help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Krishnan, A. V.
Right arrow Articles by Feldman, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Krishnan, A. V.
Right arrow Articles by Feldman, D.
Endocrinology Vol. 143, No. 5 1889-1900
Copyright © 2002 by The Endocrine Society

CANCER

A Glucocorticoid-Responsive Mutant Androgen Receptor Exhibits Unique Ligand Specificity: Therapeutic Implications for Androgen-Independent Prostate Cancer

Aruna V. Krishnan, Xiao-Yan Zhao1, Srilatha Swami, Lars Brive2, Donna M. Peehl, Kathryn R. Ely and David Feldman

Departments of Medicine (A.V.K., X.-Y.Z., S.S., D.F.) and Urology (D.M.P.), Stanford University School of Medicine, Stanford, California 94305; and The Burnham Institute (L.B., K.R.E.), La Jolla, California 92037

Address all correspondence and requests for reprints to: Dr. David Feldman, Division of Endocrinology, SUMC, Room S-005, Stanford University School of Medicine, Stanford, California 94305-5103. E-mail: . feldman{at}cmgm.stanford.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Note Added in Proof
 References
 
The cortisol/cortisone-responsive AR (ARccr) has two mutations (L701H and T877A) that were found in the MDA PCa human prostate cancer cell lines established from a castrated patient whose metastatic tumor exhibited androgen-independent growth. Cortisol and cortisone bind to the ARccr with high affinity. In the present study, we characterized the structural determinants for ligand binding to the ARccr. Our data revealed that many of the C17, C19, and C21 circulating steroids, at concentrations that are found in vivo, functioned as effective activators of the ARccr but had little or no activity via the wild-type AR or GR{alpha}. Among the synthetic glucocorticoids tested, dexamethasone activated both GR{alpha} and ARccr, whereas triamcinolone was selective for GR{alpha}. In MDA PCa 2b cells, growth and prostate-specific antigen production were stimulated by potent ARccr agonists such as cortisol or 9{alpha}-fluorocortisol but not by triamcinolone (which did not bind to or activate the ARccr). Of the potential antagonists tested, bicalutamide (casodex) and GR antagonist RU38486 showed inhibitory activity. We postulate that corticosteroids provide a growth advantage to prostate cancer cells harboring the promiscuous ARccr in androgen-ablated patients and contribute to their transition to androgen-independence. We predict that triamcinolone, a commonly prescribed glucocorticoid, would be a successful therapeutic agent for men with this form of cancer, perhaps in conjunction with the antagonist casodex. We hypothesize that triamcinolone administration would inhibit the hypothalamic-pituitary-adrenal axis, thus suppressing endogenous corticosteroids, which stimulate tumor growth. Triamcinolone, by itself, would not activate the ARccr or promote tumor growth but would provide glucocorticoid activity essential for survival.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Note Added in Proof
 References
 
THE BIOLOGICAL ACTIONS of androgens are mediated by the AR, a member of the nuclear hormone receptor superfamily (1). The AR has been implicated in the development, growth, and progression of prostate cancer (2, 3, 4, 5, 6, 7). In some prostate cancers, AR levels are elevated because of gene amplification and/or overexpression (8, 9), whereas in others, the AR is mutated (10, 11, 12). A number of mutations in the AR have been identified in metastatic prostate cancers, and these mutations are most frequently located in the ligand-binding domain (LBD) of the receptor (4, 5, 11, 12, 13, 14). ARs with LBD mutations, such as the T877A found in the LNCaP human prostate cancer cell line (15) and in many prostatic cancers (11, 12), exhibit broadened ligand specificity (14, 15, 16). For example, the T877A mutant AR is capable of responding to hydroxyflutamide, progesterone, and estrogens (15), although the circulating levels of both progesterone and estrogens in men are low and may not be clinically significant (17). The role of AR mutations in the transition of prostate cancers to androgen-independent growth and in the subsequent failure of endocrine therapy is the focus of recent studies (2, 3, 4, 5, 6, 7).

We recently identified an AR with a double mutation (L701H and T877A) in its LBD in the human prostate cancer cell lines MDA PCa 2a and MDA PCa 2b, established from a bone metastasis of a castrated patient whose prostate cancer exhibited androgen-independent growth (18, 19). This double-mutant AR binds the prostatic androgen, dihydrotestosterone (DHT), with reduced affinity, compared with the wild-type AR or AR with the T877A mutation (20). We have also shown that the double-mutant AR responds to corticosteroids such as cortisol and cortisone (20). We designated this mutant AR as the cortisol/cortisone-responsive AR (ARccr). The ARccr is a promiscuous receptor exhibiting relaxed ligand specificity, responding to glucocorticoids, androgens, progesterone, and E2, but not aldosterone (20, 21).

In the present study, we investigated the structural requirements of ligands for the ARccr, in comparison with ligands for the human GR{alpha}. We tested natural steroids in the steroidogenic pathway, as well as synthetic corticosteroids, for their potential to act as ARccr ligands. The steroids were evaluated in functional assays, which included binding to ARccr and activation of ARccr-mediated transcription. Selected corticosteroids were also tested for their ability to cause transactivation through the single-mutant L701H AR. The abilities of key steroids to regulate the growth of MDA PCa 2b cells, which harbor the ARccr, were evaluated; and their effects on the androgen-responsive target gene prostate-specific antigen (PSA) were determined. Structure-activity relationships were addressed by studying a series of structurally related steroids.

Our studies reveal that the ARccr can be activated by a number of circulating corticosteroids and their precursors. Cortisol and 9{alpha}-fluorocortisol (FluF), the most potent agonists for ARccr, stimulate the growth of MDA PCa 2b cells and PSA secretion. The presence of ARccr would therefore provide a growth advantage to prostate cancer cells harboring these mutations by responding to cortisol and other steroids in the steroidogenic pathway and thus contribute to androgen-independent growth and the progression of prostate cancer seen in androgen-ablated patients. The antiandrogen bicalutamide (casodex), as well as the GR antagonist RU38486 (RU486), acted as antagonists through the ARccr and inhibited growth and PSA stimulation in MDA PCa 2b cells. Interestingly, the synthetic glucocorticoid triamcinolone was selective for GR{alpha} and did not bind to or activate the ARccr. Because triamcinolone did not stimulate the growth of MDA PCa 2b cells or increase PSA secretion by these cells, it might be useful as a novel therapeutic agent to suppress endogenous corticosteroids in patients whose cancers express the ARccr mutant receptors.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Note Added in Proof
 References
 
Materials
All steroids were purchased from either Sigma (St. Louis, MO) or Steraloids, Inc. (Newport, RI). Tritiated DHT, cortisol, and dexamethasone were obtained from Amersham Pharmacia Biotech (Piscataway, NJ). The mouse mammary tumor virus (MMTV) reporter plasmid pMMTV-luc and expression vectors (pSG5-AR and pSG5-GR{alpha}) were gifts from Dr. Ron Evans (Salk Research Institute, San Diego, CA), Dr. Zoran Culig (University of Innsbruck, Innsbruck, Austria), and Dr. Peter Kushner (University of California, San Francisco, CA), respectively. Biological Research Faculty and Facility (BRFF)-HPC1 medium was obtained from Biological Research Faculty and Facility (Ijamsville, MD), and DMEM:F12 and LipofectAMINE were from Life Technologies, Inc. (Rockville, MD). RU486 was a kind gift from Roussel-Uclaf (Romainville, France).

Radioligand-binding assay, Scatchard analysis, and competition-binding analysis
COS-7 cells were transfected with pSG5-AR, pSG5-GR{alpha}, or pSG5-ARccr expression vectors using LipofectAMINE (Life Technologies, Inc.) (20). After 48 h, cell monolayers were harvested, and high-salt nuclear extracts were made as previously described (22, 23). Protein concentration of the extract was determined by the method of Bradford (24). Binding assays were done as described (22, 23). In a typical binding assay, 200 µl soluble extract (0.5–1 mg protein/ml) were incubated with 0–100 nM of [3H]hormone, for 16–20 h at 0 C. Bound and free hormones were separated by hydroxylapatite. Specific binding was calculated by subtracting nonspecific binding obtained in the presence of a 250-fold excess of radioinert ligand from the total binding measured in the absence of radioinert steroid. Data were expressed as femtomoles of bound hormone per milligram of protein.

Competition-binding assays were performed with extracts of COS-7 cells expressing ARccr, in the presence of 20 nM [3H]cortisol as the ligand and various nonradioactive molecules as competitors at 1-, 10-, and 100-fold excess.

Reporter assay
CV-1 monkey kidney cells (ATCC, Manassas, VA) were transfected with the expression vectors pSG5-ARccr, pSG5-GR{alpha}, or pSG5-L701H AR, as well as the reporter MMTV-luc, as previously described (20). Five nanograms of pRL-SV40 (Promega Corp., Madison, WI) renilla luciferase were cotransfected in each sample as an internal control for transfection efficiency. The cells were treated with various steroids alone or in the presence of antagonists, for 16–30 h, and luciferase activity was determined using the dual-luciferase assay system (Promega Corp.).

Cell growth and PSA assays
MDA PCa 2b cells were routinely cultured in BRFF-HPC1 medium supplemented with 20% FBS as previously described (18, 19). The BRFF-HPC1 medium contains a high concentration of cortisol (hydrocortisone, 280 nM) as well as DHT at 0.1 nM. To test the effects of ARccr agonists, such as cortisol and other steroids, on cell growth and PSA secretion, we developed a test medium whose composition was comparable to BRFF-HPC1 except for the lack of cortisol and DHT. For these assays, cells were seeded in 6-well plates (2 x 105 cells/well) in BRFF-HPC1 medium. After 48 h, the BRFF-HPC1 medium was replaced with DMEM:F12 medium supplemented with epidermal growth factor (10 ng/ml), insulin (1 µM), bovine pituitary extract (40 µg/ml), cholera toxin (25 ng/ml), phosphoethanolamine (5 µM), seleneous acid (30 nM), BSA (250 µg/ml), and trypsin inhibitor (10 µg/ml), along with 20% FBS. We refer to this medium as test medium. Various steroids were added at the indicated concentrations in test medium. Fresh test medium and compounds were replenished every 3 d. The conditioned media were collected, and the PSA levels were measured as described (22). DNA content and [3H]thymidine incorporation were assayed as measures of cell proliferation (25). The effects of casodex and RU486 on cell growth and PSA were assessed in BRFF-HPC1 medium, and their abilities to antagonize the stimulatory effects of endogenous cortisol and DHT present in the BRFF-HPC1 medium were evaluated.

Structural models of the LBDs
Molecular models were based on an AR homology model produced in an earlier study (16) using the crystal structure of PR LBD as template (Protein Data Base accession code 1A28) (26). After this study was initiated, the crystal structure of the human AR LBD was solved (27, 28). Because there is a strong structural homology between the template structure of PR LBD and AR LBD (the root mean square deviation between {alpha}-carbons is 0.84 Å), predictions about the structural effects of mutations can be made from the homology model. Based on the x-ray crystallographic findings on the T877A mutant AR, Sack and co-workers (28) modeled the double mutant AR (ARccr) bound to DHT (described in Ref. 6) and obtained results similar to our modeling data reported in this paper.

The sequences of the PR and AR LBD are 52% identical. A molecular model of the mutant ARccr LBD was produced from this model by substitution of histidine for leucine at residue 701 and alanine for threonine at residue 877. The histidine side chain was oriented using a rotamer library derived from crystallographically determined protein structures (29). For comparison, a homology model of GR{alpha} LBD (54% identical with PR) was constructed, in the present study, using essentially the same protocol described by McDonald et al. (16). Briefly, residues of PR LBD were changed to sequences of GR{alpha} at homologous sites with the program MODELLER (30), and the initial homology model was generated automatically. A molecule of cortisol and water molecules were added based on corresponding positions of steroid rings or bound waters in the template. The model was adjusted manually to optimize side chain rotamer positions (29) guided by the progesterone structure. A few local corrections employed molecular mechanics energy minimization using CHARMm (31) within QUANTA 97.0 (Molecular Simulations, Inc., San Diego, CA).

Ligand-receptor docking analyses
Molecular coordinates for steroids with crystallographically determined structures were retrieved from the Cambridge Crystallographic Database (32) for docking analyses to AR, ARccr, or GR{alpha} LBD pockets. Ligands were manually docked into the binding pocket, orienting each molecule by superimposition of steroid rings onto the position of progesterone in the PR crystal structure (26). Molecular mechanics energy minimization calculations using CHARMm were implemented, imposing harmonic restraints on all nonligand atoms. A number of starting positions/configurations were manually generated for each ligand in the binding pocket, and the structure with the lowest energy was selected for further analysis.

Statistical analysis
Data were evaluated by ANOVA using the StatView 4.5 software (Abacus Concepts, Inc., Berkeley, CA), and P < 0.05 was considered significant.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Note Added in Proof
 References
 
DHT and glucocorticoids bind to the ARccr
Because ARccr is a mutant AR that responds to cortisol, we first determined the binding affinity of androgens and glucocorticoids for ARccr, and compared the results to the wild-type AR and GR{alpha}. Dissociation constant (Kd) values of the ARccr, wild-type AR, and GR{alpha} in COS-7 cells were measured for the binding of DHT, the major prostatic androgen, cortisol, the major circulating glucocorticoid, and dexamethasone (a potent synthetic glucocorticoid). Scatchard analyses (Fig. 1AGo) revealed that [3H]DHT, [3H]cortisol, and [3H]dexamethasone bound specifically to ARccr, with Kd values of 10, 5, and 50 nM, respectively. Compared with the wild-type AR (Fig. 1BGo), the ARccr had a 50-fold reduced affinity for DHT binding (Kd = 0.2 nM for wild-type AR vs. 10 nM for the ARccr). When compared with the GR{alpha} (Fig. 1CGo), the ARccr had a 10-fold higher affinity for cortisol and a 25-fold lower affinity for dexamethasone. These results indicate that the ARccr has a unique ligand specificity distinct from either wild-type AR or GR{alpha}.



View larger version (15K):
[in this window]
[in a new window]
  		Figure 1. The ARccr binds both androgen and glucocorticoids. COS-7 cells were transfected with expression vectors for ARccr, wild-type AR, or GR{alpha}. High-salt extracts from transfected cells were incubated with various doses of radioligand, at 0 C, in equilibrium-binding assays, as described in Materials and Methods. DEX, Dexamethasone. A, Scatchard plots of [3H]DHT, [3H]cortisol, and [3H]DEX binding to ARccr; B, Scatchard plot of [3H]DHT binding to AR; C, Scatchard plots of [3H]cortisol and [3H]DEX binding to GR{alpha}.

 
Natural corticosteroids are ARccr agonists
Because the ARccr exhibits a high affinity for cortisol, we evaluated a series of cortisol-related steroids for their ability to bind to and activate the ARccr (Fig. 2Go). The structures of these steroids are depicted in Fig. 2CGo. They included natural corticosteroids like cortisol (11ß-hydroxycortisone), corticosterone, and their corresponding precursors (11-deoxycortisol and 11-deoxycorticosterone, respectively) as well as 18-hydroxycorticosterone and the mineralocorticoid hormone, aldosterone. We also tested 11{alpha}-cortisol (the biologically inactive synthetic stereoisomer of cortisol) and cortisone, which has a keto group at the 11 position (the natural metabolite of cortisol that does not bind to or activate GR{alpha}). Competition-binding analyses were performed using [3H]cortisol as the ligand and unlabeled steroids at 1, 10, and 100 molar excess as competitors. The relative binding affinity (RBA) values of these steroids for ARccr ranked as follows (Table 1Go): cortisol 100% = cortisone 100% > DHT 41% > 11{alpha}-cortisol 16% = 11-deoxycortisol 16% > corticosterone 11% > 11-deoxycorticosterone 9% >> aldosterone <1% >> 18-hydroxycorticosterone (< 0.01%).



View larger version (23K):
[in this window]
[in a new window]
  		Figure 2. The ARccr differs from the GR{alpha} in structural requirements of ligands at the C11 position. CV-1 cells were transfected with expression vectors for ARccr (panel A) or GR{alpha} (panel B) and the reporter MMTV-luc as well as renilla luciferase plasmids. Cells were treated with the indicated steroids at 10 nM in steroid-depleted medium for 30 h. Cell extracts were subsequently assayed for luciferase activity by the dual-luciferase system (Promega Corp.). Values are given as fold activation over activity found in control cells treated with ethanol. Data represent the mean of assays performed in triplicate ± SEM. Panel C, Chemical structures of the naturally occurring corticosteroids tested in panels A and B. DOC, 11-Deoxycorticosterone; B, corticosterone; 18B, 18-hydroxycorticosterone; S, 11-deoxycortisol; F, cortisol; E, cortisone; 11aF, 11{alpha}-cortisol.

 

View this table:
[in this window]
[in a new window]
  		Table 1. RBA of different molecules for ARccr

 
These steroids were also tested for their transactivation potential, using a cotransfection assay in CV-1 cells. These cells were selected because they are devoid of the steroid receptors under investigation and lack 11ß-hydroxysteroid dehydrogenase (33), the enzyme that catalyzes the reversible conversion of cortisol to cortisone. The expression plasmids for the wild-type AR, ARccr, or GR{alpha} were cotransfected into CV-1 cells with the MMTV-luc reporter. The cells were treated with 10 nM of each steroid for 30 h. We observed a significant difference in the extent of activation of the luciferase reporter between ARccr and GR{alpha}. This observation can be explained by the fact that wild-type AR has approximately 20% of the maximal transcriptional activity of GR{alpha} on the MMTV-promoter (34). The wild-type AR could be activated only by androgens such as DHT and R1881. We tested a selected panel of corticosteroids for their ability to cause transactivation through the wild-type AR, and none of them activated the wild-type AR (data not shown).

ARccr and GR{alpha} displayed distinct activation profiles in response to the various steroids (Fig. 2Go, A and B). In these transactivation assays, DHT and most of the cortisol-related steroids, except for 18-hydroxycorticosterone (18B), activated the ARccr and induced luciferase activity (Fig. 2AGo). Cortisol and cortisone were the most effective activators of the ARccr, inducing reporter levels over 30-fold above the basal level. Corticosterone increased reporter levels 20-fold. The precursor molecules of cortisol and corticosterone (11-deoxycortisol and 11-deoxycorticosterone, respectively) were also potent ARccr activators. Remarkably, the C11 isomer of cortisol, 11{alpha}-cortisol, which is inactive through GR{alpha}, also increased ARccr-mediated gene transactivation by 13-fold. Thus, the ARccr exhibited only limited stereoisomer specificity for the C11 position of the corticosteroids. In contrast, only cortisol and corticosterone, both harboring the 11ß-hydroxyl group, functioned as GR{alpha} agonists (Fig. 2BGo). Importantly, changing the stereochemistry at C11 of cortisol from the naturally occurring (ß) to the synthetic ({alpha}) configuration resulted in a complete loss of GR{alpha}-mediated transactivation, in contrast to the ARccr. Cortisone, which has a keto group at the C11 position, had no agonist activity for GR{alpha}, as expected. In contrast, it was as effective as cortisol (which has a hydroxyl group at C11) in activating the ARccr. Overall, these data suggest that the ARccr has an activation profile distinct from those of wild-type AR and GR{alpha} and that both active glucocorticoids (cortisol and corticosterone) and inactive corticosteroids (cortisone and 11{alpha}-cortisol) are potent activators of the ARccr.

Synthetic glucocorticoids exhibit differential agonist activity for the ARccr
We next tested several commonly prescribed synthetic glucocorticoids, which are potent GR{alpha} agonists, for their possible agonist activity via the ARccr. These steroids each contain the 11ß-hydroxyl group except prednisone, which has a keto group in that position (Fig. 3CGo). They also contain modified A rings that are unsaturated at C1–C2 except the mineralocorticoid/glucocorticoid FluF. In competitive binding assays (Table 1Go), FluF exhibited a 3-fold increase in binding affinity for the ARccr, compared with cortisol. The potent glucocorticoids, prednisone ({Delta}1-dehydrocortisone), prednisolone, ({Delta}1-dehydrocortisol), and dexamethasone (9{alpha}-fluoro-16{alpha}-methylprednisolone), bound to ARccr with binding affinities approximately 5-fold lower than cortisol and cortisone (Table 1Go). Thus, the double bond at C1–C2 in the A ring decreased the binding affinity of the steroids for ARccr. Interestingly, triamcinolone (9{alpha}-fluoro-16{alpha}-hydroxyprednisolone), a potent synthetic glucocorticoid, which has a hydroxyl group in the D ring of the sterol structure replacing the C16 methyl group of dexamethasone, did not bind to ARccr.



View larger version (27K):
[in this window]
[in a new window]
  		Figure 3. Synthetic glucocorticoids exhibit differential agonist activity for the ARccr. CV-1 cells were transfected with expression vectors for ARccr (A) or GR{alpha} (B) and the reporter MMTV-luc as well as renilla luciferase plasmids. Cells were treated with the indicated steroids at 5 nM in steroid-depleted medium for 30 h. Cell extracts were subsequently assayed for luciferase activity by dual-luciferase system (Promega Corp.). Values are given as fold activation over activity found in control cells treated with ethanol. Data represent the mean of assays performed in triplicate ± SEM. C, Structures of the synthetic corticosteroids tested in A and B. TRI, Triamcinolone; PSN, prednisone; PSL, prednisolone.

 
In transactivation assays, the MMTV-reporter-transfected CV-1 cells were treated with each compound at a suboptimal concentration (5 nM) to detect differences in agonist activity between cortisol and other drugs. All of these synthetic compounds are known agonists for the GR{alpha}- and activated GR{alpha}-mediated transactivation (Fig. 3BGo). Prednisone, with a keto group at C11 position, was inactive through GR{alpha} as expected, because CV-1 cells are deficient in 11ß-hydroxysteroid dehydrogenase, the enzyme that catalyzes the in vivo conversion of prednisone to the active molecule prednisolone with a hydroxyl group at the C11 position. As shown in Fig. 3AGo, FluF had a somewhat greater activity than cortisol, consistent with its increased affinity for ARccr (Table 1Go). The ARccr, which did not distinguish between a keto or hydroxyl group at C11 position, was activated by both prednisone and prednisolone. Both prednisone and prednisolone were comparable with cortisol in ARccr-mediated transcription, although they exhibited lower affinities for binding to ARccr. Dexamethasone, which contains a 16{alpha}-methyl group and a 9{alpha}-fluoro group in addition to the A-ring double bond, showed reduced activity via the ARccr. Interestingly, triamcinolone containing a C16 hydroxyl group did not promote ARccr-mediated transactivation. Thus, the C16 hydroxyl group seems to abolish ARccr binding and gene activation through this receptor.

In summary, our transactivation studies revealed that the following hormones were ARccr agonists: androgens [DHT, T, androstenedione, and R1881 (data not shown)]; corticosteroids (cortisol, cortisone 11-deoxycorticosterone, corticosterone, 11-deoxycortisol); synthetic glucocorticoids (dexamethasone, prednisone, prednisolone); and the mineralocorticoid/glucocorticoid (FluF). The synthetic glucocorticoid triamcinolone did not bind to or activate the ARccr.

Casodex and RU486 antagonize ARccr-mediated transactivation
In search of ARccr antagonists that may have therapeutic utility in the treatment of prostate cancers harboring this type of mutated receptor, we evaluated several known receptor antagonists. These included the AR antagonists hydroxyflutamide and casodex, the GR/PR antagonist RU486, and the MR/AR antagonist spironolactone. In competition-binding assays (Table 1Go), these antagonists exhibited significant binding to the ARccr. Their RBA values ranked as follows: cortisol 100% > spironolactone 30% > RU486 16.4% > hydroxyflutamide 11.3% >> casodex 0.05% (Table 1Go). Transactivation assays demonstrated that both hydroxyflutamide (20) and spironolactone (data not shown) functioned as ARccr agonists in CV-1 cells, whereas casodex and RU486 acted as antagonists through the ARccr. As shown in Fig. 4Go, both of these antagonists caused significant inhibition of R1881, cortisol, FluF, or corticosterone-induced activation of the MMTV-luc promoter in CV-1 cells. The degree of inhibition by RU486 was greater than that produced by casodex, consistent with the fact that it exhibited a higher affinity for ARccr than casodex (see RBA values in Table 1Go). Note that triamcinolone was inactive and that casodex and RU486 did not exhibit any agonist activity in this assay.



View larger version (35K):
[in this window]
[in a new window]
  		Figure 4. Casodex and RU486 inhibit ARccr-mediated transactivation. CV-1 cells were transfected with the ARccr expression vector and the reporter MMTV-luc and renilla luciferase plasmids as described in Materials and Methods. Cells were treated with the indicated steroids, at 10 nM, in steroid-depleted medium, in the presence and absence of the antagonists casodex (10 µM) or RU486 (100 nM) for 16 h. Cell extracts were subsequently assayed for luciferase activity by the dual-luciferase system (Promega Corp.). Values are given as fold activation over the activity found in cells treated with ethanol and represent mean ± SEM of three to six determinations. Relative luciferase activities (MMTV-luc/renilla luc) in cells exposed to ethanol were 0.45 ± 0.003, 0.078 ± 0.02, and 0.07 ± 0.001 in control, casodex-, and RU486-treated cells, respectively. R1881-, F-, FluF-, and B-induced activations of the reporter were significantly lower in casodex or RU486-treated cells, compared with control (P < 0.001–P < 0.0001).

 
Transactivation through the L701H AR. Effects of glucocorticoids and casodex
Earlier studies from our laboratory (20) have shown that the presence of the single mutation L701H in the AR confers glucocorticoid responsiveness to the mutated AR. The L701H AR responds to cortisol, although to a much lower degree when compared with the ARccr (20). In the present study, we attempted to characterize the responses of the L701H AR to key glucocorticoids in transactivation assays using the MMTV-luc reporter in the presence and absence of the antagonist, casodex. The results of these experiments are shown in Fig. 5Go. The L701H AR responded best to the androgen R1881 (~6-fold induction of the reporter). Both cortisol (~2-fold) and FluF (~3-fold) could elicit responses through the L701H AR. The magnitudes of these responses through the L701H AR were, however, much lower than their responses through the ARccr double mutant (Fig. 4Go). The other difference between the two mutant receptor forms was in their ability to respond to corticosterone. Whereas corticosterone produced a significant activation of the reporter through the ARccr (~6-fold), it did not cause activation through the L701H AR. Interestingly, triamcinolone activated neither the ARccr (Fig. 4Go) nor the L701H AR (Fig. 5Go). The AR antagonist casodex again caused significant inhibition of the R1881, cortisol, or FluF-mediated activation of the L701H AR.



View larger version (24K):
[in this window]
[in a new window]
  		Figure 5. L701H AR-mediated transactivation; effects of corticosteroids and casodex. CV-1 cells were transfected with the L701H AR expression vector, the reporter MMTV-luc, and renilla luciferase plasmids, as described in Materials and Methods. Cells were treated with the indicated steroids, at 10 nM, in the presence or absence of 10 µM casodex, for 16 h, in steroid-depleted medium. Cell extracts were subsequently assayed for luciferase activity by the dual-luciferase system (Promega Corp.). Values are given as fold activation over the activity found in cells treated with ethanol and represent mean ± SEM of six determinations. Relative luciferase activities (MMTV-luc/renilla luc) in cells exposed to ethanol were 0.029 ± 0.01 and 0.024 ± 0.01 in control and casodex treated cells, respectively. R1881-, F-, and FluF-induced activation of the reporter were significantly lower in casodex-treated cells, compared with control (P < 0.01–P < 0.001).

 
Effects of agonists on MDA PCa 2b cell growth
We next examined the ability of some of the key steroids to activate the endogenous ARccr expressed in MDA PCa 2b cells and thereby modulate their growth and PSA secretion. MDA PCa cells grow best in BRFF-HPC1 medium, which contains a high concentration of cortisol (280 nM), and we use this medium to routinely culture and passage these cells. To test the effects of steroids on growth, the BRFF-HPC1 medium was replaced, 48 h after plating, with the test medium (described in Materials and Methods) lacking cortisol and DHT but containing 20% FBS, because the use of serum stripped of endogenous steroids and other growth factors could not support cell growth. The initial DNA concentrations at the beginning of these experiments ranged between 1.6–2.2 µg/well. Although the cells grew in the test medium, their growth was minimal, and the DNA concentrations at the end of the 6-d period were 3.5–4.5 µg/well in various experiments. Supplementation of the test medium with cortisol or use of the BRFF-HPC1 medium with high endogenous cortisol resulted in substantial increases in cell growth, as determined by [3H]thymidine incorporation and DNA content (Fig. 6Go). When cultured in BRFF-HPC1 medium, the DNA content increased to 10–15 µg/well at the end of 6 d. When cortisol (10–200 nM) was added back to the test medium we found increases in cell growth; and at 100 nM cortisol supplementation of the test medium, the cell growth increased to approximately match that found in the BRFF-HPC1 medium. We could not demonstrate a clear-cut dose dependence for the cortisol effect. The presence of 20% serum in the test medium with its complement of endogenous steroids and the corticosteroid-binding protein might have been a confounding factor. However, the data clearly demonstrate the growth-promoting effects of cortisol on these cells.



View larger version (29K):
[in this window]
[in a new window]
  		Figure 6. Effect of cortisol on the growth of MDA PCa 2b cells. Cells were plated in BRFF-HPC1 medium. The starting DNA concentration was 1.94 ± 0.21 µg/well. After 48 h, the effects of different doses of cortisol were tested in the test medium lacking cortisol for the next 6 d, as described in Materials and Methods. Test medium was supplemented with either ethanol vehicle (Control, Con) or 10 (F10), 100 (F100), or 200 (F200) nM cortisol, respectively. BRFF represents cells in BRFF-HPC1 medium throughout the experiment. DNA content and [3H]thymidine incorporation of control cells in test medium (control) at the end of 6 d were defined as 100% and were 3.4 ± 0.73 µg/well and 696 ± 133 dpm/well, respectively. DNA content and [3H]thymidine incorporation in other experimental groups are represented as percent of control and are given as mean ± SEM from 3–12 determinations. Values were significantly higher in all the cortisol-treated groups (P < 0.02–P < 0.0001) and in the BRFF group (P < 0.0001), compared with the control in test medium.

 
Triamcinolone does not stimulate MDA PCa 2b cell growth and PSA secretion
Figure 7AGo shows the effects of various steroids on MDA PCa 2b cell growth after a 6-d treatment in the test medium. Potent ARccr agonists cortisol and FluF caused significant stimulation of cell growth, compared with cells cultured in test medium alone (control). Corticosterone also produced a significant increase in cell growth. Importantly, triamcinolone, which did not bind to ARccr (Table 1Go) or elicit a transactivational response through ARccr (Figs. 3AGo and 4Go), did not stimulate cell growth (Fig. 7AGo). The changes in secreted levels of PSA parallel the growth data (Fig. 7BGo). Cortisol and FluF significantly increased PSA levels over control. The agonistic effect of FluF was more pronounced on PSA production than growth stimulation. Paradoxically, corticosterone did not affect PSA production, although it increased cell growth. Triamcinolone, which did not stimulate cell growth, also had no stimulatory effect on PSA secretion.



View larger version (22K):
[in this window]
[in a new window]
  		Figure 7. A, Effects of ARccr agonists on cell growth. MDA PCa 2b cells were plated and treated with ethanol vehicle (control) or various steroids (50 nM) in test medium for 6 d, at the end of which, DNA levels were determined as described in Materials and Methods. DNA concentration at the beginning of the experiment was 2.09 ± 0.22 µg/well. Values are given as mean ± SEM from four to six determinations. DNA concentrations, at the end of 6 d, in F-, FluF-, or B-treated cells were significantly higher than control cells, at P < 0.01, P < 0.05, and P < 0.01, respectively. B, Effects of ARccr agonists on PSA secretion. The experimental details are as given in Fig. 7AGo. PSA levels are shown as mean ± SEM from four to six determinations. PSA concentrations in F- and FluF-treated cells were significantly higher than control cells, at P < 0.01 and P < 0.001, respectively.

 
Casodex and RU486 inhibit MDA PCa cell growth and PSA secretion
In these experiments, the cells were cultured in BRFF-HPC1 medium, which contains a high concentration of cortisol (280 nM) and DHT, at 0.1 nM, that stimulates the cells to grow and secrete PSA. Addition of the antagonists casodex (10 µM) and RU486 (100 nM) to cells growing in this medium resulted in significant inhibition of cell growth (Fig. 8AGo). RU486, which exhibited higher RBA for the ARccr (Table 1Go), was more effective at inhibiting cell growth (~78% growth inhibition) than casodex (~34% growth inhibition). Similarly, PSA secretion by the cells, in response to cortisol and DHT present in the BRFF-HPC1 medium, was significantly inhibited by RU486 (~85%) and casodex (~50%) at the doses tested, as shown in Fig. 8BGo.



View larger version (22K):
[in this window]
[in a new window]
  		Figure 8. A, Effects of ARccr antagonists on cell growth. MDA PCa 2b cells were plated and grown in BRFF-HPC1 medium in the presence of ethanol vehicle (control) or casodex (10 µM) or RU486 (100 nM) for 6 d. At the end of the experimental period, DNA levels were determined as described in Materials and Methods. Values are given as mean ± SEM from three determinations. DNA concentrations in casodex- and RU486-treated cells were significantly lower than control cells, at P < 0.05 and P < 0.001, respectively. B, Effects of ARccr antagonists on PSA secretion. The experimental details are as given in Fig. 8AGo. PSA levels are given as mean ± SEM from three determinations. PSA concentrations in casodex- and RU486-treated cells were significantly lower than control cells, at P < 0.002 and P < 0.0001, respectively.

 
Docking of steroid ligands to ARccr
Our data on the AR LBD using the homology model agree with the observations using the crystal structure of the human AR LBD reported recently (27, 28). Both our homology model (Fig. 9Go) and the model based on the AR crystal structure of the LBD of ARccr (described in Ref. 6) revealed that both mutations in this mutant receptor are located in the ligand-binding pocket. The substitutions at residue 701 (histidine for leucine) and residue 877 (alanine for threonine) are located in helix 3 and helix 11, respectively. Both residues are within 10 Å (C{alpha}-C{alpha}) at one end of the elongated-binding pocket, forming pincers on either side of steroidal D ring. The positions of the mutant residues in the ARccr pocket are shown in Fig. 9Go (top image-middle panel) along with comparable residues in the ligand-binding pockets of the wild-type AR (top image-left panel) and GR{alpha} (top image-right panel). A series of steroids with known crystal structures were docked into the ARccr pocket to evaluate the steric limits of the binding cavity and to identify the environment of positions on the steroids where different substituents may influence binding affinity. Several general observations can be made. As suggested by homology to the binding of progesterone within the binding pocket of PR (26), the C17 position of the ligand docks near residues 701 and 877. Substitution of alanine for threonine at residue 877 is expected to increase the size of the binding pocket compared with wild-type AR (16). In contrast, substitution of histidine for leucine at residue 701 alters the hydrophilic nature of the pocket. This residue is located on the side of the planar steroid ligand opposite from residue 877.



View larger version (133K):
[in this window]
[in a new window]
  		Figure 9. Structural models of the ligand-binding pockets of nuclear receptors. In the upper images, close-up views of the interior ligand pocket are shown for AR (left), ARccr (middle), and GR{alpha} (right). The atoms are represented as CPK models with all residues gray except the two substitutions in the mutant ARccr and the corresponding residues in wild-type AR and GR{alpha}. Selected homologous residues are labeled for comparison of the three receptors. Note that the hydrophobic pockets are strikingly similar, yet differ in the detailed stereochemical composition. The ligand pocket is located in the interior of the LBD; and therefore, some residues between the observer and the pocket interior have been removed to provide the close-up view. In the lower images, three ligands were docked to the binding pocket of ARccr. The view into the pocket interior is slightly different than in the ARccr shown in the middle panel in the upper images. Ligands are represented by so-called stick models, with cortisol (left), DHT (middle), and progesterone (right) shown for comparison. These ligands bind to ARccr with varying affinity: cortisol > DHT > progesterone. The substituents at C17 in these ligands differ in size, yet each one is accommodated in the ARccr pocket. Substitutions at residue 877 (alanine) and 701 (histidine) are colored for identification, and positions on the steroids that are discussed in the text are labeled.

 
The orientation of these two substituted amino acids is indicated in Fig. 9Go (lower images), along with the results of docking studies for the three ligands cortisol (left), DHT (middle), and progesterone (right) into the ligand-binding pocket of the ARccr. Our data on the docking of DHT to the ARccr-binding pocket are in agreement with the model generated by Krystek and Sack (described in Ref. 6) of the ARccr bound to DHT, which shows extra space generated by the ARccr mutations and the lack of a hydrogen bond between the receptor and DHT because of the substitution of the threonine at position 877 by an alanine residue. The binding pocket in the ARccr is highly hydrophobic but also contains five polar residues. Residues Gln711 and Arg752 are located near the C3 carbonyl group of the steroid; His701 and Asn705 are close to the C17 substituent, and the hydrophobic part of Gln783 is near C15 and C16. The substitution at amino acid 701 results in the replacement of the hydrophobic leucine with a more hydrophilic histidine residue. Histidine701 is in position to form hydrogen bonds with polar substituents at C17 or C21 on the ligand. Cortisol has an {alpha}-hydroxyl group at position C17, whereas the natural AR ligands do not. An interaction between His701 and this hydroxyl group may stabilize the binding of cortisol and cortisone to ARccr. In addition, the bulkier C21 substituents on the corticosteroids are likely accommodated by the extra space resulting from the T877A substitution. It should be noted that the mutations in ARccr do not transform the AR-binding pocket into one that resembles the wild-type GR{alpha} pocket. For example, residues corresponding to His701 and Ala877 of the ARccr are Met560 and Cys736, respectively, in wild-type GR{alpha}. These residues cannot form hydrogen bonds with the C17 or C21 substituents of steroid ligands. If hydrogen bonds are required to stabilize the binding of glucocorticoids in GR{alpha}, other residues or bound water molecules may be involved.

The substituent on C16 of the ligand determines the binding properties of dexamethasone and triamcinolone to ARccr. Dexamethasone has a methyl substituent on C16, whereas triamcinolone has a hydroxyl group at the same position. Studies on the docking of dexamethasone to the ARccr-binding pocket (data not shown) reveal that the region of the pocket around C16 of the ligand is hydrophobic. The residues closest to the C16 methyl group are Met780, Phe876, and Leu704. The van der Waals volumes of the methyl and hydroxyl groups are similar, and the binding preferences are therefore most likely attributable to the difference in polarity of the substituents. The binding of dexamethasone is stabilized by interactions between the hydrophobic C16 methyl group and the hydrophobic binding residues of the receptor. In contrast, triamcinolone has a polar hydroxyl group at this position and does not allow its binding to the ARccr-binding pocket.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Note Added in Proof
 References
 
Our data indicate a novel mechanism for prostate cancer cells to become androgen-independent. Mutations (L701H and T877A) in the AR cause the receptor to become responsive to circulating corticosteroids (ARccr), which would then drive the growth of prostate cancer cells that harbor these mutations, even in the absence of androgens in androgen-ablated patients. This process is AR dependent but androgen independent. The frequency of the ARccr mutations or other AR mutations that confer corticosteroid responsiveness to prostate cancer patients remains to be determined, and we are currently pursuing this point. Because metastatic lesions are not usually biopsied, the number of samples thus far specifically examined for the ARccr mutation is limited. The T877A mutation is common (4, 11, 12, 35), and the L701H mutation has been reported twice (36, 37) in addition to our study (20). Our current study also shows that the L701H AR responds to corticosteroids such as cortisol and FluF, although to a much lesser degree than the double-mutant ARccr, and that it can be inhibited by the AR antagonist casodex. We speculate that these mutations may be selected for by the growth advantage they confer on the cancer cells in an androgen-ablated patient. The presence of a promiscuous AR responsive to nonandrogen ligands is an important pathway for the development of androgen-independent prostate cancer (4, 6, 35).

Our studies using Scatchard analyses and competition-binding assays have revealed that the ligand-binding profile of ARccr and the stereochemistry of its binding pocket are different from that of wild-type AR or GR{alpha}. The structural requirement of the compounds that bind and activate ARccr is surprisingly broad. The ARccr binds and responds to both sex steroids and glucocorticoids. Though the ARccr binds glucocorticoids, it is more permissive than the GR{alpha} at both a qualitative level (ligand-binding specificity) and a quantitative level (ligand-binding affinity). For example, inactive corticosteroids, such as cortisone and 11{alpha}-cortisol, can activate the ARccr.

The effect of individual functional groups of the ligand for binding to ARccr is demonstrated by the ratio of binding affinities (Table 1Go) of closely related ligand pairs. The largest effect is observed when an {alpha}-hydroxyl group is introduced at the C17 position. Minor conformational adjustments are neglected, because the steroid structure is rigid, as are changes in the electronic distribution in the ligand. This hydroxyl group is the only difference between 17-hydroxyprogesterone and progesterone, where binding of the former ligand is increased 10-fold. We note that the four strongest binding ligands [FluF, cortisol, cortisone, and 17-hydroxyprogesterone (see Table 1Go)] are 17 {alpha}-OH-substituted. The difference in binding affinity is likely attributable to the formation of a hydrogen bond between the C17 hydroxyl and the side chain of His701. In support of this hypothesis, cortisol does not activate either wild-type or the T877A receptor (20), ARs that lack a hydrogen bond acceptor close to C17.

The findings, using cultured MDA PCa 2b cells to evaluate steroid actions on growth and PSA production, confirm our data from the binding and transactivation assays using COS-7 or CV-1 cells expressing ARccr, except for the following: 1) The magnitudes of growth and PSA stimulation by agonists in MDA PCa 2b cells are less than those observed in the transactivation assay. This may be attributable to higher levels of ARccr produced in the transient transfection system than present in the MDA PCa cells, as well as the increased sensitivity of the luciferase reporter assay. Other factors, such as differences in the rate of steroid metabolism between cells, might also contribute to these differences in magnitudes. 2) Prednisone and prednisolone, which have lower affinities for binding to ARccr, were good agonists in transactivation assays. Binding assays were done in COS-7 cell extracts, whereas the transactivation assay uses CV-1 cells, and factors such as the stability of the steroids in cell culture may contribute to the observed differences in potency. 3) In general, stimulation of MDA PCa cell growth by agonists is also accompanied by increases in PSA secretion, showing that changes in PSA mirror changes in cell growth. However, corticosterone, which exhibited agonist activity in transactivation and growth assays, failed to increase PSA production. Also, FluF was a more potent inducer of PSA secretion than of cell growth. A possible explanation for the divergence is that ligand-mediated regulation of a single-target gene, such as PSA, may differ from generalized effects on growth that reflect a complex interplay of actions on multiple genes. Importantly, triamcinolone, which did not bind to or activate ARccr, also had no effect on cell growth or PSA levels.

It will be difficult to treat the subset of prostate cancer patients whose cancer cells harbor glucocorticoid-responsive ARs, such as ARccr or the L701H AR, because glucocorticoid ablation would not be a feasible approach to therapy, considering that these steroids are essential for survival. One of the surprising findings that came out of our steroid-screening experiment is that triamcinolone, a potent agonist for the GR{alpha}, is unable to bind to or activate the ARccr or the L701H AR. Dexamethasone and triamcinolone are potent synthetic glucocorticoids that have been used clinically, for many years, to treat a variety of diseases (38). These two steroids differ only by the nature of the substituent at position C16. In dexamethasone, the methyl group at this position is accommodated in a hydrophobic cavity within the ARccr ligand-binding pocket. Triamcinolone has a polar hydroxyl group at C16, and binding of this ligand to the ARccr would bury a hydroxyl group, an energetically unfavorable interaction.

This selectivity of triamcinolone for GR{alpha} vs. ARccr or L701H AR may be useful for the treatment of the subset of prostate cancer patients who harbor the L701H or ARccr type promiscuous mutations in the AR. Administration of triamcinolone to these patients would have two benefits: First, triamcinolone would preserve essential glucocorticoid activity. Second, by negative feedback loops, triamcinolone would suppress the hypothalamic-pituitary-adrenal axis, thereby diminishing or eliminating the endogenous production of adrenal steroids, including cortisol. The lack of circulating corticosteroids would result in the failure of stimulation of prostate cancer cell proliferation via ARccr. Because triamcinolone does not bind or activate the ARccr, it would not promote tumor growth. Thus, the replacement of cortisol with triamcinolone represents a possible strategy to block corticosteroid activation of ARccr. However, it should be emphasized that dexamethasone, a glucocorticoid that is sometimes used in cancer treatment approaches, would stimulate the growth of this subset of prostate cancers.

Our studies show that casodex functions as an antagonist for ARccr or the L701H AR, and perhaps it can be used as a template to develop better antagonists for these mutant receptor forms. Casodex, in combination with triamcinolone, could be an effective therapeutic approach. Triamcinolone would exert its effects at the level of circulating corticosteroid ligand concentrations, eliminating endogenous glucocorticoids that act as agonists through the mutant receptors. In addition, casodex would act at the receptor level by binding to and blocking its activation. RU486 is also an effective inhibitor of ARccr-mediated transactivation as well as cortisol stimulation of MDA PCa cell growth and PSA secretion. However, its therapeutic application to treat prostate cancer may be limited, because it is also a potent GR{alpha} antagonist (38) and may have partial agonist activity in some prostate cancer cells (39). As shown by our binding and functional studies, the ARccr has a ligand specificity distinct from GR{alpha}. It is therefore possible that improved antagonists specific for ARccr, without affecting GR{alpha} signaling, could be developed and used perhaps in conjunction with triamcinolone in the treatment of the subset of prostate cancer patients that harbor the ARccr, the L701H AR, or other promiscuous AR mutant forms activated by glucocorticoids.

In summary, we have demonstrated that ARccr is a promiscuous nuclear receptor with a broad ligand-binding spectrum. The presence of AR mutations, such as L701H and ARccr, provides a growth advantage to prostate cancer cells in vivo. The use of triamcinolone to suppress corticosteroid synthesis may provide effective therapy for these patients. Furthermore, in analogy to complete androgen blockade, the addition of receptor antagonists will be potentially useful to inhibit the proliferation of prostate cancer cells containing these mutant AR forms. Thus, the combination of ARccr receptor antagonists together would a ligand suppressor (triamcinolone) represents a new therapeutic strategy for the treatment of the subset of androgen-independent prostate cancers harboring the L701H or ARccr type of promiscuous mutations.


   Note Added in Proof
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Note Added in Proof
References
 
A recent publication by Chang et al. (40) showed expanded ligand-induced transactivation of the T877A mutant AR to include many corticosteroids.


   Acknowledgments
 
We thank Drs. Ronald Evans, Zoran Culig, Peter J. Malloy, and Peter Kushner for providing plasmids pMMTV-luc, pSG5-AR, pSG5-ARccr, and pSG5-GR{alpha}, respectively.


   Footnotes
 
This work was supported by NIH Grant DK-42482, Department of the Army Grant DAMD17-02-1-0142 (to D.F.), and awards from Cap CURE (to K.R.E. and D.M.P.). L.B. was supported by a postdoctoral fellowship from the Human Frontier Science Program.

1 Current address: Berlex Biosciences, Richmond, California 94804. Back

2 Current address: The Scripps Research Institute, La Jolla, California 92037. Back

Abbreviations: ARccr, Cortisol/cortisone-responsive AR; BRFF, Biological Research Faculty and Facility; DHT, dihydrotestosterone; FluF, 9{alpha}-fluorocortisol; Kd, dissociation constant; LBD, ligand-binding domain; MMTV, mouse mammary tumor virus; PSA, prostate-specific antigen; RBA, relative binding affinity; RU486, GR antagonist RU38486.

Received September 26, 2001.

Accepted for publication January 14, 2002.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Note Added in Proof
 References
 

  1. Brinkmann AO, Blok LJ, de Ruiter PE, Doesburg P, Steketee K, Berrevoets CA, Trapman J 1999 Mechanisms of androgen receptor activation and function. J Steroid Biochem Mol Biol 69:307–313[CrossRef][Medline]
  2. Isaacs JT 1999 The biology of hormone refractory prostate cancer. Why does it develop? Urol Clin North Am 26:263–273[CrossRef][Medline]
  3. Culig Z, Hobisch A, Bartsch G, Klocker H 2000 Androgen receptor—an update of mechanisms of action in prostate cancer. Urol Res 28:211–219[CrossRef][Medline]
  4. Marcelli M, Ittmann M, Mariani S, Sutherland R, Nigam R, Murthy L, Zhao Y, DiConcini D, Puxeddu E, Esen A, Eastham J, Weigel NL, Lamb DJ 2000 Androgen receptor mutations in prostate cancer. Cancer Res 60:944–949[Abstract/Free Full Text]
  5. Buchanan G, Greenberg NM, Scher HI, Harris JM, Marshall VR, Tilley WD 2001 Collocation of androgen receptor gene mutations in prostate cancer. Clin Cancer Res 7:1273–1281[Abstract/Free Full Text]
  6. Feldman BJ, Feldman D 2001 The development of androgen-independent prostate cancer. Nature Reviews/Cancer 1:34–45
  7. Taplin ME, Ho S-M 2001 Clinical Review 134. The endocrinology of prostate cancer. J Clin Endocrinol Metab 86:3467–3477[Free Full Text]
  8. Visakorpi T, Hyytinen E, Koivisto P, Tanner M, Keinanen R, Palmberg C, Palotie A, Tammela T, Isola J, Kallioniemi OP 1995 In vivo amplification of the androgen receptor gene and progression of human prostate cancer. Nat Genet 9:401–406[CrossRef][Medline]
  9. Koivisto P, Kononen J, Palmberg C, Tammela T, Hyytinen E, Isola J, Trapman J, Cleutjens K, Noordzij A, Visakorpi T, Kallioniemi O 1997 Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation therapy failure in prostate cancer. Cancer Res 57:314–319[Abstract/Free Full Text]
  10. Culig Z, Hobisch A, Cronauer MV, Cato AC, Hittmair A, Radmayr C, Eberle J, Bartsch G, Klocker H 1993 Mutant androgen receptor detected in an advanced-stage prostatic carcinoma is activated by adrenal androgens and progesterone. Mol Endocrinol 7:1541–1550[Abstract]
  11. Gaddipati J, McLeod D, Heidenberg H, Sesterhenn I, Finger M, Moul J, Srivastava S 1994 Frequent detection of codon 877 mutation in the androgen receptor gene in advanced prostate cancers. Cancer Res 54:2861–2866[Abstract/Free Full Text]
  12. Taplin ME, Bubley GJ, Shuster TD, Frantz ME, Spooner AE, Ogata GK, Keer HN, Balk SP 1995 Mutation of the androgen-receptor gene in metastatic androgen-independent prostate cancer. N Engl J Med 332:1393–1398[Abstract/Free Full Text]
  13. Tilley WD, Buchanan G, Hickey TE, Bentel JM 1996 Mutations in the androgen receptor gene are associated with progression of human prostate cancer to androgen independence. Clin Cancer Res 2:277–285[Abstract/Free Full Text]
  14. Culig Z, Stober J, Gast A, Peterziel H, Hobisch A, Radmayr C, Hittmair A, Bartsch G, Cato AC, Klocker H 1996 Activation of two mutant androgen receptors from human prostatic carcinoma by adrenal androgens and metabolic derivatives of testosterone. Cancer Detect Prev 20:68–75[Medline]
  15. Veldscholte J, Voorhorst-Ogink MM, Vries JB-D, van Rooij HCG, Trapman J, Mulder E 1990 Unusual specificity of the androgen receptor in the human prostate tumor cell line LNCaP: high affinity for progestagenic and estrogenic steroids. Biochim Biophys Acta 1052:187–194[Medline]
  16. McDonald S, Brive L, Agus DB, Scher HI, Ely KR 2000 Ligand responsiveness in human prostate cancer: structural analysis of mutant androgen receptors from LNCaP and CWR22 tumors. Cancer Res 60:2317–2322[Abstract/Free Full Text]
  17. Griffin JE, Wilson JD 1998 Disorders of the testis and male reproductive tract. In: Wilson JD, Foster DW, Kronenberg HM, Larsen PR, eds. Williams textbook of endocrinology. Philadelphia: W. B. Saunders; 819–876
  18. Navone NM, Olive M, Ozen M, Davis R, Troncoso P, Tu S-M 1997 Establishment of two human prostate cancer cell lines derived from a single bone metastasis. Clin Cancer Res 3:2493–2500[Abstract/Free Full Text]
  19. Zhao XY, Boyle B, Krishnan AV, Navone NM, Peehl DM, Feldman D 1999 Two mutations identified in the androgen receptor of the new human prostate cancer cell line MDA PCa 2a. J Urol 162:2192–2199[CrossRef][Medline]
  20. Zhao XY, Malloy PJ, Krishnan AV, Swami S, Navone NM, Peehl DM, Feldman D 2000 Glucocorticoids can promote androgen-independent growth of prostate cancer cells through a mutated androgen receptor. Nat Med 6:703–706[CrossRef][Medline]
  21. Brinkmann AO, Trapman J 2000 Prostate cancer schemes for androgen escape. Nat Med 6:628–629[CrossRef][Medline]
  22. Zhao XY, Ly LH, Peehl DM, Feldman D 1997 1{alpha},25-Dihydroxyvitamin D3 actions in LNCaP human prostate cancer cells are androgen-dependent. Endocrinology 138:3290–3298[Abstract/Free Full Text]
  23. Zhao XY, Ly LH, Peehl DM, Feldman D 1999 Induction of androgen receptor by 1{alpha},25-dihydroxyvitamin D3 and 9-cis retinoic acid in LNCaP human prostate cancer cells. Endocrinology 140:1205–1212[Abstract/Free Full Text]
  24. Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anal Biochem 72:248–254[CrossRef][Medline]
  25. Krishnan AV, Feldman D 1991 Activation of protein kinase-C inhibits vitamin D receptor gene expression. Mol Endocrinol 5:605–612[Abstract]
  26. Williams SP, Sigler PB 1998 Atomic structure of progesterone complexed with its receptor. Nature 393:392–396[CrossRef][Medline]
  27. Matias PM, Donner P, Coelho R, Thomaz M, Peixoto C, Macedo S, Otto N, Joschko S, Scholz P, Wegg A, Basler S, Schafer M, Egner U, Carrondo MA 2000 Structural evidence for ligand specificity in the binding domain of the human androgen receptor. Implications for pathogenic gene mutations. J Biol Chem 275:26164–26171[Abstract/Free Full Text]
  28. Sack JS, Kish KF, Wang C, Attar RM, Kiefer SE, An Y, Wu GY, Scheffler JE, Salvati ME, Krystek Jr SR, Weinmann R, Einspahr HM 2001 Crystallographic structures of the ligand-binding domains of the androgen receptor and its T877A mutant complexed with the natural agonist dihydrotestosterone. Proc Natl Acad Sci USA 98:4904–4909[Abstract/Free Full Text]
  29. Dunbrack RLJ, Karplus M 1993 Back-bone dependent rotamer library for proteins. Application to side-chain prediction. J Mol Biol 230:543–574[CrossRef][Medline]
  30. Sali A, Blundell TL 1993 Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234:779–815[CrossRef][Medline]
  31. Brooks BR, Bruccoleri RE, Olafson BD, States JD, Swaminathan S, Karplus M 1983 CHARMm: a program for macromolecular energy, minimization, and dynamics calculations. J Comp Chem 4:187–217
  32. Allen FH, Bellard S, Brice MD, Cartwright BA, Doubleday A, Higgs H, Hummelink T, Hummelink-Peters BG, Kennard O, Motherwell WDS, Rodgers JR, Watson DG 1979 Cambridge crystallographic data center: computer-based search, retrieval, analysis and display of information. Acta Crystallogr B 35:2231–2339
  33. Warriar N, Page N, Govindan MV 1994 Transcriptional activation of mouse mammary tumor virus-chloramphenicol acetyl transferase: a model to study the metabolism of cortisol. Biochemistry 33:12837–12843[CrossRef][Medline]
  34. Yen PM, Liu Y, Palvimo JJ, Trifiro M, Whang J, Pinsky L, Janne OA, Chin WW 1997 Mutant and wild-type androgen receptors exhibit cross-talk on androgen-, glucocorticoid-, and progesterone-mediated transcription. Mol Endocrinol 11:162–171[Abstract/Free Full Text]
  35. Taplin ME, Bubley GJ, Ko YJ, Small EJ, Upton M, Rajeshkumar B, Balk SP 1999 Selection for androgen receptor mutations in prostate cancers treated with androgen antagonist. Cancer Res 59:2511–2515[Abstract/Free Full Text]
  36. Suzuki H, Sato N, Watabe Y, Masai M, Seino S, Shimazaki J 1993 Androgen receptor gene mutations in human prostate cancer. J Steroid Biochem Mol Biol 46:759–765[CrossRef][Medline]
  37. Watanabe M, Ushijima T, Shiraishi T, Yatani R, Shimazaki J, Kotake T, Sugimura T, Nagao M 1997 Genetic alterations of androgen receptor gene in Japanese human prostate cancer. Jpn J Clin Oncol 27:389–393[Abstract/Free Full Text]
  38. Orth DN, Kovacs WJ 1998 The adrenal cortex. In: Wilson JD, Foster DW, Kronenberg HM, Larsen PR, eds. Williams textbook of endocrinology. Philadelphia: W. B. Saunders; 517–664
  39. Terouanne B, Tahiri B, Georget V, Belon C, Poujol N, Avances C, Orio FJ, Balaguer P, Sultan C 2000 A stable prostatic bioluminescent cell line to investigate androgen and antiandrogen effects. Mol Cell Endocrinol 160:39–49[CrossRef][Medline]
  40. Chang C-Y, Walther PJ, McDonnell DP 2001 Glucocorticoids manifest androgenic activity in a cell line derived from a metastatic prostate cancer. Cancer Res 61:8712–8717[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Res.Home page
A. Yemelyanov, J. Czwornog, L. Gera, S. Joshi, R. T. Chatterton Jr., and I. Budunova
Novel Steroid Receptor Phyto-Modulator Compound A Inhibits Growth and Survival of Prostate Cancer Cells
Cancer Res., June&nb