Antiproliferative Effects of Idoxifene in a Placebo-Controlled Trial in Primary Human Breast Cancer

2260 Vol. 6, 2260–2267, June 2000 Clinical Cancer Research

Antiproliferative Effects of Idoxifene in a Placebo-controlled Trial in Primary Human Breast Cancer

Mitch Dowsett,1 J. Michael Dixon, INTRODUCTION Kieran Horgan, Janine Salter, Margaret Hills, Idoxifene is a novel SERM2 that recently has been evalu- and Eileen Harvey ated for the treatment of advanced breast cancer (1, 2). Idoxifene is the 4-iodo-substituted analogue of tamoxifen, synthesized Department of Academic Biochemistry, Royal Marsden Hospital, London SW3 6JJ [M. D., J. S., M. H.]; Edinburgh Breast Unit, using rational drug design to decrease the estrogen agonist Edinburgh EH4 2XU [J. M. D.]; Breast Unit, Leeds General activity and increase the estrogen antagonist activity of the Infirmary, Leeds LS1 3EX [K. H.]; SmithKline Beecham parent compound (3). Compared with tamoxifen, idoxifene has Pharmaceuticals, Harlow CM19 5AW [E. H.], United Kingdom been found to be metabolically more stable and to have a higher relative binding affinity for the ER. It also has reduced agonist ABSTRACT activity on breast and uterine cells, greater in vitro activity than tamoxifen, and an antiproliferative effect in tamoxifen-resistant Idoxifene is a novel selective estrogen receptor mod- breast cancer cells (3). ulator. It has reduced agonist activity on breast and At present there are limited data on idoxifene’s anti- uterine cells compared with tamoxifen and antiprolifera- breast cancer activity. In phase I studies, metastatic breast tive effects in tamoxifen-resistant breast cancer cells. Pre- cancer patients have received idoxifene at doses of 10–60 vious studies have shown that a short course of treatment mg/day for 1 week followed by treatment at a dose of 20 with other antiestrogens prior to surgery caused a signif- mg/day until disease progression. Evidence of response was icant reduction of the growth fraction when measured by seen in 2 of 14 patients in one study and 3 of 10 patients in immunohistological staining using the mouse monoclonal a second (2). In a phase II study comparing 40 mg/day antibody Ki67. In this study, we assessed the effect of idoxifene and 40 mg/day tamoxifen, idoxifene was associated idoxifene on biological markers of cell proliferation with evidence of clinical activity in patients with tamoxifen- (Ki67) and apoptosis (TdT-mediated dUTP-biotin nick resistant advanced breast cancer. Similar safety profiles were end labeling), and estrogen and progesterone receptor seen with both compounds (1). (ER/PR) expression was also evaluated. Core-cut biopsies At doses of 5 and 10 mg/day, idoxifene produced a signif- were obtained in 77 postmenopausal patients with pri- icant reduction in biochemical markers of bone turnover in mary breast cancer at diagnosis. Patients were random- postmenopausal women (4). Significant reductions were also ized to 40 mg/day idoxifene or placebo for 14–21 days observed in total cholesterol, low-density lipoprotein, and total prior to obtaining a second biopsy sample at surgical cholesterol/high-density lipoprotein ratio (5). Thus, idoxifene resection. The percentage of Ki67-positive cells fell from a has the properties of a SERM: it expresses antagonist activity in -mean 19.7 ؎ 2.7% (SE) to 13.4 ؎ 3.4% in idoxifene breast cancer cells and a low level of agonist activity on the but ,(0.0043 ؍ P ;30 ؍ treated ER-positive tumors (n uterus (in model systems) coupled with agonist activity on bone there was no significant effect in placebo-treated ER- and lipid metabolism. -No effect was seen on ER .(27 ؍ positive tumors (n It would be valuable to establish during early clinical negative tumors in either group. Idoxifene had no signif- development of SERMs (and other agents) the growth-suppres- icant effect on apoptotic index but produced a statistically sant effects of specific drug dosages and the subpopulations of significant fall in idoxifene-treated ER immunohisto- tumors in which these effects occur. We report a study that chemical score and a small increase in PR that did not explored the efficacy of idoxifene in breast cancer by means of reach statistical significance (0.05 < P < 0.10). Idoxifene the evaluation of biological markers of tumor proliferation was well tolerated in all patients. Idoxifene has an anti- (Ki67 labeling index) and apoptotic index in patients scheduled proliferative effect in ER-positive but not ER-negative for surgery with primary breast cancer. ER and PR expression breast cancers, and no significant effect on apoptosis in were also evaluated both as predictors of response and markers the short-term. of antiestrogenic effects. Earlier studies have shown that a short course of treatment with tamoxifen or the pure antiestrogen ICI 182780 prior to surgery caused a significant reduction in the Ki67 labeling index Received 1/4/00; revised 3/13/00; accepted 3/16/00. of primary breast tumors (6, 7). Recent work has demonstrated The costs of publication of this article were defrayed in part by the that responders but not nonresponders to tamoxifen therapy payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom requests for reprints should be addressed, at Department of Academic Biochemistry, Royal Marsden Hospital, London SW3 6JJ, 2 The abbreviations used are: SERM, selective estrogen receptor mod- United Kingdom. Phone: 44 171 808 2887; Fax: 44 171 376 3918; ulator; ER and PR, estrogen and progesterone receptor; ITT, intention to E-mail: [email protected]. treat.

show a significant decrease in Ki67 after 14 days (8). Thus, Study Procedures. Patients were screened within 2 weeks changes in proliferation in response to short-term challenge with prior to receiving study medication to determine eligibility for entry an antiestrogen may be predictive of clinical effectiveness. into the study. Eligible patients were randomized to receive a Significantly higher levels of apoptosis have also been blinded treatment regimen of either idoxifene or placebo and were found in human tumors during their treatment with tamoxifen or then evaluated clinically prior to surgery and 1 week after surgery/ ICI 182780 (9). This confirms the observations of estrogen withdrawal. The following procedures were carried out: complete deprivation-induced apoptosis found in xenograft studies (10, medical history and physical examination, hematology and blood 11), and these findings recently have been extended by demon- chemistry evaluations; 12-lead electrocardiogram, chest X-ray, strating that idoxifene itself induces apoptosis in xenograft computed tomography or magnetic resonance imaging scan, or systems (12). Thus, induction of apoptosis may also be a marker ultrasound if clinically indicated. Documentation of lesion size, and of pharmacological effectiveness of endocrine therapy. It is well histopathological type and grade were performed on the tumor established that breast tumor ER and PR expression is predictive excision biopsy obtained at surgery. Clinical assessments and lab- of clinical response to endocrine therapy. Response to anti- oratory tests were performed to evaluate the tolerability of study estrogen therapy in human breast cancer has been observed in treatment. In the safety evaluation, clinical interpretation was based ϳ50% of patients with ER-positive tumors and 75% that are on review of displays of adverse experiences. Principal consider- positive for both ER and PR (13). ations in this evaluation were time to onset, severity, study medication, and investigator-reported relationship of an adverse experi- ence with the study medication. MATERIALS AND METHODS Primary and Secondary Efficacy Variables. The pri- Study Design. This was a multicenter, randomized, dou- mary study end point was the change in Ki67 tumor prolifera- ble-blind study in postmenopausal women with primary breast tion score from pretreatment to posttreatment biopsy. Secondary cancer. Written informed consent to participate in the study was efficacy variables were the changes in apoptotic index, and ER obtained from all patients before enrollment. Patients received and PR expression from pretreatment to posttreatment. medication with 40 mg/day idoxifene p.o. versus placebo for Study Medication. Idoxifene was supplied as white, cir- ϳ14 days prior to surgery, to a maximum of 21 days. Tumor cular, biconvex tablets containing 40 mg of idoxifene. Placebo biopsy samples were taken before the start of treatment and at tablets were of identical appearance. Study medication was surgery for the measurement of Ki67 labeling index, apoptotic provided in plastic bottles containing 21 tablets of either idox- index, and ER and PR expression. Clinical assessments and ifene or placebo. The bottles were labeled with information including protocol number, job number, patient number, direc- laboratory tests were performed to evaluate the tolerability of tions for use, the quantity of tablets supplied, warning state- study treatment. The protocol was approved by each local Eth- ments, and storage conditions. Patients took one tablet (either 40 ical Committee. mg of idoxifene or placebo) p.o. once a day for 14–21 days. No Inclusion and Exclusion Criteria. Postmenopausal fe- instructions were given with respect to the timing of study male patients were enrolled. Postmenopausal status was defined medication in relation to meals. Treatment duration was depend- as Ͼ12 months since last menses; in patients whose last menses ent on the date of surgery. Patients continued on the study drug was Յ12 months before the start of treatment and in patients for a minimum of 14 days to a maximum of 21 days. Surgery who were amenorrheic as a result of surgery, follicle-stimulating was to take place within 21 days of commencing study medi- hormone and luteinizing hormone levels were to lie in the cation. At each study site, patient numbers were randomly Ն postmenopausal range (follicle-stimulating hormone 35 IU/L; linked to packaged medication supplies. The medication code Ն luteinizing hormone 40 IU/L). Patients had to have a histo- was to be broken only in the event of a serious adverse expe- logical or cytological diagnosis of primary breast cancer and an rience that the investigator could not adequately treat without Ն Ͻ operable lesion 2 cm in size. Lesion sizes of 2 cm could, knowledge of the identity of study medication. The numbers of however, be considered if, in the investigator’s opinion, a good bottles and tablets dispensed for each patient were maintained quality core-needle biopsy could be obtained using a 14-gauge and reconciled with the study medication record. needle. Grading was performed by central review by a single Tumor Samples. Before the start of study medication, a pathologist (Dr. N. Nasiri, Royal Marsden Hospital, London, core-cut biopsy of the primary breast tumor was obtained using a United Kingdom) on the excision biopsy specimen. Patients 14-gauge needle, and at surgery a representative sample of the were enrolled only if they had a performance status Յ2 (Eastern excision tumor was obtained from the operative specimen. Both Cooperative Oncology Group-WHO scale), a life expectancy of specimens were fixed in 10% neutral buffered formalin and em- Ն4 months, and no known renal or hepatic impairment. Patients bedded in paraffin wax. The embedded blocks were sent to the with metastatic breast cancer (stage IV) or who had received Royal Marsden Hospital, London, United Kingdom for processing prior therapy for breast cancer or with previous malignancies and analysis. Sections (3 ␮m thick) of the paraffin wax blocks were within the last 5 years at other sites were excluded. Other cut onto positively charged slides and dried overnight at 37°C exclusion criteria were concurrent medical or psychiatric prob- before being used to determine Ki67, apoptosis, ER, and PR by lems unrelated to breast cancer; treatment with another investi- previously published immunohistochemical methods. gational drug; concomitant treatment with hormone replacement Analytical Methods. Measurement of cell proliferation therapy, coumarin-type anticoagulant therapy, or oral or i.v. was by immunocytochemical assay using the MIB1 mouse corticosteroids; and a known hypersensitivity to tamoxifen or monoclonal antibody to Ki67 (14). Ten high-powered fields tamoxifen analogues. were scored. Measurement of apoptosis was by TdT-mediated

dUTP-biotin nick end labeling; the apoptotic index was Table 1 Baseline characteristics: breast cancer data and expressed as a percentage: (no. of cells displaying apoptotic histopathological grade bodies ϫ 100) Ϭ (total no. of cells) (14). Three thousand tumor Treatment group cells/section were counted for apoptotic index. Demonstration Idoxifene (n ϭ 30) Placebo (n ϭ 27) of ER expression was by the Novocastra 6F11 mouse monoclonal antibody (15) and of PR was by the Novocastra 1A6 Baseline characteristics n % n % antibody (11). ER and PR expression was assessed semiquan- Tumor characteristics titatively by determining the percentage of tumor cells stained Histopathological grade G 62000 by the primary antibody and assessing the intensity of staining, 1 G 10 33 20 74 using a score of 0 to 3, which corresponded to negative, weak, 2 G3 12 40 7 26 intermediate, and strong staining intensities in 10 high-powered Could not be assessed 1 3 0 0 fields. The percentage of tumor cells in each of these categories Missing 1 3 0 0 was used to calculate the overall H-score, ranging from 0 to 300. Nodal involvement Ͼ 015501141 Tumors with a score 20 were considered positive for either ER 1–3 9 30 8 30 or PR. In all cases, pairs of samples from the same patient were 4ϩ 517726 stained and scored in the same assay batch. Scoring was con- Not known 1 3 1 4 ducted by one analyst (M. H.) and was subject to quality control Tumor size (cm) checks by a second (J. S.). Mean 2.51 2.44 SE 0.26 0.21 Statistical Analysis. Study participants were randomly Age (years) assigned to receive either 40 mg of idoxifene or placebo using 46–55 4 3 the technique of randomly permuted blocks. 56–65 6 12 Ͼ A total of 50 patients (25 per arm) with positive ER status 65 20 12 ϳ Mean 69.0 66.6 were to be recruited for this study. It was expected that 70 SE 1.7 1.9 patients would need to be enrolled to achieve this number because in most cases ER status would be unknown at the time of study entry. The sample size was based on those of previous studies reporting decreased Ki67 levels after presurgical treat- pliance with study medication), one patient had an adverse event ment with antiestrogens (6, 7). Patients were considered to have causing therapy to be stopped; in the placebo group, three completed the study as planned and to be evaluable if they met patients had protocol violations/deviations (one with no tumor all eligibility criteria, were found to be ER positive, and had found in biopsy, two for noncompliance with study medication). both pre- and posttreatment biopsies taken that contained suffi- Of the remaining 36 patients in the idoxifene group and 34 in the cient tumor cells for analysis (efficacy evaluable population). placebo group, 30 and 27, respectively, had ER-positive tumors The primary efficacy evaluation was based on the Ki67 as assessed in the pretreatment core-cut biopsy. These 57 pa- tumor proliferation score. The arithmetic difference between tients therefore constituted the evaluable group. scores obtained from the pretreatment needle biopsy and excised Baseline characteristics for breast cancer data and for his- tumor biopsy were determined for each patient (described as topathological grade of the tumors from these 57 evaluable absolute change). These differences were compared by a Mann- patients are shown in Table 1. Twenty-one and 18 of the Whitney nonparametric analysis. Similar analyses were per- patients, respectively, were positive for PR. formed for the apoptotic index, ER, and PR. The histopathological grade assigned most frequently was Evaluation of safety data included all randomized patients grade 3 (poorly differentiated) in the idoxifene group and grade (ITT population) and were summarized as descriptive statistics. 2 (moderately differentiated) in the placebo group. Mean tumor size was very similar between the groups. The majority of patients in both treatment groups had invasive ductal tumors. RESULTS Similar percentages of patients in the idoxifene group and in the Demographic and Baseline Characteristics. Seventy- placebo group had invasive lobular tumors, and one patient (in seven patients entered the study at 12 centers and were included the idoxifene group) had Paget’s disease with accompanying in the ITT population. All patients enrolled into the study were invasive disease. Caucasian. Demographic characteristics were similar in the ITT The numbers and percentages of patients who had concom- population for both treatment groups with respect to age (mean, itant conditions other than their breast cancer on entry to the 69.2 and 68.3 years in the idoxifene and placebo groups, re- study and who received prior and concomitant medications were spectively) and weight (mean, 72.7 and 71.5 kg, respectively). similar in both treatment groups. The mean duration of study All patients were postmenopausal with mean times since the last medication was 16.5 days for patients randomized to idoxifene menstrual period of 21.5 and 19.0 years, respectively. In the ITT and 18.0 days for patients randomized to placebo. population, 40 patients in the idoxifene group and 37 in the Biomarker Data. Except where stated, all data below placebo group were screened and randomized into the study. relate only to the efficacy evaluable population, i.e., ER positive Four patients in the idoxifene group and three in the placebo and eligible for all criteria. The changes in biomarkers are group were withdrawn or excluded from analysis: in the idox- summarized in Table 2. ifene group, three patients had protocol violations/deviations Ki67 Labeling Index. Individual changes in Ki67 be- (two patients with no tumor found in biopsy, one for noncom- tween the two biopsies for both placebo and idoxifene are

Table 2 Summary of mean (ϮSE) biomarker values before and after treatment Ki67 (%) Apoptotic index (%) PR (H-score) Alla PRϩ only Alla PRϩ onlyER (H-score) Alla PRϩ only Placebo Pre 18.8 Ϯ 2.3 17.4 Ϯ 2.5 0.80 Ϯ 0.08 0.79 Ϯ 0.11 223 Ϯ 12 96 Ϯ 15 141 Ϯ 13 Post 19.8 Ϯ 2.2 18.3 Ϯ 2.7 0.85 Ϯ 0.07 0.76 Ϯ 0.08 192 Ϯ 13 93 Ϯ 14 123 Ϯ 16 Change ϩ1.0 Ϯ 1.4 ϩ0.9 Ϯ 1.9 ϩ0.04 Ϯ 0.05 Ϫ0.04 Ϯ 0.05 Ϫ30 Ϯ 8 Ϫ2 Ϯ 8 Ϫ17 Ϯ 6 n 27 18 27 18 27 27 18 Idoxifene Pre 19.7 Ϯ 2.7 18.6 Ϯ 3.5 0.88 Ϯ 0.09 0.77 Ϯ 0.08 198 Ϯ 10 95 Ϯ 14 134 Ϯ 11 Post 13.4 Ϯ 3.4 11.5 Ϯ 4.2 0.90 Ϯ 0.09 0.76 Ϯ 0.06 140 Ϯ 10 116 Ϯ 15 161 Ϯ 12 Change Ϫ6.3 Ϯ 1.9 Ϫ7.1 Ϯ 1.6 ϩ0.02 Ϯ 0.05 0.01 Ϯ 0.05 Ϫ58 Ϯ 10 20 Ϯ 826Ϯ 10 n 30 21 30 21 30 30 21 Pb 0.0043 0.0034 0.72 0.99 0.027 0.078 0.0002 a All subjects within each group, regardless of PR status. b P for the comparison of the change for each analyte between the placebo and idoxifene.

Fig. 1 Individual changes for percentage of Ki67 for the placebo (left) and idoxifene (right) ERϩ groups. One patient in the idoxifene group had pretreatment and on-treatment values of 82.3 and 91.7%, respectively, and is not plotted (shown as dashed line). Pre, pretreatment; D14/ 21, 14–21 days of treatment.

shown in Fig. 1, and the mean changes are shown in Table 2. Apoptotic Index. Individual changes in apoptotic index The change in Ki67 labeling index was significantly greater for between the two biopsies are shown for both placebo and idoxifene than placebo (P ϭ 0.0043). The change was from a idoxifene in Fig. 3, and the mean changes are shown in Table 2. mean pretreatment level of 19.7 Ϯ 2.7% (mean Ϯ SE) to 13.4 Ϯ The mean difference between pretreatment and on-treatment 3.4% giving a mean change of Ϫ6.3 Ϯ 1.9%. In the idoxifene- samples was very small for both the idoxifene group (ϩ0.02 Ϯ treated group, one patient (patient A) had much higher pretreat- 0.05%) and the placebo group (ϩ0.04 Ϯ 0.05%). The difference ment and on-treatment Ki67 than all others (82.3 and 91.7%, between the two treatment groups was not statistically signifi- respectively). There was no evidence of technical failure being cant index (P ϭ 0.7200), showing that idoxifene had no signif- responsible for these extreme results. Exclusion of this patient icant effect on the apoptotic index. The mean pretreatment level from the data analysis resulted in a mean pretreatment Ki67 in the ER-negative group was higher than in the ER-positive labeling index of 17.6 Ϯ 1.7% and posttreatment index of group (1.6% and 0.9%, respectively). The ER-negative patients 10.7 Ϯ 2.2%, with a mean change of Ϫ6.8 Ϯ 1.9%. There was showed no effect of either therapy on the apoptotic index. There no significant effect in the placebo-treated ER-positive tumors: was no significant increase in apoptosis after idoxifene treat- the mean change was ϩ1.0 Ϯ 1.4%. ment in tumors positive for both ER and PR (P ϭ 0.99). The numbers of ER-negative patients was small (n ϭ 13). ER and PR Expression. Individual changes in ER and Pretreatment Ki67 levels were higher in the ER-negative group PR expression between the two biopsies are shown for both (mean, 36.9%), and there was no significant change in levels placebo and idoxifene in Figs. 4 and 5, and the mean changes with either idoxifene (n ϭ 6) or placebo treatment (n ϭ 7). are shown in Table 2. Patients positive for both ER and PR in the idoxifene- At the end of therapy, the ER expression in the idoxifene- treated group showed a mean fall from 18.6 Ϯ 3.5% to 11.5 Ϯ treated patients decreased. A slight decrease was also observed 4.2%, giving a mean change of Ϫ7.1 Ϯ 1.6% (P ϭ 0.0034; Fig. in the placebo-treated group, but the change was significantly 2). Excluding patient A from this analysis led to a change from greater for the idoxifene-treated group (P ϭ 0.027). 15.5 Ϯ 1.5% to 7.5 Ϯ 1.4%, a mean change of Ϫ7.9 Ϯ 1.4%. Considering all patients, PR expression showed a minor Thus, the decrease as a proportion of the baseline was greater in mean increase in the immunohistochemical score in the idox- the ER-positive/PR-positive group than in the ER-positive ifene-treated group with very little change in the placebo group. group as a whole. The difference in the change between the two groups did not

Fig. 2 Individual changes for percentage of Ki67 for the placebo (left) and idoxifene (right) ERϩ/PRϩ groups. Dashed line indicates patient in the idoxifene group who had pretreatment and on-treatment values of 82.3 and 91.7%, respectively, which were not plotted. Pre, pretreatment; D14/21, 14–21 days of treatment.

Fig. 3 Individual changes in the apoptotic index (%) for the placebo (left) and idoxifene (right)ERϩ groups. Pre, pretreatment; D14/21, 14–21 days of treatment.

reach statistical significance (P ϭ 0.078). However, in those baseline levels. Comparison of these changes with those re- patients who were PR positive prior to treatment, there was an ported in the earlier studies with other antiestrogens is not valid increase in the H-score of 26 Ϯ 10 in the idoxifene group and a because, as well as being in different populations, those studies decrease in the placebo group of 2 Ϯ 8, showing a significant used a different antibody to Ki67 in frozen rather than formalin- interaction with treatment (P ϭ 0.0002). fixed biopsies. Nonetheless, the data indicate that idoxifene, like Safety. Idoxifene was well tolerated. Seven patients in these other two drugs, should be an effective anti-breast cancer the idoxifene group and six in the placebo group had adverse drug in ER-positive patients. experiences considered by the investigator as suspected as being The lower Ki67 in the ER-positive group and the greater or probably related to study treatment. The most frequently proportional change in Ki67 in the group that was also positive reported adverse event was hot flushes of mild intensity in three for PR (to ϳ50% of baseline levels) are consistent with expec- patients in the idoxifene group and dyspepsia in two patients in tations: it is known that proliferation is generally higher in the placebo group. There were very few laboratory abnormali- ER-negative tumors (6, 16) and that tumors positive for PR as ties in patients randomized to study medication, and all were well as ER have a greater likelihood of responding to endocrine considered as clinically nonsignificant. therapy (13). It is possible that the results underestimate the effects DISCUSSION that idoxifene would have at steady state. The terminal half- ϳ 3 Significant reductions in Ki67 labeling index in primary life of the drug is 5 weeks such that after a median 16.5 breast carcinomas have been reported previously after short- days of treatment only approximately half of one half-life term, presurgical treatment with tamoxifen or the pure anti- would have elapsed. The plasma levels at the time of excision estrogen ICI 182780 (6, 7). No previous studies have been would approximate to only one-fourth those of steady-state conducted with idoxifene, but it has been shown to lead to levels, and the drug exposure over the preceding 16.5 days reduced Ki67 levels in MCF7 human breast cancer xenografts would be represented by an exponential function toward (12). In this study, idoxifene led to a pronounced change in mean Ki67 in ER-positive but not ER-negative patients. Overall in the idoxifene-treated ER-positive patients, the fall was to ϳ65% of 3 Data on file at SmithKline Beecham.

Fig. 4 Individual changes in ER H-score for the placebo (left) and idoxifene (right) ERϩ groups. Pre, pretreatment; D14/21, 14–21 days of treatment.

Fig. 5 Individual changes in PR H- score for the placebo (left) and idoxifene (right)ERϩ groups. Pre, pretreatment; D14/ 21, 14–21 days of treatment.

these levels. Thus, the impact on proliferation is likely to be encing the accuracy of the data, particularly for a rare event such equivalent to that which an idoxifene dose of 10 mg/day or as apoptosis. However, the studies referenced above in which less would achieve at steady state [the pharmacokinetics of increases in apoptosis were seen had no greater statistical power idoxifene being linear (2)]. To have achieved steady-state than this study and used relatively similar clinical samples. In levels by 16.5 days in this study would have required a addition, our previously published data on the reproducibility of loading dose of idoxifene substantially above that adminis- measurements using core-cuts showed only a marginally poorer tered previously to humans. precision for apoptosis than for Ki67 (14). The data may reflect A number of studies have shown that apoptosis can be the true effect of idoxifene on breast tumors in vivo, i.e., that induced by estrogen withdrawal or by an antiestrogen, both in despite effects seen in model systems there is no effect on vitro and in animal model systems (10, 11, 17, 18). In an earlier apoptosis in patients. small study by our group, apoptosis was also found to be As noted above, the measurement of ER and PR content significantly higher in patients treated with ICI 182780 com- is of importance when evaluating the likely response to pared with controls; in patients treated with tamoxifen apoptosis endocrine therapy. The greater decrease in ER levels in the showed a small change with borderline statistical significance in patients with ER-positive tumors (9). Recently, we also have idoxifene-treated patients than in the placebo group is con- shown that idoxifene treatment of estrogen-supported MCF7 sistent with some, but not all, earlier data with tamoxifen (6, xenografts leads to increased apoptosis (12). Thus, the absence 19), but it is not as great an effect as that seen with ICI in this study of a significant increase in apoptotic index after 182780 (7), which decreases stability of the ER protein (20). treatment with idoxifene was surprising. There may be a number In vitro estrogen deprivation leads to increases in ER in of reasons why this change was not observed. It is possible that breast cancer cells because estrogen enhances ubiquitin- the exposure of the tumor to relatively low doses of idoxifene proteasomal degradation of ER (21). This is consistent with over the treatment period may be insufficient for any impact on our observations of increased ER levels in MCF7 human apoptosis. The use of pretreatment biopsies containing relatively breast cancer xenografts on estrogen withdrawal (11). The small amounts of tissue may lead to tumor heterogeneity, influ- mechanism for the decreased ER levels with idoxifene and

tamoxifen is not known but may also relate to ligand-depen- REFERENCES dent proteasomal degradation. 1. Johnston, S. R. D., Gumbrell, L., Evan, T. R. J., Coleman, R. E., The decrease in ER shown in the placebo group empha- Smith, I. E., Rea, D., Soukop, M., Howell, A., Jones, A., Dowsett, M., sizes the importance of placebo controls in this sort of study. and Coombes, R. C. A phase II randomized double-blind study of idoxifene (40mg/d) vs tamoxifen (40mg/d) in patients with locally This decrease may have occurred as a result of histological advanced metastatic breast cancer resistant to tamoxifen (20mg/d). Proc. fixation differences between core-cuts and excision biopsies. Am. Soc. Clin. Oncol., 18: 109a, 1999. Other possible explanations are that the taking of a core-cut may 2. Coombes, R. C., Haynes, B. P., Dowsett, M., Quigley, M., English, itself induce a change in ER in the tumor or that systematic J., Judson, I. R., Griggs, L. J., Potter, G. A., McCague, R., and Jarman, differences in scoring between core-cuts and excisions might M. Idoxifene: report of a phase I study in patients with metastatic breast cancer. Cancer Res., 55: 1070–1074, 1995. occur despite efforts to avoid this. A further consideration is that 3. Chander, S. K., McCague, R., Luqmani, Y., and Coombes, R. C. the change in the ER H-score may result from a conformational Pyrrolidino-4-iodotamoxifen and 4-tamoxifen, new analogues of the change in ER induced by the binding of idoxifene, which could antiestrogen tamoxifen for the treatment of breast cancer. Cancer Res., influence the binding to antibody. 51: 5851–5858, 1991. PR is an estrogen-induced gene, and its expression is 4. Delmas, P. D., Mulder, H., Eastell, R., Greenwald, M., and Mac- Donald, B. Idoxifene reduces bone turnover in osteopenic postmeno- considered as denoting an intact estrogen response mecha- pausal women. Bone, 23: S494, 1998. nism (21). By extension, decreases in PR expression may be 5. Mulder, H., Wasnich, R., and Eastell, R. The effects of idoxifene on considered indicative of an antiestrogenic effect as seen with markers of cardiovascular risk in osteopenic postmenopausal women. ICI 182780 (7). In contrast, the early increases seen in PR Eur. Heart J., 19: 237, 1998. after initiating tamoxifen treatment are considered indicative 6. Clarke, R. B., Laidlaw, I. J., Jones, L. J., Howell, A., and Anderson, of an early predominance of an agonist effect of tamoxifen [at E. Effect of tamoxifen on Ki67 labelling index in human breast tumours and its relationship to estrogen and progesterone receptor status. Br. J. least on the PR gene (22–24)], although this does not indicate Cancer, 67: 606–611, 1993. lack of efficacy: those patients showing an increase in PR 7. De Friend, D. J., Howell, A., Nicholson, R. I., Anderson, E., Dow- (after a median 13 days of tamoxifen) have a greater likeli- sett, M., Mansel, R. E., Blamey, R. W., Bundred, N. J., Robertson, J. F., hood of response (23). The trend toward an increase in PR Saunders, C., Baum, M., Walton, P., Sutcliffe, F., and Wakeling, A. E. with idoxifene but lack of a statistically significant effect Investigation of a new pure anti-estrogen (ICI 182780) in women with primary breast cancer. Cancer Res., 54: 408–414, 1994. may indicate that idoxifene has some agonist activity on this 8. Makris, A, Powles, T. J., Allred, D. C., Ashley, S. E., Ormerod, parameter in the breast, but less than tamoxifen, which is M. G., Titley, J. C., and Dowsett, M. Changes in hormone receptors and consistent with its lower uterotropic effect than tamoxifen in proliferation markers in tamoxifen treated breast cancer patients and the rodents (1). relationship with response. Br. Cancer Res. Treat., 48: 11–20, 1998. Idoxifene was well tolerated, and no major side-effects or 9. Ellis, P. A., Saccani-Jotti, G., Clarke, R., Johnston, S. R. D., Ander- changes in laboratory values were observed, consistent with son, E., Howell, A., A’Hern, R., Salter, J., Detre, S., Nicholson, R., Robertson, J., Baum, M., Smith, I. E., and Dowsett, M. Induction of other reports on its clinical usage. apoptosis by tamoxifen and ICI 182780 in primary breast cancer. Int. J. This is only the second report of a randomized biomarker Cancer, 72: 608–613, 1997. study [the first was the study of ICI 182780 (7)] during a 10. Kyprianou, N., English, H. F., Davidson, N. E., and Isaacs, J. T. prescribed period of time between diagnosis and the time at Programmed cell death during regression of the MCF human breast cancer following oestrogen ablation. Cancer Res., 51: 162–166, 1991. which a patient is scheduled for surgery. Medical treatment is 11. Detre, S., Salter, J., Barnes, D. M., Riddler, S., Hills, M., Johnston, not conventionally offered during this time, and it offers the S., Gillett, C., A’Hern, R., and Dowsett, M. Time-related effects of opportunity to assess and compare new drugs once their safety oestrogen withdrawal on proliferation and cell death-related events in has been established. Importantly, changes in Ki67 over this MCF7 xenografts. Int. J. Cancer, 81: 309–313, 1999. period by tamoxifen relate to subsequent response to therapy 12. Johnston, S. R. D, Boeddinghaus, I. M., Riddler, S., Haynes, (8). The consistency of the data derived here with those that B. P., Hardcastle, I. R., Jarman, M., and Dowsett, M. Idoxifene antagonizes estradiol-dependent MCF-7 breast cancer xenograft would be expected of an effective SERM not only supports growth through sustained induction of apoptosis. Cancer Res., 59: idoxifene being an effective anti-breast cancer agent, it also 3646–3651, 1999. supports the further exploitation of this clinical scenario with 13. Henderson, I. C. Endocrine therapy of metastatic breast cancer. In: novel therapies. J. R. Harris, S. Hellman, I. C. Henderson, and D. W. Kinne (eds.), Breast In conclusion, idoxifene showed a significant antiprolifera- Diseases, 2nd Ed., pp. 559–603. Philadelphia: J. B. Lippincott, 1991. tive effect in ER-positive breast cancer as shown by the decrease 14. Ellis, P. A., Smith, I. E., Detre, S., Burton, S. A., Salter, J., A’Hern, R., Walsh, G., Johnston, S. R. D, and Dowsett, M. Reduced apoptosis and in the Ki67 labeling index with no effect in ER-negative can- proliferation and increased bcl-2 in residual breast cancer following preop- cers. This is consistent with idoxifene being an effective anti- erative chemotherapy. Breast Cancer Res. Treat., 48: 107–116, 1998. breast cancer agent in ER-positive tumors. 15. Bevitt, D. J., Milton, I. D., Pigott, N., Henry, L., Carter, M. J., Tonas, G. L., Leonard, T. W., Westley, B., Angus, B., and Home, C. H. New monoclonal antibodies to oestrogen and progesterone receptors ACKNOWLEDGMENTS effective for paraffin section immunohistochemistry. J. Pathol., 183: 228–232, 1997. We thank the following investigators who also entered their pa- 16. Wenger, C. R., Beardslee, S., Owens, M. A., Pounds, G., Oldaker, tients into this trial: Dr. Jacques De Greve, Dr. Gilbert De Wasch, Dr. T., Vendeley, P., Pandian, M. R., Harrington, D., Clark, G. M., and A. Labrie, Dr. R. Vree, Dr. H. Sleeboom, Fiona MacNeill, R. J. Sains- McGuire, W. L. DNA ploidy, S-phase, and steroid receptors in more bury, Dr. J. Bonneterre, Dr. L. Beex, and Dr. T. Pienkowski. The than 127,000 breast cancer patients. Breast Cancer Res. Treat., 28: centralized pathology review was undertaken by Dr. N. Nasiri. 9–20, 1993.

17. Bardon, S., Vignon, F., Montcourrier, P., and Rocheport, H. Steroid 21. Nawaz, Z., Lonard, D. M., Dennis, A. P., Smith, C. L., and receptor medicated cytotoxicity of an anti-estrogen and an anti-progestin O’Malley, B. W. Proteasome-dependent degradation of the human es- in breast cancer cells. Cancer Res., 47: 1441–1448, 1987. trogen receptor. Proc. Natl. Acad. Sci. USA, 96: 1858–1862, 1999. 18. Warri, A. M., Huovinen, R. L., Laine, A. M., Martikanine, P. M., 22. Horwitz, K. B., Yoseki, Y., and McGuire, W. L. Estrogen control of and Harkonen, P. L. Apoptosis in toremifene-induced growth inhibition progesterone receptor in human breast cancer: role of estradiol and of human breast cancer cells in vivo and in vitro. J. Natl. Cancer Inst., antiestrogen. Endocrinology, 103: 1742–1751, 1978. 85: 1412–1418, 1993. 23. Howell, A., Harland, R. N. L., Barnes, D. M., Baildam, A. D., 19. Dowsett, M., Johnston, S. R. D., Detre, S., Salter, J., Humphries, S., Wilkinson, M. J., Hayward, E., Swindell, R., and Sellwood, R. A. Saccani, G-J., Smith, I. E., MacLennan, K., and Trott, P. Cytological Endocrine therapy for advanced carcinoma of the breast: effect of tumor evaluation of biological variables in breast cancer patients undergoing heterogeneity and site of biopsy on the predictive value of progesterone primary medical treatment. In: M. Motta and M. Serio (eds.), Sex Hor- mones and Antihormones in Endocrine Dependent Pathology: Basic and receptor estimations. Cancer Res., 47: 296–299, 1987. Clinical Aspects, pp. 329–336. Amsterdam: Elsevier Science BV, 1994. 24. Noguchi, S., Miyauchi, K., Nishizawa, Y., and Koyama, H. Induc- 20. Metzger, D., Berry, M., Ali, S., and Chambon, P. Effect of antag- tion of progesterone receptor with tamoxifen in human breast cancer onists on DNA binding properties of the human estrogen receptor in with special reference to its behavior over time. Cancer (Phila.), 61: vitro and in vivo. Mol. Endocrinol., 9: 579–591, 1995. 1345–1349, 1988.

Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2000 American Association for Cancer Research. Antiproliferative Effects of Idoxifene in a Placebo-controlled Trial in Primary Human Breast Cancer

Mitch Dowsett, J. Michael Dixon, Kieran Horgan, et al.

Clin Cancer Res 2000;6:2260-2267.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/6/6/2260

Cited articles This article cites 22 articles, 8 of which you can access for free at: http://clincancerres.aacrjournals.org/content/6/6/2260.full#ref-list-1

Citing articles This article has been cited by 10 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/6/6/2260.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/6/6/2260. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.