`© 1998 Elsevier Science Inc. All rights reserved.
`
`
`
`sEvIE
`
`ISSN 0006-2952/98/$19.00 + 0.00
`PII S0006—Z95Z(98)00574-1
`
`Clomiphene Analogs with Activity In Vitro and In
`Vivo against Human Breast Cancer Cells
`
`R. Jeffrey Baumann,*T Tammy L. Bush,* Doreen E. Cross»Doersen,*
`Elizabeth A. Cashman,* Paul 8. Wright,* John H. ZwoIshen,* Gregory F . Davis,*
`Donald P. Matthecws,i David M. Bender§ and Alan]. Bitonti*
`*ONcoLOOY, HOECHST MARION ROUSSEL, BRIDGEWATER, N] 08807; iENDOcRINE DEPARTMENT, ELI LILLY 8:
`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, 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 on’ or [3 ring, at the vinyl position
`or in the side chain, of which 23 were active, as defined by antiproliferation IC5O values S1 ILM. 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 ICSO values of tamoxifen were 2.0 MM against MCF—7 and 7.5 ILM 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-I4'(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 l01,986)] were 4— to 5-fold more effective than
`tamoxifen. Methylene additions up to (—CH2—)1Z 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. Intraperitoneal administration of
`46 or 57 inhibited progression of MCF—7 breast tumor xenografts in nude mice with ED5O 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
`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
`
`T Corresponding author: Dr. R. Jeffrey Baumann (c/o Dr. Paul S. Wright)
`Oncology, Hoechst Marion Roussel Route 202—206N, Bridgewater, N]
`08807. Tel. (908) 231-4000, FAX: (908) 231—2727.
`H Abbreviations: 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, 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»
`8118.
`
`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, the first reported
`active
`nonsteroidal
`antiestrogen,
`ethamoxytriphetol
`(MER—25), 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 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 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 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 Ex. 2068 p. 1
`Mylan Pharms. Inc. V. Astrazeneca AB IPR2016-01326
`
`
`
`842
`
`R. ]. Baumann er 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 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 (X ring side chain, or on the 0:’ or B ring,
`or at the vinyl Cl of clomiphene. Of the 63 analogs, 23 were
`found to have antiproliferative K250 values 51 }LM, and
`many had IC50 values lower than that of tamoxifen toward
`MCF—7 and LY2. Furthermore, several of the Z3 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)»1—butanamin€,
`4~[4~(Z»chloro—1,2—
`diphenylethenyl)phenoxy]»N ,N —diethyl—dihydrogen
`ci-
`trate] and MDL 101,986 [(E)—N»[p-(Z»chloro»1,2—diphe—
`nylvinyl)phenyl]~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
`E1350 values <0.02 mg/mouse/day, administered orally for 6
`weeks. In addition, the antiproliferative activities of both
`isomers were severalvfold 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
`
`MCl:»7 (ATCC HTB Z2) and MDA»MB~Z31 (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 Z N NaOH, extracted with
`chloroform, and analyzed by HPLC on a Porasil column
`(Waters) monitored by UV at 270 nm. The mobile solvent
`was hexanezchloroformztriethylamine (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 ML 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/CHCI3, 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.
`
`Antipmliferation 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 dilutons and all additions of drugs, cells, and
`medium to microtiter wells were made with a Perkin Elmer
`
`Cetus PRO/PETT3. Aliquots (100 pl.) of 1 X 104 l\/ICI:/7
`cells or 3 X 103 LY2 or MDA—MB~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 pL of IMEM/FBS containing drug concentrations
`from 0.078 to 10 p.M 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 pL
`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 pL 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 1C50
`values, defined as the micromolar concentration of drug
`inhibiting 50% of proliferation. To determine the effect of
`estradiol on IC5O values, compounds were assayed in me»
`dium supplemented with 0.1 p.M estradiol (Sigma).
`The following guidelines were used for making com»
`parisons on various compound activities. Active com»
`pounds had antiproliferative K350 values 51 MM toward
`any cell line, compounds were selective for MCF~7 or LY2
`if antiproliferative IC5o values against
`the cell
`lines
`differed by 25»fold, and estradiol reversal of growth
`antagonism was positive if the IC5O in estradiol—supple»
`mented medium was 23-fold the IC5O determined in
`unsupplemented medium.
`
`Astrazeneca Ex. 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 pig/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 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 96»well 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—3H]estradiol,
`114 Ci/mmol
`(Amersham). Cell extracts (15 ML) were
`added to begin the assay in final volumes of 100 pL, in
`triplicate, and incubated at 4° for 16-18 hr. Receptor
`bound [3H]estradiol was separated from unbound [3H]estra—
`diol with 100 pL 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 pl. of supernatant by subtracting the
`mean of the NSB (NSB = dpm bound in the presence of 1
`p.M nonradioactive estradiol). The IC5O values were esti-
`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 pg of plasmid DNA in 1 mL of OBTLMEM 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 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 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 chain»extended analogs and incubated for
`18-22 hr. The cells were rinsed once with HBSS and, after
`freezing at 470° 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
`[C50 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.
`MCF—7 cells (2 X 106) were inoculated s.c. 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 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].
`
`REA :
`
`lcgo estradiol
`IC5O 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 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
`
`Substitutions on the 0:’ and [5 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 (IC5O 51 MM, 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 clomi»
`phene analog. Very little improvement in the activity of 15
`against MCF—7 was generated by various substitutions at R,
`or R2 except for the hydroxy analog (22) which was 10
`times more active than 15 and 100 times more active than
`
`Astrazeneca Ex. 2068 p. 3
`
`
`
`844
`
`R. ]. Baumann er al.
`
`TABLE 1. Antiproliferative activities of analogs substituted on the 01' and B rings of 1 and 15
`
`“‘
`@
`
`R1
`
`R1
`
`H
`CH3
`oCH3
`C1
`F
`C(CH3)3
`Biphenyl
`H
`OCH3
`C1
`H
`
`OCH3
`C1
`H
`H
`oH
`oCH3
`C1
`C1
`C(CH3)3
`
`‘ ,\/©
`H
`H
`H
`OCH3
`H
`
`R
`
`R2
`
`H
`H
`H
`H
`H
`H
`H
`CH3
`CH3
`CH3
`C1
`
`C1
`C1
`O-DEAE“
`H
`H
`H
`H
`CH3
`H
`
`H
`oH
`CH3
`OCH3
`C1
`H
`
`Compound
`IT
`2
`3
`4
`5
`6
`7
`s
`9
`10
`11
`
`12
`13
`14
`15
`16
`17
`18
`19
`20
`
`21
`22
`23
`24
`25
`26
`
`R
`Hi
`H
`H
`H
`H
`H
`H
`H
`H
`H
`H
`
`H
`H
`H
`C1
`C1
`C1
`C1
`C1
`C1
`
`C1
`C1
`C1
`C1
`C1
`C1
`
`O\/\N /—
`
`MCF-7
`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
`1.3
`1.5
`0.5
`1.2
`
`1.1
`0.07
`0.8
`0.6
`0.8
`0.6
`
`1C5o* (MM)
`LYZ
`8.0
`4.0
`3.0
`6.2
`3.3
`4.5
`3.0
`2.5
`2.7
`6.2
`3.2
`
`MDA—MB-231
`\D§
`\D
`4.5
`\D
`\D
`\D
`\D
`\D
`6.0
`6.2
`\D
`
`3.5
`4.7
`5.4
`1.5
`1.0
`2.3
`3.0
`1.9
`3.3
`
`3.0
`1.6
`1.0
`0.8
`0.9
`1.4
`
`\D
`\D
`\D
`7.2
`> 10
`6.6
`4.0
`3.6
`ND
`
`ND
`> 10
`ND
`6.5
`3.1
`7.2
`
`* Most 1C50 values 51 p.M are means of at least two experiments.
`'1‘ 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.
`3; 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.
`H O(CH2)2N(C2H5)23
`
`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 oL'~4—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 LYZ 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 MCFr7 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 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 approxi»
`mately 8»fold more active than the dipropyl analog of
`clomiphene (4-1). In addition, compound 46, which differs
`from 15 by an extension of Z methylene groups in the side
`chain producing a butyl chain, was more active than 15
`
`Astrazeneca Ex. 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
`
`Comp0und*
`
`RT‘
`
`MCF-7
`
`1c5o$ (MM)
`
`LY2
`
`MDA—MB-23 1
`
`1 §
`15 §
`27
`28
`29
`30
`3 1
`32
`33
`34
`35
`36
`37
`
`H
`Cl
`Br
`F
`N02
`CN
`CONHZ
`CH3
`CHZCH3
`(CHZ)2CH3
`(CH2)3CH3
`O(CH2)2N(C2H5)2
`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
`5.0
`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.
`1 [C50 values S1 p.M are means of at least two determinations.
`§ Compounds from Table l included here for comparison.
`“ND : not determined.
`
`ND”
`7.2
`6.4
`ND
`ND
`ND
`ND
`ND
`5.4
`5.2
`ND
`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
`antiproliferatjve 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 Isumers
`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 23—fold
`over the 1C50 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 [C50 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 pLM, partial at
`5 pLM, but no reversal was seen at 10 p.M. Several E isomers
`
`Astrazeneca Ex. 2068 p. 5
`
`
`
`TABLE 3. Antiproliferative activity of side chain analogs of clomiphene
`
`R2
`
`R
`
`R
`
`1
`R1
`
`H
`
`H
`
`R
`
`R2
`
`OCH3 O\/\N >
`
`H
`H
`H
`H
`H
`H
`H
`
`H
`
`H
`
`H
`
`H
`
`H
`O(CH2)2N(C3H7)2
`O(CHZ)ZNHZ
`O(CH2)2NHCH2CH3
`OCH2CHCH3N(CH2CH3)2
`O(CHZ)3N(CHZCH3)Z
`O(CHZ)4N(CHZCH3)Z
`
`O
`
`0
`
`<
`
`>
`
`\/\N
`
`0
`
`/\\
`\/\N
`L/O
`
`O
`
`/\\
`\/\N
`K/NCH3
`
`OCH3 0\/\N
`
`H O\/\/\
`
`H 0\/\/\
`
`N
`
`N
`
`MCF-7
`> 10
`
`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 O\/\/\
`
`N
`
`NCH3
`
`> 10
`
`CompoL1nd*'l'
`38
`
`39
`
`40
`41
`42
`43
`44
`45
`46
`
`47
`
`48
`
`49
`
`50
`
`5 1
`
`52
`
`53
`
`54
`
`H
`
`H
`
`C1
`C1
`C1
`C1
`C1
`C1
`C1
`
`C1
`
`C1
`
`C1
`
`C1
`
`C
`
`C
`
`C
`
`C
`
`1%: (p1M)
`LY2
`> 10
`
`MDA-MB-231
`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
`
`ND
`
`ND
`5.8
`ND
`> 10
`6.0
`ND
`7.8
`
`6.6
`
`7.4
`
`> 10
`
`> 10
`
`ND
`
`ND
`
`ND
`
`ND
`
`H
`H
`H
`
`(CHZ)3N(CH2CH3)Z
`S(CHZ)ZN(CHZCH3)Z
`N (CH2)2N (CH2CH3)2
`
`0.3
`0.4
`0.7
`2.23
`0.03
`0.8
`
`1.6
`1.0
`4.0
`7.5
`0.62
`1.5
`
`C
`55
`C
`56
`C
`57
`Tamoxifen
`58
`4—HydroXytamoXifen
`59
`O(CH2)2N(CH2CH3)2
`H
`C
`15|1
`* Compounds are citrate sa ts except for 39, 4-2, 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.
`3; Most lC5O Values S1 p.M are means of at least two determinations.
`§ND : not determined.
`H Clomiphene.
`
`7.0
`5.9
`6.2
`7.0
`ND
`7.2
`
`Astrazeneca Ex. 2068 p. 6
`
`
`
`Clomiphene Analogs Active against MCE7 and LY2 Cells
`
`847
`
`TABLE 4. Biological and biochemical activities of isomers with consecutive side chain extensions
`.
`°\( CH2 > -N ’
`H
`\
`
`Compound
`15
`45
`46
`60
`61
`62
`63
`64
`65
`66
`67
`tam
`
`(CH2)n
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`NA
`
`Growthi
`2.2 i 0.33
`0.7 i 0.1
`0.56 i 0.1
`1.4 i 0.2
`2.7 i 0.5
`1.8 i 0.9
`1.8 i 0.7
`1.9 i 0.8
`1.6 i 0.7
`2.7 i 0.5
`3.3 i 0.4
`NA
`
`Cl
`
`E isomers*
`
`pVETLUC§
`0.5 i 0.02
`0.2 i 0.06
`0.1 i 0.03
`0.3 i 0.04
`0.6 i 0.09
`7.2 i 1.5
`1.2 i 0.2
`1.3 i 0.2
`1.8 i 0.06
`6.6 i 1.3
`12 i 1.4
`NA
`
`RBA”
`1.6 i 0.2
`6.7 i 1.4
`9.2 i 1.2
`1.9 i 0.1
`2.2 i 0.9
`2.9 i 0.1
`0.4 i 0.03
`0.22 i 0.03
`0.07 i 0.003
`0.04 i 0.003
`0.02 i 0.0
`NA
`
`Growthi
`3.8
`NAql
`4.2
`7.2
`4.6
`2.7
`3.4
`2.7
`3.1
`3.8
`5.4
`2.23 i 0.3
`
`Z isomers’?
`
`pVETLUC§
`22
`NA
`2
`8
`9
`9
`7
`3
`18
`21
`14
`0.7 i 0.08
`
`RBA”
`0.13
`NA
`0.24
`0.11
`0.17
`0.21
`0.07
`0.12
`0.07
`0.04
`0.02
`1.3 i 0.4
`
`* Means i SEM, where E isomers were tested 3 times in each assay except mat 15, 46 and mm were tested in 8, 13, and 9 antiproliferation assays, respectively.
`'l'Z isomers were tested once.
`i MCF»7 antiproliferation IC50 (p.M).
`§ [C50 (p.M) anmgonizing expression of luciferase.
`H Relative binding affinity for MCF-7 ER.
`31 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
`
`Correlations of RBA and IC50 Values
`
`RBA values were determined for several compounds by
`competition with [3 H] estradiol for MCF~7 ER and com—
`pared with antiproliferative IC50 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 K350
`values. This type of pattern has been reported previously for
`MCE7 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 (4»hydroxyclomiphene) and 59 (4—hy—
`droxytamoxifen) showed the highest RBA and the lowest
`1C5O 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
`
`ICSO (uM)
`
`MCF-7
`
`LY2
`
`Isomer
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`E
`Z
`Z1‘
`
`— 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
`
`+ E2 *
`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
`4.2
`3.8
`4.0
`4.8
`4.6
`2.9
`5.6
`5.8
`3.6
`2.3
`6.5
`6.2
`4.8
`3.4
`5.2
`8.4
`7.9
`
`Compound
`15
`
`23
`
`24
`
`27
`
`46
`
`48
`
`55
`
`56
`
`57
`
`58 (Tam)
`
`* IMEM supplemented with 0.1 p.M estradiol.
`TTamoxifen is a tram structure, but E/Z priority rules enforce the Z designation.
`
`Astrazeneca Ex. 2068 p. 7
`
`
`
`R. ]. Baumann et al.
`
`TABLE 7. Inhibition of MCF-7 tumor progression in mice by
`antiestrogens
`
`Compound*
`
`15 (clomiphene)
`46
`55
`56
`5 7
`Tamoxifen
`
`E1750
`(mg/mouse ) ']'
`
`0.22
`<0.02
`0.085
`0.29
`<0.02
`0.07
`
`* Isomer configuration, E: 15, 46, 55 and 57; E + Z: 56.
`TTumor pieces were implanted into nude mice flanks by trocar; when tumor
`volumes reached 50—100 mm3,
`treatment was begun by daily intraperitoneal
`administrations, 5 days/week for 6 weeks; N : 5—6 mice per treatment group.
`
`compounds tested, 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 vii/0 than clomiphene or tamoxifen. Of the 63
`compounds presented here, 23 were active (IC5O values S1
`MM) toward the growth of MCF—7 cells and 11 toward LY2
`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 1C5O values. Finally, it has been reported
`that TPE antiestrogens with the lowest antiproliferative
`[C50 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,
`(IC5O, 0.07
`p.M; RBA, 251). Also, several TP3s, including compound
`22, induced biphasic antiproliferaton profiles with estradi»
`ol»reversible and »nonreversible Components. Sutherland at
`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 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—3 1], and the estradiol»nonreversible component
`may involve either inhibition of protein lcinase C [38, 39]
`or calmodulin—dependent enzymes [29, 40], as the IC5O
`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 Ex. 2068 p. 8
`
`848
`
`100
`
`80
`
`_|
`E
`+Z—
`Q 60
`
`DRUG+
`ESTRADlOL
`
`20
`
`40
`
`0I
`
`-2L
`
`U
`Q)
`n:
`LU
`D.
`
`0.01
`
`0.1
`
`1
`
`10
`
`4-HYDROXYCLOMIPHENE (pM)
`
`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 rnicrotiter 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 IC5O (16) is also a 4—hydroxyclomiphene,
`but
`in 16 the hydroxy substitution is on the B ring
`(4’»hydroxyclomiphene) not the 0:’ 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
`TABLE 6. Correlation
`antiproliferative effects in MCF-7 cells
`
`affinities with
`
`Compound
`22
`59
`55
`57
`46
`56
`51
`48
`33
`24
`15
`27
`23
`58
`16
`
`lsomer
`E + Z
`E
`E
`E
`E
`E
`E
`E
`E + Z
`E
`E
`E
`E
`2
`E + Z
`
`RBA*
`251
`246
`18
`12
`9
`6
`5
`5
`3.3
`2.6
`1.6
`1.6
`1.3
`1.3
`<04
`
`* Relative binding affinities for MCF»7 ER
`TAntiproliferative activity toward MCF-7 cells.
`
`Icsdf
`(pLM)
`0.07
`0.03
`0.6
`0.25
`0.56
`0.5
`0.1
`0 8
`07
`0.8
`2.2
`1 1
`1
`2
`3
`
`
`
`Clomiphene Analogs Active against MCE—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 MCF»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 discrim-
`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
`oc 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»hydroxyl more
`distant from the phenolic sit