The Effect of F877L and T878A Mutations on Androgen Receptor

Home , Hydroxyflutamide, Mibolerone, Nilutamide

Published OnlineFirst May 16, 2016; DOI: 10.1158/1535-7163.MCT-15-0892

Cancer Biology and Signal Transduction Molecular Cancer Therapeutics The Effect of F877L and T878A Mutations on Androgen Receptor Response to Enzalutamide Stefan Prekovic1, Martin E.van Royen2, Arnout R.D.Voet3,4, Bart Geverts2, Rene Houtman5, Diana Melchers5, Kam Y.J. Zhang3, Thomas Van den Broeck1,6, Elien Smeets1, Lien Spans7, Adriaan B. Houtsmuller2,8, Steven Joniau6, Frank Claessens1, and Christine Helsen1

Abstract

Treatment-induced mutations in the ligand-binding domain that lead to a decrease in steric clashes for enzalutamide. of the androgen receptor (AR) are known to change antagonists Ligand-binding assays conﬁrmed that the F877L mutation into agonists. Recently, the F877L mutation has been described leads to an increase in relative binding afﬁnity for enzaluta- to convert enzalutamide into an agonist. This mutation was mide, but only the combination with the T878A mutation seen to co-occur in the endogenous AR allele of LNCaP cells, resulted in a strong agonistic activity. This correlated with next to the T878A mutation. Here, we studied the effects of changes in coregulator recruitment and chromatin interactions. enzalutamide on the F877L and T878A mutants, as well as Our data show that enzalutamide is only a very weak partial the double-mutant AR (F877L/T878A). Molecular modeling agonist of the AR F877L, and a strong partial agonist of the revealed favorable structural changes in the double-mutant AR double-mutant AR. Mol Cancer Ther; 15(7); 1702–12. 2016 AACR.

Introduction chemotherapy by a median of 4.8 months in comparison with the placebo group (5). On the other hand, enzalutamide has no Globally, prostate cancer is the fourth most common cancer beneficial effect in a portion (20%) of patients, while the and the fifth leading cause of cancer-related death in men (1). As majority of patients who initially respond will eventually develop this cancer is hormone-dependent, the blockade of the androgen resistance towards this therapy (6). It is imperative, therefore, to receptor (AR) signaling is an effective therapeutic strategy for men study the different mechanisms that can lead to enzalutamide with advanced metastatic prostate cancer. The discovery of novel resistance. compounds that inhibit the AR and their clinical application has Prior studies of resistance against the earlier generation anti- led to the improvement in survival time. One such compound is androgens, bicalutamide (Bic) and hydroxyflutamide (Hof), led the potent antiandrogen enzalutamide, which has been approved to the discovery of mutations that changed the ligand-binding under the name Xtandi for use in postchemotherapy patients with characteristics of the AR and thus lead to the change from an metastatic castration-resistant prostate cancer (mCRPC; refs. 2, 3). antagonistic to an agonistic behavior (7, 8). Balbas and colleagues Enzalutamide acts as a direct competitor for dihydrotestosterone transduced LNCaP cells (which carry the T878A mutation) with (DHT) binding to the ligand-binding pocket of the AR. In addi- constructs driving overexpression of a library of mutated AR and tion, it reduces the binding of the AR to DNA, and inhibits the were able to show that the F877L mutation converts enzalutamide recruitment of AR coactivators (4). Importantly, enzalutamide to an agonist (9). In addition, LNCaP xenograft tumors which significantly prolonged the survival of men with mCRPC after express the AR F877L mutant show sustained growth in castrated mice treated with enzalutamide (10). Another study in LNCaP cells pointed out that the F877L mutation had arisen in the 1Molecular Endocrinology Laboratory, Department of Cellular and endogenous AR T878A allele when cells were treated with enza- Molecular Medicine, KU Leuven, Leuven, Belgium. 2Department of lutamide for a prolonged period of time (10). It should be noted Pathology, Erasmus MC, Rotterdam, the Netherlands. 3Structural Bio- informatics Team, Division of Structural and Synthetic Biology, Center that the LNCaP cells are also mutated in the mismatch repair gene for Life Science Technologies, RIKEN, Yokohama, Japan. 4Laboratory (MSH2), along with over 1,800 other mutations and have a for Biomolecular Modeling and Design, Department of Chemistry, KU relatively unstable genome (11). However, the detection of very 5 Leuven, Leuven, Belgium. Pamgene International, the Netherlands. low levels of the same mutation in the circulating tumor DNA of 6Department of Urology, University Hospitals Leuven, Leuven, Bel- gium. 7Laboratory for Genetics of Malignant Disorders, Department of patients who were treated with ARN-509 (ARN) spiked a high Human Genetics, KU Leuven, Leuven, Belgium. 8Erasmus Optical clinical interest in the F877L mutation (12). Imaging Center, Erasmus MC, Rotterdam, the Netherlands. Several studies reported the co-occurrence of the F877L and Note: Supplementary data for this article are available at Molecular Cancer the T878A mutations (F877L/T878A, hereinafter referred to as Therapeutics Online (http://mct.aacrjournals.org/). the double-mutant AR) in cell line models (10, 12, 13). Corresponding Author: Frank Claessens, Molecular Endocrinology Laboratory, Interestingly, this double-mutant AR was described in a patient KU Leuven, Campus Gasthuisberg O&N1, Herestraat 49 Box 901, Leuven 3000, who progressed on enzalutamide treatment (14). Here, we Belgium. Phone: 321-633-0253; Fax: 321-6343-0735; E-mail: wanted to compare the agonistic activity of enzalutamide on [email protected] the AR F877L, as well as on the double-mutant AR. We dis- doi: 10.1158/1535-7163.MCT-15-0892 covered that the LNCaP-specific T878A mutation modulates the 2016 American Association for Cancer Research. F877L responses to enzalutamide.

1702 Mol Cancer Ther; 15(7) July 2016

Molecular Analysis of Androgen Receptor Mutants

Materials and Methods structure was taken from PDB entry 4OJB, and resistance mutations were introduced. Subsequently, the structure was Plasmid constructs optimized using Protonate_3D and the Amber99 force field Expression vectors, pSG5 and pEGFP-C1, containing cDNA of embedded in MOE (21). Conformations of the AR ligands wild-type flag- or EGFP-tagged human AR, respectively, (15) were (DHT, enzalutamide, ARN, and RD162) were calculated using used to generate the mutations in the ligand-binding domain (LBD) the MMFF94x force field. As a reference for the correct torsional by two-step PCR site-directed mutagenesis. The resulting PCR angle between the 3-fluoro-cyano-benzyl group and the thiox- fragments were ligated into pGEM-T Easy (Promega) and recloned oimidazolidin, PDB entry 2NW4 was utilized (22). The top five as AsuII/BglII fragments into the pSG5-flag-WT AR construct or the binding modes for all the receptor–ligand pairs were visually pEGFP-C1-WT AR construct. The WT DBD/LBD fragment of AR inspected and exhibited nearly identical binding properties inserted into pSG5 backbone (16) was used to introduce mutations with the exception of enzalutamide fitted in the WT and the via two-step PCR site–directed mutagenesis. The same cloning T878A AR LBD. As a positive control DHT and Hof were docked strategy as described for the full-length AR constructs was used. into the agonistic conformation of AR WT or T878A LBD, The construct containing double-tagged (N-ter YFP and C-ter CFP) respectively. The top docking scores exhibited a similar docking WT AR (17) has been used to generate mutations in the LBD using mode within 0.4 Å RMSD deviation of the crystallographically QuikChange II kit (Agilent Technologies). A plasmid containing determined structure. pCMV-b-Gal was acquired from Stratagene, the luciferase reporter The ligands were docked into the ligand binding pocket of the construct driven by 4 copies of a classical ARE and the VP16 AD- receptor followed by a short energy minimization allowing minor NTD(WT) expression vector have been described elsewhere "induced fit" changes to the protein and ligand conformation. The (16, 18). All expression vectors were sequence verified. binding mode was scored using the London Free Energy scoring fi Cell culture function for prediction of af nity and rescored using the DrugS- Hep3B and LAPC4 were provided by the A.B. Houtsmuller and coreX scoring function for analysis of the complementarity and colleagues (Erasmus Medical Center, Rotterdam, the Nether- clashes between the ligand and the receptor atoms (23). lands) and the Laboratory for Experimental Medicine and Endo- crinology, respectively, whereas PC-3 and COS-7 cells were Whole cell competition assay obtained from the ATCC and were authenticated by short tandem HEK293T cells were seeded at a density of 30,000 cells per well repeat DNA profiling by Genetica. HEK293T cell line was a kind in 48-well plates. Cells were transfected with 375 ng of AR b gift from the Laboratory of Biosignaling and Therapeutics (KU expression constructs, and 75 ng of pCMV- -Gal expression Leuven) in 2004. HEK293T, PC-3, and COS-7 cells were cultured vector. On the day after the transfections, medium was refreshed in DMEM supplemented with 10% FCS, while Hep3B cells were and cells were incubated overnight. Subsequently, cells were m cultured in MEMa with GlutaMAX supplement and without treated with a dilution series (0.1 nmol/L to 10 mol/L) of compound with 1 nmol/L [3H]-labeled mibolerone (Perkin nucleosides (Gibco). LAPC4 cells were cultured in Iscove's Mod- ified Dulbecco's Medium (Gibco) supplemented with 10% FCS. Elmer). After an incubation period of 90 minutes at 37 C, medium was aspirated, cells were washed three times with ice- Androgen receptor transactivation assay cold PBS, and lysed in 100 mL of Passive Lysis buffer (Promega). HEK293T or PC-3 cells were seeded at a density of 15,000 cells per Cell lysates (75 mL) were transferred to scintillation vials and 2 mL well or 10,000 cells per well, respectively, in 96-well plates. Cells were of LumaSafe Plus (PerkinElmer) was added. Radioactivity mea- transfected with 10 ng of expression vector of the receptor (wild-type surements were done in a liquid scintillation counter (LK Wallac or mutant), 100 ng of reporter construct, and 10 ng of the pCMV- 1216 RackBeta, Wallace). b-Gal expression vector per well. On the following day, medium was aspirated and fresh medium without or with compounds was Western blot analysis added. Cells were stimulated with DHT (0.1 nmol/L and 1 nmol/L) Total protein extracts (in Passive Lysis buffer) were run on a obtained from Sigma-Aldrich; enzalutamide (Sequoia Research Pro- precast NuPAGE novex 4%–12% Bis-Tris Gel (Invitrogen). After ducts Ltd, Selleckchem and Medivation, Inc.; kindly provided by electrophoresis, the proteins were blotted on a polyvinylidene Dr. Ganesh Raj, UT Southwestern Medical Center, Urology Clinic, difluoride transfer membrane (GE Healthcare). Antibodies Dallas, TX), ARN509 (ARN; MedChemExpress), Bicalutamide (Astra- against GAPDH (Santa Cruz Biotechnology) and AR (in-house Zeneca), hydroxyflutamide (Schering-Plough), abiraterone acetate antibody against N-terminus of AR as described in ref. 15), were (Sequoia Research Products Ltd), Galaterone (Sequoia Research used. The secondary antibodies anti-rabbit IgG (P0217 from Products Ltd), AZD3514 (AZD; MedChem Express), ASC-J9 (Med- Dako) and anti-Mouse IgG (P0260 from Dako) were conjugated Chem Express), EPI-001 (Epi; a kind gift from Dr. Marianne Sadar, to horseradish peroxidase, and visualized by immunodetection Department of Pathology, Laboratory Medicine, University of British combined with chemiluminescence (Western Lightning Plus- Columbia, Vancouver, British Columbia, Canada), and ARAN com- ECL; PerkinElmer). pounds [Asinex (http://www.asinex.com) and Specs] were used at concentration of 10 mmol/L. All the compounds were dissolved in Double-hybrid assay DMSO (Acros Organics). Cells were harvested in Passive Lysis buffer COS-7 cells were seeded at a density of 30,000 cells per well in (Promega)andtheluciferaseandb-galactosidase measurements 48-well plates. On the following day, cells were transfected with were done according to the protocol described previously (19). 25 ng of AR DBD/LBD expression vector with or without the mutations, 250 ng of the VP16 AD-NTD(WT) construct, 250 ng of Modeling of the androgen receptor ligand interaction the reporter construct, and 50 ng of pCMV-b-Gal construct. Cells Modeling was done using the Molecular Operating Environ- were stimulated with either DHT (10 nmol/L) or enzalutamide ment (MOE, The Chemical Computing Group; ref. 20). The AR (10 mmol/L), overnight. Cells were harvested in Passive Lysis

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1703

Prekovic et al.

buffer (Promega) and the luciferase and b-galactosidase measure- intensity before bleaching. The FRAP data were quantitatively ments were done as indicated above. analyzed by comparing the experimental data to curves generated using a previously described Monte Carlo approach (28). Acceptor photobleaching fluorescence resonance energy In short, a large set of computer-simulated FRAP curves with a transfer 3-population model was generated, containing a diffusing Acceptor photobleaching fluorescence resonance energy trans- fraction (ranging 1 to 3 m2/s) and two bound (immobile) fer (FRET) was performed using a Zeiss LSM510 confocal laser fractions (ranging from 0.65 to 0.8 s and 3 to 15 s, respectively). scanning microscope equipped with a Plan-Neofluar 40/1.3 NA Ranges in the simulation are based on quantitative FRAP oil objective (Carl Zeiss) at a lateral resolution of 100 nm (17). For analysis of AR curves in previous work (29). The top 10 excitation of CFP and YFP, an Argon laser at 358 and 514 nm was simulated curves that are fitting best to the experimental curve used. Both the CFP and YFP images were taken prior to photo- (by ordinary least squares) were averaged and provide the bleaching of the acceptor fluorophore (514 nm at high laser properties of the experimental data. power). Apparent FRET efficiency was estimated according to the instructions described in ref. (17). The abFRET datasets were tested for normality using the Kolmogorov–Smirnov tests, and Results the datasets were compared using the one-tailed Student t test. T878A enhances the partial agonistic effects of enzalutamide on AR F877L MARCoNI assay To unravel possible interferences between the F877L and the To investigate the modulation of coregulator interaction with T878A mutations, we first compared the response of the AR WT, the WT AR and mutant receptors, we used the MARCoNI assay AR F877L, AR T878A, and the AR double-mutant F877L/T878A (Pamgene). HEK293T cells were seeded at a density of 20 million to several compounds on a transactivation assay. Reporter cells per T175 flask. On the following day, cells were transfected assays were performed in HEK293T and PC-3 cells (Fig. 1 and with 85 mg of AR WT, AR F877L, AR T878A, or AR double-mutant Supplementary Fig. S1). The mutations had some effect on the expression constructs. After 24 hours, cells were incubated in maximal DHT induction: 1 nmol/L DHT resulted in induction serum-free medium without or with compounds (DHT, Hof, and factors of approximately 82, 71, 28, and 22 respectively for AR enzalutamide) for 30 minutes. Subsequently, cells were scraped WT, AR T878A, AR F877L, and the double mutant (Supple- and pelleted. The MARCoNI assay was performed with cell lysates mentary Fig. S2). To allow comparisons, the luciferase response as described previously (24). Each sample was processed on three to 1 nmol/L DHT was put at 100. As expected, the DHT-induced separate arrays. Subsequently, the binding of AR to each individ- transactivation of AR WT was inhibited by the antagonists, ual coregulator peptide was quantified. The FQNLF motif of the enzalutamide, ARN, bicalutamide, and Hof, to at least 25%. AR was discarded as the antibody to detect the AR was raised These effects were independent of the receptor levels (Supple- against a peptide that partly covered this motif. For each peptide, mentary Fig. S3). As expected, none of the antagonists had Student t test was applied to compare the receptor binding level to robust agonistic activity on WT AR. The induction of WT AR by that of the nonstimulated control. All calculations and statistics 10 mmol/L Hof rose to 6% of the 1 nmol/L DHT response; were performed using R (version 3.1.2; ref. 25). (Fig. 1A) which was comparable to the signal observed for castrate levels of DHT (0.1 nmol/L). Thus, for AR WT, enzalu- Generation of stable cell lines tamide, ARN, and bicalutamide are full antagonists, while Stable Hep3B cells were generated to constitutively express the hydroxyflutamide is a weak partial agonist. EGFP-tagged AR with either the F877L mutation or the F877L/ For the LNCaP AR T878A mutant, hydroxyflutamide gained T878A double mutation. The expression of EGFP-fusions was full agonistic activity (98% of DHT 1 nmol/L signal) and lost visualized by fluorescence microscopy and functionality of the its antagonistic properties (Fig. 1B), as reported in earlier receptor was confirmed by transfection of an ARE-driven lucifer- studies (30). The other compounds (enzalutamide, ARN, and ase reporter construct. Clones with moderate induction factor bicalutamide) efficiently antagonize AR T878A. When it comes (F877L clone: 13.3 5.5; F877L/T878A clone: 20.4 11.8) after to the AR F877L mutant, enzalutamide and ARN led to an DHT stimulation were selected (Supplementary Fig. S12). Hep3B increase in luciferase signal to approximately 10% and 14% of cell lines expressing EGFP-fused AR WT, AR T878A, or AR A574D the signal induced by 1 nmol/L DHT, respectively (Fig. 1C and were reported previously (26). Supplementary Fig. S4). It should be noted that the level of transactivation induced by these compounds was also similar Fluorescence recovery after photobleaching to that of castrate levels of DHT (0.1 nmol/L). Importantly, Cells were seeded in 6-well plates on a cover glass, ligands both enzalutamide and ARN retain their antagonistic effects on were added, and cells were incubated overnight. For each DHT-activated AR F877L. In our assays, enzalutamide and ARN experiment, a cover glass with stably transfected Hep3B cells are only weak partial agonists on AR F877L (Fig. 1C and was placed in a preheated ring. Cells were kept at 37Candused Supplementary Fig. S4), comparable with the report by Joseph for no longer than 90 minutes. A Zeiss LSM510 META confocal and colleagues (12). In case of the double mutant, however, laser scanning microscope equipped with a 40/1.3NA oil- enzalutamide and ARN were able to activate the receptor to a immersion objective, an argon laser (30 mW) and an acousto- substantial extent (33% and 34% of the response to 1 nmol/L optic tunable filter was used. Fluorescence recovery after photo- DHT, respectively) and exhibited weaker antagonistic proper- bleaching (FRAP) analysis was performed according to the ties than on AR WT, AR F877L, and AR T878A (Fig. 1D and instructions by van Royen and colleagues (27). For each treat- Supplementary Fig. S4). In other words, enzalutamide and ARN ment group, 30 cells were measured by FRAP in two indepen- are considerably stronger agonists of the double-mutant AR dent experiments. All curves were normalized to fluorescent than of the AR F877L. Surprisingly, the combination of the

1704 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Molecular Analysis of Androgen Receptor Mutants

Figure 1. T878A increases the partial agonistic effects of enzalutamide (Enza) on AR F877L. Transient transfection of AR WT or AR mutant together with an androgen- regulated luciferase reporter gene in HEK293T was performed to investigate the effect of mutations on the transactivational response to selected compounds (conc. 10 mmol/L). DHT was used in two different concentrations (1 nmol/L and 0.1 nmol/L) and all the values were normalized to DHT 1 nmol/L (100%). The mean and SEM of ﬁve independent experiments in triplicates are shown. AR WT (A), AR T878A (B), AR F877L (C), AR F877L/T878A (D), and Western blot analysis (E) showing equal expression levels of all receptor constructs with or without 1 nmol/L DHT. Hof, hydroxyﬂutamide; Bic, bicalutamide.

F877L mutation with the T878A mutation impaired the full Modeling the effect of the F877L and T878A mutations on the agonist activity of hydroxyﬂutamide (compare Fig. 1D with B). AR-enzalutamide interactions These effects are independent of AR levels and of the cell line To get an insight into the mechanisms how the T878A and the used (Fig. 1E and Supplementary Fig. S1). F877L mutations would convert the antagonist enzalutamide into

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1705

Prekovic et al.

Figure 2. The double mutant enables an optimal agonistic conformation for binding of enzalutamide (Enza). A, in silico models for enzalutamide interaction with the LBD; enzalutamide (contacting the side chain of R752) was docked into the agonistic conformation of the ligand binding pocket of WT (B), F877L (C), and the double mutant (D).

an agonist, we modeled the latter into the agonistic conformation mutant AR compared with the WT AR (Supplementary Fig. S7) of the AR ligand-binding pocket (Fig. 2A). In this agonistic and the T878A mutant (data not shown). conformation, the interactions of testosterone (T) or DHT with the R753 side chain are crucial (31, 32). However, if a similar The F877L mutation leads to an increase in the affinity of the interaction between R753 and enzalutamide is preserved, the receptor for enzalutamide latter cannot be accommodated in the agonistic conformation Whole cell competition assays with 1 nmol/L 3H-labeled of AR WT or AR T878A. The top scoring orientation for enzalu- mibolerone, were performed to investigate whether binding tamide in the WT structure significantly deviated from the binding affinities of the mutated receptors for DHT and enzalutamide mode observed in the mutant or protein in the modeled antag- have changed. A decrease was observed in relative binding affin- onistic conformation (33) with an 8 Å RMSD deviation and a non- ities (RBA) of DHT for both mutant receptors tested when com- negative predicted binding energy of 54 kcal/mol. Similarly, pared with the WT receptor [RBA(DHT)wt ¼ 100, RBA(DHT)877 ¼ oriented and bad scoring binding modes were observed in the 75.81, RBA(DHT)877/878 ¼ 77.51; Fig. 3A; Table 2]. On the other top solutions for both ARN and RD162. hand, there was a drastic increase in RBA of enzalutamide for both Indeed, there is a strong steric hindrance of enzalutamide with AR F877L and AR F877L/T878A when compared with the AR WT the phenylalanine at position 877, as well as with the threonine [RBA(Enza)wt ¼ 1.38, RBA(Enza)877 ¼ 12.77, RBA(Enza)877/878 ¼ residue at position 878 (Fig. 2B). The F877L mutation resolves 19.28; Fig. 3A; Table 2]. most of this, as leucine has a more flexible side chain compared with phenylalanine. This leucine can adopt a rotamer position Effect of mutations on the ligand-dependent N/C interactions without steric hindrance for enzalutamide, thus allowing the AR Upon binding of an agonist, the C-terminal LBD of the AR is LBD to assume an agonistic conformation (Fig. 2C). This mech- known to interact with the FQNLF motif in the N-terminal anism clearly differs from the mechanisms of agonism of the domain (35). To measure these N/C interactions, FRET and flexible bicalutamide, which fits as an agonist into a nearby cavity mammalian double-hybrid assay were performed (Fig. 3B and of the AR W741C/L mutants, thus stabilizing the agonistic con- Supplementary Fig. S8). DHT induced N/C interactions in both formation of the AR LBD in an alternative way (34). Moreover, the FRET and the mammalian double-hybrid assays (35), whereas modeling enzalutamide into the ligand-binding pocket of the no AR N/C interactions were observed for enzalutamide-bound double-mutant AR predicted even fewer steric clashes, as well as AR WT. While there was only a trend towards N/C interactions for a moderate improvement in docking scores compared with the single F877L mutant (Table 1; Supplementary Fig. S5). Similar results were obtained for ARN and RD162 (Table 1). Table 1. Docking scores of enzalutamide (Enza), ARN-509, and RD162 Modeling of hydroxyflutamide into the agonistic conformation Docking scores (kcal/mol) WT F877L F877L/T878A of the ligand-binding pocket of the double-mutant AR suggested Enza / 10.4 15.1 less favorable contacts between Leu877 and the dimethyl group ARN / 10.8 15.9 of hydroxyflutamide (Supplementary Fig. S6; Supplementary RD / 10.1 14.8 fi Table S1), predicting a less ef cient AR activation. NOTE: Docking simulations were used to predict the affinity of the compounds Modeling of DHT into the WT and mutant AR also indicated (kcal/mol) according to the top solutions for the different AR mutants. Decrease slightly reduced docking scores for the F877L and the double- of the docking score is in relation increased stability of the receptor–ligand pair.

1706 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Molecular Analysis of Androgen Receptor Mutants

Figure 3. Ligand-binding affinities and N/C interactions. A, experimental binding affinities as obtained in a whole cell competition assay with 3H-Mibolerone. Binding affinity curves for DHT and enzalutamide (Enza) on AR WT and mutant receptors. Both WT and the mutant receptors show high affinity for DHT, while for enzalutamide there is an increase in affinity for the F877L and F877L/T878A mutants. B, acceptor photobleaching FRET for AR WT and the mutant ARs stimulated with either DHT or enzalutamide. Binding of DHT-induced intramolecular N/C interactions for WT, F877L, and F877L/T878A receptors. Induction of N/C interaction by enzalutamide was only observed for the double-mutant AR. the AR F877L single mutant, high signals for N/C interactions our FRAP experiments, DHT induced nearly overlapping redis- were observed for enzalutamide-bound double-mutant AR (Fig. tribution curves from which very similar immobile fractions can 3B and Supplementary Fig. S8). be calculated (Fig. 5A and D). Hydroxyflutamide-stimulated AR F877L reacted like the inac- Comparison of enzalutamide-induced coregulator recruitment tive DNA-binding defective mutant, while the curves of the by WT and mutant receptors hydroxyflutamide-bound AR T878A and the double mutant were To explore whether DHT, enzalutamide, and hydroxyflutamide nearly superimposable on that of the DHT-bound AR WT (Fig. could induce a modulation of the interaction of WT and mutant 5B). The calculated immobilized fraction size and free diffusion AR with coregulators, we performed a MARCoNI assay for the full- time reflected these similarities in mobility behavior (Fig. 5D; length receptors. The natural ligand (DHT) induced strong inter- Supplementary Table S2). actions of coactivator peptides with both WT and all mutant We also compared the effect of enzalutamide on the mobility of receptors, while corepressor interactions were downregulated the AR mutants. The FRAP curves of the WT AR, AR T878A were (Fig. 4A and Supplementary Fig. S9). Unsupervised clustering superimposable with that of the inactive DNA-binding defective showed that a very similar binding pattern was observed for the AR A574D. The double-mutant AR showed a slower redistribution hydroxyflutamide-bound AR T878A (Supplementary Figs. S10 pattern after photobleaching when compared with AR WT, AR and S11) and the enzalutamide-bound double-mutant AR. In T878A, and AR F877L (Fig. 5C). Clearly, the AR double mutant fact, out of all enzalutamide-stimulated ARs, the double mutant approaches the FRAP profile of the DHT-stimulated AR WT, while resembled most closely to the pattern observed for DHT-stimu- the curve for the enzalutamide-bound AR F877L mutant is more lated receptors (Fig. 4A and B and Supplementary Fig. S11). In similar to that of the enzalutamide-bound AR WT (Fig. 5C). contrast, no strong effects on coregulator interactions could be Comparable results were observed for ARN-stimulated AR WT observed for enzalutamide-bound AR WT or AR T878A (Fig. 4A and mutant receptors (Supplementary Fig. S13). and B and Supplementary Fig. S11). Computational calculations of the fraction sizes of receptors that have an apparent long immobilization time revealed that the Comparison of the intranuclear mobility of AR WT and mutant AR F877L/T878A double mutant, when stimulated with enzalu- receptors tamide, behaved similarly to any of the AR variants tested with Nuclear mobility of a transcription factor can be studied by DHT, while the single mutant AR F877L behaved more like the FRAP. The extent of immobilization is strongly correlated with antagonist-bound AR WT (Supplementary Table S2). chromatin-binding characteristics (27). For all receptors tested in The double AR mutant is targetable by other AR antagonists We explored whether the double-mutant AR would be resistant Table 2. Relative binding affinities for DHT and enzalutamide (Enza) of WT AR to other antagonists nonrelated to enzalutamide or ARN. For this, and the mutant ARs we used the well-known AR axis inhibitors abiraterone acetate RBAs WT F877L F877L/T878A (primarily discovered as CYP17 inhibitors, however, it can also DHT 100 75.81 77.51 directly inhibit the AR; refs. 36, 37), galaterone, Epi (this com- Enza 1.38 12.77 19.28 pound targets the NTD of the AR; ref. 38), AZD, and ASC, as well as NOTE: The binding affinities of DHT and Enza are presented as relative to the some experimental ARAN compounds which we described earlier affinity of WT AR for DHT (set at 100), as determined in whole cell competition for their full antagonistic activity on AR W742C and AR T878A assay. (33). All of the tested compounds exhibited no agonistic and

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1707

Prekovic et al.

Figure 4. Coregulator recruitment for AR WT and mutants. A, MARCoNI assay heatmap representing AR-coregulator peptide interactions for DHT and enzalutamide (Enza)-stimulated WT AR and mutant ARs. Stimulation of receptors with DHT led to increased binding to peptides from coactivators (red) and loss of binding to corepressor peptides (blue); similar pattern was observed only for AR F877L/T878A stimulated with enzalutamide. B, data on selected coactivator peptides from the MARCoNI assay. Interaction of AR with NCOA3_609_631 and PRGC_134_154 peptides is evident upon agonistic stimulation of AR and absent in presence of antagonists. Hydroxyﬂutamide (Hof) and enzalutamide act as agonists on T878A and the double-mutant AR and are able to induce interactions in a similar fashion as DHT.

strong to moderate antagonistic activity on the AR WT, AR T878A, compared the effect of these mutations and their combination on AR F877L, and AR F877L/T878A (Fig. 6 and Supplementary Fig. the biochemical properties of the AR. S14) when tested on an androgen-responsive luciferase reporter. The only exception being ARAN-4, which slightly activated the Transactivation properties double-mutant AR, although not to the same extent as enzaluta- Enzalutamide is only a weak partial agonist for the AR F877L mide (Supplementary Fig. S14). mutation, while when this mutation co-occurs with T878A, there Proliferation studies in LAPC4 cells (Supplementary Methods) is a dramatic increase in agonistic effect of enzalutamide. A similar also show that cells transfected with the double-mutant AR were observation was made for the combination of L701H and T878A rescued from the growth-suppressing effect of enzalutamide when mutations, which resulted in greater responsiveness to cortisol – compared with WT AR transfected cells (Supplementary Fig. compared with the single L701H mutant (40). Possibly, in both S15A). Furthermore, the same rescue was observed for T878A- cases, the T878A mutation leads to a stabilization of the agonistic ﬂ transfected cells treated with hydroxy utamide (Supplementary conformation of the receptor. This might be attained by the Fig. S15B). Abiraterone acetate, however, had a similar effect on reduction in steric hindrance (discussed below) or by reposition- proliferation of all the cells regardless of the receptor transfected ing of helix (H) 11 and H12 towards the ligand-binding pocket, (Supplementary Fig. S15C). facilitating the transition into the agonistic mode of the AR as argued in refs. 30, 41. Discussion It is well documented that mutations in the AR can lead to In silico modeling clinical resistance against antiandrogens, and can even turn the The antagonists that are currently used in the clinic all act via the latter into agonists. The AR T878A mutation of the AR for instance ligand-binding pocket, preventing the correct realignment of H12 converts hydroxyﬂutamide to an agonist in both preclinical and which is recognized by coactivators. However, the details of their clinical settings (8, 30). Recently, it has been proposed that the AR mechanisms of action are different (reviewed in refs. 41, 42). It has F877L mutation leads to enzalutamide resistance, as it might been proposed that bicalutamide binds an additional binding convert enzalutamide into an agonist (9, 10, 12, 39). Here, we pocket (B-site) in the proximity of the androgen-binding pocket

1708 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Molecular Analysis of Androgen Receptor Mutants

Figure 5. Agonistic effects of enzalutamide (Enza) are reﬂected in differences in intranuclear mobility of the AR mutants. Fitted curves of data obtained in FRAP experiments (ﬁrst 30 seconds). The data represent the mean of two independent experiments (30 cells). The following compounds were used: A, DHT (10 nmol/L); B, Hof (10 mmol/L); and C, enzalutamide (10 mmol/L). D, pie- charts depicting calculated fraction size of the long and short immobile fraction, as well as mobile fraction attributed to each receptor. The long immobile fraction size of enzalutamide-bound double mutant is similar to that of agonistically activated AR WT and AR T878A.

(34). In contrast, for enzalutamide, it has been suggested that enzalutamide and ARN, while the co-occurrence with T878A when it binds to the AR it contacts residues on H11, thus (through mechanisms discussed above) is pivotal in strongly disrupting H11–H12 interactions which prevents H12 from enhancing the agonistic activity of enzalutamide. adopting the correct agonistic alignment with the LBD (9). Our in silico model indicates that the F877L mutation expands the N/C interactions and binding to coregulator peptides ligand-binding pocket, allowing for enzalutamide to fit inside and When the AR is activated by agonists it undergoes a confor- interact with R753 like the agonists DHT or T. Similarly, the mational change that leads to close positioning of the NTD and molecular dynamics simulation by Balbas and colleagues pro- LBD and thus interaction between these two domains (27, 43). posed that F877L allows repositioning of enzalutamide from an The enzalutamide-bound AR WT or F877L did not show any N/C antagonistic binding mode in such a way that steric clashes interactions. However, strong interaction was observed for the resulting in dislocation of H12 in WT AR are eliminated (9). The double-mutant AR bound by enzalutamide, or any mutant or WT modeling further proposes that the additional exchange of thre- receptor stimulated by DHT. As we detected transactivation and onine at site 878 to alanine leads to extra reduction of steric chromatin binding (discussed below) for the enzalutamide- hindrance for enzalutamide. This is corroborated by crystallo- bound F877L receptor, it might be that the absence of N/C graphic data of the bicalutamide-bound T878A mutant (34). interaction poises the F877L receptor only for specific promoters, Furthermore, according to our modeling, the F877L mutation as it was observed that deletions/mutations of FQNLF selectively leads to a reduced hydrophobic contact and putative solvent diminish the activity of the AR at specific promoter elements (44). exposure for DHT. This might be the explanation why the mutant To investigate the effect of DHT and antiandrogens on AR– AR is less responsive to DHT. coregulator interactions, we employed the MARCoNI assay (24). Our modeling may provide some insight valuable for design of The modulation of interactions by DHT was similar, but clearly next generation compounds. Putative future antiandrogens could distinct from the one recorded for the ER stimulated with estradiol for instance harbor a group incompatible with the F877L due to (45). As expected, we observed strong interactions of the agonist- steric hindrance (rigid bulky modifications) or polar incompat- bound AR with the coactivator peptides that contain FxxLF motifs. ibility with the hydrophobic contact surface of either the phenyl- The MARCoNI data for DHT on the WT and mutant receptors was alanine or the mutant leucine. very similar. The slight drop in coactivator peptide binding for the mutant receptors again strongly correlates with the decreased Ligand binding responsiveness to DHT. In concordance with the computational modeling, we observed AR WT treated with hydroxyflutamide or enzalutamide showed a decrease in affinity of DHT for the receptors containing the very little changes in coregulator binding, which is expected for F877L mutation, while the affinity of enzalutamide was increased. antagonists. Hydroxyflutamide induces some binding of AR WT, Although co-occurrence of F877L and T878A led to a strong which is in concordance with its reported partial agonist activity, increase in enzalutamide agonism, only a modest further increase and the agonist activity of hydroxyflutamide on AR T878A corre- in affinity for the double-mutant AR was observed. Therefore, we lates with the similarity of the peptide interaction pattern with propose that F877L is crucial for the increase in affinity of that of DHT-bound WT AR. The pattern of modulation of the

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1709

Prekovic et al.

Figure 6. The AR double-mutant is targetable by other AR antagonists. HEK293T cells were used to compare the activity of several antiandrogens on transactivation of AR. Transcriptional activity of none of the AR mutants was induced by abiraterone acetate (Abi), galaterone (Gal), Epi, AZD, or ASC. Moreover, these compounds successfully reduced the DHT-induced transactivation of AR for both the WT and the mutant receptors. The mean and SEM of three independent experiments are shown. AR WT (A), AR T878A (B), F877L (C), and AR F877L/T878A (D).

single mutant receptors by enzalutamide cluster more closely to as observed for AR WT and AR T878A stimulated with agonists the antagonistic enzalutamide-treated WT AR. In contrast, the and antagonists (43). Antagonist-bound ARs behave differently, enzalutamide stimulated double-mutant AR clusters together having smaller immobile fractions, resembling that of DNA- with the agonistic hydroxyflutamide-bound WT and double- binding defective receptors (26). mutant AR, further supporting the fact that enzalutamide acts as Using FRAP, we observed that while the AR F877L mutant had a a potent partial agonist of the double-mutant AR. smaller enzalutamide-induced immobile fraction, the chromatin- Clearly, all these data validate this assay for the simultaneous bound fraction of enzalutamide-stimulated double-mutant AR comparison of relative effects of agonists and antagonists on the in was similar to the fraction of the DHT-bound AR WT. vitro binding to a series of coregulator-derived peptides. In conclusion, the MARCoNI data show that ligands with different Inhibiting the F877L and T878A mutant ARs activity can be separated according to their modulation of AR Both enzalutamide and ARN share a 4-cyano-3-(trifluoro- interactions with coregulator peptides into full agonists, partial methyl)phenyl group with Bic and hydroxyflutamide required agonists, and antagonists. for specific binding to the AR. Because of the structural similarities between enzalutamide and ARN, the F877L mutation is likely to Ligand-induced intranuclear behavior of mutant receptors be enzalutamide/ARN structure-specific. We were able to show It is believed that the immobile fraction of AR measured in that structurally diverse AR inhibitors (abiraterone acetate, gala- FRAP assays corresponds to the chromatin-bound receptor during terone, and Epi) can completely abolish the DHT-activation of the the transcription activation process, because of the overlap of AR F877L and double-mutant AR (38, 46, 47). Furthermore, we have locations with the sites of RNA synthesis (17). There is also a shown that abiraterone acetate effectively inhibited proliferation strong correlation between AR activity and the immobile fraction, of cells harboring transfected double-mutant AR, while presence

1710 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Molecular Analysis of Androgen Receptor Mutants

of the double mutant led to sustained survival under enzaluta- agonist and antagonist activity of enzalutamide, hydroxyflutamide treatment. These results illustrate that targeting AR with mide, bicalutamide, and experimental BF3 binding antagonists compounds that bind differently to the LBD or with compounds on a series of AR single or combined mutations which they that target other domains rather than the LBD, might be favorable found in castration-resistant prostate cancer (39). The data and potential candidates for future treatment of enzalutamide- presented here focus on the effect of the F877L/T878A muta- resistant patients harboring F877L mutation. tion on the (ant-)agonist binding, the intramolecular N/C interactions, the coregulator interactions and the intranuclear Clinical implications mobility. Together with Lallous and colleagues, we conclude a Mutations such as T878A have been found recurrently in full molecular characterization of AR mutations that occur in mCRPC patients, in hydroxyflutamide-treated patients, but also CRPC could provide recommendations for future personalized in patients that were never treated with hydroxyflutamide (14, therapy. 48). Moreover, the antiandrogen withdrawal syndrome, which is fl known to occur in patients after cessation of antiandrogens such Disclosure of Potential Con icts of Interest fl as hydroxyflutamide, bicalutamide, or nilutamide treatment is No potential con icts of interest were disclosed. known to be partly explained by the occurrence of mutations in Authors' Contributions the AR that convert the antagonists into agonists (49). Similar Conception and design: S. Prekovic, R. Houtman, F. Claessens, C. Helsen mechanisms have been anticipated for enzalutamide. Development of methodology: M.E. van Royen, D. Melchers, A.B. Houtsmuller, For now, only a limited number of genomic studies on post- F. Claessens, C. Helsen enzalutamide patients have been reported. However, none of Acquisition of data (provided animals, acquired and managed patients, these identified the single F877L mutation in this population. provided facilities, etc.): S. Prekovic, M.E. van Royen, R. Houtman, D. Melchers, Only one patient was found to have the F877L mutation, but it T. Van den Broeck, E. Smeets, L. Spans Analysis and interpretation of data (e.g., statistical analysis, biostatistics, was co-occurring in the same allele with the T878A mutation computational analysis): S. Prekovic, M.E. van Royen, A.R.D. Voet, B. Geverts, (14, 39). In addition, withdrawal syndrome seems to be either R. Houtman, K.Y.J. Zhang, F. Claessens extremely rare or nonexistent after enzalutamide progression Writing, review, and/or revision of the manuscript: S. Prekovic, M.E. van (50, 51). All this suggests that in the clinic, the frequency of point Royen, A.R.D. Voet, R. Houtman, K.Y.J. Zhang, T. Van den Broeck, E. Smeets, mutations that convert enzalutamide to an agonist is likely low. S. Joniau, F. Claessens, C. Helsen Possibly, other events like AR overexpression or activation of Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Geverts other signaling pathways are more relevant (14, 48). Study supervision: S. Joniau, F. Claessens, C. Helsen

Conclusion Acknowledgments The authors thank H. Debruyn, R. Bollen, and D. Schollaert for their excellent In conclusion, the molecular modeling and ligand-binding technical assistance and the colleagues from the Molecular Endocrinology data, the transactivation data, the N/C interactions, the MAR- Laboratory for the helpful discussions. CoNIdata,andtheFRAPdataallconvergetothesamehypoth- esis: the F877L mutation enhances enzalutamide binding, but Grant Support has a limited effect on the activation potential of this antag- This work supported by research grants from the FWO-Vlaanderen onist. Only when F877L is combined with T878A, the receptor (G.0830.13N), KU Leuven (GOA/15/017), and Kom op tegen Kanker. F. becomes strongly activated by enzalutamide. In addition, these Claessens is the principal investigator for all the grants previously mentioned. mutant ARs can be targeted with antiandrogens such as abir- A.R.D. Voet was supported by RIKEN FPR fellowship. aterone acetate or galaterone, opening the possibility for The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked sequential inhibition of AR with structurally diverse com- advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate pounds. Further genomic studies similar to the one by Azad this fact. and colleagues should be performed to determine the identity and frequency of the AR mutations in post-enzalutamide Received November 3, 2015; revised April 18, 2016; accepted April 26, 2016; patients (14). A recent article by the same group describes the published OnlineFirst May 16, 2016.

References 1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, 6. Antonarakis ES, Lu C, Wang H, Luber B, Nakazawa M, Roeser JC, et al. AR- et al. Cancer incidence and mortality worldwide: sources, methods V7 and resistance to enzalutamide and abiraterone in prostate cancer. N and major patterns in GLOBOCAN 2012. Int J Cancer 2015;136: Engl J Med 2014;371:1028–38. E359–86. 7. Hara T, Miyazaki J-i, Araki H, Yamaoka M, Kanzaki N, Kusaka M, et al. 2. Drake C, Sharma P, Gerritsen W. Metastatic castration-resistant prostate Novel mutations of androgen receptor A possible mechanism of bicalu- cancer: new therapies, novel combination strategies and implications for tamide withdrawal syndrome. Cancer Res 2003;63:149–53. immunotherapy. Oncogene 2014;33:5053–64. 8. Taplin M-E, Bubley GJ, Ko Y-J, Small EJ, Upton M, Rajeshkumar B, et al. 3. Ha Y, Goodin S, DiPaola R, Kim I. Enzalutamide for the treatment of Selection for androgen receptor mutations in prostate cancers treated with castration-resistant prostate cancer. Drugs Today (Barc) 2013;49:7–13. androgen antagonist. Cancer Res 1999;59:2511–5. 4. Helsen C, Van den Broeck T, Voet A, Prekovic S, Van Poppel H, Joniau S, 9. Balbas MD, Evans MJ, Hosﬁeld DJ, Wongvipat J, Arora VK, Watson PA, et al. et al. Androgen receptor antagonists for prostate cancer therapy. Endocr Overcoming mutation-based resistance to antiandrogens with rational Relat Cancer 2014;21:T105–18. drug design. Elife 2013;2:e00499. 5. Cabot RC, Harris NL, Rosenberg ES, Shepard J-AO, Cort AM, Ebeling SH, 10. Korpal M, Korn JM, Gao X, Rakiec DP, Ruddy DA, Doshi S, et al. An F876L et al. Increased survival with enzalutamide in prostate cancer after chemo- mutation in androgen receptor confers genetic and phenotypic resistance therapy. N Engl J Med 2012;367:1187–97. to MDV3100 (enzalutamide). Cancer Discov 2013;3:1030–43.

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1711

Prekovic et al.

11. Spans L, Helsen C, Clinckemalie L, Van den Broeck T, Prekovic S, Joniau S, 32. Pereira de JesusTran K, Cot^ e PL, Cantin L, Blanchet J, Labrie F, Breton R. et al. Comparative genomic and transcriptomic analyses of LNCaP and C4- Comparison of crystal structures of human androgen receptor ligandbind- 2B prostate cancer cell lines. PLoS One 2014;9:e90002. ing domain complexed with various agonists reveals molecular determi- 12. Joseph JD, Lu N, Qian J, Sensintaffar J, Shao G, Brigham D, et al. A clinically nants responsible for binding affinity. Protein Sci 2006;15:987–99. relevant androgen receptor mutation confers resistance to second-gener- 33. Voet A, Helsen C, Zhang KY, Claessens F. The discovery of novel human ation antiandrogens enzalutamide and ARN-509. Cancer Discov 2013;3: androgen receptor antagonist chemotypes using a combined pharmaco- 1020–9. phore screening procedure. ChemMedChem 2013;8:644–51. 13. Yamamoto Y, Loriot Y, Beraldi E, Zhang F, Wyatt AW, Al Nakouzi N, et al. 34. Bohl CE, Gao W, Miller DD, Bell CE, Dalton JT. Structural basis for Generation 2.5 antisense oligonucleotides targeting the androgen receptor antagonism and resistance of bicalutamide in prostate cancer. Proc Natl and its splice variants suppress enzalutamide resistant prostate cancer cell Acad Sci 2005;102:6201–6. growth. Clin Cancer Res 2015:21:1675–87. 35. He B, Gampe RT, Kole AJ, Hnat AT, Stanley TB, An G, et al. Structural basis 14. Azad AA, Volik SV, Wyatt AW, Haegert A, Le Bihan S, Bell RH, et al. for androgen receptor interdomain and coactivator interactions suggests a Androgen receptor gene aberrations in circulating cell-free DNA: biomar- transition in nuclear receptor activation function dominance. Mol Cell kers of therapeutic resistance in castration-resistant prostate cancer. Clin 2004;16:425–38. Cancer Res 2015;21:2315–24. 36. Richards J, Lim AC, Hay CW, Taylor AE, Wingate A, Nowakowska K, 15. Tanner TM, Denayer S, Geverts B, Van Tilborgh N, Kerkhofs S, Helsen C, et al. Interactions of abiraterone, eplerenone, and prednisolone with et al. A 629RKLKK633 motif in the hinge region controls the androgen wild-type and mutant androgen receptor: a rationale for increasing receptor at multiple levels. Cell Mol Life Sci 2010;67:1919–27. abiraterone exposure or combining with MDV3100. Cancer Res 2012; 16. Alen P, Claessens F, Verhoeven G, Rombauts W, Peeters B. The androgen 72:2176–82. receptor amino-terminal domain plays a key role in p160 coactivator- 37. Romanel A, Tandefelt DG, Conteduca V, Jayaram A, Casiraghi N, Wetters- stimulated gene transcription. Mol Cell Biol 1999;19:6085–97. kog D, et al. Plasma AR and abiraterone-resistant prostate cancer. Sci Transl 17. van Royen ME, Cunha SM, Brink MC, Mattern KA, Nigg AL, Dubbink HJ, Med 2015;7:312re10–re10. et al. Compartmentalization of androgen receptor protein–protein inter- 38. Andersen RJ, Mawji NR, Wang J, Wang G, Haile S, Myung J-K, et al. actions in living cells. J Cell Biol 2007;177:63–72. Regression of castrate-recurrent prostate cancer by a small-molecule inhib- 18. Helsen C, Marchand A, Chaltin P, Munck S, Voet A, Verstuyf A, et al. itor of the amino-terminus domain of the androgen receptor. Cancer Cell Identification and characterization of MEL-3, a novel ar antagonist that 2010;17:535–46. suppresses prostate cancer cell growth. Mol Cancer Ther 2012;11:1257–68. 39. Lallous N, Volik SV, Awrey S, Leblanc E, Tse R, Murillo J, et al. Functional 19. Denayer S, Helsen C, Thorrez L, Haelens A, Claessens F. The rules of DNA analysis of androgen receptor mutations that confer anti-androgen resis- recognition by the androgen receptor. Mol Endocrinol 2010;24:898–913. tance identified in circulating cell-free DNA from prostate cancer patients. 20. Boyd S. Molecular operating environment. Chem World-Uk 2005;2:66-. Genome Biol 2016;17:1. 21. Hsu C-L, Liu J-S, Wu P-L, Guan H-H, Chen Y-L, Lin A-C, et al. Iden- 40. Matias PM, Carrondo MA, Coelho R, Thomaz M, Zhao X-Y, Wegg A, et al. tification of a new androgen receptor (AR) co-regulator BUD31 and Structural basis for the glucocorticoid response in a mutant human related peptides to suppress wild-type and mutated AR-mediated pros- androgen receptor (ARccr) derived from an androgen-independent prostate cancer growth via peptide screening and X-ray structure analysis. tate cancer. J Med Chem 2002;45:1439–46. Mol Oncol 2014;8:1575–87. 41. Tan ME, Li J, Xu HE, Melcher K, Yong E-l. Androgen receptor: structure, role 22. Ostrowski J, Kuhns JE, Lupisella JA, Manfredi MC, Beehler BC, Krystek Jr SR, in prostate cancer and drug discovery. Acta Pharmacol Sin 2015;36:3–23. et al. Pharmacological and x-ray structural characterization of a novel 42. Rathkopf D, Scher HI. Androgen receptor antagonists in castration-resis- selective androgen receptor modulator: potent hyperanabolic stimulation tant prostate cancer. Cancer J 2013;19:43. of skeletal muscle with hypostimulation of prostate in rats. Endocrinology 43. Farla P, Hersmus R, Geverts B, Mari PO, Nigg AL, Dubbink HJ, et al. The 2007;148:4–12. androgen receptor ligand-binding domain stabilizes DNA binding in living 23. Neudert G, Klebe G. DSX: a knowledge-based scoring function for the cells. J Struct Biol 2004;147:50–61. assessment of protein–ligand complexes. J Chem Inf Model 2011;51: 44. He B, Lee LW, Minges JT, Wilson EM. Dependence of selective gene 2731–45. activation on the androgen receptor NH2-and COOH-terminal interac- 24. Houtman R, de Leeuw R, Rondaij M, Melchers D, Verwoerd D, Ruijtenbeek tion. J Biol Chem 2002;277:25631–9. R, et al. Serine-305 phosphorylation modulates estrogen receptor alpha 45. Evers NM, Wang S, van den Berg JH, Houtman R, Melchers D, de Haan LH, binding to a coregulator peptide array, with potential application in et al. Identification of coregulators influenced by estrogen receptor subtype predicting responses to tamoxifen. Mol Cancer Ther 2012;11:805–16. specific binding of the ER antagonists 4-hydroxytamoxifen and fulvestrant. 25. Team RC. R: A language and environment for statistical computing. Chem Biol Interact 2014;220:222–30. R Foundation for Statistical Computing: Vienna, Austria; 2012. 46. Adamo V, Noto L, Franchina T, Chiofalo G, Picciotto M, Toscano G, et al. 26. Farla P, Hersmus R, Trapman J, Houtsmuller AB. Antiandrogens prevent Emerging targeted therapies for castration-resistant prostate cancer. Front stable DNA-binding of the androgen receptor. J Cell Sci 2005;118: Endocrinol 2012;3:73. 4187–98. 47. Schweizer MT, Antonarakis ES. Abiraterone and other novel androgen- 27. van Royen ME, Dinant C, Farla P, Trapman J, Houtsmuller AB. FRAP and directed strategies for the treatment of prostate cancer: a new era of FRET methods to study nuclear receptors in living cells. Methods Mol Biol hormonal therapies is born. Ther Adv Urol 2012;4:167–78. 2009;505:69–96. 48. Carreira S, Romanel A, Goodall J, Grist E, Ferraldeschi R, Miranda S, et al. 28. Geverts B, van Royen ME, Houtsmuller AB. Analysis of biomolecular Tumor clone dynamics in lethal prostate cancer. Sci Transl Med 2014;6: dynamics by FRAP and computer simulation. Methods Mol Biol 2015; 254ra125. 1251:109–33. 49. Sartor AO, Tangen CM, Hussain MH, Eisenberger MA, Parab M, Fontana JA, 29. van Royen ME, van Cappellen WA, Geverts B, Schmidt T, Houtsmuller AB, et al. Antiandrogen withdrawal in castraterefractory prostate cancer. Cancer Schaaf MJ. Androgen receptor complexes probe DNA for recognition 2008;112:2393–400. sequences by short random interactions. J Cell Sci 2014;127:1406–16. 50. Rodriguez-Vida A, Bianchini D, Van Hemelrijck M, Hughes S, Malik Z, 30. Zhou J, Liu B, Geng G, Wu JH. Study of the impact of the T877A mutation Powles T, et al. Is there an antiandrogen withdrawal syndrome with on ligandinduced helix12 positioning of the androgen receptor resulted in enzalutamide? BJU Int 2015;115:373–80. design and synthesis of novel antiandrogens. Proteins 2010;78:623–37. 51. von Klot CA, Kuczyk MA, Merseburger AS. No androgen withdrawal 31. Gao W, Bohl CE, Dalton JT. Chemistry and structural biology of androgen syndrome for enzalutamide: a report of disease dynamics in the postche- receptor. Chem Rev 2005;105:3352–70. motherapy setting. Eur Urol 2014;65:258.

1712 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

The Effect of F877L and T878A Mutations on Androgen Receptor Response to Enzalutamide

Stefan Prekovic, Martin E. van Royen, Arnout R.D. Voet, et al.

Mol Cancer Ther 2016;15:1702-1712. Published OnlineFirst May 16, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-15-0892

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2016/05/14/1535-7163.MCT-15-0892.DC1

Cited articles This article cites 50 articles, 18 of which you can access for free at: http://mct.aacrjournals.org/content/15/7/1702.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/15/7/1702.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/15/7/1702. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.