Androgen Glucuronidation: an Unexpected Target for Androgen Deprivation Therapy, with Prognosis and Diagnostic Implications

Published OnlineFirst October 11, 2013; DOI: 10.1158/0008-5472.CAN-13-1462

Cancer Molecular and Cellular Pathobiology Research

Androgen Glucuronidation: An Unexpected Target for Androgen Deprivation Therapy, with Prognosis and Diagnostic Implications

Laurent Grosse1, Sophie Paquet^ 1, Patrick Caron1, Ladan Fazli3, Paul S. Rennie3, Alain Belanger 2, and Olivier Barbier1

Abstract Androgen deprivation therapy (ADTh) remains a mainstay of prostate cancer treatment, but its efficacy is bypassed by mechanisms that are not fully understood. In human prostate cancer cells, androgen glucuronidation, catalyzed by the two UDP-glucuronosyltransferase (UGT) enzymes UGT2B15 and UGT2B17, is the major androgen inactivation pathway. In this study, we investigated the effect of ADTh on androgen glucuronidation to evaluate its potential clinical utility for prostate cancer prognosis or therapy. UGT2B15 and UGT2B17 expression was evaluated in prostate cancer specimens from untreated or treated patients and in cell models of prostate cancer exposed to clinically relevant antiandrogens. UGT2B15 and UGT2B17 protein levels in prostate were increased after 5 months of ADTh when compared with specimens from untreated patients. UGT2B15 expression remained elevated for up to 12 months, but UGT2B17 returned to initial levels as soon as after 6 months. Several androgen receptor (AR) antagonists tested caused a dose- and time-dependent stimulation of UGT2B15 and UGT2B17 expression and androgen glucuronidation in prostate cancer cell lines. The role of AR in these regulatory events was confirmed using AR-deficient LNCaP cells, in which UGT2B attenuation reduced the antiproliferative effects of AR pharmacologic antagonists. Through this combination of clinical and functional investigations, our work revealed that ADTh stimulates a local androgen metabolism in prostate cells, establish- ing a foundation to evaluate the potential of UGT2B15 and UGT2B17 as drug targets and/or molecular markers for ADTh responsiveness and maintenance in prostate cancer. Cancer Res; 73(23); 1–9. 2013 AACR.

Introduction proliferation (3). For this reason, androgen deprivation therapy Prostate cancer remains the second most frequent type of (ADTh), also called "medical castration" or "hormone therapy," fi cancer diagnosed worldwide, and the third leading cause of has become a rst-line treatment strategy for advanced pro- cancer-related death in men in industrialized countries (1). state cancer (4). ADTh is achieved using gonadotropin releas- Prostate cancer is an adenocarcinoma originating from cells of ing hormone (Gn-RH) analogs such as leuprolide, buserelin, epithelial type, but displays morphologically and genetically or goserelin (4), which inhibit testosterone synthesis (2). very heterogeneous properties (2). Androgens are among the This medical castration is often combined to antiandrogens fl main factors controlling the initiation, maintenance, and such as utamide, nitulamide, or bicalutamide (2). These AR progression of prostate cancer (3). During carcinogenesis, antagonists provide an additional blockade of androgen sig- epithelial cells transform into a malignant phenotype, where naling by preventing AR activation by the locally synthesized the androgen receptor (AR) selectively activates genes posi- androgens. Symptomatic (improvement of quality of life) and/ fi tively controlling cell viability to unrestraint epithelial cell or objective (serum prostate-speci c antigen level normaliza- tion and measurable tumor response) outcomes to ADTh are observed in approximately 80% of patients (5). However, the Authors' Affiliations: 1Laboratory of Molecular Pharmacology, CHU- duration of these improvements is highly variable, and prostate Quebec Research Centre and Faculty of Pharmacy, 2CHU-Quebec cancer relapses in the majority of patients evolving to the Research Centre and Faculty of Medicine, Laval University, Quebec; and androgen-independent phenotype within 12 to 18 months of 3Vancouver Prostate Centre at VGH, University of British Columbia, Van- couver, British Columbia, Canada therapy (5). The cancer is then commonly referred to as castrate-resistant prostate cancer (CRPC), androgen indepen- Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). dent, or hormone refractory (6). CRPC can transiently be treated with alternative ADTh, but the ultimate options for Corresponding Author: Olivier Barbier, Laboratory of Molecular Pharma- cology, CHU-Quebec Research Centre, 2705, boulevard Laurier, Quebec androgen-independent metastatic prostate cancer lie in che- G1V 4G2, Canada. Phone: 418-654-2296; Fax: 418-654-2769; E-mail: motherapy (5, 7). [email protected] The androgen axis continues to play a major role for the doi: 10.1158/0008-5472.CAN-13-1462 progression of CRPCs (8), and a deep understanding of factors 2013 American Association for Cancer Research. affecting androgen inactivation in tumor cells is required. In

www.aacrjournals.org OF1

Grosse et al.

the prostate, the active hormone dihydrotestosterone (DHT) is prostatic hyperplasia: 16 from untreated donors, 16 from extensively metabolized to inactive and easily excretable participants with CRPC, and finally 56 from patients having androstane-3a-diol-glucuronide (3a-diol-17G) and androster- received ADTh for 1 to 5 months (n ¼ 19), 6 to 8 months (n ¼ one–glucuronide (ADT-3G) derivatives (9, 10). These polar 24), or 9 to 12 months (n ¼ 13; Supplementary SM1). Tissue metabolites are formed through a conjugation reaction called samples were extracted from the whole tissue after radical glucuronidation, and correspond to the major androgen met- prostatectomy, except CRPC samples that were obtained by abolic end products found in circulation in men (11, 12). In transurethral resection of the prostate (24, 25). The therapy humans, glucuronidation is catalyzed by the 19 functional consisted of Gn-RH analogs (buserelin, leuprolide, or goserelin) UDP-glucuronosyltransferase (UGT) enzymes (13). However, and antiandrogens (flutamide, cyprosterone acetate, or bica- in prostate, androgen glucuronidation only involves two lutamide) used alone or in combination. This protocol was enzymes: UGT2B15 and UGT2B17 (14). Experimental inhibi- approved by the institutional review boards at the University of tion of these enzymes significantly improves the proprolifera- British Columbia (Vancouver, BC, Canada), and the "Centre tive properties of DHT in prostate cancer cells (14), suggesting hospitalier universitaire de Quebec" (Quebec, QC, Canada). that they are major determinants for the androgen response. Written informed consent was obtained from all participants. Accordingly, positive associations were reported between prostate cancer risk and a low-activity UGT2B15 allele or a Tissue microarrays complete UGT2B17 gene deletion (15–19). On the other hand, Microarrays and immunohistochemical (IHC) stains were various studies illustrated the negative control that androgens obtained as previously described (24, 25), with formalin-fixed exert on UGT2B15 and UGT2B17 expression in prostate cancer and paraffin-embedded 4-mm sections of the microarray, using cells (14, 20, 21). These last observations establish the AR the previously described (22, 23, 26) polyclonal anti-vimentin regulatory pathway as a major mechanism for controlling (positive control; 1:100 dilution), anti-UGT2B15 (1:300 dilu- androgen glucuronidation. We therefore sought to test the tion), and anti-UGT2B17 (1:150 dilution) antibodies (Supple- possibility that ADTh drugs affect UGT2B15 and/or UGT2B17 mentary SM2–8). The IHC staining (Supplementary SM2–8) expression or activity in prostate cancer tumor samples and was scored by a pathologist (L. Fazli) for the level of immu- cell models exposed to antiandrogens. noexpression on a scale from 0 to 3, wherein 0 was undetect- able, 1 represented a faint or focal questionably present stain, 2 Materials and Methods represented a stain of convincing intensity in a minority of Materials cells, and 3 represented a stain of convincing intensity in a UDP-glucuronic acid and all aglycons were obtained from majority of cells. Sigma or ICN-Pharmaceuticals Inc. R1881, 3a-Diol, DHT, ADT, and DHT-glucuronide (DHT-G) were purchased from Stera- Cell culture loids. ADT-3-glucuronide and 3a-Diol-17-glucuronide were Cell lines were obtained from the American Type Culture provided by the Medicinal Chemistry Division of the "Centre Collection (cell authentication through short tandem repeat de recherche du CHU-Quebec" (21). Cell culture materials, profiling), used within 6 months following reception. LNCaP blasticidin and lipofectin, were purchased from Invitrogen. cells were grown and treated in 10% FBS-supplemented RPMI Penicillin, streptomycin, and Iscove's Modified Dulbecco's 1640 medium. LAPC-4 cells were cultured in IMDM completed Medium (IMDM) were provided by Wisent, whereas puromy- with 7.5% FBS and 10 nmol/L R1881. Inducible AR-shRNA– cin, doxycyclin, bicalutamide, nilutamide, flutamide, and RPMI expressing LNCaP cells were cultured in RPMI 1640 supple- were purchased from Sigma. SYBR Green PCR Master Mix was mented with 10% FBS, puromycin (2.5 mg/mL), and blasticidin purchased from Applied Biosystem (Life Technologies). Pro- (1 mg/mL), as reported in ref. 27. AR deficiency was obtain- tein assay reagents were obtained from Bio-Rad Laboratories ed through a 48-hour pretreatment with doxycycline (DOX, Inc. The anti-calnexin antibody was purchased from Stressgen, 1 m g/mL), and the antibiotic was maintained for the complete the anti-vimentin antibody was from GenScript, and the anti- treatment duration as recommended (27). For RNA isolation, 5 5 UGT2B15 and anti-UGT2B17 antibodies were described pre- 2.5 10 LNCaP or 3.5 10 LAPC-4 cells were plated in each viously (22, 23). The secondary antibody against rabbit immu- well of 12-well plates. For glucuronidation assays and Western 6 noglobulin G (IgG) was purchased from Amersham. The blot analysis experiments, 8 10 LNCaP cells were plated chemiluminescence kit (ECL) was from Renaissance. SiRNA in 10-cm petri dishes. Cells were then treated for the indicat- probe for UGT2B15/UGT2B17 (Individual siGENOME duplex ed duration with vehicle [dimethyl sulfoxide (DMSO) or eth- D-020195-01) and Non-Target #1 were obtained from Dharma- anol; 0.1% v/v], bicalutamide (DMSO), nilutamide (ethanol), con as already described in ref. 14. flutamide (ethanol), R1881 (ethanol), and/or DHT (ethanol) at the indicated concentrations. Knockdown of UGT2B15 Tissue banks and UGT2B17 expression was obtained using a UGT2B15/ Tissues used for the microarray construction were from the UGT2B17 siRNA probe as already reported in ref. 14. Fol- Vancouver Prostate Centre Tissue Bank (http://www.prosta- lowing transfection, cells were allowed 18 hours for recovery, tecentre.com/our-research/core-facilities/biorepository), as transferred in 96-wells plates, and exposed to vehicle (DMSO; described in refs. 24, 25. Prostate samples were obtained from 0.1% v/v) or bicalutamide (10 mmol/L) for 72 hours. Cell 31 benign prostatic hyperplasia (BPH) donors, whereas pros- proliferation was then ensured using the CellTiter 96 AQueous tate cancer samples were from 88 patients with advanced One Solution Cell Proliferation Assay Kit (Promega) and

OF2 Cancer Res; 73(23) December 1, 2013 Cancer Research

Ablation Therapy Improves Androgen Glucuronidation in Prostate Cancer

Figure 1. Androgen ablation therapy modulates UGT2B15 and UGT2B17 protein levels in prostate tumors. UGT2B15 (A) and UGT2B17 (B) protein levels were quantified in prostate tumors using IHC analyses of a tissue microarray with specific antibodies as illustrated in Supplementary SM2–8. Tumor samples were from 31 BPH donors or 88 patients with advanced prostate cancer (Supplementary SM9), who were either untreated (n ¼ 16) or treated with ADTh for 1 to 5months(1–5; n ¼ 19), 6 to 8 months (6–8; n ¼ 24), and 9 to 12 months (9–12; n ¼ 13), or with castrate-resistant tumors (CRPC; n ¼ 16). The staining intensity was visually scored on a scale of 0 to 3 (see Materials and Methods), and values represent the mean SEM of the scores obtained for each experimental group (expressed relatively to untreated samples). Statistically significant differences between experimental groups are indicated by asterisks when compared with untreated and by ¥ when compared with BPH (Mann–Whitney test: , P < 0.05; , P < 0.01; , P < 0.001; ¥, P < 0.05; NS, not significant). optical density determination (490 nm) was done with a protein levels as in BPH samples (Fig. 1; Supplementary SM2– microplate reader Tecan Infinitte M1000 as recommended by 8). However, when compared with BPH and untreated samples, the supplier (Promega). both UGT proteins were significantly more abundant in tissues from ADTh-treated patients for 1 to 5 months (Fig. 1; Supple- Genotyping of the UGT2B17 gene copy number variation mentary SM2–5). The highest UGT2B15 expression was Genomic DNA was isolated from LNCaP and LAPC-4 cells observed after 6 to 8 months of treatment, and remained using the ChargeSwitch gDNA Mini Tissue Kit (Life Technol- significantly higher than in untreated samples up to 12-month ogies) and was then used as a template for UGT2B17 deletion therapy, even if a tendency to lower again was observed in these genotyping as previously described (25, 28). last samples (Fig. 1A). In contrast, UGT2B17 protein levels were reduced to basal levels (untreated) as soon as after a 6- to 8- RNA isolation and quantitative reverse transcription PCR month period of treatment and remained stable for longer Total RNA was isolated according to the Tri-Reagent acid/ exposures (Fig. 1B). In patients with CRPCs, both UGT2B15 and phenol protocol as specified by the supplier (Molecular UGT2B17 protein levels were not significantly different from Research Center Inc.). The reverse transcription (RT) and those in untreated samples. In the IHC results obtained with quantitative PCR reactions were performed as previously the positive control, vimentin (29) indicated that the quanti- described (25). fication of UGT2B protein levels was not biased by the tissue preparation (Supplementary SM2–8). Western blot analyses and in vitro glucuronidation in vitro Western blot analyses and glucuronidation assays Androgen receptor antagonism upregulates UGT2B15 were performed as previously reported (14, 21, 22). and UGT2B17 expression and activity in prostate cancer cell models Statistical analyses Ex vivo experiments were performed using prostate cancer All data are presented as mean SD, except for the IHC cell models and AR antagonists to further investigate the intensity evaluation in which results are presented as mean effects and consequences of AR blockade on androgen glucur- SEM. Comparisons between experimental groups were per- onidation (Figs. 2–4). t – U formed by a two-tailed Student test or a Mann Whitney test First, UGT2B15 and UGT2B17 mRNA and protein levels were in vivo for assessment of UGT levels using the JMP V7.01 quantified in LNCaP cells exposed to increasing doses of software (SAS Institute). bicalutamide (0.1 to 20 mmol/L) for 48 hours. The two UGTs were dose-dependently increased at both mRNA and protein Results levels (Fig. 2). Although the amount of UGT2B15 transcripts Androgen ablation therapy modulates UGT2B15 and was statistically increased in the presence of as low as 1 mmol/L UGT2B17 protein levels in prostate tumors (Fig. 2A), UGT2B17 mRNA accumulation reached the statisti- According to our previous observation (25), untreated pros- cal significance only in the presence of 5 mmol/L bicalutamide tate cancer samples presented similar UGT2B15 and UGT2B17 (Fig. 2B). In time course experiments, UGT2B15 and UGT2B17

www.aacrjournals.org Cancer Res; 73(23) December 1, 2013 OF3

Grosse et al.

Figure 2. Bicalutamide dose- and time-dependently increases UGT2B15 (A and C) and UGT2B17 (B and D) expression in LNCaP cells. A and B, LNCaP cells were treated with DMSO (vehicle, V) or increasing bicalutamide concentrations for 48 hours. C and D, LNCaP cells were treated with DMSO (vehicle) or 10 mmol/L bicalutamide for the indicated duration. UGT2B15 (A and C) and UGT2B17 (B and D) mRNA levels (top) were quantified from total RNA through quantitative RT-PCR analyses, and normalized with the housekeeping RNA 36B4. Values (mean SD) are expressed relatively to control (vehicle, V) set at 1. Statistically significant differences between control and treated cells are indicated by asterisks (Student t test: , P < 0.05; , P < 0.01; , P < 0.001; NS, not significant). UGT2B15 (A and C) and UGT2B17 (B and D) protein levels (bottom) were visualized in cell homogenates (30 mg) through immunoblotting using the anti- UGT2B15 (1:1,500 dilution; A and C) and anti-UGT2B17 (1:2,000 dilution; B and D) antibodies. The same membranes were subsequently hybridized with an anti-calnexin antibody (1:5,000 dilution) to ensure the equal loading of each lane.

transcripts were significantly increased as soon as after 12 cells(Fig.4C).Interestingly, nilutamide exhibited a cell- hours of treatment, and the maximal induction was reached specificresponsewithasignificant induction of UGT2B15 after a 48-hour exposure (Fig. 2C and D). A similar dose- and mRNA in LAPC-4 but not in LNCaP cells (Fig. 4C and D). time-dependent accumulation was observed for both UGT Furthermore, not only did nilutamide not improve UGT2B17 proteins in Western blot analyses (Fig. 2). In vitro glucuronida- expression, but it caused a non significant reduction of these tion assays confirmed that the bicalutamide-dependent transcripts in LNCaP cells (Fig. 4C). This last observation is increase in UGT expression results in improved androgen consistent with the previously reported agonistic effects of conjugation (Fig. 3). Indeed, LNCaP cells exposed to bicaluta- nilutamide in front of the mutated AR that is expressed in mide exhibited a dose- and time-dependent improved ability to LNCaP cells (30). conjugate DHT (Fig. 3A and D), ADT (Fig. 3B and E), and 3a- diol (Fig. 3C and F). The androgen receptor mediates the bicalutamide- To ascertain that results from bicalutamide-treated LNCaP dependent upregulation of UGT2B15 and UGT2B17 cells were not cell type- or antagonist specific, additional expression in LNCaP cells experiments were performed using the LAPC-4 cell model and AR contribution to the bicalutamide-dependent induction the other AR antagonists, nilutamide and flutamide (Fig. 4). In of UGT2B15 and UGT2B17 expression was ensured by using LAPC-4, bicalutamide dose- and time-dependently induced LNCaP cells expressing a DOX-inducible anti-AR short hairpin UGT2B15 expression (Fig. 4A and B). UGT2B17 transcripts RNA (shRNA; Fig. 5; ref. 27). We first confirmed that DOX were not detected in these cells. Actually, genotyping analyses treatment resulted in impaired AR expression and activity revealed that the UGT2B17 gene is absent in LAPC-4 (data not (Supplementary SM9; ref. 27). Subsequently, native (DOX ) þ shown), indicating their del/del phenotype for the previously and DOX-activated cells (DOX ) were cultured in the presence reported UGT2B17 CNV genotype (28). In both LNCaP and of vehicle, bicalutamide (10 mmol/L), or the synthetic (R1881; 1 LAPC-4 cells, bicalutamide and flutamide were efficient in nmol/L), or natural (DHT, 10 nmol/L) AR activators (Fig. 5). increasing UGT2B15 mRNA levels (Fig. 4C and D). A similar In native (DOX ) cells, DHT, R1881, and bicalutamide response was also observed for UGT2B17 transcripts in LNCaP caused the expected down- (R1881 and DHT; ref. 21) or

OF4 Cancer Res; 73(23) December 1, 2013 Cancer Research

Ablation Therapy Improves Androgen Glucuronidation in Prostate Cancer

Figure 3. Bicalutamide dose- and time-dependently stimulates androgen glucuronidation in LNCaP cells. A–C, LNCaP cells were treated with DMSO (vehicle, V) or increasing bicalutamide concentrations for 48 hours. D–F, LNCaP cells were treated with DMSO (vehicle) or 10 mmol/L bicalutamide for the indicated duration. Cell homogenates (70 to 210 mg) were used for in vitro glucuronidation assays performed for 1 hour in the presence of 100 mmol/L of the active androgen DHT (A and D) or its reduced metabolites ADT (B and E) and androstane-3a,17b-diol (3a-Diol; C and F). The formation of glucuronide derivatives was measured through liquid chromatography/electrospray ionization–tandem mass spectrometry (LC/ESI-MS/MS) analyses. Statistically significant differences between control and treated cells are indicated by asterisks (Student t test: , P < 0.05; , P < 0.01; , P < 0.001). upregulation (bicalutamide) of UGT2B15 (Fig. 5A) and transcript levels, thus revealing that inhibition of AR expres- UGT2B17 (Fig. 5B) mRNA expression. sion has similar consequences as the use of AR antagonists þ In AR-deficient (DOX ) cells, DOX alone also significantly for these UGT gene expressions. The addition of R1881 in increased the UGT2B15 (Fig. 5A) and UGT2B17 (Fig. 5B) culture media had no further impact on mRNA levels of the

Figure 4. The AR antagonists bicalutamide (A–D), nilutamide (C and D), and flutamide (C and D) differentially modulate UGT2B15 (A–D) and UGT2B17 (C) mRNA levels in LNCaP and LAPC-4 cells. A, LAPC-4 cells were treated with DMSO (vehicle) or increasing bicalutamide concentrations for 48 hours. B, LAPC-4 cells were treated with DMSO (vehicle) or 10 mmol/L bicalutamide for the indicated duration. C and D, LNCaP (C) and LAPC-4 (D) cells were treated with vehicle, bicalutamide (10 mmol/L), nilutamide (10 mmol/L), or flutamide (20 mmol/L) for 48 hours. UGT2B15 (A–D) and UGT2B17 (C) mRNA levels were quantified from total RNA through quantitative RT-PCR analyses, and normalized with the housekeeping RNA 36B4. Values (mean SD) are expressed relatively to control (vehicle) set at 1. Statistically significant differences between control and treated cells are indicated by asterisks (Student t test: , P < 0.05; , P < 0.01; , P < 0.001; NS, not significant).

www.aacrjournals.org Cancer Res; 73(23) December 1, 2013 OF5

Grosse et al.

Figure 5. shRNA-mediated knockdown of the androgen receptor modifies the ability of its agonists and antagonists to modulate UGT2B15 (A) and UGT2B17 þ (B) mRNA expression in LNCaP cells. LNCaP cells stably expressing an inducible anti-AR shRNA were pretreated (DOX ) or not (DOX ) with DOX (1 mg/mL) for 48 hours and then cultured in media enriched with DMSO or ethanol (vehicle), R1881 (1 nmol/L), DHT (10 nmol/L), or bicalutamide (10 mmol/L) for an additional 48 hours. The antibiotic was maintained at the same concentration for the complete treatment duration. UGT2B15 (A) and UGT2B17 (B) mRNA levels were quantified from total RNA through quantitative RT-PCR analyses and normalized with the housekeeping RNA 36B4. Values (mean SD) are þ expressed relatively to vehicle-treated DOX cells set at 1. Statistically significant differences between DOX and DOX (white bars vs. gray bars) cells cultured with or without AR agonists/antagonists are indicated by asterisks (Student t test: , P < 0.01; , P < 0.001; NS, not significant). The statistical significance of the differences in mRNA levels between vehicle versus R1881-, DHT-, or bicalutamide-treated DOX cells (white bars †††, þ , P < 0.001) and between vehicle versus DHT- or bicalutamide-treated DOX cells (gray bars #, P < 0.05; ###, P < 0.001) were also determined using the parametric Student t test.

two UGTs (Fig. 5). Similarly, the negative effect of DHT on deficient cells were less sensitive to the antagonist and their UGT2B15 expression was completely lost, as its mRNA levels proliferation was only 15% and 22.5% reduced in the presence þ in vehicle- and DHT-treated DOX cells were not signifi- of the drug after 3 and 4 days, respectively (Fig. 6B). Thus, cantly different (Fig. 5A). In contrast, UGT2B17 mRNA levels siRNA-mediated knockdown of UGT2B15 and UGT2B17 were found significantly more abundant in DHT-treated expression provoked 40% and 32% reductions of the antipro- þ than in vehicle DOX cells (Fig. 5B). Also surprising was liferative effects of bicalutamide in LNCaP cells exposed to the the observation that bicalutamide reduces UGT2B15 and drug for 3 and 4 days, respectively. UGT2B17 mRNA levels in cells cultured with DOX (Fig. 5). This last observation suggests that bicalutamide exerts opposite effects on UGT genes' expression in the presence Discussion þ of high (DOX )orlow(DOX ) AR levels. However, when This study evidences a novel and incidental effect by which mRNA levels were compared with those quantified in bica- androgen ablation drugs abolish the ability of carcinogenic lutamide-treated native cells, only the strong reduction in androgens to inhibit their own inactivation in prostate cancer UGT2B17 transcripts remained statistically significant (Fig. cells. This first comprehensive analysis of AR blockade con- 5B), indicating that the lossofARmainlyalteredthe sequences for androgen glucuronidation identifies the two response of this enzyme to bicalutamide. androgen-conjugating UGT2B15 and UGT2B17 enzymes as Nonetheless, the fact that AR knockdown abolishes or ADTh-positive targets, and establishes the major contribution reverses the effects of bicalutamide confirms the role played of these effects to the antiproliferative properties of ADTh by this receptor in the antiandrogen-induced upregulation of drugs. UGT2B15 and UGT2B17 expression. An interesting observation of the present study is the differential manner in which the 2 UGT genes respond to AR Knockdown of UGT2B15 and UGT2B17 reduces the blockade in and ex vivo. These differences are thought to reflect antiproliferative properties of bicalutamide variations in the AR-dependent modulation of the two UGT We next evaluated whether induction of UGT2B15 and genes. Indeed, the fact that AR agonists/antagonists fail to UGT2B17 expression participates in the antiproliferative prop- modulate UGT2B15 expression in receptor-deficient cells, erties of bicalutamide (31). For this purpose, LNCaP cells were while having inverse effects on UGT2B17 mRNA levels, sug- transfected with a nontarget (siRNA control) or the anti- gests that alternative regulatory processes drastically modify UGT2B15/UGT2B17 siRNA probe (Fig. 6A; ref. 14), and then the AR-dependent modulation of UGT2B17 expression when exposed to vehicle (DMSO) or bicalutamide (10 mmol/L) for up the androgen receptor is reduced. Although thorough inv- to 4 days (Fig. 6B). As expected (31), bicalutamide caused a estigations are required to fully grasp the mechanisms govern- significant 24.5% reduction of control cells' proliferation after 3 ing the molecular switch in the UGT2B17 response, it can be days of exposure. This reduction was further enhanced to envisioned that such UGT-specific processes may be derived 32.5% inhibition after 4 days (Fig. 6B). Interestingly, UGT2B- from nucleic acid differences in the transcriptional regulatory

OF6 Cancer Res; 73(23) December 1, 2013 Cancer Research

Ablation Therapy Improves Androgen Glucuronidation in Prostate Cancer

Figure 6. UGT2B15- and UGT2B17-deficient LNCaP cells are less sensitive to the antiproliferative effects of bicalutamide. A and B, LNCaP cells were transiently transfected with a nontarget (control) or an anti-UGT2B15/UGT2B17 siRNA probe (UGT deficient) using lipofectin for 6 hours, allowed 18 hours for recovery, and then exposed to vehicle (DMSO) or bicalutamide (10 mmol/L) for up to 2, 3, or 4 days (D). A, UGT2B15 and UGT2B17 mRNA levels were quantified from total RNA through quantitative RT-PCR analyses and normalized with the housekeeping RNA 36B4. Values (mean SD) are expressed relatively to control (nontarget siRNA probe) set at 1. B, cell viability was determined using the CellTiter 96 AQueous One Solution Cell Proliferation Assay Kit. Proliferation rate values (mean SD) are expressed as number of cells per well. Percentages of reduction in cell proliferation were determined. Statistically significant differences between siRNA control and siRNA UGT2B15/B17-transfected cells are indicated by asterisks (Student t test: , P < 0.05; , P < 0.01; , P < 0.001; NS, not significant). regions of the UGT2B15 and UGT2B17 genes. Even if the two ence that UGT2B15 and UGT2B17 polymorphisms exert on genes share considerable sequence homology (9), previous tissue androgen levels (36). The loss of UGT genes' sensitivity studies already illustrated that minor differences in their may allow prostate cancer cells to maintain sufficient DHT proximal promoters are responsible for their differential levels to activate AR and its regulated genes in CRPCs (37, 38). response to various regulatory pathways (32, 23). Such changes On the basis of these observations, one can conclude that the actually exist within the AR response elements identified in UGT induction in short-term treated samples is an important UGT2B15 and UGT2B17 promoter sequences (20), and thus biochemical component of the initial ADTh benefits (39). may play a role in their isoform-specific response under low AR Following on that point, it can be envisioned that UGT2B15 levels. Other regulatory pathways may also be involved. For and UGT2B17 are actually underexploited therapeutic targets example, EGF signaling, which downregulates UGT2B17 for androgen deprivation therapies. Until now, antiandrogenic expression without affecting UGT2B15 in LNCaP cells (23), approaches have been focusing on androgen biosynthesis and has been identified as a transduction pathway involved in activity (39). The present study validates the potential of prostate tumor growth (33, 34). targeting glucuronidation to reduce active androgen concen- Beyond these mechanistic considerations, the present trations in prostate cancer cells, thus providing a strong results are of clinical significance for prostate cancer progres- rational for searching pharmacologic agents that stimulate sion, treatment, and diagnosis. In terms of tumor progression, androgen glucuronidation in prostate cancer cells. However, the desensitization of the UGT response to ADTh, as revealed such an alternative strategy requires the identification of AR- by the similar protein levels detected in untreated tumors and independent UGT2B15 and UGT2B17 gene–inducers. CRPCs, suggests that the loss of UGT induction is involved in Thepresentworkmayalsohaveasignificant impact in tumors' transition to castration resistance. Even if resistant to terms of prostate cancer prognosis and diagnosis. Indeed, ablation therapies, CRPC progression remains dependent on the respective changes in UGT2B15 and UGT2B17 protein the androgen axis (8, 35). Because AR acts as a negative levels in response to ADTh identify these two proteins as UGT2B15 and UGT2B17 gene–regulator (14, 20, 21), the loss potential biomarkers both for evaluating the responsiveness of UGT induction detected in vivo may actually reflect an AR to antiandrogen drug therapies and preventing resistance reactivation in resistant tumors. On the other hand, as occurrence. ADTh remains the cornerstone of systemic observed with UGT knockdown experiments, the induction treatments for locally advanced prostate cancer (38), and of androgen glucuronidation may contribute to the antiproli- biochemical predictors are currently needed to assess ferative effects of AR antagonists. It is therefore tempting to whether additional treatments should be initiated (40). speculate that the loss of this induction is likely involved in the Furthermore, patients with CRPC have a poor prognosis phenotypic changes allowing tumor cells to proliferate in the and account for the majority of prostate cancer–related presence of low androgen levels, as observed in CRPCs (8). This deaths (39), thus being able to anticipate that ADTh resis- idea is supported by the inverse relationship existing between tance can be helpful in adapting treatment settings with new UGT2B15 and UGT2B17 expression and androgen-dependent effective antiandrogenic agents to prevent the development prostate cancer cell proliferation (14) and by the strong influ- of castrate-resistant tumors (39). A continuous follow-up of

www.aacrjournals.org Cancer Res; 73(23) December 1, 2013 OF7

Grosse et al.

UGT2B15 and/or UGT2B17 intraprostatic protein levels Authors' Contributions from the time of diagnosis constitutes a potential approach Conception and design: L. Grosse, O. Barbier Development of methodology: L. Grosse, P.S. Rennie, A. Belanger, O. Barbier to evaluate therapy responsiveness, and measuring Acquisition of data (provided animals, acquired and managed patients, UGT2B17 levels may also be helpful to anticipate the loss provided facilities, etc.): L. Grosse, S. Paquet,^ P. Caron, P.S. Rennie, Analysis and interpretation of data (e.g., statistical analysis, biostatistics, of optimal ADTh effects in UGT2B17-positive patients, and computational analysis): L. Grosse, S. P^aquet, A. Belanger, O. Barbier then initiate alternative therapeutics, even before the resis- Writing, review, and/or revision of the manuscript: L. Grosse, S. Paquet,^ tance occurrence. O. Barbier Administrative, technical, or material support (i.e., reporting or orga- The earlier-discussed diagnostic and therapeutic opportu- nizing data, constructing databases): L. Grosse, P. Caron, O. Barbier nities are mainly based on results from the tissue microarray Study supervision: L. Grosse, O. Barbier study, while largely supported by results from functional Pathology-TMA design and IHC evaluation: L. Fazli, investigations. However, sample heterogeneity is a major Acknowledgments problem in tissue array studies, especially in prostate cancer, The authors thank Dr. Virginie Bocher for critical reading of the manuscript, in which patients receive varied drugs or combination of drugs, and Dr. Eric Levesque for helpful discussion. and the tumor areas are small and often surrounded by normal cells (40). Therefore, the role of UGT2B15 and UGT2B17 as drug Grant Support targets and/or molecular biomarkers remains to be validated This study was supported by grants from the Canadian Institutes of Health Research (CIHR), the Terry Fox foundation, and the Canadian Foundation for through large and accurate clinical settings. Innovation. S. Paquet^ is holder of scholarships from CIHR and the Fonds pour la Nevertheless, the combination of clinical and functional Recherche en SanteduQuebec (FRSQ). L. Grosse is supported by a scholarship approaches used for the current investigation demonstrates from the Fonds pour l'Enseignement et la Recherche (FER) from the Faculty of Pharmacy, Laval University (Quebec). O. Barbier is holder of salary grants that ADTh drugs stimulate the local androgen metabolism in from the CIHR (New Investigator Award #MSH95330) and FRSQ (Junior 2 prostate cells, and identiﬁes UGT2B15 and UGT2B17 as poten- Scholarship). tial anticancer drug targets and/or biomarkers for androgen The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in ablation responsiveness and maintenance. accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Disclosure of Potential Conﬂicts of Interest Received May 22, 2013; revised August 20, 2013; accepted September 3, 2013; No potential conﬂicts of interest were disclosed. published OnlineFirst October 11, 2013.

References 1. Jemal A, Center MM, DeSantis C, Ward EM. Global patterns of cancer the androgen response in prostate cancer LNCaP cells. J Biol Chem incidence and mortality rates and trends. Cancer Epidemiol Biomar- 2007;282:33466–74. kers Prev 2010;19:1893–907. 15. Chouinard S, Yueh MF, Tukey RH, Giton F, Fiet J, Pelletier G, et al. 2. Konishi N, Shimada K, Ishida E, Nakamura M. Molecular pathology of Inactivation by UDP-glucuronosyltransferase enzymes: the end of prostate cancer. Pathol Int 2005;55:531–9. androgen signaling. J Steroid Biochem Mol Biol 2008;109:247–53. 3. Shaneyfelt T, Husein R, Bubley G, MAntzoros CS. Hormonal predic- 16. Nadeau G, Bellemare J, Audet-Walsh E, Flageole C, Huang SP, Bao tores of prostate cancer: a meta-analysis. J Clin Oncol 2000;18: BY, et al. Deletions of the androgen-metabolizing UGT2B genes have 847–53. an effect on circulating steroid levels and biochemical recurrence after 4. Sharifi N, Gulley JL, Dahut WL. Androgen deprivation therapy for radical prostatectomy in localized prostate cancer. J Clin Endocrinol prostate cancer. JAMA 2005;294:238–44. Metab 2011;96:E1550–E7. 5. Hadaschik BA, Gleave ME. Therapeutic options for hormone-refrac- 17. Park J, Chen L, Shade K, Lazarus P, Seigne J, Patterson S, et al. tory prostate cancer in 2007. Urol Oncol 2007;25:413–9. Asp85tyr polymorphism in the udp-glucuronosyltransferase (UGT) 6. Lam JS, Leppert JT, Vemulapalli SN, Shvarts O, Belldegrun AS. 2B15 gene and the risk of prostate cancer. J Urol 2004;171:2484–8. Secondary hormonal therapy for advanced prostate cancer. J Urol 18. Park JY, Tanner JP, Sellers TA, Huang Y, Stevens CK, Dossett N, et al. 2006;175:27–34. Association between polymorphisms in HSD3B1 and UGT2B17 and 7. Arianayagam M, Chang J, Rashid P. Chemotherapy in the treatment of prostate cancer risk. Urology 2007;70:374–9. prostate cancer–is there a role? Aust Fam Physician 2007;36:737–9. 19. Grant DJ, Hoyo C, Oliver SD, Gerber L, Shuler K, Calloway E, et al. 8. Balk SP, Knudsen KE. AR, the cell cycle, and prostate cancer. Nucl Association of uridine diphosphate-glucuronosyltransferase 2B gene Recept Signal 2008;6:e001. variants with serum glucuronide levels and prostate cancer risk. Genet 9. Belanger A, Pelletier G, Labrie F, Barbier O, Chouinard S. Inactivation Test Mol Biomarkers 2013;17:3–9. of androgens by UDP-glucuronosyltransferase enzymes in humans. 20. Bao BY, Chuang BF, Wang Q, Sartor O, Balk SP, Brown M, et al. Trends Endocrinol Metab 2003;14:473–9. Androgen receptor mediates the expression of UDP-glucuronosyl- 10. Hsing AW, Reichardt JKV, Stanczyk FZ. Hormones and prostate transferase 2 B15 and B17 genes. Prostate 2008;68:839–48. cancer: current perspectives and future directions. Prostate 21. Chouinard S, Pelletier G, Belanger A, Barbier O. Isoform-specific 2002;52:213–35. regulation of uridine diphosphate-glucuronosyltransferase 2B 11. Guillemette C, Belanger A. Glucuronosyltransferase activity in human enzymes in the human prostate: differential consequences for andro- cancer cell line LNCaP. Mol Cell Endocrinol 1995;107:131–9. gen and bioactive lipid inactivation. Endocrinology 2006;147:5431–42. 12. Guillemette C, Hum D, Belanger A. Regulation of steroid glucur- 22. Chouinard S, Pelletier G, Belanger A, Barbier O. Cellular specific onosyltransferase activities and transcripts by androgen in the expression of the androgen-conjugating enzymes UGT2B15 and human prostatic cancer LNCaP cell line. Endocrinology 1996;137: UGT2B17 in the human prostate epithelium. Endocr Res 2004;30: 2872–9. 717–25. 13. Dutton GJ. Glucuronidation of drugs and other compounds. Boca 23. Guillemette C, Levesque E, Beaulieu M, Turgeon D, Hum DW, Belanger Raton, FL: CRC Press; 1980. A. Differential Regulation of Two Uridine Diphospho-Glucuronosyl- 14. Chouinard S, Barbier O, Belanger A. UDP-glucuronosyltransferase transferases, UGT2B15 and UGT2B17, in Human Prostate LNCaP 2B15 (UGT2B15) and UGT2B17 enzymes are major determinants of Cells. Endocrinology 1997;138:2998–3005.

OF8 Cancer Res; 73(23) December 1, 2013 Cancer Research

Ablation Therapy Improves Androgen Glucuronidation in Prostate Cancer

24. Thomas C, Zoubeidi A, Kuruma H, Fazli L, Lamoureux F, Beraldi E, et al. 32. Kaeding J, Belanger J, Caron P, Verreault M, Belanger A, Barbier O. Transcription factor Stat5 knockdown enhances androgen receptor Calcitrol (1alpha,25-dihydroxyvitamin D3) inhibits androgen glucuro- degradation and delays castration-resistant prostate cancer progres- nidation in prostate cancer cells. Mol Cancer Ther 2008;7:380–90. sion in vivo. Mol Cancer Ther 2011;10:347–59. 33. Oosterhoff JK, Grootegoed JA, Blok LJ. Expression proﬁling of andro- 25. Paquet^ S, Fazli L, Grosse L, Verreault M, Tetu B, Rennie PS, et al. gen-dependent and -independent LNCaP cells: EGF versus androgen Differential expression of the androgen-conjugating UGT2B15 and signalling. Endocr Relat Cancer 2005;12:135–48. UGT2B17 enzymes in prostate tumor cells during cancer progression. 34. Traish AM, Morgentaler A. Epidermal growth factor receptor expres- J Clin Endocrinol Metab 2012;97:E423–E32. sion escapes androgen regulation in prostate cancer: a potential 26. Chai J, Norng M, Modak C, Reavis KM, Mouazzen W, Pham J. CCN1 molecular switch for tumour growth. Br J Cancer 2009;101:1949–56. induces a reversible epithelial-mesenchymal transition in gastric epi- 35. Attard G, Richards J, de Bono JS. New strategies in metastatic thelial cells. Lab Invest 2010;90:1140–51. prostate cancer: targeting the androgen receptor signaling pathway. 27. Cheng H, Snoek R, Ghaidi F, Cox ME, Rennie PS. Short hairpin RNA Clin Cancer Res 2011;17:1649–957. knockdown of the androgen receptor attenuates ligand-indepen- 36. Olsson M, Ekstrom L, Guillemette C, Belanger A, Rane A, Gustafsson dent activation and delays tumor progression. Cancer Res 2006; O. Correlation between circulatory, local prostatic, and intra-prostatic 66:10613–20. androgen levels. Prostate 2011;71:909–14. 28. Menard V, Eap O, Harvey M, Guillemette C, Levesque E. Copy-number 37. Montgomery RB, Mostaghel EA, Vessella R, Hess DL, Kalhorn TF, variations (CNVs) of the human sex steroid metabolizing genes Higano CS, et al. Maintenance of intratumoral androgens in metastatic UGT2B17 and UGT2B28 and their associations with a UGT2B15 prostate cancer: a mechanism for castration-resistant tumor growth. functional polymorphism. Hum Mutat 2009;30:1310–9. Cancer Res 2008;68:4447–54. 29. Satelli A, Li S. Vimentin in cancer and its potential as a molecular target 38. Mostaghel EA, Page ST, Lin DW, Fazli L, Coleman IM, True LD, et al. for cancer therapy. Cell Mol Life Sci 2011;68:3033–46. Intraprostatic androgens and androgen-regulated gene expression 30. Veldscholte J, Berrevoets CA, Brinkmann AO, Grootegoed JA, Mulder persist after testosterone suppression: therapeutic implications for E. Anti-androgens and the mutated androgen receptor of LNCaP cells: castration-resistant prostate cancer. Cancer Res 2007;67:5033–41. differential effects on binding afﬁnity, heat-shock protein interaction, 39. Ferraldeschi R, Pezaro C, Karavasilis V, de Bono J. Abiraterone and and transcription activation. Biochemistry 1992;31:2393–9. novel antiandrogens: overcoming castration resistance in prostate 31. Colquhoun AJ, Venier NA, Vandersluis AD, Besla R, Sugar LM, Kiss A, cancer. Annu Rev Med 2013;64:1–13. et al. Metformin enhances the antiproliferative and apoptotic effect of 40. Lehmusvaara S, Erkkila T, Urbanucci A, Waltering K, Seppala J, Larjo bicalutamide in prostate cancer. Prostate Cancer Prostatic Dis A, et al. Chemical castration and anti-androgens induce differential 2012;15:346–52. gene expression in prostate cancer. J Pathol 2012;227:336–45.

www.aacrjournals.org Cancer Res; 73(23) December 1, 2013 OF9

Androgen Glucuronidation: An Unexpected Target for Androgen Deprivation Therapy, with Prognosis and Diagnostic Implications

Laurent Grosse, Sophie Pâquet, Patrick Caron, et al.

Cancer Res Published OnlineFirst October 11, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-1462

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/10/11/0008-5472.CAN-13-1462.DC1

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2013/11/21/0008-5472.CAN-13-1462. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.