Influence of the nuclear hormone receptor axis in the progression and treatment of hormone dependent cancers

A dissertation submitted to the Division of Research and Advanced Studies at the University of Cincinnati in partial fulfillment of the requirements for the degree of

Doctorate of Philosophy (Ph.D.)

In the Department of Cell and Cancer Biology of the College of Medicine

2007

by

Janet K. Hess-Wilson

B.A., Wittenberg University, 2000

Committee Chair: Karen E. Knudsen, Ph.D. Abstract

Due to its pivotal role in prostatic growth and survival, the androgen receptor

(AR) is the primary target of disseminated prostate cancer (CaP), as achieved via androgen deprivation therapy (ADT). Unfortunately, ADT is circumvented by restoration of AR activity, resulting in ADT resistant tumors for which there is no alternate treatment option. Through multiple mechanisms, reactivation of the AR specifically underlies the progression to therapy resistant tumors. The environmentally prevalent endocrine disrupting compound (EDC), (BPA) is able to activate specific somatically mutated ARs commonly found in CaP, resulting in androgen-independent proliferation of CaP cells. To directly assess the effect of BPA on ADT, we used an in vivo xenograft model of CaP that expresses a BPA-sensitive mutant AR, and mimics standard human cytotoxic response to ADT, followed by subsequent tumor re-growth. When tumor- bearing animals were exposed to environmentally relevant levels of BPA during ADT, the tumors failed therapy more rapidly (compared to placebo controls), with AR re- activation and concomitant increased tumor cell proliferation. These data suggest that environmentally relevant exposure to EDCs may reduce the efficacy of mainline ADT for

CaP. We next determined that the environmentally persistent pesticide, DDE, was able to activate select AR mutants, and in in vitro models of CaP, DDE induces cellular proliferation in the absence of androgen, demonstrating that this observed response was not unique to BPA. Strikingly, the mechanism of DDE impact on CaP cells was distinct from BPA, in that at low doses this agent also activated the MAPK pathway, and requires this activation for mitogenesis. Given the context specific impact of AR activation by EDCs, and the deleterious effect of exposure to BPA on ADT in vivo, we next addressed the impact of AR action on taxane-based therapy. These cytotoxic agents have recently been shown to improve survival outcome for patients with advanced CaP, and are potentially new second line therapies, however the impact of

AR action on this cytotoxic modality had yet to be assessed. Contrary to the stigma of

AR as a survival factor, we found that AR activation by both endogenous and exogenous agonists (i.e. DHT and BPA), synergized with taxanes to decrease cell survival, through p53-mediated, caspase dependent apoptosis. This synergistic action was attributed directly to the AR-dependent mitogenic capacity of these agents.

Importantly, these data further support the conclusion that the environmental impact on

AR action and CaP therapeutic outcome is context specific. We further evaluated the complexity of EDC affects by assessing the impact of these agents under clinically relevant conditions in another hormone-dependent tissue, breast cancer. BPA and the , (COU), were able to activate the receptor (ERα), but this activation was restricted to estrogen-depleted conditions, and could be blocked by the standard ER antagonist, . Additionally, BPA and COU have disparate affects in multiple analyses including mutant ERα activation, and specific coactivator deregulation. In conclusion, the environmental impact on nuclear receptor signaling and hormone dependent cancer progression and treatment is highly molecular context specific. Therefore, the impact of EDCs on hormone dependent cancer treatment needs to be identified under specific and well defined, clinically relevant, molecular contexts.

Acknowledgements

I would like to thank my family, friends, and co-workers for their support and encouragement. I would like to especially thank my wife, Carrie, and daughter Kaia, for their patience, understanding, and unwavering support. They are my daily examples of the truly important things in life that no publication or degree could replace. My family, parents and brother have shown me unconditional love and support, and without that, I would not have had the confidence or stamina to push through. In addition, I could not have succeeded without the help from my fellow lab mates, past and present – thanks to them for their constructive criticism, technical support, and friendships. I also want to thank my committee members for their ideas, comments and guidance. Finally, I would like to thank my advisor, Karen Knudsen, for everything that she has taught me over the last four and a half years, and for all of the opportunities that she has provided me. I truly admire her knowledge, strength, commitment to science and her lab, and who she is as a person. Thank you!  Table of Contents

List of Figures and Tables ………………………………………………………….. 2

Chapter I: Introduction A: Prostate cancer: the role of the androgen receptor (AR) in disease progression and treatment…………………………………………… 4 B: Endocrine disrupting compounds and prostate cancer……………….. 11

Chapter II: Bisphenol A facilitates bypass of androgen ablation therapy in prostate cancer A: Abstract…………………………………………………………………….. 24 B: Introduction………………………………………………………………… 24 C: Materials and Methods…………………………………………………… 25 D: Results……………………………………………………………………… 27 E: Discussion………………………………………………………………….. 29 F: Acknowledgements……………………………………………………….. 32 G: References………………………………………………………………… 32

Chapter III: DDE modulates proliferation of prostate cancer cells through MAPK and mutant AR pathways A: Abstract…………………………………………………………………….. 35 B: Introduction………………………………………………………………… 38 C: Materials and Methods…………………………………………………… 40 D: Results……………………………………………………………………… 45 E: Discussion………………………………………………………………….. 57 F: Acknowledgements……………………………………………………….. 61 G: References………………………………………………………………… 61

Chapter IV: Mitogenic action of the androgen receptor sensitizes prostate cancer cells to taxane-based cytotoxic insult A: Abstract…………………………………………………………………….. 75 B: Introduction………………………………………………………………… 75 C: Materials and Methods…………………………………………………… 75 D: Results……………………………………………………………………… 76 E: Discussion…………………………………………………………………. 82 F: Acknowledgements…………………………………………………….…. 84 G: References………………………………………………………………… 84

Chapter V: action in breast cancer: impact on ER-dependent transcription and mitogenesis A: Abstract…………………………………………………………………….. 87 B: Introduction………………………………………………………………… 87 C: Materials and Methods…………………………………………………… 88 D: Results……………………………………………………………………… 90 E: Discussion………………………………………………………………….. 95 F: Acknowledgements……………………………………………………….. 98 G: References………………………………………………………………… 98

Chapter VI: Summary and Future Directions…………………………………….. 102

List of Tables and Figures Chapter I-A: Figure 1 ……………………………………………………………………… 5 Figure 2 ……………………………………………………………………… 7

Chapter I-B: Figure 1 ……………………………………………………………………… 12 Table 1.……………………………………………………………………… 14 Figure 2 ……………………………………………………………………… 16

Chapter II: Figure 1 ……………………………………………………………………… 27 Figure 2 ……………………………………………………………………… 28 Figure 3 ……………………………………………………………………… 28 Figure 4 ……………………………………………………………………… 30 Figure 5 ……………………………………………………………………… 31

Chapter III: Figure 1 ……………………………………………………………………… 67 Figure 2 ……………………………………………………………………… 68 Figure 3 ……………………………………………………………………… 69 Figure 4 ……………………………………………………………………… 70 Figure 5 ……………………………………………………………………… 71 Figure 6………………………………………………………………………. 72 Figure 7……………………………………………………………………….. 73

Chapter IV: Figure 1 ……………………………………………………………………… 77 Figure 2 ……………………………………………………………………… 78 Figure 3 ……………………………………………………………………… 79 Figure 4 ……………………………………………………………………… 80 Figure 5 ……………………………………………………………………… 82 Figure 6………………………………………………………………………. 83

Chapter V: Figure 1 ……………………………………………………………………… 89 Figure 2 ……………………………………………………………………… 91 Figure 3 ……………………………………………………………………… 92 Table 1 ……………………………………………………………………….. 93 Figure 4 ……………………………………………………………………… 93 Figure 5………………………………………………………………………. 94 Figure 6………………………………………………………………………. 95 Table 2 ………………………………………………………………………. 95

Chapter VI: Figure 1...... 106

Chapter I: Introduction

A. Prostate cancer: the role of the androgen receptor (AR) in disease progression and treatment

B. Endocrine disrupting compounds and prostate cancer

I-A: Prostate cancer: the role of the androgen receptor (AR) in disease progression and treatment

Prostate Cancer Prevalence and Risk:

In the United States, prostate cancer (CaP) is the most commonly diagnosed malignancy, and the second leading cause of cancer related deaths in men (1). The causes of CaP remain undefined; however, the most significant risk factor is age (2).

CaP is regulated to older men, as incidence of prostate cancer in men under the age of

40 is rare and the risk for development of this disease in men over the age of 65 is a striking 1 in 9 (2). Other known risk factors include race, diet, family history and exposure to environmental contaminants. African American men have the highest incidence of CaP; however in general, all Western populations have the highest prevalence, most likely due to diet, life span, and overall higher health screening practices (3, 4).

Prostate Cancer Therapy:

For early detected and organ confined CaP, therapy is highly successful.

Options for these patients include radical prostatectomy and/or radiation based therapies, and survival rates reach near 100% (5, 6). Conversely, for disseminated disease there are few treatment options. The first line of therapeutic intervention for these tumors is androgen deprivation therapy (ADT). Various treatment methods of

ADT include limiting endogenous androgen levels through gonadotropin release hormone (GnRH) antagonists and/or direct antagonists of AR (bicalutamide/Casodex)

(7, 8), which induce a molecular paradigm of target gene repression through repressor complex formation. ADT is initially successful, however, within a median time of 18-24

months, the tumors will inevitably reform (8). ADT is continuously circumvented by restoration of androgen receptor (AR) activity, an event central in the development of

ADT-resistant tumors (9, 10). Sadly, for these patients with ADT-resistant tumors, there is currently no treatment option other than palliative therapies.

The androgen receptor:

Charles Huggins published the Nobel Prize winning finding in 1941 that removing the male androgen ( and , DHT) results in

CaP tumor regression in castrated dogs. It has since been known that in humans, androgen deprivation therapy (ADT) manifest at the cellular level by proliferative arrest and apoptosis (8). Hormone dependent tumors initially respond to ADT due to the

requirement of prostatic Figure 1 epithelium (and malignant

derivation there of) for androgen

signaling. Androgens are

required for the normal growth

and function of the prostate, and

mediate their biological activity

through activation of the

androgen receptor (AR) (10).

Both testosterone and DHT can bind and activate the AR, however DHT forms a more stable complex and has a higher binding affinity for AR (11). AR is a member of the class I hormone receptors, which are ligand-induced transcription factors. Once bound by ligand, active AR translocates into the nucleus, where it binds androgen

responsive elements (AREs) in the regulatory regions of target genes, recruits a series of coactivators, and induces a program of gene transcription (10) (Figure 1). The biological outcome of AR activation is divergent dependent on ligand bound, the milieu of coactivators and cellular context. Although, the specific profile of AR target genes has yet to be well defined in CaP, PSA (prostate specific antigen; a protease) is a serum marker for prostate cancer progression. PSA is utilized clinically to monitor disease progression, as the serum levels of PSA are thought to directly correlate to AR activity and therefore therapeutic efficacy (12). The overall affect of androgens in AR positive cells is an increase in cyclin-dependent kinase activity and stimulation for cells to enter S phase of the cell cycle, thereby facilitating cell proliferation (13). However, it is clear that AR transcriptional regulation of target genes is critical in prostate cancer progression, as directly blocking AR transcriptional activity through antagonists successfully inhibits disease progression (in early stage disease) (14). Additionally, in

CaP tissues from patients who failed first line ADT, AR is nuclear (indicative of an active transcription factor) and PSA secretion is re-elevated (9).

Hormone Independent CaP:

The question of how prostate cancer cells evade ADT, forming hormone independent tumors is the focus of most the research in prostate cancer. It is clear that tumors maintain AR expression, as the AR is present in the majority of CaP both at the primary and metastatic sites (15, 16). At least four mechanisms of AR reactivation and

ADT resistance have been elucidated (Figure 2). First, it is know that the AR gene can become amplified, increasing the amount of AR in the cell and rendering AR hypersensitive to lower levels of androgen (17-19). Increase in AR (mRNA and protein)

is necessary and proficient to transition CaP to hormone independent and ADT-resistant disease (17) and reducing AR levels inhibits proliferation and delays tumor progression

(20, 21). Secondly, ligand independent, growth factor pathways (e.g. IL-6, IGF, EGF) can be altered in ADT-resistant cells, which may facilitate inappropriate AR action and

disease progression (22). Third, Figure 2 deregulation of AR coactivators

has been observed, and it is

hypothesized that this

deregulation may sensitize AR to

low ligand concentrations (16, 23).

Lastly, somatic mutations within

the AR ligand-binding domain

arise in between 8-25% of

hormone independent tumors,

which cause broadened ligand

specificity, specifically enabling

activation by alternate ligands,

such as estrogen, and cortisole (24-29). The oncogenic nature of AR, itself, is evident in that altered AR function and/or expression is sufficient to induce, and correlates with, prostate cancer initiation, invasion and metastatsis (18, 30, 31). Importantly, regardless of the mechanistic pathway, in tumors that are hormone independent and ADT-resistant, in the absence of endogenous androgen, AR is reactivated (Figure 2), as evident by nuclear

staining of the receptor and reactivation of AR targets genes (PSA) (9, 15, 16).. These findings clearly define AR as the regulator of prostate cancer progression. Therefore, it is essential to explore possible modes of AR activation in the absence of androgen to elucidate how these tumors re-emerge. As there are limited other treatment options currently available for patients with hormone independent tumors, it is imperative to determine how AR signaling maybe be impacting therapeutic strategies.

1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55(1):10-30. 2. Bostwick DG, Burke HB, Djakiew D, et al. Human prostate cancer risk factors. Cancer 2004;101(10 Suppl):2371-490. 3. Hayes RB, Ziegler RG, Gridley G, et al. Dietary factors and risks for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev 1999;8(1):25-34. 4. Reddy S, Shapiro M, Morton R, Jr., Brawley OW. Prostate cancer in black and white Americans. Cancer Metastasis Rev 2003;22(1):83-6. 5. Damber JE, Khatami A. Surgical treatment of localized prostate cancer. Acta Oncol 2005;44(6):599-604. 6. Kirkpatrick JP, Anscher MS. Radiotherapy for locally recurrent prostate cancer. Clin Adv Hematol Oncol 2005;3(12):933-42. 7. Dehm SM, Tindall DJ. Molecular regulation of androgen action in prostate cancer. J Cell Biochem 2006. 8. Denmeade SR, Isaacs JT. A history of prostate cancer treatment. Nat Rev Cancer 2002;2(5):389-96. 9. Feldman BJ, Feldman D. The development of androgen-independent prostate cancer. Nat Rev Cancer 2001;1(1):34-45. 10. Trapman J, Brinkmann AO. The androgen receptor in prostate cancer. Pathol Res Pract 1996;192(7):752-60. 11. Russell DW, Berman DM, Bryant JT, et al. The molecular genetics of steroid 5 alpha-reductases. Recent Prog Horm Res 1994;49:275-84. 12. Stephan C, Jung K, Diamandis EP, Rittenhouse HG, Lein M, Loening SA. Prostate-specific antigen, its molecular forms, and other kallikrein markers for detection of prostate cancer. Urology 2002;59(1):2-8. 13. Knudsen KE, Arden KC, Cavenee WK. Multiple G1 regulatory elements control the androgen-dependent proliferation of prostatic carcinoma cells. J Biol Chem 1998;273(32):20213-22.

14. Reid P, Kantoff P, Oh W. Antiandrogens in prostate cancer. Invest New Drugs 1999;17(3):271-84. 15. Gregory CW, Hamil KG, Kim D, et al. Androgen receptor expression in androgen- independent prostate cancer is associated with increased expression of androgen- regulated genes. Cancer Res 1998;58(24):5718-24. 16. Gregory CW, He B, Johnson RT, et al. A mechanism for androgen receptor- mediated prostate cancer recurrence after androgen deprivation therapy. Cancer Res 2001;61(11):4315-9. 17. Chen CD, Welsbie DS, Tran C, et al. Molecular determinants of resistance to antiandrogen therapy. Nat Med 2004;10(1):33-9. 18. Koivisto P, Kononen J, Palmberg C, et al. Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation therapy failure in prostate cancer. Cancer Res 1997;57(2):314-9. 19. Visakorpi T, Hyytinen E, Koivisto P, et al. In vivo amplification of the androgen receptor gene and progression of human prostate cancer. Nat Genet 1995;9(4):401-6. 20. Cheng H, Snoek R, Ghaidi F, Cox ME, Rennie PS. Short hairpin RNA knockdown of the androgen receptor attenuates ligand-independent activation and delays tumor progression. Cancer Res 2006;66(21):10613-20. 21. Haag P, Bektic J, Bartsch G, Klocker H, Eder IE. Androgen receptor down regulation by small interference RNA induces cell growth inhibition in androgen sensitive as well as in androgen independent prostate cancer cells. J Steroid Biochem Mol Biol 2005;96(3-4):251-8. 22. Craft N, Shostak Y, Carey M, Sawyers CL. A mechanism for hormone- independent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase. Nat Med 1999;5(3):280-5. 23. Fujimoto N, Mizokami A, Harada S, Matsumoto T. Different expression of androgen receptor coactivators in human prostate. Urology 2001;58(2):289-94. 24. Bohl CE, Gao W, Miller DD, Bell CE, Dalton JT. Structural basis for antagonism and resistance of bicalutamide in prostate cancer. Proc Natl Acad Sci U S A 2005;102(17):6201-6. 25. Culig Z, Hobisch A, Cronauer MV, et al. Mutant androgen receptor detected in an advanced-stage prostatic carcinoma is activated by adrenal androgens and progesterone. Mol Endocrinol 1993;7(12):1541-50. 26. Fenton MA, Shuster TD, Fertig AM, et al. Functional characterization of mutant androgen receptors from androgen-independent prostate cancer. Clin Cancer Res 1997;3(8):1383-8. 27. Gottlieb B, Beitel LK, Wu JH, Trifiro M. The androgen receptor gene mutations database (ARDB): 2004 update. Hum Mutat 2004;23(6):527-33. 28. Hara T, Miyazaki J, Araki H, et al. Novel mutations of androgen receptor: a possible mechanism of bicalutamide withdrawal syndrome. Cancer Res 2003;63(1):149- 53. 29. Suzuki H, Sato N, Watabe Y, Masai M, Seino S, Shimazaki J. Androgen receptor gene mutations in human prostate cancer. J Steroid Biochem Mol Biol 1993;46(6):759- 65.

30. Han G, Buchanan G, Ittmann M, et al. Mutation of the androgen receptor causes oncogenic transformation of the prostate. Proc Natl Acad Sci U S A 2005;102(4):1151- 6. 31. Tilley WD, Buchanan G, Hickey TE, Bentel JM. Mutations in the androgen receptor gene are associated with progression of human prostate cancer to androgen independence. Clin Cancer Res 1996;2(2):277-85.

Cancer Letters 241 (2006) 1–12 www.elsevier.com/locate/canlet Mini-review Endocrine disrupting compounds and prostate cancer

J.K. Hess-Wilson a, K.E. Knudsen a,b,* a Department of Cell Biology, University of Cincinnati College of Medicine, P.O. Box 670521, 3125 Eden Ave., Cincinnati, OH 45267-0521, USA b Center for Environmental Genetics, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0521, USA

Received 26 September 2005; accepted 7 October 2005

Abstract

Prostate cancer is a major health concern and is treated based on its hormone dependence. Agents that alter hormone action can have substantial biological effects on prostate cancer development and progression. As such, there is significant interest in uncovering the potential effects of endocrine disrupting compound (EDC) exposure on prostate cancer. The present review is focused on agents that alter hormone action in the prostate and how they may impact cancer growth or treatment. q 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Endocrine disrupting compound; Prostate cancer; Estrogen; Androgen; Environmental; Phytoestrogen

1. Introduction all patients undergo a period of remission. However, within a median time of 2–3 years a majority of patients Prostate cancer is the most commonly diagnosed develop ADT-resistant tumors, for which no effective malignancy in men and the second leading cause of or curative treatment has been identified [4,5]. There- cancer deaths among males in the United States. In fore, understanding the mechanisms by which ADT 2005, it is estimated that over 230,000 men will be resistance is acquired and identifying the factors that newly diagnosed with prostate cancer, and over 30,000 influence ADT efficacy are imperative to improve the will die of the disease [1]. Prostate cancer is relatively outcome of prostate cancer treatment. As described unique in that it is dependent on androgen for below, several endocrine disrupting compounds have development, growth, and survival [2]. While patients been proposed to influence prostate cancer develop- with organ-confined prostate cancer can be effectively ment, progression and management. treated through radical prostatectomy or radiation therapy, this form of cancer is relatively indolent, and most invasive tumors are resistant to general cytotoxic 2. Androgen action in prostate cancer therapy [3]. Therefore, the mainline therapeutic intervention, androgen deprivation therapy (ADT), Androgens exert their biological activity through the relies on the androgen dependence of prostatic androgen receptor (AR) [2,6], and ADT is designed to adenocarcinoma. ADT is highly effective as almost ablate AR function [7]. AR is a member of the nuclear receptor superfamily and functions as a ligand-depen- dent transcription factor. Prior to ligand activation, the * Corresponding author.Address: Department of Cell Biology, AR is present diffusely throughout the cell and is held University of Cincinnati College of Medicine, P.O. Box 670521, 3125 Eden Ave., Cincinnati, OH 45267-0521, USA. Tel.: C1 513 558 inactive through association with inhibitory heat shock 7371; fax: C1 513 558 4454. proteins. Testosterone is the most prevalent AR ligand in E-mail address: [email protected] (K.E. Knudsen). serum, but is converted to dihydrotestosterone (DHT) in

0304-3835/$ - see front matter q 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.canlet.2005.10.006 2 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 prostatic epithelial or adenocarcinoma cells through [24], and subsets of these mutations effectively convert the action of the enzyme 5-alpha reductase [8,9]. DHT is therapeutic AR antagonists into agonists (e.g. fluta- a high-affinity ligand for AR, and DHT binding mide) [20,22,25–27]. Combined, these observations stimulates displacement of heat shock proteins, demonstrate that agents that alter AR action can have receptor homodimerization, and rapid translocation to substantial effects on prostate cancer development and the nucleus. Activated AR binds to specific DNA treatment. sequences called androgen responsive elements (AREs), and recruits co-activators to initiate a program of gene transcription [2]. Through its transcrip- 3. Endocrine disrupting chemicals and the prostate tional transactivation function, AR activity results in a myriad of biological outcomes dependent on Several environmental and industrial compounds cellular context, including differentiation, secretion, mimic, alter, or block endogenous hormone action. and cellular proliferation. Given the relationship These compounds are generally classified as endocrine between AR activity and prostate cancer proliferation, disrupting compounds (EDCs), and there is an expression of the most well characterized AR target increasing interest in elucidating their biological gene, prostate-specific antigen (PSA), is utilized effects. Numerous EDC activities have been identified, clinically to diagnose prostate cancer and monitor including activation of nuclear hormone receptors. This tumor progression [10]. is particularly true of EDCs that harbor estrogenic It is well-established that AR function is required for activity, referred to as EEDCs (estrogenic endocrine prostate cancer development and survival [2,4]. For disrupting compounds) [28]. Similar to endogenous invasive disease, ADT blocks AR activity either , several EEDCs have been hypothesized to through preventing ligand (androgen) synthesis or alter androgen action or prostate growth and develop- through the use of direct AR antagonists [7,11].At ment [29]. These agents can be classified into naturally the cellular level, ADT triggers a cascade of events occurring compounds (dietary EDCs or phytoestro- leading to either programmed cell death or cell cycle gens) or synthetically generated agents (Fig. 1). arrest [7,11]. Clinically, efficacy of therapeutic inter- Human exposure to EEDCs can be significant, vention is demonstrated by significant reductions in dependent on diet and environment. The largest class PSA and tumor remission; however, recurrent ADT- of dietary EEDCs is represented by the resistant tumors typically arise [4,5]. Development of [30]. Soy-rich foods, fruits and selected nuts contain ADT-resistance is attributed to restored AR activity substantial levels of isoflavones (e.g. and (detected by a rise in PSA) and a concomitant ), and these compounds are increasingly resumption of tumor growth [2,4,12,13]. Investigations ingested as supra-dietary supplements [31,32]. Other into the mechanisms leading to AR reactivation and phytoestrogens include (found in red grapes ADT-resistance have revealed at least four putative and peanuts) and the and mechanisms (reviewed in [4,14,15]). First, AR can be , which are formed from fermentation of amplified or overexpressed in a subset of recurrent -derived precursors in the gut [33,34]. As will be tumors, presumably boosting basal activity to levels sufficient to induce a biological response [13,16,17]. Second, overexpression of AR co-activators has been observed, and it is hypothesized that co-activator deregulation can sensitize AR to low ligand concen- trations [18,19]. Third, ligand-independent AR acti- vation is thought to occur through induction of growth factor or cytokine pathways (e.g. IL-6, IGF, EGF, KGF), although the mechanisms underlying this activation event are poorly understood (reviewed in [20]). Lastly, ADT is known to select for mutations of Fig. 1. Illustration of the different sources of hormonally active the AR ligand binding domain that change the compounds that may affect prostate cancer (CaP) development, conformation of this domain and allow activation by progression or treatment. DES, ; EE, ethinylestra- diol; BPA, bisphenol A; PCB, polychlorinated bisphenols; HCB, other steroid hormones (e.g. estrogen, cortisol, pro- hexachlorobenzene; DDT, dichlorodiphenyltrichloroethane; DDE, gesterone) [21–23]. A substantial number of AR dichlorodiphenyltrichloroethylene; NP, ; OP, octophe- mutations have been identified in ADT-resistant tumors nol; Cd, cadmium. J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 3 discussed in a subsequent section, the phytoestrogens outcomes that may impact prostate cancer risk or may have protective capabilities for the prevention of progression. prostate cancer. Several EEDCs elicit biological endpoints that may The synthetic xenoestrogen class of EEDCs is be significant for prostate cancer development (sum- comprised of a myriad of unrelated compounds marized in Table 1). Developmental exposure (in utero shown to harbor estrogenic activity [35,36], and several or neonatal) to low dose DES, EE, BPA, PCB, or of these have a potential impact on the male hexachlorobenzene (HCB) has been reported in some reproductive tract. The pharmacological subclass of studies to increase prostate size or weight, thus EEDCs includes diethylstilbestrol (DES) and ethiny- suggesting that these agents influence early growth lestradiol (EE), which are used in prostate cancer patterns [47,53–58]. A particular emphasis in uncover- therapy and oral contraceptives, respectively [37,38]. ing the potential effects of BPA has emerged, as the Industrial EEDCs consist of bisphenols, organochlor- majority of studies report prostate-specific effects of ines, and alkylphenols [35,36,39], and exposure to this EEDC. It should be noted that this result was not re- these compounds can be frequent. Bisphenol A (BPA) capitulated in every model system [49,59,60], and is utilized in the manufacture of polycarbonate plastics recent articles analyzed the putative causes of this and epoxy resins and is leached from plastic containers disparity [48,61]. The effects of neonatal exposure to a into food [40], and certain dental sealants are also a natural estrogen on male rat reproductive organ major source of BPA exposure [41]. In the US, a recent development have also been assessed, and in these study demonstrates that 95% of adults have detectable studies the observed transient increase in prostate size levels of BPA in urine [42]. Pesticide-derived organo- was attributed to a significant acceleration of puberty chlorines such as the polychlorinated bisphenols [62,63]. These observed effects did not persist into (PCBs) and dichlorodiphenyltrichloroethane (DDT) or adulthood, and demonstrate that estogenic agents can metabolites of DDT are EDCs known to persist for long indirectly (albeit temporarily) influence prostate size periods in the environment and have been implicated in through alteration of developmental timing. increased cancer susceptibility [43]. Interestingly, the Cell-specific effects of EEDCs have been recently environmentally persistent DDT metabolite dichlor- observed, wherein prenatal exposure to low dose BPA or odiphenyldichloroethylene (DDE) has been shown to DES increased both number and size of prostatic harbor anti-androgenic activity [44]. Lastly, alkylphe- dorsolateral ducts, with an increased rate of proliferation nol surfactants are EEDCs used in textile and oil seen in the basal epithelia [55]. A separate study manufacturing [45]. Human exposure to alkylphenols demonstrated that low dose BPA and DES effects can be substantial through soil and water contami- growth using prostate explants [54].Thesedata nation, as these agents have been shown to bioaccu- indicated that the proliferative functions of BPA and mulate [45]. Each of these agents has been assessed for DES on the developing prostate can be direct and do not modulating prostate cancer risk or therapeutic outcome. require androgen stimulation. Interestingly, multiple studies show high doses of DES induced reduction in prostate size or weight, thus underscoring the repeated 4. In vivo action of EEDCs on prostate development observation that EEDC effects are highly dose depen- dent [47,53,55]. Similar effects were observed when Given the importance of hormone action for prostate growth of the urogenital sinus was utilized as an growth and development, there is an increasing focus endpoint [54,55]. Reductions in ventral prostate size on identifying potential effects of EEDCs in this tumor were shown after rats were exposed to high dose type. Studies have been highly diverse in experimental nonylphenol (NP) by intraperitoneal injection [64], design, with variables including model system, agent however, these results were not confirmed in an alternate utilized, timing of exposure, dosage, and endpoints rat strain nor were they reproducible by oral exposure of examined. As such, a uniform depiction of EEDC NP [65]. The discrepancies between these studies action in the prostate has yet to emerge, and some highlight the complexity and specificity of compounds, observations remain highly controversial. Recent species/strains differences, duration and method of reviews have addressed the importance of considering exposure, and endpoints examined. Additionally, these low dose effects and have attempted to clarify apparent studies clearly illustrate the biphasic dose response of discrepancies in outcome [46–52]. Here, a review of the EEDCs wherein receptor mediated responses first literature monitoring EEDCs influence on prostate increase and then decrease as dose is increased, development is summarized, with a focus on those exhibiting an ‘inverted U’ response [28]. Collectively, 4 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12

Table 1 EEDCs and factors proposed to influence prostate cancer development J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 5 a potentially significant impact of EEDCs on prostate BPA also induces histoarchitectural changes in the growth and development has been observed. Although periductal stroma of the ventral prostate, wherein a the mechanisms of EEDC action have yet to be reduced number of AR positive cells were observed rigorously examined, several EEDCs have been shown [58], and an increase in the ratio of fibroblasts to smooth to influence AR or (ER) expression or muscle cells was also reported [58,66]. These obser- action [28]. vations may be particularly pertinent to prostate cancer Given the importance of AR for both prostate development, as the stromal cells play important roles development and tumorigenic growth, a subset of in prostate cancer progression [70–72]. Inclusion of EEDC exposure studies monitored effects on AR these endpoints in subsequent studies of EEDC expression and AR function in vivo. Both BPA and exposure may help to clarify the mechanism of action low dose EE induce an increase in AR expression in and the window of sensitivity for biological effect. dorsolateral and ventral prostate lobes [56,59,66]. Together, these data suggest that EEDCs modify a However, at least one report demonstrated no change number of processes (e.g. proliferation, differentiation, after BPA exposure in the ventral prostate [58]. In vivo, androgen action, etc.) that may impact tumorigenesis in the BPA effect was also accompanied by an increase in the prostate. Although the above studies support the AR activity [66], but this parameter was not monitored contention that EEDCs may alter AR expression and in the EE study [56]. Increases in AR activity have also influence prostate development, the contribution of been reported after exposure to low doses of DES, these events for prostate cancer risk, growth or pro- HCB, or PCB [53,54,57,66]. The DES effect may be gression has yet to be established. manifested in part through both AR and ER, or through ER-mediated increases in EGF and IGF-1 [53]. Some 5. Effect of EEDCs on AR activity and CaP of the observed inductions of AR expression or action therapeutic response were transient, again indicating that the biological effect of EEDCs may be dependent on timing of Recent studies have examined the impact of select exposure during development [66]. Transient altera- EEDCs on AR regulation in prostate cancer cells and tions in testosterone and prolactin levels and permanent response to ADT. First, BPA has been shown to affect increases in ERb expression have been reported after therapeutic response in a subset of tumors that harbor BPA exposure, although it has yet to be established AR ligand binding domain mutations. While BPA fails whether these alterations are a cause or effect of the to activate the wild-type AR, a number of tumor- proliferative response. Interestingly, EEDCs associated derived AR mutants are activated by BPA and stimulate with decreased prostate growth (i.e. high dose DES or PSA expression in the absence of androgen [73,74]. HCB) also reduced AR expression or activity Tumor-derived mutants of AR that can be activated by [47,54,57,67]. In contrast, neonatal exposure to BPA include the prevalent T877A mutation, which is benzoate has been shown to induce pre- selected for during ADT therapy. Low dose BPA also cancerous lesions of the prostate (PIN, prostatic sensitizes prostate cancer cells harboring the AR- intraepithelial neoplasia) that was accompanied by T877A mutant to low levels of androgen, and low AR degradation in the ventral prostate [67]. In sum, the dose BPA was also shown to enhance androgen action mitogenic effect of EEDCs often correlates with in the presence of this mutant [74]. The biological changes in AR expression or function. Future studies impact of these observations was demonstrated in that are clearly needed to elucidate the relationship between BPA restored proliferation of prostate cancer cells observed changes in proliferation and AR status, as the expressing AR-T877A mutant under conditions of preponderance of studies have shown that EEDCs do androgen deprivation [73,74]. Together, these obser- not directly stimulate AR activity (reviewed in [68]). vations indicate that BPA exposure may inappropri- The observed effects of EEDCs on prostate growth ately activate AR in a subset of tumors treated by ADT, and AR function have also been associated with as well as cooperate with endogenous androgens to alterations in differentiation patterns or other develop- enhance AR activity in untreated patients. Confirmation mental defects [57,58,66]. While the minority of of these results in animal model systems has yet to be studies monitored those endpoints, BPA, low dose established and will be essential for defining the DES, phthalates, and PCBs also alter anogenital relative risk of BPA exposure in patients undergoing distance [53,54,69]. This parameter is considered to ADT. be a developmental marker of androgen action, again Several other EEDCs have also been implicated in suggesting that selected EEDCs can alter AR activity. modulating AR function or prostate cancer cell biology 6 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12

Although these hypotheses have yet to be examined for prostate cancer cells, there is evidence in breast cancer cells that EEDCs can initiate altered co-activator recruitment and differential gene expression [84,85]. Deregulation of growth factor pathways is another known mechanism utilized by prostate cancer cells in the transition to ADT resistance, and several EEDCs have been examined for their effect on growth factor pathways. Low concentrations of both genistein and resveratrol can enhance AR activity in prostate cancer Fig. 2. Therapeutic bypass and potential EEDC contributions. ADT cells via activation of the Raf–MEK–ERK pathway resistance is known to arise through at least four distinct pathways [86] and several agents (BPA, DES and octophenol) that enhance AR activity [4]. These pathways include AR influence IGF-I and EGF levels in vivo [54,87]. Thus, amplification and hypersensitivity, mutations in the AR, alterations in growth factor associated pathways, and deregulation of AR co- these collective findings indicate that environmentally activators. Where shown, selected EEDCs have been hypothesized to relevant concentrations of EEDCs may impact path- impinge on these pathways. ways contributing to therapeutic resistance. (as summarized in Fig. 2). Pesticide-derived DDE 6. EEDCs as potential mutagens activates wild-type AR in transient assays at physio- logical concentrations [75]. Cadmium also activates Distinct from their role as endocrine disrupting both wild-type and tumor-derived AR, thus implicating compounds, several EEDCs harbor mutagenic activity, this heavy metal as a putative effecter of the androgen including the bisphenols, alkylphenols, organochlor- pathway in prostate cancer [76,77]. Given the preva- ines and phthalates. Most in vitro assays of mutageni- lence of both DDE and cadmium in the environment city and genotoxicity have not supported a role for BPA [78,79], these observations highlight the need to as a direct mutagen [88,89]; however, a recent report delineate the contribution of each compound to prostate linked BPA exposure to the development of meiotic cancer growth. Strikingly, low levels of specific aneuploidy [90]. In this study, oocytes of female mice phytoestrogens (genistein and resveratrol) can activate exposed to short-term, low-dose BPA harbored signifi- both wild-type and mutant ARs and stimulate prolifer- cant meiotic defects, including a sharp rise in meiotic ation in prostate cancer cells [80]. However, in keeping aneuploidy and congression failure during metaphase I. with the dose-dependent effects of these agents, high These unexpected effects occurred through incidental dose BPA, genistein, and resveratrol inhibit AR exposure to BPA, and indicate that BPA exposure function [81,82]. These data underscore the need to below the accepted exposure dose may influence the monitor both receptor activity and proliferative genetic quality of offspring and induce transgenera- response across multiple doses of EEDC exposure. tional effects. Although the mechanisms of BPA action Moreover, high throughput screens currently examin- have yet to be elucidated, it was hypothesized that the ing EEDC effects on AR ligand binding and function effects of BPA were elicited through alteration of should be complemented with biological endpoints. estrogen action. Additional studies have also indicated In addition to AR mutation, there are multiple that BPA may negatively affect microtubule function pathways by which prostate cancer cells circumvent and chromosome stability, dependent on experimental ADT and progress to therapy resistant tumors (sum- or cellular context [91–95]. marized in Fig. 2). While not yet challenged in prostate DNA adduct formation has also been indicated as a cancer models, the demonstrated ability of EEDCs to potential BPA effect, although these properties of BPA influence AR action may impact distinct pathways of typically required relatively high exposure levels. In ADT resistance. For example, when AR is over- hamster embryonic cells, BPA initiated DNA adduct expressed in tumor recurrence, minimal AR activation formation in a dose dependent manner [96,97], and by select EEDCs may be sufficient to induce biological these changes were associated with the development of changes. Moreover, deregulation or overexpression of aneuploidy and cellular transformation. In a separate AR co-activators may also potentiate utilization of study, BPA exposure triggered K-Ras mutation and select EEDCs for inappropriate AR activity, as AR co- measurable changes in DNA synthesis, thus indicating activators are known to alter response or sensitize that BPA can elicit genetic alterations associated with hormone receptors to low ligand environments [14,83]. cancer development [98]. Subsequent studies revealed J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 7 that the mutagenic properties of BPA are shared by high in the plasma and urine of men residing in eastern several derivatives of BPA [97]. BPA can be oxidized countries [107,108]. These observations led to investi- into quinone derivatives that at millimolar concen- gation of the effects of isoflavones (most notably trations form DNA adducts in vitro and in the cells of genistein and daidzein) and lignans on prostate cancer treated rats [99], thus revealing that modifications of incidence. EEDCs may alter their bioactivity. Additionally, BPA Analyses of phytoestrogen effects on prostate cancer can be nitrosylated, which can also initiate DNA adduct risk are complicated by the discrete biological activities formation and mutations. Cadmium, another well- of metabolites and by variations in the metabolic established mutagen with EEDC abilities [100,101], profiles of different individuals. From numerous can induce pre-neoplastic lesions and cancer of the rat investigations, it is clear that phytoestrogens do not prostate, initiate transformation of human prostatic act via a single mechanism of action, nor do different epithelia in vitro, and has been linked through phytoestrogens produce the same profile of biological epidemiological studies to increases in prostate cancer responses. However, several studies have revealed the risk [93,102]. The impetus to investigate the DNA potential beneficial effects of these compounds. Studies altering effects of EDCs has been bolstered by recent in the 1980s initially described an association between findings that selected agents can induce alterations of diets rich in soy, , or and decreased prostate methylation patterns of germline DNA, which persisted cancer risk [109–112]. One such study of 14,000 men through multiple generations [103]. These epigenetic revealed that a select subset incurred a significantly changes resulted in decreased sperm number and decreased risk for prostate cancer when consumption of viability and decreased male fertility, implicating the soy products was increased [110]. Protective effects potential role of these biological actions in male disease have also been correlated with dietary intake of [103]. Other EDCs with anti-androgenic properties (e.g. daidzein, coumestrol and genistein [113–117]. phthalates [104,105]), have been linked to develop- Although the mechanisms of action have yet to be mental aberrations in the male reproductive tract [69]. elucidated, some studies demonstrated that phytoestro- Combined, these studies suggest that the mutageni- gens could alter endogenous hormone levels. An city of EDCs may be enhanced or altered by inverse association between soy food consumption environmental context, and demonstrate that epigenetic and serum and testosterone levels in Japanese effects can persist through multiple generations. While men has been reported [118], but not all studies have the contribution of these activities to prostate cancer validated this conclusion [119–121]. The protective risk has not been clearly defined, these data collectively effect of phytoestrogens has also been challenged in indicate that studies of EDC action should take into animal models, and the majority of studies have consideration cellular environment, the potential for demonstrated that a diet rich in phytoestrogens results derivative formation, and the possibility of epigenetic in reduced incidence of prostate tumors in rats or mice alterations. [122–126], and prostate tumor development was inhibited in rats fed an isoflavone-rich diet [127]. 7. Protective effects of dietary EEDCs Growth and metastatic potential of tumors was also dramatically reduced in a transgenic mouse model of Although there is significant concern regarding the prostate cancer (TRAMP) as a result of genistein impact of EEDCs on prostate cancer development and exposure [128]. Similarly, genistein and daidzein progression, evidence indicates that some naturally induce reduction in ventral prostate carcinoma inci- occurring EEDCs (phytoestrogens) may have preven- dence [118]. Expression of AR, ERa and ERb is tative effects on prostate cancer development. Initially, reduced in the developing rat dorsolateral prostate after epidemiological data provided the rationale for exam- oral exposure to genistein [129], revealing one possible ining the link between diets rich in soy and rye based mechanism behind the lower incidence of prostate products and the prevention of prostate cancer. cancer in populations consuming high levels of Although the incidence of latent or non-infiltrative phytoestrogens. Resveratrol has also been examined prostate cancer is similar worldwide, mortality associ- in animal models of prostate cancer and has been ated with the disease is substantially higher in western determined to act as a chemo-preventative agent [130]. countries [106]. Therefore, environmental factors such In addition to lowering prostate cancer risk, as diet have been implicated in altering prostate cancer phytoestrogen intake may improve outcome in patients risk. Phytoestrogens are consumed in large amounts in being treated for prostate cancer. Intake of an Asian populations, and isoflavones and lignan levels are isoflavone-supplemented diet for a minimum of 3 8 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 months was shown to decrease the rise of PSA levels in Direct effects on prostate cancer incidence and both ADT-sensitive and ADT-refractory prostate treatment are also supported by substantial in vitro, cancers [131]. A separate study examined the effect in vivo, and epidemiological studies. However, of phytoestrogen consumption prior to radical prosta- reductions in prostate cancer incidence and cooperation tectomy, and observed that the resected specimen with standard therapy have been associated with the displayed significant apoptosis, suggesting that some dietary EEDCs. In summary, EEDC exposure may have tumor regression had occurred [132]. Not all studies a significant impact on prostate cancer development have observed alterations in biomarkers for pro- and progression, and further studies will be required to gression, thus indicating that agent, dosage, and identify the cellular and environmental factors that tumor status may affect patient outcome. In addition contribute to the denouement of EEDC exposure. to biomarker analyses, phytoestrogens influence specific aspects of the steroid hormone axis, including Acknowledgements the production, metabolism and biological activity of sex hormones, and the regulation of numerous The authors would like to thank Drs Shuk-Mei Ho, intracellular steroid-metabolizing enzymes (reviewed Sohaib Khan, Erik Knudsen, Gail Prins, Alvaro Puga, in [108,133,134]). In a pilot study of prostate cancer and Frederick vom Saal for critical discussions and patients who consumed a low-fat, phytoestrogen- insightful commentary. We are also grateful to Drs Lisa supplemented diet, decreased total testosterone and Morey, Christopher Mayhew, and Clay Comstock, in androgen levels were observed [135]. Furthermore, addition to Kevin Link, William Zagorski, Nick soy-based diets significantly decreased circulating Olshavsky, and Jon Lenihan, for assistance with editing testosterone levels and prostate weights in rats [136]. and for ongoing discussions. Due to space constraints, Given the importance of androgen deprivation for we regret omissions of additional studies. tumor regression, these data indicate that the effect of phytoestrogens on ADT should be monitored. Numer- ous in vitro studies have also revealed that high References concentrations of phytoestrogens (both isoflavones and lignans) inhibit the growth of prostate cancer [1] A. Jemal, T. Murray, E. Ward, A. Samuels, R.C. Tiwari, A. Ghafoor, et al., Cancer statistics, 2005, CA Cancer J. Clin. cells in a dose dependant manner regardless of AR 55 (1) (2005) 10–30. expression or hormone dependency (reviewed in [2] J. Trapman, O.A. Brinkmann, The androgen receptor in [21,137–146]). Although these data indicate that the prostate cancer, Pathol. Res. Pract. 192 (7) (1996) 752–760. dietary EEDCs can have general protective effects for [3] S. Goodin, V.K. Rao, R.S. DiPaola, State-of-the-art treatment prostate cancer, these studies have not addressed the of metastatic hormone-refractory prostate cancer, Oncologist 7 ability of selected EEDCs to activate mutant ARs from (4) (2002) 360–370. [4] B.J. Feldman, D. Feldman, The development of androgen- recurrent tumors. The efficacy of dietary EEDC action independent prostate cancer, Nat. Rev. Cancer 1 (1) (2001) 34– in patients with AR mutations has yet to be assessed, 45. and may provide insight into identifying patients for [5] S. Leewansangtong, S. Soontrapa, Hormonal ablation therapy whom dietary EEDCs may be indicated. In summary, for metastatic prostatic carcinoma: a review, J. Med. Assoc. current evidence indicates that phytoestrogens may Thai. 82 (2) (1999) 192–205. [6] C. Lee, D.M. Sutkowski, J.A. Sensibar, D. Zelner, I. Kim, influence multiple pathways impacting both prostate I. Amsel, et al., Regulation of proliferation and production of cancer development and progression. prostate-specific antigen in androgen-sensitive prostatic cancer cells, LNCaP, by dihydrotestosterone, Endocrinology 136 (2) 8. Conclusions (1995) 796–803. [7] P. Reid, P. Kantoff, W. Oh, Antiandrogens in prostate cancer, Invest. New Drugs 17 (3) (1999) 271–284. A large body of evidence demonstrates that human [8] D.W. Russell, J.D. Wilson, Steroid 5 alpha-reductase: two exposure to highly prevalent endocrine disrupting genes/two enzymes, Annu. Rev. Biochem. 63 (1994) 25–61. compounds may alter hormone function. Given the [9] D.W. Russell, D.M. Berman, J.T. Bryant, K.M. Cala, importance of hormone action in prostate cancer, it is D.L. Davis, C.P. Landrum, et al., The molecular genetics of imperative to identify factors that influence cancer risk, steroid 5 alpha-reductases, Recent Prog. Horm. Res. 49 (1994) progression, and treatment. As discussed in the present 275–284. [10] C. Stephan, K. Jung, E.P. Diamandis, H.G. Rittenhouse, review, estrogenic EDCs cause aberrations in develop- M. Lein, S.A. Loening, Prostate-specific antigen, its molecular ment of the male reproductive tract, which could be forms, and other kallikrein markers for detection of prostate hypothesized to impact prostate cancer initiation. cancer, Urology 59 (1) (2002) 2–8. J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 9

[11] L. Klotz, Hormone therapy for patients with prostate in prostate cancer, Proc. Natl Acad. Sci. USA 102 (17) (2005) carcinoma, Cancer 88 (12) (2000) 3009–3014. 6201–6206. [12] N. Craft, C.L. Sawyers, Mechanistic concepts in androgen- [28] W.V. Welshons, K.A. Thayer, B.M. Judy, J.A. Taylor, dependence of prostate cancer, Cancer Metastasis Rev. 17 (4) E.M. Curran, F.S. vom Saal, Large effects from small (1998) 421–427. exposures. I. Mechanisms for endocrine-disrupting chemicals [13] T. Visakorpi, E. Hyytinen, P. Koivisto, M. Tanner, with estrogenic activity, Environ. Health Perspect. 111 (8) R. Keinanen, C. Palmberg, et al., In vivo amplification of the (2003) 994–1006. androgen receptor gene and progression of human prostate [29] R.A. Jarred, B. Cancilla, G.S. Prins, K.A. Thayer, G.R. Cunha, cancer, Nat. Genet. 9 (4) (1995) 401–406. G.P. Risbridger, Evidence that estrogens directly alter [14] Z. Culig, H. Steiner, G. Bartsch, A. Hobisch, Mechanisms of androgen-regulated prostate development, Endocrinology 141 endocrine therapy-responsive and -unresponsive prostate (9) (2000) 3471–3477. tumours, Endocr. Relat. Cancer 12 (2) (2005) 229–244. [30] J.P. Basly, M.C. Lavier, Dietary phytoestrogens: potential [15] D. Navarro, O.P. Luzardo, L. Fernandez, N. Chesa, B.N. Diaz- selective estrogen enzyme modulators?, Planta Med. 71 (4) Chico, Transition to androgen-independence in prostate (2005) 287–294. cancer, J. Steroid Biochem. Mol. Biol. 81 (3) (2002) 191–201. [31] C.E. Piersen, Phytoestrogens in botanical dietary supplements: [16] P. Koivisto, J. Kononen, C. Palmberg, T. Tammela, implications for cancer, Integr. Cancer Ther. 2 (2) (2003) 120– E. Hyytinen, J. Isola, et al., Androgen receptor gene 138. amplification: a possible molecular mechanism for androgen [32] L.A. Fitzpatrick, Alternatives to estrogen, Med. Clin. North deprivation therapy failure in prostate cancer, Cancer Res. 57 Am. 87 (5) (2003) 1091–1113. (2) (1997) 314–319. [33] I. Rowland, M. Faughnan, L. Hoey, K. Wahala, G. Williamson, [17] C.D. Chen, D.S. Welsbie, C. Tran, S.H. Baek, R. Chen, A. Cassidy, Bioavailability of phyto-oestrogens, Br. J. Nutr. 89 R. Vessella, et al., Molecular determinants of resistance to (Suppl 1) (2003) S45–S58. antiandrogen therapy, Nat. Med. 10 (1) (2004) 33–39. [34] L.Q. Wang, Mammalian phytoestrogens: enterodiol and [18] C.W. Gregory, B. He, R.T. Johnson, O.H. Ford, J.L. Mohler, enterolactone, J. Chromatogr. B Analyt. Technol. Biomed. F.S. French, E.M. Wilson, A mechanism for androgen Life Sci. 777 (1–2) (2002) 289–309. receptor-mediated prostate cancer recurrence after androgen [35] N. Olea, P. Pazos, J. Exposito, Inadvertent exposure to deprivation therapy, Cancer Res. 61 (11) (2001) 4315–4319. , Eur. J. Cancer Prev. 7 (Suppl 1) (1998) S17– [19] N. Fujimoto, A. Mizokami, S. Harada, T. Matsumoto, Different S23. expression of androgen receptor coactivators in human [36] D.W. Singleton, S.A. Khan, Xenoestrogen exposure and prostate, Urology 58 (2) (2001) 289–294. mechanisms of endocrine disruption, Front. Biosci. 8 (2003) [20] B. Chatterjee, The role of the androgen receptor in the s110–s118. development of prostatic hyperplasia and prostate cancer, Mol. [37] W.K. Oh, The evolving role of estrogen therapy in prostate Cell Biochem. 253 (1–2) (2003) 89–101. cancer, Clin. Prostate Cancer 1 (2) (2002) 81–89. [21] A. Shih, S. Zhang, H.J. Cao, S. Boswell, Y.H. Wu, H.Y. Tang, [38] S.J. Keam, A.J. Wagstaff, /drospirenone: a et al., Inhibitory effect of epidermal growth factor on review of its use as an oral contraceptive, Treat. Endocrinol. 2 resveratrol-induced apoptosis in prostate cancer cells is (1) (2003) 49–70. mediated by protein kinase C-alpha, Mol. Cancer Ther. 3 [39] S. Mitchell, P. Abel, M. Ware, G. Stamp, E. Lalani, Phenotypic (11) (2004) 1355–1364. and genotypic characterization of commonly used human [22] H. Suzuki, N. Sato, Y. Watabe, M. Masai, S. Seino, prostatic cell lines, Br. J. Urol. Int. 85 (7) (2000) 932–944. J. Shimazaki, Androgen receptor gene mutations in human [40] L.A. Haighton, J.J. Hlywka, J. Doull, R. Kroes, B.S. Lynch, prostate cancer, J. Steroid Biochem. Mol. Biol. 46 (6) (1993) I.C. Munro, An evaluation of the possible carcinogenicity of 759–765. bisphenol A to humans, Regul. Toxicol. Pharmacol. 1 (2002) [23] Z. Culig, A. Hobisch, M.V. Cronauer, A.C. Cato, A. Hittmair, 238–254. C. Radmayr, et al., Mutant androgen receptor detected in an [41] N.D. Ruse, Xenoestrogenicity and dental materials, J. Can. advanced-stage prostatic carcinoma is activated by adrenal Dent. Assoc. 63 (11) (1997) 833–836. androgens and progesterone, Mol. Endocrinol. 7 (12) (1993) [42] A.M. Calafat, Z. Kuklenyik, J.A. Reidy, S.P. Caudill, J. Ekong, 1541–1550. L.L. Needham, Urinary concentrations of bisphenol A and [24] B. Gottlieb, L.K. Beitel, J.H. Wu, M. Trifiro, The androgen 4-nonylphenol in a human reference population, Environ. receptor gene mutations database (ARDB): 2004 update, Hum. Health Perspect. 113 (4) (2005) 391–395. Mutat. 23 (6) (2004) 527–533. [43] M.P. Longnecker, W.J. Rogan, G. Lucier, The human health [25] M.A. Fenton, T.D. Shuster, A.M. Fertig, M.E. Taplin, effects of DDT (dichlorodiphenyltrichloroethane) and PCBS G. Kolvenbag, G.J. Bubley, S.P. Balk, Functional character- (polychlorinated biphenyls) and an overview of organochlor- ization of mutant androgen receptors from androgen-indepen- ines in public health, Annu. Rev. Public Health 18 (1997) 211– dent prostate cancer, Clin. Cancer Res. 3 (8) (1997) 1383– 244. 1388. [44] W.R. Kelce, C.R. Stone, S.C. Laws, L.E. Gray, [26] T. Hara, J. Miyazaki, H. Araki, M. Yamaoka, N. Kanzaki, J.A. Kemppainen, E.M. Wilson, Persistent DDT metabolite M. Kusaka, M. Miyamoto, Novel mutations of androgen p,p’-DDE is a potent androgen receptor antagonist, Nature 375 receptor: a possible mechanism of bicalutamide withdrawal (6532) (1995) 581–585. syndrome, Cancer Res. 63 (1) (2003) 149–153. [45] G.G. Ying, B. Williams, R. Kookana, Environmental fate of [27] C.E. Bohl, W. Gao, D.D. Miller, C.E. Bell, J.T. Dalton, alkylphenols and alkylphenol ethoxylates–a review, Environ. Structural basis for antagonism and resistance of bicalutamide Int. 28 (3) (2002) 215–226. 10 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12

[46] S.O. Mueller, Xenoestrogens: mechanisms of action and literature showing adverse effects at low doses: a response to detection methods, Anal. Bioanal. Chem. 378 (3) (2004) attempts to mislead the public, Toxicology 212 (2/3) (2005) 582–587. 244–252. [47] F.S. vom Saal, B.G. Timms, M.M. Montano, P. Palanza, [62] O. Putz, C.B. Schwartz, S. Kim, G.A. LeBlanc, R.L. Cooper, K.A. Thayer, S.C. Nagel, et al., Prostate enlargement in mice G.S. Prins, Neonatal low- and high-dose exposure to estradiol due to fetal exposure to low doses of estradiol or diethyl- benzoate in the male rat: I. Effects on the prostate gland, Biol. and opposite effects at high doses, Proc. Natl Acad. Reprod. 65 (5) (2001) 1496–1505. Sci. USA 94 (5) (1997) 2056–2061. [63] O. Putz, C.B. Schwartz, G.A. LeBlanc, R.L. Cooper, [48] F.S. vom Saal, C. Hughes, An extensive new literature G.S. Prins, Neonatal low- and high-dose exposure to estradiol concerning low-dose effects of bisphenol A shows the need benzoate in the male rat: II. Effects on male puberty and the for a new risk assessment, Environ. Health Perspect. 113 (8) reproductive tract, Biol. Reprod. 65 (5) (2001) 1506–1517. (2005) 926–933. [64] P.C. Lee, Disruption of male reproductive tract development [49] I.F. Purchase, Fraud, errors and gamesmanship in experimental by administration of the xenoestrogen, nonylphenol, to male toxicology, Toxicology 202 (1–2) (2004) 1–20. newborn rats, Endocrine 9 (1) (1998) 105–111. [50] H.A. Milman, M.C. Bosland, P.D. Walden, J.E. Heinze, [65] J. Odum, J. Ashby, Neonatal exposure of male rats to Evaluation of the adequacy of published studies of low-dose nonylphenol has no effect on the reproductive tract, Toxicol. effects of bisphenol A on the rodent prostate for use in human Sci. 56 (2) (2000) 400–404. risk assessment, Regul. Toxicol. Pharmacol. 35 (3) (2002) [66] J.G. Ramos, J. Varayoud, L. Kass, H. Rodriguez, L. Costabel, 338–346. M. Munoz-De-Toro, E.H. Luque, Bisphenol a induces both [51] S. Pflieger-Bruss, H.C. Schuppe, W.B. Schill, The male transient and permanent histofunctional alterations of the reproductive system and its susceptibility to endocrine hypothalamic–pituitary–gonadal axis in prenatally exposed disrupting chemicals, Andrologia 36 (6) (2004) 337–345. male rats, Endocrinology 144 (7) (2003) 3206–3215. [52] J. Ashby, H. Tinwell, J. Odum, P. Lefevre, Natural variability [67] C. Woodham, L. Birch, G.S. Prins, Neonatal estrogen down- and the influence of concurrent control values on the detection regulates prostatic androgen receptor through a proteosome- and interpretation of low-dose or weak endocrine toxicities, mediated protein degradation pathway, Endocrinology 144 Environ. Health Perspect. 112 (8) (2004) 847–853. (11) (2003) 4841–4850. [53] C. Gupta, Reproductive malformation of the male offspring [68] J.A. McLachlan, Environmental signaling: what embryos and following maternal exposure to estrogenic chemicals, Proc. evolution teach us about endocrine disrupting chemicals, Soc. Exp. Biol. Med. 224 (2) (2000) 61–68. Endocr. Rev. 22 (3) (2001) 319–341. [54] C. Gupta, The role of estrogen receptor, androgen receptor and [69] S.H. Swan, K.M. Main, F. Liu, S.L. Stewart, R.L. Kruse, growth factors in diethylstilbestrol-induced programming of A.M. Calafat, et al., Decrease in anogenital distance among prostate differentiation, Urol. Res. 28 (4) (2000) 223–229. male infants with prenatal phthalate exposure, Environ. Health [55] B.G. Timms, K.L. Howdeshell, L. Barton, S. Bradley, Perspect. 113 (8) (2005) 1056–1061. C.A. Richter, F.S. vom Saal, Estrogenic chemicals in plastic [70] E. Castellon, K. Venegas, L. Saenz, H. Contreras, C. Huidobro, and oral contraceptives disrupt development of the fetal mouse Secretion of prostatic specific antigen, proliferative activity prostate and urethra, Proc. Natl Acad. Sci. USA 102 (19) and androgen response in epithelial-stromal co-cultures from (2005) 7014–7019. human prostate carcinoma, Int. J. Androl. 28 (1) (2005) 39–46. [56] K.A. Thayer, R.L. Ruhlen, K.L. Howdeshell, D.L. Buchanan, P.S. Cooke, D. Preziosi, et al., Altered prostate growth and [71] M.S. Condon, The role of the stromal microenvironment in daily sperm production in male mice exposed prenatally to prostate cancer, Semin. Cancer Biol. 15 (2) (2005) 132–137. subclinical doses of 17alpha-ethinyl oestradiol, Hum. Reprod. [72] M.S. Condon, M.C. Bosland, The role of stromal cells in 16 (5) (2001) 988–996. prostate cancer development and progression, In Vivo 13 (1) [57] J.L. Ralph, M.C. Orgebin-Crist, J.J. Lareyre, C.C. Nelson, (1999) 61–65. Disruption of androgen regulation in the prostate by the [73] Y.B. Wetherill, C.E. Petre, K.R. Monk, A. Puga, environmental contaminant hexachlorobenzene, Environ. K.E. Knudsen, The xenoestrogen bisphenol A induces Health Perspect. 111 (4) (2003) 461–466. inappropriate androgen receptor activation and mitogenesis [58] J.G. Ramos, J. Varayoud, C. Sonnenschein, A.M. Soto, in prostatic adenocarcinoma cells, Mol. Cancer Ther. 1 (7) M. Munoz De Toro, E.H. Luque, Prenatal exposure to low (2002) 515–524. doses of bisphenol A alters the periductal stroma and glandular [74] Y.B. Wetherill, N.L. Fisher, A. Staubach, M. Danielsen, cell function in the rat ventral prostate, Biol. Reprod. 65 (4) R.W. de Vere White, K.E. Knudsen, Xenoestrogen action in (2001) 1271–1277. prostate cancer: pleiotropic effects dependent on androgen [59] H. Tinwell, J. Haseman, P.A. Lefevre, N. Wallis, J. Ashby, receptor status, Cancer Res. 65 (1) (2005) 54–65. Normal sexual development of two strains of rat exposed in [75] T.J. Schrader, G.M. Cooke, Examination of selected food utero to low doses of bisphenol A, Toxicol. Sci. 68 (2) (2002) additives and organochlorine food contaminants for andro- 339–348. genic activity in vitro, Toxicol. Sci. 53 (2) (2000) 278–288. [60] H. Yoshino, T. Ichihara, M. Kawabe, N. Imai, A. Hagiwara, [76] V. Verougstraete, D. Lison, P. Hotz, Cadmium, lung and M. Asamoto, T. Shirai, Lack of significant alteration in the prostate cancer: a systematic review of recent epidemiological prostate or testis of F344 rat offspring after transplacental and data, J. Toxicol. Environ. Health B Crit. Rev. 6 (3) (2003) 227– lactational exposure to bisphenol A, J. Toxicol. Sci. 27 (5) 255. (2002) 433–439. [77] A.E. Sahmoun, L.D. Case, S.A. Jackson, G.G. Schwartz, [61] F.S. Vom Saal, S.C. Nagel, B.G. Timms, W.V. Welshons, Cadmium and prostate cancer: a critical epidemiologic Implications for human health of the extensive bisphenol a analysis, Cancer Invest. 23 (3) (2005) 256–263. J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12 11

[78] M.C. Henson, P.J. Chedrese, Endocrine disruption by the microtubule network in chinese hamster V79 cells in cadmium, a common environmental toxicant with paradoxical culture and in sertoli cells in rats, Teratog. Carcinog. Mutagen. effects on reproduction, Exp. Biol. Med. (Maywood) 229 (5) 21 (6) (2001) 453–462. (2004) 383–392. [94] T. Ochi, Induction of multiple microtubule-organizing centers, [79] K. Jaga, C. Dharmani, Global surveillance of DDT and DDE multipolar spindles and multipolar division in cultured V79 levels in human tissues, Int. J. Occup. Med. Environ. Health 16 cells exposed to diethylstilbestrol, estradiol-17beta and (1) (2003) 7–20. bisphenol A, Mutat. Res. 431 (1) (1999) 105–121. [80] M. Maggiolini, A. Vivacqua, A. Carpino, D. Bonofiglio, [95] M. Metzler, E. Pfeiffer, Effects of estrogens on microtubule G. Fasanella, M. Salerno, et al., The mutant androgen receptor polymerization in vitro: correlation with estrogenicity, T877A mediates the proliferative but not the cytotoxic dose- Environ. Health Perspect. 103 (Suppl. 7) (1995) 21–22. dependent effects of genistein and on human LNCaP [96] T. Tsutsui, Y. Tamura, E. Yagi, K. Hasegawa, M. Takahashi, prostate cancer cells, Mol. Pharmacol. 62 (5) (2002) 1027– N. Maizumi, et al., Bisphenol-A induces cellular transform- 1035. ation, aneuploidy and DNA adduct formation in cultured [81] S.H. Mitchell, W. Zhu, C.Y. Young, Resveratrol inhibits the syrian hamster embryo cells, Int. J. Cancer 75 (2) (1998) 290– expression and function of the androgen receptor in LNCaP 294. prostate cancer cells, Cancer Res. 59 (23) (1999) 5892–5895. [97] T. Tsutsui, Y. Tamura, E. Yagi, J.C. Barrett, Involvement of [82] J.N. Davis, O. Kucuk, F.H. Sarkar, Expression of prostate- genotoxic effects in the initiation of estrogen-induced cellular specific antigen is transcriptionally regulated by genistein in transformation: studies using syrian hamster embryo cells prostate cancer cells, Mol. Carcinog. 34 (2) (2002) 91–101. treated with 17beta-estradiol and eight of its metabolites, Int. [83] Z. Culig, B. Comuzzi, H. Steiner, G. Bartsch, A. Hobisch, J. Cancer 86 (1) (2000) 8–14. Expression and function of androgen receptor coactivators in [98] S. Takahashi, X.J. Chi, Y. Yamaguchi, H. Suzuki, S. Sugaya, prostate cancer, J. Steroid Biochem. Mol. Biol. 92 (4) (2004) K. Kita, et al., Mutagenicity of bisphenol A and its suppression 265–271. by interferon-alpha in human RSa cells, Mutat. Res. 490 (2) [84] E.J. Routledge, R. White, M.G. Parker, J.P. Sumpter, (2001) 199–207. Differential effects of xenoestrogens on coactivator recruitment [99] A. Atkinson, D. Roy, In vivo DNA adduct formation by by estrogen receptor (ER) alpha and ERbeta, J. Biol. Chem. bisphenol A, Environ. Mol. Mutagen. 26 (1) (1995) 60–66. 275 (46) (2000) 35986–35993. [100] M.P. Waalkes, M. Anver, B.A. Diwan, Carcinogenic effects of [85] H. Inoshita, H. Masuyama, Y. Hiramatsu, The different effects cadmium in the noble (NBL/Cr) rat: induction of pituitary, of endocrine-disrupting chemicals on estrogen receptor- testicular, and injection site tumors and intraepithelial mediated transcription through interaction with coactivator proliferative lesions of the dorsolateral prostate, Toxicol. Sci. TRAP220 in uterine tissue, J. Mol. Endocrinol. 31 (3) (2003) 52 (2) (1999) 154–161. 551–561. [101] M.P. Waalkes, M.R. Anver, B.A. Diwan, Chronic toxic and [86] S. Gao, G.Z. Liu, Z. Wang, Modulation of androgen receptor- carcinogenic effects of oral cadmium in the Noble (NBL/Cr) dependent transcription by resveratrol and genistein in prostate rat: induction of neoplastic and proliferative lesions of the cancer cells, Prostate 59 (2) (2004) 214–225. adrenal, kidney, prostate, and testes, J. Toxicol. Environ. [87] C.B. Herath, W. Jin, G. Watanabe, K. Arai, A.K. Suzuki, Health A 58 (4) (1999) 199–214. K. Taya, Adverse effects of environmental toxicants, octyl- [102] W.E. Achanzar, B.A. Diwan, J. Liu, S.T. Quader, phenol and bisphenol A, on male reproductive functions in M.M. Webber, M.P. Waalkes, Cadmium-induced malignant pubertal rats, Endocrine 25 (2) (2004) 163–172. transformation of human prostate epithelial cells, Cancer Res. [88] J.L. Ivett, B.M. Brown, C. Rodgers, B.E. Anderson, 61 (2) (2001) 455–458. M.A. Resnick, E. Zeiger, Chromosomal aberrations and sister [103] M.D. Anway, A.S. Cupp, M. Uzumcu, M.K. Skinner, chromatid exchange tests in Chinese hamster ovary cells Epigenetic transgenerational actions of endocrine disruptors in vitro. IV. Results with 15 chemicals, Environ. Mol. and male fertility, Science 308 (5727) (2005) 1466–1469. Mutagen. 14 (3) (1989) 165–187. [104] N.P. Moore, The oestrogenic potential of the phthalate esters, [89] H. Schweikl, G. Schmalz, K. Rackebrandt, The mutagenic Reprod. Toxicol. 14 (3) (2000) 183–192. activity of unpolymerized resin monomers in salmonella [105] J.S. Fisher, Are all EDC effects mediated via steroid hormone typhimurium and V79 cells, Mutat. Res. 415 (1/2) (1998) receptors? Toxicology 205 (1–2) (2004) 33–41. 119–130. [106] R. Yatani, I. Chigusa, K. Akazaki, G.N. Stemmermann, [90] P.A. Hunt, K.E. Koehler, M. Susiarjo, C.A. Hodges, A. Ilagan, R.A. Welsh, P. Correa, Geographic pathology of latent R.C. Voigt, et al., Bisphenol a exposure causes meiotic prostatic carcinoma, Int. J. Cancer 29 (6) (1982) 611–616. aneuploidy in the female mouse, Curr. Biol. 13 (7) (2003) 546– [107] H. Adlercreutz, H. Honjo, A. Higashi, T. Fotsis, 553. E. Hamalainen, T. Hasegawa, H. Okada, Urinary excretion of [91] E. Pfeiffer, B. Rosenberg, S. Deuschel, M. Metzler, Inter- lignans and isoflavonoid phytoestrogens in Japanese men and ference with microtubules and induction of micronuclei in vitro women consuming a traditional Japanese diet, Am. J. Clin. by various bisphenols, Mutat. Res. 390 (1–2) (1997) 21–31. Nutr. 54 (6) (1991) 1093–1100. [92] L. Lehmann, M. Metzler, Bisphenol A and its methylated [108] H. Adlercreutz, Epidemiology of phytoestrogens, Baillieres congeners inhibit growth and interfere with microtubules in Clin. Endocrinol. Metab. 12 (4) (1998) 605–623. human fibroblasts in vitro, Chem. Biol. Interact. 147 (3) (2004) [109] R.K. Severson, A.M. Nomura, J.S. Grove, G.N. Stemmermann, 273–285. A prospective study of demographics, diet, and prostate cancer [93] M. Nakagomi, E. Suzuki, K. Usumi, Y. Saitoh, S. Yoshimura, among men of Japanese Ancestry in Hawaii, Cancer Res. 49 T. Nagao, H. Ono, Effects of endocrine disrupting chemicals on (7) (1989) 1857–1860. 12 J.K. Hess-Wilson, K.E. Knudsen / Cancer Letters 241 (2006) 1–12

[110] P.K. Mills, W.L. Beeson, R.L. Phillips, G.E. Fraser, Cohort the incidence of poorly differentiated prostatic adenocarcinoma study of diet, lifestyle, and prostate cancer in adventist men, in transgenic mice (TRAMP), Cancer Res. 61 (18) (2001) Cancer 64 (3) (1989) 598–604. 6777–6782. [111] T. Hirayama, Nutrition and cancer–a large scale cohort study, [129] W.A. Fritz, J. Wang, I.E. Eltoum, C.A. Lamartiniere, Dietary Prog. Clin. Biol. Res. 206 (1986) 299–311. genistein down-regulates androgen and estrogen receptor [112] S. Graham, B. Haughey, J. Marshall, R. Priore, T. Byers, expression in the rat prostate, Mol. Cell. Endocrinol. 186 (1) T. Rzepka, et al., Diet in the epidemiology of carcinoma of the (2002) 89–99. prostate gland, J. Natl Cancer Inst. 70 (4) (1983) 687–692. [130] M. Jang, J.M. Pezzuto, Cancer chemopreventive activity of [113] E. Giovannucci, A. Ascherio, E.B. Rimm, M.J. Stampfer, resveratrol, Drugs Exp. Clin. Res. 25 (2/3) (1999) 65–77. G.A. Colditz, W.C. Willett, Intake of carotenoids and retinol in [131] M. Hussain, M. Banerjee, F.H. Sarkar, Z. Djuric, M.N. Pollak, relation to risk of prostate cancer, J. Natl Cancer Inst. 87 (23) D. Doerge, et al., Soy isoflavones in the treatment of prostate (1995) 1767–1776. cancer, Nutr. Cancer 47 (2) (2003) 111–117. [114] S.K. Clinton, Lycopene: chemistry, biology, and implications [132] F.O. Stephens, Phytoestrogens and prostate cancer: possible for human health and disease, Nutr. Rev. 1 (2 Pt 1) (1998) 35– preventive role, Med. J. Aust. 167 (3) (1997) 138–140. 51. [133] H. Adlercreutz, W. Mazur, Phyto-oestrogens and western [115] L.N. Kolonel, A.M. Nomura, R.V. Cooney, Dietary fat and diseases, Ann. Med. 29 (2) (1997) 95–120. prostate cancer: current status, J. Natl Cancer Inst. 91 (5) [134] M.S. Kurzer, X. Xu, Dietary phytoestrogens, Annu. Rev. Nutr. (1999) 414–428. 17 (1997) 353–381. [116] J.H. Cohen, A.R. Kristal, J.L. Stanford, Fruit and vegetable [135] W. Demark-Wahnefried, D.T. Price, T.J. Polascik, intakes and prostate cancer risk, J. Natl Cancer Inst. 92 (1) C.N. Robertson, E.E. Anderson, D.F. Paulson, et al., Pilot (2000) 61–68. study of dietary fat restriction and flaxseed supplementation in [117] S.S. Strom, Y. Yamamura, C.M. Duphorne, M.R. Spitz, men with prostate cancer before surgery: exploring the effects R.J. Babaian, P.C. Pillow, S.D. Hursting, Phytoestrogen intake on hormonal levels, prostate-specific antigen, and histopatho- and prostate cancer: a case-control study using a new database, logic features, Urology 58 (1) (2001) 47–52. Nutr. Cancer 33 (1) (1999) 20–25. [136] K.S. Weber, K.D. Setchell, D.M. Stocco, E.D. Lephart, Dietary soy-phytoestrogens decrease testosterone levels and prostate [118] C. Nagata, Ecological study of the association between soy weight without altering LH, prostate 5alpha-reductase or product intake and mortality from cancer and heart disease in testicular steroidogenic acute regulatory peptide levels in adult Japan, Int. J. Epidemiol. 29 (5) (2000) 832–836. male sprague-dawley rats, J. Endocrinol. 170 (3) (2001) 591–599. [119] R.C. Habito, J. Montalto, E. Leslie, M.J. Ball, Effects of [137] K. Kato, S. Takahashi, L. Cui, T. Toda, S. Suzuki, replacing meat with soyabean in the diet on sex hormone M. Futakuchi, et al., Suppressive effects of dietary concentrations in healthy adult males, Br. J. Nutr. 84 (4) (2000) and on rat prostate carcinogenesis, Jpn. J. Cancer Res. 557–563. 91 (8) (2000) 786–791. [120] C. Nagata, N. Takatsuka, H. Shimizu, H. Hayashi, [138] X. Lin, B.R. Switzer, W. Demark-Wahnefried, Effect of T. Akamatsu, K. Murase, Effect of soymilk consumption on mammalian lignans on the growth of prostate cancer cell lines, serum estrogen and androgen concentrations in japanese men, Anticancer Res. 21 (6A) (2001) 3995–3999. Cancer Epidemiol. Biomarkers Prev. 10 (3) (2001) 179–184. [139] P.J. Magee, I.R. Rowland, Phyto-oestrogens, their mechanism [121] N.E. Allen, P.N. Appleby, G.K. Davey, T.J. Key, Soy milk of action: current evidence for a role in breast and prostate intake in relation to serum sex hormone levels in British men, cancer, Br. J. Nutr. 91 (4) (2004) 513–531. Nutr. Cancer 41 (1/2) (2001) 41–46. [140] J.H. Mitchell, S.J. Duthie, A.R. Collins, Effects of phytoestro- [122] M. Pollard, P.H. Luckert, Influence of isoflavones in soy gens on growth and DNA integrity in human prostate tumor cell protein isolates on development of induced prostate-related lines: PC-3 and LNCaP, Nutr. Cancer 38 (2) (2000) 223–228. cancers in L-W rats, Nutr. Cancer 28 (1) (1997) 41–45. [141] Y.A. Kim, S.H. Rhee, K.Y. Park, Y.H. Choi, Antiproliferative [123] M. Pollard, Prevention of prostate-related cancers in lobund- effect of resveratrol in human prostate carcinoma cells, J. Med. wistar rats, Prostate 39 (4) (1999) 305–309. Food. 6 (4) (2003) 273–280. [124] M. Onozawa, T. Kawamori, M. Baba, K. Fukuda, T. Toda, [142] J. Hempstock, J.P. Kavanagh, N.J. George, Growth inhibition H. Sato, et al., Effects of a isoflavone mixture on of prostate cell lines in vitro by phyto-oestrogens, Br. J. Urol. carcinogenesis in prostate and seminal vesicles of F344 rats, 82 (4) (1998) 560–563. Jpn. J. Cancer Res. 90 (4) (1999) 393–398. [143] G. Peterson, S. Barnes, Genistein and inhibit the [125] M. Landstrom, J.X. Zhang, G. Hallmans, P. Aman, A. Bergh, growth of human prostate cancer cells but not epidermal J.E. Damber, et al., Inhibitory effects of soy and rye diets on the growth factor receptor tyrosine autophosphorylation, Prostate development of dunning R3327 prostate adenocarcinoma in 22 (4) (1993) 335–345. rats, Prostate 36 (3) (1998) 151–161. [144] H.R. Naik, J.E. Lehr, K.J. Pienta, An in vitro and in vivo study [126] A. Bylund, J.X. Zhang, A. Bergh, J.E. Damber, A. Widmark, of antitumor effects of genistein on hormone refractory prostate A. Johansson, et al., Rye bran and soy protein delay growth and cancer, Anticancer Res. 14 (6B) (1994) 2617–2619. increase apoptosis of human LNCaP prostate adenocarcinoma [145] J.F. Santibanez, A. Navarro, J. Martinez, Genistein inhibits in nude mice, Prostate 42 (4) (2000) 304–314. proliferation and in vitro invasive potential of human prostatic [127] M. Pollard, W. Wolter, Prevention of spontaneous prostate- cancer cell lines, Anticancer Res. 17 (2A) (1997) 1199–1204. related cancer in lobund-wistar rats by a soy protein [146] J.N. Davis, N. Muqim, M. Bhuiyan, O. Kucuk, K.J. Pienta, isolate/isoflavone diet, Prostate 45 (2) (2000) 101–105. F.H. Sarkar, Inhibition of prostate specific antigen expression [128] R. Mentor-Marcel, C.A. Lamartiniere, I.E. Eltoum, by genistein in prostate cancer cells, Int. J. Oncol. 16 (6) (2000) N.M. Greenberg, A. Elgavish, Genistein in the diet reduces 1091–1097.

Chapter II:

Bisphenol A facilitates bypass of androgen ablation

therapy in prostate cancer

3181

Bisphenol A facilitates bypass of androgen ablation therapy in prostate cancer

Yelena B. Wetherill,1 Janet K. Hess-Wilson,1 androgen deprivation, BPA enhanced both cellular prolifer- Clay E.S. Comstock,1 Supriya A. Shah,1 ation rates and tumor growth. These effects were C. Ralph Buncher,2 Larry Sallans,3 mediated, at least in part, through androgen receptor Patrick A. Limbach,3 Sandy Schwemberger,4 activity, as prostate-specific antigen levels rose with George F. Babcock,4 and Karen E. Knudsen1,5,6 accelerated kinetics in BPA-exposed animals. Thus, at levels relevant to human exposure, BPA can modulate 1Department of Cell and Cancer Biology, 2Department of tumor cell growth and advance biochemical recurrence in Environmental Health, 3Rieveschl Laboratories for Mass tumors expressing the AR-T877A mutation. [Mol Cancer 4 Spectrometry, Department of Chemistry, Department of Ther 2006;5(12):3181–90] Surgery/Shriners Hospital for Children, 5Center for Environmental Genetics, and 6University of Cincinnati Cancer Center, University of Cincinnati College of Medicine, Cincinnati, Ohio Introduction Prostate cancer is the second leading cause of cancer death Abstract among males in the United States and the most commonly diagnosed malignancy (1). Organ-confined prostate cancer Prostatic adenocarcinomas depend on androgen for growth is associated with favorable outcome, as local disease can and survival. First line treatment of disseminated disease be effectively treated through radical prostatectomy or exploits this dependence by specifically targeting androgen radiation therapies (reviewed in refs. 2–5). However, receptor function. Clinical evidence has shown that disseminated disease is treated based on the androgen androgen receptor is reactivated in recurrent tumors dependence of the tumor, and androgen deprivation despite the continuance of androgen deprivation therapy. therapy (ADT) is the first line of therapeutic intervention Several factors have been shown to restore androgen (6, 7). Although ADT is initially effective, within 18 to 30 receptor activity under these conditions, including somatic months recurrent tumors ultimately arise, for which no mutation of the androgen receptor ligand-binding domain. effective treatment has been identified (8–10). As such, We have shown previously that select tumor-derived there is an urgent need to identify factors that influence the mutants of the androgen receptor are receptive to imminent transition to therapy resistance. activation by bisphenol A (BPA), an endocrine-disrupting ADT centers on ablation of androgen receptor function, compound that is leached from polycarbonate plastics and an androgen-dependent transcription factor whose activi- epoxy resins into the human food supply. Moreover, we ties are required for prostate cancer growth and progres- have shown that BPA can promote cell cycle progression in sion (11). In ADT, androgen receptor function is blocked cultured prostate cancer cells under conditions of androgen either through inhibition of androgen synthesis as achieved deprivation. Here, we challenged the effect of BPA on the through chemical or surgical means (6) and/or through the therapeutic response in a xenograft model system of use of androgen receptor antagonists that inhibit androgen prostate cancer containing the endogenous BPA-respon- receptor activity via direct interaction (12–15). ADT sive AR-T877A mutant protein. We show that after strategies are typically successful, as monitored by both tumor regression and a reduction in prostate-specific antigen (PSA) expression, a known androgen receptor target gene used clinically to monitor prostate cancer progression (16). Received 5/11/06; revised 9/5/06; accepted 10/13/06. A plethora of evidence shows that androgen receptor Grant support: NIH grant RO1-CA 093404 (K.E. Knudsen), NIH grant RR019900 (to P.A. Limbach), National Institute of Environmental Health activity is inappropriately restored in recurrent tumors Sciences Center for Environmental Genetics core grant E30-ES-06096, leading to therapeutic resistance. The androgen receptor is and National Institute of Environmental Health Sciences Environmental Mutagenesis and Cancer training grant ES-07250-16 (Y.B. Wetherill and aberrantly activated in late-stage prostate cancer through J.K. Hess-Wilson). multiple mechanisms, including androgen receptor ampli- The costs of publication of this article were defrayed in part by the fication, ligand-independent androgen receptor activation, payment of page charges. This article must therefore be hereby marked or gain-of-function androgen receptor mutations (17–22). advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Several tumor-derived androgen receptor mutations are Note: Y.B. Wetherill and J.K. Hess-Wilson contributed equally to this known to change the conformation of the ligand-binding work. domain and allow activation of the receptor by non- Requests for reprints: Karen E. Knudsen, Department of Cell and Cancer canonical ligands. For example, the AR-T877A mutant Biology, Vontz Center for Molecular Studies, University of Cincinnati results in the ability of the receptor to use androgen, College of Medicine, 3125 Eden Avenue, ML 0521, Cincinnati, OH 45267-0521. Phone: 513-558-7371; Fax: 513-558-4454. estrogen, progestins, and anti-androgens as ligands (23–26). E-mail: [email protected] This specific mutation is reported to occur in up to 31% of Copyright C 2006 American Association for Cancer Research. advanced prostate tumors (27) and is present in the LNCaP doi:10.1158/1535-7163.MCT-06-0272 model system of prostate cancer.

Mol Cancer Ther 2006;5(12). December 2006 3182 BPA Advances Progression to Therapy Resistance

We showed previously that a subset of prostate cancer inoculated s.c. in the right flanks of 6- to 8-week-old male cell lines harboring androgen receptor mutations might be nude mice. Mice were housed and monitored for tumor susceptible to growth stimulation by the known endocrine- growth every week. Most tumors were established 6 to disrupting compound bisphenol A (BPA; refs. 28, 29). BPA 8 weeks postinjection with a tumor take of f60%, is used in the manufacture of polycarbonate plastics and consistent with what has been reported in the literature epoxy resins and is leached from these materials into food (32). Tumor volumes were measured using a caliper and and water supplies (30). A recent study shows that 95% of expressed in mm3 using the following formula: volume adults in the United States have detectable BPA in their (mm3) = long diameter (mm)  short diameter (mm)2  urine (31). In previous studies, we showed that select 0.5236 (33). Once tumors reached 50 to 100 mm3, the mice androgen receptor tumor-derived mutants can be activated were surgically castrated under general anesthesia with by BPA, and BPA can cooperate with low-level androgen to and randomized into two cohorts and then enhance mutant androgen receptor activity (29). Moreover, implanted s.c. between the shoulders with 21-day release we showed that this action of BPA stimulated cell cycle pellets, either BPA (n = 11; 12.5 mg) or placebo (n =12; progression in LNCaP cells under conditions of androgen Innovative Research of America, Saratoga, FL). This deprivation (28). Although these studies suggested that constituted day 0 of treatment. After 35 days postcastra- exposure to endocrine-disrupting compounds may influ- tion, mice were euthanized according to University ence the cellular response to ADT, the effect on therapeutic protocol, at which point tumors were harvested, fixed in response in vivo had yet to be assessed. Here, we formalin, and embedded in paraffin for further histo- challenged the consequence of BPA on ADT in vivo using chemical analyses. a well-established xenograft model of prostate cancer. Our Serum Assays data indicate that after androgen ablation, environmentally Blood samples for determination of serum PSA, testos- relevant levels of BPA stimulated tumor cell proliferation terone, and BPA levels were obtained weekly by tail vein and enhanced tumor growth. In addition, animals exposed incisions in mice anesthetized with isoflurane. Serum PSA to BPA exhibited a shorter time to therapeutic resistance, levels were determined by an ELISA kit with a lower limit shown by a more rapid increase in PSA. Together, these sensitivity of 1 ng/mL (BioCheck, Inc., Foster City, CA) data show that incidental environmental exposure to BPA according to the manufacturer’s protocol. PSA doubling may advance the transition to therapeutic resistance in a time (PSA-DT) was calculated using the following formula:  À subset of prostate cancers harboring BPA-responsive PSA-DT = time (days) loge (2) / [loge (PSA2) loge mutations of the androgen receptor. (PSA1); ref. 34]. Serum testosterone levels were determined using the testosterone ELISA kit (Neogen, Lexington, KY) according to the manufacturer’s protocol. Materials and Methods Levels of BPA in serum were determined by high mass Animals accuracy-liquid chromatography/mass spectrometry, NCR/nu/nu (athymic) male mice, 6 to 8 weeks of age, wherein a 15 AL serum sample was mixed in a deactivated A were purchased from the Animal Production Area of the glass vial with 30 L of a deuterium-labeled BPA-d16 National Cancer Institute-Frederick Cancer Research Facil- (6.7 ng/mL) internal standard (Sigma-Aldrich, St. Louis, ity (Frederick, MD). The mice were housed in cages fitted MO) in an aqueous solution of 40% acetonitrile. A full with a high efficiency filter top in animal facilities loop (2 AL) was injected onto a Waters (Milford, MA) approved by the American Association for Accreditation XBridge 3.5 Am C18 column (100 mm  2.1 mm i.d.) and of Laboratory Animal Care. The room was kept at 25jC eluted under isocratic conditions using 3.7 mN ammonium with a 12-h light-dark cycle. All of the animal studies were hydroxide with 40% acetonitrile at 208 AL/min using a conducted in accordance with the principles and proce- Thermo Finnigan (Waltham, MA) Surveyor MS pump dures outlined by the NIH guidelines and the Institutional coupled with a Micro AS autosampler. Subsequent analysis Animal Care and Use Committee of the University of was done using a Thermo Finnigan LTQ-FT mass spec- Cincinnati (Cincinnati, OH). trometer. Electrospray conditions were as follows: capillary Assessment of Tumor Growth In vivo temperature, 325jC; source voltage, 4.5 kV; sheath gas LNCaP cells between passages 32to 37 were cultured in (nitrogen), 30 units; and auxiliary gas (nitrogen), 10 units. improved MEM (Biofluids, Rockville, MD) supplemented Instrument voltages were optimized using tetradecanoate with 5% fetal bovine serum (HyClone, Logan, UT), anion (m/z 227.20165) from column bleed. Spectra were 100 units/mL /streptomycin, and 2mmol/L acquired in the Fourier transform ion cyclotron resonance L- (Mediatech, Herndon, VA). Cells were grown portion of the instrument in negative ion profile mode; a j at 37 Cina5%CO2 humidified incubator. Before broad mass range of m/z 220.11 to m/z 250.11 was inoculation, LNCaP cells were washed, trypsinized, and required to observe both the deprotonated BPA ion (m/z resuspended in phenol red–free improved MEM supple- 227.10775) and its labeled analogue (m/z 241.19563). mented with 5% charcoal dextran–treated (CDT) serum. Resolution was set at 12,500 to provide a rapid duty cycle LNCaP cells were then combined with Matrigel (Becton while still providing ample separation from interferences Dickinson, Bedford, MA) at 3:1 volume ratio of medium to due to column bleed. The FTMS SIMS AGC Target was set Matrigel, and a total of 2  106 cells in 100 AL was at 500,000 with a maximum ion injection time of 1,500 ms.

Mol Cancer Ther 2006;5(12). December 2006 Molecular Cancer Therapeutics 3183

BPA standards 1.0, 2.5, 10, 20, 25, 30, 40, 50, 100, and 250 BrdUrd [BU1/75 (ICR1) Accurate Chemical & Scientific, ng/mL were run before and after the serum samples to Westbury, NY] in antibody dilution buffer. All remaining verify drift and used to calculate serum BPA levels. The steps were conducted at room temperature. Sections were limit of detection was 2ng/mL in serum. rinsed in PBS and incubated for 30 min using a 1:50 dilution H&E Staining of biotinylated rabbit anti-rat IgG in PBS containing 0.05% Paraffin-embedded LNCaP xenograft tumors were sliced sodium azide. Following a PBS wash, BrdUrd was into 5-Am sections using a Leica (Bannockburn, IL) visualized using a 1:200 dilution of horseradish peroxi- microtome (model RM 2125). Tissue sections were depar- dase-streptavidin in PBS containing 0.05% thimerosal. The affinized by immersion in xylene using 3 Â 5 min washes biotinylated secondary and streptavidin conjugate were and rehydrated stepwise to 70% ethanol, with a final from Vector Laboratories. Antigen development, counter- incubation in deionized water. Slides were stained in staining, dehydration, and clearing were done as described Harris’ hematoxylin for 30 s, rinsed in water for 4 min, above. and then stained in Harris eosin for 45 s. Both reagents Terminal Deoxynucleotidyl Transferase ^Mediated were purchased from Poly Scientific (Bay Shore, NY). dUTP Nick End Labeling Staining Subsequently, sections were incubated stepwise to 100% Sections from LNCaP xenograft tumors were deparaffi- ethanol, cleared in xylene using 3 Â 5 min washes, and nized and rehydrated as described above and then stained mounted using Permount (Fisher Scientific, FairLawn, NJ). using the DeadEnd Colorimetric terminal deoxynucleotidyl Immunohistochemistry transferase–mediated dUTP nick end labeling (TUNEL) Following the deparaffinization and rehydration proce- System according to the manufacturer’s specifications dure described above, antigen retrieval was achieved by (Promega, Madison, WI). The immersion of slides into boiling the slides in a 600 W microwave using an antigen paraformaldehyde was omitted and a 10-min incubation of unmasking solution (Vector Laboratories, Inc., Burlingame, proteinase K was optimal for permeabilization of the tissue. CA) for 5 min at full power and then for 20 min at 30% To generate a positive control, a random tumor section was power. Slides were removed from the antigen retrieval treated with DNase I (1.0 unit/mL) for 10 min. solution after 1 h at room temperature and rinsed in PBS, Counting and Statistical Assessment and peroxidase activity was subsequently quenched by Images were digitally captured (1,315 Â 1,033 pixels) 15 min of incubation in H2O2 blocking reagent (DAKO, using Spot version 4.09 (Diagnostic Instruments, Sterling Carpinteria, CA). Tissue sections were processed using Heights, MI) on a Nikon (Melville, NY) Microphot-FXA a Vectastain Elite avidin-biotin complex method rabbit microscope at Â20orÂ40 magnification. Triplicate images staining kit according to the manufacturer’s specifications from each slide for BrdUrd were counted blind using (Vector Laboratories). The following primary rabbit poly- MetaMorph version 6.3 (Molecular Devices, Sunnyvale, clonal antibodies, from Santa Cruz Biotechnology (Santa CA), and to ensure accuracy, a digital 7 Â 7 grid was Cruz, CA), were diluted in PBS and incubated overnight at applied to each image. Only the center 5 Â 5 grid was 4jC: 1:12,500 anti-androgen receptor (N-20), 1:1,000 anti- counted. Statistical analyses for BrdUrd were determined p21 (C-19), and 1:1,000 anti-cyclin D3 (C-16). The antigen using Student’s t test. Statistical analyses for tumor growth was visualized using a 3,3¶-diaminobenzidine substrate kit and serum PSA were determined used repeated measures for peroxidase (Vector Laboratories) using 2min of one-way ANOVA followed by a Tukey’s multiple compar- development. Omission of the primary antibodies yielded ison test. Statistical values of P < 0.05 were considered no 3,3¶-diaminobenzidine reactivity (data not shown). significant. Data are expressed as means F SE. Nuclei were visualized by counterstaining with Harris Cell Cycle Analyses hematoxylin and further processed as described above. Asynchronous LNCaP cells were seeded at 3.5 Â 105 per Bromodeoxyuridine Staining 6-cm dish into appropriate culture conditions, CDT + Sections from LNCaP xenograft tumors that had been ethanol (vehicle control), CDT plus 1 Amol/L BPA, CDT labeled in vivo with 10 mmol/L bromodeoxyuridine plus 1 nmol/L BPA, or 1 Amol/L paclitaxel (apoptotic (BrdUrd) for 1 h before sacrifice were deparaffinized, positive control), and allowed to propagate for either 24 or rehydrated, and peroxidase quenched as described above. 72h. Adherent cells and culture medium were collected via Antigen retrieval was accomplished by 10 min of incuba- trypsinization and fixed in 80% ice-cold ethanol. To tion in a 37jC humidified chamber using 1:6 dilution determine percentage of cells with sub-2N DNA content of prewarmed trypsin concentrate/trypsin diluent 1B for each culture condition, fixed cells were resuspended in (93-3943, Zymed Laboratories, San Francisco, CA) and f1 mL propidium iodide staining solution (50 Ag/mL rinsed in deionized water. The DNA was denatured with propidium iodide and 40 Ag/mL RNase). DNA content was 2N HCl and rinsed in PBS. Sections were blocked for then determined by flow cytometry analysis using a 30 min at room temperature with 2% normal rabbit serum Coulter Epics XL (Beckman Coulter, Miami, FL) with in an antibody dilution buffer [0.01 mol/L PBS (pH 7.2), a 488-nm argon-ion laser. 1% bovine serum albumin, 0.1% fish skin gelatin, 0.05% Reverse Transcription-PCR 6 sodium azide] containing 0.1% Triton X-100 and 0.05% LNCaP cells (10 ) were seeded onto poly-L-lysine– Tween 20. Sections were then probed for 1 h in a 37jC coated plates and incubated for 48 h in CDT medium as humidified chamber using a 1:100 dilution of rat anti- described above. Subsequently, cells were washed with

Mol Cancer Ther 2006;5(12). December 2006 3184 BPA Advances Progression to Therapy Resistance

PBS, and CDT medium with BPA (1 nmol/L) was added postcastration). Mass spectrometry was used as a reliable, for 24 h. Trizol reagent was used to extract total RNA, of sensitive assay to quantify BPA levels in serum, similar to which 5 Ag were used to generate the cDNA template previous reports (38, 39). As shown in Fig. 1B, BPA levels using the ThermoScript Reverse Transcription-PCR Sys- were readily detected as a single peak (Fig. 1B, inset), and tem according to the random hexamer protocol (Invitro- confidence in quantification was achieved at a lower limit of gen, Carlsbad, CA). PCRs were done with the following 2ng/mL (data not shown). No BPA was detected in the intron spanning primer sets (amplicon sizes are indicat- placebo-treated mice (data not shown), whereas mice ed): vascular endothelial growth factor (VEGF; 226 bp), implanted with the BPA release pellets exhibited a gradual 5¶-CTTTCTGCTGTCTTGGGTGCATTG-3¶ (forward) and decline in BPA over the time course (Fig. 1B). Average BPA 5¶-CACAGGATGGCTTGAAGATGTACTCG-3¶ (reverse) levels were highest at day 7 (27 ng/mL) and undetectable by and glyceraldehyde-3-phosphate dehydrogenase (597 day 35. The levels of BPA exposure match well with bp), 5¶-CCACCCATGGCAAATTCCATGGCA-3¶ (forward) reported doses for human exposure (reviewed in ref. 40). and 5¶-TCTAGACGGCAGGTCAGGTCCACC-3¶ (reverse). Therefore, the system used models the consequence of Amplifications were done using Taq DNA polymerase in environmentally relevant BPA exposure on prostate cancer buffer B (Promega) and conditions for each primer set growth after androgen deprivation. were as follows: VEGF, 95jC for 5 min; 45 cycles of 95jC To assess tumor growth, tumor volumes from each for 15 s, 57jC for 15 s, and 70jC for 30 s, with a final cohort were monitored weekly for the duration of the extension of 72jC for 5 min. Glyceraldehyde-3-phosphate dehydrogenase, 94jC for 2min; 27cycles of 94 jC for 30 s, 57jC for 45 s, and 72jC for 50 s, with a final extension of 72jC for 5 min.

Results Previously, we have shown that nanomolar levels of BPA provide mitogenic stimuli to androgen-dependent prostatic adenocarcinoma cells expressing somatic mutations of the androgen receptor (28, 29). Given the importance of the androgen receptor for prostate cancer growth and progres- sion, it was imperative to examine the effect of BPA on the response to standard therapeutic intervention in vivo. For these experiments, the LNCaP xenograft model was used, which expresses an endogenous BPA- responsive androgen receptor mutant (AR-T877A; ref. 28) and requires androgen receptor activation for tumor growth (24). After androgen ablation, this xenograft model under- goes a cytostatic response, which persists until recurrent tumors form (35). Therefore, we assessed the influence of BPA on recurrent tumor formation. For these studies, LNCaP cells were injected s.c. into flanks of intact athymic male mice and tumors allowed to form. Once tumors reached a volume approximating 100 mm3, the mice were surgically castrated and randomized into two cohorts. The first cohort was implanted with 21-day BPA time-release pellets (12.5 mg total BPA) at the time of castration. Concurrently, the second cohort was implanted with placebo pellets. To validate conditions approximating ADT, serum Figure 1. Effects of treatment on serum testosterone and BPA levels. A, testosterone levels were measured postcastration, as shown blood samples were collected from either intact male athymic mice or mice in Fig. 1A. Testosterone levels in age-matched intact male bearing LNCaP prostate tumors treated with either placebo or BPA 21-day time-release pellets (12.5 mg) following 35 d of androgen withdrawal. mice were found to be in the range of 1.2ng/mL, which is Testosterone serum concentrations were measured using an enzymatic within the physiologic range reported previously (36, 37). immunoassay. The data represent individual animal measurements: intact As expected, testosterone was virtually undetectable mice, n = 2; placebo-treated mice, n = 3; and BPA-treated mice, n =3. B, parallel studies of mice implanted with either placebo or BPA time- (0.003–0.162ng/mL) in castrated mice regardless of the release pellets (n = 4 for each group) were used to monitor serum BPA treatment (placebo or BPA), showing that established levels over the experimental time course by high mass accuracy-liquid LNCaP tumors in both groups were exposed to minimal chromatography/mass spectrometry. Inset, representative ion chromato- circulating androgen. Serum BPA levels of mice implanted gram from day 21 serum taken from a BPA-implanted mouse. No detectable BPA was found in placebo mice (data not shown). Columns, in parallel with the BPA or placebo pellets were examined average serum BPA levels in mice implanted with BPA time-release pellets; over the proposed time course of the experiment (0–35 days bars, SE.

Mol Cancer Ther 2006;5(12). December 2006 Molecular Cancer Therapeutics 3185

androgen withdrawal. In the BPA-treated cohort, tumor growth was also apparent by 21 days. This enhanced tumor growth trend continued through day 35 and resulted in significantly larger tumors (284 F 89 and 507 F 137.5 mm3 for placebo- and BPA-treated groups, respectively; P < 0.01). Thus, these data suggest that at relatively low levels, BPA is capable of promoting growth of prostate tumors in vivo in the absence of testicular androgen, suggesting that BPA may adversely affect the magnitude of the response to ADT. To determine if the increase in tumor size was attributed to differences in histologic architecture, analysis of tumor microanatomy was done, as shown in Fig. 3. As evident, all tumors exhibited highly disorganized and multicapsulated architecture with extensive hemorrhage throughout, re- gardless of the treatment (Fig. 3A, top, H&E; data not shown). Additionally, there was little difference in cell size Figure 2. BPA exposure reduces the efficacy of androgen deprivation and number, and both cohorts displayed densely packed on tumor growth. Once xenograft tumors achieved the approximate volume of 100 mm3, mice were surgically castrated, randomly selected, cells with evident mitotic figures and areas of necrosis (Fig. and implanted s.c. with either placebo or BPA 21-d time-release pellets 3A; data not shown). Androgen receptor expression was (12.5 mg). Tumor volume was measured weekly and calculated by the subsequently examined, as it has been shown that recurrent formula: volume (mm3) = long diameter (mm) Â short diameter (mm)2 Â 0.5236. Columns, average tumor size among the groups (n = 12 animals (ADT resistant) tumors frequently show more intense in the placebo-treated group; n = 11 animals for BPA-treated group); androgen receptor immunostaining than the primary bars, SE. *, P < 0.01; **, P < 0.001. lesion, and amplification and/or overexpression are caus- ative for therapeutic relapse (41–46). Androgen receptor was expressed in tumors from both treatment groups, and experiment. As shown in Fig. 2, LNCaP tumor volumes no discernable differences in androgen receptor levels were remained relatively static throughout 14 to 21 days observed (Fig. 3A; data not shown). Thus, alterations in postcastration, in agreement with previously published androgen receptor expression are unlikely to underlie observations (33, 37). Following this time point, tumor observed changes in tumor volume. It has been reported growth resumed and steadily increased in the placebo that BPA can induce VEGF mRNA expression in the uterus, group, reflecting the progression to therapeutic resistance vagina, and pituitary of Sprague-Dawley rats (47). Thus, reported for this model system under conditions of we examined VEGF mRNA levels after BPA exposure

Figure 3. LNCaP tumors treated by ADT in the presence of BPA exhibit increased proliferation. A, top and middle top, general LNCaP prostate tumor architecture in both treatment groups, H&E staining (magnifications, Â20 and Â40); bottom and middle bottom, immunohistochemical detec- tion of androgen receptor expression (magnifications, Â20 and Â40). B, VEGF reverse transcription-PCR on mRNA from MDA-MB-468 (positive control; lane 2) or LNCaP cells F1 nmol/L BPA (lanes 3 and 4). Glyceral- dehyde-3-phosphate dehydrogenase (GAPDH) mRNA was used to control for concentration of RNA in each reac- tion. C, BrdUrd incorporation (top; magnification, Â20) and graphical rep- resentation (bottom) of BrdUrd-positive cells from placebo- or BPA-exposed mice.

Mol Cancer Ther 2006;5(12). December 2006 3186 BPA Advances Progression to Therapy Resistance

in LNCaP cells (Fig. 3B). In this model system, VEGF therefore, to validate our studies, we measured serum expression was present but similar in both BPA-treated or PSA in both cohorts (Fig. 5A). In the control (placebo control-stimulated cells (Fig. 3B, compare lanes 3 and 4), as treated) group, PSA levels dropped 75% postcastration, in may be predicted based on pathologic inspection of the accordance with the published literature (Fig. 5A, compare tumors. Breast cancer MDA-MB-468 cells were used as a striped columns; days 0 and 7; refs. 32, 33). By day 21 in the positive control (Fig. 3B, lane 2; ref. 48), and glyceralde- placebo cohort, PSA levels rose to 40 ng/mL and remained hyde-3-phosphate dehydrogenase served as the internal static at this position through the remainder of the testing control for mRNA concentration (Fig. 3B, bottom). Com- period. In the BPA-treated group, the nadir PSA levels bined, these data indicate that the histopathologic features after androgen deprivation were higher compared with of the LNCaP xenografts are not significantly altered by the placebo-treated group (day 7), and elevated PSA levels BPA exposure. in the BPA group continued until the end point of To determine if the differences in tumor size were the experiment, wherein PSA levels were increased 50% attributed to enhanced proliferation, animals from both (P < 0.01). Moreover, the increase in PSA in the BPA- cohorts were injected with BrdUrd before sacrifice, and exposed group from days 21 to 35 was significant tumors were analyzed for BrdUrd incorporation by (P < 0.01), whereas the placebo-treated group did not immunohistochemistry (Fig. 3C). BrdUrd labeling yield a statistically significant increase in PSA. revealed that the BPA-treated tumors exhibited increased The rate at which PSA levels rose also increased in the proliferative indices compared with the placebo-treated BPA cohort compared with placebo-treated animals cohort (20% compared with 12%, respectively; P < 0.01; (Fig. 5B). PSA doubling time (PSA-DT) is often used Fig. 3C, bottom). These findings show that tumors exposed clinically to monitor ADT relapse or biochemical failure to BPA have a higher proportion of cells actively engaged (34, 50, 51). In Fig. 5B, serum PSA levels were plotted as in the cell cycle, thus contributing to the increased relative change over the nadir and the following formula Volume and enhanced growth rate of BPA-exposed was used to determine the average doubling time for PSA: Â À tumors. Together, these data strongly implicate BPA as a PSA-DT = time (days) loge (2) / [loge (PSA2) loge mitogenic stimulus for prostate tumors harboring the (PSA1); ref. 34]. BPA-treated tumors secreted PSA at a AR-T877A mutation. dramatically elevated rate (PSA-DT, 0.68 days) compared Increased tumor size in the BPA-treated animals could be with placebo-treated tumors (PSA-DT, 1.4 days). As this attributed to mechanisms other than enhanced cellular measurement is correlated with ADT resistance and tumor proliferation, such as a reduction in apoptotic rates. To test recurrence (50), these data show that environmentally this, cells were initially examined for sub-2N DNA content, relevant levels of circulating BPA could exacerbate progres- an indication of apoptotic DNA fragmentation and degra- sion from androgen-sensitive disease state to androgen- dation. BPA failed to induce quantifiable changes in independent, therapy-resistant cancer in prostate tumors apoptosis in LNCaP cells in culture at any BPA dose tested harboring the AR-T877A mutation. (1 Amol/L–1 nmol/L) following either 24 or 72 h of exposure (Fig. 4A, middle). Conversely, the chemothera- peutic agent paclitaxel induced accumulation of marked Discussion sub-2N DNA content and provided a positive control This study shows that environmentally relevant levels of (Fig. 4A, bottom). The apoptotic index was also monitored exposure to a prevalent endocrine-disrupting compound, in vivo through standard TUNEL assays in the recovered BPA, could promote prostate cancer growth and advance xenograft tumors. As shown in Fig. 4B, no perceptible biochemical recurrence in tumors harboring the endoge- distinctions in apoptotic cells between the BPA and placebo nous AR-T877A mutation. Thus, these results suggest that groups were apparent (Fig. 4B, middle and left, respec- incidental exposure to BPA may facilitate the transition of a tively). Overall, apoptotic rates were low in both cohorts, subset of prostate tumors to ADT resistance and therefore although the positive control (DNase treated) tumors may influence the duration and magnitude of the thera- showed high levels of positivity (Fig. 4B, right). Together, peutic response in prostate cancer patients. these data indicate that BPA exposure is unlikely to alter Endocrine-disrupting compounds have pleiotropic influ- rates of cell death in prostate cancer cells. ences on prostate cancer risk, development, and progres- In patients, tumor burden is typically monitored using a sion (reviewed in ref. 52). The various effects include serum biomarker, PSA, which is a well-defined androgen initiating changes in prostatic development, acting as receptor target gene (34). PSA detection is a powerful potential mutagens, or directly modulating the androgen measure of tumor size/growth and progression to hormone receptor pathway. Previous studies of BPA action in therapy-resistant prostate cancer and is invariably associat- prostate cancer cell lines revealed that the effect of BPA ed with rapidly rising PSA levels, termed biochemical on androgen-independent proliferation is likely confined to recurrence or failure (49). Previous studies have shown tumors expressing certain somatic androgen receptor that BPA can activate AR-T877A to induce PSA expression mutations (29). ADT is known to select for mutations of (28, 29). Thus, these data indicate that BPA may affect the the androgen receptor, most commonly in the ligand- timing of biochemical recurrence that precedes visible binding domain (53). These somatic mutations change the tumor formation. PSA expression is human specific; conformation of this domain, thereby facilitating activation

Mol Cancer Ther 2006;5(12). December 2006 Molecular Cancer Therapeutics 3187

Figure 4. Apoptotic rates are indistinguishable in place- bo- or BPA-treated cohorts. A, LNCaP cells treated with either ethanol (EtOH; vehicle control; top), 1 nmol/L or 1 Amol/L BPA (middle), or 1 Amol/L paclitaxel (PTX; positive apoptotic con- trol) for indicated time points were analyzed for sub-2N DNA content by propidium iodide staining and fluorescence-acti- vated cell sorting analyses. Quantification of cell cycle pro- file. B, representative TUNEL images from xenograft tumor sections from placebo-exposed (left) or BPA-exposed (middle) mice (arrows, positive nuclei). DNase-treated section included as a positive control (right). Magnification, Â40.

Mol Cancer Ther 2006;5(12). December 2006 3188 BPA Advances Progression to Therapy Resistance

relevant to human exposure (between 2and 27ng/mL; Fig. 1B; ref. 40) for 35 days after surgical castration showed an increased tumor size relative to the placebo- treated group (Fig. 2). Vascular density in prostate tumors directly correlates with advanced tumor stage and metastatic development (59, 60). Given that angiogenesis can be promoted by estrogen (61) and that BPA is a known estrogen mimic, histoarchitectural features were evaluated between the groups to address the possibility that altered tumor size was attributed to unbalanced vascular density. No distinctions were observed between the two groups, and no variances in VEGF production were detected following BPA exposure (Fig. 3), indicating that the disparity in size was likely attributed to enhanced proliferation. This hypothesis was supported by analyses of proliferative index, wherein BPA-exposed tumors showed a higher overall BrdUrd incorporation rate (Fig. 3C). Thus, the data shown indicate that BPA mediates increased cellular proliferation, most likely through AR-T877A. This prediction is strongly supported by in vitro analyses of BPA effects on AR-T877A (28, 29) and in the observation that elevated PSA levels in the BPA-treated cohort preceded significant induction of tumor size (Figs. 2and 5). It has been shown previously that BPA can alter Figure 5. BPA exposure advances progression to therapy resistance. expression of several key cell cycle regulators, including Blood samples were collected weekly starting at day 0 of the experiment, which corresponded to surgical castration. Serum PSA levels were p21, cyclin A, and the D type cyclins (29). Although determined by enzymatic immunoassay. A, tumor-derived serum PSA preliminary immunohistochemical investigations revealed concentration. After initial decrease in PSA levels (day 7 postcastration some alteration in p21 expression as a function of BPA versus day 0), PSA concentrations started to increase again at day 21 of exposure (data not shown), these trends did not reach androgen ablation, reaching statistical significance between placebo- and BPA-treated groups at day 35. *, P < 0.01. n = 12 in the placebo-treated statistical significance. AKT activity was also monitored, group and n = 11 in BPA-treated group. Results were average and BPA exposure failed to alter the levels of AKT concentration. Columns, PSA (ng/mL); bars, SE. B, rate of PSA increase phosphorylation in this model system (data not shown). for both placebo-and BPA-exposed animals. Change in PSA was plotted relative to the nadir for each cohort. PSA-DT was calculated as PSA-DT= However, it should be noted that LNCaP cells harbor  time (days) loge (2) / [loge (PSA2) À loge (PSA1); ref. 34]. constitutively active AKT (62, 63) that is not further activated by androgen receptor ligands (including dihy- drotestosterone; ref. 64). Additionally, a recent study by by noncanonical endogenous steroid hormones (e.g., Ho et al. (65) found that mice exposed to low-level BPA estrogen, cortisol, and progesterone; refs. 27, 54) as well during development exhibited alternations in DNA as converting therapeutic androgen receptor antagonists methylation patterns in cell signaling genes, contributing into agonists (e.g., flutamide; refs. 27, 55–57). Several to increased incidence of precancerous lesions within the tumor-derived androgen receptor mutants are responsive mouse prostate. These data indicate that altered methyl- to BPA, including AR-T877A, AR-H874Y, AR-V715M, and ation of genes associated with proliferation may, in part, AR-T877S (29). AR-T877A, the most frequently reported underlie the response to BPA. Therefore, the precise somatic mutation in prostate cancer, is suggested to arise in mechanisms by which BPA influences cell cycle progres- 8% to 31% of therapy-resistant tumors (27, 58). The present sion in vivo will be the subject of future analyses. study addresses the influence of BPA in the presence of The transition to therapeutic resistance in prostate cancer this prevalent androgen receptor mutant protein. Current- is frequently associated with enhanced androgen receptor ly, the influence of BPA on prostate cancer therapeutic expression (43, 44, 46, 66). Increased androgen receptor is response in the presence of the remaining androgen both necessary and sufficient to render prostate cancer cells receptor mutants is hindered by the paucity of appropriate resistant to anti-androgen therapy, as amplified androgen model systems. Thus, further investigation is needed to receptor enhances the basal level response to low residual determine whether the effect of BPA on androgen hormoneaswellasmodifiestheandrogenreceptor deprivation is limited to tumors expressing the AR- response to antagonists (42). Exposure to low levels of T877A mutant. estrogenic compounds, including BPA, during neonatal The ability of BPA to impinge on the AR-T877A mutant development alters androgen receptor signaling and results is proposed to underlie its proliferative effect on tumor in changes in prostate development and size (67). More- growth in vivo. Mouse tumors exposed to BPA levels over, it is notable that BPA has been shown to alter

Mol Cancer Ther 2006;5(12). December 2006 Molecular Cancer Therapeutics 3189

androgen receptor levels in vivo (68). Therefore, we Acknowledgments analyzed androgen receptor expression in the xenograft We thank Dr. Monica P. Revelo (UC Department of Pathology) for tumors and observed no differences in androgen receptor pathologic assessment; Dr. Eric R. Hugo (UC Department of Cell and Cancer Biology) for providing the VEGF primers and MDA-MB-468 cDNA; levels between BPA- and placebo-exposed tumors (Fig. 3A; Drs. Erik Knudsen, Craig Burd, and Lisa Morey as well as Nick Olshavsky, data not shown). Therefore, the transition to therapeutic Kevin Link, William Zagorski, and Ankur Sharma (UC Department of Cell and Cancer Biology) for insightful commentary and assistance with editing resistance and PSA elevation was likely associated with of the manuscript. AR-T877A activation, rather than elevation of androgen receptor protein levels. References The observation that BPA-exposed tumors exhibited 1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer increased PSA secretion is consistent with previous J Clin 2005;55:10 – 30. reports in vitro,whereinBPAinducedPSAmRNA 2. Sharifi N, Farrar WL. Androgen receptor as a therapeutic target for expression in LNCaP cells (28, 29). This finding is of androgen independent prostate cancer. Am J Ther 2006;13:166 – 70. importance, as progression to therapy-resistant prostate 3. Damber JE, Khatami A. Surgical treatment of localized prostate cancer. Acta Oncol 2005;44:599 – 604. cancer is often associated clinically with rapidly rising 4. Kirkpatrick JP, Anscher MS. Radiotherapy for locally recurrent prostate PSA levels (49). Given the relationship between androgen cancer. Clin Adv Hematol Oncol 2005;3:933 – 42. receptor activity and prostate cancer proliferation, PSA 5. Zeliadt SB, Ramsey SD, Penson DF, et al. Why do men choose one serum levels are used clinically to diagnose prostate treatment over another? A review of patient decision making for localized cancer and monitor tumor progression (16). Moreover, in prostate cancer. Cancer 2006;106:1865 – 74. 6. Dehm SM, Tindall DJ. Molecular regulation of androgen action in both tissue culture and xenograft model systems of prostate cancer. J Cell Biochem 2006;99:333 – 44. prostate cancer, a reduction in PSA follows inactivation 7. Eder IE, Haag P, Bartsch G, Klocker H. Targeting the androgen receptor of androgen receptor (either through direct androgen in hormone-refractory prostate cancer-new concepts. Future Oncol 2005; receptor antagonists or androgen deprivation treatment) 1:93 – 101. and a recurrence of PSA expression is seen after sup- 8. Denmeade SR, Isaacs JT. A history of prostate cancer treatment. Nat Rev Cancer 2002;2:389 – 96. plementation with androgen or androgen receptor ago- 9. Denis LJ. The role of active treatment in early prostate cancer. nists and is concomitant with the onset of ADT resistance. Radiother Oncol 2000;57:251 – 8. Throughout the duration of this study period, serum PSA 10. Dyrstad SW, Shah P, Rao K. Chemotherapy for prostate cancer. Curr levels were monitored and BPA-exposed tumors elicited Pharm Des 2006;12:819 – 37. an earlier and more dramatic elevation in PSA compared 11. Trapman J, Brinkmann AO. The androgen receptor in prostate cancer. with the placebo-treated animals. PSA-DT was signifi- Pathol Res Pract 1996;192:752 – 60. cantly enhanced in the BPA-treated group versus the 12. Shang Y, Myers M, Brown M. Formation of the androgen receptor transcription complex. Mol Cell 2002;9:601 – 10. placebo control group. PSA-DT is used clinically to moni- 13. Mendelsohn LG. Prostate cancer and the androgen receptor: tor ADT relapse, and rapidly rising PSA after intervention strategies for the development of novel therapeutics. Prog Drug Res is referred to as biochemical failure (34, 50, 51). Thus, the 2000;55:213 – 33. examined xenograft model system exhibited a shorter time 14. Hirawat S, Budman DR, Kreis W. The androgen receptor: structure, to biochemical recurrence after BPA exposure (Fig. 5). mutations, and antiandrogens. Cancer Invest 2003;21:400 – 17. 15. Freeman SN, Mainwaring WI, Furr BJ. A possible explanation for the Although these studies implicate BPA as a contributing peripheral selectivity of a novel non-steroidal pure antiandrogen, Casodex factor for therapy resistance in tumors challenged by ADT (ICI 176,334). Br J Cancer 1989;60:664 – 8. [particularly with gonaclotropin-releasing hormone 16. Stephan C, Jung K, Diamandis EP, Rittenhouse HG, Lein M, Loening (GnRH) agonists or finasteride] and expressing BPA- SA. Prostate-specific antigen, its molecular forms, and other kallikrein sensitive androgen receptor mutants, it is imperative to markers for detection of prostate cancer. Urology 2002;59:2 – 8. 17. Feldman BJ, Feldman D. The development of androgen-independent note that direct androgen receptor antagonists, such as prostate cancer. Nat Rev Cancer 2001;1:34 – 45. bicalutamide, effectively counter BPA activity in cultured 18. Taplin ME, Bubley GJ, Shuster TD, et al. Mutation of the androgen- cells (28, 29) Thus, any increase in PSA that is attributed to receptor gene in metastatic androgen-independent prostate cancer. N Engl BPA exposure would be expected to be reversed through J Med 1995;332:1393 – 8. the administration of such androgen receptor antagonists. 19. Taplin ME, Bubley GJ, Ko YJ, et al. Selection for androgen receptor mutations in prostate cancers treated with androgen antagonist. Cancer Additional studies to address the long-term effect of BPA Res 1999;59:2511 – 5. exposure on the response to androgen receptor antagonists 20. Taplin ME, Rajeshkumar B, Halabi S, et al. Androgen receptor are currently under way. mutations in androgen-independent prostate cancer: Cancer and Leukemia In summary, this study suggests that exposure of Group B Study 9663. J Clin Oncol 2003;21:2673 – 8. advanced prostate tumors expressing somatic mutations 21. Taplin ME, Balk SP. Androgen receptor: a key molecule in the progression of prostate cancer to hormone independence. J Cell Biochem in the androgen receptor gene (that frequently occur in 2004;91:483 – 90. ADT; refs. 18, 19, 21) to a highly prevalent endocrine- 22. Tilley WD, Buchanan G, Hickey TE, Bentel JM. Mutations in the disrupting compound can facilitate ADT bypass and androgen receptor gene are associated with progression of human prostate advance tumor recurrence. Hence, our study lays a cancer to androgen independence. Clin Cancer Res 1996;2:277 – 85. foundation for defining the clinical variables wherein 23. Zhao XY, Malloy PJ, Krishnan AV, et al. Glucocorticoids can promote androgen-independent growth of prostate cancer cells through a mutated BPA and other endocrine-disrupting chemicals may con- androgen receptor. Nat Med 2000;6:703 – 6. tribute to prostate cancer progression and alter treatment 24. Veldscholte J, Berrevoets CA, Brinkmann AO, Grootegoed JA, Mulder success. E. Anti-androgens and the mutated androgen receptor of LNCaP cells:

Mol Cancer Ther 2006;5(12). December 2006 3190 BPA Advances Progression to Therapy Resistance

differential effects on binding affinity, heat-shock protein interaction, and 46. Culig Z, Steiner H, Bartsch G, Hobisch A. Mechanisms of endocrine transcription activation. Biochemistry 1992;31:2393 – 9. therapy-responsive and -unresponsive prostate tumours. Endocr Relat Cancer 2005;12:229 – 44. 25. Veldscholte J, Berrevoets CA, Ris-Stalpers C, et al. The androgen receptor in LNCaP cells contains a mutation in the ligand binding domain 47. Long X, Burke KA, Bigsby RM, Nephew KP. Effects of the which affects steroid binding characteristics and response to antiandro- xenoestrogen bisphenol A on expression of vascular endothelial growth gens. J Steroid Biochem Mol Biol 1992;41:665 – 9. factor (VEGF) in the rat. Exp Biol Med (Maywood) 2001;226:477 – 83. 26. Veldscholte J, Ris-Stalpers C, Kuiper GG, et al. A mutation in the 48. Weigand M, Hantel P, Kreienberg R, Waltenberger J. Autocrine ligand binding domain of the androgen receptor of human LNCaP cells vascular endothelial growth factor signalling in breast cancer. Evidence affects steroid binding characteristics and response to anti-androgens. from cell lines and primary breast cancer cultures in vitro. Angiogenesis Biochem Biophys Res Commun 1990;173:534 – 40. 2005;8:197 – 204. 27. Suzuki H, Sato N, Watabe Y, Masai M, Seino S, Shimazaki J. 49. Balk SP, Ko YJ, Bubley GJ. Biology of prostate-specific antigen. J Clin Androgen receptor gene mutations in human prostate cancer. J Steroid Oncol 2003;21:383 – 91. Biochem Mol Biol 1993;46:759 – 65. 50. Buyyounouski MK, Hanlon AL, Eisenberg DF, et al. Defining 28. Wetherill YB, Petre CE, Monk KR, Puga A, Knudsen KE. The biochemical failure after radiotherapy with and without androgen depriva- xenoestrogen bisphenol A induces inappropriate androgen receptor tion for prostate cancer. Int J Radiat Oncol Biol Phys 2005;63:1455 – 62. activation and mitogenesis in prostatic adenocarcinoma cells. Mol Cancer 51. Pickles T. Prostate-specific antigen (PSA) bounce and other fluctua- Ther 2002;1:515 – 24. tions: which biochemical relapse definition is least prone to PSA false 29. Wetherill YB, Fisher NL, Staubach A, Danielsen M, de Vere White RW, calls? An analysis of 2030 men treated for prostate cancer with external Knudsen KE. Xenoestrogen action in prostate cancer: pleiotropic effects beam or brachytherapy with or without adjuvant androgen deprivation dependent on androgen receptor status. Cancer Res 2005;65:54 – 65. therapy. Int J Radiat Oncol Biol Phys 2006;64:1355 – 9. 30. Haighton LA, Hlywka JJ, Doull J, Kroes R, Lynch BS, Munro IC. An 52. Hess-Wilson JK, Knudsen KE. Endocrine disrupting compounds and evaluation of the possible carcinogenicity of bisphenol A to humans. Regul prostate cancer. Cancer Lett 2006;241:1 – 12. Toxicol Pharmacol 2002;35:238 – 54. 53. Gottlieb B, Beitel LK, Wu JH, Trifiro M. The androgen receptor gene 31. Calafat AM, Kuklenyik Z, Reidy JA, Caudill SP, Ekong J, Needham LL. mutations database (ARDB): 2004 update. Hum Mutat 2004;23:527 – 33. Urinary concentrations of bisphenol A and 4-nonylphenol in a human 54. Culig Z, Hobisch A, Cronauer MV, et al. Mutant androgen receptor reference population. Environ Health Perspect 2005;113:391 – 5. detected in an advanced-stage prostatic carcinoma is activated by adrenal 32. Sato N, Gleave ME, Bruchovsky N, Rennie PS, Beraldi E, Sullivan LD. androgens and progesterone. Mol Endocrinol 1993;7:1541 – 50. A metastatic and androgen-sensitive human prostate cancer model using 55. Hara T, Miyazaki J, Araki H, et al. Novel mutations of androgen intraprostatic inoculation of LNCaP cells in SCID mice. Cancer Res 1997; receptor: a possible mechanism of bicalutamide withdrawal syndrome. 57:1584 – 9. Cancer Res 2003;63:149 – 53. 33. Gleave M, Hsieh JT, Gao CA, von Eschenbach AC, Chung LW. 56. Chatterjee B. The role of the androgen receptor in the development of Acceleration of human prostate cancer growth in vivo by factors produced prostatic hyperplasia and prostate cancer. Mol Cell Biochem 2003;253: by prostate and bone fibroblasts. Cancer Res 1991;51:3753 – 61. 89 – 101. 34. Haukaas SA, Halvorsen OJ, Daehlin L, Hostmark J, Akslen LA. Is 57. Bohl CE, Gao W, Miller DD, Bell CE, Dalton JT. Structural basis for preoperative serum prostate-specific antigen level significantly related to antagonism and resistance of bicalutamide in prostate cancer. Proc Natl clinical recurrence after radical retropubic prostatectomy for localized Acad Sci U S A 2005;102:6201 – 6. prostate cancer? BJU Int 2006;97:51 – 5. 58. Marcelli M, Ittmann M, Mariani S, et al. Androgen receptor mutations 35. Lim DJ, Liu XL, Sutkowski DM, Braun EJ, Lee C, Kozlowski JM. in prostate cancer. Cancer Res 2000;60:944 – 9. Growth of an androgen-sensitive human prostate cancer cell line, LNCaP, 59. Weidner N, Carroll PR, J, Blumenfeld W, Folkman J. Tumor in nude mice. Prostate 1993;22:109 – 18. angiogenesis correlates with metastasis in invasive prostate carcinoma. 36. Wang J, Eltoum IE, Lamartiniere CA. Genistein alters growth factor Am J Pathol 1993;143:401 – 9. signaling in transgenic prostate model (TRAMP). Mol Cell Endocrinol 60. Bigler SA, Deering RE, Brawer MK. Comparison of microscopic 2004;219:171 – 80. vascularity in benign and malignant prostate tissue. Hum Pathol 1993;24: 37. Jin RJ, Wang Y, Masumori N, et al. NE-10 neuroendocrine cancer 220 – 6. promotes the LNCaP xenograft growth in castrated mice. Cancer Res 61. Seo KH, Lee HS, Jung B, et al. Estrogen enhances angiogenesis 2004;64:5489 – 95. through a pathway involving platelet-activating factor-mediated nuclear 38. Sajiki J, Takahashi K, Yonekubo J. Sensitive method for the factor-nB activation. Cancer Res 2004;64:6482 – 8. determination of bisphenol-A in serum using two systems of high- 62. Chen X, Thakkar H, Tyan F, et al. Constitutively active Akt is an performance liquid chromatography. J Chromatogr B Biomed Sci Appl important regulator of TRAIL sensitivity in prostate cancer. Oncogene 1999;736:255 – 61. 2001;20:6073 – 83. 39. Sambe H, Hoshina K, Hosoya K, Haginaka J. Direct injection analysis 63. Nesterov A, Lu X, Johnson M, Miller GJ, Ivashchenko Y, Kraft AS. of bisphenol A in serum by combination of isotope imprinting with liquid Elevated AKT activity protects the prostate cancer cell line LNCaP from chromatography-mass spectrometry. Analyst 2005;130:38 – 40. TRAIL-induced apoptosis. J Biol Chem 2001;276:10767 – 74. 40. Welshons WV, Nagel SC, vom Saal FS. Large effects from small 64. Xu Y, Chen SY, Ross KN, Balk SP. Androgens induce prostate cancer exposures. III. Endocrine mechanisms mediating effects of bisphenol A at cell proliferation through mammalian target of rapamycin activation and levels of human exposure. Endocrinology 2006;147:S56 – 69. post-transcriptional increases in cyclin D proteins. Cancer Res 2006;66: 41. Culig Z, Hoffmann J, Erdel M, et al. Switch from antagonist to agonist 7783 – 92. of the androgen receptor bicalutamide is associated with prostate tumour 65. Ho SM, Tang WY, Belmonte de Frausto J, Prins GS. Developmental progression in a new model system. Br J Cancer 1999;81:242 – 51. exposure to estradiol and bisphenol A increases susceptibility to prostate 42. Chen CD, Welsbie DS, Tran C, et al. Molecular determinants of carcinogenesis and epigenetically regulates phosphodiesterase type 4 resistance to antiandrogen therapy. Nat Med 2004;10:33 – 9. variant 4. Cancer Res 2006;66:5624 – 32. 43. Visakorpi T, Hyytinen E, Koivisto P, et al. In vivo amplification of the 66. Zhang H, Yi X, Sun X, et al. Differential gene regulation by the SRC androgen receptor gene and progression of human prostate cancer. Nat family of coactivators. Genes Dev 2004;18:1753 – 65. Genet 1995;9:401 – 6. 67. Gupta C. Reproductive malformation of the male offspring following 44. Koivisto P, Kononen J, Palmberg C, et al. Androgen receptor gene maternal exposure to estrogenic chemicals. Proc Soc Exp Biol Med 2000; amplification: a possible molecular mechanism for androgen deprivation 224:61 – 8. therapy failure in prostate cancer. Cancer Res 1997;57:314 – 9. 68. Ramos JG, Varayoud J, Sonnenschein C, Soto AM, Munoz De Toro 45. Gregory CW, He B, Johnson RT, et al. A mechanism for androgen M, Luque EH. Prenatal exposure to low doses of bisphenol A alters the receptor-mediated prostate cancer recurrence after androgen deprivation periductal stroma and glandular cell function in the rat ventral prostate. therapy. Cancer Res 2001;61:4315 – 9. Biol Reprod 2001;65:1271 – 7.

Mol Cancer Ther 2006;5(12). December 2006

Chapter III:

DDE modulates proliferation of prostate cancer cells through

MAPK mutant AR pathways DDE modulates proliferation of prostate cancer cells through MAPK mutant AR

pathways

Janet K. Hess-Wilson1, Supriya E. Shah1, Hannah E. Daly1, Soibhan Webb1, Karen E.

Knudsen1,2,3

1Department of Cell and Cancer Biology, 2Center for Environmental Genetics, 3UC

Cancer Center, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0521

Financial Support: NIH grant RO1-CA 093404 (to K.E.K.), NIEHS Environmental Mutagenesis and Cancer training grant ES-07250-16 (to J.K.H.)

Running Title: DDE induces prostate cancer cell proliferation

Keywords: androgen receptor, tumor-derived mutants, environmental exposure, DDE, organochlorine

The text of this chapter is primary research to be submitted for publication

Abstract.

Somatic, gain-of-function mutations of the androgen receptor (AR) that occur during prostate cancer progression are known to promote recurrent tumor formation and therapeutic resistance. These mutations enable promiscuous AR activation by endogenous non-canonical ligands (e.g. estrogen, progesterone, cortisol). In addition, an estrogenic endocrine disrupting compound present in the environment, bisphenol A, can activate selected tumor-derived AR mutants to induce androgen-independent proliferation and promote therapeutic bypass. Since these data indicate that AR mutation can sensitize the receptor to hormonally active compounds in the environment, the present study challenged the capacity of known endocrine disruptors to modulate tumor-derived AR mutant function. Using a yeast colorimetric reporter assay, these data revealed that several tumor-derived AR mutants are significantly receptive to activation by the pesticide, p,p’ DDE. Further investigation revealed that DDE increases mutant recruitment to the regulatory regimes of AR target genes and androgen- independent gene activation. Furthermore, DDE stimulated androgen independent cellular proliferation only within prostate cancer cells expressing the DDE-responsive

AR-T877A mutant (LNCaP cells). DDE did not effect the proliferation of androgen- independent, AR-H874Y expressing 22Rv1 prostate cancer cells. The mitogenic activity of DDE occurred in the nanomolar range, showing that low, environmentally relevant levels are required to stimulate growth. Consistent with previous reports, activation of the MAPK pathway was also observed upon DDE exposure in prostate cancer cells; therefore, the relative impact of mutant AR and MAPK activation for androgen-independent cell proliferation was assessed. There data herein demonstrate that while each pathway contributes to DDE-induces cellular proliferation, MAPK inhibitors effectively ablated androgen or DDE induced cell cycle progression.

Combined, these data indicate that DDE acts as an inappropriate mitogen in androgen dependent (AR-T877A dependent) prostate cancer cells, through its ability to modulate both mutant AR and MAPK pathways.

Introduction.

Prostatic adenocarcinoma is the second leading cause of cancer death in the

United States (1). At early stages, prostate tumors are dependant on androgen for growth (2). Androgen deprivation therapy (ADT) is therefore the first line of therapy for such patients, and is achieved via bilateral orchiectomy and/or through the use of anti- androgen chemotherapeutics (3-5). This treatment is initially effective, as the majority of patients (>80%) incur tumor remission (2). However, tumor remission is transient, as most patients relapse and develop therapy resistant, androgen-independent cancer, in which the androgen receptor (AR) has been inappropriately reactivated (6). At present, there is no effective treatment for therapy resistant prostate cancer, however, it is clear that AR reactivation is critical for the development of therapy resistant tumors.

AR is a member of the nuclear receptor superfamily of ligand dependent transcription factors and has a major role in regulating function, growth and differentiation of the prostate gland (7). Prior to ligand binding, AR is held inactive in the cytoplasm bound to heat shock proteins. In the prostate, testosterone is converted to dihydrotestosterone (DHT), which has high affinity for the AR (8). Binding of DHT to the receptor triggers dislocation of the inhibitory heat shock proteins, and rapid translocation to the nucleus. Once in the nucleus the complex binds to specific androgen responsive elements on target genes, which, with the aid of coactivators, results in target gene transcription (7, 9). The expression of a specific AR target gene, prostate specific antigen (PSA) is important as a clinical marker for prostate cancer progression. Serum

PSA levels directly correlate to AR activity, therefore, monitoring PSA levels provides a correlative measure of tumor cell proliferation and disease progression (10). As prostate cancers are dependent on AR activity for survival and proliferation, ablation of AR action via depletion of endogenous ligand (e.g. bilateral orchiectomy or gonadotropin releasing hormone antagonists) (11) or through the use of direct AR antagonist (e.g. bicalutamide) (12, 13) remains the first line of treatment for disseminated disease. Strikingly, there is no effective treatment for recurrent tumors, which lead to patient death (14). As such there is significant need to identify the factors that contribute to AR re-activation and recurrent tumor formation. Multiple mechanisms for AR re- activation have been identified including (i) AR amplification (ii) overexpression of AR co-activators (iii) ligand-independent AR activation or (iv) gain-of-function AR somatic mutations (6, 15-17). The competence of these mechanisms to initiate tumor recurrence is strongly responsive to factors that enhance AR signaling.

Previously, we showed that low levels of a common environmental contaminant, bisphenol A (BPA) could negatively affect prostate cancer treatment. BPA activated a specific subset of tumor derived AR mutants (18), and through this mechanism, stimulated prostate cancer cell proliferation under conditions of androgen depletion (19).

Moreover, environmentally relevant levels of BPA were sufficient to induce ADT bypass in in vivo models of prostate cancer. These data begin to examine the impact of endocrine disrupting compounds (EDCs) on prostate cancer management and have important clinical implications. Given these previous findings, we sought to determine the scope of environmental impact on this specific mode of AR reactivation (gain-of- function mutation) in prostate cancer therapy resistance. Strikingly, the known endocrine disrupting compound p,p’DDE (2,2-Bis(4-chlorophenyl)-1,1-dichloroethylene; a stable metabolite of the pesticide DDT) was able to activate select tumor derived mutant ARs. Despite the ban of its use in many countries, DDE is still a persistent environmental pollutant and can be detected in between 50 and 99% of human serum analysis globally (20). In prostate cancer model systems expressing DDE-responsive mutant ARs, DDE was only moderately able to induce AR transactivation, dependent upon cellular context. However, DDE could induce androgen independent cell proliferation at very low and environmentally relevant levels. The mitogenic affects are mediated in combination through both AR-T877A and MAPK pathways, as DDE was able to induce Erk phosphorylation, and moderately required AR activation and but was completely dependent upon MAPK activation for the mitogenic response. Combined, these data indicate that for tumors harboring select AR mutants, exposure to environmentally relevant levels of p,p’DDE may promote androgen-independent tumor cell proliferation thereby impact the rate of tumor progression.

Materials and Methods.

Reagents

Dihydrotestosterone (DHT), 1,1-Bis(4-chlorophenyl)-2,2,2-trichloroethane (P,P’ DDT),

2,2-Bis(4-chlorophenyl)-1,1-dichloroethylene (P,P’ DDE), resveratrol (Res), coumestrol

(Cou), and cadmium (Cd) were obtained from Sigma-Aldrich (St. Louis, MO). All reagents were solubilized in ethanol and stored at –20oC. Casodex (bicalutamide) was a generous gift from AstraZeneca Pharmaceuticals (London, United Kingdom) and was solubilized in DMSO to 10-2M and stored at –20oC. The MEK inhibitor, U0126, was from

Promega (Madison, WI) and used as per manufactor’s protocol. Cell Culture and treatments

LNCaP cells were obtained from American Type Culture Collection (Rockville, MD) and were utilized between passage 28 and 40 and maintained in IMEM (Cellgro, Mediatech,

Herdon, VA) containing 5% heat-inactivated fetal bovine serum (FBS; Biofluids,

Rockville, MD). The 22Rv1 cell line was the gift of Dr. J. Jacobberger (Case Western

Reserve University, Cleveland, OH) and was maintained in DMEM containing 10% heat-inactivated FBS. CV1 cells were also maintained in DMEM containing 10% heat- inactivated FBS. Media for all cell types were supplemented with 100 units/mL penicillin-streptomycin and 2 mmol/L l-glutamine (Mediatch, Herndon, VA). Cells were

o grown at 37 C in a 5% CO2 humidified incubator. For culture in steroid-free conditions, cells were seeded in phenol red-free IMEM (LNCaP) or DMEM (22Rv1 and CV1) containing charcoal/dextran-treated fetal bovine serum (CDT; Hyclone Laboratories,

Logan, UT).

Colorimetric Yeast Screen for AR activity

Yeast strains with the genotype MATade2-1 leu2-3, 112trp1-1his3-11, 15can1-100ura3-

1 URA3 3xARE::pCYC1::ADE2 expressing distinct AR alleles were the generous gift of

Ralph W. deVere White (University of California at Davis School of Medicine,

Sacramento, CA) and were maintained and grown as previously described (Wetherill

CR05). Briefly, single yeast colonies were grown for 3 days at 30oC in 3mL minimal selective medium (Difco yeast nitrogen base without amino acids, Becton Dickinson,

Franklin Lakes, NJ) containing 0.5% adenine, 1%histidine, and 1% leucine and supplemented with argenine, valine, penylalanine, aspartic acid, isoleucine, serine, methionine, threonine, and glutamic acid. Yeast culture in log phase of growth was diluted into sterile PBS 1:2 and 3uL inoculated onto selection plates containing indicated concentration of test compound (either 10-8M DHT, 10-5M resveratrol, 10-5M coumestrol,

10-5M cadmium, 10-5M DDT or 10-5M DDE) and grown at 35oC for 3 days. Based on the color of the colony, the transactivational activity of AR was measured as previously described.

Chromatin Immunoprecipitation Assay (ChIP)

LNCaP or 22Rv1 cells (4x106) were seeded into poly-L-Lysine coated 15cm dishes in phenol red free IMEM containing 5% CDT. After 72 hours of incubation in serum free condition cells were treated with 0.1% EtOH or 10nM DHT or 10nM DDE. After three hours of treatment cells were fixed in 1% formaldehyde solution in PBS to allow for

DNA-protein cross-linking. Cross-linking process was stopped using 125mM glycine.

Cells were collected and treated with 750uL cell lysis buffer (5 mM PIPES [pH 8.0], 85 mM KCl, 0.5% NP-40, + 4ug benzamidine, 25ug 1,10-phenanthroline, 25ug aprotinin,

25ug leupeptine) for 10 min. 400uL nuclear lysis buffer was added and samples sonicated to shear DNA to 300-500bp fragments. Lysates were precleared with rotating incubation on Protein A beads for 24 hrs at 4OC. For immunoprecipitation, 10uL of lysate was used for input and 100uL of lysate was incubated with either 0.5ug AR antibody (N-20; Santa Cruz) or 0.5ug control IgG antibody (preimmune sera). The immunoprecipitation was carried out in 400uL RIPA buffer (150 mM NaCl, 1.0% NP-40,

0.5% deoxycholate, 0.1% SDS, 50mM Tris [pH 8.0]) for two hours at 4OC. At this point, protein-A Sepharose (Amersham) was added overnight at 4OC with rotation. Protein-A beads were washed with Super-RIPA (RIPA buffer plus 150mM NaCl), then thrice with

RIPA buffer and once with TE buffer (10mM Tris [pH 8.0], 1mM EDTA, with 10 minute roations between each wast at room temperature. To extract DNA, all samples

(including inputs) were incubated with 150uL of ChIP extraction buffer (1% SDS, 0.1M

NaHCO3) along with 10uL of 5M NaCl and Rnase A. This solultion was incubated at

65o overnight. DNA was purified using the QIAquick PCR purification kit (Qiagen). PCR amplification was performed on DNA recovered from the immunoprecipitated samples as well as input chromatin samples. For PCR amplification, the PSA promoter was amplified with previously published primers (9). The PCR conditions used were: 94oC for 5 min; followed by 35 cycles of 94oC for 30s, 49oC for 30s, and 72oC for 30s. The

PCR product obtained from these primers is nearly 200bp and was visualized using 2% agarose gel.

Reverse Transcriptase PCR

Cells were seeded onto 6cm plates with appropriate CDT media at approximately 4 – 5 x105 cells per plate. The following day, ethanol, DHT, or DDE was added for 24 hours.

Trizol Reagent was used to extract total RNA, of which 5 ug was used to generate cDNA with random hexamers using the ThermoScript RT-PCR system (Invitrogen,

Carlsbad, CA). Polymerase chain reactions were performed with the following primer sets: PSA forward: CTTGTAGCCTCTCGTGGCAC PSA reverse:

GACCTTCATAGCATCCGTGAG GAPDH forward: CCACCCCATGGCAAATTCCATGCA GAPDH reverse TCTAGACGGCAGGTCAGGTCCACC. Amplifications were performed using Taq

DNA polymerase with the following conditions: 940C 2 min followed by 25 cycles of

940C 30 sec, 540C 30 sec, and 720C 30 sec, with a final extension of 720C for 10 min. Immunoblotting

LNCaP cells (1x106) in 10-cm dishes were seeded in 5% CDT serum IMEM. After 24 hours, cells were supplemented with either vehicle (0.1% ethanol), 0.1nM DHT, or DDE.

Following 24 hours of treatment, cells were pelleted and whole cell lysates prepared.

Lysates were subjected to brief sonication and clarified by centrifugation. Equal protein concentrations (20ug) were loaded and subjected to SDS PAGE. Proteins were transferred to Immobilon membrane (Millipore, Corp., Bedford, MA) and immunoblotted for p-ERK (Santa Cruz, CA, P-ERK (E-4) Sc-7383, mouse monoclonal 1:250), total ERK

(Santa Cruz, CA, ERK 1 (K-23) SC-94 rabbit polyclonal 1:1000) and Lamin B (Santa

Cruz, CA, Lamin B (M-20), Sc-6217, goat polyclonal 1:1000). Goat anti-rabbit (Alexa

Fluor 680 A21076, 1:10,000, Molecular Probes, Eugene, OR; for detecting total ERK), donkey anti-Goat (Alexa fluor 680 A21084 1:10,000, Molecular Probes, Eugene, OR; for detecting Lamin B), and goat anti-mouse (1:10,000, Rockland 610-132-121,

Gilbertsville, PA: for detecting p-ERK) secondary antibody were used to visualize the antibody-antigen complex. Quantification of band intensity performed on Odysset IR imagines System (LI-COR Biosciences).

Bromodeoxyuridine incorporation assay

LNCaP and 22Rv1 cells were seeded in six well dishes on poly-L-lysine coated coverslips at a density of 2.5x105 cells per well in appropriate CDT media. Cells were then supplemented with either vehicle (0.1% ethanol), 0.1nm DHT, or various doses of

DDE. Following 72 hours of treatment, cells were labeled with cell proliferation labeling reagent (Amersham, Buchkinghamshire, UK, Cell Proliferation Labelling Reagent) according to manufacturer’s protocol. Labeling continued for 16 hours and cells were then processed to detect bromodeoxyuridine (BrdU) via indirect immunofluorescence

(as described, citation) Experiments were performed at least twice, each with duplicate independent biological replicates. At least 250 cells per experiment were counted for each condition. Averages and standard deviation are shown.

Cell growth Assay

LNCaP cells were seeded to approximately 3x105 cells per well in six-well dishes into appropriate medium containing CDT serum. Approximately 18 hours later, indicated concentrations of either ethanol (0.3%, vehicle control), DHT or DDE were added. Cells were cultured in designated conditions for indicated time points and fresh reagents added every 48 hours. After treatment, viable cells were counted using a hemacytometer and trypan blue exclusion. Cell doubling time (DT) was determined by

DT=t*ln2/ln (Ct-Co), where t=time interval in hours, Ct=final cell count and Co=initial cell count, as previously described (21).

Statistical Assessment

Quantitative results are expressed as average +/- standard deviation. Statistical analyses were performed using One-way ANOVA and Neuman Kuels post test. The criterion for statistical significance was p<0.05.

Results.

The pesticide p,p’DDE facilitates transcriptional activation of select tumor-derived

AR mutants.

It has been shown that the endocrine disrupting compound (EDC) bisphenol A

(BPA) is able to activate the tumor-derived AR mutant, AR-T877A, resulting in ligand independent receptor activation, induction of endogenous AR target gene transcription,

AR-T877A-dependent cellular proliferation, and subsequent therapeutic bypass for prostate cancer (18, 19). Given the impact of this EDC on mutant AR activity and resultant cellular proliferation, a more comprehensive screen was utilized to identify mutant ARs receptive to other endocrine disrupting compounds. For these studies, a yeast colorimetric screen for AR transactivity was employed (22). The yA(G)RE yeast strain (ade2 deficient) contains an integrated wt ADE2 sequence as a reporter gene under the control of an ARE-promoter. Yeast were transformed with plasmids that express functional ARs. Failure of AR to transactivate ADE2 results in the formation of red colonies, due to the accumulation of an intermediate red pigment in the adenine biosynthesis pathway. Upon AR activation by ligand and ADE2 transactivation, adenine biosynthesis is initiated and the red pigment is enzymatically reduced, resulting in the formation of white or pink yeast colonies, and the intensity of the color correlates directly to the extent of AR activation (22). Transactivational potential was assessed based on the yeast colony colorimetric score and results were reproduced in at least three independent experiments (Figure 1). This methodology provides a reliable and rapid method for screening a large number of test compounds on our selected AR mutants and wild type AR. Each compound was tested against the same set of AR proteins, which included one constitutively inactive mutant originally isolated from a patient with complete androgen insensitivity syndrome (C784Y) as a negative control, and one constituently active mutant (K580R) detected in advanced prostate cancer. The other four mutants were all human-prostate tumor derived, have mutations within the ligand binding domain, and have been utilized in this assay for transactivational activity demonstrating their responsiveness to estrogen (17-Β-estradiol) (T877A, H874Y, L701H and V715M).

A total of seven ARs (including wtAR) were screened for transactivational activity utilizing seven ligands. Results are depicted in Figure 1. Ethanol (EtOH) was utilized as the negative (vehicle) control and DHT served as the positive control, as the effect of this canonical ligand on each AR has already been established (22). As can be seen,

EtOH failed to activate the constitutively inactive C784Y, did not enhance the constitutively active K580R (positive control), and did not induce transactivation of wtAR, T877A, H874Y, L701H or V715M. Additionally, DHT mediated the known responses in the presence of each AR protein. There was no response to DHT by

C784Y and full activity demonstrated by both the constitutively active AR (K580R) as well as wt AR. As expected, DHT also induced robust activation of T877A, H874Y,

L701H and V715M, which maintain androgen responsiveness (22).

Through epidemiological studies it is clear that one hypothesis for the variation in the geographical incidence of prostate cancer may be due to the differences in consumption on phytoestrogens (23). Specifically, diets rich in soy products (such as

Asian countries) have a reduced prostate cancer related death rate compared to

Western men (24, 25). Additionally, data exists describing direct affects of select phytoestrogens on prostate cancer development, cellular proliferation and PSA secretion (26-28). Given the recent attention towards phytoestrogens as possible dietary supplements for chemoprevention and chemotherapeutic means of controlling prostate cancer (29, 30), the phytoestrogens and known estrogenic EDCs, resveratrol

(Res) and Coumestrol (Cou) were screened for their ability to activate AR. Neither phytoestrogen activated the negative control AR, or dramatically altered K580R transactivation potential, compared to vehicle control. Both compounds had no effect on wtAR, T877A, H874Y, L701H or V715M. Next, AR mutants were examined for their responsiveness to the metal, cadmium (Cd), as this compound is implicated in increased risk for prostate cancer development and reported to activated AR in prostate cancer cells (31-33). Surprisingly, Cd did not activate C784Y, nor alter the activity of

K580R, and had no effect on T877A, H874Y, L701H or V715M.

Lastly, the influence of the pesticide, p,p’DDT and its stable metabolite p,p’DDE was examined for mutant AR activity. There is strong epidemiological data suggesting that long-term exposure to specific organochlorines (such as DDT/DDE) may contribute to an increased risk of prostate cancer development (34-38). Furthermore, anti- androgenic properties of p,p’DDE on wtAR have been well characterized (39), with

-5 known binding kinetics (IC50 of 1.53x10 (40)). Therefore, we explored the capacity of mutant ARs to be activated by DDT and DDE in the yeast reporter system. As expected, these agents had no effect on C784Y, however, similar to DHT, they modestly enhanced AR-K580R activity. DDT and DDE also facilitated moderate activation of wtAR (weak activation suggested by pinkish-red colonies). This finding is not surprising given that p,p’-DDE has been shown to bind to wtAR (39, 40). Strikingly, the other tumor-derived mutant ARs also demonstrated robust transactivational activity in the presence of DDT and DDE. Specifically, V715M was moderately activated by DDE and

DDT (white pink colonies) and L701H was completely activated by DDT and DDE, as demonstrated by white colony formation. Additionally, H874Y was moderately activated by DDT and DDE, and T877A was consistently only partially activated by DDT and DDE as evident by pink colony formation. Representative images are shown in the bottom panel (Figure 1). These data reveal that in yeast, the pesticide DDT and its stable metabolite, DDE, are able to enhance the inherent transactivation function of K580R, and importantly, initiate wtAR, L701H, T877A, H874Y, and V715M activity in the absence of androgen. To determine if this effect is of importance for human risk to these compounds, the effects of these agents need to be tested under environmentally relevant conditions and in mammalian systems.

DDE effects on mutant AR activity in mammalian systems are dose dependent and molecular context specific.

Although the yeast system supported DHT-mediated activation of each receptor, it is well established that the cell-specific coregulatory milieu can dramatically alter the effect of ligand-mediated mutant AR activation (41, 42), necessitating the impact of DDE on endogenous mutant receptors in prostate cancer cells to be explored. Unfortunately, due to the paucity of prostate models expressing mutant ARs, the affects of DDE in prostate cancer cell lines expressing only two of the mutant ARs (AR-T877A and AR-

H874Y) could be elucidated. To examine endogenous receptors, 22Rv1 cells

(expressing AR-H874Y) and LNCaP cells (expressing AR-T877A) were analyzed for endogenous AR activation, monitored by rt-PCR analysis of PSA transcript levels. For these studies, cells were cultured in androgen-depleted media for 24 hours, then treated for 24 hours with EtOH, DHT 10-9M (positive control), or increasing doses of DDE (10-13,

10-11, and 10-8M) (Figure 2B). RNA was isolated from either 22Rv1 or LNCaP, and subjected to PCR with primers specific for GAPDH for internal control, and PSA. Shown in figure 2B, the PCR reactions were subjected to quantitative PCR, and relative

PSA/GAPDH shown. As expected and previously shown, PSA levels are sustained in

22Rv1 cells deprived of androgen, as these cells maintain AR-H874Y activity in the absence of agonists. However, basal AR-H874Y activity was further increased by DHT,

(approximately 2.5 fold, data not shown). Importantly, DDE exposure induced PSA transcription in a dose dependent manner in this cell system. Specifically, DDE at 10-

11M did not increase PSA mRNA expression, and although DDE 10-8M was consistently higher than EtOH control (average 1.2 fold), this increase did not reach statistical significance. However, DDE 10-5M did facilitate an increase in PSA mRNA approximately 2 fold over no ligand (EtOH) control. These data indicate that only high doses of DDE increase PSA expression in cells expressing AR-H874Y. To determine if this finding was recapitulated in prostate cancer cells, which express an endogenous

AR-T877A, PSA mRNA regulation by DDE was examined in the LNCaP model system.

As these cells are dependent upon AR agonist and AR-T877A action for PSA expression, in the absence of ligand (EtOH) little PSA mRNA was produced (Figure 2B, right panel, white bar). DHT dramatically enhanced the PSA mRNA levels

(approximately 36 fold induction over EtOH control, data not shown). Importantly, PSA mRNA was also regulated by DDE exposure. Interestingly, the DDE 10-8M dose produced the most significant enhancement of PSA mRNA (~1.5 fold over EtOH), while lower dose (DDE 10-11M) did not alter PSA mRNA levels and the highest dose, which mediated the most increased in PSA mRNA in 22Rv1 cells, consistently showed a reduction in PSA mRNA. These data indicate that DDE is able to induce PSA mRNA expression only at higher (10-5M) doses in cells expressing an endogenous AR-H874Y, however in cells with endogenous AR-T877A, high DDE doses inhibit PSA production, while DDE 10-8M enhances AR target gene expression. Combined, DDE is able to induce PSA transcription in cells expressing DDE-responsive mutants, however there is discrepancy in the required dose, dependent on cell type.

DDE induces AR-H874Y and AR-T877A recruitment to target gene regulatory regions.

Given that DDE facilitated increase in AR-target gene transcription in prostate cancer cells, and the ability of DDE to activate select tumor-derived AR mutants in yeast, these data suggest that DDE impinges on mutant AR to enhance target gene activation. To challenge this concept directly, the ability of DDE to induce AR recruitment to the PSA locus was monitored by standard chromatin immunoprecipitation assays (ChIP) in 22Rv1 and LNCaP prostate cancer cell lines. Briefly, either 22Rv1 or

LNCaP cells were seeded in the absence of hormone (CDT media) and 24 hours later stimulated with EtOH control, DHT 10-8M or DDE 10-6 or 10-8M for three hours. Cells were then harvested and processed for ChIP analysis. AR was immunoprecipitated, and primers for PSA utilized in the PCR.

As can be seen in Figure 3A, given the androgen-independent characteristic of the 22Rv1 cell line, there was residual AR-H874Y residency at the PSA enhancer region in the absence of ligand (EtOH, top lane, second column, IgG was negative control). However, DHT 10-8M treatment further enhanced AR-H874Y occupation in this regulatory region (second row). Importantly, DDE 10-6M was able to enhance recruitment of AR-H874Y to PSA enhancer regions, indicating that DDE exposure directly affects AR-H874Y function. We next analyzed the impact of DDE on AR-

T877A residency on PSA regulatory regions in chromatin (Figure 3B). LNCaP cells were utilized for these assays, and as can be seen in the absence of ligand (EtOH) there was no recruitment of AR-T877A to the PSA enhancer (compare top row, lane 2 to lane 3, IgG control). DHT 10-8M facilitated an increase in AR-T877A occupancy on

PSA regulatory regions. Treatment of these cells with DDE 10-8M recruited AR-T877A to PSA enhancer regions; therefore, DDE directly impacts AR-T877A function by inducing occupancy on target gene regulatory regions. Combined, these data indicate that DDE exposure increases AR target gene transcription by directly inducing AR residency on regulatory promoter regions.

DDE induces proliferation in AR-T877A expressing prostate cancer cells

Activation of the AR is known to promote proliferation in androgen dependent prostate cancer cells (43). Therefore, the ability of DDE to induce cell cycle progression was examined via bromodeoxyuridine (BrdU) incorporation. Briefly, cells were seeded onto glass coverslips into hormone depleted, CDT media. EtOH control, DHT 10-10M, or increasing doses of DDE (10-13, 1011, 108, or 10-5M) were added to the culture media 24 hours after seeding. Cells were treated with indicated ligand for 24 hours, and pulsed with BrdU for the last 12 hours of culture. The ability of DDE to induce cell cycle progression in AR-H874Y expressing 22Rv1 cells was first examined. Androgen- independent BrdU incorporation rates (EtOH control) were set to 1 so as to quantify the impact of DDE on proliferative indices. As expected, given the androgen-independent nature of this specific model system, DHT 10-10M did not significantly increased cell cycle progression (approximately 1.3 fold over EtOH control). Tested doses of DDE

(10-13, 10-11, 10-8, and 10-5M) also did not alter the amount of DNA synthesis in these cells compared to the proliferative rate in the absence of ligand (Figure 4A). Although, there was a consistent trend for increased BrdU incorporation with increasing doses, these results did not reach statistical significance. These data confirm the androgen independence of this model system, and suggest that despite that ability of DDE to impinge on AR-H874Y, DDE is unable to statistically enhance the cell cycle progression of these cells.

Contrary to 22Rv1s, LNCaP cells are dependent on ligand induced AR-T877A activity for cellular proliferation (44). As described above, androgen-independent BrdU incorporation rates (EtOH control) were set to 1 so as to quantify the impact of DDE on proliferative indices. As expected, given the androgen-dependent characteristics of this specific model system, DHT 10-10M significantly increased cell cycle progression to approximately 3 fold over EtOH control. Importantly, DDE exposure increased BrdU incorporation in a dose dependent manner, specifically demonstrating that lower doses had a significant increase in BrdU incorporation. As can be seen in Figure 4B, 10-5M did not increase BrdU incorporation compared to EtOH control (0.6 fold), however 10-

13M, 10-11M and 10-18M DDE facilitated a 2.2, 2.1 and 1.8 fold increase in DNA synthesis, respectively.

To verify that the DDE-mediated increase in DNA synthesis observed in Figure

4B with LNCaP cells actually resulted in increased cell number, LNCaP cells were treated with ligand (EtOH, DHT 10-10M, DDE 10-8, 10-11, 10-13M) for 24, 48, 72, 96 and

120 hours, and at each time point LNCaPs were collected and counted via trypan blue exclusion on a hemacytometer. As can be seen in Figure 4C, cells treated with EtOH failed to undergo cell doubling throughout the time course examined. Conversely, DHT increased cell number as expected. Congruent with the BrdU data, 10-5M DDE did not increase LNCaP cell number, however, there was a dose dependent affect of DDE 10-8,

10-11, and 10-13M, with the DDE 10-11 and 1013M doses both providing the paramount mitogenic stimuli. Combined, these data indicate that DDE is able to increase cell cycle progression and proliferation in AR-T877A dependent but not AR-H874Y expressing prostate cancer cells.

DDE activates MAPK in LNCaP cells.

Recent evidence has demonstrated that in select model systems, including prostate cancer cells, organochlorines are able to activate the MAPK pathway (45).

Additionally, AR activation through ligand-independent MAPK signaling cascades is an important mechanism in the development of hormone independent prostate cancers

(46). Given the known mode of AR action involving MAPK signaling and induction of proliferation, as well as the finding that in specific cell types, select organochlorines

(including DDT) can activate the MAPK pathway, the impact of DDE on the MAPK pathway in our androgen dependent prostate cancer cell line (LNCaP) was examined.

Figure 5 shows a representative immunoblot wherein LNCaP cells were seeded into

CDT media and challenged with either EtOH, DHT 10-10M, DDE 10-13 or DDE 10-11M 24 hours after initial seeding into androgen-depleted media. At the indicated time points, cells were harvested and total protein probed for phosphorylated Erk, a measure of

MAPK pathway activation. In EtOH treated LNCaPs, minimal p-Erk was observed (Figure 5, top lane, lines 1 and 4); however, as previously shown (citations) DHT exposure increased Erk phosphorylation after 40 minutes (lane 2). We also see that

DDE treatment causes induction of p-Erk at 40 minutes (lane 3). Strikingly, after two hours of exposure, DHT stimulated cells showed reduced p-Erk (lane 5); however DDE treated cells maintained ErK phosphorylation (lane 6). Total Erk levels were not altered by treatment (EtOH, DHT or DDE) and lamin B was used as the protein loading control.

These data demonstrate that DDE can activate the MAPK pathway in prostate cancer cells.

DDE utilizes MAPK and mutant AR pathways for induction of cell cycle progression in LNCaP cells

DDE exposure in LNCaP, AR-T877A expressing cells resulted in increased AR activity and dose-dependent proliferation. Additionally, it is known that activation of the

MAPK signaling pathway can lead to enhanced proliferation in prostate cancer cells (45,

46). Therefore, we next examined whether MAPK activation and AR activation were necessary for DDE-mediated LNCaP proliferation. For these sets of experiments,

LNCaP cells were seeded as described for Figure 4A, however prior to addition of ligand, either 10-6M Casodex (CSDX; direct AR antagonist) and/or 10-6M U0126 (MEK inhibitor) was added to the culture media. The last 12 hours of treatment, cells were pulsed with BrdU labeling agent to score cells with new DNA synthesis. Cells were fixed and coverslips processed for BrdU incorporation as described for Figure 4A. The amount of BrdU incorporation for each ligand alone was set to 100% to delineate the effect of each inhibitor on ligand-induced cell cycle progression. As can be seen, the direct AR antagonists, CSDX, did not alter cell cycle progression in the presence of

EtOH (striped bar). This is attributed to the already low BrdU incorporation in this cell line with inactive AR-T877A. However, inhibition of MEK via U0126 abolished residual

BrdU incorporation in the presence of EtOH (gray bar; 5% compared to androgen depletion alone). Co-administration of CSDX and U1026 also demonstrated very low

BrdU incorporation in the absence of androgen (black bar; 3%). As expected given the known requirement for DHT-mediated AR-T877A activation to result in proliferation, the direct AR antagonists, CSDX, reduced DHT mediated BrdU incorporation by 60%

(striped bar, 37% compared to DHT alone). Interestingly, the MEK inhibitor also demonstrated a similar level of BrdU decrease (30% compared to DHT alone, no statistically difference between singular treatments of CSDX and U0126, p>0.05) and the combination of the two inhibitors resulted in even further reduction in BrdU to only

9% of cell incorporating BrdU in the presence of DHT with both CSDX and U0126.

Lastly, the contribution of AR-T877A and MAPK pathway to DDE mediated BrdU incorporation was examined. CSDX did DDE 10-13M BrdU (80%, p<0.05), and also modestly reduced DDE 10-11 BrdU incorporation from 100% to 71%, (p<0.05).

Therefore, it appears that AR-T877A is only partially responsible for DDE mediated cell cycle progression. In contrast, administration of the MEK inhibitor, U1026 drastically reduced DDE-mediated BrdU incorporation (5% BrdU incorporation for both DDE 10-13 and 10-11M, p<0.001), indicating that this pathway contributes to the mitogenic action of

DDE in this cell type. The combination of CSDX and U1026 was not statistically different that of U1026 alone (p>0.05), indicating that inhibiting the MAPK pathway is paramount to blocking AR-T877A-mediated proliferation induced by low level DDE. Representative images of cells labeled with BrdU are shown in Figure 6, bottom panel.

These data demonstrate that DDE may be acting predominantly through the MAPK pathway to initiate cell cycle progression in prostate cancer cells, however the cross-talk between MAPK signaling and AR remains poorly understood.

Discussion.

The studies herein demonstrate that p,p’DDE is able to activate select mutant

ARs that often arise during prostate cancer progression in a yeast model of AR transactivational activity. In prostate cancer model systems expressing the endogenous

AR mutants, AR-H874Y or AR-T877A, DDE exposure induced PSA mRNA increase, as well as AR residency on the PSA promoter, suggesting that in prostatic model systems,

DDE can activate endogenous AR mutants. However, androgen-independent prostate cancer cells expressing AR-H874Y were not sensitive to DDE for proliferation, although this specific AR mutant was activated by DDE for target gene regulation. When AR-

T877A expressing androgen-dependent prostate cancer cells were exposed to low levels of DDE, cells progressed through cell cycle in the absence of androgen. In these cells, DDE mediated proliferation was facilitated through both AR activation, as well as

MAPK pathway activation, however seemed to be more dependant upon MAPK activation, since blocking MEK had a more severe inhibition on proliferation than hindering AR action through Casodex. Combined, these data indicate that exposure

DDE may cause prostate tumor bypass of androgen deprivation therapy through inappropriate proliferation, only in cells that are androgen-dependent and harbor a DDE- responsive mutant AR. The data herein detail a novel mode of EDC impact on prostate cancer therapy:

MAPK signaling pathway activation by Erk phosphorylation. In prostate cancer cells, it has been shown that Erk is stimulated by growth factors and cytokines to mediate proliferation (47), additionally regulation of Erk (via phosphorylation status) has been implicated to correlate with prostate cancer disease state and grade, wherein inactivation of Erk coincided with increased Gleason score and differentiation (48).

DDE mediated proliferation in prostate cancer cells was slightly attenuated by inhibition of AR function, however, completely blocked by inhibition of MAPK signaling; however,

DDE directly acted upon mutant AR by facilitating AR recruitment to target gene (PSA) regulatory regions and increasing AR-dependent PSA transcription. It is unknown how

MAPK pathways and AR activation cooperate given a specific cellular and promoter context, however the data described herein indicate an clear distinction. Therefore, these data posse an interesting paradigm illustrated in Figure 7, wherein DDE-mediated prostate cancer proliferation may be directly through AR ligand binding and AR transactivation; and/or indirectly through phosphorylation cascades involving MAPK signaling pathways. However, it is undetermined what the effects of MAPK pathways are on AR directly, or what the role of MAPK is in AR phosphorylation and activation status. As it is known that MAPK signaling pathways are important processes in prostate cancer disease progression (46), it is imperative to determine the relative contribution of both AR activation as well as MAPK activation in proliferation and therapeutic bypass by EDCs.

Given the disparate results comparing DDE-mediated AR activity in the yeast reporter system (activation of wtAR and all tumor-derived AR mutants tested), DDE- induced AR-dependent target gene transcription (in prostate cancer cells harboring AR-

H874Y and AR-T877A), and DDE-facilitated proliferation (only in cells dependant upon

AR-T877A for proliferation), it is of interest to explore how specific AR mutations dictates DDE-mediated AR activation and proliferation through exploration of coregulatory molecule usage, as well as target gene profile differences. The ability of

DDE to bind to wtAR has been well documented (39). Most recently, it has been shown

-5 that p,p’DDE binds to recombinant rat wt AR with an IC50 of 1.53x10 (40). p,p’DDE has also been shown to have dose dependent competitive inhibition of DHT (linear slope plots) in competitive AR binding assays (39). Therefore, our finding that DDE was able to activate wtAR in yeast is not surprising. However, it is known that AR utilizes numerous coactivator molecules to mediate its transactivational properties (9, 41) and that ligands confer specificity to target genes profiles and biological endpoints by influencing the coregulatory milieu (49). It is likely that the requirement of coactivator molecules for AR specificity underlies the antagonist action of DDE on wtAR in prostate cancer cell model systems.

p,p’DDE is a stable breakdown product of DDT, and is of particular interest due to its persistence in the environment and its continued detection in human tissues (50-

52). Additionally, there is a plethora of epidemiological evidence linking DDT/DDE and chlorinated pesticides, in general, to adverse affects on male reproductive organs and fertility in both wildlife and human populations (51). Globally, between 50 and 99% of human serum analysis have detected DDE (20). As such, exposure to DDE by prostate cancer patients is likely and of concern; however this is the first study, to date, to address how DDE may impact a very specified molecular context, potentially adversely affecting prostate cancer therapy and progression. The data herein demonstrate a distinct and disparate mode of action for p,p’DDE, than has previously shown for the

EDC bisphenol A, as this compound is also able to activate AR, inducing AR-dependent gene transcription and proliferation in the absence of androgen. However, in the absence of androgen, DDE is able to potentiate proliferation of prostate cancer cells expressing the DDE-responsive AR-T877A, through AR- and MAPK-dependent pathways. This unique mode of action for DDE highlights the compound specific actions of EDCs and indicates that the direct AR antagonists, Casodex, is unable to completely block the mitogenic action of DDE. These data highlight the need for further investigation into relative contribution of AR versus MAPK signaling towards prostate cancer cell proliferation.

The impact of DDE on prostate cancer has been previously examined under higher concentrations and only within the context of a wtAR. Kelce WR et al found that

DDE is a potent androgen receptor antagonist, as higher doses (uM) inhibit androgen binding to the AR, inhibit androgen-induced transcriptional activity, and inhibit androgen action in developing, pubertal and adult male rats (39). The present student found that the most pronounced affects of DDE occur at the very low, nanamolar concentrations.

These concentrations fall within the known human exposure range for these compounds

(between 6.6 and 19.36 ug/L in plasma samples; (50). Although it is not without precedent for low doses of ligand to bind to and activate steroid nuclear receptors, one possible mechanism of low dose affects of DDE in LNCaP cells is that recently DDE has been found to increase aromatase activity in endometrial stromal cells in culture, resulting in locally higher estrogen (53). As LNCaP cells express an estrogen responsive AR (AR-T877A), this increase in local estrogen may drive AR activity and

AR dependent proliferation. Aromatase expression is regulated through several different promoter regions in a tissue-specific manner (54), so it is yet to be determined if and how DDE may be regulating aromatase activity in LNCaP cells.

The emerging conclusion on the impact of p,p’DDE on prostate cancer progression and therapy is that the action of this EDC is regulated by a complex interaction of genomic signaling through responsive AR mutants, and non-genomic signaling through the MAPK pathway. Whether the exposure to environmentally relevant levels of DDE may negatively affect prostate cancer patients undergoing either androgen deprivation therapy or combined androgen blockade (ADT plus Casodex) is yet to be determined and will be the focus of future studies.

References.

1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55(1):10-30. 2. Klotz L. Hormone therapy for patients with prostate carcinoma. Cancer 2000;88(12 Suppl):3009-14. 3. Sharifi N, Farrar WL. Androgen receptor as a therapeutic target for androgen independent prostate cancer. Am J Ther 2006;13(2):166-70. 4. Damber JE, Khatami A. Surgical treatment of localized prostate cancer. Acta Oncol 2005;44(6):599-604. 5. Kirkpatrick JP, Anscher MS. Radiotherapy for locally recurrent prostate cancer. Clin Adv Hematol Oncol 2005;3(12):933-42. 6. Feldman BJ, Feldman D. The development of androgen-independent prostate cancer. Nat Rev Cancer 2001;1(1):34-45. 7. Trapman J, Brinkmann AO. The androgen receptor in prostate cancer. Pathol Res Pract 1996;192(7):752-60. 8. Russell DW, Wilson JD. Steroid 5 alpha-reductase: two genes/two enzymes. Annu Rev Biochem 1994;63:25-61. 9. Shang Y, Myers M, Brown M. Formation of the androgen receptor transcription complex. Mol Cell 2002;9(3):601-10. 10. Stephan C, Jung K, Diamandis EP, Rittenhouse HG, Lein M, Loening SA. Prostate- specific antigen, its molecular forms, and other kallikrein markers for detection of prostate cancer. Urology 2002;59(1):2-8. 11. Dehm SM, Tindall DJ. Molecular regulation of androgen action in prostate cancer. J Cell Biochem 2006. 12. Mendelsohn LG. Prostate cancer and the androgen receptor: strategies for the development of novel therapeutics. Prog Drug Res 2000;55:213-33. 13. Hirawat S, Budman DR, Kreis W. The androgen receptor: structure, mutations, and antiandrogens. Cancer Invest 2003;21(3):400-17. 14. Denmeade SR, Isaacs JT. A history of prostate cancer treatment. Nat Rev Cancer 2002;2(5):389-96. 15. Taplin ME, Bubley GJ, Shuster TD, et al. Mutation of the androgen-receptor gene in metastatic androgen-independent prostate cancer. N Engl J Med 1995;332(21):1393-8. 16. Taplin ME, Bubley GJ, Ko YJ, et al. Selection for androgen receptor mutations in prostate cancers treated with androgen antagonist. Cancer Res 1999;59(11):2511-5. 17. Tilley WD, Buchanan G, Hickey TE, Bentel JM. Mutations in the androgen receptor gene are associated with progression of human prostate cancer to androgen independence. Clin Cancer Res 1996;2(2):277-85. 18. Wetherill YB, Fisher NL, Staubach A, Danielsen M, de Vere White RW, Knudsen KE. Xenoestrogen action in prostate cancer: pleiotropic effects dependent on androgen receptor status. Cancer Res 2005;65(1):54-65. 19. Wetherill YB, Petre CE, Monk KR, Puga A, Knudsen KE. The xenoestrogen bisphenol A induces inappropriate androgen receptor activation and mitogenesis in prostatic adenocarcinoma cells. Mol Cancer Ther 2002;1(7):515-24. 20. Charlier C, Desaive C, Plomteux G. Human exposure to endocrine disrupters: consequences of gastroplasty on plasma concentration of toxic pollutants. Int J Obes Relat Metab Disord 2002;26(11):1465-8. 21. Hess-Wilson JK, Boldison J, Weaver KE, Knudsen KE. Xenoestrogen action in breast cancer: impact on ER-dependent transcription and mitogenesis. Breast Cancer Res Treat 2006;96(3):279-92. 22. Shi XB, Ma AH, Xia L, Kung HJ, de Vere White RW. Functional analysis of 44 mutant androgen receptors from human prostate cancer. Cancer Res 2002;62(5):1496-502. 23. Adlercreutz H. Phyto-oestrogens and cancer. Lancet Oncol 2002;3(6):364-73. 24. Clinton SK, Giovannucci E. Diet, nutrition, and prostate cancer. Annu Rev Nutr 1998;18:413-40. 25. Cohen JH, Kristal AR, Stanford JL. Fruit and vegetable intakes and prostate cancer risk. J Natl Cancer Inst 2000;92(1):61-8. 26. Mentor-Marcel R, Lamartiniere CA, Eltoum IE, Greenberg NM, Elgavish A. Genistein in the diet reduces the incidence of poorly differentiated prostatic adenocarcinoma in transgenic mice (TRAMP). Cancer Res 2001;61(18):6777-82. 27. Landstrom M, Zhang JX, Hallmans G, et al. Inhibitory effects of soy and rye diets on the development of Dunning R3327 prostate adenocarcinoma in rats. Prostate 1998;36(3):151-61. 28. Pollard M. Prevention of prostate-related cancers in Lobund-Wistar rats. Prostate 1999;39(4):305-9. 29. Cos P, De Bruyne T, Apers S, Vanden Berghe D, Pieters L, Vlietinck AJ. Phytoestrogens: recent developments. Planta Med 2003;69(7):589-99. 30. Stephens FO. Phytoestrogens and prostate cancer: possible preventive role. Med J Aust 1997;167(3):138-40. 31. Sahmoun AE, Case LD, Jackson SA, Schwartz GG. Cadmium and prostate cancer: a critical epidemiologic analysis. Cancer Invest 2005;23(3):256-63. 32. Achanzar WE, Diwan BA, Liu J, Quader ST, Webber MM, Waalkes MP. Cadmium- induced malignant transformation of human prostate epithelial cells. Cancer Res 2001;61(2):455- 8. 33. Verougstraete V, Lison D, Hotz P. Cadmium, lung and prostate cancer: a systematic review of recent epidemiological data. J Toxicol Environ Health B Crit Rev 2003;6(3):227-55. 34. Blair A, Zahm SH. Agricultural exposures and cancer. Environ Health Perspect 1995;103 Suppl 8:205-8. 35. Fleming LE, Bean JA, Rudolph M, Hamilton K. Cancer incidence in a cohort of licensed pesticide applicators in Florida. J Occup Environ Med 1999;41(4):279-88. 36. Mills PK, Beeson WL, Phillips RL, Fraser GE. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer 1989;64(3):598-604. 37. Ritchie JM, Vial SL, Fuortes LJ, Guo H, Reedy VE, Smith EM. Organochlorines and risk of prostate cancer. J Occup Environ Med 2003;45(7):692-702. 38. Schreinemachers DM, Creason JP, Garry VF. Cancer mortality in agricultural regions of Minnesota. Environ Health Perspect 1999;107(3):205-11. 39. Kelce WR, Stone CR, Laws SC, Gray LE, Kemppainen JA, Wilson EM. Persistent DDT metabolite p,p'-DDE is a potent androgen receptor antagonist. Nature 1995;375(6532):581-5. 40. Fang H, Tong W, Branham WS, et al. Study of 202 natural, synthetic, and environmental chemicals for binding to the androgen receptor. Chem Res Toxicol 2003;16(10):1338-58. 41. Jenster G. The role of the androgen receptor in the development and progression of prostate cancer. Semin Oncol 1999;26(4):407-21. 42. Culig Z, Comuzzi B, Steiner H, Bartsch G, Hobisch A. Expression and function of androgen receptor coactivators in prostate cancer. J Steroid Biochem Mol Biol 2004;92(4):265- 71. 43. Knudsen KE, Arden KC, Cavenee WK. Multiple G1 regulatory elements control the androgen-dependent proliferation of prostatic carcinoma cells. J Biol Chem 1998;273(32):20213- 22. 44. Kim IY, Kim JH, Zelner DJ, Ahn HJ, Sensibar JA, Lee C. Transforming growth factor- beta1 is a mediator of androgen-regulated growth arrest in an androgen-responsive prostatic cancer cell line, LNCaP. Endocrinology 1996;137(3):991-9. 45. Tessier DM, Matsumura F. Increased ErbB-2 tyrosine kinase activity, MAPK phosphorylation, and cell proliferation in the prostate cancer cell line LNCaP following treatment by select pesticides. Toxicol Sci 2001;60(1):38-43. 46. Craft N, Shostak Y, Carey M, Sawyers CL. A mechanism for hormone-independent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase. Nat Med 1999;5(3):280-5. 47. Ghosh PM, Malik S, Bedolla R, Kreisberg JI. Akt in prostate cancer: possible role in androgen-independence. Curr Drug Metab 2003;4(6):487-96. 48. Malik SN, Brattain M, Ghosh PM, et al. Immunohistochemical demonstration of phospho-Akt in high Gleason grade prostate cancer. Clin Cancer Res 2002;8(4):1168-71. 49. Smith CL, O'Malley BW. Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev 2004;25(1):45-71. 50. Pant N, Mathur N, Banerjee AK, Srivastava SP, Saxena DK. Correlation of chlorinated pesticides concentration in semen with seminal vesicle and prostatic markers. Reprod Toxicol 2004;19(2):209-14. 51. Turusov V, Rakitsky V, Tomatis L. Dichlorodiphenyltrichloroethane (DDT): ubiquity, persistence, and risks. Environ Health Perspect 2002;110(2):125-8. 52. Younglai EV, Holloway AC, Lim GE, Foster WG. Synergistic effects between FSH and 1,1-dichloro-2,2-bis(P-chlorophenyl)ethylene (P,P'-DDE) on human granulosa cell aromatase activity. Hum Reprod 2004;19(5):1089-93. 53. Holloway AC, Stys KA, Foster WG. DDE-induced changes in aromatase activity in endometrial stromal cells in culture. Endocrine 2005;27(1):45-50. 54. Simpson ER, Zhao Y, Agarwal VR, et al. Aromatase expression in health and disease. Recent Prog Horm Res 1997;52:185-213; discussion -4.

Figure Legends.

Figure 1. The pesticide p,p’DDE facilitates transcriptional activation of select tumor-derived AR mutants. Yeast strains containing the ARE-driven ADE2 reporter gene and expressing individual AR mutants were grown on selective medium plates as described in Materials and Methods, supplements with either vehicle (o.1% ethanol), 10- 8M DHT, 10-5M resveratrol (Res), 10-5M coumestrol (Cou), 10-5M cadmium (Cd), 10-5M p,p’DDE, or 10-5M p,p’DDT. AR-mediated transactivational activity was scored based on the color of the yeast colonies. White, complete activation of AR; pink, weak AR activation; red, no AR activation. Experiments were performed in three biological replicates with representative images shown in bottom panel.

Figure 2. DDE effects on mutant AR activity in mammalian systems are dose dependent and molecular context specific. 22Rv1 (left panel) or LNCaP (right panel) cells were treated with either vehicle control (0.1% ethanol) or increasing doses of DDE (10-11, 10-8, or 10-5M) for 24 hours and RNA harvested and used for cDNA synthesis followed by quantitative real time PCR.

Figure 3. DDE induces AR-H874Y and AR-T877A recruitment to target gene regulatory regions. A. 22Rv1 cells were treated with either vehicle control (0.1% ethanol), 10-8M DHT, or 10-6M DDE for 24 hours prior to Chromatine Immunoprecipitation Assay (ChIP) B. LNCaP cells were treated as in A, with either vehicle control (0.1% ethanol, DHT 10-8M, or DDE 10-8M.

Figure 4. DDE induces proliferation in AR-T877A expressing prostate cancer cells. 22Rv1 cells (A.) or LNCaP cells (B.) were cultured in the absence of steroid for 24 hours, then either vehicle control (0.1% ethanol), DHT 10-10M, or increasing doses of DDE (10-13, 10-11, 10-8, or 10-5M) for 24 hours. Cells were labeled with BrdU for 12 hours and then BrdU incorporation was monitored via indirect immunofluorescence. Level of BrdU incorporation in the absence of ligand (vehicle control) for each cell type was set to 1 (dotted line). C. LNCaP cells were cultured in the absence of steroid for 24 hours prior to addition of ligand (ethanol vehicle control, DHT 10-10M, or DDE (10-13, 10-11, 10-8, or 10-5M) for indicated time periods (24, 48, 72, 96, or 120 hours) with fresh ligand added every 48 hours. At each time point, cells were collected and cell number and viability determined via trypan blue exclusion. Experiment repeated with three biological replicates performed in triplicate.

Figure 5. DDE activates MAPK in LNCaP cells. LNCaP cells were cultured in the absence of steroid for 24 hours, then, either vehicle control (0.1% ethanol), 10-9M DHT, or 10-13M DDE added. Following 40 or 120 minutes, cell lysates were collected and subjected to SDS-PAGE followed by immunoblotting for total Erk 1/2, pp-Erk 1/2, or lamin B (loading control). Numbers correspond to fluorescence intensity of pp-Erk relative to total Erk in the same conditions.

Figure 6. DDE utilizes MAPK and mutant AR pathways for induction of cell cycle progression in LNCaP cells. LNCaP cells were cultured in the absence of steroid plus either no inhibitory treatment (0.1% DMSO vehicle control; open bars), 10-6M Casodex (CSDX; black striped bars), 10-6M MEK inhibitor (U0126; gray bars), or combined 10-6M Casodex and 10-6M U0126 (black bars). Following 24 hours of initial treatment, cells were stimulated with either ethanol vehicle control, DHT 10-10M or DDE (10-13 and 10- 11M) for 24 hours and labeled with BrdU labeling agent for the last 12 hours of treatment. BrdU incorporation was detected via indirect immunofluorescence. BrdU incorporation for each ligand in the absence of inhibitory challenge was set to 100%. Experiment was performed with two biological replicates, in triplicate. ** p<0.01, *p<0.05

Figure 7. Schematic illustrating the interaction between DDE - AR, DDE - MAPK and MAPK - AR. Studies herein have shown that DDE acts in part through mutant AR, and predominately through MAPK activation to facilitate prostate cancer cell proliferation. However, the interaction between MAPK and mutant AR is not well understood.

Figure 1. The pesticide p,p’DDE facilitates transcriptional activation of select tumor- derived AR mutants

C784Y K580R WtAR T877A H874Y L701H V715M EtOH - +++ - - - - - DHT 10-8 - ++++ ++++ ++++ ++++ ++++ ++++ Res 10-5 - ++ - - - - - Cou 10-5 - ++ - - - - - CD 10-5 - ++ - - - - - DDE 10-5 - ++++ + ++ +++ ++++ +++ DDT 10-5 - ++++ ++ ++ +++ ++++ +++

EtOH White/++++ = complete activation of AR DHT 10-8 White-pink /+++ = moderate activation of AR Pink/++ = partial activation of AR -5 DDE 10 Pink-Red/+ = weak activation of AR DDT 10-5 Red/- = No activation of AR

EtOH (ethanol), DHT (dihydrotestosterone), DDE (dichlorodiphenyldichloroethylene), DDT (dichlorordiphenyldicholorethalene), Res (Resveretrol), Cou (Coumesterol), CD (cadmium) Figure 2. DDE effects on mutant AR activity in mammalian systems are dose dependent and molecular context specific.

22Rv1 LNCaP 2.5 2.0 2.0 1.6 1.5 1.2 Q-PCR 1.0 Q-PCR 0.8 0.5 Relative PSA/GAPDH Relative PSA/GAPDH 0.4 0.0 0.0 -11 -8 -6 EtOH 10-11 10-8 10-5 EtOH 10 10 10 DDE DDE Figure 3. DDE induces AR-H874Y and AR-T877A recruitment to target gene regulatory regions

A. 22Rv1 B. LNCaP IP: IP:

Input AR IgG Input AR IgG EtOH EtOH

DHT DHT

DDE DDE

1 2 3 1 2 3 Figure 4. DDE induces proliferation in AR-T877A expressing prostate cancer cells.

A. B. 4 22Rv1 4 LNCaP 3.5 3.5 3 3 * * BrdU BrdU 2.5 2.5 * in 2 in 2 Δ Δ 1.5 1.5 1

Fold 1 Fold 0.5 0.5 0 0 EtOH DHT 10-13 10-11 10-8 10-5 EtOH DHT 10-13 10-11 10-8 10-5 10-10 10-10 DDE DDE

C.

80 LNCaP

4 70 DHT 10-10 60 DDE 10-11 50 DDE 10-13 40 DDE 10-8

30 EtOH

Cell Number x10 -5 20 DDE 10 10 0 24 48 72 96 120 Hours Figure 5. DDE activates MAPK in LNCaP cells

40 minutes 120 minutes

EtOH DHT DDE EtOH DHT DDE 10-9 10-13 10-9 10-13

pp-Erk

Erk 1/2

Lamin B 1 2 3 4 5 6 Figure 6. DDE utilizes MAPK and mutant AR pathways for induction of cell cycle progression in LNCaP cells.

* 140 * no treatment 120 CSDX

100 U0126

BrdU combined 80 ** 60 *

% Change 40

20 (relative to untreated control)

0 EtOH DHT 10-10 10-13 10-11 DDE Figure 7. Schematic depicting links between DDE -AR -MAPK and MAPK-AR in prostate cancer cells

DDE

MEK inhibitor ? MAPK AR ?

AR antagonist

U0126 PROLIFERATION CSDX

Chapter IV:

Mitogenic action of the androgen receptor sensitizes prostate cancer cells to taxane-based cytotoxic insults

Research Article

Mitogenic Action of the Androgen Receptor Sensitizes Prostate Cancer Cells to Taxane-Based Cytotoxic Insult

Janet K. Hess-Wilson,1 Hannah K. Daly,1 William A. Zagorski,2 Christopher P. Montville,3 and Karen E. Knudsen1,4,5

Departments of 1Cell Biology, 2Surgery, and 3Obstetrics and Gynecology, 4Center for Environmental Genetics, and 5University of Cincinnati Cancer Center, University of Cincinnati College of Medicine, Cincinnati, Ohio

Abstract in prostatic epithelia or adenocarcinoma cells (4). DHT binding stimulates displacement of heat shock proteins from AR, receptor Prostate cancer cells are dependent on androgen for growth and survival; as such, inhibition of androgen receptor (AR) dimerization, and rapid translocation of AR into the nucleus. activity is the first line of intervention for disseminated Activated AR associates with specific DNA sequences, termed disease. Recently, specific cytotoxic agents have been shown to androgen-responsive elements, and subsequently recruits coacti- extend survival times in patients with advanced disease. Given vators to initiate target gene transcription (2, 5). Through these the established ability of androgen to modify cell survival in events, androgen elicits numerous biological outcomes dependent prostate cancer cells, it is imperative to determine the effect on cellular context, including proliferation, survival, and differen- of the hormonal environment on cytotoxic response. Here, tiation (6). we show that the response of prostate cancer cells to taxane- To exploit the dependence of prostate cancer on AR function, induced cell death is significantly enhanced by androgen androgen deprivation therapy is implemented by either surgical (bilateral orchiectomy) or pharmacologic (gonadotropin-releasing stimulation in AR-positive, androgen-dependent prostate cancer cells. Similar results were observed on androgen- hormone agonists) methodologies (3, 5, 7). These therapies are independent AR activation. By contrast, AR-positive yet initially effective and induce both cell cycle arrest and apoptosis in androgen-independent or AR-negative cells were refractory tumor cells (3). However, recurrent tumors ultimately arise wherein to androgen influence on taxane function. The ability of AR activity has been restored (5, 7–9). Until recently, no therapeutic androgen to potentiate taxane activity was dependent on its strategy had been identified that yielded a significant survival mitogenic capacity and was separable from overall AR activity, advantage for patients with recurrent prostate cancer (7). Recently completed clinical trials showed that microtubule-stabilizing as coadministration of AR antagonists, G1 cyclin-dependent kinase inhibitors, or high-dose (growth inhibitory) androgen agents (e.g., taxanes) improve clinical outcome in recurrent disease nullified the proapoptotic function of androgen. Observed (10). Although encouraging, the benefits were relatively modest. As induction of cell death was attributed to caspase-dependent such, recent attention has been directed toward the optimization apoptosis and correlated with p53 activation. Combined, these of treatment strategies using these cytotoxic agents (10). data indicate that the cytotoxic effects of taxanes are sub- Given that androgens and the AR play significant roles in regulation of proliferation and apoptosis in prostatic epithelium stantially influenced by the hormonal environment and/or status of AR activity in prostate cancer cells and provide the (11–14), we investigated the role of AR in the response to cytotoxic foundation for refinement and optimization of cytotoxic inter- insult induced by taxanes. Our data show that in androgen- dependent prostate cancer cells, AR activation synergizes with vention in prostate cancer. (Cancer Res 2006; 66(24): 11998-2008) paclitaxel to enhance cell death. This function of AR is exquisitely dependent on its mitogenic capacity as shown through multiple Introduction analyses. By contrast, efficacy of paclitaxel was severely diminished Prostate cancer is the most commonly diagnosed malignancy when AR activity was nullified or under conditions of forced cell and the second leading cause of cancer-related deaths among U.S. cycle arrest. Similarly, in AR-deficient cells, or AR-proficient but men (1). Locally confined tumors are treated by radical prostatec- hormone-independent cells, the response of paclitaxel was tomy or radiation therapies; however, treatment for disseminated refractory to the hormonal milieu. Combined, these data indicate disease remains a major clinical challenge. Conventional therapy that the cytotoxic effects of paclitaxel are dependent on the for metastatic disease is reliant on the androgen dependence of mitogenic function of AR and provide the foundation for further prostatic adenocarcinomas, as this tumor type requires androgen refinement of combinatorial therapy for prostate cancer. for growth and survival (2, 3). Androgens mediate their action through activating the androgen Materials and Methods receptor (AR), a ligand-dependent transcription factor. The most prevalent AR ligand in serum is testosterone, which is converted Reagents. DHT, 17h-estradiol (E2), , bisphenol A (BPA; ¶ through the action of 5-a-reductase to dihydrotestosterone (DHT) 4,4 -isopropylidenediphenol), paclitaxel, and docetaxel were purchased from Sigma Chemical Co. (St. Louis, MO). Casodex (bicalutamide) was a generous gift from AstraZeneca Pharmaceuticals (London, United Kingdom). Cholesterol was dissolved in to 1 mol/L and À Requests for reprints: Karen E. Knudsen, Department of Cell Biology, Vontz then to 10 2 mol/L in 100% ethanol (EtOH) for storage at À20jC. À Center for Molecular Studies, University of Cincinnati College of Medicine, 3125Eden Paclitaxel and docetaxel were dissolved in DMSO to 10 2 mol/L. All À Avenue, ML 0521, Cincinnati, OH 45267-0521. Phone: 513-558-7371; Fax: 513-558-4454; other compounds were reconstituted to 10 2 mol/L in EtOH. E-mail: [email protected]. h I2006 American Association for Cancer Research. Recombinant human heregulin 1 protein (HRG) was purchased from doi:10.1158/0008-5472.CAN-06-2249 NeoMarkers (Fremont, CA) and stored at À20jC. Ac-VAD-CHO caspase

Cancer Res 2006; 66: (24). December 15, 2006 11998 www.aacrjournals.org Androgen Receptor Action Synergizes with Taxanes inhibitor, roscovitine, and aphidicolin were purchased from Calbiochem Reverse transcription-PCR. 22Rv1 cells were seeded in 6-cm dishes in À (La Jolla, CA), reconstituted in DMSO, and used at the indicated either 10% CDT, 10% FBS, or 10% CDT plus 10 10 mol/L DHT. After concentrations. The following antibodies were used: rabbit polyclonal 48 hours, cells were harvested and total RNA was isolated via Trizol reagent poly(ADP-ribose) polymerase (PARP) and Ser15 p53 phosphorylated (Life Technologies, Gaithersburg, MD) as recommended by the manufac- specific (Cell Signaling, Danvers, MA), Bax (N-20) and cyclin-dependent turer. Reverse transcription-PCR was done followed by PCR of the cDNA kinase (CDK) 4 (H-22; Santa Cruz Biotechnology, Santa Cruz, CA), E2F-1 using primers for prostate-specific antigen (PSA; primer pair: 5¶- (Ab-3; NeoMarkers), p53 (Ab-6; Calbiochem), and bcl-2 and Hsp27 CTTGTAGCCTCTCGTGGCAG-3¶ and 5¶-GACCTTCATAGCATCCGTGAG-3¶) (Stressgen, Ann Arbor, MI). and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; loading control; Cell culture and treatment. PC-3 and LNCaP cells were obtained from primer pair: 5¶-CCACCCATGGCAAATTCCATGGCA-3¶ and 5¶-TCTAGACGG- the American Type Culture Collection (Rockville, MD) and used between CAGGTCAGGTCCACC-3¶). PCR conditions were as follows: PSA, 94jC for 2 passages 30 and 45. 22Rv1 cells were the generous gift of Dr. J. Jacobberger minutes; 35cycles of 94 jC for 30 seconds, 54jC for 30 seconds, and 72jCfor (Case Western Reserve University, Cleveland, OH). LNCaP cells were 30 seconds; and 72jC for 10 minutes; GAPDH, 94jC for 2 minutes; 25cycles maintained in Iscove’s modified Eagle’s medium (IMEM; Cellgro, Mediatech, of 94jC for 30 seconds, 57jC for 30 seconds, and 72jC for 30 seconds; and Herndon, VA) containing 5% heat-inactivated fetal bovine serum (FBS; 72jC for 10 minutes. Biofluids, Rockville, MD). PC-3 and 22Rv1 cells were maintained in DMEM Flow cytometry. LNCaP cells were seeded in 6-cm dishes in either 5% supplemented with 10% heat-inactivated FBS. For growth in steroid-free CDT or FBS medium. After f24 hours, indicated concentrations of conditions, cells were seeded in phenol red–free IMEM (LNCaP) or DMEM aphidicolin or roscovitine or DMSO control were added to each culture (PC-3 or 22Rv1) containing charcoal/dextran-treated FBS (CDT serum, 5% condition for 24 hours. Cells were harvested and fixed in 80% ice-cold EtOH. for LNCaP cells and 10% for PC-3 and 22Rv1 cells; BioSource, Rockville, Following fixation, cells were stained with propidium iodide (0.2 Ag/AL) and MD). All media for cell types were supplemented with 100 units/mL subjected to flow cytometry to detect propidium iodide intensity. Samples penicillin-streptomycin and 2 mmol/L L-glutamine (Mediatech). Cells were were analyzed and quantified on a Beckman Coulter (Fullerton, CA) Cell cultured at 37jCina5%CO2 humidified incubator. Lab Quanta SC flow cytometer. Histograms represent f10,000 cells. Cell growth and survival assessment. Cells were seeded to approxi- Statistical assessment. Quantitative results are expressed as mean F mately 3.5 Â 105 per well in six-well dishes into appropriate medium with SD. Statistical analyses were done using one-way ANOVA followed by indicated concentration of hormone. For cell growth assays, cells were Newman-Keuls’ multiple comparison post test. The criterion for statistical cultured in designated conditions for 48 hours. For survival assays, cells significance was P < 0.05. were seeded as above and indicated doses of paclitaxel, docetaxel, or DMSO control were added to each well after 24 hours in culture and challenged for Results indicated hours (24 or 72 hours). After treatment, viable cells were counted using a hemacytometer and trypan blue exclusion. To define the effect of AR activity reduces cell survival in response to taxanes. taxanes on cell survival, the number of cells remaining after taxane Given the influence of AR activity on cell survival and proliferation treatment was set relative to each condition without cytotoxic challenge in prostate cancer, the effect of androgen on the response to (100% survival). For experiments analyzing the effect of bicalutamide taxanes in prostate cancer cells was determined. For these initial (Casodex) or aphidicolin/roscovitine on paclitaxel-mediated cell death, studies, PC-3 (AR negative, androgen independent) and LNCaP (AR À 10 6 mol/L Casodex or 2 Ag/mL aphidicolin or 5 Ag/mL roscovitine was positive, androgen dependent) cells were cultured in the absence or added to indicated medium after cells had adhered to surface (before presence of steroid hormones (CDT or FBS sera, respectively). After administration of paclitaxel). The Ac-VAD-CHO pan-caspase inhibitor was 24-hour pretreatment, parallel cultures were treated with either included at a final concentration of 50 Amol/L and administered 1 hour À6 À 10 mol/L paclitaxel or vehicle (DMSO) control. Following a before paclitaxel exposure. Following 24 hours of 10 6 mol/L paclitaxel 24-hour incubation, cell viability was assessed using trypan blue treatment, cell survival was determined as above. Total cell number was determined for each condition in triplicate samples, and each experiment exclusion. As shown in Fig. 1A (left), cell viability was unaffected by f was replicated at least thrice. hormone background in PC-3 cells, wherein 50% of cells survived Quantification of micronucleated cells. Cells were seeded on poly-L- in either CDT or FBS (P > 0.05). This magnitude of cell death and lysine–coated coverslips under conditions indicated. After a 24-hour survival is consistent with previous reports (15) and indicates that À attachment period, 10 6 mol/L paclitaxel or DMSO control was added to sensitivity of PC-3 cells to paclitaxel is unaffected by steroid the cells. Cells were treated for 16 hours and subsequently fixed in 3.7% hormones. Conversely, LNCaP cells cultured in steroid hormones formaldehyde. Cells were permeabilized in 0.3% Triton X-100 at room (FBS) showed substantially reduced cell survival in the presence of temperature for 20 minutes before addition of Hoechst 33258 (Sigma- paclitaxel (25% survival) compared with those depleted of steroid Aldrich, St. Louis, MO). Hoechst was added to a final concentration of 0.1 hormone (55–60% survival; P < 0.001; Fig. 1A, right). These data Ag/mL in PBS and incubated for 30 minutes at 37jC. Coverslips were indicate that the sensitivity of AR-dependent prostate cancer cells mounted on glass slides, and nuclei were visualized by indirect immunofluorescence. Experiments were done twice in triplicate, and at to paclitaxel may be strongly influenced by the hormonal milieu least 200 cells were scored per condition. and implicate AR activity as a potential effector of paclitaxel- Immunoblotting. Cells were treated as described for survival assays. For mediated cell death. To challenge this concept directly, a specific protein analyses, total cells were harvested and lysis was done in AR antagonist, bicalutamide (Casodex), was used. For these studies, radioimmunoprecipitation assay buffer supplemented with protease LNCaP cells pretreated with either CDT or FBS were cosupple- À inhibitor mixture and phenylmethylsulfonyl fluoride. Lysates were subjected mented with 10 6 mol/L Casodex and subsequently challenged to brief sonication and clarified by centrifugation. Protein concentration with paclitaxel using the strategy outlined in Fig. 1A. As shown, was determined using Bio-Rad DC Protein Assay (Bio-Rad, Hercules, CA), Casodex had no significant effect on cell survival in the presence of and equal protein was loaded and subjected to SDS-PAGE. Proteins were CDT, wherein AR activity is already inhibited as a function of ligand transferred to Immobilon-P membrane (Millipore Corp., Bedford, MA) and depletion (Fig. 1B). By contrast, Casodex increased cell survival in immunoblotted for the indicated proteins. Antigen-antibody complexes were visualized using enhanced Western Lightning chemiluminescence the presence of steroid hormone, improving cell survival from 25% (Perkin-Elmer Life Sciences, Wellesley, PA). Ser15-p53 fluorescence was to 50% (P < 0.01). determined using Alexa Fluor 680 rabbit secondary (Molecular Probes/ To further confirm that AR activity augments taxane cytotoxicity, Invitrogen, Carlsbad, CA) and imaged and quantified on Odyssey IR Imaging we sought to define whether the specific taxane, dose, and/or System (LI-COR Biosciences, Lincoln, NB). timing affect the change in cytotoxic efficacy imparted by AR www.aacrjournals.org 11999 Cancer Res 2006; 66: (24). December 15, 2006 Cancer Research

Figure 1. AR activity reduces cell survival in response to taxanes. A, PC-3 or LNCaP cells were seeded either in CDT or FBSmedium. After 24 hours, 10À6 mol/L paclitaxel or DMSO vehicle control was added to each set of cells. Percentage cell survival for each cell type after 24 hours of paclitaxel treatment was determined by trypan blue exclusion comparing cell number after paclitaxel exposure with cell number in DMSO control for each condition. B, percentage survival following paclitaxel treatment in LNCaP cells was determined as in (A); however, cells were initially seeded in CDT or FBS F 10À6 mol/L Casodex (CSDX). C, percentage survival following paclitaxel or docetaxel treatment at 10À9,10À8,10À7, and 10À6 mol/L in LNCaP was determined as in (A) for 24 hours of exposure and also at 72 hours after addition of cytotoxic challenge, wherein after 24 hours of exposure fresh medium was replaced and cells were allowed to propagate for an additional 48 hours. The number of viable cells was set relative to each condition without taxane. All experiments were done at least thrice in triplicate. ***, P < 0.001; **, P < 0.01. action. For these studies, LNCaP cells were seeded as described in ligand is linear, hormone-dependent cells (including LNCaP) Fig. 1A in either CDT or FBS conditions. The following day, various exhibit a biphasic dose response to steroid stimulation (17). À À À À À doses of paclitaxel (10 9,108,107,or106 mol/L) or DMSO Specifically, physiologic doses of androgen (10 10 mol/L DHT) vehicle control were added. After 24 hours of exposure, cells were facilitate mitogenesis, and higher doses inhibit cellular prolifer- either counted and scored for viability, as in Fig. 1A, or placed into ation, although still facilitate AR transactivation (17). Therefore, it fresh CDT or FBS medium and cultured for an additional 48 hours was imperative to determine whether the ability of AR to enhance and then counted as described above. As can be seen in Fig. 1C paclitaxel cytotoxicity was attributed to AR activation overall or (top), paclitaxel-mediated cell death was enhanced in the presence was specific to the mitogenic function of AR. To assess this, of hormone (FBS) for all doses tested, and this disparity in survival mitogenic steroid hormone doses were validated to induce AR- persisted for up to 72 hours after the recovery period (Fig. 1C, top). dependent proliferation in prostate cancer cells. Cells were seeded In parallel experiments, LNCaP cells were challenged with in triplicate at a density of 3.5 Â 105 per well under each docetaxel to determine the effect of hormone on this cytotoxic condition (dotted line). As expected, under steroid-depleted response. As can be seen in Fig. 1C (bottom), hormone exacerbated conditions (CDT plus EtOH vehicle; Fig. 2A, white columns), docetaxel-mediated cell death at each dose tested. LNCaP cells LNCaP cells failed to undergo cell doubling during the growth proved to be generally more sensitive to docetaxel cytotoxicity period. Also consistent with previous results, DHT maximally À compared with paclitaxel especially when combined with hormone. enhanced proliferation at 10 10 mol/L, whereas higher doses À À Together, these data indicate that AR activation synergizes with failed to induce proliferation (10 7 and 10 5 mol/L, respectively; taxanes to enhance cell death. ref. 17). A similar response was observed with the commonly used À Mitogenic doses of AR ligands synergize with paclitaxel to DHT analogue R1881. For this compound, 10 11 mol/L was the reduce cell survival. Because our data indicate that hormone optimal dose (P < 0.05), and higher concentration levels inhibited enhances paclitaxel-mediated cytotoxicity, we hypothesized that proliferation. Lastly, the effect of E2 was examined, as LNCaP cells this synergy is attributed to the ability of AR activation to express a somatic mutant of AR (AR-T877A) that commonly stimulate proliferation in AR-dependent cells. The central arises during prostate cancer disease progression. This mutant cytotoxic action of paclitaxel is mediated through microtubule renders the receptor amenable to activation by alternate steroid stabilization in mitosis (16). Thus, the ability of AR to bolster hormones, especially estrogen (18). As shown, E2 induced À paclitaxel action could be dependent on its mitogenic capacity. maximal cellular proliferation at 10 9 mol/L (P < 0.001) and Interestingly, although the transcriptional response of AR to higher doses reduced proliferation.

Cancer Res 2006; 66: (24). December 15, 2006 12000 www.aacrjournals.org Androgen Receptor Action Synergizes with Taxanes

Based on this information, the effect of individual AR ligands on AR-mediated gene transcription and cellular proliferation (19). the response to paclitaxel was monitored. LNCaP cells were We have previously shown that BPA optimally induces AR- À pretreated for 24 hours with either CDT supplemented with vehicle mediated cellular proliferation at 10 9 mol/L, whereas BPA inhibits À control (EtOH), the steroid precursor cholesterol (negative control), cell proliferation at micromolar doses (10 6 mol/L or higher; ref. À À mitogenic doses of AR ligands (10 10 mol/L DHT, 10 11 mol/L 20). As shown in Fig. 2C, pretreatment with the mitogenic dose of À À À R1881, or 10 9 mol/L E2), or nonmitogenic doses (10 7 mol/L DHT, BPA (10 9 mol/L) enhanced the effect of paclitaxel-mediated À À 10 7 mol/L R1881, or 10 6 mol/L E2; Fig. 2B). Cells were then cytotoxicity compared with vehicle-treated cells (40% compared À À challenged with 10 6 mol/L paclitaxel or DMSO control for with 60% cell survival). High-level BPA (10 6 mol/L), which inhibits 24 hours. As expected, cholesterol has no effect on paclitaxel- proliferation, did not statistically alter the paclitaxel-mediated mediated cell death (Fig. 2B, light gray columns) and resulted in reduction in cell survival (just under 70% compared with the 60% f60% cell survival similar to paclitaxel effects in the presence of vehicle control). Collectively, these data strongly support a model CDT plus vehicle treatment alone. However, when the mitogenic wherein AR activity significantly enhances the response to À À doses of AR ligands were used (10 10 mol/L DHT, 10 11 mol/L paclitaxel; however, this synergistic function of AR on paclitaxel- À R1881, and 10 9 mol/L E2), cell death following 24-hour paclitaxel mediated cell death is exquisitely dependent on the mitogenic treatment was enhanced (35%, 45%, and 25% cell survival capacity of AR. A table summarizing the dose-dependent effects of compared with no-paclitaxel treatment, respectively). Conversely, AR agonists on the response to paclitaxel is shown in Fig. 2D. À À at the nonmitogenic doses (10 7 mol/L DHT, 10 7 mol/L R1881, Androgen-independent activation of AR synergizes with À and 10 6 mol/L E2), there was no change in cell survival following paclitaxel to reduce cell survival. The data herein suggest that paclitaxel exposure (60%, 75%, and 60% cell survival, respectively). AR ligands synergize with paclitaxel to reduce prostate cancer cell These data indicate that AR ligands are capable of enhancing survival dependent on the cell cycle progression function of AR paclitaxel-mediated cell death only at doses in which cellular activation. Ligand-independent modes of AR activation may play a proliferation is induced. To validate this concept, an exogenous AR- role in prostate cancer disease progression, as activation of AR via T877A ligand was used, BPA. BPA is a prevalent environmental growth factor pathways results in cell cycle progression under compound and activates AR-T877A, inducing dose-dependent androgen deprivation conditions (21). Therefore, we aimed to

Figure 2. Mitogenic doses of AR ligand synergize with paclitaxel to reduce cell survival. A, 3.5 Â 105 LNCaP cells (dotted line) were seeded into CDT medium supplemented with EtOH control, DHT (10À13,10À10,10À7, and 10À5 mol/L), R1881 (10À13,10À11,10À10, and 10À7 mol/L), or E2 (10À12,10À9,10À6, and 10À5 mol/L). Viable cells were counted after 48 hours of culture. Experiments were done at least twice in triplicate. B, LNCaP cells were seeded into CDT medium F cholesterol (CHL;10À9 or 10À6 mol/L), DHT (10À10 or 10À7 mol/L), R1881 (10À11 or 10À7 mol/L), or E2 (10À9 or 10À6 mol/L). After 24 hours, 10À6 mol/L paclitaxel or DMSO control was added for 24 hours. Cell survival in paclitaxel was determined by comparing the number of viable cells exposed to paclitaxel relative to each culture condition in DMSO alone. ***, P < 0.001; *, P < 0.05. C, LNCaP cells were seeded as described above in CDT medium containing either 10À9 or 10À6 mol/L BPA. Cell survival following 10À6 mol/L paclitaxel challenge was determined as in (B). **, P < 0.01. All cell survival assays were done at least twice in triplicate. D, summary of fold changes in cell survival following paclitaxel challenge relative to androgen-depleted conditions (CDT). www.aacrjournals.org 12001 Cancer Res 2006; 66: (24). December 15, 2006 Cancer Research determine the effect of a previously shown and clinically relevant androgen-independent growth stimulation on paclitaxel-mediated cell death. Androgen-independent proliferation of prostate cancer cells can occur through HRG activation of HER2-HER3 or HER2-HER4 heterodimers, which has been shown to induce AR transactivation and phosphorylation (22, 23) and also increase AR stability and DNA binding (24). LNCaP cells were seeded as described in Fig. 2A and supplemented with recombinant HRG (50 ng/mL) for 48 hours in the absence of hormone (CDT). Consistent with previous reports, treatment with HRG resulted in increased cell proliferation compared with vehicle control con- ditions (Fig. 3A; compare white columns with black columns). Parallel experiments were done in the presence of the direct AR À antagonist, Casodex, wherein 10 6 mol/L Casodex was added before the supplementation with HRG (Fig. 3A, stripped columns). Casodex had no effect on the growth of LNCaP cells in hormone- depleted medium; however, blocking AR with Casodex resulted in partial inhibition of HRG-mediated growth. These data are consistent with previous observations, wherein HRG-induced AR activation is only partially sensitive to AR antagonist action (21, 23). To determine whether HRG activity is sufficient to synergize with paclitaxel, LNCaP cells were cultured as described À for Fig. 3A but challenged with 10 6 mol/L paclitaxel for 24 hours before viability assessment. Consistent with Fig. 1A, LNCaP cells Figure 3. Androgen-independent AR activation synergizes with paclitaxel to exposed to paclitaxel for 24 hours in the absence of hormone reduce cell survival. A, 3.5 Â 105 LNCaP cells (dotted line) were seeded into showed f60% survival (Fig. 3B). Importantly, this survival was CDT medium supplemented with DMSO control or 50 Ag/mL recombinant HRG. greatly attenuated by growth factor receptor activation, wherein Viable cells were counted after 48 hours of culture. ***, P < 0.001. B, LNCaP cells were seeded into CDT medium F 50 Ag/mL HRG. After 24 hours, LNCaP cells supplemented with 50 ng/mL HRG showed diminished 10À6 mol/L paclitaxel or DMSO control was added and culture was continued cell survival following paclitaxel challenge (f50% cells remaining; for an additional 24 hours. Cell survival in paclitaxel was determined by Fig. 3B). These data show that, in addition to ligand-induced AR comparing the number of viable cells exposed to paclitaxel relative to each culture condition in DMSO alone. *, P < 0.05. Experiments were done at least activation, the efficacy of taxanes can be enhanced by androgen- twice in triplicate. independent, AR-dependent proliferation. The synergistic effect of paclitaxel and AR activation is mediated through caspase-dependent apoptosis and p53 To directly monitor induction of apoptosis, the abundance of activation. These data indicate that AR-mediated mitogenic cleaved PARP in each hormonal environment was determined in action in prostate cancer cells enhances sensitivity to paclitaxel- the presence or absence of paclitaxel treatment. PARP is mediated cytotoxicity; therefore, the effect of AR function on the enzymatically cleaved by active caspases as a final step in cellular apoptotic index was monitored. The precise cytotoxic mechanisms apoptosis, and the relative amount of cleaved PARP to total PARP of paclitaxel-mediated cell death in prostate cancer have not been reflects the proportion of the population undergoing apoptosis completely defined, and paclitaxel action is heavily dependent on (26). Paclitaxel challenge has been shown to induce cleavage and cellular context and genetic background (16). To assess the activation of caspases (25). As expected, no PARP cleavage was influence of AR on paclitaxel-induced cell death, the appearance observed in the absence of paclitaxel challenge, regardless of of micronuclei was monitored. Cells develop multiple micronuclei hormonal environment (Fig. 4B, lanes 1, 3, and 5). Cells cultured in after paclitaxel exposure due to nuclear membrane reformation androgen-depleted conditions and treated with paclitaxel exhibited around unsegregated chromosomes, and this process leads to only a marginal increase in cleaved PARP fragment as evident by apoptosis (25). LNCaP cells were treated as in Fig. 2B, and the the appearance of the 89-kDa band. However, more pronounced percentage of micronucleated cells in each hormone condition induction of the cleaved PARP band was observed in conditions of was analyzed via nuclear Hoechst stain (Fig. 4A). As expected, AR activation (FBS or DHT; Fig. 4B, compare lanes 1 and 2 with cells cultured in a steroid-depleted condition showed no micro- lanes 3 and 4 and lanes 5 and 6) when normalized to the loading nucleated cells in the absence of paclitaxel challenge. Identical control (CDK4). Combined, these data indicate that AR agonists results were observed on hormone stimulation (FBS or DHT). synergize with paclitaxel to induce an apoptotic response. À However, following 24 hours of 10 6 mol/L paclitaxel challenge, Furthermore, these data indicate that this synergistic effect is micronuclei were readily discerned and counted. In the steroid- manifest through caspase-dependent mechanisms. depleted condition (CDT plus EtOH vehicle), f20% of the To validate the concept that AR activation enhances caspase- population showed signs of micronucleation (Fig. 4A, right). dependent apoptosis in response to paclitaxel, cell survival assays This apoptotic index was markedly increased in the presence of were done in the presence of a caspase inhibitor. Specifically, cells AR agonists (FBS or DHT), wherein the micronuclear index was were treated as in Fig. 2B but were pretreated for 1 hour before raised to 55% (P < 0.01) and 43% (P < 0.05), respectively. The paclitaxel administration with 50 Amol/L Ac-VAD-CHO, a pan- observed increase in micronucleated cells is consistent with the caspase inhibitor (Fig. 4C). Paclitaxel-mediated reduction in cell hypothesis that AR activation enhances cell death in response survival was not affected by Ac-VAD-CHO in cells pretreated in to paclitaxel. androgen-depleted conditions (CDT plus EtOH; f65%). However,

Cancer Res 2006; 66: (24). December 15, 2006 12002 www.aacrjournals.org Androgen Receptor Action Synergizes with Taxanes

Ac-VAD-CHO effectively nullified the synergistic effect of androgen shown in Fig. 4D, left, basal levels of p53 were identical in the on paclitaxel-mediated cell death. As shown, pretreatment in absence of paclitaxel, regardless of hormonal status (Fig. 4D, FBS before paclitaxel exposure reduced cell survival to 25%, but compare lanes 1, 3, and 5). Although after paclitaxel challenge, p53 Ac-VAD-CHO restored cell survival to levels comparable with the levels were induced in each condition (Fig. 4D, lanes 2, 4, and 6); no-hormone condition (60%). Similar effects were observed with interestingly, this induction was modestly enhanced by the DHT, wherein cell survival was restored to 65% from 35% on Ac- presence of AR agonists (FBS and DHT). A similar profile was VAD-CHO administration. These data indicate that caspase observed when phosphorylation-specific antisera against Ser15 of activation is required for the ability of AR agonists to enhance p53 were used to identify active p53 (Fig. 4D, second row, compare paclitaxel-mediated cell death. lane 2 with lanes 4 and 6). Fold increase of active p53 AR agonists modulate the expression of selected apoptotic (phosphorylated Ser15) for each hormone condition challenged proteins (14, 27–30). In addition, paclitaxel treatment of cells can with paclitaxel was determined by signal intensity relative to also alter expression of apoptotic factors dependent on cellular no-paclitaxel treatment control. In androgen-depleted conditions, context (31, 32). Therefore, protein levels of suggested AR and p53 was activated 1.67-fold, whereas under mitogenic conditions paclitaxel effectors were assessed under each hormone condition. (FBS and DHT) p53 was activated 18-fold and 10-fold, respectively. The p53 tumor suppressor gene has been implicated as a mediator Equal loading was confirmed using CDK4 as a control (Fig. 4D, of cellular sensitivity to paclitaxel, as p53 is a known modulator of bottom row). These data indicate that paclitaxel challenge activated cell death following DNA damage (33). However, paclitaxel does not p53 under all hormonal conditions but that p53 activation is activate p53 in all cell types (34). To assess the effect of paclitaxel enhanced by the presence of AR agonists. on p53 status, in the LNCaP model system, cells were treated as in In specific cell types, the bcl-2 family of apoptotic proteins has Fig. 2B and harvested and protein was analyzed by immunoblot. As also been suggested to be modified by both AR and paclitaxel

Figure 4. The synergistic effect of paclitaxel and AR activation is attributed to enhanced apoptosis. A, LNCaP cells were seeded on coverslips in CDT + EtOH, FBS, or CDT + 10À10 mol/L DHT. The cells were exposed to 10À6 mol/L paclitaxel (PTX) or DMSO control for 24 hours. Cells were fixed and nuclei were stained with Hoechst DNA dye. Left, representative images; right, percentage micronucleated nuclei from at least three experiments done in triplicate. B, LNCaP cells were seeded in CDT + EtOH, FBS, or CDT + 10À10 mol/L DHT, and adherent and floating cells were collected after 24 hours of 10À6 mol/L paclitaxel or DMSO control. Isolated protein was immunoblotted with an antibody specific for both full-length and cleaved PARP (106 and 89 kDa, respectively). CDK4 is the loading control. C, LNCaP cells were seeded as described. One hour before paclitaxel exposure, 50 Amol/L Ac-VAD-CHO pan-caspase inhibitor was added to the medium. Following 24 hours of 10À6 mol/L paclitaxel treatment in the presence of Ac-VAD-CHO, cell survival was determined as described previously. D, LNCaP cells were seeded, harvested, and collected as in (B). Isolated protein was immunoblotted with the following: (left) antisera against p53, Ser15 p53, bcl-2, and Bax (numbers correspond to fluorescence intensity of Ser15 p53 relative to no paclitaxel control in the same condition); (middle) antisera against p21CIP1 and Hsp27; (right) antisera against E2F-1. CDK4 served as the loading control. www.aacrjournals.org 12003 Cancer Res 2006; 66: (24). December 15, 2006 Cancer Research

(10, 32). Therefore, expression levels of conventional apoptotic paclitaxel stimulation (Fig. 4D, right, compare lanes 2, 4, and 6), family members were monitored. Increased bcl-2 levels are corre- similar to reports in head and neck cancer cells after effective lated with disease progression in prostate cancer and androgen docetaxel treatment (43). Combined, these data indicate that the independence, suggesting a mechanism for resistance to apoptosis AR-enhanced cell death is attributed to caspase-dependent (35). As shown in Fig. 4D (left), bcl-2 was activated equally by apoptosis and associated with p53 activation. Consistent with the paclitaxel in all conditions tested (as evident by the phosphorylated, hypothesis that the mitogenic function of AR underlies enhanced slower mobility upper bands; lanes 2, 4, and 6) and androgen cell death, AR-induced cells expressed higher levels of E2F-1 and treatment did not induce detectable changes of bcl-2 or its were sensitized to paclitaxel. phosphorylated forms (compare third row, lanes 1 and 2, 3 and 4, The mitogenic action of AR is necessary for reduction of cell and 5 and 6). This finding matches similar results reported for this survival in the presence of paclitaxel. Based on the preceding model system (27), wherein paclitaxel challenge phosphorylates bcl- observations, we speculated that the ability of AR to increase 2, resulting in decreased ability of bcl-2 to form heterodimers with caspase-dependent cell death after paclitaxel challenge results from Bax (27, 36). Bax is a proapoptotic factor that, on oligomerization, an increased percentage of the population passing through mitosis, permeabilizes the mitochondria and induces cytochrome c release therefore sensitizing those cells to paclitaxel action. To challenge (37). No significant alteration in Bax was observed under any this hypothesis, 22Rv1 cells were used. These cells express a treatment condition examined (Fig. 4D, left). These data are functional endogenous AR, but androgen stimulation is dispensable congruent with previous observations that paclitaxel does not alter for mitogenic progression (44). Androgen independence was verified Bax levels in prostate cancer cells (32). Moreover, p53 is known to by monitoring cellular proliferation in the presence or absence of induce proapoptotic activity by both transcriptional and non- androgen. As shown in Fig. 5A, and consistent with previous reports, transcriptional mechanisms in prostate cancer cells (27). 22Rv1 cells proliferated identically under conditions of no hormone, In selected models, androgen has been suggested to modulate physiologic androgen stimulation, and even in the presence of high- the expression of two genes implicated in cell survival: p21 and dose androgen (45). However, AR is still activated by ligand in this Hsp27 (13, 14, 38). Therefore, the relevance of these factors for cell line as validated by monitoring expression of an AR target gene, androgen facilitated cell death after paclitaxel challenge was PSA (Fig. 5B). As can be seen, PSA mRNA levels were low under examined. As we and others have previously shown, androgens conditions of androgen depletion (Fig. 5B, lane 1) but induced on induce expression of the CDK2 inhibitor p21CIP1 (13, 38). As culture in FBS or CDT plus DHT (Fig. 5B, lanes 2 and 3). Therefore, expected, an increase in p21 under mitogenic conditions was seen, although androgen is dispensable in 22Rv1 cells for proliferation, highlighting the AR-mediated cell cycle regulation of this protein ligands induce AR target gene activation, including PSA. This model (Fig. 4D, middle, compare lane 1 with lanes 3 and 5). Although system was subsequently used to monitor the effect of androgen p21CIP1 can also be induced by p53 (39), p21 levels actually on paclitaxel-induced cytotoxicity. As shown in Fig. 5C, 22Rv1 decreased after paclitaxel challenge in FBS and DHT conditions cells were equally sensitive to paclitaxel treatment, regardless of (Fig. 4D, middle, compare lanes 3 with lanes 4 and 5 with 6). This androgen stimulation (compare EtOH with FBS or DHT; no change in p21 on paclitaxel challenge is interesting, as p21 statistical difference, P > 0.05). Moreover, this cell line expresses a induction has been shown to protect LNCaP derivatives from promiscuous variant mutant receptor (AR-H874Y) that can arise paclitaxel-mediated apoptosis (40). Thus, observed deregulation of during tumor progression and renders the cells sensitive to p21 may induce a permissive state for cell death that is enhanced additional ligands, including estrogen (44). Similar to the results by AR activity. It has been shown that androgen depletion induces observed with androgen, estrogen stimulation had no effect on up-regulation of survival proteins, such as Hsp27, whose expression 22Rv1 proliferation (data not shown) or on the magnitude of correlates with increased survival in response to cytotoxic stimuli paclitaxel-induced cell death (approximately 45–50% cell survival (14), and these observations were also recapitulated in this model for all agents; Fig. 5C). To validate that AR activation in the (Fig. 4D, middle, compare lanes 1 and 2 with lanes 3 to 6). androgen-independent cells does not alter the paclitaxel response, However, Hsp27 was unaffected by paclitaxel exposure, suggesting the effect of Casodex on paclitaxel-mediated cell death was that androgen deprivation conditions increase Hsp27 levels, monitored. As shown, addition of the AR antagonist Casodex had thereby protecting the cells from cytotoxic insult, and that this no effect on the cytotoxic effects of paclitaxel regardless of hormone protective response is impermeable to paclitaxel challenge. environment (Fig. 5D). These data indicate that, although androgen- Collectively, these data indicate that the synergistic effect of AR independent cells express active AR, activation of AR affords no activation with paclitaxel is dependent on mitogenic doses of AR enhancement of paclitaxel-mediated cell death. As such, these ligand and the proapoptotic function of AR is likely dependent on observations support the conclusion that AR enhances paclitaxel cell cycle progression. action through facilitating cell cycle progression. The hypothesis that enhanced cytotoxicity was a result of AR- Cell cycle progression is necessary for paclitaxel-mediated induced proliferation was borne out in the examination of E2F-1 cell death. It has been hypothesized that the ability of paclitaxel to (Fig. 4D, right). As expected, E2F-1 was induced under conditions interfere with cytokinesis would make dividing cells more sensitive; of AR activation (FBS and DHT; Fig. 4D, right, compare lane 1 with however, there has been significant variation in cellular sensitivity lanes 3 and 5). These data are consistent with previous to paclitaxel (46). The data herein suggest that the mitogenic action observations that androgen initiates RB inactivation and induction of AR enhances paclitaxel-induced apoptosis in prostate cancer of E2F-dependent target genes (38) and suggest that the ability of cells and therefore implicate AR-mediated cell cycle progression androgen to stimulate E2F-1 and subsequent cellular proliferation as the underlying mechanism of enhanced cell death. Therefore, likely underlies its ability to synergize with paclitaxel. E2F-1 the necessity of cell cycle progression in prostate cancer cells for expression is also known to be proapoptotic (41), and loss of E2F-1 paclitaxel-mediated apoptosis was challenged using pharmacologic can protect against chemotherapeutic-induced cell death in inhibitors of cell cycle progression. For these experiments, prostate cancer cells (42). Interestingly, E2F-1 was lost after asynchronously proliferating LNCaP cells were validated to exhibit

Cancer Res 2006; 66: (24). December 15, 2006 12004 www.aacrjournals.org Androgen Receptor Action Synergizes with Taxanes

Figure 5. The mitogenic action of the AR is necessary for reduction of cell survival in the presence of paclitaxel. A, 3.5 Â 105 22Rv1 cells were seeded (dotted line)in CDT + EtOH, FBS, or CDT supplemented with DHT (10À13,10À10,10À7, and 10À5 mol/L). Viable cells were counted following 48 hours of culture. B, 22Rv1 cells were seeded in CDT, FBS, or CDT supplemented with DHT (10À10 mol/L). Total RNA was isolated and converted to cDNA and then subjected to PCR using primers specific to GAPDH (loading) or PSA. C, 22Rv1 cells were seeded in CDT + EtOH, FBS, CDT + 10À10 mol/L DHT, or CDT + 10À9 mol/L E2. After 24 hours of 10À6 mol/L paclitaxel treatment, percentage survival for each condition was determined by comparing cell viability with paclitaxel to the DMSO control in the same condition. D, 22Rv1 cells were seeded into CDT or FBSmedium supplemented with 10À6 mol/L Casodex or DMSO control. Following 24 hours of 10À6 mol/L paclitaxel, cell survival was determined as in (C). All experiments were done at least twice in triplicate.

a normal cell cycle profile (FBS; Fig. 6A). Proliferating cells treated highest efficiency (48), whereas Tang et al. showed that taxane with the DNA polymerase inhibitor aphidicolin exhibited an early S (docetaxel) treatment was most efficacious as a neoadjuvant before phase (92.01% 2N), and similar effects were observed with the CDK castration (49). Given this discrepancy, it is imperative to define the inhibitor roscovitine (83.56% 2N: Fig. 6A), which arrests cells effect of androgen action and AR activity on the response to predominately in G1 but also in S phase through its ability to inhibit taxane-based cytotoxic insult. CDK2 activity (47). These conditions were used to challenge the Herein, we assessed the molecular and cellular consequence of requirement of cell cycle progression on the cytotoxic response to AR activity on the response to taxanes, cytotoxic agents recently paclitaxel (Fig. 6B). For these studies, LNCaP cells were pretreated in validated for use in patients with recurrent prostate cancer. Our the presence of mitogenic AR agonists (FBS or DHT) and either data show that AR activity is a critical determinant of the cellular 2 Ag/mL aphidicolin or 5 Ag/mL roscovitine. Subsequently, pacli- response to paclitaxel. We show that AR activation significantly taxel or DMSO control was administered, and after 24 hours of enhances the response to paclitaxel in AR-proficient, androgen- treatment, cell survival was assayed as described previously (Fig. 2B). dependent prostate cancer cells, whereas no effect of androgen was In the presence of AR agonists (FBS or DHT), both aphidicolin and observed in AR-proficient, androgen-refractory cells or AR- roscovitine markedly enhanced cell survival in the presence of deficient prostate cancer cells (Figs. 1 and 2). Additionally, paclitaxel. In both cases, forced G1 arrest nullified the cytotoxic ligand-independent AR activation was also sufficient to enhance effect of paclitaxel. These observations validate a model in which the cytotoxic effects of paclitaxel (Fig. 3). AR-enhanced loss of the mitogenic capacity of AR is required to synergize with paclitaxel survival was attributed to caspase-induced cell death and was to sensitize prostate cancer cells to cell death. associated with p53 activation (Fig. 4). Strikingly, the ability of AR to bolster cell death in response to paclitaxel precisely requires its mitogenic function and was not simply a consequence of AR- Discussion dependent transcriptional activity (Figs. 5and 6). Combined, these Prostate cancer growth and recurrence is dependent on AR data indicate that the mitogenic action of AR on cell cycle activity, and ablation of AR activity is the first-line therapeutic progression is requisite for efficient paclitaxel-induced cell death intervention in disseminated disease. However, recent studies and lay the molecular foundation for design of efficacious indicate that adjuvant, cytotoxic therapies could potentially extend therapeutic intervention. patient survival (reviewed in ref. 10). Given the need to improve The mechanism of AR-enhanced cell death revealed herein is outcome for prostate cancer patients with metastatic disease, it is inextricably linked to its mitogenic function. The concept that AR essential to discern the effect of androgen ablation therapy on induces cell cycle progression is well validated (38, 50), and cytotoxic regimens. Recent preclinical investigations assessing the paclitaxel is known to enhance cell death, at least in part, through effect of combinatorial and/or sequential use of androgen ablation its ability to prevent chromosome segregation in mitosis (16, 46) and taxane administration on overall survival and tumor volume in and in prostate cancer cells to require G2-M CDK activity (51). The xenograft models have led to divergent conclusions about present data show that the ability of AR to enhance cell death in androgen action. For example, Eigl et al. concluded that concurrent response to paclitaxel requires its capability to induce cell cycle administration of the taxane paclitaxel and castration provided the progression, as both ligand-dependent and ligand-independent www.aacrjournals.org 12005 Cancer Res 2006; 66: (24). December 15, 2006 Cancer Research

Figure 6. Cell cycle progression is necessary for enhancement of paclitaxel-mediated apoptosis, as cell cycle inhibitors reduce paclitaxel efficacy. A, representative cell cycle histograms of LNCaP cells either in CDT + EtOH, FBS, or FBS + 2 Ag/mL aphidicolin (APH)or5Ag/mL roscovitine (ROSC). Percentage of cells in each phase of cell cycle per treatment is noted. B, percentage survival of LNCaP cells seeded into FBSor CDT + 10À10 mol/L DHT with 2 Ag/mL aphidicolin, 5 Ag/mL roscovitine, or DMSO control. Following 24 hours of 10À6 mol/L paclitaxel treatment, cell survival was determined comparing paclitaxel-exposed cells with DMSO control in each culture condition. All experiments were done at least thrice in triplicate.

mechanisms induced AR activity and enhanced paclitaxel cytotox- requirement for prostate cancer maintenance (5). Moreover, icity. Additionally, high doses of androgen that strongly stimulate androgen-dependent prostate cancer cells in culture exhibited AR function but block cellular proliferation failed to bolster enhanced survival in the presence of androgen after challenge with paclitaxel function. Moreover, the proapoptotic effect of androgen agents, such as tumor necrosis factor (TNF), Fas ligand, and was nullified by AR antagonist administration or through forced okadaic acid (12). In the TNF and Fas ligand study, androgen cell cycle inhibition, thus validating the concept that the pro- attenuated proapoptotic Bax expression and prevented caspase- proliferative function of AR underlies its ability to enhance induced cell death (12). However, the present study failed to show paclitaxel action. Parallels have been observed in other systems, any influence of androgen on Bax expression, and caspase- wherein advances in cell cycle progression can increase sensitivity dependent cell death was enhanced in the presence of androgen to specific cytotoxic agents. For example, loss of the retinoblastoma after paclitaxel administration (Fig. 4). Similarly, although expres- tumor suppressor protein, RB, sensitizes cells to cell death induced sion of the antiapoptotic protein bcl-2 is increased in androgen- by cisplatin and etoposide (52). These observations are relevant to independent prostate cancer and potentially contributes to the present study, as androgen induces cell cycle progression therapeutic resistance (6, 59, 60), no alterations in bcl-2 expression through induction of CDK-mediated RB phosphorylation and or phosphorylation were discerned as a function of hormonal inactivation (38). RB inactivation is known to derepress E2F-1 status (Fig. 4D). Together, the present data do not support a role for function (reviewed in ref. 53), and indeed, E2F-1 expression was AR in altering paclitaxel-induced alterations in Bax or bcl-2 levels. elevated in unchallenged cells under conditions of androgen- Elevation in the expression of the cochaperone Hsp27 has also been induced cell cycle progression (Fig. 4D). Interestingly, E2F-1 was documented as a prognostic for poor outcome and is associated specifically down-regulated in androgen-stimulated cells after with enhanced survival (14, 59). Although Hsp27 levels seemed paclitaxel administration; this result is consistent with observa- higher in cells deprived of androgen, no significant alteration of tions of effective taxane treatment in head and neck cancers (43). basal Hsp27 levels was observed after paclitaxel administration Although the implications of this event have yet to be identified, it under any hormonal condition (Fig. 4D). Thus, it is unlikely that is intriguing that E2F-1 can regulate expression of multiple factors Hsp27 contributes to the ability of AR activation to enhance that control apoptosis, including Apaf-1, p73, PUMA, and Bim paclitaxel-mediated cell death. (54–57). In addition, the E2F-1 regulated gene MDM2 has been In contrast, several factors associated with the DNA damage implicated in controlling AR function (58); thus, potential cross- response and cell cycle progression were significantly altered by talk between these pathways is worthy of future study. Combined, androgen and paclitaxel. First, p21CIP1 was induced by androgen these data indicate that the ability of AR to enhance paclitaxel- (Fig. 4D), consistent with previous reports and consistent with the CIP1 mediated cell death is a direct result of its mitogenic capacity. requisite ability of p21 to activate CDK4/cyclin D1 activity in G1 The concept that AR enhances cell death is striking, as previous phase and promote cell cycle progression (61). Although p21CIP1 studies have implicated androgen as a survival factor under can inhibit CDK2 activity in late G1 in specific conditions, evidence disparate conditions. For example, prostate cancer cells undergo has shown that p21CIP1 induction correlates with enhanced cellular apoptosis on androgen withdrawal in vivo, thus implicating AR proliferation in prostate cancer cells (38) and is required for CDK4

Cancer Res 2006; 66: (24). December 15, 2006 12006 www.aacrjournals.org Androgen Receptor Action Synergizes with Taxanes function (61). Interestingly, when paclitaxel was administered in In summary, although it is indisputable that androgen and the cycling (androgen stimulated) cells, p21CIP1 was markedly reduced, AR act as survival factors during the response to androgen perhaps reflecting the alteration in cell cycle position (G2-M ablation, our data show that the ability of AR activity to induce enrichment) and enhanced cell death on taxane exposure. cellular proliferation can enhance cell death on genotoxic insult. Consistent with this idea, p53 activation was more pronounced The present data indicate that AR activation cooperates with in these same conditions as shown by increases in overall p53 levels paclitaxel to enhance cell death and that this function of AR is and Ser15-phosphorylated (activated) forms of the protein (Fig. 4D). exquisitely dependent on its ability to promote cellular In addition to cellular stress, p53 is induced in cells held in mitosis proliferation. Combined, these studies indicate that paclitaxel and is an indicator of ‘‘mitotic timing’’ (62). Our data are consistent is likely to be most efficacious under conditions wherein AR with the hypothesis that increased p53 activation results from exerts its mitogenic function, thus indicating that the response enhanced paclitaxel-mediated DNA damage and/or G2-M accumu- to cell cycle–dependent cytotoxic insult may be more pro- lation in cycling (androgen stimulated) rather than arrested nounced before androgen deprivation or under specific molec- (androgen depleted) cells. ular environments of heightened AR activity. Therefore, these The fact that p53 and p21 expression levels were reciprocally studies provide the impetus for future investigations directed at induced was surprising, as p21 is a target gene of p53 (39). delineating the effect of AR status on the response to cytotoxic However, this result is not without precedent, as the therapeutic insult. agent and CDK inhibitor flavopiridol induces p53 activity with concurrent p21CIP1 down-regulation in prostate cancer cells (63). Interestingly, p21CIP1 has been shown to both protect against Acknowledgments apoptosis [e.g., as induced by doxorubicin (64)] and enhance Received 6/19/2006; revised 9/22/2006; accepted 9/29/2006. apoptosis (e.g., as induced by green tea ) in prostate Grant support: NIH grants RO1-CA 099996 and RO1-CA 93404 (K.E. Knudsen), CIP1 National Institute of Environmental Health Sciences (NIEHS) Center for Environmen- cancer cells (65). Thus, although the mechanisms by which p21 tal Genetics core grant E30-ES-06096, and NIEHS Environmental Mutagenesis and may affect the apoptotic response have not been elucidated, the Cancer training grant ES-07250-16 (J.K. Hess-Wilson). present data indicate that induction of p53 and attenuated p21CIP1 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 accordance expression correlate with paclitaxel-induced cell killing in andro- with 18 U.S.C. Section 1734 solely to indicate this fact. gen-dependent cells. Thus, further studies should be directed at We thank all the members of the K. Knudsen lab and Dr. Erik Knudsen for critical CIP1 discussions on the study and article; Drs. Lisa Morey, Clay Comstock, and Kevin Link determining whether p53 and p21 status alter hormonal for critical reading and editing of the article; and Drs. Sohaib Khan and Robin influence on paclitaxel function. Therakan and other members of the Khan lab for reagents and collegial support.

References 14. Rocchi P, So A, Kojima S, et al. Heat shock protein 27 induction of androgen target genes through MAP increases after androgen ablation and plays a cytopro- kinase in prostate cancer cells. Proc Natl Acad Sci 1. Jemal A, Tiwari RC, Murray T, et al. Cancer statistics, tective role in hormone-refractory prostate cancer. U S A 1999;96:5458–63. 2004. CA Cancer J Clin 2004;54:8–29. Cancer Res 2004;64:6595–602. 24. Mellinghoff IK, Vivanco I, Kwon A, Tran C, Wongvipat 2. Shang Y, Myers M, Brown M. Formation of the 15. Shankar S, Chen X, Srivastava RK. Effects of J, Sawyers CL. HER2/neu kinase-dependent modulation androgen receptor transcription complex. Mol Cell sequential treatments with chemotherapeutic drugs of androgen receptor function through effects on DNA 2002;9:601–10. followed by TRAIL on prostate cancer in vitro and binding and stability. Cancer Cell 2004;6:517–27. 3. Klotz L. Hormone therapy for patients with prostate in vivo. Prostate 2005;62:165–86. 25. Schatten H, Ripple M, Balczon R, et al. Androgen and carcinoma. Cancer 2000;88:3009–14. 16. Blagosklonny MV, Fojo T. Molecular effects of taxol cause cell type-specific alterations of centrosome 4. Russell DW, Wilson JD. Steroid 5a-reductase: two paclitaxel: myths and reality (a critical review). Int J and DNA organization in androgen-responsive LNCaP genes/two enzymes. Annu Rev Biochem 1994;63:25–61. Cancer 1999;83:151–6. and androgen-independent DU145prostate cancer cells. 5. Trapman J, Brinkmann AO. The androgen receptor in 17. Kim IY, Kim JH, Zelner DJ, Ahn HJ, Sensibar JA, Lee C. J Cell Biochem 2000;76:463–77. prostate cancer. Pathol Res Pract 1996;192:752–60. Transforming growth factor-h1isamediatorof 26. Decker P, Muller S. Modulating poly (ADP-ribose) 6. Wang G, Reed E, Li QQ. Apoptosis in prostate cancer: androgen-regulated growth arrest in an androgen- polymerase activity: potential for the prevention and progressive and therapeutic implications (Review). Int J responsive prostatic cancer cell line, LNCaP. Endocri- therapy of pathogenic situations involving DNA damage Mol Med 2004;14:23–34. nology 1996;137:991–9. and oxidative stress. Curr Pharm Biotechnol 2002;3: 7. Feldman BJ, Feldman D. The development of andro- 18. Taplin ME, Bubley GJ, Ko YJ, et al. Selection for 275–83. gen-independent prostate cancer. Nat Rev Cancer 2001; androgen receptor mutations in prostate cancers treated 27. Lin Y, Kokontis J, Tang F, et al. Androgen and its 1:34–45. with androgen antagonist. Cancer Res 1999;59:2511–5. receptor promote Bax-mediated apoptosis. Mol Cell Biol 8. Visakorpi T, Hyytinen E, Koivisto P, et al. In vivo ampli- 19. Wetherill YB, Petre CE, Monk KR, Puga A, Knudsen 2006;26:1908–16. fication of the androgen receptor gene and progression of KE. The xenoestrogen bisphenol A induces inappropri- 28. Bruckheimer EM, Kyprianou N. Dihydrotestosterone human prostate cancer. Nat Genet 1995;9:401–6. ate androgen receptor activation and mitogenesis in enhances transforming growth factor-h-induced apo- 9. Craft N, Sawyers CL. Mechanistic concepts in andro- prostatic adenocarcinoma cells. Mol Cancer Ther 2002;1: ptosis in hormone-sensitive prostate cancer cells. gen-dependence of prostate cancer. Cancer Metastasis 515–24. Endocrinology 2001;142:2419–26. Rev 1998;17:421–7. 20. Wetherill YB, Fisher NL, Staubach A, Danielsen M, 29. Briehl MM, Miesfeld RL. Isolation and characteriza- 10. Petrylak DP. The current role of chemotherapy in de Vere White RW, Knudsen KE. Xenoestrogen action tion of transcripts induced by androgen withdrawal and metastatic hormone-refractory prostate cancer. Urology in prostate cancer: pleiotropic effects dependent on apoptotic cell death in the rat ventral prostate. Mol 2005;65:3–7; discussion -8. androgen receptor status. Cancer Res 2005;65:54–65. Endocrinol 1991;5:1381–8. 11. Leung S, Miyake H, Zellweger T, Tolcher A, Gleave 21. Craft N, Shostak Y, Carey M, Sawyers CL. A 30. Rokhlin OW, Taghiyev AF, Guseva NV, et al. Androgen ME. Synergistic chemosensitization and inhibition of mechanism for hormone-independent prostate cancer regulates apoptosis induced by TNFR family ligands via progression to androgen independence by antisense through modulation of androgen receptor signaling by multiple signaling pathways in LNCaP. Oncogene 2005; Bcl-2 oligodeoxynucleotide and paclitaxel in the LNCaP the HER-2/neu tyrosine kinase. Nat Med 1999;5:280–5. 24:6773–84. prostate tumor model. Int J Cancer 2001;91:846–50. 22. Gregory CW, Whang YE, McCall W, et al. Heregulin- 31. Blagosklonny MV, Schulte TW, Nguyen P, Mimnaugh 12. Kimura K, Markowski M, Bowen C, Gelmann EP. induced activation of HER2 and HER3 increases EG, Trepel J, Neckers L. Taxol induction of p21WAF1 Androgen blocks apoptosis of hormone-dependent androgen receptor transactivation and CWR-R1 human and p53 requires c-raf-1. Cancer Res 1995;55:4623–6. prostate cancer cells. Cancer Res 2001;61:5611–8. recurrent prostate cancer cell growth. Clin Cancer Res 32. Haldar S, Chintapalli J, Croce CM. Taxol induces bcl-2 13. Lu S, Liu M, Epner DE, Tsai SY, Tsai MJ. Androgen 2005;11:1704–12. phosphorylation and death of prostate cancer cells. regulation of the cyclin-dependent kinase inhibitor p21 23. YehS,LinHK,KangHY,ThinTH,LinMF,Chang Cancer Res 1996;56:1253–5. gene through an androgen response element in the C. From HER2/Neu signal cascade to androgen 33. Wahl AF, Donaldson KL, Fairchild C, et al. Loss of proximal promoter. Mol Endocrinol 1999;13:376–84. receptor and its coactivators: a novel pathway by normal p53 function confers sensitization to Taxol by www.aacrjournals.org 12007 Cancer Res 2006; 66: (24). December 15, 2006 Cancer Research

increasing G2/M arrest and apoptosis. Nat Med 1996;2: squamous cell carcinoma using cDNA microarray and 53. Nevins JR. The Rb/E2F pathway and cancer. Hum 72–9. PowerBlot. Clin Cancer Res 2002;8:3910–21. Mol Genet 2001;10:699–703. 34. Srivastava RK, Srivastava AR, Korsmeyer SJ, 44. Attardi BJ, Burgenson J, Hild SA, Reel JR. Steroid 54. Hershko T, Ginsberg D. Up-regulation of Bcl-2 Nesterova M, Cho-Chung YS, Longo DL. Involvement hormonalregulationofgrowth,prostatespecificantigen homology 3 (BH3)-only proteins by E2F1 mediates of microtubules in the regulation of Bcl2 phosphoryla- secretion, and transcription mediated by the mutated apoptosis. J Biol Chem 2004;279:8627–34. tion and apoptosis through cyclic AMP-dependent androgen receptor in CWR22Rv1 human prostate carci- 55. Stiewe T, Putzer BM. Role of the p53-homologue p73 protein kinase. Mol Cell Biol 1998;18:3509–17. noma cells. Mol Cell Endocrinol 2004;222:121–32. in E2F1-induced apoptosis. Nat Genet 2000;26:464–9. 35. McDonnell TJ, Troncoso P, Brisbay SM, et al. Expres- 45. Hartel A, Didier A, Ulbrich SE, Wierer M, Meyer HH. 56. Irwin M, Marin MC, Phillips AC, et al. Role for the sion of the protooncogene bcl-2 in the prostate and its Characterisation of steroid receptor expression in the p53 homologue p73 in E2F-1-induced apoptosis. Nature association with emergence of androgen-independent human prostate carcinoma cell line 22RV1 and quan- 2000;407:645–8. prostate cancer. Cancer Res 1992;52:6940–4. tification of androgen effects on mRNA regulation of 57. Furukawa Y, Nishimura N, Satoh M, et al. Apaf-1 is a 36. Longuet M, Serduc R, Riva C. Implication of bax in prostate-specific genes. J Steroid Biochem Mol Biol 2004; mediator of E2F-1-induced apoptosis. J Biol Chem 2002; apoptosis depends on microtubule network mobility. Int 92:187–97. 277:39760–8. J Oncol 2004;25:309–17. 46. Abal M, Andreu JM, Barasoain I. Taxanes: microtu- 58. Lin HK, Wang L, Hu YC, Altuwaijri S, Chang C. 37. Gross A, Jockel J, Wei MC, Korsmeyer SJ. Enforced bule and centrosome targets, and cell cycle dependent Phosphorylation-dependent ubiquitylation and degra- dimerization of BAX results in its translocation, mito- mechanisms of action. Curr Cancer Drug Targets 2003;3: dation of androgen receptor by Akt require Mdm2 E3 chondrial dysfunction, and apoptosis. EMBO J 1998;17: 193–203. ligase. EMBO J 2002;21:4037–48. 3878–85. 47. Meijer L, Borgne A, Mulner O, et al. Biochemical and 59. McKenzie S, Kyprianou N. Apoptosis evasion: the role 38. Knudsen KE, Arden KC, Cavenee WK. Multiple G1 cellular effects of roscovitine, a potent and selective of survival pathways in prostate cancer progression and regulatory elements control the androgen-dependent inhibitor of the cyclin-dependent kinases cdc2, cdk2, therapeutic resistance. J Cell Biochem 2006;97:18–32. proliferation of prostatic carcinoma cells. J Biol Chem and cdk5. Eur J Biochem 1997;243:527–36. 60. Catz SD, Johnson JL. BCL-2 in prostate cancer: a 1998;273:20213–22. 48. Eigl BJ, Eggener SE, Baybik J, et al. Timing is minireview. Apoptosis 2003;8:29–37. 39. Ewen ME, Miller SJ. p53 and translational control. everything: preclinical evidence supporting simulta- 61. LaBaer J, Garrett MD, Stevenson LF, et al. New Biochim Biophys Acta 1996;1242:181–4. neous rather than sequential chemohormonal ther- functional activities for the p21 family of CDK 40. Wang LG, Ossowski L, Ferrari AC. Overexpressed apy for prostate cancer. Clin Cancer Res 2005;11: inhibitors. Genes Dev 1997;11:847–62. androgen receptor linked to p21WAF1 silencing may be 4905–11. 62. Blagosklonny MV. Prolonged mitosis versus tetra- responsible for androgen independence and resistance 49. Tang Y, Khan MA, Goloubeva O, et al. Docetaxel ploid checkpoint: how p53 measures the duration of to apoptosis of a prostate cancer cell line. Cancer Res followed by castration improves outcomes in LNCaP mitosis. Cell Cycle 2006;5:971–5. 2001;61:7544–51. prostate cancer-bearing severe combined immunodefi- 63. Blagosklonny MV, Darzynkiewicz Z, Figg WD. Fla- 41. Kowalik TF, DeGregori J, Leone G, Jakoi L, Nevins JR. cient mice. Clin Cancer Res 2006;12:169–74. vopiridol inversely affects p21(WAF1/CIP1) and p53 and E2F1-specific induction of apoptosis and p53 accumu- 50. Debes JD, Tindall DJ. The role of androgens and the protects p21-sensitive cells from paclitaxel. Cancer Biol lation, which is blocked by Mdm2. Cell Growth Differ androgen receptor in prostate cancer. Cancer Lett 2002; Ther 2002;1:420–5. 1998;9:113–8. 187:1–7. 64. Martinez LA, Yang J, Vazquez ES, et al. p21 modulates 42. Libertini SJ, Tepper CG, Guadalupe M, Lu Y, Asmuth 51. Perez-Stable C. 2-Methoxyestradiol and paclitaxel threshold of apoptosis induced by DNA-damage and DM, Mudryj M. E2F1 expression in LNCaP prostate have similar effects on the cell cycle and induction of growth factor withdrawal in prostate cancer cells. cancer cells deregulates androgen dependent growth, apoptosis in prostate cancer cells. Cancer Lett 2006;231: Carcinogenesis 2002;23:1289–96. suppresses differentiation, and enhances apoptosis. 49–64. 65. Agarwal R. Cell signaling and regulators of cell Prostate 2006;66:70–81. 52. Knudsen KE, Booth D, Naderi S, et al. RB-dependent cycle as molecular targets for prostate cancer 43. Yoo GH, Piechocki MP, Ensley JF, et al. Docetaxel S-phase response to DNA damage. Mol Cell Biol 2000;20: prevention by dietary agents. Biochem Pharmacol induced gene expression patterns in head and neck 7751–63. 2000;60:1051–9.

Cancer Res 2006; 66: (24). December 15, 2006 12008 www.aacrjournals.org

Chapter V:

Xenoestrogen action in breast cancer: impact on ER-

dependent transcription and mitogenesis

Breast Cancer Research and Treatment (2006) 96: 279–292 Springer 2005 DOI 10.1007/s10549-005-9082-y

Xenoestrogen action in breast cancer: impact on ER-dependent transcription and mitogenesis

J.K. Hess-Wilson1, J. Boldison1, K.E. Weaver1, and K.E. Knudsen1,2 1Department of Cell Biology, University of Cincinnati College of Medicine, Cincinnati, Ohio, USA; 2Center for Environmental Genetics, University of Cincinnati College of Medicine, Cincinnati, Ohio, USA

Key words: breast cancer, bisphenol A, co-activators, coumestrol, estrogen receptor, ER-D351Y, xenoestrogens

Summary Several estrogen mimics (xenoestrogens) inappropriately activate the estrogen receptor (ER) in the absence of endogenous ligand. Given the importance of the ER in breast cancer growth and regulation, delineating the impact of these agents under conditions related to tumor treatment is of significant importance. We examined the effect of two prevalent xenoestrogens (bisphenol A and coumestrol) on ER activation and ER-dependent mitogenesis in breast cancer cells. We show that the ability of these agents to induce mitogenesis was restricted to conditions of estrogen depletion, and that these agents failed to cooperate with estradiol to induce MCF-7 breast cancer cell growth. These observations are consistent with the impact of each agent specifically on exogenous ER activation as monitored in HeLa cells, wherein the xenoestrogens activated the receptor in the absence of estradiol but failed to cooperate with estrogen. Tamoxifen blocked bisphenol A and coumestrol-mediated ER activation, indicating that exposure to these agents is unlikely to disrupt such therapeutic intervention. The response of tumor-derived ER alleles to these xenoestrogens was also examined. Although the xenoestrogens failed to alter ER-Y537S function, the ER-D351Y mutant demonstrated an enhanced response to bisphenol A. Moreover, tamoxifen enhanced the agonistic effects of xenoestrogens on ER-D351Y. Lastly, we examined the impact of ER co-activator overexpression on xenoestrogen response. Bisphenol A and coumestrol exhibited differential responses to co-activators with regard to ER activation. However, when using mitogenesis as an endpoint, these co-activators were insufficient to provide a significant growth advantage. Combined, these data demonstrate that bisphenol A and coumestrol can impact ER activity and ER-dependent proliferation in breast cancer cells, but the influence of these agents is restricted to conditions of estrogen depletion, selective mutation of the ER, and expression of specific co-activators.

Introduction It has been well established that estrogen plays a pivotal role in both the development and progression of Breast cancer is the most commonly diagnosed malig- breast cancer [6,7]. Estrogen putatively promotes cancer nancy in women throughout the world, with an esti- progression by increasing the rate of cellular prolifera- mated 217,440 new cases diagnosed in the US during tion in breast epithelium [8]. The biological effects of 2004 [1]. Moreover, the incidence of breast cancer estrogen are mediated through binding and activating worldwide has been slowly rising in those age 50 and the cognate nuclear receptor, either estrogen receptor over for the past two decades [1–3]. Numerous genetic alpha (ERa) or (ERb) [9]. Once and lifestyle risk factors are associated with breast bound by estrogens, the active complex is able to initiate cancer, which include exposure to exogenous estrogens gene transcription by either directly binding estrogen (including synthetic and industrial estrogenic com- responsive elements within promoter regions of target pounds), obesity, and consumption [3]. The genes or by tethering to DNA bound transcription fac- most notable risk factors, however, are those related to tors [10]. Early stage breast cancer cells are dependent the reproductive cycle and estrogen hormone levels, on estrogen for proliferation and survival [11]. Thus, including early onset of menarche, delayed first full-term limiting ERa activity represents the major strategy for pregnancy, and late menopause [4]. Strikingly, the risk the prevention and treatment of ERa positive breast of developing and dying from breast cancer dramatically cancer [12,13]. For patients with ERa positive tumors, increases with age, as this disease is most prevalent in agents that either antagonize ER activity (e.g. selec- postmenopausal women over the age of 65 [5]. tive estrogen receptor modulators, SERMs, such as 280 JK Hess-Wilson et al. tamoxifen or ) or prevent estrogen synthesis [4,44,45]. Interestingly, many abundant environmental (e.g. aromatase inhibitors, AIs) are utilized clinically to and commercial estrogenic compounds, termed xenoes- block tumor growth. Tamoxifen, an ER antagonist in trogens (XEs), promote proliferation of breast cancer mammary tissue [14], binds the ER and alters the cells through inappropriate ER activation under condi- protein conformation such that the receptor is tions of hormone ablation [4,44–47]. In this study, we compromised for both co-activator recruitment and examined the xenoestrogenic compounds bisphenol A initiation of target gene transcription [15,16]. AIs (e.g. (BPA) and coumestrol (COU). The effect of these XEs anastrozole, letrozole, and exemestane) function by on activation of mutant ER alleles, mutant ER antag- blocking the action of enzymes responsible for con- onist action, or within the context of deregulated ER verting androgens to estrogens such as 17-b-estradiol co-activators has yet to be fully determined. BPA is a (E2), a high affinity ER ligand [17]. This treatment non-planar compound used in the production of poly- strategy is effective, as it reduces endogenous estradiol carbonate plastics (e.g. food containers, baby bottles, and thus limits ligand-dependent ER activation. and dental sealants) and is released from these materials Despite the success of SERM and AI treatment for in microgram quantities [48,49]. COU, one of the most ERa positive breast tumors, a large number of patients estrogenic phytoestrogens, is found in alfalfa, soya become refractory to therapy [17–19]. The mechanisms plant, and red clover [44,45,50]. Previous studies have leading to therapy relapse are likely diverse and at established that both BPA and COU are non-endoge- present are incompletely understood. Several genetic nous ligands for the ER, capable of stimulating mod- aberrations are linked to inappropriate ER activity and erate estrogen-independent proliferation in breast often render tumor cells therapy resistant, including ER cancer cells [46,47,51–53]. It is well documented that the loss, altered co-activator expression, and ER mutation binding of alternate ligands to ERa results in confor- [20]. ER alterations occur in 3–10% of breast cancer mational alterations in the receptor, thereby modifying patients, and several ER variants have been isolated the interaction of ER with co-regulatory proteins and from mammary tumors [21–23]. Several of these muta- with target estrogen responsive elements (EREs) tions alter the effect of ligand upon ER function, thereby [25,54–57]. Due to the potential biological consequences enhancing receptor response to low-affinity or alterna- of XE exposure during breast cancer initiation, pro- tive ligands [22]. A tumor-derived allele, ER-Y537S, gression, and treatment, it is imperative to fully under- encodes a constitutively active receptor, despite retain- stand how BPA and COU may modulate ER action. ing the ability to bind ligand [24–27]. A second tumor- The experiments described herein examine the impact derived allele, ER-D351Y, changes the estrogenic of BPA and COU under clinically relevant conditions activity of such as tamoxifen, rendering related to breast cancer. We confirm that while these agents agonists of ER [28,29]. xenoestrogens impact breast cancer cell growth, the ef- An additional mechanism by which breast cancer fects are largely confined to conditions of estrogen cells circumvent therapeutics is manifested by deregu- depletion, and therefore the stimulatory action of BPA lated expression of ER co-activators [30]. The ability of and COU on ER activity can be effectively countered by ER to transactivate target genes is known to require tamoxifen. However, we show that select tumor-derived recruitment of distinct co-activators that serve to alleles of ER respond differentially to BPA and COU, enhance receptor action. Several ER co-activators and tamoxifen enhanced the agonistic effects of the XEs expressed in breast cancer (e.g. cyclin D1, AIB-1, SRC- on the ER-D351Y mutant. Lastly, we show that co- 1, and TIF-2) are thought to enhance ER-dependent activators can selectively potentiate XE action in ER mitogenesis in breast tumors [10,30–35]. Cyclin D1, activation but not ER-dependent mitogenesis. Com- which is overexpressed in 30–50% of breast tumors, was bined, these studies establish the impact of BPA and originally demonstrated to activate the ER indepen- COU on ER-dependent signaling and mitogenesis in dent of ligand and to confer tamoxifen resistance breast cancer. [31,33,36,37]. However, the prognostic value of cyclin D1 overexpression with regards to tumor progression is controversial [38–40]. Additionally, enhanced expression Methods of SRC-1 has been shown to modulate ERa ligand uti- lization, allowing antagonists, including tamoxifen, to Plasmids function as agonists [10,41]. Expression levels of ERa have also been shown to correlate with TIF-2 expression Expression plasmids for pCMV5-hERa wild type throughout multiple stages of breast cancer [42]. estrogen receptor a and pCMV5-ERaY537S were gen- Moreover, AIB-1 expression has been specifically linked erous gifts from Dr. B. S. Katzenellenbogen (University to late stage cancers [30,43]. of Illinois, Urbana, Illinois). The expression plasmid While it is apparent that mutations in the ER and/or encoding ERa-D351Y (pSG5: D351Y) was kindly pro- overexpression of co-activators may contribute to the vided by Dr. V.C. Jordon (Northwestern University, bypass of therapeutic strategies, exposure to exogenous Chicago, Illinois). The estrogen responsive element estrogenic compounds has been shown to facilitate reporter plasmid, pBS3 3ÂERE/TATA/Luc, was a gift inappropriate ER activation and tumor progression from Dr. S. Khan (University of Cincinnati, Cincinnati, Xenoestrogens and breast cancer 281

OH). The CMV-b-galactosidase, pRc/CMV-cyclin D1, measured as an internal control for transfection effi- pcDNA3.1, and GFP-tagged histone H2B constructs ciency using Galacto-Star reagent (Tropix, Bedford, have been previously described [58]. The expression MA). For wtERa and ER-D351Y reporter assays, basal construct for pCMVTIF-2 was a generous gift from Dr. ER activity (vehicle-alone treated samples) was set to A. Puga (University of Cincinnati, Cincinnati, OH). ‘1’. Relative activity in response to steroid treatment is shown. In reporter assays examining the effect of BPA Cell culture and treatment and COU on tamoxifen-mediated wtERa or ER-D351Y function, activity of each ligand alone is set to ‘1’. Fold MCF-7 and HeLa cells were obtained from American change with the addition of tamoxifen is set relative to Type Culture Collection (Rockville, MD) and maintained each ligand. In reporter assays examining ligand con- at 37 C in a 5% CO2 humidified incubator. MCF-7 cells tribution to co-activator mediated activation, for each were cultured in Dulbecco’s modified Eagle’s medium condition empty vector was set to ‘1’ and changes in (DMEM) supplemented with 5% heat-inactivated fetal activation with the addition of each co-activator set bovine serum (DFBS; Atlanta Biologhicals, Norcross, relative to empty vector. Averages and standard devia- GA), 100 units/ml penicillin/streptomycin (Mediatech, tions from at least six independent experiments are de- Herndon, VA), and 2 mM L-glutamine. HeLa cells were picted. Appropriate p-values were obtained using cultured in DMEM supplemented with 10% FBS ANOVA followed by a Newman–Kuels Multiple plus 100-units/ml penicillin/streptomycin and 2 mM Comparison post-test. L-glutamine. For steroid-depleted conditions, 5 or 10% charcoal-dextran treated FBS (CDT; Hyclone Labora- BrdU Incorporation and cell growth assay tories, Logan UT) and phenol red-free DMEM was utilized. MCF-7 cells were seeded in 5% CDT/DMEM in six- well dishes with a concentration of 5Â105 cells per well. Chemicals and reagents These cells were propagated in the absence of steroid for 48 h, and then stimulated with the indicated treatments Estradiol (E2), tamoxifen, bisphenol A (BPA; 4,4¢-iso- for 48 h. Cells were labeled with bromodeoxyuridine propylidenediphenol), and coumestrol (COU) were (BrdU) the last 16 h of treatment. For experiments obtained from Sigma-Aldrich (St. Louis, MO). All involving overexpressed co-activators, MCF-7 cells were reagents were solubilized in ethanol and utilized at the transfected the day after being seeded in 5% CDT indicated concentrations. DMEM using the BES/calcium phosphate protocol with H2B-GFP (0.25 lg) (control for transfection efficiency). Transfection and transcriptional reporter assays A total of 2 lg of co-activator expression plasmid was used. Parental vector was added to bring each trans- HeLa cells were seeded at 8Â104 cells per well of a six- fection to a total of 4 lg DNA. Following transfection, well plate in the absence of steroid hormones (10% CDT the media was replaced and cells allowed to recover for DMEM) and transfected the following day with 4 lgof 24 h. MCF-7 cells were then stimulated with either 0.1% DNA using the BES/calcium phosphate transfection ethanol vehicle control, E2, BPA, or COU (as indicated) protocol (N,N-bis(2-hydroxyethyl)-2-amino-ethanesulf- for 48 h. Transfected cells were pulsed for the last 18 h onic acid-buffered saline/calcium phosphate) as previ- with bromodeoxyuridine (BrdU; Amersham Bioscienc- ously described [59]. For reporter assays utilizing wtERa es), fixed, and stained as previously described [58]. The or ER-Y537S, cells were co-transfected with pCMV5- percentage of cells proliferating was determined by hERa (0.25 lg), pBS3-3ÂERE/TATA/Luc reporter scoring transfected cells (GFP positive) for BrdU construct (1.1 lg), and CMV-b-galactosidase (0.25 lg). incorporation. All BrdU experiments were performed at For reporter assays with ER co-activators, 0.75 lgof least in triplicate with at least 200 cells per treatment plasmid DNA encoding each co-activator (pRc/CMV- scored for incorporation. Averages and standard devi- cyclin D1 or pCMVTIF-2) was also added. For reporter ations are shown. Appropriate p-values were obtained assays using ER-D351Y mutant allele, cells were using ANOVA followed by Newman–Kuels Multiple co-transfected with pSG5:D351Y (0.25 lg), pBS3- Comparison post-test. 3ÂERE/TATA/Luc (1.1 lg), and TK-Renilla (0.25 lg) To assay cell growth, approximately 4Â105 MCF-7 as an internal control for transfection efficiency. In all cells were seeded in duplicate wells of a six-well dish in transfections, parental vector was used to bring the total 5% CDT medium supplemented with the indicated transfection content to 4 lg. Following overnight treatment (10)8 M E2, 10)6 M BPA, 10)8 M COU, or transfection, cells were washed with phosphate buffered 0.1% EtOH vehicle). Cells were then trypsinized and saline four times. Cells were then stimulated with hor- counted using trypan blue exclusion after 24, 72, 168, mone (17-b-estradiol, bisphenol A, coumestrol, tamox- and 216 h treatments. Media containing the indicated ifen) or 0.1% ethanol vehicle as indicated for 48 h. Cells treatments was replenished every 72 h and experiments were then harvested, lysed, and monitored for luciferase were performed at least twice in triplicate. Doubling activity following the manufacturer’s protocol. For the time was determined using the following equation: Promega luciferase assays, b-galactosidase activity was dt=t*ln2/ln (Ct–Co), where t=time interval in hours, 282 JK Hess-Wilson et al.

Ct=final cell count and Co=initial cell count. Data compounds, and BPA is a non-planar compound with shown reflect the averages and standard deviations. two phenolic rings. To define the mitogenic potential of BPA and COU in the presence of estradiol, the MCF-7 model system was utilized. These cells were derived from Results human invasive ductal carcinoma [60], express wild type ERa and exhibit retarded growth upon estrogen with- The mitogenic action of xenoestrogens is restricted to drawal [61,62]. To verify the mitogenic action of these conditions of estrogen deprivation compounds singularly, asynchronously proliferating MCF-7 cells were cultured in steroid-free medium Xenoestrogenic compounds have been shown to induce (DMEM+5%CDT) for 48 h to deplete residual steroid estrogen-independent mitogenesis in breast cancer cells; hormones, then supplemented either with E2 (10)11– however, these compounds are numerous and structur- 10)5 M), BPA (10)11–10)6 M), COU (10)11–10)5 M) or ally diverse [4,44,45]. As a result, it has been proposed 0.1% EtOH vehicle. Doses utilized encompass the that these agents may contribute to inappropriate breast known human exposure range for BPA and phytoes- cancer proliferation in patients undergoing hormone trogens such as COU (both within the nM range) ablation therapy. The impact of these agents on breast [49,63–66]. Following 48 h of treatment with each agent, cancer progression in the presence of estrogen (as ob- MCF-7 cells were labeled with bromodeoxyuridine served in patients treated with ER antagonists) or in the (BrdU) for approximately 18 h, at which time cells were presence of low levels of estrogen (as observed fixed and BrdU incorporation was detected via indirect in patients treated with aromatase inhibitors) remained immunofluorescence. As shown in Figure 1b, MCF-7 largely unresolved. As such, we chose to examine cells cultured in EtOH vehicle (estrogen ablation) the structurally divergent compounds bisphenol A exhibited a low rate of BrdU incorporation (15%), (BPA) and coumestrol (COU) (Figure 1a). As shown consistent with the dependence of these cells on E2 for in Figure 1a, COU and estradiol are both planar robust proliferation [61,62]. As expected, addition of E2

Figure 1. The mitogenic action of BPA and COU is restricted to conditions of estrogen depletion. (a) Structure of 17-b-estradiol (E2) compared to the xenoestrogens bisphenol A (BPA) and coumestrol (COU). (b) MCF-7 cells were propagated for 48 hrs in 5% CDT then supplemented for 48 h with either 0.1% EtOH vehicle, increasing E2 concentrations (10)11–10)5 M), increasing BPA concentrations (10)11–10)5 M), or increasing COU concentrations (10)11–10)5 M). Cells were then labeled with BrdU as described, and BrdU incorporation was detected via indirect immunofluorescence. Data shown are the average of at least three independent experiments in which at least 200 cells/experiment were analyzed. (c) MCF-7 cells cultured for 48 h in 5% CDT were then supplemented with either 0.1% EtOH, 10)8 M E2, 10)6 M BPA or 10)8 M COU. Cell number was analyzed at the indicated times by counting and trypan blue exclusion. Data shown are the averages and standard deviations of three independent experiments performed in triplicate. (d) MCF-7 cells were cultured as described for A, then supplemented with either 0.1% EtOH vehicle, 10)11 M E2, increasing BPA (10)11–10)5 M) plus 10)11 M E2, or increasing COU (10)11–10)5 M) plus 10)11 M E2. After BrdU labeling, cells were scored for BrdU incorporation via indirect immunofluorescence. Xenoestrogens and breast cancer 283 significantly induced MCF-7 proliferation with maximal COU to cooperate with E2 for ER activation. To mitogenesis occurring at 10)8 M E2 (65% BrdU posi- delineate the impact of BPA and COU on E2-induced tive). Consistent with previous findings, both BPA and gene expression, HeLa cells were transiently transfected COU stimulated MCF-7 cell cycle progression in the in the absence of steroid hormone with expression absence of E2, with maximal proliferation seen at plasmids encoding wtERa, b-galactosidase (as an inter- ) ) 10 6 M BPA (65% BrdU positive) and 10 8 MCOU nal control for transfection efficiency), and the 3ÂERE- (55% BrdU positive) [47,51,52]. Moreover, a dose TATA/Luc-reporter gene construct. Post transfection, dependent, biphasic effect of E2, BPA and COU on cells were stimulated with the indicated ligands for 48 h, cellular proliferation was observed. This non-linear re- at which time cells were harvested and analyzed for both sponse is often exhibited by hormone dependent cells luciferase and b-galactosidase activity. As expected, after exposure to estrogenic agents (reviewed in [67]). treatment with 10)8 M E2 resulted in an approximate To verify that the observed mitogenic concentrations 10-fold increase in transactivation of the reporter gene of BPA or COU (10)6 M and 10)8 M, respectively) re- over vehicle control (Figure 2a). This is consistent with sulted in cellular proliferation, cell number was moni- previously published data utilizing this model reporter tored as a function of time. Asynchronous MCF-7 cells system [69]. Also, consistent with the literature, 10)6 M were cultured in the presence of indicated compounds, BPA and 10)8 M COU stimulated modest, statistically as previously described (Figure 1b), and replenished significant (p<0.05 compared to EtOH treatment) every 72 h. Subsequently, cell number was quantified by activation of the reporter construct (3 and 4 fold in- counting following 24, 72, 168, and 216 h of culture in crease respectively, Figure 2a). Similar results were each condition through trypan blue exclusion. Consis- achieved using different estrogen response element tent with our findings and previously published data, (ERE) reporter constructs, (e.g. pS2-Luc and C3-Luc, MCF-7 cells cultured in the absence of steroid hormones data not shown) indicating that BPA and COU action is doubled approximately every 47 h [68]. As expected, conserved on multiple promoters. Thus, these data 10)8 M E2 reduced the doubling time to 26 h (Fig- confirm that like E2, both BPA and COU stimulate ERa ure 1c). In addition, as expected, 10)6 M BPA added to activity at physiologically relevant doses that stimulate MCF-7 culture medium resulted in a doubling time of mitogenesis in breast cancer cells. approximately 29 h. Similarly, MCF-7 cells supple- To determine whether BPA and COU can cooperate mented with 10)8 M COU doubled in 28 h. (Figure 1c) with E2 to enhance ERa transactivation in our system, Thus, 10)6 M BPA and 10)8 M COU stimulate both cell reporter experiments were repeated in the presence of cycle progression and cellular proliferation under con- limiting E2 plus a range of doses for either BPA or COU ditions of estrogen deprivation. (Figure 2b). Post transfection, cells were treated with To examine potential cooperation between the either the dose range for BPA and COU alone, or xenoestrogens and estradiol, MCF-7 cells were initially concomitant with limiting E2. Exposure to 10)14 ME2 cultured in steroid-depleted conditions for 48 h at which induced approximately 3.5 fold activation of the ER time suboptimal E2 (10)11 M; as determined by the dose reporter construct over vehicle control (p<0.001 com- response curve in Figure 1b) was supplemented with pared to EtOH vehicle, Figure 2b). Additionally, acti- either BPA (10)11–10)5 M) or COU (10)11–10)5 M). vation of wtERa with varying doses of BPA or COU Treatments continued for 48 h, at which time cells were (10)11,10)8 and 10)6) produced a dose dependent in- pulsed overnight with BrdU. BrdU incorporation was crease in reporter activation. At all doses tested, BPA monitored via indirect immunofluorescence and plotted and COU failed to enhance the effect of E2 on ERa as percent BrdU incorporation. As can be seen in Fig- function, and addition of these agents actually decreased ure 1d, only 20% of cells supplemented with vehicle ER activity in the presence of E2 (Figure 2b). These control entered S-phase, and addition of E2 (10)11 M) observations are likely attributed to competition stimulated increased proliferation (40%). Noticeably, between low dose E2 and the weaker agonists. Together, although BPA and COU stimulate proliferation in the these data demonstrate that the estrogenic action of absence of E2 (Figure 1b, c), these agents failed to BPA and COU is manifest only under conditions of enhance the proliferative effect of limiting E2 (Fig- estrogen deprivation. ure 1d). Taken together, these data verify that although BPA and COU stimulate mitogenesis of breast cancer BPA and COU elicit disparate responses in tumor-derived cells, the mitogenic potential of these xenoestrogens is ER mutants restricted to conditions of estrogen depletion. As discussed above, mutations in the ER have been re- Estrogen receptor agonist activity of BPA and COU fails ported to occur during tumor progression, and several of to synergize with estradiol these mutations alter the effect of ligand on ER function [20,24,25,70,71]. Moreover, it has been observed that While active ER is known to promote cell cycle pro- planar (e.g. COU) and non-planar (e.g. BPA) com- gression in breast cancer cells, retention of ER status is a pounds have different mechanisms of estrogen action, favorable prognostic indicator of disease outcome and therefore may be differentially utilized by ER mu- [11,13]. Therefore, we examined the ability of BPA and tant with alternate molecular conformations [46]. To 284 JK Hess-Wilson et al.

Figure 2. Estrogen receptor agonist activity of BPA and COU fails to synergize with estradiol. (a) HeLa cells were transfected in the absence of steroid with plasmids encoding wtERa, b-galactosidase, and the 3ÂERE/TATA-Luc reporter. After transfection, cells were treated with either 0.1% EtOH, 10)8 M E2, 10)6 M BPA or 10)8 M COU for 48 h. Cells were then harvested, lysed and analyzed for luciferase activity (to monitor ER activity) and b-galactosidase activity (to normalize for transfection efficiency). Relative luciferase activity detected in EtOH-treated cells was set to 1. Experiments were performed on three independent experiments performed in triplicate. Averages and standard deviation are shown. *** indicates a p value of <0.001 (E2) and * denotes p<0.05 (BPA and COU) comparing each treatment to EtOH. (b) HeLa cells were cultured as described for A and treated with either 0.1% EtOH, 10)14 M E2, increasing doses of BPA (10)11–10)6 M) alone and plus 10)14 M E2, or increasing doses of COU (10)11–10)6 M) alone and plus 10)14 M E2. Cells were harvested, lysed and analyzed as described in A. Experiments were performed at least in triplicate and averages and standard deviations are shown. ** indicates p<0.01 for E2 10)14 M compared to EtOH alone. determine if binding of BPA or COU alters the trans- Amino acid 351 is important for SERM activity and activation potential of tumor-derived ER mutants, re- upon point mutation (D351Y) renders the receptor porter analyses were performed as described using the responsive to tamoxifen [28,29,72]. Given the disparate indicated expression plasmid and the 3ÂERE/TATA- response to ligand of this tumor-derived ER mutant Luc reporter. As expected, the ER-Y537S mutant dem- compared to wtERa, we examined the ability of BPA onstrated increased activity as compared to wtERa in the and COU to stimulate receptor activity using the absence of ligand (EtOH vehicle), and E2 failed to sig- 3ÂERE/TAT-Luc reporter construct as a read out for nificantly increase receptor activity (Figure 3a), consis- transcriptional activation. ER-D351Y showed an ) tent with the reported constitutive activity of this ER approximate 17 fold induction after 10 8 M E2 stimu- mutant [25]. Similar responses were observed with lation (Figure 3b; p<0.001 over EtOH) as compared to ) ) 10 6 M BPA or 10 8 M COU, thus indicating that BPA the 8 fold activation of the wtERa observed in and COU do not alter the constitutively active nature of Figure 2a under identical conditions (p<0.05 compar- the ER-Y537S mutant. ing E2 activation of ER-D351Y to wtERa). After Next, we examined the transcriptional activity of the 10)6 M BPA exposure ER-D351Y activity was stimu- ER-D351Y mutant in the presence of BPA and COU. lated 7 fold over EtOH control (p<0.05), whereas Xenoestrogens and breast cancer 285

Figure 3. BPA and COU elicit disparate responses in tumor-derived ER mutants. (a) HeLa cells were transfected as described in Figure 2, with either expression plasmids encoding wtERa or ER-Y537S. Cells transfected with wtERa were treated with 0.1% EtOH vehicle, while cells transfected with ER-Y537S were then treated with either 0.1% EtOH, 10)8M E2, 10)6M BPA or 10)8M COU for 48 h and monitored for luciferase and b-galactosidase activity as described. ER-Y537S activity was set relative to wtERa EtOH treated sample. Data shown represent averages and standard deviations from at least three independent experiments performed in triplicate. (b) HeLa cells were transfected in steroid free conditions with plasmids encoding ER-D351Y, TK-Renilla, and the 3ÂERE/TATA-Luc reporter. After transfection, cells were harvested, lysed and analyzed for renilla-luciferase activity (to normalize for transfection efficiency) or luciferase activity (to monitor ER-D351Y activity). Experiments were performed at least in triplicate and averages and standard deviations are shown. *** indicates a p value of <0.001 comparing E2 treatment to EtOH alone. * indicates p<0.05 comparing BPA or COU to EtOH treatment alone. + indicates a p value of <0.05 comparing BPA to COU treatment. wtERa was stimulated only 4 fold under identical tamoxifen reduced E2-mediated wtERa activity 0.5 fold conditions. Thus, ER-D351Y maintains an enhanced (Figure 4b, left panel), and similar results were observed agonist activity for BPA as compared to wtERa after BPA and COU stimulation (0.5 and 0.7 fold, (p<0.05). Interestingly, this mutant demonstrated a respectively, Figure 4b, left panel). The antagonistic ef- similar response to 10)8 M COU as wtERa (approxi- fects of tamoxifen were also conserved in the presence of mately 4 fold induction for both, with no statistical each xenoestrogen using a higher, more clinically rele- distinction between the alleles). Thus, these data show vant dose of tamoxifen (10)6 M) (Figure 4b, right pa- that the ER-D351Y mutant allele demonstrates an nel). Combined, these data confirm that tamoxifen is an enhanced transactivational response to E2 and BPA as effective antagonist of XE action on wtERa at all con- compared to wtERa. centrations examined. By contrast, tamoxifen exhibited a cooperative effect Tamoxifen cooperates with xenoestrogens for activation with select xenoestrogens for activation of the ER- of ER-D351Y D351Y mutant. As shown, low dose tamoxifen signifi- cantly enhanced COU-mediated ER-D351Y activation Tamoxifen acts as an agonist for the tumor derived ER- by approximately 50% (Figure 4b, p<0.05, left panel), D351Y mutant [28]. Thus, the response of this ER mu- thus indicating that these two agents may cooperate. No tant to tamoxifen was examined in the presence of BPA cooperative effect was observed after BPA or E2 sti- and COU. For these studies, cells were transfected as in mulation. However, with concentrations of tamoxifen Figure 2a with expression plasmids encoding wtERa or similar to those used in breast cancer management, ER-D351Y and the 3ÂERE/TATA-Luc reporter con- significant cooperation of tamoxifen with each estro- struct, and subsequently stimulated with the indicated genic agent was observed (Figure 4b, right panel). Thus, treatments (E2, BPA or COU plus or minus a low and the stimulatory effect of these compounds may enhance high dose of tamoxifen). Initially, a dose response of the agonistic effect of tamoxifen on clinically relevant tamoxifen alone on each receptor was analyzed. EtOH ER mutants. The effects of BPA and COU on ER activity was set to ‘1’ for both wtERa and ER-D351Y activation, tamoxifen response and cooperation with and relative luciferase activity after addition of either estradiol are summarized in Table 1. 10)6 Mor10)8 M tamoxifen is plotted in Figure 4a. Consistent with the literature, ER-D351Y activity was ER co-activators commonly expressed in breast cancer significantly stimulated by low dose 10)8 M tamoxifen facilitate disparate ER activation by BPA and COU (1.5 fold induction over vehicle alone, p<0.01, Figure 4a) [28]. The agonistic properties of tamoxifen The ER-dependent mitogenic stimulation of E2 in vivo is were dose dependent, as ER-D351Y activity dramatically enhanced through the action of ER co-activators, several increased at higher doses (10)6 M, Figure 4a, p<0.001) of which are aberrantly expressed in breast cancer [30]. to an approximate 25-fold increase in receptor activity. Expression of these proteins (e.g. AIB-1, TIF-2 and cyclin To examine the response of ER-D351Y to tamoxifen D1) is thought to further enhance ER activation and in the presence of BPA and COU, ER activation for provide a mitogenic advantage for tumor cells. Although each ligand was set to ‘1’ and the fold change induced by these co-activators have been previously shown to tamoxifen is plotted in Figure 4b. As expected, 10)8 M potentiate ER activity in the presence of E2 [35,73,74], 286 JK Hess-Wilson et al.

Table 1. The effect of BPA and COU on ER activation, tamoxifen determine the impact of ER co-activator overexpression response and cooperation with estradiol on ER responsiveness to BPA and COU, cells were co- transfected as described for Figure 2a with expression E2 BPA COU plasmids encoding wtERa, the 3ÂERE/TAT-Luc re- ER activation wt ERa ›› › › porter gene and b-galactosidase (as an internal control) in Y537S – – – addition to either cyclin D1 or TIF-2. Impact of each co- D351 Y ››› ›› › activator was calculated as the fold increase over vector ) 10 8 M Tam response wt ERa flflfl control. As shown, only TIF-2 enhanced the activity of D351 Y – – › wtERa over vector control in the absence of ligand ) 10 6 M Tam response wt ERa flflfl (approximately 2 fold induction over basal activity, D351 Y ›› › › p<0.05, Figure 5 top left panel). However, E2 signifi- E2 cooperation ER activity N/A – – cantly enhanced wtERa activity in the presence of both Mitogenesis N/A – – co-activators (Figure 5 top right panel), thus confirming the efficacy of each ER co-activator. Interestingly, BPA ‘›’=slight increases, ‘››’=modest increases, ‘›››’=strong increases, was refractory to potentiation by cyclin D1 expression, ‘fl’=slight decreases, ‘–’ =no effect, ‘N/A’=not applicable. but the agonistic properties of BPA were significantly enhanced by TIF-2 (approximately 2.5 fold, bottom left panel, p<0.05 compared to pcDNA). By contrast, COU- their actions in the presence of XEs remain largely un- mediated ER activation was refractory to TIF2 expres- characterized. Such analyses are crucial, as it has been sion, but was enhanced by the overexpression of cyclin D1 shown that ERa demonstrates ligand-dependent differ- (bottom right panel, p<0.01 compared to pcDNA). To- ences in co-activator binding and recruitment [54]. To gether, these data indicate that co-activators commonly

Figure 4. Tamoxifen cooperates with xenoestrogens for activation of ER-D351Y. (a) HeLa cells were transfected as described above for either wtERa or ER-D351Y. Transfected cells were then treated for 48 h with 0.1% EtOH vehicle, plus 10)8 Mor10)6 M tamoxifen. Reporter analysis was then performed as described previously. EtOH activity for each receptor was set to ‘1’. * indicates a p value of p<0.05 for ER-D351Y treated with 10)8 M tamoxifen compared to EtOH. *** indicated a p value of p<0.001 for ER-D351Y treated with 10)6 M tamoxifen compared to EtOH alone. (b) HeLa cells were transfected as described for ‘a’ and then treated with either, 10)8 M E2, 10)6 M BPA, 10)8 M COU or these same concentrations plus 10)6 Mor10)8 M tamoxifen. Activity of each receptor with each ligand alone is set to one and fold chance with 10)8 Mor 10)6 M tamoxifen is plotted. Averages and standard deviations are shown. * indicates a p value of p<0.05 for ER-D351Y treated with COU plus 10)8 M tamoxifen compared to COU alone (left panel). ** indicates a p value of p<0.01 for E2 plus 10)6 M tamoxifen compared to the activity of E2 alone on ER-D351Y. * indicates p<0.05 for BPA plus 10)6 M tamoxifen and COU plus 10)6 M tamoxifen compared to the activity of these ligands on ER-D351Y alone. Xenoestrogens and breast cancer 287

Figure 5. Overexpressed ER co-activators differentially potentiate the agonist functions of BPA and COU. HeLa cells were transfected in the absence of steroid with plasmids encoding wtERa, the co-activator cyclin D1 or TIF-2, b-galactosidase, and 3ÂERE/TATA-Luc. After trans- fection, cells were treated with either 0.1% EtOH vehicle, 10)14 M E2, 10)6 M BPA or 10)8 M COU for 48 h then harvested, lysed and analyzed for luciferase activity to monitor ER activity and b-galactosidase activity to normalize for transfection efficiency. ER activity in the presence of each ligand without co-activators (empty pcDNA vector control) was set to 1. The fold change resulting from addition of overexpressed co- activators was set relative to vector control for each ligand. Averages and standard deviation from at least three independent experiments performed in triplicate are shown. * indicates p<0.05 for cyclin D1 compared to pcDNA in the presence of E2 and TIF-2 enhanced activation over pcDNA control in the presence of both EtOH and BPA while ** indicates p<0.01 for the increased activation mediated by TIF-2 in the presence of E2 and cyclin D1 compared to pcDNA in the presence of COU. expressed in breast cancer differentially regulate the ac- cycle (through association with cyclin dependent kinase tions of BPA and COU as ER agonists. 4 or 6) [75]. In most model systems, however, it has been Given the variant activities of BPA and COU in re- reported that overexpression of cyclin D1 alone is sponse to ER co-activators, we analyzed the impact of insufficient to advance cell cycle progression [75]. Con- each co-activator on xenoestrogen-dependent mitogen- sistent with this hypothesis, cyclin D1 failed to increase esis of human breast cancer cells. For these experiments, the mitogenic potential of both xenoestrogens. Overex- MCF-7 cells were transiently transfected with expression pression of TIF2 also failed to change the mitogenic plasmids encoding the indicated co-activators (TIF2 or potential of either BPA or COU. These data indicate cyclin D1) and a marker for transfection efficiency that while overexpression of ER co-activators is suffi- (histone H2B fused to GFP). Following transfection, cient to alter xenoestrogen-mediated ER function, these cells were steroid depleted for 48 h then stimulated with individual events fail to alter the proliferative potential either 0.1% EtOH vehicle, 10)8 M E2, 10)6 M BPA or of breast cancer cells, in the presence of either XE or 10)8 M COU for 48 h. During the last 18 h of treat- estradiol. Table 2 summarizes these data comparing the ment, cells were labeled with BrdU to monitor S-phase effect of BPA and COU on co-activator mediated ER progression. GFP positive cells were scored for BrdU activity and mitogenesis. Thus, there is a distinction incorporation. Results are depicted as fold increase in between co-activator mediated ER activation and proliferation over vector alone for each ligand. Over- co-activator mediated cellular proliferation. expression of TIF-2 had no measurable impact on cel- lular proliferation, whereas cyclin D1 significantly increased the proliferative index by 50% in the presence Discussion of vehicle (EtOH), p<0.05 compared to pcDNA (Figure 6). While this action of cyclin D1 may be par- The incidence of breast cancer in postmenopausal wo- tially attributed to its ER co-activation potential, it men has continued to rise over the past decade [1]. This should also be noted that cyclin D1 itself impacts the cell parallels the increase in human exposure to estrogen 288 JK Hess-Wilson et al.

Figure 6. Proliferative effects of xenoestrogens are refractory to co-activator de-regulation. MCF-7 cells were transfected with expression plasmids encoding histone H2B tagged with GFP and cyclin D1, TIF-2 or pcDNA vector control. Cells were propagated in absence of steroid for 24 h then stimulated with either 0.1% EtOH, 10)8 M E2, 10)6 M BPA or 10)8 M COU for 48 h. Transfected cells were labeled for BrdU incorporation as described. Positively transfected cells (GFP positive) were scored for BrdU incorporation by indirect immunofluorescence. The fold change in BrdU incorporation mediated by overexpressed co-activators was set relative to vector control for each ligand. Experiments were performed at least in triplicate. Averages and standard deviations are shown. * denotes p<0.05 for cyclin D1 BrdU incorporation compared to pcDNA.

Table 2. Comparison of BPA and COU on Cyclin D1 and TIF-2 transactivational responses to BPA and COU (Fig- mediated ER activity and mitogenesis ure 3), thus suggesting that mutations in the ER during tumor progression may alter the response to xenoes- No hormone E2 BPA COU trogens. Further highlighting the importance of these Cyclin D1 ER activity – › – ›› findings, tamoxifen was shown to potentiate the ago- Mitogenesis › –– – nistic effect of both xenoestrogens in the presence of ER- TIF-2 ER activity ›› ›› ›› – D351Y (Figure 4). Lastly, BPA and COU effects were Mitogenesis – – – – monitored under conditions of co-activator deregula- tion. In these studies, BPA and COU exhibited differ- ‘›’=slight increases, ‘››’=modest increases, ‘–’=no effect. ential responses to co-activators with regard to ER activation, however they failed to provide a significant mimics, termed xenoestrogens (XE), found in , mitogenic advantage in the presence of overexpressed pesticides, pharmaceuticals and industrial by-products. co-activators in breast cancer cells (Figures 5 and 6). The experiments described herein aim to examine the These data highlight a disparity between co-activator effect of two highly prevalent XEs, bisphenol A (BPA) functions in ER regulation versus ER-dependent mito- and coumestrol (COU), within physiologically relevant genesis. Combined, these data specify that the influence conditions related to breast cancer progression and of BPA and COU on ER activation and ER-dependent treatment. The data presented indicate that the ability of proliferation in breast cancer cells is limited to condi- these agents to induce mitogenesis is restricted to con- tions of estrogen depletion, selective mutation of the ditions of estrogen deprivation, as these agents were ER, and overexpression of selected co-activators. unable to cooperate with estradiol (E2) for mitogenesis While the influence of both BPA and COU on of breast cancer cells (Figures 1 and 2). XE-mediated reproductive development and fertility has been exten- wtERa activation was blocked by tamoxifen (Figure 4), sively investigated, studies of the impact of these XEs on indicating that these agents are not likely to disrupt breast cancer growth and progression have been limited. tamoxifen-based therapeutics. However, a tumor-de- It is known that human exposure to both BPA and rived mutant ER (ER-D351Y) demonstrated disparate COU can be significant. Over 800 million kg of the Xenoestrogens and breast cancer 289 plasticizer BPA is generated annually in the United ER activity (the presumptive mechanism of mitogenic States alone [76] and human exposure to this compound stimulation) is effectively antagonized by tamoxifen in largely occurs through leaching from canned food lin- the presence of wild-type receptor, thus indicating that ings and polycarbonate plastics [48]. Few studies have tamoxifen therapy is unlikely to be effected by BPA or extensively analyzed the accumulation of BPA; however COU exposure. the serum concentrations in adult females have been While therapeutic regimens utilize ER selective reported to range between 0.1 and 2 nM [49,63]. Con- antagonists (SERMs, e.g. tamoxifen) or agents to block sumption of the phytoestrogen COU occurs through ER ligand synthesis (e.g. AIs), failure of these treatment ingestions of red clover, soy plant and alfalfa [45,50] and modalities ultimately occurs in a significant number of diets rich in these plants have reportedly led to sera clinical cases. Approximately one-half of patients with levels approaching 5 lM in range [64–66,77]. It should ER positive breast tumors will relapse with recurrent be noted that phytoestrogens such as COU exist mainly tumors retaining ER expression but that are refractory as , and their bioavailability requires hydro- to therapy [18]. Of these tumors, somatic point muta- lysis of the sugar moiety by b-glucosidases [78]. More- tions in ER have been reported (3–10%), which may over, only a small amount of this free aglycone form has alter the response of ER to ligand [20]. For example, the been detected in human blood, therefore exact concen- ER-Y537S tumor derived allele is constitutively active, tration of bioactive phytoestrogens in humans has not no longer requiring ligand to stimulate transactivation been accurately defined [78]. However, Zubik et al. [79] (although this allele is still able to bind ligand) reported an estimated level of aglycones in humans be- [24,25,27]. As shown herein, BPA and COU failed to tween 2–40 mg per day dependent on diet and gut mi- impact ER-Y537S activity, similar to the response of croflora composition. As COU makes up only a small this allele to E2. In contrast, the ER-D351Y tumor-de- fraction of known phytoestrogens, the actual exposure rived allele demonstrated a heightened response to E2 to active COU represents only a portion of this total. compared to wtERa, consistent with previous reports Future studies directed at assessing levels of bioavailable [95]. Additionally, ER-D351Y demonstrated an en- BPA and COU will assist determining the relative im- hanced response to BPA but not COU, thus indicating pact of each compound on receptor activity and cancer that this mutant receptor is selectively responsive to cell growth. specific XEs. The effect of BPA and COU on the pro- Both BPA and COU harbor estrogenic activity and liferation of breast cancer cells expressing mutant ERs are weak agonists of ERa and ERb, with binding has yet to be determined, as no model system exists to affinities that are approximately 1000-fold (BPA) and sufficiently test this endpoint. However, ER-D351Y is 10-fold (COU) less than estradiol [52,80]. Moreover, known to be activated by tamoxifen exposure [28,29], these agents induce estrogen-independent mitogenesis in and this altered ligand response is proposed to con- cultured breast cancer cells [48,51,81–85]. Given the tribute to therapeutic relapse. Therefore, the ability of importance of hormone therapy and ER antagonists in XE to enhance the agonistic effects of tamoxifen were breast cancer therapy [11,86–88], analysis of XE action assessed (Figure 4). We demonstrate that while low dose using physiologically relevant concentrations of BPA tamoxifen (10–8 M) failed to synergize with BPA and E2, and COU in the context of conditions related to breast this dose of tamoxifen significantly enhanced ER- cancer treatment and progression is well justified. It has D351Y activation by COU. By contrast, a more clini- been reported that BPA and COU stimulate estrogen- cally relevant dose of tamoxifen (10)6 M) sensitized the independent proliferation of breast carcinoma cells ER-D351Y mutant to each estrogenic agent (E2, BPA [44,46,48]. Our analyses demonstrated that the mito- and COU), thus indicating that the cooperative effects genic capacity of both agents fell within the range of may be dose dependent. The effect of BPA and COU on human exposure (1 lM for BPA and 10 nM for COU) ER activation and tamoxifen response comparing [49,63–66,77]. Interestingly, high doses of BPA and wtERa and select tumor-derived mutant alleles is sum- COU (10 lM) failed to induce mitogenesis, indicating marized in Table 1. Combined, the data presented that XE elicit the described ‘biphasic’ dose response on indicate that XEs may harbor alternate effects on proliferation (reviewed in [67]). However, as shown tumor-derived alleles of the ER, and that XEs may summarized in Table 1, the mitogenic action of BPA change the response to tamoxifen therapy in this context and COU was observed only under conditions of of tumor-derived mutant ER expression. estrogen deprivation, as no cooperative effects were Lastly, we analyzed the impact of the XEs under observed in the presence of 17-b-estradiol (E2). In breast conditions of deregulated ER co-activator expression, cancer therapy, estrogen depletion is achieved by the use which is known to occur in breast cancer progression. of aromatase inhibitors (AIs, e.g. anastrozole, letrozole, These studies may hold particular clinical significance, exemestane) in postmenopausal women. These agents as ER is known to selectively utilize co-activators, reduce E2 levels by 80–95%, and are highly effective dependent on ligand bound [54]. Cyclin D1 is overex- therapeutics [19,89–94]. The present data suggest that pressed in 30–50% of breast cancers [31,36,75], while exposure to xenoestrogenic agents may circumvent such TIF-2 expression tightly correlates with disease stage estrogen ablation therapies. However, these data also [42]. The data shown reveal that the co-activators cyclin confirm that the ability of BPA and COU to stimulate D1 and TIF-2 differentially regulate the actions of BPA 290 JK Hess-Wilson et al. and COU when ER activity was utilized as an end- 3. Key TJ, Verkasalo PK, Banks E: Epidemiology of breast cancer. point. Specifically, the agonsistic properties of BPA Lancet Oncol 2(3): 133–40, 2001 were selectively enhanced by TIF-2, whereas COU 4. Coyle YM: The effect of environment on breast cancer risk. Breast Cancer Res Treat 84(3): 273–88, 2004 mediated Cyclin D1 facilitated ER activation. How- 5. Tilson-Mallett N, Santner SJ, Feil PD, Santen RJ: Biological ever, the potentiating effects of each co-activator were significance of aromatase activity in human breast tumors. J Clin not recapitulated when proliferation was assessed as Endocrinol Metab 57(6): 1125–8, 1983 the endpoint. As is summarized and highlighted in 6. Huff J, Boyd J, Barrett JC: Hormonal carcinogenesis and envi- Table 2, there is a disparity between co-activator ronmental influences: background and overview. Prog Clin Biol Res 394: 3–23, 1996 function with regard to ER activation and ER-depen- 7. Davis DL, Pongsiri MJ, Wolff M: Recent developments on the dent mitogenesis. avoidable causes of breast cancer. Ann N Y Acad Sci 837: 513–23, In summary, the data presented define the action of 1997 two prevalent xenoestrogens, BPA and COU, in condi- 8. Platet N, Cathiard AM, Gleizes M, Garcia M: Estrogens and their tions clinically relevant for breast cancer progression receptors in breast cancer progression: a dual role in cancer pro- liferation and invasion. Crit Rev Oncol Hematol 51(1): 55–67, 2004 and management. The data confirm that in the presence 9. McDonnell DP: The molecular determinants of estrogen receptor of wtERa, BPA and COU effects are largely confined to pharmacology. Maturitas 48(Suppl 1): 7–12, 2004 conditions of estrogen depletion. Importantly, tamoxi- 10. Shang Y, Brown M: Molecular determinants for the tissue speci- fen reversed the ER activating effects of each XE, thus ficity of SERMs. Science 295(5564): 2465–8, 2002 indicating that this antagonist can clinically reverse BPA 11. Russo J, Yang X, Hu YF, Bove BA, Huang Y, Silva ID, Tahin Q, Wu Y, Higgy N, Zekri A, Russo IH: Biological and molecular and COU activity. However, a tumor-derived mutant basis of human breast cancer. Front Biosci 3: D944–60, 1998 ER (ER-D351Y) harbored an altered response to select 12. Clark GM, Osborne CK, McGuire WL: Correlations between XEs. Specifically, the D351Y mutation rendered the estrogen receptor, progesterone receptor, and patient characteris- receptor hypersensitive to BPA, and the agonistic tics in human breast cancer. J Clin Oncol 2(10): 1102–9, 1984 properties of BPA and COU were significantly enhanced 13. Clark GM, McGuire WL: Steroid receptors and other prognostic factors in primary breast cancer. Semin Oncol 15(2 Suppl 1): 20–5, by tamoxifen. Lastly, we showed that deregulated co- 1988 activator expression differentially cooperates with each 14. Jordan VC: The role of tamoxifen in the treatment and prevention XE for ER activation but not estrogen independent of breast cancer. Curr Probl Cancer 16(3): 129–76, 1992 proliferation. Combined, these studies establish the 15. Brzozowski AM, Pike AC, Dauter Z, Hubbard RE, Bonn T, effects of BPA and COU on breast cancer cells under Engstrom O, Ohman L, Greene GL, Gustafsson JA, Carlquist M: Molecular basis of agonism and antagonism in the oestrogen conditions of estrogen deprivation, selected ER muta- receptor. Nature 389(6652): 753–8, 1997 tion, and co-activator deregulation. As such, this study 16. Shiau AK, Barstad D, Loria PM, Cheng L, Kushner PJ, Agard begins to define the subset of clinical conditions for DA, Greene GL: The structural basis of estrogen receptor/coac- which XE activity may be relevant. tivator recognition and the antagonism of this interaction by tamoxifen. Cell 95(7): 927–37, 1998 17. Sainsbury R: Aromatase inhibition in the treatment of advanced breast cancer: is there a relationship between potency and clinical Acknowledgements efficacy? Br J Cancer 90(9): 1733–9, 2004 18. Hayes, DF, Tamoxifen: Dr. Jekyll and Mr. Hyde? J Natl Cancer Inst 96(12) (2004). 895-7 We would like to thank Drs. Erik Knudsen, Lisa Morey, 19. Arora A, Potter JF: Aromatase inhibitors: current indications and Christin Petre, and Yelena Wetherill, in addition to future prospects for treatment of postmenopausal breast cancer. J Am Geriatr Soc 52(4): 611–6, 2004 Kevin Link, William Zagorski, Craig Burd, and Jon 20. Herynk, M.H. and SA, Fuqua, Estrogen Receptor Mutations in Lenihan for the technical assistance and critical reading Human Disease. Endocr Rev (2004) of the manuscript. The authors gratefully acknowledge 21. Fuqua SA, Wiltschke C, Zhang QX, Borg A, Castles CG, Fried- Drs. James Fagin and Sohaib Khan for discussions and richs WE, Hopp T, Hilsenbeck S, Mohsin S, O’Connell P, Allred constructive comments. DC: A hypersensitive estrogen receptor-alpha mutation in pre- malignant breast lesions. Cancer Res 60(15): 4026–9, 2000 Funding This work was funded without conflict of 22. Fuqua SA, Chamness GC, McGuire WL: Estrogen receptor interest by NIEHS Training Grant ES-07250–13 Envi- mutations in breast cancer. J Cell Biochem 51(2): 135–9, 1993 ronmental Mutagenesis and Cancer (to J.K.H), NIH 23. Sommer S, Fuqua SA: Estrogen receptor and breast cancer. Semin Grant R01 CA93404 and RO1 CA 099996 (to K.E.K.), Cancer Biol 11(5): 339–52, 2001 and NIEHS Core Grant E30-ES-06096 Center for 24. Carlson KE, Choi I, Gee A, Katzenellenbogen BS, Katzenel- lenbogen JA: Altered ligand binding properties and enhanced Environmental Genetics. stability of a constitutively active estrogen receptor: evidence that an open pocket conformation is required for ligand interaction. Biochemistry 36(48): 14897–905, 1997 References 25. Weis KE, Ekena K, Thomas JA, Lazennec G, Katzenellenbogen BS: Constitutively active human estrogen receptors containing 1. Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, Ward E, amino acid substitutions for tyrosine 537 in the receptor protein. Feuer EJ, Thun MJ: Cancer statistics, 2004. CA Cancer J Clin Mol Endocrinol 10(11): 1388–98, 1996 54(1): 8–29, 2004 26. Skafar DF: Formation of a powerful capping motif corresponding 2. Lacey JV Jr., Devesa SS, Brinton LA: Recent trends in breast to start of ‘‘helix 12’’ in agonist-bound estrogen receptor-alpha cancer incidence and mortality. Environ Mol Mutagen 39(2–3): contributes to increased constitutive activity of the protein. Cell 82–8, 2002 Biochem Biophys 33(1): 53–62, 2000 Xenoestrogens and breast cancer 291

27. Anghel SI, Perly VG, Melancon A, Barsalou S, Chagnon A, Ro- 48. Krishnan AV, Stathis P, Permuth SF, Tokes L, Feldman D: senauer WH, Miller Jr., Mader S: Aspartate 351 of estrogen Bisphenol-A: an estrogenic substance is released from polycar- receptor alpha is not crucial for the antagonist activity of anties- bonate flasks during autoclaving. Endocrinology 132(6): 2279–86, trogens. J Biol Chem 275(27): 20867–72, 2000 1993 28. Levenson AS, Catherino WH, Jordan VC: Estrogenic activity is 49. Inoue K, Kato K, Yoshimura Y, Makino T, Nakazawa H: increased for an by a natural mutation of the estrogen Determination of bisphenol A in human serum by high-perfor- receptor. J Steroid Biochem Mol Biol 60(5–6): 261–8, 1997 mance liquid chromatography with multi-electrode electrochemical 29. Catherino WH, Wolf DM, Jordan VC: A naturally occurring detection. J Chromatogr B Biomed Sci Appl 749(1): 17–23, 2000 estrogen receptor mutation results in increased estrogenicity of a 50. Cos P, De Bruyne T, Apers S, Vanden Berghe D, Pieters L, tamoxifen analog. Mol Endocrinol 9(8): 1053–63, 1995 Vlietinck AJ: Phytoestrogens: recent developments. Planta Med 30. Hudelist G, Czerwenka K, Kubista E, Marton E, Pischinger K, 69(7): 589–99, 2003 Singer CF: Expression of sex steroid receptors and their co-factors 51. Olsen CM, Meussen-Elholm ET, Samuelsen M, Holme JA, in normal and malignant breast tissue: AIB1 is a carcinoma-spe- Hongslo JK: Effects of the environmental oestrogens bisphenol A, cific co-activator. Breast Cancer Res Treat 78(2): 193–204, 2003 tetrachlorobisphenol A, tetrabromobisphenol A, 4-hydroxybi- 31. Barnes DM, Gillett CE: Cyclin D1 in breast cancer. Breast Cancer phenyl and 4,4¢-dihydroxybiphenyl on oestrogen receptor binding, Res Treat 52(1–3): 1–15, 1998 cell proliferation and regulation of oestrogen sensitive proteins in 32. Neuman E, Ladha MH, Lin N, Upton TM, Miller SJ, DiRenzo J, the human breast cancer cell line MCF-7. Pharmacol Toxicol Pestell RG, Hinds PW, Dowdy SF, Brown M, Ewen ME: Cyclin 92(4): 180–8, 2003 D1 stimulation of estrogen receptor transcriptional activity inde- 52. Morito K, Aomori T, Hirose T, Kinjo J, Hasegawa J, Ogawa S, pendent of cdk4. Mol Cell Biol 17(9): 5338–47, 1997 Inoue S, Muramatsu M, Masamune Y: Interaction of phytoes- 33. Zwijsen RM, Wientjens E, Klompmaker R, van der Sman J, trogens with estrogen receptors alpha and beta (II). Biol Pharm Bernards R, Michalides RJ: CDK-independent activation of Bull 25(1): 48–52, 2002 estrogen receptor by cyclin D1. Cell 88(3): 405–15, 1997 53. Schafer WR, Zahradnik HP: Organochlorine compounds and 34. Zwijsen RM, Buckle RS, Hijmans EM, Loomans CJ, Bernards R: xenoestrogens in human endometrium. Adv Exp Med Biol 444: 5– Ligand-independent recruitment of steroid receptor coactivators 8, 1998 to estrogen receptor by cyclin D1. Genes Dev 12(22): 3488–98, 54. Routledge EJ, White R, Parker MG, Sumpter JP: Differential 1998 effects of xenoestrogens on coactivator recruitment by estrogen 35. Onate SA, Boonyaratanakornkit V, Spencer TE, Tsai SY, Tsai receptor (ER) alpha and ERbeta. J Biol Chem 275(46): 35986–93, MJ, Edwards DP, O’Malley BW: The steroid receptor coactivator- 2000 1 contains multiple receptor interacting and activation domains 55. Onate SA, Tsai SY, Tsai MJ, O’Malley BW: Sequence and char- that cooperatively enhance the activation function 1 (AF1) and acterization of a coactivator for the steroid hormone receptor AF2 domains of steroid receptors. J Biol Chem 273(20): 12101–8, superfamily. Science 270(5240): 1354–7, 1995 1998 56. Horwitz KB, Jackson TA, Bain DL, Richer JK, Takimoto GS, 36. Bartkova J, Lukas J, Muller H, Lutzhoft D, Strauss M, Bartek J: Tung L: Nuclear receptor coactivators and corepressors. Mol Cyclin D1 protein expression and function in human breast cancer. Endocrinol 10(10): 1167–77, 1996 Int J Cancer 57(3): 353–61, 1994 57. Halachmi S, Marden E, Martin G, MacKay H, Abbondanza C, 37. Wilcken NR, Prall OW, Musgrove EA, Sutherland RL: Inducible Brown M: Estrogen receptor-associated proteins: possible media- overexpression of cyclin D1 in breast cancer cells reverses the tors of hormone-induced transcription. Science 264(5164): 1455–8, growth-inhibitory effects of antiestrogens. Clin Cancer Res 3(6): 1994 849–54, 1997 58. Petre CE, Wetherill YB, Danielsen M, Knudsen KE: Cyclin D1: 38. Umekita Y, Ohi Y, Sagara Y, Yoshida H: Overexpression of mechanism and consequence of androgen receptor co-repressor cyclinD1 predicts for poor prognosis in estrogen receptor-negative activity. J Biol Chem 277(3): 2207–15, 2002 breast cancer patients.. Int J Cancer 98(3): 415–8, 2002 59. Chen C, Okayama H: High-efficiency transformation of mam- 39. Hwang TS, Han HS, Hong YC, Lee HJ, Paik NS: Prognostic malian cells by plasmid DNA. Mol Cell Biol 7(8): 2745–52, 1987 value of combined analysis of cyclin D1 and estrogen receptor 60. Engel LW, Young NA: Human breast carcinoma cells in contin- status in breast cancer patients. Pathol Int 53(2): 74–80, 2003 uous culture: a review. Cancer Res 38(11 Pt 2): 4327–39, 1978 40. Gillett C, Smith P, Gregory W, Richards M, Millis R, Peters G, 61. Kasid A, Lippman ME, Papageorge AG, Lowy DR, Gelmann EP: Barnes D: Cyclin D1 and prognosis in human breast cancer. Int J Transfection of v-rasH DNA into MCF-7 human breast cancer Cancer 69(2): 92–9, 1996 cells bypasses dependence on estrogen for tumorigenicity. Science 41. Berns EM, van Staveren IL, Klijn JG, Foekens JA: Predictive 228(4700): 725–8, 1985 value of SRC-1 for tamoxifen response of recurrent breast cancer. 62. Prall OW, Rogan EM, Sutherland RL: Estrogen regulation of cell Breast Cancer Res Treat 48(1): 87–92, 1998 cycle progression in breast cancer cells. J Steroid Biochem Mol 42. Kurebayashi J, Otsuki T, Kunisue H, Tanaka K, Yamamoto S, Biol 65(1–6): 169–74, 1998 Sonoo H: Expression levels of estrogen receptor-alpha, estrogen 63. Takeuchi T, Tsutsumi O: Serum bisphenol a concentrations receptor-beta, coactivators, and corepressors in breast cancer. Clin showed gender differences, possibly linked to androgen levels. Cancer Res 6(2): 512–8, 2000 Biochem Biophys Res Commun 291(1): 76–8, 2002 43. Zhang H, Yi X, Sun X, Yin N, Shi B, Wu H, Wang D, Wu G, 64. Xu X, Wang HJ, Murphy PA, Cook L, Hendrich S: Daidzein is a Shang Y: Differential gene regulation by the SRC family of co- more bioavailable soymilk isoflavone than is genistein in adult activators. Genes Dev 18(14): 1753–65, 2004 women. J Nutr 124(6): 825–32, 1994 44. Adlercreutz H: Phytoestrogens and breast cancer. J Steroid Bio- 65. Xu X, Harris KS, Wang HJ, Murphy PA, Hendrich S: Bioavail- chem Mol Biol 83(1–5): 113–8, 2002 ability of soybean isoflavones depends upon gut microflora in 45. Adlercreutz H: Phyto-oestrogens and cancer. Lancet Oncol 3(6): women. J Nutr 125(9): 2307–15, 1995 364–73, 2002 66. Morton MS, Wilcox G, Wahlqvist ML, Griffiths K: Determi- 46. Bentrem D, Fox JE, Pearce ST, Liu H, Pappas S, Kupfer D, Zapf nation of lignans and isoflavonoids in human female plasma JW, Jordan VC: Distinct molecular conformations of the estrogen following dietary supplementation. J Endocrinol 142(2): 251–9, receptor alpha complex exploited by environmental estrogens. 1994 Cancer Res 63(21): 7490–6, 2003 67. Welshons WV, Thayer KA, Judy BM, Taylor JA, Curran EM, 47. McLachlan JA: Environmental signaling: what embryos and vom Saal FS: Large effects from small exposures. I. Mechanisms evolution teach us about endocrine disrupting chemicals. Endocr for endocrine-disrupting chemicals with estrogenic activity. Envi- Rev 22(3): 319–41, 2001 ron Health Perspect 111(8): 994–1006, 2003 292 JK Hess-Wilson et al.

68. Levenson AS, Jordan VC: MCF-7: the first hormone-responsive 83. Schafer TE, Lapp CA, Hanes CM, Lewis JB, Wataha JC, Schuster breast cancer cell line. Cancer Res 57(15): 3071–8, 1997 GS: Estrogenicity of bisphenol A and bisphenol A dimethacrylate 69. Hall JM, McDonnell DP: The estrogen receptor beta-isoform in vitro. J Biomed Mater Res 45(3): 192–7, 1999 (ERbeta) of the human estrogen receptor modulates ERalpha 84. Olea N, Pulgar R, Perez P, Olea-Serrano F, Rivas A, Novillo- transcriptional activity and is a key regulator of the cellular Fertrell A, Pedraza V, Soto AM, Sonnenschein C: Estrogenicity of response to estrogens and antiestrogens. Endocrinology 140(12): resin-based composites and sealants used in dentistry. Environ 5566–78, 1999 Health Perspect 104(3): 298–305, 1996 70. Lazennec G, Ediger TR, Petz LN, Nardulli AM, Katzenellenbo- 85. Schmitt E, Dekant W, Stopper H: Assaying the estrogenicity of gen BS: Mechanistic aspects of estrogen receptor activation probed phytoestrogens in cells of different estrogen sensitive tissues. with constitutively active estrogen receptors: correlations with Toxicol In Vitro 15(4-5): 433–9, 2001 DNA and coregulator interactions and receptor conformational 86. Forbes JF: The incidence of breast cancer: the global burden, changes. Mol Endocrinol 11(9): 1375–86, 1997 public health considerations. Semin Oncol 24(1 Suppl 1): S1-20– 71. White R, Sjoberg M, Kalkhoven E, Parker MG: Ligand-inde- S135, 1997 pendent activation of the oestrogen receptor by mutation of a 87. Forbes JF: The control of breast cancer: the role of tamoxifen. conserved tyrosine. Embo J 16(6): 1427–35, 1997 Semin Oncol 24(1 Suppl 1): S1-5–S119, 1997 72. Levenson AS, Svoboda KM, Kwaan HC, Jordan VC: Agonist 88. Feigelson HS, Henderson BE: Estrogens and breast cancer. Car- activity of antiestrogen-receptor complexes to regulate urokinase cinogenesis 17(11): 2279–84, 1996 plasminogen activator (uPA) and plasminogen activator inhibitor 89. Wiseman LR, Adkins JC: Anastrozole. A review of its use in the type 1 (PAI-1) endogenous gene expression in breast cancer cells. management of postmenopausal women with advanced breast Cancer Lett 125(1–2): 215–20, 1998 cancer. Drugs Aging 13(4): 321–32, 1998 73. Tai H, Kubota N, Kato S: Involvement of nuclear receptor co- 90. Nabholtz JM, Buzdar A, Pollak M, Harwin W, Burton G, Man- activator SRC-1 in estrogen-dependent cell growth of MCF-7 galik A, Steinberg M, Webster A, von Euler M: Anastrozole is cells. Biochem Biophys Res Commun 267(1): 311–6, 2000 superior to tamoxifen as first-line therapy for advanced breast 74. Voegel JJ, Heine MJ, Zechel C, Chambon P, Gronemeyer H: cancer in postmenopausal women: results of a North American TIF2, a 160 kDa transcriptional mediator for the ligand-depen- multicenter randomized trial. Arimidex Study Group. J Clin Oncol dent activation function AF-2 of nuclear receptors. Embo J 15(14): 18(22): 3758–67, 2000 3667–75, 1996 91. Ingle JN, Suman VJ: Aromatase inhibitors versus tamoxifen for 75. Fu M, Wang C, Li Z, Sakamaki T, Pestell, RG: Minireview: Cyclin management of postmenopausal breast cancer in the advanced D1: Normal and Abnormal Functions. Endocrinology (2004) disease and neoadjuvant settings. J Steroid Biochem Mol Biol 76. Zoeller RT, Bansal R, Parris, C: Bisphenol-A, an Environmental 86(3-5): 313–9, 2003 Contaminant that Acts as a Thyroid Hormone Receptor Antag- 92. Buzdar AU, Hortobagyi G: Update on endocrine therapy for onist In Vitro, Increases Serum Thyroxine and Alters RC3/Neu- breast cancer. Clin Cancer Res 4(3): 527–34, 1998 rogranin Expression in the Developing Rat Brain. Endocrinology 93. Dombernowsky P, Smith I, Falkson G, Leonard R, Panasci L, (2004) Bellmunt J, Bezwoda W, Gardin G, Gudgeon A, Morgan M, 77. Setchell KD, Zimmer-Nechemias L, Cai J, Heubi JE: Exposure of Fornasiero A, Hoffmann W, Michel J, Hatschek T, Tjabbes T, infants to phyto-oestrogens from soy-based infant formula. Lancet Chaudri HA, Hornberger U, Trunet PF: Letrozole, a new oral 350(9070): 23–7, 1997 aromatase inhibitor for advanced breast cancer: double-blind 78. Rowland I, Faughnan M, Hoey L, Wahala K, Williamson G, randomized trial showing a dose effect and improved efficacy and Cassidy A: Bioavailability of phyto-oestrogens. Br J Nutr tolerability compared with megestrol acetate. J Clin Oncol 16(2): 89(Suppl 1): S45–58, 2003 453–61, 1998 79. Zubik L, Meydani M: Bioavailability of soybean isoflavones from 94. Kaufmann M, Bajetta E, Dirix LY, Fein LE, Jones SE, Zilembo aglycone and glucoside forms in American women. Am J Clin N, Dugardyn JL, Nasurdi C, Mennel RG, Cervek J, Fowst C, Polli Nutr 77(6) : 1459–65, 2003 A, Salle E, Arkhipov A, Piscitelli G, Miller LL, Massimini G: 80. Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, van Exemestane Improves Survival in Metastatic Breast Cancer: Re- der Saag PT, van der Burg B, Gustafsson JA: Interaction of sults of a Phase III Randomized Study. Clin Breast Cancer estrogenic chemicals and phytoestrogens with estrogen receptor 1(Suppl 1): S15–8, 2000 beta. Endocrinology 139(10): 4252–63, 1998 95. Webb P, Nguyen P, Valentine C, Weatherman RV, Scanlan TS, 81. Rivas A, Lacroix M, Olea-Serrano F, Laios I, Leclercq G, Olea N: Kushner PJ: An antiestrogen-responsive estrogen receptor-alpha Estrogenic effect of a series of bisphenol analogues on gene and mutant (D351Y) shows weak AF-2 activity in the presence of protein expression in MCF-7 breast cancer cells. J Steroid Bio- tamoxifen. J Biol Chem 275(48): 37552–8, 2000 chem Mol Biol 82(1): 45–53, 2002 82. Vivacqua A, Recchia AG, Fasanella G, Gabriele S, Carpino A, Rago V, Di Gioia ML, Leggio A, Bonofiglio D, Liguori A, Address for offprints and correspondence: K.E. Knudsen, Department Maggiolini M: The food contaminants bisphenol A and 4-nonyl- of Cell Biology, University of Cincinnati College of Medicine, 3125 phenol act as agonists for in MCF7 breast Eden Ave., 670521, Cincinnati, Ohio, 45267-0521, USA; Tel.: +1-513- cancer cells. Endocrine 22(3): 275–84, 2003 558-7371; Fax: +1-513-558-4454; E-mail: [email protected] Chapter VI:

Summary and Future Directions VI: Summary and Future Directions

A large body of evidence reveals that human exposure to environmentally prevalent endocrine disrupting compounds may impact hormone function (1). Given the crucial link between hormone action and hormone-dependent cancer progression and therapy, we sought to define the agents and molecular context in which EDCs negatively affect hormone dependent cancer treatment. The data herein delineate specific molecular framework in which select EDCs impact hormone-dependent cancer therapies.

Reactivation of the AR plays an important role in the progression of prostate cancer to therapy resistant disease (2, 3); therefore, understanding modes of inappropriate AR activation is critical to advancing the life expectancy of patients with these tumors. Therapy resistant growth of prostate cancer cells is often associated with gain-of-function AR mutation (4-6). The environmentally prevalent endocrine disrupting compound (EDC), bisphenol A (BPA) is able to activate specific somatically mutated

ARs commonly found in ADT-resistant tumors, resulting in androgen-independent, AR- dependent cell cycle progression and proliferation of CaP cells (7, 8). Strikingly, BPA- responsive xenograft tumor-bearing animals exposed to environmentally relevant levels of BPA following ADT, demonstrated accelerated kinetics towards therapy failure, with

AR re-activation (increase in PSA production) and concomitant increased tumor cell proliferation, compared to placebo controls. These data suggest that environmentally relevant exposure to BPA may reduce the efficacy of mainline ADT for prostate cancer, resulting in shorter time to therapy resistance. Although the dose used throughout these experiments is relevant to human exposure ranges, one significant caveat of these experiments is the duration of BPA exposure on the animals. The experimental paradigm utilized allowed for a regulated window of constant BPA exposure to the adult animal, however it is known that human exposure to EDCs occurs throughout life, and in fact, prenatal exposure to BPA has significant effects on prostate development and subsequent disease risk (9-12). A critical question to address is how long-term

(including prenatal) exposure to EDCs such as BPA may impact prostate cancer therapeutic efficacy. It will also be pertinent to address timing and duration of exposure of BPA with ADT therapeutic efficacy as an endpoint, as exposure during puberty, for example, likely has disparate effects than exposure during neonatal or adult stages.

Moreover, it is unlikely that humans are exposed to single agents, whereas exposure to mixture contaminants as well as multiple single agents simultaneously is more probable.

As such, experiments using environmentally relevant combinations of compounds would more closely mimic actual human exposure and may reveal synergistic, cooperative, or antagonistic tendencies of compound mixtures and the pathways in which they impact.

As human exposure to EDCs is widespread, the logical and most straightforward next analysis is to elucidate additional compounds commonly found in the environment that might have similar effects on AR activation and prostate cancer progression.

Strikingly, the environmentally persistent pesticide p,p'DDE activated select AR mutants. Similarly, in in vitro models of CaP, DDE induced AR activation and cellular proliferation in the absence of androgen, however only in cells dependent upon AR activation for mitogenesis. Conversely, the mechanism of DDE impact on CaP cells was distinct from BPA, in that at low doses this agent also activated the MAPK pathway, which was requisite for the mitogenic action of DDE. Combined, these data imply that the environmental influence on AR action in CaP is impacted by select EDCs, and through specific and divergent mechanisms. The necessity for clearly elucidating the precise modes of action by which specific EDCs may impinge upon hormone dependent cancer therapy is highlighted by these findings.

Given the context specific influence of AR activation by EDCs, and the deleterious effect of exposure to BPA on mainline ADT, the impact of AR action on taxane-based therapy (potential second line therapies already in advanced clinical trials for CaP treatment) was assessed. It is a widely held belief that both androgen and AR action act as survival factors for prostatic epithelia and carcinoma. Studies have indeed found that AR action decreased the capacity of specific cytotoxic agents to induce cell death, specifically okadaic acid, and TRAIL (13), however these cytotoxic insults are not clinically relevant for CaP. Strikingly, AR activation by both endogenous and exogenous agonists (i.e. DHT, and BPA), synergized with taxanes to decrease cell survival. This reduced cell survival was facilitated through p53-mediated, caspase dependent apoptosis and was attributed directly to the AR-dependent mitogenic capacity of AR ligands. These data further support the conclusion that the EDC impact on prostate cancer therapy is context specific. Within the context of taxane based treatments, pharmaceutical EDCs (such as diethylstilbestrol; DES) that could drive proliferation of prostate cancer cells at mitogenic doses, may be explored for their ability to cooperate with taxanes and enhance the cytotoxic impact of this treatment modality.

These data highlight the necessity for pre-clinical in vivo animal studies and clinical trials comparing taxane-based treatment on divergent prostate cancer patient populations in different disease states. Studies such as these may improve the clinical usage and outcome of taxanes and provide an additional treatment option for prostate cancer patients.

The data herein specifically demonstrate that AR mutation facilitates EDC impact on CAP progression and therapy, using clinically relevant prostate cancer tissue models and in vivo models expressing EDC-responsive AR mutants. However, how EDCs impact the other pathways involved in CaP progression (i.e. amplification and overexpression of AR, deregulated AR coactivators, deregulated ligand independent

(growth factor) pathways, loss of function and/or mutation of the retinoblastoma tumor suppressor or PTEN) remains undetermined. Additionally, given the large focus on the effects of EDCs whose action may be mediated by binding to steroid hormone receptors, it is imperative to address other potential mechanisms of EDC action. EDCs have the ability to induce endocrine interference via routes that involve disrupting the biosynthesis and or metabolism of (14), and also activating growth factor or kinase signaling pathways (15-17) . A broad approach is needed when considering

EDC action, including the biological context in which exposure occurs, mechanism of action, endpoints examined, timing and duration of exposure, etc.

Similar to CaP, breast cancer is often treated based on its dependence on estrogen for proliferation, therefore, EDC exposure on breast cancer cells were examined under clinically relevant molecular conditions, to address the possibility of negatively impacting breast cancer therapy and progression. BPA and the phytoestrogen coumestrol (COU) demonstrated the known ability to activate the estrogen receptor (ERa), but this activation was restricted to estrogen-depleted conditions, and could be blocked by the standard ER antagonist, tamoxifen (18). Although BPA and COU could activate ERa-dependent proliferation, this action would not be deleterious to most patients, as tamoxifen is the standard therapy. However with the increased usage of aromatase inhibitors, which limit endogenous estrogen levels, and further insight into specific molecular contexts related to breast cancer progression, more attention needs to be directed toward delineating the effects of EDCs on breast cancer therapy and progression. Additionally, these studies highlight that other hormone-dependent diseases can be impacted by EDC exposure. Therefore, the molecular context underlying the development, progression and treatment of diseases such as endometriosis, polycystic ovarian syndrome, androgen insensitivity syndrome, as well as reproductive and fertility diseases need to be fully assessed for sensitivity to

EDC exposure.

The studies presented in this dissertation begin to address what molecular context of hormone dependent cancers are susceptible to EDC adverse effects

(summarized in Figure 1). Through these analyses, numerous disparities between both canonical ligand (DHT) and EDCs (i.e. BPA and DDE or BPA and COU) were identified.

Characteristics that are unique to select agents, (i.e. activation of MAPK pathway by

DDE) were demonstrated. In-depth molecular studies are needed to address these discrepancies, which would provide tremendous insight into both steroid receptor and endogenous hormone function within the hormone dependent tissue, as well as distinct

EDC mode of action. Specifically, AR structure – function analysis could reveal why only select AR mutants are responsive to very specific EDCs. Additionally, preliminary data in our lab reveals that although BPA and DHT have similar biological effects on

AR-dependent proliferation, and do share some similar AR-mediated gene targets, Figure 1 Summary of EDC impact on androgen deprivation therapy and taxane therapy for prostate cancer.

Within the context of mutant AR expression, bisphenol A (BPA) induces AR activation and AR-dependent proliferation, and is able to drive the progression towards recurrent disease. Dichlorodiphenyldichloroethylene (DDE) utilizes both mutant AR and MAPK-dependent pathways to increase proliferation of prostate cancer cells. The impact of other EDCs on disease progression has yet to be determined. The efficacy of taxane therapy is increased by the presence of AR-T877A agonists, including BPA. Although other EDCs have not been assessed for cooperation with taxanes, the utilization of pharmaceutical EDCs such as diethylstilbestrol (DES), which may drive the proliferation of prostate cancer cells, should enhance the cytotoxicity of taxane-based treatments. microarray data surprisingly shows that they also facilitate very unique mRNA signatures in prostate cancer cells. For example, microarray analysis revealed that in

LNCaP cells, BPA downregulated ER-beta mRNA, however DHT did not (data not shown). An analysis of co-regulatory molecules recruited to the AR in the presence of different ligands would provide valuable information on mechanism of distinct target gene profiles, as well as biology of select co-activator usage by AR. The more information is learned about the biology and modes of AR action in the prostate, as well as the specific points of alteration in prostate cancers, the more refined the analysis of

EDC impact on prostate cancer development and treatment can be.

The environmental impact on nuclear receptor signaling and hormone dependent cancer progression and treatment is highly molecular context specific. As shown in

Figure 1, we have shown that for CaP, exposure to the environmental compounds BPA and DDE may activate select ARs, inducing AR-mediated cell proliferation and thereby reducing the efficacy of ADT. Conversely, the cytotoxic action of taxane-based regimens is improved upon by AR-dependent cell cycle progression in CaP cells, indicating that inappropriate AR activation by exogenous EDCs may not be a threat to the efficacy of this therapeutic modality. As it is known that these diseases are heterogeneous and distinct molecular environments may exist in different patients, speculations on the impact of EDCs on hormone dependent cancer treatment, needs to be identified under specific and well-defined, clinically relevant, molecular contexts.

In summary, the prevailing message suggests that the action of steroid hormones upon target cells and organs is regulated by a complex interplay of genomic and non-genomic signaling. Any one of these mechanisms, singularly and in complex mixtures, could be disrupted by EDC exposure. To fully understand EDC risk, the critical issue of timing of exposure and the idea of sensitive periods of development or disease state needs to be completely identified. Moreover, studies such as those described herein analyzing the change in biological consequence dependent upon the association between nuclear hormone receptor signaling and drug/environmental agent, can be manipulated clinically when designing or prescribing treatment for hormone dependent diseases. As knowledge of mode of action of EDCs in specific molecular context increases, we will be better enabled to hone the definition of human risk to these agents and more effectively counter their adverse effects clinically.

1. Hess-Wilson JK, Knudsen KE. Endocrine disrupting compounds and prostate cancer. Cancer Lett 2005. 2. Feldman BJ, Feldman D. The development of androgen-independent prostate cancer. Nat Rev Cancer 2001;1(1):34-45. 3. Trapman J, Brinkmann AO. The androgen receptor in prostate cancer. Pathol Res Pract 1996;192(7):752-60. 4. Culig Z, Hobisch A, Cronauer MV, et al. Mutant androgen receptor detected in an advanced-stage prostatic carcinoma is activated by adrenal androgens and progesterone. Mol Endocrinol 1993;7(12):1541-50. 5. Gottlieb B, Beitel LK, Wu JH, Trifiro M. The androgen receptor gene mutations database (ARDB): 2004 update. Hum Mutat 2004;23(6):527-33. 6. Suzuki H, Sato N, Watabe Y, Masai M, Seino S, Shimazaki J. Androgen receptor gene mutations in human prostate cancer. J Steroid Biochem Mol Biol 1993;46(6):759- 65. 7. Wetherill YB, Fisher NL, Staubach A, Danielsen M, de Vere White RW, Knudsen KE. Xenoestrogen action in prostate cancer: pleiotropic effects dependent on androgen receptor status. Cancer Res 2005;65(1):54-65. 8. Wetherill YB, Petre CE, Monk KR, Puga A, Knudsen KE. The xenoestrogen bisphenol A induces inappropriate androgen receptor activation and mitogenesis in prostatic adenocarcinoma cells. Mol Cancer Ther 2002;1(7):515-24. 9. Ho SM, Tang WY, Belmonte de Frausto J, Prins GS. Developmental exposure to estradiol and bisphenol A increases susceptibility to prostate carcinogenesis and epigenetically regulates phosphodiesterase type 4 variant 4. Cancer Res 2006;66(11):5624-32. 10. Gupta C. Reproductive malformation of the male offspring following maternal exposure to estrogenic chemicals. Proc Soc Exp Biol Med 2000;224(2):61-8. 11. Ramos JG, Varayoud J, Sonnenschein C, Soto AM, Munoz De Toro M, Luque EH. Prenatal exposure to low doses of bisphenol A alters the periductal stroma and glandular cell function in the rat ventral prostate. Biol Reprod 2001;65(4):1271-7. 12. Timms BG, Howdeshell KL, Barton L, Bradley S, Richter CA, vom Saal FS. Estrogenic chemicals in plastic and oral contraceptives disrupt development of the fetal mouse prostate and urethra. Proc Natl Acad Sci U S A 2005;102(19):7014-9. 13. Kimura K, Markowski M, Bowen C, Gelmann EP. Androgen blocks apoptosis of hormone-dependent prostate cancer cells. Cancer Res 2001;61(14):5611-8. 14. Sharpe RM, Irvine DS. How strong is the evidence of a link between environmental chemicals and adverse effects on human reproductive health? Bmj 2004;328(7437):447-51. 15. Agarwal R. Cell signaling and regulators of cell cycle as molecular targets for prostate cancer prevention by dietary agents. Biochem Pharmacol 2000;60(8):1051-9. 16. Ray G, Dhar G, Van Veldhuizen PJ, et al. Modulation of cell-cycle regulatory signaling network by 2-methoxyestradiol in prostate cancer cells is mediated through multiple signal transduction pathways. Biochemistry 2006;45(11):3703-13. 17. Tessier DM, Matsumura F. Increased ErbB-2 tyrosine kinase activity, MAPK phosphorylation, and cell proliferation in the prostate cancer cell line LNCaP following treatment by select pesticides. Toxicol Sci 2001;60(1):38-43. 18. Hess-Wilson JK, Boldison J, Weaver KE, Knudsen KE. Xenoestrogen action in breast cancer: impact on ER-dependent transcription and mitogenesis. Breast Cancer Res Treat 2006;96(3):279-92.