`© 1998 Elsevier Science Inc. All rights reserved.
`
`
`
`SI
`
`ISSN 0006~2952/98/$19.00 + 000
`PH $0006~2952(98)00574~1
`
`Clomiphene Analogs with Activity In Vitro and In
`Vivo against Human Breast Cancer Cells
`
`R. Jeffrey Baumann,*”r Tammy L. Bush,* Doreen E. Croseroersen,*
`Elizabeth A. Cashmanf“ Paul S. Wright,* John H. Zwolshen,* Gregory F. Davisf“
`Donald P. Matthews,i David M. Bender§ and Alan]. Bitonti*
`*ONCOLOGY, HOECHST MARION ROUSSEL, BRIDGEWATER, N] 08807; iENDOCRINE DEPARTMENT, ELI LILLY &
`Co., INDIANAPOLIS, IN 46285; AND §DEPARTMENT OF CHEMISTRY, COLORADO STATE UNIVERSITY,
`FT. COLLINS, CO 80523, U.S.A.
`
`ABSTRACT. Six hundred triphenylethylenes were assayed for antiproliferative activity against MCF~7, LYZ,
`and MDA~MB1231 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 04’ or [3 ring, at the vinyl position
`or in the side chain, of which 23 were active, as defined by antiproliferation ICSO values :1 LLM. 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 1C5O values of tamoxifen were 2.0 LLM against MCF~7 and 7.5 LLM against LY2 and
`MDA—MB~231. Estradiol reversed antiproliferative activities of several E isomers but not their Z isomer
`counterparts. Clomiphene side chain analogs 46 [(E)~1~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 (—CHZ—)[2 in the clomiphene side chain showed that analog 46
`[(—CHZ—)4 side chain] had maximal antiproliferative activity, binding affinity, and inhibition of transcription of
`an estrogen response element luciferase construct in transfected MCF—7 cells, lntraperitoneal administration of
`46 or 57 inhibited progression of MCF~7 breast tumor xenografts in nude mice with EDSO values of <0.02
`mg/mouse/day. Both analogs may hold promise for treating ER positive breast cancer and are of interest for
`fuither development.
`BIOCHEM PHARMACOL 556841—851, 1998. © 1998 Elsevier Science Inc.
`
`KEY WORDS. clomiphene analogs; antiestrogens; breast cancer; MCF—7; LYZ; 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 hormoneedependent
`breast cancer cells are steroidal antiestrogens such as ICI
`164,3 84 [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
`
`T Corresponding author: Dr. R. Jeffrey Baumann (c/o Dr. Paul S. Wright)
`Oncology, Hoechst Marion Roussel Route 202~20’6N, Bridgewater, N]
`08807. Tel. (908) 2314000, FAX: (908) 2312727.
`HAbbreviations: CSCS, charcoal stripped 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, Sedimethylthiazol—Z—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 shore and longeterm sideeeffects. Howe
`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, the first reported
`active
`nonsteroidal
`antiestrogen,
`ethamoxytriphetol
`(MEREZS), was synthesized at the Wm. S. Merrell Co. [9]
`and subsequently shown to have antifertility activity [10]
`and antitumor activity [11]. These findings were the incene
`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 Roussel as 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 come
`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 IPR2017-00900
`
`
`
`842
`
`R. ]. Baumann et 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’
`ifenaresistant breast cancer cells, we screened the TPEs for
`antiproliferative activity against three human breast cancer
`cell
`lines: MCF’7 (ER positive,
`tamoxifen’susceptible),
`LY2 (ER positive, tamoxifen’resistant), and MDA’MB’231
`(ER negative, tamoxifen’resistant).
`In this paper, we present
`results on 63 TPEs with
`substitutions in the or ring side chain, or on the (x’ or B ring,
`or at the vinyl Cl of clomiphene. Of the 63 analogs, 23 were
`found to have antiproliferative [€50 values £1 MM, and
`many had IC5O values lower than that of tamoxifen toward
`MCF’7 and LY2. Furthermore, several of the 23 analogs
`were 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)’1rbutanamine, 4’[4’(2’chloroa1,2¢
`diphenylethenyl )phenoxy]/N ,N Idiethyl’dihydrogen
`ci’
`trate] and MDL 101,986 [(E)¢N¢[p¢(2¢chloroa1,2¢diphe»
`nylvinyl)phenyl] IN , N Idiethylethylenediamine 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
`ED5O 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
`
`MCF’7 and LY2 cell lines. The biological activities of these
`analogs suggest potential utility for the treatment of estro’
`gen’dependent breast cancer,
`
`MATERIALS AND METHODS
`Cell Lines
`
`NICE/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. Marc 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 NaOH, extracted with
`chloroform, and analyzed by HPLC on a Porasil column
`(Waters) monitored by UV at 270 nm. The mobile solvent
`was hexanezchloroform:triethylamine (201801002 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 ML containing 25 mg of compound The individual
`isomer peaks were collected and identified by NMR and
`mass spectrometry. The E isomer of 22 (4rhydroxyclomir
`phene) isomerized to an approximately equal mixture of E
`and Z isomers within 2 weeks in hexane/CHCl3, 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.
`
`Antip‘roliferation 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 MTT assay, 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 ML) of 1 X 104 MCF’7
`cells or 3 X 103 LY2 or MDAIMB¢231 cells were dispensed
`in duplicate into 96~well microtiter plates and incubated at
`37° in 5% CO2 for 20—24 hr, and the medium was replaced
`with 100 ML of IMEM/FBS containing drug concentrations
`from 0.078 to 10 uM 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 ML
`of 10% trichloroacetic acid, rinsed five times with water,
`and dried. The fixed cells were stained for 30 min at room
`
`temperature with 100 ML 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 ML of 10 mM Tris
`base, pH 10.5. Absorbances were determined at 492 nm
`with a Titertek Multiscan MCC/340 plate reader. Concena
`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 [€50 values, compounds were assayed in me,
`dium supplemented with 0.1 MM estradiol (Sigma).
`The following guidelines were used for making com;
`parisons on various compound activities. Active com;
`pounds had antiproliferative IC5O values :1 MM toward
`any cell line, compounds were selective for MCFr7 or LY2
`if antiproliferative ICSO values against
`the cell
`lines
`differed by 25’fol‘d, and estradiol reversal of growth
`antagonism was positive if the ICSO in estradiol’supple’
`mented medium was 23’fold the IC5O determined in
`unsupplemented medium.
`
`AstraZeneca Exhibit 2068 p. 2
`
`
`
`Clomiphene Analogs Active against MCF~7 and LY2 Cells
`
`843
`
`Extraction of ER
`
`MCFr7 or LY2 cells were cultured for 15—30 passages in
`IMEM supplemented with 5% CSCS (Cocalico Biologi»
`cals) and 4 [Lg/mL bovine insulin (Gibco BRL), since
`preliminary assays indicated the ER yield was 2’ to 3efold
`greater if CSCS was substituted for FBS. The monolayers
`were rinsed 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 of high salt extraction
`buffer [10% (v/v) glycerol, 500 mM KCl, 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 96ewell microtiter plates with
`conical wells [22]. Drugs were prepared as 10 mM stock
`solutions in DMSO, and further dilutions were made with
`Tris/EDTA 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¢°H]estradiol,
`114 Ci/mmol
`(Amersham). Cell extracts (15 ML) were
`added to begin the assay in final volumes of 100 ML, in
`triplicate, and incubated at 4° for 16—18 hr. Receptor
`bound [°H]estradiol was separated from unbound [°H]estrae
`diol with 100 ML 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 net disintegrations per minute were
`determined in 160 ML of supernatant by subtracting the
`mean of the NSB (NSB = dpm bound in the presence of 1
`MM nonradioactive estradiol). The K350 values were estie
`mated from percent control versus concentration curves,
`and RBAs were calculated from the expression:
`
`Martin. MCF¢7 cells were transfected with the pVETLUC
`plasmid by electroporation. Cells (2 X 10°) were combined
`with 50 [Lg of plasmid DNA in 1 mL of OPTLMEM 1
`medium in an electroporation chamber (Gibco BRL). The
`suspension was subjected to a charge of 500 V/cm, 800
`microfarads, at 0° and low resistance. Following a lamin
`recovery period,
`the cells were resuspended in growth
`medium, viability was assessed by trypan blue exclusion,
`and cells were dispensed into 96ewell microtiter plates at
`approximately 1 X 104 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 chaineextended analogs and incubated for
`18—22 hr. The cells were rinsed once with HBSS and, after
`freezing at —70°
`for 15 min, 150 ML 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
`IC5O values were determined from logelog 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.
`MCF¢7 cells (2 X 10°) were inoculated so into the flanks
`of female nu/nu mice, and tumors were allowed to develop.
`Tumors of 400—500 mm3 were taken from maintenance
`mice, cut into 2emm3 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 mm3, 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].
`
`REA 2
`
`IC5O estradiol
`1cso TPE
`
`X 100, according to Korenman [24].
`
`RESULTS
`
`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 inhibe
`ited expression of an estrogen responsive gene, the pGL
`2ebasic vector (Promega) was digested with Small and Xhol,
`and a DNA fragment containing two copies of the vitele
`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
`
`Substitutions on the a’ and [3 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 (1C5O £1 HM, 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 HCl salt of clomiphene, not a clomie
`phene analog. Very little improvement in the activity of 15
`against MCF¢7 was generated by various substitutions at R1
`or R2 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. ]. Baumann et al.
`
`TABLE 1. Antiproliferative activities of analogs substituted on the OH and B rings of 1 and 15
`
`O\/\N /—
`\_
`
`R
`
`R2
`
`MCF—7
`
`LY2
`
`MDA—MB—23 1
`
`ICso* (MM)
`
`
`
`VD§
`VD
`4.5
`VD
`VD
`VD
`VD
`VD
`6.0
`6.2
`ND
`
`VD
`VD
`VD
`7.2
`> 10
`
`R2
`R1
`
`R1
`
`H
`CH3
`OCH3
`Cl
`F
`C(CH3)3
`Biphenyl
`H
`OCH3
`Cl
`H
`
`OCH3
`C1
`H
`H
`OH
`
`Compound
`
`l'i'
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`
`12
`13
`14
`15
`16
`
`R
`
`Hi
`H
`H
`H
`H
`H
`H
`H
`H
`H
`H
`
`H
`H
`H
`Cl
`Cl
`
`H
`H
`H
`H
`H
`H
`H
`CH3
`CH3
`CH3
`Cl
`
`Cl
`C1
`O—DEAEH
`H
`H
`
`8.0
`4.0
`2.0
`5.8
`1.4
`1.3
`0.85
`3.4
`1.0
`4.1
`3.2
`
`3.0
`3.0
`2.7
`0.8
`3.0
`
`8,0
`4.0
`3.0
`6.2
`3.3
`4.5
`3.0
`2.5
`2.7
`6.2
`3.2
`
`3.5
`4.7
`5.4
`1.5
`1.0
`
`
`
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`Cl
`C1
`C1
`Cl
`C1
`Cl
`Cl
`Cl
`Cl
`Cl
`
`OCH3
`C1
`C1
`C(CH3)3
`\OVQ
`H
`H
`H
`OCH3
`H
`
`H
`H
`CH3
`H
`H
`OH
`CH3
`OCH3
`Cl
`H
`
`1.3
`1.5
`0.5
`1.2
`1.1
`0.07
`0.8
`0.6
`0.8
`0.6
`
`2.3
`3.0
`1.9
`3.3
`3.0
`1.6
`1.0
`0.8
`0.9
`1.4
`
`6.6
`4.0
`3.6
`ND
`ND
`> 10
`ND
`6.5
`3.1
`712
`
`* Most {€50 values :1 11M are means of at least two experiments
`TCompounds 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.
`It lsomer configuration: isomeric mixtures except for 1, 4, 6, 1.3, and 14 are one isomer of unknown configuration; 7 is E configuration; 10 and 11, isomer status unknown,
`§ ND : not determined.
`ll 0(CHZ)ZN(CZH5)Z,
`
`11 Furthermore, compound 22 was about 40 times more
`active than compound 16 against MCF’7, demonstrating
`the superior antiproliferative activity of the 8’4’OH over
`the (x’I4’OH.
`
`Several analogs of 15 with various substitutions for the
`vinyl C1 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 211 In reference to the
`17 analogs (41—57), 11 showed either the same activity as
`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 approxii
`mately 8’fold more active than the dipropyl analog of
`clomiphene (41). In addition, compound 46, which differs
`from 15 by an extension of 2 methylene groups in 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
`
`TABLE 2. Antiproliferative activity of vinyl substituted clomiphene analogs against breast cancer cells
`
`O\/\N’
`
`R
`
`Compound*
`
`R'l'
`
`MCF—7
`
`10501 (pLM)
`
`LY2
`
`MDA—MB—231
`
`VDH
`7.2
`6.4
`VD
`VD
`VD
`
`VD
`VD
`5.4
`5.2
`VD
`VD
`VD
`
`
`
`
`
`1§
`15§
`27
`28
`29
`30
`
`3 1
`32
`33
`34
`35
`36
`37
`
`H
`Cl
`Br
`F
`N02
`CN
`
`CONH2
`CH3
`CHZCH3
`(CH2)2CH3
`(CHZ)3CH3
`O(CHZ)ZN(CZH5)Z
`C6H5
`
`8.0
`0.8
`0.8
`2.7
`2.0
`6.0
`
`2.5
`2.4
`0.7
`1.0
`2.5
`6.0
`3.0
`
`8.0
`1.5
`0.7
`2.3
`2.0
`50
`
`2.5
`2.4
`0.7
`1.0
`2.5
`6.0
`3.5
`
`* Compounds are citrate salts except for 33 which is an HCl salt, 36 which is an oxalate salt, and 31 and 37 which are free bases.
`T lsomer configurations: isomer mixtures except for 30 and 35 are one isomer of unknown configuration, and the isomer smtus is unknown for 36.
`ingO values :1 uM are means of at least two determinations.
`§ Compounds from Table 1 included here for comparison.
`HND : not determined.
`
`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 IC5O 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 IC5O by Z3rfold
`over the [(350 in unsupplemented medium. Thus, growth
`antagonism by E isomers 15, 23, 24, and 27 was reversed
`in LY2 cells but not in MCFr7 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 (4ahydroxyclomia
`phene) had the lowest IC5O 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 MM, partial at
`5 MM, but no reversal was seen at 10 pLM. Several E isomers
`
`AstraZeneca Exhibit 2068 p. 5
`
`
`
`TABLE 3. Antiproliferative activity of side chain analogs of clomiphene
`
`R2
`
`R
`
`R1
`
`R1
`
`R
`
`R2
`
`MCF—7
`> 10
`
`H
`
`H
`
`OCH3 O\/\N >
`
`H
`H
`H
`H
`H
`H
`H
`
`H
`
`H
`
`H
`
`H
`
`H
`O(CHZ)ZN(C3H7)2
`O(CHZ)ZNHZ
`O(CHZ)ZNHCHZCH3
`OCHZCHCH3N(CHZCH3)Z
`0(CHZ)3N(CHZCH3)Z
`O(CHZ)4N(CHZCH3)Z
`
`O
`
`0
`
`<
`
`>
`
`\/\N
`
`o
`
`/\\
`\/\N
`\\/O
`
`O
`
`/\\
`\/\N
`\\/NCH3
`
`OCH3 O\/\N
`
`H OW\
`
`H O\/\/\
`
`N
`
`N
`
`2.1
`
`> 10
`5.4
`7.0
`0.9
`1.0
`0.97
`0.64
`
`0.6
`
`0.2
`
`3.0
`
`0.1
`
`0.1
`
`2.3
`
`4.3
`
`H
`
`H
`
`C1
`C1
`C1
`C1
`C1
`C1
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`Compound*'l'
`38
`
`39
`
`40
`41
`42
`43
`44
`45
`46
`
`47
`
`48
`
`49
`
`50
`
`51
`
`52
`
`53
`
`54
`
`55
`56
`57
`58
`59
`15||
`
`
`
`Icsoi (MM)
`
`LYZ
`>10
`
`MDA—MB—23 1
`ND§
`
`3.5
`
`>10
`2.5
`8.0
`5.0
`3.0
`4.5
`0.6
`
`2.0
`
`1.0
`
`3.0
`
`1.4
`
`0.9
`
`ND
`
`ND
`
`ND
`
`1.6
`1.0
`4.0
`7.5
`062
`1.5
`
`ND
`
`ND
`5.8
`ND
`> 10
`6.0
`ND
`7.8
`
`6.6
`
`7.4
`
`> 10
`
`> 10
`
`ND
`
`ND
`
`ND
`
`ND
`
`7.0
`5.9
`6.2
`7.0
`ND
`7.2
`
`\
`H ON /
`NCH3
`N
`\_./
`
`>10
`
`C1
`C1
`C1
`Tamoxifen
`4~Hydroxytamoxifen
`C1
`
`H
`H
`H
`
`H
`
`(CHZ)3N(CHZCH3)Z
`S(CHZ)ZN(CHZCH3)Z
`N(CHZ)ZN(CHZCH3)Z
`
`O(CHZ)ZN(CHZCH3)Z
`
`0.3
`0.4
`0.7
`2.23
`0.03
`0.8
`
`* Compounds are citrate salts except for 39, 42, 49 and 51 which are HCl salts.
`T lsomer 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
`rules enforce Z isomer designations.
`Ili Most {€50 values £1 uM are means of at least two determinations.
`§ND : not determined.
`ll 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
`
`O\( CH2 > -N
`11 ¥
`
`/_
`
`Cl
`
`
`E isomers*
`Z isomers'l'
`
`Compound
`(CH2)n
`Growthi
`pVETLUC§
`RBAH
`Growthi‘
`pVETLUC§
`RBAH
`15
`2
`2.2 i 0.33
`0.5 i 0.02
`1.6 i 0.2
`3.8
`22
`0.13
`45
`3
`0.7 i 0.1
`0.2 i 0.06
`6.7 i 1.4
`NACIT
`NA
`NA
`46
`4
`0.56 i 0.1
`0.1 i 0.03
`9.2 i 1.2
`4.2
`2
`0.24
`60
`5
`1.4 i 0.2
`0.3 i 0.04
`1.9 i 0.1
`7.2
`8
`0.11
`61
`6
`2.7 i 0.5
`0.6 i 0.09
`2.2 i 0.9
`4.6
`9
`0.17
`62
`7
`1.8 i 0.9
`7.2 i 1.5
`2.9 i 0.1
`2.7
`9
`0.21
`63
`8
`1.8 i 0.7
`1.2 i 0.2
`0.4 i 0.03
`3.4
`7
`0.07
`64
`9
`1.9 i 0.8
`1.3 i 0.2
`0.22 i 0.03
`2.7
`3
`0.12
`65
`10
`1.6 i 0.7
`1.8 i 0.06
`0.07 i 0.003
`3.1
`18
`0.07
`66
`11
`2.7 i 0.5
`6.6 i 1.3
`0.04 i 0.003
`3.8
`21
`0.04
`67
`12
`3.3 i 0.4
`12 i 1.4
`0.02 i 0.0
`5.4
`14
`0.02
`tam
`NA
`NA
`NA
`NA
`2.23 i 0.3
`0.7 i 0.08
`1.3 i 0.4
`
`* Means i 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.
`TZ isomers were tested once.
`513 MCF77 antiproliferation IC50 (MM).
`§IC50 (uM) antagonizing expression of luciferase.
`H Relative binding affinity for MCF’7 ER.
`91 NA : not available.
`
`showed biphasic concentration—response curves similar to
`that of compound 22 (data not shown).
`
`TABLE 5. Estradiol reversal of the antiproliferative effects of
`E and Z isomers
`
`Icso (HM)
`
`MCF—7
`
`LY2
`
`
`
`Correlations of RBA and [€50 Values
`
`RBA values were determined for several compounds by
`competition with [3H] estradiol for MCF¢7 ER and com;
`pared with antiproliferative ICSO values for those analogs, as
`shown in Table 6. The RBA values are listed in descending
`order and correlate with an ascending order of the ICSO
`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 same site in a non’cooperative manner.
`Compound 22 (4ahydroxyclomiphene) and 59 (4’hya
`droxytamoxifen) showed the highest RBA and the lowest
`1CSO 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
`
`Compound
`15
`
`23
`
`24
`
`27
`
`46
`
`48
`
`55
`
`56
`
`57
`
`58 (Tam)
`
`lsomer
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`Z'l'
`
`— E2
`2.2
`4.5
`1.0
`2.9
`0.76
`3.4
`1.1
`2.1
`0.56
`2.3
`0.76
`1.9
`0.62
`2.7
`0.45
`2.0
`0.25
`2.2
`2.23
`
`+ 132*
`4.4
`7.0
`2.2
`2.1
`2.2
`4.1
`3.1
`1.8
`3.5
`2.9
`2.7
`1.7
`3.9
`3.5
`3.0
`2.5
`3.0
`2.5
`5.8
`
`— E2
`1.7
`3.4
`0.72
`2.6
`0.62
`2.7
`0.62
`2.6
`0.62
`3.5
`0.4
`1.8
`0.31
`4.2
`0.9
`2.9
`0.45
`7.2
`7.5
`
`+E2*
`5.6
`4.3
`42
`3.8
`4.0
`4.8
`4.6
`2.9
`5.6
`5.8
`3.6
`2.3
`6.5
`6.2
`48
`3.4
`52
`8.4
`7.9
`
`* IMEM supplemented with 0.1 uM estradiol.
`TTamoxifen is a tram structure, but E/Z priority rules enforce the Z designation.
`
`AstraZeneca Exhibit 2068 p. 7
`
`
`
`848
`
`100
`
`80
`
`DRUG +
`ESTRADIOL
`
`
`
`PERCENTCONTROL
`
`60
`
`40
`
`20
`
`0.01
`
`0.1
`
`1
`
`10
`
`4-HYDROXYCLOMIPHENE (uM)
`
`FIG. 1. Antiproliferative activity of 4—hydroxyclomiphene (22)
`and reversal by 0.1 [LM estradiol. MCF—7 cells (1 X 104/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 ICSO (16) is also a 4’hydroxyclomiphene,
`but
`in 16 the hydroxy substitution is on the B ring
`(4'Ihydroxyclomiphene) not the OL' 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
`
`Icsol-
`
`Compound
`Isomer
`RBA*
`([LM)
`22
`E + Z
`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
`0.5
`51
`E
`5
`0.1
`48
`E
`5
`0.8
`33
`E + Z
`3.3
`0.7
`24
`E
`2.6
`0.8
`15
`E
`1.6
`2.2
`27
`E
`1.6
`1.1
`23
`E
`1.3
`1
`58
`Z
`143
`2
`16
`E + Z
`<04
`3
`
`* Relative binding affinities for MCF’7 ER.
`TAntiproliferative activity toward MCF’7 cells.
`
`R. ]. Baumann et al.
`
`TABLE 7. Inhibition of MCF—7 tumor progression in mice by
`antiestrogens
`
`ED50
`
`Compound*
`(mg/mouse )1‘
`
`15 (clomiphene)
`46
`5 5
`5 6
`5 7
`Tamoxifen
`
`0.2 2
`< 0.02
`0.085
`0.29
`< 0.02
`0.07
`
`* Isomer configuration, E: 15, 4-6, 55 and 57; E + Z: 56.
`[Tumor pieces were implanted into nude mice flanks by trocar; when tumor
`volumes reached 507100 mm},
`treatment was begun by daily intraperitoneal
`administrations, 5 days/week for 6 weeks; N : 576 mice per treatment group.
`
`compounds tested, 46 and 57 were greater than 3’fol‘d 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 W120
`than clomiphene or tamoxifen. Of the 63
`compounds presented here, 23 were active (ICSO values :1
`MM) toward the growth of MCF’7 cells and 11 toward LYZ
`cells. TPE antiestrogens reported to have the most potent
`antiproliferative activity toward breast cancer cells contain
`side chains on the or ring [33], a feature common to 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 1CSO values. Finally, it has been reported
`that TPE antiestrogens with the lowest antiproliferative
`IC5O 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 (4Ihydroxyclomiphene) against
`MCF’7 cells and with MCF¢7 ER, respectively, (ICSO, 0.07
`MM; RBA, 251). Also, several TPEs, including compound
`22, induced biphasic antiproliferation profiles with estradi’
`ol’reversible and Inonreversible components. Sutherland et
`al. [29] reported 4’hydroxyclomiphene to be the most active
`of several hydroxy TPEs against MCF’7, and that biphasic
`concentration»response profiles are helpful in defining difi
`ferent mechanisms of antiestrogen growth antagonism [29,
`32]. The estradiolrreversible 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 l<inase C [38, 39]
`or calmodulin’dependent enzymes [29, 40], as the 1C5O
`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 compel
`
`AstraZeneca Exhibit 2068 p. 8
`
`
`
`Clomiphene Analogs Active against MCF~7 and LY2 Cells
`
`849
`
`tition for the ER, since the activities were reversed by
`estradiol. In addition, we tested 19 of the 23 active analogs
`against the ER negative cell line MDA’MB’231 and all
`were inactive; therefore, it seems unlikely that the antipro’
`liferative effects toward MCE’7 and LY2 were due to
`
`nonspecific cytotoxic effects. Models have been proposed
`for the binding of estradiol and antiestrogens to the ER.
`The model proposed by Katzenellenbogen et al. [41] suggests
`two major sites, a ligand binding site and a ligand discrima
`inating site. The model proposed by Jordan [15] accounts
`for both E and Z isomers, an antiestrogen region that
`accommodates binding of the alkyl ether side chain and a
`phenolic site responsible for low or high affinity binding, to
`which,
`it has been proposed,
`the 3’hydroxy group of
`estradiol binds. While both models have merit, the latter
`model helped put into perspective the markedly different
`RBAs of each 4’hydroxyclomiphene, 16 and 22, Briefly,
`occupancy of the ER phenolic site by the 4’hydroxyl on the
`(x ring of 22 (E isomer) would correlate with high’affinity
`binding; however, with 16 (E isomer)
`the 4»hydroxyl
`resides on the [3 ring,
`thus placing the 4’hy