Introduction

Selective estrogen receptor modulators (SERMs), previously known as “antiestrogens,” are a category of therapeutic agents used for the treatment of not only breast cancer, but also common diseases such as osteoporosis. SERMs act as an antagonist for the estrogen receptor (ER) in breast cancer, while acting as an agonist in some other tissues such as bone [1]. Currently, there are two classes of clinically used SERMs. One is the triphenylethylene derivatives, such as tamoxifen and toremifene, and the other is the benzothiophene derivatives, such as raloxifene (Fig. 1). The newly developed SERMs, arzoxifene, droloxifene, and idoxifene, have higher affinity for ER than tamoxifen and were expected to be at least equal to and possibly better than tamoxifen for the treatment of breast cancer [2, 3]. However, neither arzoxifene, droloxifene, nor idoxifene were shown to be superior to tamoxifen with respect to time to progression or response rate in patients with locally advanced or metastatic breast cancer who had not been previously treated for metastatic disease [4–6]. Thus, it is obvious that the 30-year-old veteran SERM, tamoxifen, remains the gold standard for the current breast cancer treatment. In this article, we present an overview of SERMs in breast cancer treatment and related topics.

Fig. 1
figure 1

Structure of selective estrogen receptor modulator (reproduction based on figure courtesy from Prof. S. Inoue and reference [42])

Full size image

How long should tamoxifen be used in breast cancer treatment?

Among women with ER-positive breast cancer, tamoxifen reduces the risk of recurrence and death after surgery and can provide palliation in those with metastatic disease. In the EBCTCG overview published in 2005, 5 years of tamoxifen was associated with a 41% proportional reduction in the annual risk of recurrence and a 34% reduction in the annual risk of death [7]. Proportional risk reductions were similar between younger and older women, and between node negative and node-positive breast cancer. This benefit was also independent of the use of chemotherapy. Compared to chemotherapy alone, the addition of tamoxifen after completion of chemotherapy reduced the risk of annual recurrence to 31% and the risk of annual death to 34% [7]. Several trials with predominantly postmenopausal patients have shown significantly better outcomes from 5 years, as compared to 2 years of tamoxifen administration [8–11]. The EBCTCG overview also confirmed that 5 years is significantly more effective than 1–2 years of tamoxifen (HR = 0.82, SE 0.03, 2 P < 0.00001, for recurrence, HR = 0.91, SE 0.04, 2 P = 0.01 for breast cancer mortality) [7]. In cases of ER-positive breast cancer patients, 5 years of tamoxifen is the reliable choice as adjuvant endocrine therapy.

There have been three large trials comparing 5 years versus longer tamoxifen treatment duration. The results from the NSABP B-14 trial [12] and the Scottish trial [11] demonstrated that no additional benefit was observed in those randomly assigned to continue taking tamoxifen beyond 5 years. Moreover, relatively worse prognosis were observed in patients with longer treatments. However, in the ECOG trial, longer duration of treatment was associated with significantly favorable relapse-free survival in ER-positive patients (P = 0.014), although the survival difference for this subgroup was not statistically significant (P = 0.81) [13]. The difference between the former two studies and the ECOG trial is nodal involvement of enrolled patients. All patients in the B-14 trial and the majority of the Scottish trial attendants were node-negative, whereas all women were node-positive in the ECOG trial. It is interesting that in the NCIC-CTG MA17 trial, which evaluated the impact of the addition of 5 years of treatment with the aromatase inhibitor letrozole to the patients completing the initial 5 years of tamoxifen treatment, patients with node-positive disease achieved a significant survival benefit compared to those receiving placebo (HR 0.61, 95% CI = 0.38–0.98) [14]. The most recently reported large randomized study (n = 11,500), the ATLAS trial, showed that 10 years of tamoxifen resulted in a 12% reduction in annual risk of recurrence in the 5–9 years after surgery and in a 22% reduction in the 10–14 years after surgery compared to 5 years of tamoxifen [15]. Very recently, the preliminary result of aTTom trial was reported in ASCO 2008. Although there was non-significant reduction in recurrence of breast cancer, as a meta-analysis of previous five trials (aTTom, ATLAS, Scottish, B-14, ECOG) it was shown that 10 years tamoxifen reduced risk of recurrence among ER-positive or unknown patients compared to 5 years (OR=0.90, 95% CI=0.84–0.98). These observations suggest that prolonged treatment of over 5 years of tamoxifen may have a survival benefit in a subset of patients. This notion might be important for premenopausal patients with nodal involvement who have survived the early recurrence period and are still premenopausal after completion of 5 years of initial endocrine treatment.

SERMs for ductal carcinoma in situ (DCIS) and prevention

It is clear that the aromatase inhibitors offer an advantage over tamoxifen as adjuvant therapy for postmenopausal patients [16]. Although tamoxifen is still the recommended choice for certain patients, such those with severe osteoporosis or those with low risk of recurrence, these seem to be the minority of cases. The current most frequent applications of tamoxifen are for premenopausal patients and patients with ductal carcinoma in situ (DCIS).

Two trials have directly studied the use of tamoxifen in women with DCIS (Table 1). The NSABP B-24 trial randomly assigned 1,804 women with DCIS undergoing partial excision plus radiation therapy (RT) to 5 years of tamoxifen or placebo [17].

Table 1 New breast events in the tamoxifen randomized trials for DCIS
Full size table

Tamoxifen reduced the 5-year cumulative risk of all breast cancer events by 5.2% (P = 0.0009). The relative reduction with tamoxifen was statistically significant only for ipsilateral invasive tumors (HR = 0.56, 95% CI = 0.32–0.95) and contralateral non-invasive breast cancer (HR = 0.22, 95% CI = 0.04–0.81). In the UK/ANZ Trial, which randomly assigned 1,701 women to one of four groups following excision of DCIS with clear margins (excision alone, excision plus RT, excision plus tamoxifen, and excision plus RT plus tamoxifen), a benefit from tamoxifen was not observed either in terms of the risk reduction of all breast cancer events or a reduction in contralateral breast cancer (HR = 0.52, 95% CI = 0.25–1.07) [18]. The benefit was limited to risk reduction of non-invasive breast tumors (HR = 0.68, 95% CI = 0.49–0.96).

Thus, the available trial data are conflicting as to the benefit of tamoxifen in DCIS. This discrepancy is thought to depend on the difference in margin involvement of patients with lumpectomy and the presence or absence of radiation. Tamoxifen may be indicated as post-operative therapy in patients who wish to reduce the risk of contralateral breast cancer or to reduce the risk of ipsilateral breast recurrence after breast conservation therapy with a suspicious margin.

In the same vein as these findings, the application of tamoxifen as chemoprevention in women at increased risk of the development of breast cancer has been studied (Table 2). The NSABP Breast Cancer Prevention trial (protocol P-1) of 13,388 such women showed that the cumulative rate of invasive breast cancer was reduced from 42.5 per 1,000 women in the placebo group to 24.8 per 1,000 women in the tamoxifen group (RR = 0.57, 95% CI = 0.46–0.70) [19]. In addition, the cumulative rate of noninvasive breast cancer was reduced from 15.8 per 1,000 women in the placebo group to 10.2 per 1,000 women in the tamoxifen group (RR = 0.63, 95% CI = 0.45–0.89). For all types of breast cancer, a smaller degree of benefit (RR = 0.73, 95% CI = 0.58–0.91) was reported in a second smaller trial, the IBIS-I trial, involving 7,145 women [20, 21], while two other smaller trials, the Royal Marsden (n = 2,494) [22] and Italian (n = 5,408) [23] trials, have had negative results for their primary objectives. Although the results of these four trials are inconsistent, at least for the prevention of ER-positive invasive breast cancer, all trials except for the Italian trial suggested a statistically significant decrease in risk. Also in the Italian trial, when analyses were restricted to women at high risk of hormone receptor (HR)-positive breast tumors (taller than 160 cm, at least one intact ovary, <14 years of age at menarche, no full term pregnancy before age 24 years), tamoxifen groups showed statistically significant reduction of breast cancer incidence (RR = 0.24, 95% CI = 0.10–0.59) [23]. Since none of the individual trials showed whether a reduction in incidence of breast cancer will lead to a reduction in breast cancer-related mortality, the use of tamoxifen for prevention should be limited to the specific population at higher risk of breast cancer. Appropriate candidates for breast cancer chemoprevention could be women who meet the eligibility criteria for the P-1 trial [19]: age over 60, a history of LCIS, or those between 35 and 59 years of age with an estimated risk of breast cancer of 1.66% or higher over 5 years as estimated by the Gail model. However, it should be noted that the Gail model is not fully applicable to the Asian population, and therefore currently we do not have an established risk population indicated for chemoprevention for Asian. In daily practice, the indication for 5 years of tamoxifen may be patients who have undergone surgery for DCIS or LCIS and who wish to prevent newly developed breast cancer.

Table 2 Summary and rate of breast events in the tamoxifen and raloxifene prevention trials
Full size table

Raloxifene, currently approved as the therapeutic drug for postmenopausal osteoporosis in Japan, has an agonistic effect on bone and lipids, but an antagonistic effect on the breasts and uterus. Raloxifen has been studied in four trials as a chemopreventative agent for breast cancer without an increase in endometrial cancer (Table 2). The multiple outcomes of the raloxifene evaluation (MORE) trial was originally designed to test the hypothesis that raloxifene would lower the risk of osteoporotic fractures [24]. The study involved 7,705 postmenopausal women under the age of 80 who had osteoporosis. Subjects were randomly assigned to receive placebo or raloxifene at a dose of 60 or 120 mg per day. With 4 years of follow-up, there was an 84% relative reduction (RR = 0.16, 95% CI 0.09–0.30) in the incidence of ER-positive invasive breast cancer among the 5,129 women administered raloxifene compared to the 2,576 receiving placebo. In the continuing outcomes relevant to Evista (CORE) trial, women who had been randomly assigned to receive raloxifene (either 60 or 120 mg/day) in MORE were assigned to receive raloxifene (60 mg/day) (n = 3,510), and women who had been assigned to receive placebo in MORE continued on placebo [25]. The incidences of invasive breast cancer and ER-positive invasive breast cancer were reduced by 66% (HR = 0.34; 95% CI = 0.22–0.50) and 76% (HR = 0.24; 95% CI = 0.15–0.40), respectively, in the raloxifene group compared with the placebo group. In the Raloxifene Use for The Heart (RUTH) Trial, in which 10,101 postmenopausal women with coronary heart disease or risk factors for coronary heart disease were randomly assigned to raloxifene (60 mg/day) or placebo, raloxifene reduced the risk of invasive breast cancer compared with placebo (RR = 0.56; 95% CI = 0.38–0.83), and this benefit was primarily due to a reduced risk of estrogen-receptor-positive invasive breast cancers [26].

The NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial was a large prospective randomized clinical trial directly comparing raloxifene and tamoxifen in 19,747 postmenopausal women at high risk of breast cancer [27]. Women were randomly assigned to receive 60 mg of raloxifene or 20 mg of tamoxifen daily for 5 years. The number of invasive breast cancers that developed in each group was statistically equivalent (RR = 1.02, 95% CI 0.82–1.28). Taken together, the results of the MORE/CORE, RUTH, and STAR trials suggest that raloxifene can reduce the incidence of invasive breast cancers in high-risk women.

In the STAR trial, the raloxifene group had more cases of noninvasive breast cancer than the tamoxifen group, although the difference was not statistically significant (RR 1.40, 95% CI 0.98–2.00) [27]. Raloxifen did not decrease the incidence of noninvasive breast cancer in the MORE/CORE and RUTH trials. The reason for the lesser preventive effect for DCIS of raloxifene compared to tamoxifen is unclear. Jordan [28] in his commentary paper noted that this may due to the extremely short duration of action of raloxifene compared to tamoxifen, referring to the pharmacological lifespan of the two agents. In poor compliance patients, raloxifene failed to maintain adequate concentration for the prevention of the early steps of carcinogenesis. Beside this notion, an understanding of the molecular differences between raloxifene and tamoxifen on the carcinogenesis of DCIS and invasive ductal cancer might be important. We are interested in the difference in activation of the AP-1 responsive element through ERα and ERβ by raloxifene and tamoxifen [29].

Prevention: for whom?

Exactly who develops hormone-dependent breast cancer is the upcoming area of interest in breast cancer prevention, as current promising SERMs such as tamoxifen and raloxifene only prevent hormone-receptor-positive breast cancer and have no effect on receptor-negative tumors (Table 2). Moreover, in the Italian prevention trial, tamoxifen treatment for women at low risk of hormone receptor-positive tumors (described in the previous section) was associated with a non-statistically significant increase of breast cancer incidence (RR = 1.46, 95% CI = 0.84–2.53) [23]. Some of the reported chemoprevention trials with SERMs have used the Gail model for defining high-risk women for breast cancer; however, this model estimates the overall breast cancer risk and is not hormone-dependent specific.

For instance, it is widely known that obesity is a risk factor for postmenopausal breast cancer, due to the possible contribution of local production of estrogen in fat tissue [30]. Obesity increases the risk of hormone-dependent breast cancer in postmenopausal women, whereas it has no contribution to the incidence of hormone-receptor negative ones [31]. More interestingly, from an epidemiologic study with a Swedish cohort, this effect is obvious for ER + PR + tumors, but not for ER + PR- cases [32]. This has also been observed in a meta-analysis of previous reports.

The most frequent age for the incidence of ER + PR + and ER + PR- breast cancer varies. In the database of a white non-Hispanic population, the peak age of ER + PR + patients is younger than that of ER + PR- breast cancer [33]. A report from a single Japanese institution comprising 3,600 cases showed that ER + PR + patients were observed more frequently in premenopausal women, while ER + PR- patients were dominantly post-menopausal [34]. The biological properties of breast cancers also differ, in that Her2 and/or EGFR positive tumors are more frequently ER + PR- cancers than ER + PR + cancers [35]. These observations imply that the biological nature of these subtypes differs, either in the occurrence of breast cancer cells or in possible acquired modifications to breast cancer character due to circumstances such as estrogen concentration. These issues need to be clarified by further research to develop a chemo-prevention strategy for a suitable population and with suitable timing.

Differences in the effects of SERMs between premenopausal and postmenopausal women

The differing effects of SERMs between pre- and postmenopausal women is interesting. An increase in the incidence of endometrial cancer by long-term tamoxifen was one of the important side effects of tamoxifen used in an adjuvant setting. However, this observation was mainly limited to women over the age of 50, and there was no obvious sign of this when tamoxifen alone was used in premenopausal women [36]. Therefore, the recommendation by the American College of Obstetricians and Gynecologists stated that premenopausal women treated with tamoxifen require no additional annual monitoring beyond routine gynecologic care [37].

It is widely known that tamoxifen and raloxifen increase bone mineral density in postmenopausal women; however, these drugs decrease it when used in premenopausal women [38]. Because tamoxifen is a competitive inhibitor against estrogen, the circulating level of estrogens influences both the harmful and beneficial effects of tamoxifen. In the absence of estrogen, tamoxifen seems to exert the full range of its agonistic and antagonistic properties. In metastatic breast cancer, the LH-RH analogue plus tamoxifen is more effective than tamoxifen alone or LH-RH analogue alone [39, 40]. All these observations suggest that the concentration of circulating estrogen affects both the favorable and unfavorable effects of tamoxifen.

In postmenopausal women, the same issue arises. In the ATAC trial, there was significant benefit in risk of recurrence after 5 years of anastrozole versus tamoxifen [41]. However, anastrozole with tamoxifen had an almost equal effect as tamoxifen with respect to disease-free survival. In this case, a reduction in peritumoral estrogen concentration by anastrozole did not enhance the effect of tamoxifen; moreover, tamoxifen reduced the benefit of anastrozole. There has been previous discussion regarding accelerated metabolization in combination usages [41]. In a phase III trial for advanced or metastatic diseases, toremifene with the aromatase inhibitor, atamestane, has equal efficacy to letrozole [42]. In this case, although the study design was complex, a reduction in peritumoral estrogen levels may have enhanced the anti-tumor effect of toremifene, or toremifene did not affect the anti-tumor effect of atamestane. In postmenopausal women, the relationship between circulating and peritumoral estrogen levels and the effects of SERMs are unclear.

Novel application of high-dose SERMs

SERM is generally used for inhibition or modification of ER function. This is mainly due to its property as both an agonist and antagonist for estrogen signaling. However, SERMs possibly have cross reactions with other receptor systems when used in relatively high doses. These features render the SERMs positive or negative therapeutic bullets in different circumstances.

There have been several reports regarding the effect of high-dose SERMs for ER-independent cells. For instance, around 1–10 μM tamoxifen and toremifene inhibits the proliferation of ER-negative breast cancer cell lines [43, 44]. Hui et al. [45] reported that 10 μM tamoxifen and CC-8490, a novel benzopyranone with SERM activity, induce apoptosis in glioma cell lines in a dose-dependent manner, although high-dose fulvestrant failed to do so. All examined cell lines, U87MG, U138MG, and U373MG glioma cells, were negative for ERα and ERβ.

Since around 1990, there have been several reports of the effects of estrogens, tamoxifen, and other SERMs on the reversal of chemotherapy drug-resistance of cancer cells. In one study, 10 μM of tamoxifen enhanced the chemo-sensitivity of bladder cancer cell lines to methotrexate, vinblastin, doxorubicin and cisplatin [46], and multidrug-resistant variant MCF-7/ADR cells to mitoxantrone [47]. Rao et al. [48] have reported that multiple panels of SERMs and steroid hormones can be active substrates transported by p-glycoprotein (Pgp), which is related to multiple drug resistance. They examined tamoxifen, droloxifen, toremifene, clomifene, nafoixdine, diethylstilbestrol, estradiol, progesterone, hydrocortisone, and corticosterone. Mubashar et al. [49] attempted to define the ability of 3 days of toremifene (780 mg/day) to inhibit the effect of Pgp in 20 breast cancer patients using (99 m)Tc-sesatamibi imaging. They reported that modulation of Pgp function by toremifene may be due to toremifene acting as a competitive substrate against other agents in vivo. The mechanisms of these actions are not clearly understood. Some reports have suggested that post-transcriptional modification of Pgp [50], inhibition of protein kinase C activity [51], and inhibition of ceramide glycosylation [52] are involved.

The question that arises regarding high-dose SERM usage is what drug serum concentration can be achieved in vivo. For instance, from a phase I trial of toremifene, 120 mg/day use of toremifene produced a constant 5 μM toremifene-1 (TOR-1), an active metabolite of toremifene [53]. Wurz et al. [54] reported that 360 mg/day for 5 days of toremifene achieved a 10–15 μM plasma concentration in their patients. In a phase II study in Japanese breast cancer patients, 240 mg/day toremifene over 8 weeks was able to be used without severe adverse events [55]. Stuart et al. [56] tried to treat 26 patients with relapsed or drug-resistant cancer with a combination of oral etoposide (300 mg/day for 3 days) and tamoxifen (480 or 720 mg/day for 6 days). They reported two partial response cases of high-grade lymphoma and adenocarcinoma in this study, and the serum level of tamoxifen reached 3–3.5 μM. A phase I trial of 200–700 mg/day tamoxifen for 7 days in 14 acute leukemia patients in combination with daunorubicin produced plasma tamoxifen levels approaching 7 μM at the two highest dose levels [57]. In a phase I/II study with renal cell cancer patients administered 780 mg/day toremifene for 3 days every 2 weeks in combination with vinblastine, the mean serum concentration of toremifene was 7.82 μM, which exceeded that known to reverse multi-drug resistance in vitro [58].

Due to the current lack of data regarding the adverse effects of high-dose SERMs when used long term, any attempt to use high-dose SERMs should be over a short period and in a research setting only. However, effective drug combinations or well-designed studies may reveal new useful aspects of SERM usage.