Steroid Receptors and Selective Estrogen Receptor Modulation in Mammary and Gynecologic Malignancy

Malignancy of the breast, ovary, and endometrium is linked to endogenous hormones. Observations over the past 100 years have established this causal relationship. In 1900, Beatson and Boyd independently established that the removal of the ovaries from premenopausal women with metastatic breast cancer would, roughly one third of the time, cause regression of disease with improved prognosis.^{1, 2} Therefore, an ovarian factor was implied in the promotion of carcinogenesis.

Allen and Doisey identified the “estrus stimulating principle”, endogenous estrogens that could activate tumors of the breast, and subsequent research in synthetic chemistry identified nonsteroidal estrogens.³ Dodds and associates described an extremely potent estrogen in animals, diethylstilbestrol (DES),⁴ but Haddow and colleagues⁵ first tested high-dose DES as a treatment for advanced breast cancer in postmenopausal women. Use of high-dose estrogen treatment is somewhat paradoxical, but effective in one third of patients. The mechanism of action is unknown. Alternatively, reducing endogenous steroids by adrenalectomy improves prognosis of some postmenopausal women in similar proportions to oophorectomy.⁶ It is known that adrenal androstenedione and testosterone are converted to estrone and estradiol by peripheral aromatase enzyme systems.

Numerous studies have linked breast cancer risk to prolonged, unopposed exposure to estrogen as in early menarche,⁷ late menopause,^{8, 9} and increased age at first pregnancy.¹⁰ Other hormonal risk factors are less well established. Reports on abortion and the risk of breast cancer have been equivocal.^{11, 12, 13, 14} Studies of the effect of lactation on breast cancer risk have suggested that a longer duration of lactation reduces risk in premenopausal women.¹⁵ Postmenopausal obesity has been shown to increase risk.^{16, 17}

Similarly, obesity¹⁸ is a well-defined risk factor for endometrial cancer stemming from increased availability of peripheral estrogens especially in postmenopausal women with a compounding lack of progesterone.^{19, 20} Estrogen is a proliferative hormone, whereas progesterone is a differentiating hormone in the uterus. Perhaps most important, periodic withdrawal bleeding sheds any malignant epithelial cells. Analogous factors related to high or unopposed estrogen levels also elevate the risk of endometrial cancer; these include early menarche, late menopause,^{21, 22} nulliparity,²³ and unopposed estrogen replacement therapy.^{23, 24}

The effects of exogenous hormones in the form of oral contraceptives and hormone replacement therapy (HRT) on breast cancer have been studied extensively. Some studies suggest that the long-term use of oral contraceptives in young women before their first pregnancy may increase breast cancer risk.^{25, 26} Two meta-analyses of this effect demonstrate small but statistically significant increases in risk for HRT users.^{27, 28} Steinberg and coworkers²⁷ noted that after five years of estrogen use there is a proportional yearly increase in risk, whereas Sillero-Arenas and associates²⁸ did not observe a significant association between duration of HRT and breast cancer risk. The Iowa Women's Health Study found a longer duration of HRT was associated with increased risk of invasive carcinoma with a favorable histology, with a relative risk (RR) of 1.81 for HRT use of less than five years duration versus a RR of 2.65 for use longer than five years.²⁹ Even so, women who develop breast cancer while using HRT at the time of diagnosis have a similar prognosis in terms of type, size, or grade of tumor as that of nonusers.³⁰ In breast epithelia, progesterone, like estradiol, also has a strong proliferative effect. Progesterone accelerates the appearance, growth rate, and number of tumors in carcinogen-induced rat mammary cancers.³¹ Furthermore, use of progestins in combination with estradiol for HRT has been linked with a higher risk of developing breast cancer than use of estradiol alone.³²

In 1962, Jensen and Jacobson³³ demonstrated that [³H] estradiol bound to and was retained by estrogen target tissues including the uterus, vagina, and pituitary gland in the female rat. In contrast, tissues that did not respond to estrogen did not retain [³H] estradiol. Jensen³³ proposed that an estrogen receptor (ER) in estrogen target tissues must capture circulating steroids and initiate the cascade of biochemical events associated with estrogen action in that particular tissue. Toft and respective coworkers in two studies^{34, 35} first identified the ER as an extractable protein from the rat uterus. Subsequently, Gorski and associates³⁶ and Jensen and coworkers³⁷ independently proposed subcellular models of estrogen action in target tissues. Jensen and associates³⁸ then proposed the clinical ER assay to predict hormone responsive breast cancer. This established a mechanistic link between estrogen action and the growth of breast cancer.

Monoclonal antibodies to the ER subsequently established that the ER is a nuclear protein,³⁹ and the technology of immunocytochemistry is now standard for the determination of receptor status in breast biopsies.^{40, 41} The ER status of the patient is highly predictive of a treatment response to endocrine ablative surgery⁴² or antiestrogen therapy.⁴³

Similar experimental techniques have been used with human endometrial tissue. The uptake of radiolabeled estrogen is highest during the first two weeks of the menstrual cycle,⁴⁴ that is, during the time of unopposed estrogen action. In addition, quantitation of the human ER protein revealed that concentrations are markedly higher during the proliferative phase than in the secretory phase.^{45, 46, 47, 48} In contrast, uptake of radiolabeled progesterone is higher in the secretory phase than in the proliferative phase of the menstrual cycle.⁴⁹ The progesterone receptor (PgR) level is low during the proliferative phase, peaks at midcycle, and becomes lower again during the secretory phase.^{46, 48}

Endometrial cancers contain ERs and an inverse relationship exists between ER levels and tumor grade.⁵⁰ However, cytosol receptor assays are inherently inaccurate in endometrial carcinoma because of the contribution from normal stromal and myometrial cells, which also contain receptors.^{51, 52, 53} Development of monoclonal antibodies to the PgRs allowed the quantitation of hormone receptor levels in endometrial cancer.⁵⁴

Molecular cloning of the cDNAs for the estrogen⁵⁵ and progesterone receptors^{56, 57} led to the realization that these proteins are members of the steroid and thyroid hormone receptor superfamily of ligand-responsive transcription factors.⁵⁸ Both receptors are encoded by eight exons, which correspond to domains of functional significance.^{59, 60, 61}

Biologic effects of estrogen are mediated by two receptors, ERα and ERβ.^{62, 63} These two ERs share a conserved structure with six functional domains, A to F. ERβ is homologous to ERα at the ligand-binding domain (58%) and DNA-binding domain (95%). The remaining domains are not well conserved.⁶⁴ Discovery of ERβ has advanced our understanding of estrogen signaling and may explain tissue responses to estrogen in which ERα is undetectable.⁶⁵ Furthermore, the existence of ERα and β subtypes provides a possible explanation for the tissue selectivity of selective ER modulators (SERMs).

The responsiveness of neoplastic tissues to steroid hormones and antiestrogen therapy correlates with hormone receptor concentration in target cells. Neoplastic cells that retain a full or partial amount of steroid receptors are expected to respond to endocrine therapy whereas cells lacking receptors do not. Response to endocrine therapy in breast cancer correlates with levels of ER and PgR in the tumor,⁴³ which makes receptor status useful in the management of breast cancer. In general, women who are ER- and PgR-positive have a longer disease-free interval and longer survival time than women with receptor-negative tumors.^{66, 67}

More than three fourths of endometrial carcinomas contain ER and PgR.^{50, 68, 69, 70, 71, 72, 73} ER levels in hyperplastic and neoplastic tissues are comparable with those found in early proliferative endometrium.⁷⁴ Ehrlich and coworkers^{71, 75} found a correlation between the response of patients with advanced endometrial cancer to progestin therapy and the presence of PgR. Early stage endometrial cancers are more likely to be receptor-positive and have higher receptor levels than more advanced tumors.⁶⁸ Receptor concentrations are greater in well or moderately differentiated tumors than in anaplastic carcinomas. Receptor-poor tumors are generally more aggressive with decreased five-year survival rates.⁶⁸ Positive hormone receptor status is associated with improved disease-free and overall survival rates.^{13, 76, 77} When considered separately, positive PgR status in endometrial cancer appears to represent a more reliable predictor of prognosis than positive ER status.^{71, 72, 78, 79} These studies suggest that there is a tendency for PgR-positive status to correlate with early stage tumors, that is, type I endometrial carcinoma. Receptor status does not correlate with invasion and metastasis found in advanced stages of disease.

Both ERα and ERβ contain activation functions (AFs) that contribute to the ER's transcriptional activity. AF-1 is located in the amino-terminal region within the A/B region and is believed to be constitutively active and ligand-independent in ERα^{80, 81} (Fig. 1). AF-2 is present at the carboxy-terminus, and is believed to be ligand dependent. Mutational analyses demonstrate the importance of this region for ER transcription because AF-2 can interact with a number of transcriptional coactivators in a ligand-dependent manner.^{82, 83} AF-1 and AF-2 can each activate transcription independently but in most cases they synergize with one another as promoters within the specific cell context. The ER activates gene expression through binding to estrogen response elements (EREs) in responsive genes through the action of AF-1 and AF-2.⁸⁴ ERβ also activates transcription of target genes through EREs.⁸⁵ Additionally, estrogen can induce an AP-1 site in a reporter construct through ERα, but it is inactive through ERβ. Interestingly, ER antagonists activate ERβ to induce activity through an AP-1 site.⁸⁶ These two receptors, ERα and ERβ, form functional heterodimers on DNA and stimulate transcription of a target gene. The existence of two rather than one ER portends a more complex mechanism of action than previously thought and remains a primary therapeutic target for new and existing SERMs.

One of the functional areas of the ER, the E region, is the steroid-binding domain that undergoes a conformational change on binding with estrogen, locking the ligand into the hydrophobic pocket of the receptor. This allows the ER to dissociate from heat shock proteins, dimerize, and bind to specific DNA sequences of estrogen responsive genes. Changes in conformation of the ER allow for interaction with coregulatory proteins (coactivators or corepressors) that act as signaling intermediates between the ER and the transcriptional machinery.⁸⁷ The crystal structure of the ligand binding domain of ERα was determined with estradiol⁸⁸ and DES.⁸⁹ A key feature of the agonist–receptor complex is the availability of a portion of the ligand binding site of ERα, helix 12, to lock the steroid in the hydrophobic pocket and allow for recruitment of coactivators to the AF2 site. The repositioning of helix 12 after ligand binding is an important mechanism for full estrogen action at ERα.^{90, 91}

The nuclear steroid receptors must associate with other nuclear proteins to form a transcription complex. Several coactivators, such as ERAP160⁹² and RIP140,⁹³ have been identified based on their ability to interact with the agonist-bound receptor and not to the antagonist-bound receptor at AF-2. However, the SRC-1 protein has been identified and it also interacts with AF-1⁹⁴ and with ERb through phosphorylation of AF-1 by the MAP-kinase signaling cascade.⁹⁵ It has been demonstrated that SRC-1 and another protein, p300/CBP, contain intrinsic acetyltransferase activity and can interact with other histone acetyl transferases (HATs). Acetylation by SRC-1 of histones bound at specific promoters could be a mechanism by which the AFs of ER and associated coactivators activate transcription of specific genes by enhancing formation of a stable preinitiation complex.⁹⁶

Antiestrogens are competitive inhibitors of estrogen action; the shape of the ligand is essential to reducing the estrogenic efficacy. The side chain of the antiestrogen prevents helix 12 from closing so that activation cannot occur.^{88, 97, 98} It is known that tamoxifen silences AF-2 whereas AF-1 remains constituitively activated.^{81, 99} Other antiestrogens, such as raloxifene, silence both AF-1 and AF-2.^{100, 101} This is a possible explanation for the promiscuous estrogen-like actions of tamoxifen compared with raloxifene.

Lerner and coworkers¹⁰² described the pharmacologic properties of a low potency nonsteroidal antiestrogen, ethamoxytriphetol (MER 25). In laboratory animals, MER 25 and the related compound MRL 37 demonstrated antifertility actions^{103, 104, 105}stimulating a search for more potent agents for clinical applications. Clomiphene nafoxidine,^{106, 107, 108} nitromifene¹⁰⁹and tamoxifen^{110, 111} were all the result of that search, but clinical application as postcoital contraceptives was found to be inappropriate. The drugs stimulate ovulation in subfertile women.¹¹² Clomiphene remains available as a profertility agent, inducing ovulation in subfertile women.

Further study revealed that tamoxifen is a SERM.  It has both pro-estrogenic and anti-estrogenic effects on different tissues. In the breast, it serves as a competitive inhibitor of ER receptor activity, and therefore inhibits growth of estrogen receptor-positive breast cancers. In the uterus, it has partial agonist effects and therefore is associated with a two to fourfold increase in the incidence of uterine cancer.^{113, 114, 115, 116, 117}

In the early 1970s, tamoxifen was tested successfully as an antiestrogen for the treatment of advanced breast cancer in postmenopausal women.^{118, 119} Tamoxifen is now the endocrine therapy of choice for the treatment of all stages of ER-positive breast cancer. A five-year course is currently considered to be optimal adjuvant therapy and produces a profound increase in disease-free and overall survival.⁴³ The 1998 Oxford Overview Analysis⁴³ combined 55 randomized trials that began before 1990, comparing adjuvant tamoxifen therapy to no tamoxifen before recurrence. The study population, of 37,000 women with node-positive and node-negative breast cancer, comprised 87% of the known randomized clinical trials with tamoxifen. Of these women, fewer than 8,000 had a very low or zero level of ER, and 18,000 were classified as ER-positive. The ER status of the remaining nearly 12,000 women was unknown. Based on the normal distribution of ER in random populations, the authors estimated that two thirds would be ER-positive.

The recurrence reductions with tamoxifen in ER-positive patients are all significant whereas the therapeutic effect on ER-negative patients is minimal. In five-year trials of tamoxifen treatment, the reductions of recurrence were 43% and 60% for patients with less than and greater than 100 fmol/mg cytosol of ER protein, respectively, translating to reductions in mortality rates of 23% and 36%.

The Overview Analysis also provided unequivocal proof of the laboratory principle that longer tamoxifen adjuvant therapy is more beneficial.^{120, 121, 122} Duration of therapy is extremely important for the ER-positive premenopausal woman with large amounts of circulating estrogen that can rapidly reverse the effect of short-term tamoxifen treatment (Fig. 2). The benefit of one year of tamoxifen in the premenopausal woman is virtually nonexistent, with a 20-fold increase in effectiveness with five years of therapy. In contrast, the effect of tamoxifen duration on women over the age of 60 is less dramatic because one year of tamoxifen is much more effective in postmenopausal women. The principle that longer therapy is more beneficial than shorter therapy is also demonstrated in the control of contralateral breast cancer (Fig. 3). Tamoxifen has no significant effect if one year of treatment is used but extending therapy to five years produces a 47% reduction in breast cancer incidence. This is a powerful demonstration of the effect of tamoxifen as a chemopreventive.

In 1996, Fischer et al published a study which demonstrated that tamoxifen treatment longer than five years did not appear to improve outcomes.¹²³ In that trial, when therapy was continued beyond five years, there was a decrease in DFS and DDFS. Overall survival was unchanged. There was a higher incidence of thrombembolic events and endometrial cancer in patients who took tamoxifen longer than five years. Since that trial, five years of adjuvant tamoxifen has been considered the optimal duration of therapy. New and contradictory data were recently presented at the 2007 San Antonio Breast Cancer Symposium. The ATLAS trial (Adjuvant Tamoxifen, Longer Against Shorter) found a small, but significant lower annual rate of recurrence in breast cancer patients who took tamoxifen for 10 years (2.9% per year, versus 3.4% per year).¹²⁴ There was no difference in overall survival. The optimal duration of antiestrogen therapy remains controversial. 

It is well known that tamoxifen produces a partial agonist action in the rat uterus,¹¹⁰ but until recently, little information was available about the actions of tamoxifen in the normal human uterus. Although tamoxifen has been used to treat endometrial cancer,¹²⁵ the drug causes stromal thickening in the uterus^{126, 127} and would be expected to encourage growth of preexisting endometrial cancer.¹²⁸ Tamoxifen treatment is associated with a fourfold increase in the incidence of endometrial carcinoma in postmenopausal women^{126, 127, 129, 130, 131} (Fig. 4). Most important, the stage and grade of endometrial cancers observed in women taking tamoxifen are the same as those in the general population.¹²⁹ It is known that the uterus harbors five times the amount of occult disease as is reported clinically.¹³² Any 'estrogen-like' molecule enhances growth and promotes selection of hormone responsive disease. A recent report by Bernstein and colleagues¹³³ analyzed the incidence of endometrial cancer associated with tamoxifen therapy with known risk factors for the disease. This population of women warrant careful evaluation before the use of tamoxifen. The women most likely to develop endometrial cancer with tamoxifen therapy have previously used HRT and have known risk factors, such as obesity. There is no reported increase in risk of endometrial cancer in premenopausal women¹³³ inasmuch as menstrual cycles continue during tamoxifen treatment for most women.

Another life-threatening side effect of tamoxifen is thromboembolic disease. Women taking tamoxifen have more strokes, pulmonary embolisms, and deep-vein thrombosis than age-matched controls. According to the NSABP B-14, women taking tamoxifen had a 0.9% incidence of thromboembolism, compared to 0.2% in the placebo group.¹¹³  One of those patients died of a pulmonary embolism. 

In terms of non life-threatening side effects, tamoxifen use results in an elevated risk of cataracts (RR=1.14, 95% CI=1.01–1.29), which is associated with an increased rate of cataract surgery.¹²⁹ Women taking tamoxifen also have more hot flashes, vaginal dryness, and vaginal discharge. 

Within the past decade, aromatase inhibitors (AI) have become an option for post-menopausal women with hormone receptor-positive breast cancer.  After menopause, a woman's major source of estrogen is through the peripheral conversion of adrenal steroid hormones to estrogen.  The AIs inhibit the aromatase enzyme responsible for this conversion, which facilitates a very complete suppression of estrogen.   

Over 25 years ago, the first AI, aminoglutethimide, became available.  The drug had significant toxicity which limited its clinical use.¹¹⁴ Third generation AIs were evaluated in Phase I and II trials for women with metastatic breast cancer.  Once these drugs were shown to be effective in the metastatic setting, they were evaluated in the adjuvant setting. The three AIs currently being used in clinical practice are anastrozole, letrazole, and exemestane.

The landmark trial which established the superiority of AIs over tamoxifen was the ATAC trial (Arimidex, Tamoxifen, Alone or in Combination) published in 2003.  This study was a prospective randomized controlled trial that compared adjuvant therapy with tamoxifen versus arimidex, and was limited to postmenopausal women with hormone receptor-positive breast cancer.¹³⁴  Patients were randomized to either tamoxifen, arimidex, or a combination of both.  The combination arm was stopped early because it did not demonstrate any benefit over tamoxifen alone.  At four years, arimidex was superior to tamoxifen in terms of DFS (86.9% vs. 84.5% with a HR= 0.86, 95% CI 0.76–0.99, p=0.03). Arimidex was also superior in terms of time to recurrence (TTR), with a HR=0.83 (95% CI 0.71–0.96; p=0.015).  There was no statistically significant difference between treatments regarding overall survival (OS).   

In 2005, an updated analysis of these data was released with 68 months follow-up.¹³⁵ Treatment with anastrozole continued to demonstrate a significant improvement in DFS with a HR = 0.87 (95% CI 0.73–0.94) and TTR with a HR = 0.74 (95% CI = 0.64–0.87). There was still no statistically significant difference in overall survival.  Fewer patients withdrew from anastrozole treatment than from tamoxifen treatment. The conclusion of the updated analysis was that anastrozole was the treatment of choice for post-menopausal hormone receptor-positive breast cancer patients.

Since the ATAC was published, several other trials have been initiated to study the role of AIs in the treatment of breast cancer. The BIG 1-98 trial compared another AI, letrozole, with tamoxifen and also demonstrated an improved DFS.¹¹⁵ Other trials have been initiated to determine the benefit of adding an AI after an abbreviated course of tamoxifen.  Sequential trials are looking at a five-year course of tamoxifen, followed by treatment with an AI. These trials will help to determine the optimal use and sequencing of aromatase inhibitor therapy. 

Arimidex (anastrozole) has a very different side effect profile to tamoxifen. The ATAC trial provided a useful comparison of drug side effects amongst similar groups of patients. Since anastrozole does not have partial agonist effects on the uterus, it is not associated with an increased risk of endometrial cancer.  In the ATAC trial, the risk of endometrial cancer with anastrozole was 0.1%, which is comparable to age-matched controls. Accordingly, anastrozole offers a significant risk reduction for this co-morbidity. Anastrozole is associated with a lower risk of thromboembolic events compared to tamoxifen (2.1% vs. 3.5%, p=0.0006).  There were also fewer strokes and pulmonary embolisms in the anastrozole group.

Anastrozole was associated with an increased risk of musculoskeletal problems.  Arthralgias were much more common and often led to discontinuation of the drug. Osteoporotic fractures were increased with anastrozole use (5.9% vs. 3.7%, p=<0.0001).  Although the rate of hip fractures was equivalent, there were more spine and wrist fractures in the anastrozole group.  In terms of quality of life, hot flashes, vaginal discharge, and vaginal bleeding were much less common in patients taking anastrozole.

The role of neoadjuvant chemotherapy in the treatment of breast cancer was established by the NSABP B-18 and the EORTC 10902 trials.^{116, 117} The NSABP B-18 randomized patients with primary operable breast cancer to either preoperative or postoperative adriamycin and cytoxan. The conclusion of the study was that an additional 12% of lumpectomies could be performed after neoadjuvant chemotherapy due to downsizing of the tumor, with a 37% increase in the incidence of pathologically negative axillary lymph nodes. The EORTC trial randomized patients to preoperative or postoperative fluorouracil, epirubicin, and cyclophosphamide. The conclusion of this trial was that no difference in overall survival or time to local regional recurrence was found for patients treated in the neoadjuvant setting. Based on these two trials, and subsequently several others, patients with larger primary tumors are offered preoperative chemotherapy to potentially facilitate breast conservation as a surgical option.

Subsequent to the success of neoadjuvant chemotherapy, trials were designed to evaluate the potential role of neoadjuvant hormonal therapy. A study by Eiermann et al compared neoadjuvant letrozole with tamoxifen in hormone receptor-positive patients who either required mastectomy or had locally advanced or inoperable breast cancers.¹³⁶ Clinical objective response rates (ORR) were better with letrozole than with tamoxifen (55% vs. 36%, p<0.001). More patients in the letrozole group were treated with breast conserving surgery (45% vs. 35%, p<0.022). There have been two trials comparing neoadjuvant anastrozole with tamoxifen, the IMPACT and PROACT trials.^{137, 138} A combined analysis of these trials failed to demonstrate a significant difference in ORR between the different treatment groups. There was an improvement in ORR for patients treated with anastrozole who were either inoperable or were thought to require a mastectomy (47% vs. 35%, p=0.026). These trials would suggest that aromatase inhibitors may be more effective than SERMs in reducing tumor size. The American College of Surgeons Oncology Group (ACOSOG) and the Cancer Trials Support Unit are currently accruing patients for the Z1031 trial which will compare the effectiveness of letrozole, anastrozole, and exemestane in the neoadjuvant setting. Based on the available data, neoadjuvant endocrine therapy appears to be effective in regard to the ORR and the ability to downstage tumors with limited side effects.¹³⁹

Neoadjuvant therapy also has significant implications for research. Tumor tissue can be studied prior, during, and after the completion of systemic therapy to evaluate the response of the tumor to specific treatments.  Valuable data can be acquired using biomarkers, RNA profiling, p53 microarray analysis, and comparative genomic hybridization, which will hopefully help to identify patients who will have a better response to therapy. 

The decrease in contralateral breast cancers following adjuvant therapy and favorable toxicity profile in clinical practice suggested tamoxifen could play a role in breast cancer prevention. Furthermore, estrogenic hormones act over many years to promote a population of initiated cells. Given that the promotional stage is characterized by reversibility, much interest has been shown in preventing the consequences of prolonged, unopposed estrogenic stimulation of the breast.

The National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1¹²⁹ opened in the United States and Canada in May 1992 with an accrual goal of 16,000 women from 100 North American sites. The specific aim was to test the utility of tamoxifen as a preventive for breast cancer. It closed after accruing 13,338 women in 1997. Women over the age of 60, or women between the ages of 35 and 59 whose 5-year risk of developing breast cancer, as predicted by the Gail model,¹⁴⁰ was equal to that of a 60-year-old woman, were eligible for inclusion. Additionally, women over age 35 with a diagnosis of lobular carcinoma in situ (LCIS) treated by biopsy alone were eligible for entry. In the absence of LCIS, the risk factors necessary to enter the study varied with age, such that a 35-year-old woman needed an RR of 5.07, whereas the required RR for a 45-year-old woman was 1.79. Routine endometrial biopsies were performed in both arms of the study.

The breast cancer risk of women enrolled in the study was generally extremely high. Recruitment was also age balanced, with about one third younger than 50 years, one third between 50 and 60 years, and one third older than 60 years. Secondary endpoints of the study include the incidence of fractures and cardiovascular deaths.

The first results of the NSABP study were reported in September 1998, after a mean followup of 47.7 months.  The data safety monitoring committee prematurely unblinded the study due to a significant reduction in the incidence of breast cancer among patients taking tamoxifen. In total, 363 invasive and noninvasive breast cancers occurred in the participants; 124 in the tamoxifen group and 239 in the placebo group. There was a 49% reduction in the risk of invasive breast cancer and a 50% reduction in the risk of noninvasive breast cancer in patients treated with tamoxifen. A subset analysis demonstrated that women with LCIS had a 56% reduction in risk. Women with atypical hyperplasia had an 86% reduction in risk.

The benefits of tamoxifen were observed in all age groups, with an RR of breast cancer ranging from 0.45 in women aged 60 and older to 0.49 for those in the 50–59 year-old group, and 0.56 for women aged 49 and younger (Fig. 5). The benefit of tamoxifen was observed in all levels of breast cancer risk within the study and was not confined to a particular lower- or higher-risk subset.

As expected, tamoxifen affected the incidence of ER-positive tumors, which were reduced by 69% per year. The rate of ER-negative tumors in the tamoxifen group (1.46 per 1,000 women) did not significantly differ from that of the placebo group (1.20 per 1,000 women) (Fig. 6). Tamoxifen reduced the rate of invasive cancers of all sizes, but the greatest reduction was in the incidence of tumors 2.0 cm or smaller. Tamoxifen also reduced the incidence of both node-positive and node-negative breast cancer. The beneficial effects of tamoxifen were observed for each year of follow-up in the study. After year 1, the risk was reduced by 33%, and in year 5, by 69%.

Tamoxifen also reduced the incidence of osteoporotic fractures of the hip, spine, and radius by 19% (Fig. 7),¹²⁹ approaching, but not reaching statistical significance. This reduction was greatest in women who were 50 years and older at study entry. No differences in the risks of myocardial infarction, angina, coronary artery bypass grafting, or angioplasty were noted between groups.¹²⁹ The prevention study also confirmed the association between tamoxifen and endometrial carcinoma.¹¹³ The relative risk of endometrial cancer in the tamoxifen group was 2.5, with women over age 50 having an RR of 4.01. All endometrial cancers in the tamoxifen group were grade 1 and none of the women receiving tamoxifen died as a result of endometrial cancer. One endometrial cancer death occurred in the placebo group. Although no doubt remains that tamoxifen increases the risk of endometrial cancer, it is important to recognize that this increase translates to an incidence of 2.3 women per 1,000 per year.¹²⁹

More women in the tamoxifen group developed deep vein thrombosis (DVT),¹²⁹ with the excess risk confined to women over 50 years. The RR of DVT in the older age group was 1.71 (95% confidence interval [CI], 0.85 to 3.58). An increase in pulmonary emboli was also seen in the older women taking tamoxifen, with an RR of approximately 3. Three deaths from pulmonary emboli occurred in the tamoxifen study arm, all in women with significant comorbidities. An increased incidence of stroke, RR 1.75, was also seen in the tamoxifen group, but this did not reach statistical significance.

An assessment of quality of life showed similar depression scores between groups. Hot flashes were reported in 81% of the women taking tamoxifen compared with 69% of the placebo group, and the tamoxifen-associated hot flashes appeared to be of comparable severity. Twenty-nine per cent of the women in the tamoxifen group and 13% in the placebo group reported moderate or severe vaginal discharge. No differences in the occurrence of irregular menses, nausea, fluid retention, skin changes, or weight gain or loss were reported.

In 2005, an update of the NSABP P-1 trial was published with 7 years of follow-up.¹⁴¹ The results were very similar to the original data in terms of the decreased incidences of invasive and noninvasive breast cancer and osteoporotic fractures. There were comparable increases in endometrial cancer, thromboembolic events, and cataracts in women taking tamoxifen, which did not differ from the original statistics. Despite the increased length of follow-up, there was no survival benefit demonstrated. The authors contend that the study was not designed to detect a survival benefit, and that 15–20 years of follow-up would be needed.

The ATAC trial also demonstrated that in the adjuvant setting, anastrozole decreased the incidence of contralateral breast cancers.  The more favorable side effect profile of AIs makes them a more appealing option for healthy high-risk women who are trying to modulate their breast cancer risk. The International Breast Cancer Intervention Study II (IBIS-II)¹⁴² is a trial which is currently accruing high-risk post-menopausal women and randomizing them to anastrozole versus placebo. 

Raloxifene is a SERM currently approved for the treatment of osteoporosis. Evaluation of 60-mg daily raloxifene on bone homeostasis in postmenopausal women found that although the suppression of remodeling was greater for estrogen, the remodeling balance was the same for the two agents.¹⁴³ Raloxifen increases bone by 2.4% in the lumbar spine and hip, and has been shown to decrease spine fractures by 40%.¹⁴⁴

Raloxifene inhibits the growth of 7,12-dimethylbenz[a]-anthracene (DMBA)-induced rat mammary carcinoma,¹⁴⁵ and more important, for the proposed evaluation as a preventive therapy, raloxifene reduces the incidence of N-nitrosomethylurea-induced tumors if given after the carcinogen, but before the appearance of palpable tumors.¹⁴⁶ As would be anticipated with a drug that has a short biologic half-life, raloxifene is not superior to tamoxifen at equivalent doses. Raloxifene and its analogs are clearly effective and potent inhibitors of the growth of breast cancer cells in culture,^{147, 148} but the complication of first pass metabolism in vivo reduces potency. For this reason, doses above 60 mg daily have been tested in clinical trials to prevent osteoporosis.

Based on the hypothesis that raloxifene reduces incidence of breast cancer as a beneficial side effect of the prevention of osteoporosis,¹⁴⁹ placebo-controlled trials with raloxifene are ongoing. The Multiple Outcomes of Raloxifene Evaluation (MORE) randomized 7,704 postmenopausal women, mean age 66.5 years, with osteoporosis defined by hip or spine bone density at least 2.5 standard deviations (SDs) below normal or vertebral fractures, and no history of breast or endometrial cancer, to placebo, 60 mg, or 120 mg of raloxifene daily. Results at three years, with a total of 40 cases of breast cancer confirmed, indicate a 76% reduction in the risk of breast cancer.¹⁵⁰ Another database pooled all placebo-controlled trials and included 10,553 women monitored for, on average, three years. In this group there was a 54% reduction in the incidence of breast cancer in the raloxifene-treated patients.¹⁵¹ Similar to tamoxifen, raloxifene reduces the incidence of ER-positive breast cancer only.

Raloxifene's effect in the human uterus is currently being evaluated. A study in women without preexisting endometrial abnormalities shows that raloxifene, unlike estrogen, does not increase endometrial thickness.¹⁵² Raloxifene is less estrogenic than tamoxifen and only increases the growth of human endometrial carcinomas under laboratory conditions by approximately 50% of tamoxifen.¹²⁸ Raloxifene has not been found to increase incidence of endometrial cancer during trials of osteoporosis treatment.¹⁵⁰

Preliminary data on the ability of raloxifene to decrease the risk of breast cancer and the risk of endometrial cancer led to the NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial.¹⁵³ This trial was a phase III, prospective, double-blind, randomized trial that assigned 19,747 postmenopausal high-risk women to either five years of tamoxifen or raloxifene. The results of this study were published in 2006 and demonstrated that raloxifene was as effective as tamoxifen in decreasing the incidence of invasive breast cancer (RR=1.02, 95% CI, 0.82–1.28).  Tamoxifen appeared to be more effective at decreasing noninvasive breast cancers compared to raloxifene (RR=1.40, 95% CI, 0.98–2.00), although these results did not reach statistical significance. Raloxifene was associated with a lower risk of thromboembolic events, with a 30% decrease in pulmonary embolism and DVT (RR, 0.79; 95% CI, 0.54–0.91). There were fewer cataracts in the raloxifene group (RR=0.79, 95% CI, 0.68–0.92), which also resulted in fewer cataract surgeries. Although there was a trend towards a decrease in the risk of endometrial cancer and endometrial hyperplasia, this decrease did not reach statistical significance. The risk of developing other cancers, bone fractures, and ischemic heart disease was similar between the two treatment groups.

The central issue for research on hormone receptor pharmacology is the discovery of mechanisms of target site specificity for the modulation of estrogenic and antiestrogenic response. The model must encompass the sum of pharmacologic consequences of signal transduction through ERα and ERβ with the simultaneous competition from endogenous estrogens at both sites. Dissection of the underlying pathways responsible for this activity may permit the development of tissue-selective ER modulators.¹⁰⁰

The development of ERβ monoclonal antibodies, the classification of target sites for the protein, and the continuing evaluation of ERβ and ERα,β knock-out mice will identify new therapeutic targets to modulate physiologic functions.¹⁵⁴ Clearly, the successful crystallization of ERα with raloxifene⁸⁸ has acted as a stimulus for the crystallization of ERβ with SERMs.¹⁵⁵ However, the description of an agent that produces agonist or antagonist effects exclusively at ERα or ERβ will determine the therapeutic usefulness of ERβ as a target for disease prevention.

Hormonal therapy has made a major contribution to the treatment of breast cancer. Reduced recurrence rates and longer survival will continue to drive the need for newer agents to use after first-line hormonal therapy. Duration of therapy continues to be a key issue. Additional trials and longer follow-up of existing trials are needed to answer this question. The results of sequencing and switching trials will be essential to help determine optimal hormonal therapy regimens.  Individualized treatment plans will continue to be important to minimize side effects and maximize benefit for patients.  Additionally, resistance to tamoxifen has emerged as a clinical problem, requiring the development of novel agents to circumvent resistance in this patient population.

Major clinical questions remain regarding the best application of tamoxifen, raloxifene, or AIs as chemoprevention. The P-1 and P-2 trials have established the benefits of tamoxifen and raloxifene in high-risk patients. The results of the IBIS-II will help to define the role of AIs as chemopreventive agents. When the goal is prevention rather than cure, the balance between risk and benefit is particularly critical, necessitating greater efforts towards minimizing side effects.

The era of gene sequencing and molecular analysis has expanded our knowledge of breast cancer tremendously.  Stratification of recurrence risk using the Oncotype Dx and other genomic assays will allow physicians to tailor therapy to the individual cancer patient. Cytotoxic chemotherapy will ultimately be reserved for those patients most likely to recur.  Further research will continue to identify the differences in cancers which can then be exploited for therapeutic purposes.

Breast cancer mortality has declined slightly for the first time in 30 years.^{156, 157} Of particular impact has been the appropriate and systematic use of antiestrogens in more recently diagnosed patients. Hopefully, the development of new chemopreventives will reveal an even greater potential for improved patient outcomes.