`
`ICI 182,780 on Endometrial Cancer Growth
`
`Ruth M. 0 ’Regan, Angela Cisneros, Gale A/I. England, Jennifer]. MacGregor,
`Henry D. Muenzner, Vasilios J. Assikis, Malcolm A/I. Bilimoria, Michael Piette,
`Yvonne P. Dragan, Henry C. Pitot, Robert Chatterton, V. Craig Jordan
`
`Background: Tamoxifen has been shown to promote the
`growth of human endometrial tumors implanted in athymic
`mice, and it has been associated with a twofold to threefold
`increase in endometrial cancer. Toremifene, a chlorinated
`derivative of tamoxifen, and ICI 182,780, a pure antiestro-
`gen, are two new antiestrogens being developed for the treat-
`ment of breast cancer. The effects of these drugs on endo-
`metrial cancer are currently unknown. Our objective was to
`evaluate the effects of toremifene and ICI 182,780 on the
`growth of human endometrial cancer in athymic mice. Meth-
`ods: Athymic, ovariectomized mice were implanted with hu-
`man endometrial tumors and treated with estrogen, tamoxi-
`fen, or the new antiestrogens. Results: The effects of
`tamoxifen and toremifene on the growth of either tamoxifen-
`stimulated or tamoxifen-naive endometrial tumors in athy-
`mic mice were not substantially different. ICI 182,780 inhib-
`ited the growth of tamoxifen-stimulated endometrial cancer,
`in both the presence and the absence of estrogen. Conclu-
`sions: Toremifene and tamoxifen produce identical effects in
`our endometrial cancer models. Therefore, it is possible that
`toremifene, like tamoxifen, may be associated with an in-
`creased incidence of endometrial cancer. In contrast, ICI
`182,780 inhibited tamoxifen-stimulated endometrial cancer,
`both in the presence and in the absence of estrogen, suggest-
`ing that this drug may be safe with regard to the endome-
`trium, even if it is used following tamoxifen, and that it may
`not result in an increased incidence of endometrial cancer.
`
`Indeed, it is even possible that ICI 182,780 may prove useful
`as an adjuvant agent in early stage endometrial cancer. [J
`Natl Cancer Inst 1998;90:1552—8]
`
`In 1988, we demonstrated that the antiestrogen tamoxifen
`exhibited target site-specific actions in breast and endometrial
`cancers (1). Athymic mice were co-transplanted with the estro-
`gen-responsive breast tumor, MCF-7, and the estrogen receptor
`(ER)-positive endometrial carcinoma, EnCal0l. Treatment with
`estradiol and tamoxifen demonstrated that the antiestrogen com-
`pletely inhibited the estrogen-stimulated growth of the breast
`tumor but stimulated growth of the endometrial carcinoma (1).
`From these observations, we concluded that women who were
`being treated with long-term adjuvant tamoxifen therapy should
`be screened for pre-existing endometrial carcinoma, which is
`known to be present in five times as many women as is detected
`clinically (2). Although tamoxifen had proven benefits in breast
`cancer at that time (3), we suggested that pre-existing endome-
`trial cancer would not be controlled (1). Our finding of the target
`site-specific actions of tamoxifen was subsequently demon-
`
`strated in patients. Since the original clinical report by Fornander
`et al. (4) in 1989 showing that tamoxifen significantly decreased
`the incidence of contralateral breast cancer but increased the
`
`the topic of the association
`incidence of endometrial cancer,
`between tamoxifen and endometrial carcinoma has been a sub-
`
`ject of intense investigation and some controversy. Recently, we
`surveyed the world literature to determine the extent of the prob-
`lem and to survey gynecologic recommendations based on cur-
`rent knowledge (5). It is clear that tamoxifen causes a twofold to
`threefold increase in the incidence of endometrial cancer (5).
`This increase translates to about two to three cases per thousand
`postmenopausal patients per annum. The disease is the same
`stage and grade as endometrial cancer in the general population
`(5). As a result of the rarity of detection, no special gynecologic
`monitoring, other than routine annual checkups and the follow-
`up of suspicious spotting and bleeding, has been recommended
`(6). Indeed, the International Agency for Research on Cancer
`(IARC), an agency of the World Health Organization, recently
`stated that no patient should stop taking tamoxifen because of
`concerns about the risk of endometrial cancer and that the ben-
`
`efits of tamoxifen use far outweigh any risks (7).
`Concerns about the uterine safety of tamoxifen have naturally
`provoked a search for agents that might control the growth of
`both breast and endometrial carcinomas. Toremifene (Fig. 1), a
`chlorinated derivative of tamoxifen, has shown promise in the
`treatment of advanced breast cancer in postmenopausal women
`(8—I 0). The drug has been evaluated at numerous doses, ranging
`from 60 mg daily to 260 mg daily in postmenopausal women,
`and the general consensus is that responses, particularly in ER-
`positive breast cancer, are equivalent to those seen with tamoxi-
`fen at doses of 20 or 40 mg daily in postmenopausal women
`(I 1). Based on its clinical and toxicologic profiles, toremifene at
`a dose of 60 mg daily has been approved by the U.S. Food and
`Drug Administration for the treatment of advanced breast cancer
`in postmenopausal women.
`ICI 182,780 (Fig. 1) is an example of a pure antiestrogen,
`which, like tamoxifen, acts through the ER but has no demon-
`strated estrogen agonist effects. ICI 182,7 80 inhibits tamoxifen-
`
`Afliliations of authors: R. M. O’Regan (Division of Hematology/Oncology),
`A. Cisneros, J. I. MacGregor, H. D. Muenzner, V. J. Assikis, M. Piette, V. C.
`Jordan (Robert H. Lurie Comprehensive Cancer Center), G. M. England, M. M.
`Bilimoria (Department of Surgery), R. Chatterton (Department of Obstetrics and
`Gynecology), Northwestern University, Chicago, IL; Y. P. Dragan, H. C. Pitot,
`McArdle Laboratory, University of Wisconsin, Madison.
`Correspondence to: V. Craig Jordan, Ph.D., D.Sc., Robert H. Lurie Compre-
`hensive Cancer Center and Northwestern University Medical School, 8258 O1-
`son, 303, E. Chicago Ave., Chicago, IL 60611.
`See “Notes” following “References.”
`
`© Oxford University Press
`
`1552 ARTICLES
`
`Journal of the National Cancer Institute. Vol. 90. No. 20. October 21, 1998
`
`
`
`
`
`9103‘L1ounfII01S9‘n3/{q/310's[I2u1no[p1o;xo'iou[//zduqmoi;poptzommoq
`
`
`
`
`
`|nnoPharma Exhibit 1009.0001
`
`
`
`ICI 132,730
`
`Toremifene
`
`Tamoxifen
`
`Fig. 1. Structures of antiestrogens—tamoXifen, toremifene, and ICI 182,780—
`listed in text. Toremifene and tamoxifen differ only in the presence of a chloride
`group. The methyl groups are designated by two open lines on nitrogen atoms in
`tamoxifen and toremifene. The position —4 is marked by (—X) in these structures,
`where hydroxyl groups are introduced during metabolism.
`
`stimulated breast cancer growth in mice (12). Clinically, it must
`be given by depot intramuscular injection because of low oral
`potency. ICI 182,780 has shown promising results clinically in
`Europe, with high response rates of almost 70% in tamoxifen-
`failed, advanced breast cancer (13), and a large randomized,
`international clinical trial is under way.
`However,
`the endometrial safety of toremifene and ICI
`182,7 80 has not been examined satisfactorily, primarily because
`endometrial cancer is a rare event. Moreover, there are not the
`same stringent requirements for a drug that is used as a palliative
`therapy in advanced disease compared with drugs that are used
`for long-terin adjuvant therapy.
`As a result of claims that toremifene is safer than tamoxifen
`
`because it does not produce liver tumors in rats (14, 15), we have
`used the endometrial cancer model, which was so instructive for
`tamoxifen, to provide information about the potential uterine
`safety of toremifene. Our aim was to replicate the situation seen
`with two clinical scenarios: 1) where toremifene will be used as
`first-line adjuvant therapy and 2) where toremifene will be used
`after adjuvant tamoxifen therapy. In addition, we have compared
`and contrasted the effects of tamoxifen with those of ICI
`
`182,7 80 on the growth of tamoxifen-stimulated endometrial car-
`cinomas. Clearly, an evaluation of uterine safety is important to
`reassure patients in clinical trials.
`
`MATERIALS AND METHODS
`
`Athymic Mouse Model
`
`Six-week-old athymic, ovariectomized mice were implanted with endometrial
`EnCal0l carcinomas (1). These tumors originated from an ER-positive, well-
`differentiated human endometrial tumor (16); they have been serially passaged
`in mice treated with tamoxifen and grow in response to tamoxifen (0.5 mg per
`animal per day) and estrogen (1-cm estrogen capsule per animal given every 6
`weeks) (tamoXifen-stimulated/estrogen-responsive model). A second model
`(tamoxifen-naive/estrogen-responsive) was developed by passaging the tamoxi-
`fen-stimulated (0.5 mg per animal per day) endometrial EnCal0l tumors in mice
`that had been treated with estrogen (1-cm estrogen capsule given every 6 weeks)
`
`and had not been exposed to tamoxifen for at least three passages. These estro-
`gen-stimulated tumors are more responsive to estrogen for growth. Pieces of
`tumor (1 mm“) were implanted bilaterally with a trochar into the mammary fat
`pads.
`The mice were divided into groups of five or 10 and were treated with
`estrogen, antiestrogens, or the vehicle. Silastic estradiol capsules were made as
`described previously (17),
`implanted subcutaneously, and replaced after 6—8
`weeks of treatment. Estrogen capsules were either 1 cm or 0.3 cm in length.
`Tamoxifen and toremifene were each suspended in a solution of 90% CMC
`(1% carboxymethylcellulose in double-distilled water) and 10% PEG 400/Tween
`80 (99.5% polyethyleneglygol 400 and 0.5% Tween 80). Tamoxifen was ad-
`ministered orally, i.e., by mouth, at a dose of 0.5 mg per mouse daily for 5 days
`each week. Toremifene was administered orally at a dose of 0.5, 1.5, or 5 mg per
`animal. ICI 182,780 was dissolved in ethanol and administered in peanut oil
`(following the evaporation of ethanol under N2) to a final concentration of 50
`mg/mL. ICI 182,780 was injected subcutaneously at a dose of 5 mg (0.1 mL
`peanut oil) per animal each week.
`The tumors were measured weekly with calipers. The cross-sectional area was
`determined by use of the following formula: length X breadth/4 X ’1T.
`All procedures involving animals were approved by the Animal Care and Use
`Committee of Northwestern University.
`
`Quantitation of Antiestrogens
`
`The mice were killed, their livers, hearts, and uteri were harvested, and serum
`was obtained by decapitation. Serum samples (150 pL) were deproteinated with
`equal volumes of 100% acetonitrile, followed by centrifugation (Model J2 HC;
`Beckrnan Instruments, Westbury, NY) at 21 200g for 5 minutes at 0 “C. Super-
`natant layers were transferred to vials. Samples were stored at —80 “C.
`Tissue samples (15 mg) were homogenized in 2% acetic acid in methanol
`(vol/vol) and centrifuged at 502g for 10 minutes at room temperature, and the
`supernatant layer was transferred to a glass tube and dried under N2 at 37 “C. The
`precipitates were re-extracted with 100% acetone and centrifiiged at 502g for 10
`minutes at room temperature, and the organic layer was combined with the
`methanolic extract and then redried. Dried samples were reconstituted in their
`respective mobile phases for the high-performance liquid chromatography
`(HPLC) assay (Waters Corporation, Milford, MA) of toremifene and tamoxifen
`(18). Samples were derivatized after separation by an in-column in-line photo-
`chemical reaction, and the highly fluorescent phenathrene derivatives were quan-
`tified by fluorescence detection. Toremifene and metabolites were separated by
`using the Prodigy 5-ODS3 column (Phenomenex, Torrance, CA) (0.1% dieth-
`ylamine [DEA] [Fisher Chemicals, Fairlawn, NJ] in 57% acetonitrile [HPLC
`grade; Fisher Chemicals] in H2O for 15 minutes and 0.1% DEA in 76% ace-
`tronile in H2O at 0.1 mL/minute for 40 minutes) (19). Tamoxifen and metabo-
`lites were separated by column switching to a coupled analytical column (Rex-
`chrom 5 p.-CN; Regis Chemicals, Morton Grove, IL) and eluted by reversed
`phase ion exchange in 34% acetontrile and 66% of 20 p.M potassium dibasic
`phosphate (HPLC grade; J. T. Baker, Phillipsburg, NJ) (pH 3.1) at 1.2 mL/
`minute. Both assays were conducted on Hitachi HPLC systems (Hitachi Instru-
`ments, Inc., San Jose, CA) (20).
`
`Quantitation of Estrogen
`
`levels were assayed in mouse serum by use of a time-resolved
`Estradiol
`immunofluorescence procedure (Delphia assay; Wallac, Gaithersburg, MD).
`Mouse serum gives responses parallel to those of the reference preparation up to
`a concentration of 1300 pg/mL; thereafter, the serum responses are blunted. The
`intra-assay coefficient of variation was 5.2%. All samples were measured in a
`single assay.
`
`Statistical Methods
`
`Differences in the mean tumor area between the treatment and control groups
`were measured by analysis of variance followed by unpaired Student’s t test,
`performed at the last week of each experiment. Significance is reported as
`two-sided P values.
`
`RESULTS
`
`Preliminary data demonstrated that parent toremifene levels
`are low at 24 hours (Table 1), and we have observed that 4-
`hydroxylation is the major route of toremifene metabolism
`
`
`
`
`
`9103‘L1ounfII01S9‘n3/{q/310'sjvuinofpiojxoqouf//zduquloijpoptzojumoq
`
`
`
`
`
`
`
`Journal of the National Cancer Institute, Vol. 90, No. 20. October 21. 1998
`
`ARTICLES 1553
`
`|nnoPharma Exhibit 1009.0002
`
`
`
`Table 1. Levels of toremifene in tissues from mice killed 6 or 24 hours after
`
`Tissue
`
`Serum, ng/mL
`Tumor, ug/g
`Heart, ug/g
`Liver, ug/g
`
`Time, h
`
`final dosing*
`
`6
`
`795 :: 3591‘
`11.1 :: 2.11‘
`6.1:: 3.31‘
`13.0 :: 8.01‘
`
`24
`
`58 :: 39?‘
`2.3 :: 1.01‘
`<1
`<1
`
`*Athymic mice (n = 10) were treated with toremifene at 1.5 mg (60 mg/kg)
`daily for 9 weeks. Levels of toremifene were measured 6 and 24 hours after final
`dosing (five mice per time point).
`‘(Values = mean :: standard deviation.
`
`(mean serum levels :: standard deviation for 4-hydroxy deriva-
`tive at 6 hours = 2879 :: 1647 ng/mL) in athymic mice (data not
`shown).
`To characterize the relative metabolism of tamoxifen and
`
`toremifene, we performed an experiment in which athymic,
`ovariectomized mice without tumors were treated with tamoxi-
`
`fen at a dose of 0.5 mg or 1.5 mg daily or toremifene at a dose
`of 0.5 mg or 1.5 mg daily for 3 weeks. Based on our preliminary
`data, drug levels in serum were measured 4 or 8 hours after final
`dosing. At both doses, serum levels of toremifene were higher
`than those of tamoxifen, and the major route of metabolism for
`both drugs (particularly toremifene) in mice appears to be 4-
`hydroxylation (Table 2). Results for both the 0.5-mg and the
`1.5-mg doses are shown in Table 2.
`At the 0.5-mg doses, parent toremifene levels were signifi-
`cantly higher at 4 hours (P = .02) but not at 8 hours (P = .25),
`compared with parent tamoxifen levels. Levels of the 4-hydroxy
`metabolite were significantly higher for toremifene than for
`tamoxifen at 4 hours (P = .002) and at 8 hours (P = .001).
`There was no significant difference in levels of N-desmethyl
`metabolites between tamoxifen and toremifene at 4 hours (P =
`.17) and at 8 hours (P = .12).
`At the 1.5-mg dose, there was no significant difference be-
`tween parent levels of tamoxifen and toremifene at 4 hours (P =
`.27) and at 8 hours (P = .8). Levels of the 4-hydroxytoremifene
`
`Table 2. Levels of tamoxifen, toremifene, and their metabolites in serum of
`mice 4 and 8 hours after final dosing*
`
`Time after dose
`
`4 h
`
`8 h
`
`Drug and metabolite
`
`Tamoxifen, ng/mL
`Tamoxifen
`N-Desmethyltamoxifen
`4-Hydroxytamoxifen
`Toremifene, ng/mL
`Toremifene
`N-Desmethyltoremifene
`4-Hydroxytoremifene
`
`0.5 mg
`daily‘?
`
`50 :: 27
`60 :: 33
`58 :: 34
`
`1.5 mg
`daily‘?
`
`208 :: 81
`249 :: 73
`198 :: 54
`
`0.5 mg
`daily‘?
`
`58 :: 7
`36 :: 36
`11 :: 19
`
`1.5 mg
`daily‘?
`
`203 :: 100
`431 :: 107
`236 :: 70
`
`135 :: 77
`106 :: 80
`566 :: 301
`
`302 :: 192
`207 :: 120
`1117 :: 781
`
`80 :: 49
`83 :: 68
`331 :: 174
`
`181 :: 151
`161 :: 100
`708 :: 489
`
`were significantly higher at 4 hours (P = .02) but not at 8 hours
`(P = .l5), compared with those of tamoxifen. P values were
`calculated by analysis of variance followed by unpaired Stu-
`dent’s t tests.
`
`To confirm that the 1-cm and 0.3-cm estrogen capsules re-
`sulted in levels of estradiol approximating premenopausal and
`postmenopausal levels, we performed a separate experiment in
`which athymic, ovariectomized mice without tumors were un-
`treated or were implanted with 1-cm or 0.3-cm estradiol capsules
`for 2 weeks (Fig. 2). Mean estradiol levels :: standard errors
`were 379.5 :: 101.2 pg/mL and 83.8 :: 34.6 pg/mL for the 1-cm
`and 0.3-cm capsules, respectively (Fig. 2). The 1-cm capsule
`produces serum estradiol
`levels approximating those in pre-
`menopausal women, which vary throughout the menstrual cycle,
`between 150 pg/mL and 350 pg/mL (21). The 0.3-cm capsule
`results in levels similar to those in postmenopausal women, in
`whom the majority of circulating estrogen is in the form of
`estrone, which is secreted at an average of 35 ptg/day to 40
`ptg/day (22). Although these levels are much higher than physi-
`ologic estrogen levels in mice, we wanted to provide levels
`similar to levels in premenopausal and postmenopausal women
`because the tumors implanted were of human origin.
`For the evaluation of the impact of estradiol and toremifene
`on the growth of a tamoxifen-stimulated endometrial tumor,
`mice were treated with vehicle, with estrogen (1-cm capsule),
`with tamoxifen at a dose of 0.5 mg daily, or with toremifene at
`a dose of 0.5 mg, 1.5 mg, or 5 mg daily. A broad range of
`toremifene doses was used to cover the range used clinically rela-
`tive to tamoxifen, i.e., three to 10 times the dose of tamoxifen.
`There was no significant difference between tamoxifen and
`toremifene (at all three doses) on tumor growth at 9 weeks (P =
`.438) (Fig. 3). Both antiestro gens stimulated tumor growth com-
`pared with that in the untreated animals (P<.05) but to a lesser
`extent than estrogen (P = .02) (Fig. 3).
`We had observed that toremifene produces higher serum lev-
`els than tamoxifen in mice that had not been implanted with
`tumors (Table 2). It was, therefore, possible that lower serum
`
`El Control
`
`I E21cm capsule
`
`E2 0.3cm capsule
`
`
`
`
`
`
`
`9103‘L1ounfuo1s9n3/{q/310's[I2u1no[p1o;xo'rou[//zduqulolgp9pI2o[uAAoq
`
`
`
`
`
`
`
`*Athymic mice (n = 10) were treated with tamoxifen or toremifene at 0.5 mg
`or 1.5 mg per animal (20 mg/kg or 60 mg/kg, respectively) daily for 3 weeks.
`Serum levels of parent drug and metabolites were measured 4 and 8 hours after
`final dosing.
`‘(Values = means :: standard deviation.
`
`Fig. 2. Serum estrogen levels were measured for different capsule sizes. The
`estrogen level (mean :: standard error) for the 1-cm and 0.3-cm estrogen (E2)
`capsules were 379.5 :: 101.2 pg/mL and 83.8 :: 34.6 pg/mL, respectively, after
`implantation of the capsules for 2 weeks in athymic, ovariectomized mice (n =
`10 per group) without tumors.
`
`1554 ARTICLES
`
`Journal of the National Cancer Institute. Vol. 90. No. 20. October 21, 1998
`
`|nnoPharma Exhibit 1009.0003
`
`
`
`
`
`
`
`9103‘L1ounfII01S9‘n3/{q/310's[I2u1no[p1o;xo'rou[//zduqulolgp9pI2o[uAAoq
`
`
`
`
`
`
`
`Control
`
`TAM 0.5mglday
`- E2 1cm capsule
`TOR 0.5mglday
`- TOR1.5mglday
`- TOR Smglday
`
`K
`
`-13- Control
`--<>-- EZ1cm capsule
`_ "A" TOR1.5mglday
`-0- TAM 0.5mglday
`
`.2Tumor
`
`sizecm
`
`2
`
`EU
`
`E’10
`I-
`
`OE5
`
`I-
`
`Fig. 3. Athymic, ovariectomized mice were divided into groups of 10, implanted
`with tamoxifen-stimulated endometrial tumors (1-mmz tumor piece [one piece]
`per mammary fat pad per animal), and treated with estrogen (E2) (1-cm capsule),
`tamoxifen (TAM) at a dose of 0.5 mg (20 mg/kg) daily, or toremifene (TOR) at
`a dose of 0.5 mg, 1.5 mg, or 5 mg daily (20 mg/kg, 60 mg/kg, or 200 mg/kg,
`respectively) or untreated (control). Tamoxifen and toremifene at all three doses
`significantly stimulated tumor growth (P<.05), although to a lesser extent than
`estrogen (P = .02), compared with control. There was no significant difference
`between tamoxifen and toremifene at all three doses (P = .438). The results are
`expressed as means :: standard error. The results were analyzed by analysis of
`variance test followed by unpaired two-sided Student’s t test.
`
`Fig. 4. Athymic, ovariectomized mice were divided into groups of 10, implanted
`with tamoxifen-naive/estrogen-responsive endometrial tumors (1-mmz tumor
`piece [one piece] per mammary fat pad per animal) and were treated with
`estrogen (E2) (1-cm capsule), tamoxifen (TAM) at a dose of 0.5 mg (20 mg/kg)
`daily, or toremifene (TOR) at a dose of 1.5 mg (60 mg/kg) daily or untreated
`(control). Neither tamoxifen nor toremifene significantly stimulated tumor
`growth compared with control (two-sidedP = .09 for tamoxifen; two-sided P =
`.06 for toremifene). The results are expressed as means :: standard error. Error
`bars are not shown for the toremifene, tamoxifen, and control groups because the
`values were too low. The results were analyzed by analysis of variance test
`followed by unpaired two-sided Student’s t test.
`
`levels of toremifene may be associated with less tumor growth
`than lower serum levels of tamoxifen. To examine this possibil-
`ity further, we performed an experiment in which athymic mice
`were implanted with tamoxifen-stimulated/estrogen-responsive
`endometrial tumors and treated daily with tamoxifen at a dose of
`either 0.5 mg or 1.5 mg. The tumor area was measured weekly,
`and serum levels of tamoxifen and metabolites were assayed 4
`hours after the last dosing (Table 3). We were surprised to ob-
`serve that the 1.5-mg dose resulted in less tumor growth than the
`0.5-mg dose, despite higher serum levels (Table 3).
`To evaluate the action of tamoxifen or toremifene on the
`
`growth of tamoxifen-naive/estrogen-responsive endometrial tu-
`mors, we treated the mice with vehicle, with estrogen (1-cm
`capsule), with tamoxifen (0.5 mg daily), or with toremifene (1.5
`mg daily). A ratio of 1 : 3 of tamoxifen to toremifene was chosen
`because clinical trials have demonstrated that 60 mg of toremi-
`fene is equivalent in efficacy to 20 mg of tamoxifen (9). There
`was no significant difference in tumor growth between tamoxi-
`fen and toremifene after 9 weeks of treatment (P = .833) (Fig.
`4). Estrogen significantly stimulated tumor growth compared
`with control (P = .0002); however, in contrast to the tamoxifen-
`
`stimulated/estrogen-responsive model, neither antiestrogen sig-
`nificantly stimulated tumor growth compared with control (P =
`.09 for tamoxifen; P = .06 for toremifene) (Fig. 4).
`Finally, mice (five per group) were implanted with tamoxi-
`fen-stimulated/estrogen-responsive endometrial tumors. The
`mice were treated with vehicle, with postmenopausal levels of
`estrogen (provided by a 0.3-cm estrogen capsule), or with ICI
`182,780 at a dose of 5 mg weekly, with and without estrogen
`(0.3-cm capsule). As can be seen in Fig. 5, estrogen stimulated
`tumor growth compared with control at 10 weeks. However, ICI
`182,7 80 inhibited tumor growth in the presence of estrogen com-
`pared with control (Fig. 5), and ICI 182,780 when given alone
`did not stimulate tumor growth.
`
`D1sCUss1oN
`
`Tamoxifen is an effective therapy approved for all stages of
`breast cancer. Toremifene, or chlorotamoxifen, shows efficacy
`in the treatment of endocrine therapy-naive, postmenopausal pa-
`tients with advanced disease (9); however, it demonstrates cross-
`resistance with tamoxifen, even when high doses (as high as 10
`times the dose of tamoxifen) are administered (23).
`
`Table 3. Tumor growth and serum levels of tamoxifen and metabolites in mice receiving 0.5-mg or 1.5-mg doses per animal per day for 7 weeks*
`
`Tamoxifen and metabolite‘?
`
`Dose, mg
`0.5
`1.5
`
`Tumor area, cm2‘l‘
`1.1:: 0.9
`0.6 :: 0.7
`
`Tamoxifen, ng/mL
`50.9 :: 23
`334.3 :: 60.7
`
`N-Desmethyltamoxifen, ng/mL
`27.6 :: 13.8
`337.7 :: 74.9
`
`4-Hydroxytamoxifen, ng/mL
`64.4 :: 50.4
`477.8 :: 127.9
`
`*Athymic mice (n = 10) were treated with tamoxifen at a dose of 0.5 mg per animal (20 mg/kg) or 1.5 mg per animal (60 mg/kg) daily. The tumor area at 7
`weeks and levels of tamoxifen and metabolites, 4 hours after final dosing, are shown.
`‘(Values = means :: standard deviation.
`
`Journal of the National Cancer Institute, Vol. 90, No. 20. October 21. 1998
`
`ARTICLES 1555
`
`|nnoPharma Exhibit 1009.0004
`
`
`
`
`
`
`
`9103‘L1ounfuo1s9n3/{q/310's[I2u1no[p1o;xo'tou[//zduqulolgpoptzqumoq
`
`
`
`
`
`
`
`Tamoxifen and toremifene clearly have cross-resistance for
`EnCal0l growth. We chose a broad range of oral doses of
`toremifene to ensure that the large doses that have been used
`clinically were not,
`in fact, inhibitory for endometrial cancer
`(9,26). Equivalent, three times, and 10 times the daily dose of
`tamoxifen all supported the growth of the EnCal0l tumors;
`however, in all cases, growth was not as rapid as that observed
`with estradiol. Clearly, the known estrogen-like properties of
`toremifene (2 7) in animals translate to estrogen-like effects to
`support the accelerated growth of pre-existing endometrial can-
`cer.
`
`Much has been made of a potential link between DNA adduct
`formation and the carcinogenesis of high doses of tamoxifen in
`rat liver and the potential for carcinogenesis in humans (28-
`30). Toremifene has not been demonstrated to form DNA ad-
`ducts in rat liver (1 4), and it was thought, therefore, that it would
`be less likely than tamoxifen to result in an increased risk for
`endometrial cancer. Our data suggest that this theory is not the
`case and that toremifene stimulates endometrial tumor growth in
`athymic mice to the same extent as tamoxifen.
`However, there is little evidence for a link between liver
`tumorigenesis in rats and endometrial cancer in women with
`tamoxifen. First, extensive investigations of human metabolism
`and adduct formation have demonstrated that there are funda-
`
`mental differences between rats and humans (31). Second, stud-
`ies of DNA adduct formation with tamoxifen in human liver (32)
`and human uterus (33) have been negative, although an intrigu-
`ing study from Scandinavia (34) suggests uterine adduct forma-
`tion during tamoxifen therapy. Obviously, on the face of it, this
`theory would seem to be of concern, but it is inconsistent with
`the known genesis of human cancer. If the DNA adduct hypoth-
`esis is correct, endometrial cancer would be predicted to occur
`after several years of tamoxifen exposure. Initiation and promo-
`tion of human cancer may require even a decade. However,
`this is inconsistent with the facts. Nearly all
`tamoxifen-
`associated endometrial cancers occur within the first 5 years
`of exposure, and half of them are detected after fewer than 2
`years of treatment. We have suggested that this is consistent with
`the activation and detection of pre-existing disease (5). The
`model would be that estrogen-induced endometrial cancer un-
`dergoes clonal selection during tamoxifen or toremifene treat-
`ment and is subsequently detected on follow-up of gynecologic
`symptoms. In addition, a recent report (35) noted similar chro-
`mosome changes and gene rearrangements in tamoxifen-
`associated and control polyps. If tamoxifen is a carcinogen and
`if endometrial hyperplasia and polyps are part of a stepwise
`process resulting in cancer, tamoxifen-associated polyps should
`have genomic abnormalities different from those of polyps oc-
`curring in patients not receiving tamoxifen (35). Our data sug-
`gest that any woman exposed to tamoxifen, who had a pre-
`existing endometrial cancer, would have continued growth of
`disease during toremifene treatment. This theory is consistent
`with the similar estrogen-like effects of tamoxifen and toremi-
`fene on the human uterus (36).
`In contrast, ICI 182,780 inhibited tamoxifen-stimulated en-
`dometrial growth in the presence of postmenopausal levels of
`estradiol, and, when administered alone, it did not increase the
`growth of endometrial cancer. This observation suggests that,
`even in patients with pre-existing endometrial cancer that has
`
`_
`
`-13- Control
`--<>- E2 0.3cm capsule
`-0- ICI 5mgIweek
`—A-
`ICI 5mglweekIE2 0.3cm capsule
`
`I-
`
`2
`
`Eu
`a:
`.5U’!&
`
`O E3
`
`im-
`Fig. 5. Athymic, ovariectomized mice were divided into groups of five,
`planted with tamoxifen-stimulated endometrial tumors, and treated with estrogen
`(E2) (0.3-cm capsule) or ICI 182,780 (ICI) at a dose of 5 mg (200 mg/kg) given
`as a single subcutaneous dose once weekly, with or without estrogen (0.3-cm
`capsule) or untreated (control). ICI 182,780 inhibited tumor growth both in the
`presence (two-sided P = .10) and in the absence (two-sided P = .17) of
`estrogen, although this did not reach statistical significance because of small
`numbers of animals per group. The results were analyzed by analysis of variance
`test followed by two-sided unpaired Student’s I test.
`
`The major route of metabolism for both antiestrogens
`(tamoxifen and toremifene) in the mouse appears to be 4-
`hydroxylation. It is interesting that tamoxifen seems to be
`cleared more rapidly than toremifene in the mouse, resulting in
`the lower tissue and serum levels seen in mice with and without
`tumors.
`
`The question we wanted to address was whether toremifene
`and ICI 182,780 can support the growth of EnCal0l endometrial
`cancer in athymic mice. This model was used previously to
`justify clinical studies to detect an association between tamoxi-
`fen and endometrial carcinoma in patients receiving adjuvant
`therapy; therefore,
`the current evaluation is important at the
`outset of exposure of patients to any new agent because endo-
`metrial cancer is so rare and, as has been noted with tamoxifen,
`only rigorous studies can detect even a modest association. In-
`deed, numerous early studies from the U.K. (24, 25) showed no
`association between tamoxifen and endometrial cancer; there-
`fore, well-designed preclinical studies are essential to avoid pa-
`tients being inadvertently uninformed of the risks. Clearly, a
`woman should not be led to believe that no risks exist because
`
`inadequate and early clinical studies are being reported.
`In our tamoxifen-naive endometrial cancer model, which
`simulates the antiestrogen-naive woman, neither tamoxifen nor
`toremifene significantly stimulated tumor growth compared with
`control. We used the higher dose of toremifene, since the rec-
`ommended dose is three times that of tamoxifen in the treatment
`
`of breast cancer and we had observed that higher doses of
`tamoxifen, paradoxically, result in less tumor growth (Table 3).
`Our results suggest that, in women who have not been treated
`with tamoxifen, either antiestrogen would be safe, at least ini-
`tially, even if she has pre-existent endometrial cancer. It is likely
`that tumor growth would occur eventually because of clonal
`selection in a continuous antiestrogen environment.
`
`1556 ARTICLES
`
`Journal of the National Cancer Institute. Vol. 90. No. 20. October 21, 1998
`
`|nnoPharma Exhibit 1009.0005
`
`
`
`been stimulated to grow with tamoxifen, ICI 182,780 would be
`safe and would inhibit further endometrial tumor growth. ICI
`182,780 is not cross-resistant with tamoxifen. Studies show that
`ICI 182,780 has no estrogemc actions on the rodent uterus (3 7)
`or on the primate uterus (38), and preliminary screening of
`women who are treated for advanced breast cancer has not dem-
`
`onstrated uterine hypertrophy (I 3). There is every indication that
`ICI 182,780 will control growth of both breast cancer and en-
`dometrial cancer in patients.
`In summary, toremifene appears to have identical effects as
`tamoxifen on the growth of endometrial tumors in athymic mice.
`This observation suggests that, in humans, toremifene will sup-
`port the growth of pre-existent endometrial cancer. In addition,
`if toremifene is ever used as an adjuvant agent, we would an-
`ticipate an increase in the detection of endometrial cancer simi-
`lar to that seen with tamoxifen. In contrast, ICI 182,7 80 inhibits
`endometrial cancer, both in the presence and in the absence of
`estrogen, suggesting that it will prevent further tumor growth in
`patients with tamoxifen-stimulated endometrial cancer. ICI
`182,7 80 should not be associated with an increase in endometrial
`cancer and could even be considered in the treatment of endo-
`metrial cancer.
`
`REFERENCES
`
`(1) Gottardis MM, Robinson SP, Satyaswaroop PG, Jordan VC. Contrasting
`actions of tamoxifen on endometrial and breast tumor growth in the athy-
`mic mouse. Cancer Res 1988;48:812-5.
`(2) Horwitz RI, Feinstein AR, Horwitz SM, Robboy SJ. Necropsy diagnosis of
`endometrial cancer and detection-bias in case/control studies. Lancet 1981;
`2:66-8.
`
`(3) Early Breast Cancer Trialists’ Collaborative Group. Effects of adjuvant
`tamoxifen and of cytotoxic therapy on mortality in early breast cancer. An
`overview of 62 randomized trials among 28,896 women. N Engl J Med
`1988;319:1681-92.
`(4) Fornander T, Rutqvist LE, Cederrnark B, Glas U, Mattsson A, Silfversward
`C, e