Over 80% of women suffering from advanced breast cancer ultimately develop bone metastases [1,2]. As revealed by observations published more than a decade ago [3], patients with estrogen receptor (ER)-positive neoplasms are more prone to develop skeletal secondaries. Metastatic breast cancer cells stimulate osteoclast-mediated bone resorption, inducing a marked osteolysis that is responsible for considerable morbidity [4,5].

Bisphosphonates are potent inhibitors of osteoclast-mediated osteolysis [6] and have, therefore, emerged as a rational approach for the management of bone metastases [7,8]. These drugs are synthetic analogs of pyrophosphate. They show high affinity for bone mineral and preferentially accumulate at sites of active bone remodeling. The most potent bisphosphonates are nitrogen-containing compounds (e.g. ibandronate, zoledronic acid) that interfere with the mevalonate pathway, leading to inhibition of the post-translational prenylation of proteins [9,10]. From cell culture studies, it is known that they inhibit the resorptive activity and induce the apoptosis of mature osteoclasts [10,11].

Moreover, there is now compelling in vitro evidence that bisphosphonates may also act directly on tumor cells. They inhibit proliferation and induce apoptosis in cell lines derived from various neoplasms, such as breast [12,13] and prostate carcinomas [14,15]. Bisphosphonates may also antagonize the growth stimulation induced by bone-derived growth factors on human breast cancer cells [16]. Furthermore, recent animal data indicate that bisphosphonates inhibit bone metastasis growth through promotion of apoptosis in cancer cells [17,18]. Bisphosphonates also reduce tumor cell invasiveness [19] and cell adhesion to bone [20].

In the clinical setting, bisphosphonates are often combined with conventional endocrine agents for the treatment of patients with metastatic bone disease, especially as endocrine therapy is often preferred to chemotherapy for patients with soft tissue and bone metastases [21]. The extent to which such bisphosphonate and antiestrogen combination affects tumor cell growth has not yet been examined, however, and it is unknown which interactions are operating. The triphenylethylene antiestrogen tamoxifen is the classic hormonal treatment for the management of breast cancers expressing ERs [22]. On the other hand, ICI 182,780 [23] (now called fulvestrant or Faslodex™) is the only steroidal antiestrogen that has reached clinical development [24]. Both compounds are competitive inhibitors for the binding of 17β-estradiol (E₂) to ER, but their mechanisms of action are quite different [25]. Tamoxifen, a partial ER antagonist, inhibits the activation function-2 (AF-2)-mediated transactivation, probably via the recruitment of corepressors [26,27]. Yet this type of antagonist does not interfere with AF-1-mediated transactivation. Tamoxifen, as well as its active metabolite 4-hydroxytamoxifen, has also been shown to cause ER nuclear accumulation [28]. By contrast, ICI 182,780, a pure ER antagonist, suppresses both AF-1 and AF-2 ER transactivation functions, and prevents nuclear transport of the receptor [29]. In addition, such pure antagonists reduce the half-life of ER protein, leading to a decrease in receptor content (down-regulation) [30].

In the present study, we assessed the anti-proliferative properties of ibandronate, a newly developed nitrogen-containing bisphosphonate, on ER-positive breast cancer cells. These in vitro experiments were conducted in steroid-free medium (SFM) to allow for the assessment of estrogenic responses and for the measurement of ER content and activity. Besides, it is known that ER antagonists exert a growth-inhibitory effect on MCF-7 cells even in the absence of estrogenic stimulation [31-34]. We thus tested ibandronate in combination with antiestrogens in order to identify possible additive or synergistic interactions.

All experiments were performed in plastic flasks, dishes and multi-well plates obtained from Nunc (Naperville, IL, USA). Cells were cultured at 37°C in a humidified 95% air and 5% CO₂ atmosphere. For routine maintenance, cells were cultured in 75 cm² flasks containing RPMI medium 1640 (Gibco BRL, Life Technologies, Merelbeke, Belgium) with Phenol Red, supplemented with 10% (v/v) heat-inactivated FCS, and containing standard concentrations of L-glutamine, penicillin and streptomycin (Gibco BRL). Cells were harverested by trypsinization (0.05% (w/v) trypsin, 0.53 mM EDTA.4Na) twice a week. For experiments, cells were plated in SFM made up of RPMI medium 1640 without Phenol Red supplemented with 10% (v/v) FCS stripped of endogenous estrogens by a dextran-coated charcoal treatment as previously described [36]. One day later, the seeding medium was replaced by fresh SFM containing ibandronate (gift from Hoffmann-LaRoche (Basel, Switzerland), E₂ (Sigma, St Louis, MO, USA), 4-hydroxytamoxifen (Sigma), ICI 182,780 (Tocris, Bristol, UK) or vehicle for 1 to 6 days.

CI_A+B = [(D_A/A+B)/D_A] + [(D_B/A+B)/D_B] + [α(D_A/A+AB × D_B/A+B)/D_AD_B]

where CI_A+B = CI for a fixed effect (F) for the combination of cytotoxic A and cytotoxic B; D_A/A+B = concentration of cytotoxic A in the combination A + B giving an effect F; D_B/A+B = concentration of cytotoxic B in the combination A + B giving an effect F; D_A = concentration of cytotoxic A alone giving an effect F; D_B = concentration of cytotoxic B alone giving an effect F; α = parameter with value 0 when A and B are mutually exclusive and 1 when A and B are mutually non-exclusive.

The CI indicates synergism for values lower than 0.8, additivity for values included between 0.8 and 1.2, and antagonism for values higher than 1.2. Values of 0.8 and 1.2 suggest slight synergistic and additive cytotoxic activities, respectively.

Figure 1 Dose-response analyses of ibandronate on MCF-7, IBEP-2 and MDA-MB-231 cells. Breast cancer cells were cultured for three days in steroid-free medium containing ibandronate (Iban) or vehicle (control). Cell proliferation was determined by crystal violet (more ...)

Table 1 Cell growth determination

As shown by time-course experiments over 6 days, ibandronate at 10^-5 M only produced a weak inhibition of MCF-7 cell growth, detectable from day 4. By contrast, higher drug concentrations drastically affected cell growth kinetics: 10^-4 M ibandronate induced rapid cytostatic effects up to day 6, whereas 10^-3 M ibandronate clearly exerted cytotoxic effects (Fig. 2).

Figure 2 Time-course experiments of ibandronate (Iban) on MCF-7 cells. Cells were cultured for up to six days in steroid-free medium containing Iban or vehicle (control). Cell growth was determined as described in Fig. 1. Statistical analysis on log transformed (more ...)

Moreover, high ibandronate concentrations induced apoptotic cell death, as documented by the detection of annexin-positive and propidium iodide-negative MCF-7 cells (Fig. 3a). The percentages of annexin-positive cells after 5 days of incubation with 10^-4 M and 10^-3 M ibandronate were 23.5 ± 2.4 (mean ± SD) and 50.0 ± 8.7, respectively, indicating that the inhibition of cell growth was caused, at least in part, by cell apoptosis (Fig. 3b). Of note, 10^-4 M ibandronate did not induce significant apoptosis at day 3. No significant cell death was observed using 10^-5 M bisphosphonate.

Figure 3 Ibandronate-induced apoptosis in MCF-7 cells assessed by annexin-V fluorescein isothiocyanate and propidium iodide double staining (An/PI). Cancer cells were cultured for three to six days in steroid-free medium containing ibandronate (Iban) or vehicle (more ...)

The influence of increasing treatment duration on cell growth estimated at 72 h is illustrated by pulse exposure studies in Fig. 4. A 2 h exposure to ibandronate did not significantly affect cell proliferation, whereas bisphosphonate exposure for 4 h resulted in a significant decrease in cell growth by 21%. In addition, a 45% inhibition of cell growth was already seen after a 24 h treatment and the effect was not significantly greater when the incubation with ibandronate was prolonged up to 48 or 72 h (43% and 53% growth inhibition, respectively). This suggests that the growth inhibition recorded after a 24 h exposure is already irreversible and that 24 to 72 h exposure does not augment inhibitory efficacy.

Figure 4 Effects of various durations of exposure of MCF-7 cells to ibandronate. Cells were cultured in steroid-free medium and exposed to 10-4 M ibandronate or vehicle (control) for 1 to 72 h (pulse exposures). Cell growth measurements were performed after 72 (more ...)

Figure 5 Effects of ibandronate (Iban) on the stimulation of MCF-7 and IBEP-2 cell growth induced by 17β-estradiol (E2). Cells were incubated with 10-4 M Iban and/or 10-8 M E2 or vehicle (control) for 3 days. Cell growth determination was performed after (more ...)

ER expression and activity were examined in the same culture conditions by western blot analysis. Ibandronate alone had no effect on ER content whereas E₂ induced a marked receptor down-regulation reflected by a 63% decline in ER steady-state level after 24 h (Fig. 6a). Ibandronate did not affect ER decrease induced by E₂. The expression of PgR was also investigated as it is estrogen-inducible and viewed as a classic marker of ER activation [40]. Thus, the effect of ibandronate on PgR expression was assessed by western blot analysis using an antibody raised against the A/B isoforms of PgR. Exposure of MCF-7 cells for 72 h to 10^-9 M E₂ increased by more than three-fold the level of PgR B isoform (Fig. 6b). Ibandronate modified neither PgR baseline level nor the increase in PgR level induced by E₂. As previously reported [39], the A isoform of PgR was not detectable in untreated MCF-7 cells and only a small amount of this isoform was observed after E₂ stimulation.

Figure 6 Estrogen receptor (ER) and progesterone receptor (PgR) expression in MCF-7 cells exposed to ibandronate (Iban) and/or 17β-estradiol (E2) as assessed by western blot. (a) For ER determination, MCF-7 cells were incubated for 24 hours with 10-4 (more ...)

Overall, these data indicate that ibandronate completely blocks the proliferative response induced by E₂, without affecting ER regulation and activity. From these results, it can be inferred that ibandronate does not directly act on the ER pathway but interferes with E₂-induced mitogenicity at steps downstream or independent of ER-mediated signaling. Indeed, in MCF-7 cells exposed to ibandronate, ER kept its transactivation ability.

Figure 7 Effect of combined treatment of MCF-7, IBEP-2 and MDA-MB-231 cells with ibandronate (Iban) and 4-hydroxytamoxifen (TAM) or ICI 182,780 (ICI). Cells were incubated with 10-4 M Iban in the presence or absence of 10-7 M TAM or ICI for 72 hours. Cell proliferation (more ...)

To better characterize the interactions between ibandronate and ER antagonists in MCF-7 cells, drug combinations were evaluated over a wide range of concentrations (from 10^-6 to 10^-3 M for ibandronate and from 10^-10 to 10^-7 M for the antiestrogens). Dose-response curves are illustrated in Fig. 8. These data were submitted to isobolographic analysis to calculate the CIs at 50% and 75% inhibition of cell growth, according to the analytical procedure developed by Chou and Talalay [38]. Results are summarized in Table 2. According to the CI values, the effects of ibandronate and antiestrogen were additive, suggesting that these compounds act through distinct mechanisms.

Figure 8 Dose-response curves of ibandronate (Iban from 10-6 to 10-3 M) in combination with increasing concentrations of 4-hydroxytamoxifen (TAM from 10-10 to 10-7 M) or ICI 182,780 (ICI from 10-10 to 10-7 M). MCF-7 cells were incubated for 72 hours. Cell proliferation (more ...)

Table 2 Correspondence of combination index values to combined drug effects in the MCF-7 cell line

Bisphosphonates, including ibandronate, are widely used for the treatment of bone diseases involving enhanced osteoclast-mediated bone resorption, such as osteoporosis, tumor-induced hypercalcemia and, most importantly, cancer-induced bone disease [7,41]. Bisphosphonates thus constitute a major advance in the supportive care of cancer patients who develop skeletal metastases [42]. They exert a clinically significant analgesic activity and reduce by up to 40% the frequency of cancer-induced bone complications in patients with bone metastasis from breast cancer [43]. The beneficial effects of ibandronate, a new and potent bisphosphonate, in patients sufffering from breast cancer-induced osteolysis have recently been reported [44-46]. A substantial morbidity nevertheless persists even with the use of such new potent agents, indicating that bisphosphonates are not fully effective in blocking tumor-induced osteolysis. More studies are needed, notably to better understand the activity of bisphosphonates on tumor cells.

Recent data from in vitro work on breast cancer-derived cell lines indicate that nitrogen-containing bisphosphonates can directly inhibit tumor cell growth, primarily by inducing apoptosis [12,13,47-49]. However, these studies were performed in medium supplemented with whole serum, where the presence of steroids does not allow the evaluation of possible interactions of bisphosphonates with ER-mediated signaling. In the present study, we investigated the effects of ibandronate on the growth of MCF-7, IBEP-2 and MDA-MB-231 cells in medium supplemented with charcoal-stripped serum, providing a steroid-free environment suitable for the assessment of estrogenic responses. In fact, these experimental conditions may reflect more accurately the clinical situation because the vast majority of patients suffering from breast cancer are postmenopausal women with low circulating estrogen levels.

In the first part of this work, we showed that ibandronate (a nitrogen-containing bisphosphonate) induced a dose-dependent decrease in the growth of the three tested breast cancer cells cultured in SFM. These results were consistent with previous reports using steroid-containing medium (complete medium) [12,13,47], but contrast with our recent data showing that clodronate (a non-nitrogen-containing bisphosphonate) can stimulate the proliferation of MCF-7 cells in SFM [37]. In fact, this mitogenic effect of clodronate appears to be mediated by the activation of ERs and is completely suppressed by antiestrogens. Moreover, focusing on MCF-7 cells, we showed that 10^-3 M ibandronate exerted strong cytotoxicity partly through apoptosis induction. A lower concentration (10^-4 M) of ibandronate exerted cytostatic effects associated with moderate apoptosis induction, suggesting that cell proliferation was exactly balanced by cell death. Of note, significant apoptosis was only observed after four days of incubation with 10^-4 M ibandronate. Hence, at day 3, cell growth inhibition could be explained by cell cycle arrest, this step preceding apoptosis induction. Furthermore, in our experimental conditions, short-term exposures (four hours) of cells to 10^-4 M ibandronate were sufficient to induce a significant inhibition of MCF-7 cell proliferation. In addition, the presence of ibandronate for only 24 hours led to an irreversible loss of cell proliferative capacity, as shown by measurement two days later, and growth inhibition was comparable to a full-time exposure. This is consistent with the observations of Jagdev et al. [47] showing that incubation of MCF-7 cells with zoledronic acid, even for a short period of time, results in a significant reduction in cell number and a sizeable increase in apoptosis. These results might be important insofar as serum ibandronate concentrations are maintained for only a few hours after oral bisphosphonate administration [50] and raise the exciting possibility of short-term effects of bisphosphonates in non-osseous sites. On the other hand, in the particular microenvironment of bone metastases, bisphosphonates accumulate at the surface of bone resorption sites. This creates a compartment where cells undergo prolonged exposure to high drug concentrations. In vivo data revealed that effective local concentrations of bisphosphonates at sites of active bone resorption are much higher than serum levels, and may reach up to 10^-3 M in the resorption lacunae [51].

In the second part of our study, we examined the effects of ibandronate on estrogenic stimulation of MCF-7 and IBEP-2 cells. Interestingly, our experiments conducted in SFM showed that, at a concentration affecting cell proliferation (10^-4 M) but not at a lower concentration (10^-6 M), ibandronate suppressed the mitogenic effect induced by E₂ and inhibited cell growth regardless of the presence of estrogen. As the nitrogen-containing bisphosphonate ibandronate is known to act through the inhibition of the mevalonate pathway and subsequent protein prenylation [9], the estrogenic stimulation of cell proliferation might require some prenylated proteins. Indeed, it has recently been reported that prenylated proteins play a role in estradiol-induced stimulation of cell proliferation through activation of the Src/Ras/Erk pathway [52]. Nevertheless, as shown by the current observations using MCF-7 cells, ibandronate had no direct effect on ER expression and did not affect E₂-induced receptor down-regulation. Similarly, ibandronate did not affect the baseline level of progesterone receptor and did not interfere with E₂-induced expression of this receptor. Altogether, these data indicate that ibandronate did not alter the regulation and the activity of ERs in MCF-7 cells, while it could totally prevent estrogen-induced cell proliferation in ER-positive breast cancer cells, suggesting that it acts downstream of ER-mediated gene transactivation.

Cancer therapy mostly relies on the use of drug combinations, which generally improve the therapeutic index, that is, give better responses with less toxicity. In this setting, bisphosphonates are most often used concomitantly with endocrine therapy. The presence of functional ERs in MCF-7 cells exposed to ibandronate suggests that these cells should remain sensitive to antiestrogenic agents. In the third part of our work, we thus tested the effects of ibandronate, antiestrogens (4-hydroxytamoxifen and ICI 182,780) and combinations thereof on MCF-7, IBEP-2 and MDA-MB-231 cell growth. Of note, our data show that antiestrogens alone inhibited the proliferation of MCF-7 and IBEP-2 cells in SFM. Used as a negative control, MDA-MB-231 cell growth was not affected by antiestrogens. These results confirm previous studies showing that ER antagonists are able to reduce the growth of ER-positive breast cancer cells, even in the absence of estrogenic stimulation [53]. It is conceivable that ER in estrogen-deprived MCF-7 cells is activated by phosphorylation due to cross-talk with other signaling pathways [54,55]. Thus, antiestrogens would completely abolish this activation. Moreover, when ER-positive breast cancer cells were exposed to ibandronate and antiestrogens, we observed additive inhibition of cell proliferation. The mechanism by which this occurred has not been fully elucidated but inhibition of the mevalonate pathway and subsequent protein prenylation due to ibandronate treatment, and prevention of cell proliferation by antiestrogens might result in this additive effect. Indeed, nitrogen-containing bisphosphonates are known to act as analogs of isoprenoid diphosphate lipids and to inhibit farnesyl pyrophosphate synthase, an enzyme of the mevalonate pathway [56,57]. This inhibition results in decreased isoprenoid lipid production (farnesyl pyrophosphate and geranylgeranyl pyrophosphate) and prevents protein prenylation [58]. On the other hand, antiestrogens are reported to cause the arrest of MCF-7 cells in the G1 phase of the cell cycle, resulting in a lower proportion of cells in S phase [59,60]. Moreover, cells treated with partial and pure antiestrogens show a significant decrease in cyclin D1 mRNA, which suggests that cyclin may be a target of antiestrogens, thus blocking entry into S phase [61]. Because of these different mechanisms of action, it is not a surprise that we found additive growth inhibitory effects when ibandronate and antiestrogens were added concomitantly. The mechanisms involved in the additive effect of ibandronate and antiestrogens are still elusive, however, and require further investigations.

Interestingly, in accordance with our results, recent data indicate that tamoxifen and the farnesyl transferase inhibitor FTI-277, acting through distinct pathways, exert an additive effect on MCF-7 cells, inhibiting cell cycle progression and cell proliferation [62]. Of note, farnesyl transferase functions closely downstream of farnesyl pyrophosphate synthase (the target for nitrogen-containing bisphosphonates) as it catalyzes the covalent attachment of farnesyl groups to prenylated proteins.

Our results indicate that ibandronate inhibits breast cancer cell growth, both in the presence and absence of estrogenic stimulation. Moreover, the growth inhibition induced by classic antiestrogens, such as 4-hydroxytamoxifen and ICI 182,780, is larger on cancer cells expressing ERs when they are combined with ibandronate. These data suggest for the first time the existence of additive interactions between bisphosphonates and antiestrogens. Thus, our in vitro data provide a rationale for the combined use of ibandronate and antiestrogens in breast cancer patients suffering from bone metastases.

AF = activation function; CI = combination index; E₂ = 17β-estradiol; ER = estrogen receptor; FCS = fetal calf serum; FITC = fluorescein isothiocyanate; PBS = phosphate-buffered saline; PgR = progesterone receptor; SD = standard deviation; SFM = steroid-free medium.

The authors declare that they have no competing interests.

FJ designed the experiments, performed the analysis and interpretation of the data and drafted the manuscript. CC carried out cell culture experiments, cell growth determination and western blot analysis. NM participated in combination index calculation. HD carried out apoptosis determination. GL carried out cell growth experiments and critically revised the manuscript. JJB participated in the design of the experiments, discussed the results and revised the manuscript. All authors read and approved the final manuscript

This study received financial support from F Hoffmann-La Roche AG (Basel, Switzerland), from the 'Fondation Medic', from the Belgian Fund for Medical Scientific Research (grants 3.4563.02 and 3.4512.03), and from 'Les Amis de l'Institut Bordet'. Guy Laurent is Senior Research Associate of the National Fund for Scientific Research. This study was presented in part at The IV^th International Conference on Cancer-Induced Bone Diseases, 7 to 9 December 2003, San Antonio, TX (abstract number 67), and at the workshop 'What is New in Bisphosphonates?', 24 to 26 March 2004, Davos, Switzerland [63].