Definition of the intron/exon boundaries has permitted the synthesis of specific oligonucleotides for use in the amplification of segments of the androgen receptor gene from samples of total genomic DNA. This technique allows the analysis of all segments of the androgen receptor gene except a small region of exon 1 that encodes the glycine homopolymeric segment. Using these methods we have analyzed samples of DNA prepared from a patient with complete androgen resistance and have detected a single nucleotide substitution at nucleotide 1924 in exon 3 of the androgen receptor gene that results in the conversion of a lysine codon into a premature termination codon at amino acid position 588. The introduction of a termination codon into the sequence of the normal androgen receptor cDNA at this position leads to a decrease in the amount of mRNA encoding the human androgen receptor and the synthesis of a truncated receptor protein that is unable to bind ligand and is unable to activate the long terminal repeat of the mouse mammary tumor virus in cotransfection assays.

Molecular defects in the androgen receptor causing androgen resistance.

Patients with androgen resistance exhibit a spectrum of abnormalities of male sexual development ranging from 46,XY phenotypic women (complete testicular feminization) to undervirilized fertile men. The definition of the androgen receptor gene structure has permitted the identification of the defects causing androgen resistance in a number of patients. In some individuals androgen resistance is caused by large-scale structural alterations in the androgen receptor gene. In most patients, however, the androgen receptor mutation is the result of single nucleotide substitutions, which cause premature termination or amino acid replacement that result in the synthesis of defective androgen receptor proteins. These amino acid substitutions identify residues crucial to the normal function of the androgen receptor protein.

A steroidogenic factor-1-binding site and cyclic adenosine 3',5'-monophosphate response element-like elements are required for the activity of the rat aromatase promoter in rat Leydig tumor cell lines.

Although transcription initiation within CYP19 (cytochrome P450 aromatase) occurs immediately 5' to the initiator methionine (proximal promoter) in two rat Leydig tumor cell lines (R2C and H540) that express high aromatase activity and in rat ovary, the patterns of aromatase expression in the two cell types are distinctive. To define mechanisms controlling different patterns of expression of the rat aromatase proximal promoter, we performed transient transfection and gel mobility shift assays. Transfection experiments using different sized promoter fragments fused to a reporter gene were used to identify regions that are functionally important for transcriptional regulation in steroidogenic cell lines [R2C, H540, and Y1 (mouse adrenocortical cells that express low aromatase activity)]. These experiments indicate that the cAMP response element (CRE) at -231 and the steroidogenic factor-1 (SF1) motif are both required for expression of the reporter gene in each steroidogenic cell line and that the CRE at -169 is similarly required in R2C cells. Gel mobility shift assays confirm binding of nuclear proteins from the steroidogenic cell lines to the SF1 motif and to CRE (-231). Leydig tumor cells also contain nuclear proteins that bind to the CRE (-169), but nuclear extracts from R2C cells produce a uniquely shifted band compared with H540 cells. These results suggest that differences in proteins that bind to distinct elements within the rat aromatase promoter may be responsible for different patterns and levels of aromatase expression in these steroidogenic cell lines.

Time-specific androgen blockade with flutamide inhibits testicular descent in the rat.

Inhibition of androgen action by flutamide, a nonsteroidal antiandrogen, blocked testicular descent in 40% of the testes exposed to this agent continuously from gestational day 13 through postpartal day 28. By contrast, only 11% of the testes failed to descend when blocked by 5 alpha-reductase inhibitors during the same period. Flutamide administration over narrower time intervals (gestational day 13-15, 16-17, or 18-19) revealed maximal interference with testicular descent after androgen inhibition during gestational days 16-17. No significant differences in testicular or epididymal weights were evident between descended and undescended testes; furthermore, no correlation was detected between the presence of epididymal abnormalities and testicular descent. These findings indicate that androgen inhibition during a brief period of embryonic development can block testicular descent. The mechanism through which this inhibition occurs remains to be elucidated.

Reversal of flutamide-induced cryptorchidism by prenatal time-specific androgens.

The recent discovery of time-specific antiandrogens (flutamide) to induce undescent of the testes has provided us the ability to study androgen therapy for cryptorchidism in the first specific antiandrogen animal model for testicular descent. Postpartal administration of pharmacological doses of testosterone, dihydrotestosterone, or human chorionic gonadotropin failed to reverse antiandrogen-induced cryptorchidism. Reduction in the frequency of testicular undescent occurred only when these agents were administered before parturition. Indeed, statistically significant reductions in the incidence of undescended testes occurred only when androgens were administered simultaneously with the antiandrogens on gestational days 16-17. However, the successful ability of prenatal dihydrotestosterone or testosterone to prevent cryptorchidism was associated with significant complications of prenatal androgen therapy. Specifically, 100% (46 of 46) of the animals that had inhibition of flutamide-induced cryptorchidism manifested findings of hypogonadotropic hypogonadism. The findings of descended diminutive testes and epididymis in these latter animals did, however, substantiate that testicular weight does not play a significant role in descent of this organ. In summary, these studies delineate a narrow window of time in which androgens act to effect testicular descent that is independent of testicular size.

Localization of the androgen receptor in the developing rat gubernaculum.

The distribution of androgen receptor expression within the developing rat gubernaculum was examined at different times during fetal and neonatal development using polyclonal antibodies directed at the amino-terminus (amino acids 1-21) and the carboxy-terminus (amino acids 898-917) of the androgen receptor. These studies reveal a high level expression of androgen receptor in the undifferentiated cells that comprise the mesenchymal core of the gubernaculum in early development. However, during the morphological alterations of the gubernaculum into a muscular structure, the level of androgen receptor detected in these cells declines, with minimal immunoreactivity present 2 weeks postpartum. Whether the loss of staining of the mesenchyme is related to decreases in androgen receptor expression, death of the mesenchymal cells, or differentiation of these cells into myoblasts remains to be determined.