`
`PERGAMON
`
`Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`TheJournal of
`Steroid Biochemistry
`&
`Molecular Biology
`
`EM-652 (SCH 57068), a third generation SERM acting as pure
`antiestrogen in the mammary gland and endometrium”
`
`Fernand Labrie*, Claude Labrie, Alain Bélanger, Jacques Simard, Sylvain Gauthier,
`Van Luu-The, Yves Mérand, Vincent Giguere, Bernard Candas, Shouqi Luo,
`Céline Martel, Shankar Mohan Singh, Mare Fournier, Agnés Coquet, Virgile Richard,
`Ronald Charbonneau, Gilles Charpenet, André Tremblay, Gilles Tremblay,
`Lionel Cusan, Raymonde Veilleux
`
`Oncology and Molecular Endocrinology Research Center, Centre Hospitalier Universitaire de Québec (CHUQ), Pavilion CHUL, Department of
`Medicine, Laval University, Québec, GIV 4G2, Canada
`
`
`
`Abstract
`
`Breast cancer is the most frequent cancer in women while it is the second cause of cancer death. Estrogens are well recognized
`to play the predominant role in breast cancer development and growth and muchefforts have been devoted to the blockade of
`estrogen formation and action. The most widely used therapy of breast cancer which has shown benefits at all stages of the
`disease is the use of the antiestrogen Tamoxifen. This compound, however, possesses mixed agonist and antagonist activity and
`major efforts have been devoted to the development of compounds having pure antiestrogenic activity in the mammary gland
`and endometrium. Such a compound would avoid the problem of stimulation of the endometrium and the risk of endometrial
`carcinoma. We have thus synthesized an orally active non-steroidal antiestrogen, EM-652 (SCH 57068) and the prodrug EM-800
`(SCH57050) which are the most potent of the known antiestrogens. EM-652 is the compound having the highest affinity for the
`estrogen receptor,
`including estradiol. It has higher affinity for the ER than ICI 182780, hydroxytamoxifen,
`raloxifene,
`droloxifene and hydroxytoremifene. EM-652 has the most potent inhibitory activity on both ERa and ERf compared to any of
`the other antiestrogens tested. An important aspect of EM-652 is that it inhibits both the AF1 and AF2 functions of both ERa
`and ERf while the inhibitory action of hydroxytamoxifen is limited to AF2, the ligand-dependent function of the estrogen
`receptors. AF1 activity is constitutive, ligand-independent and is responsible for mediation of the activity of growth factors and
`of the ras oncogene and MAP-kinase pathway. EM-652 inhibits Ras-induced transcriptional activity of ERa and ERf and
`blocks SRC-1-stimulated activity of the two receptors. EM-652 was also found to block the recruitment of SRC-1 at AFI of
`ER§,this ligand-independent activation of AF1 being closely related to phosphorylation of the steroid receptors by protein
`kinase. Most importantly, the antiestrogen hydroxytamoxifen has no inhibitory effect on the SRC-1l-induced ERf activity while
`the pure antiestrogen EM-652 completely abolishes this effect, thus strengthening the need to use pure antiestrogens in breast
`cancer therapy in order to control all known aspects of ER-regulated gene expression. In fact, the absence of blockade of AF2
`by hydroxytamoxifen could explain why the benefits of tamoxifen observed up to 5 years become negative at longer time
`intervals and why resistance develops to tamoxifen. EM-800,
`the prodrug of EM-652, has been shown to prevent
`the
`development of dimethylbenz(a)anthracene (DMBA)-induced mammary carcinomain the rat, a well-recognized model of human
`breast cancer. It is of interest that the addition of dehydroepiandrosterone, a precursor of androgens,
`to EM-800,
`led to
`complete inhibition of tumor development in this model. Not only the development, but also the growth of established DMBA-
`induced mammary carcinoma was inhibited by treatment with EM-800. An inhibitory effect was also observed when
`medroxyprogesterone was added to treatment with EM-800. Uterine size was reduced to castration levels in the groups of
`animals treated with EM-800. An almost complete disappearance of estrogen receptors was observed in the uterus, vaginum and
`
`* Proceedings of Xth International Congress on Hormonal Steroids, Quebec, Canada, 17-21 June 1998.
`* Corresponding author. Tel.: + (418) 654-2704; fax: + (418) 654-2735.
`E-mail address: fernand. labrie@crchul.ulaval.ca (F. Labrie)
`
`0960-0760/99/$ - see front matter © 1999 Published by Elsevier Science Ltd. All rights reserved.
`PII: $0960-0760(99)00065-5
`
`AstraZeneca Exhibit 2034 p. 1
`InnoPharma Licensing LLC v. AstraZeneca AB IPR2017-00904
`Fresenius-Kabi USA LLC v. AstraZeneca AB IPR2017-01910
`
`
`
`52
`
`F. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`tumors in nude mice treated with EM-800. EM-652 was the most potent antiestrogen to inhibit the growth of human breast
`cancer ZR-75-1, MCF-7 and T-47D cells in vitro when compared with ICI 182780,
`ICT 164384, hydroxytamoxifen, and
`droloxifene. Moreover, EM-652 and EM-800 have no stimulatory effect on the basal levels of cell proliferation in the absence of
`E2 while hydroxytamoxifen and droloxifene had a stimulatory effect on the basal growth of T-47D and ZR-75-1 cells. EM-652
`was also the most potent inhibitor of the percentage of cycling cancer cells. When human breast cancer ZR-75-1 xenografts were
`grown in nude mice, EM-800 led to a complete inhibition of the stimulatory effect of estrogens in ovariectomized mice while
`tamoxifen was less potent and even stimulated the growth of the tumors in the absence of estrogens,
`thus illustrating the
`stimulatory effect of tamoxifen on breast cancer growth. When incubated with human Ishikawa endometrial carcinomacells,
`EM-800 had no stimulatory effect on alkaline phosphatase activity, an estrogen-sensitive parameter. Raloxifene, droloxifene,
`hydroxytoremifene and hydroxytamoxifen, on the other hand, all stimulated to various extent, the activity of this enzyme. The
`stimulatory effect of all
`four compounds was blocked by EM-800,
`thus confirming their estrogenic activity in human
`endometrial
`tissue. When administered to ovariectomized animals, EM-800 prevents bone loss,
`the effect on bone mineral
`density, trabecular bone volume, and trabecular separation being 5—10 times more potent than raloxifene. EM-800 lowers serum
`cholesterol and triglyceride levels in the rat as well as in women. In a Phase IJ study performed in patients with breast cancer
`showing failure on tamoxifen, 1 patient had a complete response while 5 patients had a partial response and stable disease for at
`least three months has been observed in an additional 13 patients for a total of 19 positive responses out of 43 evaluable
`patients (44.2%). No significant secondary effect related to the drug was observed. A phase 3 international clinical
`trial
`is
`currently being performed in tamoxifen failure patients where EM-800 (SCH 57050) is compared to Arimidex. The detailed
`information obtained at the preclinical level with EM-652 or EM-800 indicates that these orally active compounds are highly
`potent and pure antiestrogens in the mammary gland and endometrium while they prevent bone loss and lower serum
`cholesterol and triglyceride levels. Preclinical and clinical data clearly suggest the interest of studying this compound in the
`neoadjuvant and adjuvant settings and, most importantly, for the prevention of breast and uterine cancer in which settings they
`should provide additional benefits on bone and lipids. © 1999 Published by Elsevier Science Ltd. All rights reserved.
`
`Keywords: Pure antiestrogen; Breast cancer; Uterine cancer; EM-652; SCH 57068; EM-800; SCH 57050; Osteoporosis; Selective estrogen receptor
`modulator (SERM); Cholesterol; Triglycerides; Prevention; Risk reduction; Adjuvant; Neoadjuvant
`
`1. Introduction
`
`1.1. Breast cancer
`
`Breast cancer is the most frequent cancer in women,
`with 176,300 new cases and 43,700 deaths predicted in
`the United States in 1999 [1]. Present
`therapies in
`breast cancer achieve meaningful clinical
`results in
`only 30-40% of patients, with response duration
`usually limited to 12-18 months [2—5]. Five-year survi-
`val in women with metastatic disease is still only 10—
`40%.
`Amongall risk factors, estrogens are well recognized
`to play the predominant role in breast cancer develop-
`ment and growth [6-9]. However, existing surgical or
`medical ablative procedures do not result in complete
`elimination of estrogens in women [10], due to the con-
`tribution of the adrenal glands that secrete high levels
`of dehydroepiandrosterone (DHEA) and DHEA-sul-
`fate which are converted into estrogens in peripheral
`target tissues [11-13]. Considerable attention has thus
`been focused on the development of blockers of estro-
`gen biosynthesis and action [14-20].
`Since the first step in the action of estrogens in tar-
`get tissues is binding to the estrogen receptor [21,22], a
`logical approach for the treatment of estrogen-sensitive
`breast cancer is the use of antiestrogens, or compounds
`which block the interaction of estrogens with their
`specific receptor. Until very recently, no agent with
`
`pure antiestrogenic activity under in vivo conditions
`has been available.
`
`1.2. Tamoxifen
`
`the antiestrogen most widely used for
`Tamoxifen,
`the treatment of women with breast cancer has shown
`clear clinical benefit in advanced breast cancer, its effi-
`cacy being comparable to that achieved with ablative
`and additive therapies [23]. In the first clinical studies
`initiated in 1969,
`tamoxifen was found to achieve
`remissions in advanced breast carcinoma similar
`to
`those observed following estrogen therapy but with
`fewer side effects [24]. Since then, because ofits favor-
`able safety profile and clinical efficacy comparable to
`other endocrine therapies,
`including oophorectomy
`and estrogens, tamoxifen has become the treatment of
`choice for patients with advanced or metastatic breast
`cancer [25—27]. This compound, however, is known to
`possess mixed estrogenic and antiestrogenic activities
`[19,23,28] which are species-,
`tissue-, cell-, and even
`gene-specific [29,30]. In support of the clinical evidence
`for
`the estrogenic activity of tamoxifen on human
`breast cancer growth [31,32], tamoxifen and its active
`metabolite 4-OH-tamoxifen have been found to stimu-
`late the growth of human breast cancer cells in vitro
`and in vivo [29,3340]. Tamoxifen may act as an estro-
`gen agonist more frequently than generally thought
`and this may explain some of the apparent paradoxes
`
`AstraZeneca Exhibit 2034 p. 2
`
`
`
`F. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`53
`
`
`
`EM-652 R=H
`EM-800 R=COC(CH3)3
`
`X=Y=Z=H, Ry=R2=CH3
`Tamoxifen
`Droloxifene X=H,Y=OH, Z=H, Ri=R2=CH3
`Toremifene X=Y=H, Z=Cl, R}=R2=CH3
`Idoxifene
`X=l, Y=Z=H, Ry,Ro=CyHg
`
`OH
`
`tl
`
`HO
`
`™ACH2),R
`
`1C}164,384
`EM-139
`1C1182,780
`
`X=H, n=10, R=CON(CH3)C,H9
`X=Cl, n=10, ReCON(CH;)C,H9
`X=H, n=9, R=SO(CH.2),C2Fs
`
`
`
`Raloxifene
`
`Fig. 1. Antiestrogens—molecular structures.
`
`of endocrine treatments such as response to second en-
`docrine therapy and withdrawal responses [27].
`are
`Additionally, while
`benefits
`of
`tamoxifen
`observed on breast cancer in up to 40% of patients,
`the long-term use of this compound has recently been
`recognized as being associated with a
`significant
`increase in the incidence of endometrial carcinoma
`[41-55], an effect which is likely caused by the intrinsic
`estrogenic activity of the compound and possibly
`because of its genotoxic action on the DNA, by form-
`ing DNA adducts. The close analogs of tamoxifen,
`namely toremifene,
`Idoxifene and droloxifene, also
`possess estrogenic effects analogous to those of tamox-
`ifen [56,57 data not shown].
`
`2. Need for an orally active pure antiestrogen in the
`mammary gland and endometrium
`
`Since clinical data suggest that long-term (5 years)
`tamoxifen adjuvant therapy is preferable to the short-
`term (2 years) use of the antiestrogen [58,59] and stu-
`dies are in progress on the long-term use of tamoxifen
`as a chemo-preventive in breast cancer [54,60,61],
`it
`has become important to develop a pure antiestrogen
`to avoid the negative effects of the partial estrogenic
`activity of Tamoxifen and thus make available a com-
`pound having activities limited to the desired thera-
`peutic action. The first class of pure antiestrogens
`obtained were
`7u-substituted estradiol
`derivatives
`
`[5,14,16,18,19,62,63], especially ICI 164,384, EM-139,
`and ICI 182,780 (Fig. 1).
`These compounds have been shown to possess pure
`and potent antiestrogenic activity in most well recog-
`
`EM 652
`
`Fig. 2. Structure of EM-652 (SCH 57068).
`
`AstraZeneca Exhibit 2034 p. 3
`
`
`
`54
`
`Table 1
`
`fF. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`Comparison of the estrogen receptor affinity of a series of antiestrogens and related compoundswith estradiol (E) and diethylstilbestrol (DES)
`in human breast cancer and normal human uterine cytosol*
`
`Breast Cancer
`Uterus
`
`
`ethanol
`DMF
`ethanol
`DMF
`
`
`Compounds
`K; (nM) (max)
`RBA
`K; (max)
`RBA
`K; (max)
`RBA
`K; (max)
`RBA
`
`
`0.181
`0.120
`0.113
`100
`0.138
`E,
`-
`0.128
`-
`110
`0.126
`DES
`0.069
`0.042
`0.076
`291
`0.047
`(S)-6(EM-652)
`-
`1.89
`-
`6.62
`2.09
`(R)-6(EM-651)
`-
`11.14
`-
`2.32
`4.71
`(S)-1(EM-800)
`-
`-
`-
`< 0.04
`> 270
`(R)-1(EM-776)
`10.3
`1.76
`2.33
`1.53
`3.00
`4.60
`ICI 164,384
`27.2
`0.668
`-
`-
`0.755
`1.81
`7.63
`ICI 182,780
`-
`-
`43.8
`0.346
`-
`-
`43.8
`0.249
`(Z)-4-OH-Tamoxifen
`
`
`
`
`
`
`
`
`11.9 0.92 - - 34.4 0.92 -Tamoxifen -
`
`100
`93.5
`284
`6.34
`1.32
`
`-
`
`5.15
`
`100
`-
`264
`-
`-
`-
`
`100
`-
`150
`-
`-
`-
`
`746
`15.1
`
`* Incubations were performed at room temperature for 3 h using 100 uL of cytosol, 100 wL of PHIE2 (5 nM Ep,final) and 100 pL ofthe indi-
`cated unlabelled compounds leading to final concentrations of 3.3% ethanol or 2.5% dimethylformamide (DMF). The apparent inhibition con-
`stant (Kj) and relative binding affinity (RBA) values were calculated as described [73,223]. The apparent inhibition constant K; values were
`calculated according to the following equation: K;=ICsq/(1+S/K) where S represents the concentration of labelled E, and Kis the Kp value of
`E> (0.14 nM)for the estrogen receptor. RBA values were calculated as follows: RBA =ICs9 of Es/ICs9 of tested compound x100 [56].
`
`cancer and normal human uterine cytosol as described
`[56,73]. As measured by competition studies in human
`breast cancer tissue, the affinity of EM-652 (K;=0.047
`+0.003 nM, RBA =291, relative to 17f-estradiol set at
`100) studied in the presence of ethanol was 2.9 higher
`
`100
`
`oa
`
`a
`
`40
`
`80
`
`60
`
`O-© E2
`@-@ EM-s52
`
`-- Icl 182780
`
`20 mm ici 164384
`
`ZZ. DROLOXIFENE
`
`A TOREMIFENE
`
`
`
`-11
`
`-10
`
`9
`
`“8
`
`7
`
`6
`
`“5
`
`including human
`nized in vitro and in vivo systems,
`breast cancer cells
`[14,16,18,19,64,65]. The 7«-alkyl
`estradiol derivative ICI 164384, however, has been
`found to possess some estrogenic agonistic activity in
`guinea pig uterine cells
`[66,67]. Furthermore, both
`OH-tamoxifen and ICI-164384 can stimulate CAT ac-
`tivity in MCF-7 cells transfected with a pS2-tkCAT
`fusion gene [68]. Moreover, such 7a-alkyl estradiol de-
`rivatives are difficult to synthesize and their bioavail-
`ability by the oral route is very low, thus necessitating
`parenteral administration. We therefore concentrated
`our efforts on the synthesis of non-steroidal com-
`pounds having oral activity in order to overcomethis
`difficulty.
`In order to facilitate large-scale purification, EM-
`800 (SCH 57050), the bipivalate derivative of EM-652
`was synthesized. EM-800 is rapidly transformed into
`EM-652 in intact cells and following in vivo adminis-
`tration. The other derivative currently used in our stu-
`dies is EM-652.HCl (SCH 57068.HCI). In an effort to
`develop an orally active agent, EM-652 was
`syn-
`thesized (Fig. 2). As will be discussed later, the active
`compound EM-652 derived from EM-800 or EM-
`652.HCI behaves as a highly potent and pure antiestro-
`gen in human breast and uterine cancer cells in vitro
`as well as in vivo in nude mice [56,69—72].
`
`3. Binding characteristics to the estrogen receptors «
`and B
`
`The estrogen receptor affinity of EM-652, the active
`drug of EM-800, was first measured in human breast
`
`[3H]E2BOUND(%)
`
`CONCENTRATION(LOG M)
`
`ICI
`Fig. 3. Effect of increasing concentrations of EM-652, E2,
`182780, Droloxifene, ICI 164384, and Toremifene on [*H] 17f-estra-
`diol binding to the rat uterine estrogen receptor. The incubation was
`performed with 5 nM [*H] 17f-estradiol (E>) for 2 h at room tem-
`perature in the presence or absence of the indicated concentrations
`of unlabeled compounds[224].
`
`AstraZeneca Exhibit 2034 p. 4
`
`
`
`fF. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`55
`
`Binding properties of estrogen receptors
`
`
`
`Foldinduction
`
`2 —
`
`0.5
`
`0.4
`
`oO @
`
`° NS
`
`Bound/Free
`
`4
`6
`8
`-10
`12
`E2 concentration (logM)
`
`E1=
`3c39
`
`a
`
` =>
`
`0
`
`100
`
`200
`
`0
`
`50
`=6«—0
`300
`Bound (fmol)
`
`100
`
`150
`
`Fig. 4. Dose-response and binding properties ofmERa and mER#. (A) Cos-1 cells were transfected with 500 ng mER® (open circles) or MERa
`(closed circles) expression vectors and 1 pg vitA,-ERE-TKLuc and then incubated for 12 h with increasing concentrations of E, as indicated. (B)
`Specific binding of [2,4,6,7-3H]-17f-estradiol (PHJE:) to mERf was determined using receptors generated from rabbit reticulocyte lysates.
`Binding was determined over a concentration range of 0.01-3 nM [*HJE, in the absenceor presence of a 200-fold excess of unlabeled E>. Thesat-
`uration plot is shown in the inset, and results were plotted by the method of Scatchard. Each point was determined in triplicate in each exper-
`iment, and the aboveresults are representative of at least two separate experiments. (C) Specific binding to mER« using the conditions described.
`in panel B [65].
`
`(RBA=6.62). Similar
`itself
`than that of estradiol
`results were obtained on the human uterine estrogen
`receptor (Table 1). It can be seen in the same table
`that ICI 182,780 has about 10 times lower affinity than
`EM-652 to displace PHJE, from the human estrogen
`receptor while (Z)-4-OH-Tamoxifen is about 6 times
`less potent under the experimental conditions used.
`The new antiestrogen EM-652 thus shows the highest
`affinity for the human estrogen receptorofall the com-
`poundstested [56] (Table 1).
`It can be seen in Fig. 3 that EM-652 is 7- to 8-fold
`more potent
`than E2 and ICI 182780 in displacing
`[HJE2 from the rat uterine estrogen receptor (ICso
`values of 0.52, 4.13, and 3.59 nM for EM-652, E>, and
`ICI 182780, respectively). ICI 164384 and Droloxifene
`are 21-fold less potent than EM-652 while Toremifene
`is 400 times less potent than EM-652.
`Over the past decade, all the studies on the elucida-
`tion of the molecular events underlying the mode of
`ER action as well as the antiestrogen-designed therapy
`have focused on the ER« identified and cloned several
`years ago [21,74,75]. Recently, a second estrogen recep-
`tor, designated ERf, has been described and shown to
`share common structural and functional characteristics
`with ER« [65,76,77]. Based on amino acid sequence
`comparison, ERf shares with ER« the same modular
`
`[78].
`(A-F)
`six domains
`composed of
`structure
`Domain C, which contains the two zinc fingers respon-
`sible for DNA binding, is the most conserved followed
`by domain E, responsible for ligand binding, homodi-
`merization and nuclear localization. Domain E also
`contains a ligand-dependent activation function (AF-2)
`involved in trans-activation by the ERs. A second acti-
`vation function, AF-1, resides in the A/B domain and
`acts in a ligand-independent manner [79-81].
`Both ERs recognize a specific estrogen response el-
`ement (ERE) composed of two AGGTCA motif half-
`sites configured as a palindrome spaced by three
`nucleotides [65]. ER« has also been shown to interact
`with a number of coregulators via the AF-2 domain,
`and these protein-protein interactions promote tran-
`scriptional
`regulation
`of
`target
`genes
`[82-85].
`Following cloning of mouse ERf [65], comparison
`could be made of the activity of ER« and ERand
`measurement could be made of the affinity of the two
`ERsfor various ligands, especially, antiestrogens.
`Wefirst tested the activity of both receptors in
`the presence of increasing E, concentrations using
`the
`vitA2-ERE-TKLuc
`reporter
`in Cos-1
`cells.
`Comparison of the dose-response curves of Fig. 4A
`shows a shift of approximately 4-fold of the E, con-
`centration required to achieve half of the maximal
`
`AstraZeneca Exhibit 2034 p. 5
`
`
`
`56
`
`fF. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`1C1182,780
`IC1182,780
`HN [art]A2|COOH
`100
`
`
`
`repressionofE2response(%)
`
`A
`
`B
`
` °o
`
`aNoai
`
`NOoi
`
`-
`
`E2 -11-10 -9 -8
`
`-7
`
`-
`
`E2 -11-10 -9
`
`-8
`
`-7
`
`antagonist concentration (logM)
`
`Fig. 5. Dose-response of antagonists on ERa- and ERf-mediated
`transactivation. Comparison of the dose—-responses of the antagonists
`in the presence of 10 nM E> on the transcriptional activity of ER«
`(A) and ERP (B) using the vitA,ERETKLuc reporter in COS-1 cells.
`Results represent the mean +SEM of three separate experiments and
`are expressed as percentages of the maximal induction by E, alone
`(set arbitrarily at 100%; filled bars) for the two ERs. The untreated
`ERoa and ERfbasallevels are also shown [105].
`
`level of induction between the two receptors, the ER«
`being more sensitive to Ep.
`The above-indicated results already suggested that
`ERf may have lower affinity for E, than mER«. To
`verify if the difference in E, responsiveness was due to
`a difference in ligand binding, we performed a binding
`analysis on both mERf and mER«. [’H]E> was used
`to conduct binding studies with mERf, and results
`were plotted by the method of Scatchard. As shown in
`Fig. 4B,
`this analysis yielded an average dissociation
`constant (Ka) of 0.5 nM for E, when performed on
`ERf prepared from rabbit reticulocyte lysates. This
`value is comparable to that obtained for the rat ERS,
`which was reported to be 0.6 nM [76]. On the other
`hand, we obtained an average Kg of 0.2 nM for mER«
`(Fig. 4C), which is well within the range of previously
`published determinations for the cloned human recep-
`tor [86]. Therefore,
`this slightly reduced affinity of
`mERf for E; may provide an explanation for the shift
`in E, responsiveness indicated by the dose-response
`curves (Fig. 4A).
`To further evaluate the potency of various antiestro-
`gens, we compared their dose-dependent inhibition of
`E,-induced
`ERa
`and ER activity
`using
`vitA,ERETKLuc in COS-1 cells (Fig. 5). When com-
`pared to ICI182,780, EM-652 was highly effective,
`achieving a complete blockade of the E,-inducedeffect
`of ER« (Fig. 5A) and ERf (Fig. 5B) at concentrations
`of 10-* M and above. Comparison of the apparent
`ICsq values showed that under the conditions used,
`EM-652 was more potent in repressing ER«& activity
`(Csg=2 nM) than ICI182,780 (Cs 9=20 nM). Both
`
`ERoa, ERB
`
`Ligand-
`dependent
`+
`
`coactivators
`
`Ligand-
`independent
`Constitutive
`
`But activated by
`dopamine
`growth factors
`PKAactivators
`
`Cyclic AMP
`Ras-MAPkinase pathway
`
`Fig. 6. Schematic representation of the activation functions 1 and 2
`of ERa and ER. AF-1 is ligand-independent but is activated by
`dopamine, growth factors, cyclic AMP, MAPkinase, PKA activators
`and RAS. AF-2 on the other hand, is activated by estrogenic com-
`pounds. EM-652 blocks both AF-1 and AF-2 completely while OH-
`Tamoxifen blocks AF-2 only.
`
`antiestrogens were more effective to inhibit ERf than
`ERe« function with ICs9 values of 0.4 nM and 8 nM
`for EM-652 and ICI182,780, respectively. In addition,
`lower concentrations of EM-652 in the 107'°-107'! M
`range contributed already to a 25-30% reduction in
`the E, response of both ERs, and even when added at
`1073 M, EM-652 already showed a 20-25% repres-
`sion (data not shown).
`
`4. EM-652 inhibits both AF-1 and AF-2 functions of
`ERo and ERB
`
`As mentioned above, the two ERs share many func-
`tional characteristics based on their well conserved
`modular structure (Fig. 6). AF-2 is responsible for es-
`trogen-dependent activation through recruitment of
`coactivator proteins including members of the steroid
`receptor coactivator (SRC) family [85,87—93]. On the
`other hand, AF-1 activity is constitutive and ligand-
`independent [79-81].
`In addition to the classical hormoneactivation path-
`way, a numberofsteroid receptors including ER and
`B have been shown to be activated by non steroidal
`agents (Fig. 6) including dopamine, growth factors and
`PKAactivators [65,94—98].
`
`4.1. EM-652 inhibits RAS-induced transcriptional
`activity of ERa and ERB
`
`Potential phosphorylation of serine 118 in human
`ERe« [96,99,100] and serine 60 in mouse ERf [65]
`through activation of the Ras-MAPK_ pathway has
`been shown to further maximize the E, response of
`both estrogen receptors. To investigate whether EM-
`652 could efficiently block this effect, we used the wild-
`type H-Ras and its dominant active form H-Ras”!? in
`
`AstraZeneca Exhibit 2034 p. 6
`
`
`
`fF. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`37
`
`B
`
`A
`
`fle
`my
`:
`ogy SEEERGDEF GST-mEAIEF
`input
`Hi Controi
`
`
`Es BibsoeSesl le SE te om te ee a
`
`
`
`
`
`
`Oo
`
`
`
`foldactivation 3a
`
`a RXRe—_ASMSSMHAiAN 3
`xe 0 -12-11-10 -9 -8 -7
`
`eMs520- ee ee
`
`wra7eg
`
`|= - - -
`
`
`
`
`
`
`
`
`
`
`
`
`O (AB110 of «8 -
`a eed
`sthagenial canbead logit}
`
`128 4
`
`
`
`o B
`
`aan
`
`cS
`
`ofeaxiealreaperpee ealoe
`
`652 (filled squares) and ICI 182,780 (open squares) in the presence of
`10 nM E> on ER« activity in COS-1 cells
`transfected with
`vitA,ERETKLuc reporter and SRC-1 expression plasmid. The maxi-
`mal induction by E, alone wasset arbitrarily at 100%. (G), same as
`in (F) except that ERf expression vector was used [105].
`
`AstraZeneca Exhibit 2034 p. 7
`
`.
`
`E2+
`F2 EM-652 For"
`
`-
`
`Fe EM-652 Tei
`
`w 125
`on
`
`D
`
`11182,780
`
`1C1182,780
`
`68
`
`100
`2= 75
`
`o £
`
`50©
`£
`‘5 25
`
`0 -12-11-10 -9 -8 -7
`
`antagonist concentration (logM)
`
`Fig. 7. EM-652 blocks the Ras-induced ERa and ERftranscriptional
`activity.
`(A) COS-1
`cells were
`cotransfected with
`1
`ug
`vitAZERETKLuc and 500 ng pCMX-ERz«in the presence or absence
`of 1 yg Ha-Ras or Ha-Ras‘!? expression plasmids. The cells were
`then grown in the presence or absence of 10 nM E> or 100 nM of
`EM-652 or ICI 182,780 (ICI). The basal activity of ERa in the
`absence of estradiol was set arbitrarily at 1.0. (B), same as in (A),
`except that ER# expression vector was used. (C). Dose responses of
`EM-652 (filled squares) and ICI 182,780 (open squares) in the pre-
`sence of 10 nM E, on ER«activity in COS-1 cells transfected with
`vitA;ERETKLuc reporter and Ha-Ras’? expression plasmid. The
`maximal induction by E, alone was set arbitrarily at 100%.
`(D)
`same as in (C) except that ERB expression vector was used [105].
`
`
`
`Fig. 8. EM-652 blocks the estrogen and SRC-1-stimulated AF2 ac-
`tivity of ERa and ER. (A) GST pull-down experiments. The puri-
`fied fusion proteins were incubated with labeled SRC-1 in the
`absence (lanes 3 and 7) or presence of 5 nM Ez (lanes 4-6 and 8-10)
`in addition to a 100-fold excess of EM-652 (lanes 5 and 9) and ICI
`182,780 (lanes 6 and 10). The input lane (lane 1) represents 20% of
`the total amount of labeled SRC-1 used in each binding reaction. An
`equivalent amount of protein was used in the sample containing only
`GST (lane 2).
`(B) COS-1 cells were cotransfected with 1
`ug
`vitAZERETKLuc and 500 ng pCMX-ERz«in the presence or absence
`of 1 ug SRC-1 expression plasmid. Cells were incubated with or
`without 10 nM E> or 100 nM antagonist as indicated. Results are
`expressed as fold response over basal levels set arbitrarily at 1.0. (C),
`same as in (B), except that ERf expression vector was used. (D) and
`(E), same as in (B) and (C) respectively, except that pS2Luc reporter
`and HeLacells were used in transfections. (F), dose response of EM-
`
`
`
`58
`
`F. Labrie et al. | Journal of Steroid Biochemistry and Molecular Biology 69 (1999) 51-84
`
`our transfection experiments, as indicated in Fig. 7. As
`observed previously [65,100],
`the addition of H-Ras
`contributed to increase the activity of ERa in the pre-
`sence of E5, with an even stronger response when H-
`Ras’ ’? was used (Fig. 7A). These inductions by both
`Ras forms were completely abolished with the addition
`of EM-652 in the medium, as with ICI 182,780,
`suggesting that EM-652 is effective in blocking the
`AF-1 activity of ER. The same experiment was also
`conducted on ERf where H-Ras and H-RasY”” aug-
`mented the Ez response in a similar manner (Fig. 7B).
`Again, EM-652 and ICI 182,780 abolished the Ras
`effect on ERin the presence of E>. Interestingly, we
`observed a ligand independent effect of Ras on ERP
`basal activity where a 2—3-fold induction occurred
`with H-Ras’’” (Fig. 7B). On the other hand, noeffect
`of Ras was seen on basal levels of ERa. The Ras in-
`duction of unliganded ERf was blocked by EM-652
`and ICI 182,780 (data not shown). We werealso inter-
`ested to test whether EM-652 wasefficient in blocking
`ER responsiveness on a natural promoter. The pS2
`promoter has been extensively studied in respect to its
`ERe mediated regulation [101]. We previously showed
`that ERf can also modulate transactivation of a repor-
`ter gene driven by the pS2 promoter in HeLa cells,
`and that the E, response was potentiated by H-Ras
`[65]. The effects of Ras on liganded ER« and f activi-
`ties are completely abrogated by EM-652 (data not
`shown). Dose response analyses were also performed
`to further evaluate the potency of EM-652 to inhibit
`the effect of Ras on ER activities in the presence of
`E,. EM-652 was
`slightly more effective than ICI
`182,780 in blocking H-RasY’* inductions of ERa and
`ER§, especially at lower concentrations (Fig. 7C and
`D).
`
`4.2. EM-652 blocks SRC-1 induced activity of both
`ERa and ERB
`
`The co-activator SRC-1 has been shown to interact
`with and promote the transcriptional activity of a
`number of nuclear receptors including ER« [85,102].
`More recently, we have demonstrated that SRC-1 also
`stimulates ERP activity through a direct
`interaction
`with its ligand-binding domain (LBD) where the AF-2
`domain resides [65]. We took advantage of this effect
`of SRC-1 to study whether EM-652 could block the
`E,-activated AF-2 function of ERe and ER#.
`Wefirst generated glutathione-S-transferase (GST)
`fusion proteins with the E and F domains of mERf
`(GST-mERf EF) and domains D-F of mER« (GST-
`mERe DEF) for use in GST-pull down experiments
`(Fig. 8A). GST-mERf# EF and GST-mER« DEF were
`expressed in FE. coli, purified with GST-Sepharose and
`incubated with [°S] methionine labeled SRC-1. As
`shown in Fig. 8A,
`the LBD of mER« interacted
`
`(lane 3)
`weakly with SRC-1 in the absence of E,
`whereas addition of E, caused an increase in inter-
`action between the two proteins (lane 4). Both EM-652
`(lane 5) and ICI 182,780 (lane 6) efficiently blocked the
`ligand-dependent SRC-1 interaction, with a stronger
`effect for EM-652. A similar
`inhibition of the E,-
`dependent interaction between SRC-1 and the LBD of
`ERf was also observed whereas ICI182,780 was less
`efficient
`(see Fig. 8A)
`lanes 7-10). We also demon-
`strate that the stimulatory effect of SRC-1 on the E,
`response of both ERs in COS-1 cells was completely
`abolished with the addition of EM-652 in the medium
`as did ICI 182,780 at the concentration used (Fig. 8B,
`C). Furthermore, as observed with Ras (see above),
`SRC-1, under
`the present experimental conditions,
`enhanced the basal activity of ERf but not that of
`ER« in the absence of ligand. This ligand-independent
`effect of SRC-1 on ERf was blocked by EM-652.
`Similar results were obtained using HeLacells trans-
`fected with a pS2Luc reporter construct (Fig. 8D, E).
`Dose response analyses were also performed to
`further evaluate the potency of EM-652 to inhibit the
`potentiating effect of SRC-1 on ER activities in the
`presence of E,. EM-652 was very effective in blocking
`SRC-1 potentiation of ligand-dependent ERe and ERf
`transcriptional activities with apparent ICs9 values of
`10~'° M and 107° M, respectively (Fig. 8F and G).
`ICI 182,780 was less potent to inhibit the SRC-1 in-
`duction of ERa and ERfactivities with an ICs value
`of 10~° M for both receptors.
`The present study describes the molecular action of
`EM-652, the active metabolite of EM-800, on ER tran-
`scriptional functions. We present evidence that EM-
`800, and its active metabolite EM-652, act as pure es-
`trogen antagonists on ERa« and ER/# transcriptional
`activities. This pure antiestrogenic profile is of primary
`Importance in endocrine-based breast cancer therapy,
`since the objective, as mentioned earlier, is to develop
`a compound having both activities, while the widely
`available antiestrogen currently available, Tamoxifen,
`acts as a mixed agonist—antagonist on ER function
`and does not inhibit the AF-1 function. Besides a rela-
`tively good clinical record in inducing remission of
`ER-positive breast cancer and in post-surgical adju-
`vant therapy, resistance to tamoxifen, a phenomenon
`likely due to its