`© 1998 Elsevier Science Inc. All rights reserved.
`
`
`
`ELSEVIER
`
`ISSN 0006-2952/98/$19.00 + 0.00
`PIT S0006-2952(98)00574-1
`
`Clomiphene Analogs with Activity In Vitro and In
`Vivo against Human Breast Cancer Cells
`R. Jeffrey Baumann,*} Tammy L. Bush,* Doreen E. Cross-Doersen,*
`Elizabeth A. Cashman,* Paul S. Wright,* John H. Zwolshen,* Gregory F. Davis,*
`Donald P. Matthews,# David M. Bender§ and Alan J. Bitonti*
`*OncoLocy, Horcust Marion RousseL, BripcewaTerR, NJ 08807; {ENpocRINE DEPARTMENT, Et Litty &
`Co., INDIANAPOLIS, IN 46285; AND §DEPARTMENT OF CHEMISTRY, COLORADO STATE UNIVERSITY,
`Fr. Cotzins, CO 80523, U.S.A,
`
`ABSTRACT.Six hundred triphenylethylenes were assayed for antiproliferative activity against MCF-7, LY2,
`and MDA-MB-231 breast cancer cells using sulforhodamine B dye to measure proliferation. Here we report on
`just 63 of the compounds, mostly clomiphene analogs, with substitutions on the a’ or B ring, at the vinyl position
`or in the side chain, of which 23 were active,as defined by antiproliferation 1C;, values =1 uM. Activity profiles
`showed that 23 and 11 analogs were active toward MCF-7 and LY2, respectively, but none were active against
`MDA-MB-231. The ICsq values of tamoxifen were 2.0 uM against MCF-7 and 7.5 uM against LY2 and
`MDA-MB-231. Estradiol reversed antiproliferative activities of several E isomers but not their Z isomer
`counterpatts. Clomiphene side chain analogs 46 [(E)-l-butanamine, 4-[4-(2-chloro-1,2-diphenylethenyl)
`phenoxy]-N,N-diethyl-dihydrogen citrate (MDL 103,323)] and 57 [(E)-N-[p-(2-chloro-1,2-diphenylvinyl)
`phenyl]-N,N-diethylethylenediamine dihydrogen citrate (MDL 101,986)] were 4- to 5-fold more effective than
`tamoxifen. Methylene additions up to (-CH,-),,
`in the clomiphene side chain showed that analog 46
`[(-CH,-),4 side chain] had maximal antiproliferative activity, bindingaffinity, and inhibition of transcription of
`an estrogen response element luciferase construct in transfected MCF-7 cells. Intraperitoneal administration of
`46 ot 57 inhibited progression of MCF-7 breast tumor xenografts in nude mice with EDs, values of <0,02
`mg/mouse/day. Both analogs may hold promise for treating ER positive breast cancer and are of interest for
`further development.
`BIOCHEM PHARMACOL 55;6:841-851, 1998. © 1998 Elsevier Science Inc.
`
`KEY WORDS. clomiphene analogs; antiestrogens; breast cancer; MCF-7; LY2; tamoxifen resistance; estrogen
`receptor
`
`The number of new cases of breast cancer diagnosed every
`year is approximately 180,000, making this disease the most
`prevalent cancer in women[1]. Breast cancer is a hormone-
`dependent cancer, and an estimated 70% of tumors are
`positive for the ER! [2]. Two classes of antiestrogen drugs
`known to antagonize the growth of hormone-dependent
`breast cancer cells are steroidal antiestrogens such as ICI
`164,384 [3] and nonsteroidal antiestrogens such as the TPEs
`clomiphene [4],
`toremifene [5], and tamoxifen [6, 7].
`Tamoxifen is used worldwide by over a million women for
`the treatment of ER positive breast cancer. Its continued
`use has reduced both the annual rate of death and disease
`
`+ Corresponding author: Dr. R. Jeffrey Baumann (c/o Dr. Paul S. Wright)
`Oncology, Hoechst Marion Roussel Route 202-206N, Bridgewater, NJ
`08807. Tel. (908) 231-4000, FAX: (908) 231-2727.
`Abbreviations: CSCS, chatcoal sttipped calf serum; ER, estrogen
`receptor; FBS, fetal bovine serum; HBSS, Hanks’ Balanced Salt Solution;
`IMEM,Improved Minimum Essential Medium, Eagle’s; MTG, monothio-
`glycerol; MTT,3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
`mide; NSB, nonspecific binding; RBA, relative binding affinity; SRB,
`sulforhodamine B dye; tam, tamoxifen citrate; and TPE, triphenylethyl-
`ene.
`
`Received 21 January 1997; accepted 12 September 1997.
`
`recurrence among breast cancer patients [8], and it has a
`low incidence of short- and long-term side-effects. How-
`ever, tamoxifen resistance eventually develops, resulting in
`the failure of tamoxifen therapy, thus creating the need for
`additional nontoxic therapeutic modalities.
`Prior to the development of tamoxifen, thefirst reported
`active
`nonsteroidal
`antiestrogen,
`ethamoxytriphetol
`(MER-25), was synthesized at the Wm. S. Mertell Co. [9]
`and subsequently shown to have antifertility activity [10]
`and antitumoractivity [11]. These findings were the incen-
`tive for more intensive efforts leading to the synthesis of
`clomiphene [12], Upjohn’s nafoxidine [13], and later ta-
`moxifen at Imperial Chemical Industries [14]. For a review
`of antiestrogens, see Ref. 15. Clomiphene is marketed by
`Hoechst Marion Rousselas a fertility agent, but it has been
`used in clinical trials for the treatment of breast cancer[16].
`Cumulative results were published from various clinical
`trials, and objective responses were noted in 28% of late
`stage breast cancer patients given clomiphene, which com-
`pared favorably with responses to tamoxifen (27%)
`in
`unselected patients [17]. Because of our earlier efforts in
`antiestrogen research, approximately 600 TPEs were avail-
`
`AstraZeneca Exhibit 2068 p. 1
`InnoPharma Licensing LLC v. AstraZeneca AB IPR2017-00904
`Fresenius-Kabi USA LLC v. AstraZeneca AB IPR2017-01910
`
`
`
`842
`
`R. J. Baumannet al.
`
`able in the Hoechst Marion Roussel chemical inventory.
`Having set goals to seek compounds significantly more
`potent than tamoxifen, and with efficacy against tamox-
`ifen-resistant breast cancer cells, we screened the TPEs for
`antiproliferative activity against three humanbreast cancer
`cell
`lines: MCF-7 (ER positive,
`tamoxifen-susceptible),
`LY2 (ER positive, tamoxifen-resistant), amd MDA-MB-231
`(ER negative, tamoxifen-resistant).
`In this paper, we present
`results on 63 TPEs with
`substitutions in the a ring side chain, or on the e’ or B ring,
`ot at the vinyl Cl of clomiphene. Of the 63 analogs, 23 were
`found to have antiproliferative ICs) values =1 pM, and
`many had ICs9 values lower than that of tamoxifen toward
`MCE-7 and LY2. Furthermore, several of the 23 analogs
`wete separated into pure E and Z isomers to define isomer
`specificities in terms of biological and biochemical activi-
`ties including antiproliferative assays, antitumor activity in
`nude mice, and ER relative binding activities.
`MDL 103,323 [(E)-1l-butanamine, 4-[4-(2-chloro-1,2-
`diphenylethenyl)phenoxy|-N ,N-diethyl-dihydrogen
`_ci-
`trate] and MDL 101,986 [(E)-N-[p-(2-chloro-1,2-diphe-
`nylvinyl)phenyll-N,N-diethylethylenediamine dihydrogen
`citrate] are E isomers of the clomiphene analogs 46 and 57,
`respectively. Both showed significant antitumor activity
`against MCF-7 human tumor xenografts in nude mice with
`EDsq values <0.02 mg/mouse/day, administered orally for 6
`weeks. In addition, the antiproliferative activities of both
`isomers were several-fold better than tamoxifen toward
`MCE-7 and LY2cell lines. The biological activities of these
`analogs suggest potential utility for the treatment of estro-
`gen-dependent breast cancer.
`
`MATERIALS AND METHODS
`Cell Lines
`
`MCE-7 (ATCC HTB 22) and MDA-MB-231 (ATCC
`HTB 26) were obtained from the American Type Culture
`Collection. LY2 cells were provided by Dr. Mare Lippman
`[18]. Cells were maintained in Costar T75 flasks containing
`IMEM without phenol red (Biofluids, Inc.) supplemented
`with 4 mM glutamine and 5% FBS (Gibco BRL).
`
`TPEs
`
`All compounds were analyzed for structural integrity and
`spectral purity by NMR prior to their use in assays. The
`majority of TPEs had been synthesized as mixtures of
`isomers, and were screened without further modifications.
`However, several of the isomer mixtures that showed
`interesting biological activity were resolved into pure iso-
`mers and retested. To obtain pure E and Z isomers, isomeric
`mixtures were made basic with 2 N NaQH,extracted with
`chloroform, and analyzed by HPLC on a Porasil column
`(Waters) monitored by UV at 270 nm. The mobile solvent
`was hexane:chloroform:triethylamine (20:80:0.02). For pre-
`parative HPLC, a 19 X 300 mm semipreparative column
`was used with a flow of 15 mL/min and an injection volume
`
`of 50 wL containing 25 mg of compound. The individual
`isomer peaks were collected and identified by NMR and
`mass spectrometry. The E isomer of 22 (4-hydroxyclomi-
`phene) isomerized to an approximately equal mixture of E
`and Z isomers within 2 weeks in hexane/CHCl,, but not in
`DMSO.Binding assays with the E isomer were performed
`within 3 days of its dissolution in DMSO. Tamoxifen
`citrate (tam) and 4-hydroxytamoxifen citrate were ob-
`tained from ICI America, Inc.
`
`Antiproliferation Assay
`
`SRB stains protein and is used to measure cell growth.
`Because the SRB assay is suitable for large-scale screening
`with several practical advantages over the MTTassay, the
`National Cancer Institute adopted this assay for use in
`routine antiproliferative screening.
`Antiproliferative assays were performed using SRB
`(Sigma) as described [19] with modifications. Cells were
`harvested when nearly confluent from IMEM/FBS using
`trypsin/EDTA, washed once with serum-free IMEM, and
`resuspended in IMEM/FBS. Stock drug solutions (10 mM)
`were prepared in DMSO and diluted with serum-free
`IMEM. Drug dilutions and all additions of drugs, cells, and
`medium to microtiter wells were made with a Perkin Elmer
`Cetus PRO/PETTE. Aliquots (100 wL) of 1 x 10* MCF-7
`cells or 3 X 10° LY2 or MDA-MB-231 cells were dispensed
`in duplicate into 96-well microtiter plates and incubated at
`37° in 5% CO, for 20-24 hr, and the medium wasreplaced
`with 100 wL of IMEM/FBS containing drug concentrations
`from 0.078 to 10 pM in duplicate. After 4 days of
`incubation, the medium and drugs were replaced. After a
`total of 8 days of incubation, the medium was removed and
`the cell monolayers were fixed for 60 min at 4° with 100 wL
`of 10% trichloroacetic acid, rinsed five times with water,
`and dried. The fixed cells were stained for 30 min at room
`temperature with 100 wL of 0.4% SRB in 1% acetic acid,
`rinsed four times with 1% acetic acid, and dried, and the
`SRB was extracted for 5 min with 100 wL of 10 mM Tris
`base, pH 10.5. Absorbances were determined at 492 nm
`with a Titertek Multiscan MCC/340 plate reader. Concen-
`tration—response curves were constructed to estimate IC5o
`values, defined as the micromolar concentration of drug
`inhibiting 50% of proliferation. To determine the effect of
`estradiol on ICs values, compounds were assayed in me-
`dium supplemented with 0.1 2M estradiol (Sigma).
`The following guidelines were used for making com-
`parisons on various compound activities. Active com-
`pounds had antiproliferative 1Csy values =1 uM toward
`any cell line, compounds were selective for MCF-7 or LY2
`if antiproliferative 1Csy values against
`the cell
`lines
`differed by =5-fold, and estradiol reversal of growth
`antagonism was positive if the Ics9 in estradiol-supple-
`mented medium was =3-fold the ICs) determined in
`unsupplemented medium.
`
`AstraZeneca Exhibit 2068 p. 2
`
`
`
`Clomiphene Analogs Active against MCF-7 and LY2 Cells
`
`843
`
`Extraction of ER
`
`MCF-7 or LY2 cells were cultured for 15-30 passages in
`IMEM supplemented with 5% CSCS (Cocalico Biologi-
`cals) and 4 we/mL bovine insulin (Gibco BRL), since
`preliminary assays indicated the ER yield was 2- to 3-fold
`greater if CSCS was substituted for FBS. The monolayers
`were tinsed with HBSS (Gibco BRL), scraped into HBSS
`containing 0.1% (v/v) MTG,and centrifuged for 10 min at
`800 g. To extract total ER (cytosolic + nuclear), cells were
`resuspended in 2 packed cell volumes ofhighsalt extraction
`buffer [10% (v/v) glycerol, 500 mM KCI, 25 mM HEPES
`buffer, pH 7.8] [18, 20, 21], frozen and thawed three times,
`mixed for 30 min at 4°, and centrifuged at 4° for 30 min at
`12,000 g. Supernatants were retained as the source of ER
`and stored at —80°.
`
`Relative Binding Affinities
`
`RBAs were determined in 96-well microtiter plates with
`conical wells [22]. Drugs were prepared as 10 mM stock
`solutions in DMSO, and further dilutions were made with
`TrisEDTA buffer (TE buffer) containing the following
`supplements and final concentrations of 8 mM Tris, pH 7.4,
`1 mM EDTA,0.4% BSA [23], 12.5% (v/v) dimethylform-
`amide, 0.1% (v/v) MTG and 2 nM [2,4,6,7-Hlestradiol,
`114 Ci/mmol
`(Amersham). Cell extracts (15 wL) were
`added to begin the assay in final volumes of 100 wL, in
`triplicate, and incubated at 4° for 16-18 hr. Receptor
`bound [H]estradiol was separated from unbound [PHlestra-
`diol with 100 wL of TE buffer, pH 7.4, containing 0.1%
`(v/v) MTG, 0.5% BSA, 0.05% dextran T70, and 0.5%
`Norit A at 4° for 15 min and centrifuged at 4° for 20 min
`at 1200 g. The mean netdisintegrations per minute were
`determined in 160 wL of supernatant by subtracting the
`mean of the NSB (NSB = dpm boundin the presence of 1
`»M nonradioactive estradiol). The ics) values were esti-
`mated from percent control versus concentration curves,
`and RBAs were calculated from the expression:
`
`ABA ICsq estradiol
`ICsy TPE
`
`X 100, according to Korenman[24].
`
`Concentration—response curves of several E isomers were
`analyzed for parallelism using Graph Pad software (Prism
`Version 2.01).
`
`Transfection
`
`To determine whether isomers of side chain analogs inhib-
`ited expression of an estrogen responsive gene, the pGL
`2-basic vector (Promega) was digested with Smal and Xhol,
`and a DNA fragment containing two copies of the vitel-
`logenin estrogen response element [25], adjacent to a 180
`bp fragment encoding the thymidine kinase promoter[26],
`was inserted upstream to the luciferase gene. This plasmid,
`pVETLUC, was provided by Drs. Steven Busch and Gary
`
`Martin. MCF-7 cells were transfected with the pVETLUC
`plasmid by electroporation. Cells (2 x 10°) were combined
`with 50 pg of plasmid DNA in 1 mL of OPTI-MEM 1
`mediumin an electroporation chamber (Gibco BRL). The
`suspension was subjected to a charge of 500 V/cm, 800
`microfarads, at O° and low resistance. Following a 1-min
`recovery period,
`the cells were resuspended in growth
`medium, viability was assessed by trypan blue exclusion,
`and cells were dispensed into 96-well microtiter plates at
`approximately 1 X 10* cells/well. The culture medium was
`replaced with fresh medium after 4 hr of incubation at 37°,
`and after 24 hr with fresh medium containing 1 nM
`estradiol and side chain-extended analogs and incubated for
`18-22 hr. The cells were tinsed once with HBSSand,after
`freezing at —70°
`for 15 min, 150 pL of lysis buffer
`(Promega) was added and the plates were agitated for 20
`min at ambient temperature. The lysates were analyzed for
`luciferase (Promega assay system) in a luminometer. The
`ICsq values were determined from log-log curve fits using
`Biolinks software from Dynatech.
`
`Antitumor Effects of TPEs
`
`Nude mice were housed in microisolator cages under
`positive air pressure, and all surgical manipulations and
`drug treatments were performed in a laminar flow cabinet.
`MCE-7 cells (2 X 10°) were inoculated s.c. into the flanks
`of female nu/nu mice, and tumors were allowed to develop.
`Tumors of 400-500 mm? were taken from maintenance
`mice, cut into 2-mm? pieces, and transplanted into the
`flanks of naive mice using a 13-Ga trocar. These xenografts
`were allowed to grow to volumes of 50-100 mm’, at which
`time mice (N = 6) were assigned randomly to control or
`drug treatment groups. To assess tumor growth and the
`effects of TPEs, tumors were measured weekly with Vernier
`calipers in two dimensions as described previously [27].
`TPEs were administered daily, 5 days/week, as a solution in
`6% ethanol, 4% Tween 80, 0.8% NaCl, and 0.68 mM citric
`acid (0.2 mL/dose) [28].
`
`RESULTS
`Substitutions on the a’ and B Rings of 1 and 15, and
`at the Vinyl Cl of 15
`
`The compounds in Table 1 consist of two groups, analogs of
`1,R = H (2-14) and analogs of 15, R = Cl (clomiphene)
`(16-25). The necessity of Cl for activity (ICs) =1 uM, see
`Materials and Methods) is clearly shown, since compound
`15 was ten and five times more active than compound 1
`toward MCF-7 and LY2, respectively. Further comparisons
`show that just 2 analogs of 1 (7 and 9) and 6 analogs of 15
`(16, 19, and 22-25) were active. Compound 26 was also
`active, but it is the HCI salt of clomiphene, not a clomi-
`phene analog. Very little improvement in the activity of 15
`against MCF-7 was generated by various substitutions at R,
`or R, except for the hydroxy analog (22) which was 10
`times more active than 15 and 100 times more active than
`
`AstraZeneca Exhibit 2068 p. 3
`
`
`
`844
`
`R. J. Baumannet al.
`
`TABLE 1. Antiproliferative activities of analogs substituted on the a’ and PB rings of 1 and 15
`
`ON es
`XQ
`
`IC59* (uM)
`
`DYZ
`
`MDA-MB-231
`
`
`
`
`MCE-7
`
`8.0
`4.0
`2.0
`5.8
`1
`13
`0.85
`3,4
`1.0
`4.1
`Bd:
`3.0
`3.0
`Zl
`0.8
`3.0
`LS
`145
`0.5
`1.2
`
`8.0
`4.0
`3:0
`6.2
`Bd
`4.5
`3.0
`25
`Dil
`6.2
`BZ
`3.5
`Asd
`5.4
`1.5
`1.0
`2.3
`3:0
`1.9
`33
`
`ND§
`ND
`4.5
`ND
`ND
`ND
`ND
`ND
`6.0
`6.2
`ND
`ND
`ND
`ND
`72
`> 10
`6.6
`4.0
`3.6
`ND
`
`R
`
`© 3
`
`Bi
`
`R,
`
`H
`CH;
`OCH:
`Cl
`F
`C(CH3)5
`Biphenyl
`H
`OCH,
`Cl
`H
`OCH,
`Cl
`H
`H
`OH
`OCH;
`Cl
`Cl
`C(CH3)5
`
`R,
`
`H
`H
`tial
`H
`H
`H
`H
`CH,
`CH;
`CH;
`cl
`Cl
`Cl
`O-DEAE!
`H
`H
`tial
`EL
`CH;
`H
`
`Compound
`
`1f
`Z
`3
`4
`5
`6
`a
`8
`5
`10
`il
`12
`13
`14
`15
`16
`17
`18
`19
`20
`
`R
`
`Ht
`H
`H
`H
`H
`H
`H
`H
`H
`H
`H.
`H
`H
`H
`Cl
`cl
`Cl
`Cl
`Cl
`Cl
`
`21
`
`22
`23
`24
`25
`26
`
`Cl
`
`Cl
`Cl
`Cl
`Cl
`Cl
`
`aged
`
`H
`H
`H
`OCH,
`H
`
`H
`
`OH
`CH;
`OCH;
`Cl
`EL
`
`i
`
`0.07
`0.8
`0.6
`0.8
`0.6
`
`3.0
`
`1.6
`1.0
`0.8
`0.9
`1.4
`
`ND
`
`> 10
`ND
`6.5
`3.1
`Ta
`
`* Most ICs values =1 j4M are means of at least two experiments.
`+ Compounds are citrate salts except for 1, 2, 4, 7, 13, 19, 25, and 26 which are hydrochloride salts and 6 which is a free base.
`£ Isomer configuration: isomeric mixtures except for 1, 4, 6, 13, and 14 are one isomer of unknown configuration; 7 is E configuration; 10 and 11, isomer status unknown.
`§$ ND = not determined.
`"O(CH,)2N(CjHs)2.
`
`1. Furthermore, compound 22 was about 40 times more
`active than compound 16 against MCF-7, demonstrating
`the superior antiproliferative activity of the B-4-OH over
`the a’-4-OH.
`Several analogs of 15 with various substitutions for the
`vinyl Cl are shown in Table 2. The only improvements in
`antiproliferative activity seen with these analogs were with
`compounds 27, 33, and 34, which were at best 2-fold more
`active than 15 against LY2 cells.
`
`Variations on the Alkyl Ether Side Chain
`
`A number of substitutions were made on the alkyl ether
`side chain of clomiphene, as shown in Table 3. Compounds
`38 and 40 are included for comparison to show that
`
`without side chains the structures were devoid of activity
`against MCF-7 or LY2. The unsubstituted vinyl compound
`(39) is shown for comparison to 51 to emphasize the
`potency differential between 39 and 51, that is, 51 was
`more active than 39 by a factor of 21. In reference to the
`17 analogs (41-57), 11 showed either the same activityas
`15, or were more active than 15, against MCF-7. Among
`those were clomiphene side chain analogs that markedly
`affected antiproliferative activity. For example, variations
`on the alkylamino groups showed the monoethyl analog of
`clomiphene (43), as well as clomiphene,
`to be approxi-
`mately 8-fold more active than the dipropyl analog of
`clomiphene (41). In addition, compound 46, whichdiffers
`from 15 by an extension of 2 methylene groupsin the side
`chain producing a butyl chain, was more active than 15
`
`AstraZeneca Exhibit 2068 p. 4
`
`
`
`Clomiphene Analogs Active against MCF-7 and LY2 Cells
`
`845
`
`TABLE2. Antiproliferative activity of vinyl substituted clomiphene analogs against breast cancer cells
`OLSy
`
`R
`
`
`TCgo (uM)
`
`Compound*
`Rt
`MCE-7
`LY2
`MDA-MB-231
`
`1§
`15§
`27
`28
`29
`30
`31
`32
`33
`34
`35
`36
`37
`
`H
`Cl
`Br
`F
`NO,
`CN
`CONH,
`CH,
`CH,CH,
`(CH,),CH;
`(CH,),CH,
`O(CH,)N(C,Hs)5
`C.Hs
`
`8.0
`0.8
`0.8
`AT
`2.0
`6.0
`25
`2.4
`07
`1.0
`23
`6.0
`3.0
`
`8.0
`1.5
`on
`Da
`2.0
`5.0
`2.5
`LA
`0.7
`1.0
`25
`6.0
`3.5
`
`* Compoundsare citrate salts except for 33 which is an HCIsalt, 36 whichis an oxalate salt, and 31 and 37 whichare free bases.
`+ Isomer configurations: isomer mixtures except for 30 and 35 are one isomer of unknown configuration, and the isomer status is unknown for 36.
`f cso values =1 uM are meansof at least two determinations.
`§ Compounds from Table 1 included hete for comparison.
`"ND = not determined.
`
`
`ND
`
`ND!
`ted
`6.4
`ND
`ND
`ND
`ND
`ND
`5.4
`52
`ND
`ND
`
`against both MCF-7 and LY2. Therefore, to investigate the
`biological effects of longer side chains, additional analogs
`with side chains extending from 5 to 12 methylene groups
`were synthesized. E and Z isomers of the analogs were
`purified and assayed for antagonism of MCF-7 growth,
`estradiol-enhanced expression of luciferase in transiently
`transfected MCF-7 cells and estrodial binding to MCF-7
`ER, as shown in Table 4. Maximal activity against cell
`growth,
`luciferase expression, and in the competitive
`binding assay correlated with the E isomer of the 4
`carbon side chain analog (46). Side chain extensions
`longer than the butyl side chain did not improve activity,
`and regardless of side chain length the Z isomers were
`uniformly less active.
`Further observations from Table 3 show that the activity
`of clomiphene against MCF-7, but not LY2, was either
`maintained or improved upon by a number of analogs with
`heterocyclic groups such as pyrrolidyl (47), piperidyl (48,
`51) and 4-methylpiperazinyl (50). Therefore, we synthe-
`sized analogs with pyrrolidyl
`(52), piperidyl
`(53) or
`4-methylpiperazinyl (54) groups and butyl side chains in
`anticipation that
`the butyl side chain would enhance
`antiproliferative activity. However, activities were not
`improved upon, since the ICsp values of the butyl side chain
`analogs were substantially greater than those of the respec-
`tive ethyl side chain counterparts. Additional side chain
`analogs substituting O with C (55), S (56), or N (57)
`
`produced activities the same as or 2—3 times greater than
`that of clomiphene against MCF-7. However, the activities
`of these analogs against LY2 were the same as or not as
`potent as clomiphene.
`
`Estradiol Reversal of Activities of Pure Isomers
`and of 22
`
`Pure E and Z isomers of several analogs were assayed for cell
`growth antagonism and reversal by estradiol, as shown in
`Table 5. While both isomers of each analog elicited some
`degree of growth antagonism,estradiol reversal was positive
`if estradiol supplementation increased the ICs, by =3-fold
`over the ICs) in unsupplemented medium. Thus, growth
`antagonism by E isomers 15, 23, 24, and 27 was reversed
`in LY2 cells but not in MCF-7 cells. However, growth
`antagonism by E isomers 46, 48, 55, 56, and 57 was
`reversed in both cell lines. None of the Z isomer-induced
`growth antagonism was reversed by estradiol.
`As mentioned above, compound 22 (4-hydroxyclomi-
`phene) had the lowest Icsq of all clomiphene analogs tested
`against MCF-7. The data in Fig. 1 show a biphasic concen-
`tration—response curve of compound 22 mediated growth
`antagonism and a monophasic concentration response
`showing reversal by estradiol. Reversal of antagonism was
`complete at concentrations of 22 up to 2.5 wM, partial at
`5 pM, but no reversal was seen at 10 ~M. Several E isomers
`
`AstraZeneca Exhibit 2068 p. 5
`
`
`
`TABLE3. Antiproliferative activity of side chain analogs of clomiphene
`
`R,
`
`R
`
`Ry
`
`
`ICsot (uM)
`
`
`Compound*+
`R
`Ri
`R3
`MCE-7
`LY2
`MDA-MB-231
`38
`> 10
`= 1D
`NDS§
`
`H
`
`H
`
`H
`
`39
`
`40
`41
`42
`43
`44
`45
`46
`
`47
`
`48
`
`49
`
`50
`
`1
`
`52
`
`H
`
`cl
`Cl
`Cl
`cl
`cl
`cl
`cl
`
`Cl
`
`cl
`
`Cl
`
`cl
`
`cl
`
`cl
`
`OCH, OLANN
`
`»
`
`H
`H
`H
`H
`H
`H
`H
`
`H
`
`H
`
`H
`O(CH,),N(C3H;)>
`O(CH,),NH,
`O(CH,),NHCH,CH,
`OCH,CHCH,N(CH,CH,),
`O(CH,);N(CH,CH;)>
`O(CH,),N(CH,CH,),
`
`O
`
`O
`
`(
`
`\
`
`“Nn
`
`H
`
`H
`
`Oo
`“—\
`Se“Nn
`L_/?
`“—~\
`“nN
`L_NCHs
`OCH, OL™y
`
`O
`
`21
`
`> 10
`5.4
`7.0
`0.9
`1.0
`0.97
`0.64
`
`0.6
`
`0.2
`
`3.0
`
`01
`
`O41
`
`3.5
`
`> 10
`25
`8.0
`5.0
`3.0
`4.5
`0.6
`
`2.0
`
`1.0
`
`3.0
`
`14
`
`0.9
`
`H On es)
`N
`
`ND
`
`ND
`
`ND
`5.8
`ND
`> 10
`6.0
`ND
`7.8
`
`6.6
`
`7.4
`
`> 10
`
`> 10
`
`ND
`
`ND
`
`53 H OLA™4,3Cl ND ND
`
`
`
`
`
`
`N
`
`
`
`54
`
`55
`56
`57
`58
`59
`15||
`
`cl
`
`cl
`cl
`cl
`Tamoxifen
`4-Hydroxytamoxifen
`cl
`
`N
`H OLA /
`NCH3
`N
`\_/f
`(CH,);N(CH,CH,),
`S(CH,),N(CH,CH,)>
`N(CH,),N(CH,CH;)5
`
`H
`H
`H
`
`H
`
`O(CH,),N(CH,CH;)>
`
`> 10
`
`0.3
`04
`0.7
`Deh
`0,03
`0.8
`
`ND
`
`1.6
`1.0
`4.0
`tS
`0.62
`15
`
`ND
`
`7.0
`5.9
`6.2
`Ea
`ND
`72
`
`* Compoundsare citrate salts except for 39, 42, 49 and 51 which ate HCl salts.
`+ Isomer configurations: all isomeric mixtures except for the configurations of 39 and 42 are unknown; 51 is an E isomer; 58 and 59 have E configurations, but E/Z priority
`tules enforce Z isomer designations.
`+ Most ICs values =1 pM are means of at least two determinations.
`§ND = not determined.
`| Clomiphene.
`
`AstraZeneca Exhibit 2068 p. 6
`
`
`
`Clomiphene Analogs Active against MCF-7 and LY2 Cells
`
`847
`
`TABLE 4. Biological and biochemical activities of isomers with consecutive side chain extensions
`
`Nc) -N
`
`a“
`
`nN
`
`Cl
`
`
`E isomers*
`Z isomers
`
`Compound
`(CH3),
`Growtht
`pVETLUC§
`RBA|
`Growth?
`pVETLUC§
`RBA!
`15
`2
`22+ 0.33
`05: + 0.02
`Lé+ 62
`3.8
`2A
`0.13
`45
`3
`0.7 + 0.1
`0.2 + 0.06
`6.7 + 1.4
`NAT
`NA
`NA
`46
`4
`0.56 + 0.1
`0.1 + 0.03
`92 12
`4.2
`2
`0,24
`60
`5
`1.4 +02
`0.3 + 0.04
`LO = 0,1
`ia
`8
`0.11
`61
`6
`2. £05
`0.6 + 0,09
`i+ 08
`4.6
`9
`C17
`62
`7
`18+ 0.9
`t+ 15
`29 + O1
`2
`9
`0.21
`63
`8
`1.8 + 0.7
`L2.+ 02
`0.4 + 0.03
`3.4
`7
`0.07
`64
`9
`19+08
`1,3: O2
`0.22 + 0.03
`27
`5
`0.12
`65
`10
`1.6+0.7
`1.8 + 0.06
`0.07 + 0.003
`31
`18
`0.07
`66
`11
`2. £05
`6.6 + 13
`0.04 + 0.003
`3.8
`21
`0.04
`67
`12
`3.3 + 0.4
`12+14
`0.02 + 0.0
`5.4
`14
`0.02
`tam
`NA
`NA
`NA
`NA
`2.23: = 0.3
`0.7 + 0.08
`1.3 +0.4
`
`* Means + SEM, where E isomers were tested 3 times in each assay except that 15, 46 and tam were tested in 8, 13, and 9 antiproliferation assays, respectively.
`+Z isomers were tested once.
`= MCF-7 antiproliferation 159 (uM).
`$1Cs59 (uM) antagonizing expression of luciferase.
`Relative binding affinity for MCF-7 ER.
`TNA = notavailable.
`
`showed biphasic concentration-response curves similar to
`that of compound 22 (data not shown).
`
`TABLE5. Estradiol reversal of the antiproliferative effects of
`E and Z isomers
`
`Correlations of RBA and Ics, Values
`
`RBA values were determined for several compounds by
`competition with PH] estradiol for MCF-7 ER and com-
`pared with antiproliferative 1C;, values for those analogs, as
`shown in Table 6. The RBA valuesare listed in descending
`order and correlate with an ascending order of the ics
`values. This type of pattern has been reported previously for
`MCF-7 cells
`[29-32]. Analyses of the concentration—
`response curves of the analogs in Table 6 showed all curves
`to be parallel (slope values were in a range of —25 to —43
`and were not significantly different), indicating that the
`analogs bind to the samesite in a non-cooperative manner.
`Compound 22 (4-hydroxyclomiphene) and 59 (4-hy-
`droxytamoxifen) showed the highest RBA and the lowest
`ICsg values; however, compound 22 is a mixture of E and Z
`isomers. Therefore,
`to determine whether the elevated
`binding activity of 22 was due primarily to the E isomer,
`the isomers of compound 22 were purified by HPLC and
`the configurations were verified by NMR. The RBA values
`were 285 for the E isomer and 16 for the Z isomer (data not
`shown). Paradoxically, the compound with the lowest RBA
`
`
`IC5q (4M)
`
`MCE-7
`LY2
`
`
`Compound
`Isomer
`—E2
`+E2*
`—E2
`+E2*
`
`15
`
`23
`
`24
`
`2a
`
`46
`
`48
`
`55
`
`56
`
`57
`
`58 (Tam)
`
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`Zt
`
`22
`4.5
`1.0
`2.9
`0.76
`3.4
`1.1
`Pek
`0.56
`23
`0.76
`1.9
`0.62
`pai
`0.45
`2.0
`O25
`os
`Lad
`
`4.4
`TO
`22
`Bd.
`ne
`4.1
`Sul
`1.8
`35
`2.9
`Zul
`1,7
`3.9
`3.5
`3,0
`25
`3.0
`25
`5.8
`
`Le
`3.4
`0.72
`2.6
`0.62
`of
`0.62
`2.6
`0.62
`3.5
`0.4
`1.8
`0.31
`4.2
`0.9
`2.9
`0.45
`Te
`13
`
`56
`43
`4.2
`3.8
`4.0
`4.8
`4.6
`2.9
`5.6
`5,8
`3.6
`Zuo
`6.5
`6.2
`4.8
`3.4
`52
`8.4
`7.9
`
`*IMEMsupplemented with 0.1 wM estradiol.
`+ Tamoxifen is a trans structure, but E/Z priority rules enforce the Z designation.
`
`AstraZeneca Exhibit 2068 p. 7
`
`
`
`848
`
`
`
`PERCENTCONTROL
`
`100
`
`80
`
`60
`
`40
`
`DRUG +
`ESTRADIOL
`
`20
`
`0.01
`
`0.1
`
`1
`
`10
`
`4-HYDROXYCLOMIPHENE (uM)
`
`FIG. 1. Antiproliferative activity of 4-hydroxyclomiphene (22)
`and reversal by 0.1 pM estradiol. MCF-7 cells (1 x 10*/well)
`were dispensed into 96-well microtiter plates, and after 24 and
`96 hr the medium and drugs were renewed. After a total of 7
`days of incubation,
`the cells were fixed and stained with
`sulforhodamine B, and the stain was extracted from the cells to
`determine absorbancies (492 nm) and percent control values.
`
`and the highest Icsy (16) is also a 4-hydroxyclomiphene,
`but
`in 16 the hydroxy substitution is on the B ting
`(4'-hydroxyclomiphene) not the a’ ring as in 22.
`
`Antitumor Activity
`
`Several analogs active in the antiproliferative assay were
`tested for activity toward MCF-7 tumor xenografts in nude
`mice. The results in Table 7 show that among the five
`
`of ER binding
`6. Correlation
`TABLE
`antiproliferative effects in MCF-7 cells
`
`affinities with
`
`ICsof
`
`Compound
`Isomer
`RBA*
`(4M)
`22
`B+2Z
`251
`0.07
`59
`E
`246
`0.03
`55
`E
`18
`0.6
`57
`E
`2
`0.25
`46
`E
`9
`0.56
`56
`E
`6
`O5
`51
`B
`5
`0.1
`48
`EB
`5
`0.8
`33
`E+2Z
`a
`0.7
`24
`EB
`2.6
`0.8
`15
`E
`1.6
`2.
`27
`EB
`1.6
`lel
`23
`E
`L2
`Ll
`58
`va
`Lo
`2
`16
`B+2Z
`<0.4
`5
`
`* Relative bindingaffinities for MCF-7 ER.
`+ Antiproliferative activity toward MCF-7 cells.
`
`R. J. Baumannet al.
`
`TABLE7. Inhibition of MCF-7 tumor progression in mice by
`antiestrogens
`
`EDso
`
`Compound*
`(mg/mouse)*
`
`15 (clomiphene)
`46
`55
`56
`57
`Tamoxifen
`
`0,22
`<0.02
`0,085
`0.29
`20.07
`0,07
`
`*Tsomer configuration, E: 15,46, 55 and 57; EF + Z: 56.
`+ Tumor pieces were implanted into nude mice flanks by trocar; when tumor
`volumes reached 50-100 mm’,
`treatment was begun by daily intraperitoneal
`administrations, 5 days/week for 6 weeks; N = 5—6 mice per treatment group.
`
`compoundstested, 46 and 57 were greater than 3-fold more
`active than tamoxifen and greater than 10-fold more active
`than clomiphene.
`
`DISCUSSION
`
`In an antiproliferative screen of nearly 600 TPEs, we
`discovered clomiphene analogs with better activity in vitro
`and in vive than clomiphene or tamoxifen. Of the 63
`compounds presented here, 23 were active (ICs) values <1
`uM) toward the growth of MCF-7 cells and 11 toward LY2
`cells. TPE antiestrogens reported to have the most potent
`antiproliferative activity toward breast cancercells contain
`side chains on the « ring [33], a feature commonto all the
`compounds we identified as active. In addition, TPE anti-
`estrogen-mediated growth antagonism is known to be
`reversible by estradiol [8, 34-36], and the growth antago-
`nism by E isomers that we report here was reversed by
`estradiol supplementation as manifested by 3- to 20-fold
`increases in the ICs) values. Finally, it has been reported
`that TPE antiestrogens with the lowest antiproliferative
`ICsq values have the highest RBA values for the ER [8, 31,
`37]. We found this correlation in a group of 15 compounds,
`mainly E isomers of clomiphene analogs.
`Maximal antiproliferative and ER binding activity was
`shown with compound 22 (4-hydroxyclomiphene) against
`MCF-7 cells and with MCF-7 ER,respectively, (ICs, 0.07
`wM; RBA, 251). Also, several TPEs, including compound
`22, induced biphasic antiproliferation profiles with estradi-
`oL-reversible and -nonreversible components. Sutherlandet
`al. [29] reported 4-hydroxyclomipheneto be the most active
`of several hydroxy TPEs against MCF-7, and that biphasic
`concentration-response profiles are helpful in defining dif-
`ferent mechanisms of antiestrogen growth antagonism [29,
`32]. The estradiol-reversible component suggests competi-
`tion between the inhibitor and estradiol for binding to the
`ER [29-31], and the estradiol-nonreversible component
`may involve either inhibition of protein kinase C [38, 39]
`or calmodulin-dependent enzymes [29, 40], as
`the ics
`values of antiestrogens against these enzymes are in micro-
`molar concentrations.It is apparent that the mechanism of
`growth antagonism by our most active analogs was compe-
`
`AstraZeneca Exhibit 2068 p. 8
`
`
`
`Clomiphene Analogs Active against MCF-7 and LY2 Cells
`
`849
`
`versus LY2 remains unexplained but may be of interest for
`tition for the ER, since the activities were reversed by
`estradiol. In addition, we tested 19 of the 23 active analogs
`consideration of mechanisms of antiestrogen resistance.
`against the ER negative cell line MDA-MB-231 andall
`Clomiphene can sustain more substitutions in the side
`were inactive; therefore, it seems unlikely that the antipro-
`chain than at the vinyl Cl position without concomitant
`liferative effects toward MCF-7 and LY2 were due to
`reductions in antiproliferative activity toward MCF-7, For
`nonspecific cytotoxic effects. Models have been proposed
`example, of the 17 clomipheneside chain analogs (40-50,
`for the binding of estradiol and antiestrogens to the ER.
`52-57), 3 had about the same activity as clomiphene, 7
`The model proposed by Katzenellenbogenetal. [41] suggests
`were more active, and 7 were less active than clomiphene.
`two majorsites, a ligand binding site and a ligand discrim-
`In contrast, with the 12 vinyl Cl analogs, just 3 (27, 33,
`inating site. The model proposed by Jordan [15] accounts
`and 34) were nearly as potent as clomiphene and the
`for both E and Z isomers, an antiestrogen region that
`remaining 9 analogs were less active than clomiphene.
`accommodates binding of the alkyl ether side chain and a
`Regarding 5 of the side chain analogs more active than
`phenolic site respo