Results presented in part at the poster plenary session of the American Society for Bone and Mineral Research meeting, St. Louis, Missouri, U.S.A., 1999.
Breast cancer frequently spreads to bone and is almost always associated with osteolysis. This tumor-induced osteolysis is caused by increased osteoclastic bone resorption. Bisphosphonates are used successfully to inhibit bone resorption in tumor bone disease and may prevent development of new osteolytic lesions. The classical view is that bisphosphonates only act on bone cells. We investigated their effects on breast cancer cells using three human cell lines, namely, MCF-7, T47D, and MDA.MB.231, and we tested four structurally different bisphosphonates: clodronate, pamidronate, ibandronate, and zoledronate. We performed time course studies for each bisphosphonate at various concentrations and found that all four compounds induced a nonreversible growth inhibition in both MCF-7 and T47D cell lines in a time- and dose-dependent manner. The MDA.MB.231 cell line was less responsive. Bisphosphonates induced apoptosis in MCF-7 and cell necrosis in T47D cells. The inhibition of MCF-7 cell proliferation could be reverted almost completely by the benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethyl ketone (z-VAD-fmk) inhibitor of caspases, suggesting that the apoptotic process observed in the MCF-7 cell line is mediated, at least partly, by the caspase system. Caspase activity was little changed by bisphosphonates in T47D cells and the inhibitor of caspase did not modify bisphosphonates effects. In summary, we found that bisphosphonates inhibit breast cancer cell growth by inducing cell death in vitro. Such effects could contribute to the beneficial role of bisphosphonates in the treatment and the prevention of tumor-induced osteolysis.
BISPHOSPHONATES ARE synthetic analogs of pyrophosphate in which the oxygen bridge has been replaced by a carbon atom that allows the attachment of various side chains.(1) Some compounds exhibit short side chains such as etidronate or clodronate. The length of side chains can be increased and ornamented with amino groups at their end. Such aminobisphosphonates, including pamidronate and ibandronate, are 100- to 1000-fold more potent inhibitors of bone resorption than clodronate or etidronate in animal models.(2) Zoledronate, a cyclic bisphosphonate of the last generation, contains a nitrogen atom in an imidazole ring and is the most potent compound described so far.(3) Thus, we now dispose of a large panel of structurally distinct bisphosphonates, resulting in different ranges of activities.
Bisphosphonates exhibit a high affinity for calcified matrices, such as hydroxyapatite in bone,(4) and are used successfully as powerful inhibitors of increased bone resorption in several bone diseases including Paget's disease of bone, osteoporosis, and tumor-associated bone diseases.(5–8) This inhibition in bone resorption can be explained by a decrease in osteoclast function because, under bisphosphonate treatment, osteoclasts form fewer and smaller cavities.(9) Bisphosphonates act by inhibiting the recruitment, proliferation, and differentiation of preosteoclasts, by adhesion to the mineralized matrix,(10–12) and, most importantly, by resorptive activity of mature osteoclasts.(13–15) They also shorten the life span of osteoclasts by inducing their programmed cell death (apoptosis).(16) However, in addition to these direct effects on osteoclastic cells, the inhibitory effects of bisphosphonates also may be mediated by other cells, such as cells of the osteoblastic lineage or the macrophage family.(12,17,18)
Breast cancer frequently spreads to bone and is almost always associated with osteolysis. Bisphosphonates constitute the standard treatment for cancer hypercalcemia and a new form of medical therapy for metastatic bone disease.(8) Adjuvant trials have now been started, and results of the first trials indicate that bisphosphonates can indeed prevent the appearance of bone and, surprisingly, of nonbone metastases as well.(19) Because bisphosphonates localize in bone, this decreased incidence of nonbone metastases is surprising. In animal models, bisphosphonates can inhibit the development of bone metastases and reduce the tumor burden in bone when injected before or at the same time as breast cancer cells,(20) suggesting that the production of bone-destroying substances by the cancer cells can set up a vicious cycle that can be interrupted by anti-osteoclastic drugs. A decrease in tumor cell growth in the skeleton could lead to a reduction in the appearance of secondary metastases in other organs. The metastatic cascade is quite a complex phenomenon, but it is known that metastases can themselves metastasize, amplifying disease progression.(21)
In this study, we have investigated the effects of four structurally different bisphosphonates on the proliferation and cell survival of three human breast cancer cell lines in vitro. All four bisphosphonates inhibited cell growth after short-term incubation, resulting from the induction of a combination of cell necrosis and cell apoptosis.
MATERIALS AND METHODS
RPMI-1640 medium with or without phenol red, penicillin/streptomycin (10,000 U/ml and 10,000 μg/ml, respectively), L-glutamine, trypsin-EDTA, fetal calf serum (FCS), and plastic culture materials were obtained from Grand Island Biological Co. (Merelbeke, Belgium). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT), dimethylsulfoxide (DMSO), bovine serum albumin (BSA), calf thymus DNA, Hoechst 33258, acridine orange, and ethidium bromide were obtained from Sigma-Aldrich SA (Bornem, Belgium). The synthetic peptide N-acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin (ac-DEVD-AMC) was obtained from Pharmingen (Becton Dickinson Benelux NV, Erembodegem-Aalst, Belgium), and 7-amino-4-methylcoumarin (AMC) was obtained from Bachem Feinchemikalien AG (Bubendorf, Switzerland). The benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethyl ketone (z-VAD-fmk) was from Enzyme Systems Products (Livermore, CA, U.S.A.) and Annexin/Propidium Iodide (PI) kit detection from Boehringer Ingelheim Bioproducts (Ingelheim, Germany).
We studied the effects of four different bisphosphonates: pamidronate (3-amino-1 hydroxypropylidene bisphosphonic acid) and zoledronate (1-hydroxy-2,1 imidazol-1-yl-ethylidene bisphosphonic acid) were provided by Novartis (Basel, Switzerland), and clodronate (dichloromethylene bisphosphonic acid) and ibandronate (1-hydroxy-3 methylpentylaminopropylidene bisphosphonic acid) were provided by Roche/Boehringer Mannheim (Mannheim, Germany).
A stock solution (10−2 M) of each bisphosphonate was prepared in phosphate-buffered saline (PBS) containing 0.1% BSA. For short-term cultures, serial dilutions (10−3−10−10 M) were prepared in serum and phenol red-free RPMI-1640 containing PBS/BSA at final concentrations of 10%/0.1%. Control medium contained PBS/BSA (10%/0.1%) diluted in RPMI-1640 without phenol red. For long-term cultures, serial dilutions (10−3−10−10 M) were prepared in phenol red-free RPMI-1640 containing 5% FCS. Corresponding control medium also was supplemented with 5% FCS.
Cell line cultures
Human breast cancer cell lines (MCF-7, T47D, and MDA-MB-231) were obtained from the American Type Culture Collection (Rockville, MD, U.S.A.). Cells were cultured routinely in RPMI-1640 medium supplemented with 2 mM L-glutamine, 1% penicillin/streptomycin, and 10% FCS at 37°C in a humidified atmosphere containing 5% CO2 in air. Culture media were changed every 2–3 days.
Effects of bisphosphonates on cell proliferation
Cells were plated at 10,000 cells/cm2 in multiwell plates and cultured for 24 h in RPMI-1640 supplemented with 10% FCS to allow cell adhesion. The next day, the medium was removed and cells were washed once with serum-free medium and then incubated with or without bisphosphonates. Cells were treated under serum-free conditions for up to 48 h. Cell proliferation was evaluated by the MTT test, bromodeoxyuridine (BrdU) incorporation, or by cell counting.
MTT test and BrdU incorporation
Cell viability was measured by the colorimetric microassay described by Mosmann.(22) Ten-microliter tetrazolium salt (MTT, 3 mg/ml) was added to the culture medium for the last 2 h of incubation. Culture media were then removed and cells were lysed in 100 μl DMSO. Microplates were read using a multiwell scanning spectrophotometer (Titertek Elisa reader) at wavelength 540 nm. We used six replicates for each condition and experiments were repeated at least three times. Results are means ± SEM of treated/control ratios. BrdU incorporation was used to confirm the results obtained with the MTT test. MCF-7 cells were cultured for 24 h in the presence of increasing concentrations of bisphosphonates (10−9−10−4 M) and labeled with BrdU for the last 6 h (kit from Amersham, Roosendaal, The Netherlands).
Cells were plated in 48-well plates and incubated with the different bisphosphonates at 10−6 M. At appropriate times, cell layers were rinsed with PBS, harvested by trypsinization, and cell number was determined using a hemocytometer in 4–6 wells per condition in at least two independent cultures. Results were expressed as percentages of control values.
For long-term studies, cells were cultured in 48-well plates and incubated in the presence of each bisphosphonate at 10−4, 10−6, or 10−8 M for up to 6 days. Total DNA content was evaluated at different time points by fluorometry using Hoechst 33258 reagent, as previously described.(23) Briefly, supernatants were removed and cell layers were washed in PBS and lysed in 10 mM EDTA (pH 12.3) for 30 minutes at 37°C. Cell lysates were then stored at 4°C. Aliquots of cell lysates were neutralized by addition of 1 M KH2PO4 (10% vol/vol) before incubation for 1 h at 4°C with 20 ng/ml Hoechst 33258 reagent (vol/vol). Fluorescence was measured using a spectrofluorometer at excitation and at emission wavelengths 350 nm and 455 nm, respectively. Calf thymus DNA was used as standard. We used tri- or tetrareplicates for each condition and each experiment was repeated at least twice. Results are shown for one representative experiment.
Reversibility of bisphosphonates effects
The possible reversibility of the effects of bisphosphonates on breast cancer cells was tested by incubating the cells for 24 h in the presence of 10−6 M bisphosphonate. Cells were then rinsed and further cultured without bisphosphonate for 48 h. In parallel, cells also were cultured for the full 72 h in the presence of 10−6 M bisphosphonate. All media were changed at day 1 for the switch to bisphosphonate-free medium or for the continuous incubation.
Apoptotic cell death detection
To assess the effects of bisphosphonates on the process of apoptosis, breast cancer cells were incubated for 24 h with or without 10−6 M bisphosphonates.
Acridine orange/ethidium bromide staining
A double staining with ethidium bromide and acridine orange(24) was performed to visualize and quantify the number of viable cells (green nuclei), apoptotic cells (nuclei condensed and colored in orange), and necrotic cells (red nuclei). Briefly, 2 μl of the dye mixture (100 μg/ml acridine orange and 100 μg/ml ethidium bromide) was added to 20 μl of cell suspension and immediately examined with a 40× oil immersion objective using a Leitz DMRB fluorescence microscope (Green/Red filter, 100 Watt lamp) equipped with a photometrics CCD camera and the Logikon image analysis system (Numeris Benelux SA, Ath, Belgium). Several fields, randomly chosen, were digitalized and at least 600–800 nuclei for each sample were counted and scored. Results are expressed as the relative percentages of viable, apoptotic, and necrotic cells to the total number of cells scored.
Cell death also was assessed by the annexin V/propidium iodide (SUPPRESS) double staining assay, according to the manufacturer's recommendations. Briefly, cells were incubated with or without bisphosphonates (10−6 M) for 6–48 h, washed once in PBS, and then incubated with 195 μl binding buffer containing 5 μl of fluorescein isothiocyanate (FITC)-labeled annexin V for 10 minutes at room temperature. Cells were washed once in PBS, and incubated in 190 μl of binding buffer containing 10 μl of PI. Cells were then detached from culture dishes in PBS/1% gelatin and analyzed by fluorocytometry. More than 5000 cells were scored for each condition. Experiments were repeated at least three times. Results are expressed as ratios of apoptotic (annexin+, PI−) or necrotic (secondary to apoptosis; annexin+, PI+) cells to untreated cells.
Measurement of caspase protease activity
Interleukin-1β-converting enzymelike proteases are crucial components of cell death pathways. The activity of this family of cysteine proteases (caspases) was assessed by the cleavage of SUPPRESS, a synthetic fluorogenic substrate containing the amino acid sequence recognized by caspase (DEVD, Asp-Glu-Val-Asp) combined to a fluorophore (AMC, 7-amino-4-methylcoumarin). On cleavage of the substrate by caspase, free AMC fluorescence emission was detected using a spectrofluorometer.
Cells were plated at 500,000 cells/well in 6-well plates and allowed to attach for 24 h in RPMI supplemented with 10% FCS. Cultures were then washed once in PBS and incubated in serum free-RPMI with or without bisphosphonates (10−6 M). After 3, 8, 15, 24, or 48 h of incubation, cells were trypsinized and counted. Cell pellets were then lysed in 400 μl lysis buffer (10 mM Tris, pH 7.4, 200 mM NaCl, 5 mM EDTA, 10% glycerol, and 1% NP-40) for 30 minutes on ice and stored at −20°C. For the assay, aliquots of 200 μl were incubated for 2 h at 37°C with 400 μl reaction buffer (0.1 mM phenylmethylsulfonyl fluoride [PMSF], 10 mM dithiothreitol [DTT], and 10 mM HEPES/NaOH, pH7.4) containing 5 μl Ac-DEVD-AMC substrate (20 μM final concentration). The fluorescence released in samples was measured by exciting at 367 nm and reading at 440 nm. The negative control was buffer mixed and the positive control was free AMC (10 μM in PBS). Results are expressed as nanomoles of AMC/h per 106 cells.
Inhibition of caspase activity
To confirm the implication of caspase activity in the inhibitory effects of bisphosphonates on cell viability, MCF-7 and T47D cell lines were treated for 24 h with both caspase inhibitor peptide (z-VAD-fmk) at a final concentration of 20 μM and bisphosphonates (10−6 M). Cell survival was evaluated using the MTT test as described previously.
Results are expressed as means ± SEM and have been analyzed using two-tailed Student's t-test. Minimal statistical significance was described as p < 0.05.
Effects of bisphosphonates on breast cancer cell viability in short-term culture
MCF-7 cell line
The effects of four bisphosphonates, at doses ranging from 10−9 to 10−4 M, were evaluated on MCF-7 cell viability using the MTT test (Fig. 1) and confirmed by cell counting. All four bisphosphonates decreased MCF-7 cell viability. High doses of clodronate (10−6−10−5 M) induced an inhibition of about 10% after 6 h incubation. A stronger decrease was observed at 10−4 M (−28%, p < 0.0001) and was still detected thereafter (−23%, p < 0.04 at 24 h). After 24 h, an inhibitory effect was already present at 10−8 M. Pamidronate also induced a decrease in cell viability after 24 h at concentrations above 10−9 M but the effects were somewhat more marked (−13% to −23%, p < 0.001). High doses (above 10−6 M) were already inhibitory after 6 h. Ibandronate exerted a similar dose-dependent inhibitory effect on MCF-7 cell viability after 6 h and 24 h incubation from 10−9 to 10−4 M (−12% to −22%, p < 0.002). For zoledronate, we observed a strong inhibition after only 2 h of incubation with doses ≥ 10−8 M (−13% to −16%, p < 0.02, data not shown). This effect was maintained at 6 h and 24 h (Fig. 1). These results were confirmed by measuring BrdU incorporation (data not shown).
Cell counts were performed on cells plated at the same density and incubated with 10−6 M bisphosphonates for 24 h (Table 1). We found similar results to the MTT test because pamidronate, ibandronate, and zoledronate decreased cell number by about 25% but the effects of clodronate were not significant.
Table Table 1.. Short-Term Effects of Bisphosphonates on Breast Cancer Cell Number
T47D cell line
Figure 2 shows the effects of bisphosphonates on T47D cell viability in short-term culture (6-24 h). We found that all four bisphosphonates inhibited cell viability, but at various degrees. Clodronate exerted weak inhibitory effects after 24 h at the doses of 10−7 M and 10−5 M (about −10%, p < 0.05). The highest dose of clodronate induced a strong inhibition already after 6 h of incubation (−28%, p < 0.0001). Pamidronate exerted its inhibitory effects after 24 h at doses of 10−8 M and above (−13%, p < 0.005). Ibandronate induced the strongest dose-dependent inhibition of cell viability even after only 6 h incubation (up to 40%, p < 0.0001) and zoledronate was again the most rapid compound to inhibit cell viability (−11% after 6 h, p < 0.03), but the effects were less marked than for ibandronate after 6 h and 24 h of incubation. For all four bisphosphonates the highest dose tested (10−4 M) was toxic for cell viability because cells became granular and detached from the plastic culture dish.
Cell countings performed after 24 h of incubation in the presence of 10−6 M bisphosphonates again confirmed the MTT test (Table 1). Clodronate exerted a nonsignificant effect and pamidronate and zoledronate induced a 20% reduction in T47D cell number (p < 0.01), whereas ibandronate had the strongest inhibitory effect with a 36% decrease in cell number (p < 0.0002).
MDA.MB.231 cell line
This cell line was less responsive to bisphosphonate inhibitory effects than the two other cell lines. Inhibitory effects were quite small (4-9%, p < 0.01; data not shown) even in the presence of high doses of bisphosphonates (10−4 M). This weak effect was confirmed by evaluation of cell number (Table 1).
Effects of bisphosphonates on breast cancer cell proliferation in long-term culture
MCF-7 cell line
The effects of bisphosphonates on the MCF-7 cell line were then evaluated in long-term culture (1-6 days) in the presence of 5% FCS. In control conditions, MCF-7 cells exhibit a high proliferation rate (30-fold increase in total DNA content between days 2 and 6; Fig. 3) and no significant effect of bisphosphonates could be detected before day 4. The highest dose (10−4 M) of pamidronate, ibandronate, or zoledronate was toxic for cell survival because cell proliferation was blocked completely. Ibandronate was the most potent compound at 10−6 M. The lowest dose of bisphosphonate (10−8 M) inhibited cell growth by about one-half after 6 days of incubation for pamidronate, clodronate, and ibandronate, except for zoledronate.
T47D cell line
For the T47D cell line, we also observed a high proliferation rate in the presence of 5% FCS (9-fold increase between days 1 and 6). The highest dose tested (10−4 M) diminished cell growth significantly after only 2 days incubation and induced between 45% and 56% inhibition (p < 0.002; Table 2) for ibandronate and pamidronate, respectively, and a maximum of 70% inhibition (p < 0.0003) for zoledronate. Clodronate was again the least potent compound, inducing only a 35% reduction in cell proliferation after 6 days (p < 0.01; Table 2). The lowest concentration (10−8 M) of pamidronate, ibandronate, or zoledronate induced a significant inhibition of cell growth after 6 days of incubation (−25% to −35%, p < 0.01; Table 2). Finally, the intermediate concentration (10−6 M) induced weak and nonsignificant (5-15%) reduction in cell growth up to 4 days of incubation (data not shown) and between 20% and 30% inhibition (p < 0.02; Table 2) after 6 days.
Table Table 2.. Long-Term Effects of Bisphosphonates on T47D Cell Growth
MDA.MB.231 cell line
MDA.MB.231 cells proliferate markedly in the presence of 5% FCS because we observed an increase of about 10-fold in total DNA content between days 1 and 6 of culture (data not shown). After 6 days of incubation, doses of 10−8 M bisphosphonates reduced cell growth by 18% for zoledronate (p < 0.01) and by up to 30% for ibandronate (p < 0.02). The highest dose (10−4 M) of pamidronate or zoledronate induced an almost complete inhibition of cell growth (−95%, p < 0.0001). Clodronate was again the least potent compound because 10−4 M reduced cell proliferation by only 35% (p < 0.005). The dose of 10−6 M induced intermediate effects because cell growth was reduced by 23% (p < 0.03) for clodronate and up to 45% (p < 0.04) for zoledronate.
Mechanisms of antiproliferative effects of bisphosphonates
Irreversibility of bisphosphonate effects
To detect a possible reversibility of the inhibitory effects of bisphosphonates on breast cancer cell proliferation, cells were pulsed for 24 h with 10−6 M bisphosphonates and then chased with bisphosphonate-free medium for an additional 48 h. Parallel cultures were treated for the full 72 h with 10−6 M bisphosphonates. Cell growth was evaluated using the MTT test. The inhibition of cell proliferation was not reversed at all after removal of bisphosphonates in the MCF-7 or in the T47D cell line (data not shown).
Changes in nuclei morphology after bisphosphonate treatment
A double staining with a mixture of ethidium bromide and acridine orange was used to visualize and quantify the number of viable, necrotic, and apoptotic cells. Viable cells exhibit large green nuclei whereas apoptotic cells show signs of nuclear condensation or nuclear bead formation and are colored in orange. Necrotic cells have red nuclei without signs of nuclear condensation (Fig. 4).
We found that all bisphosphonates used at the intermediate concentration of 10−6 M decreased the percentages of viable cells in MCF-7 or T47D cell lines (Tables 3 and 4, respectively). The percentage of apoptotic cells observed in the MCF-7 cell line was increased to 12% in ibandronate-treated cells compared with 4% in control conditions. The percentage of necrotic cells was not much changed (Table 3). However, in the T47D cell line, both apoptotic and necrotic cells percentages were increased after 24-h treatment but the percentage of apoptotic cells was low. The MDA.MB.231 cell line was again less sensitive to bisphosphonates with no marked changes in the percentages of viable cells on treatment (data not shown).
Table Table 3.. Breast Cancer Cell Death Induction by Bisphosphonates
Table Table 4.. Breast Cancer Cell Death Induction by Bisphosphonates
Kinetic evolution of annexin/PI staining
Because bisphosphonates induced MCF-7 cell death by apoptosis and T47D cell death rather by cell necrosis, we further investigated the effects of bisphosphonates on the MCF-7 cell line. Phosphatidylserine exposure serves as a sensitive marker for early stages of apoptosis and can be detected by annexin V. A parallel staining with PI indicates if the membranes are still intact (no staining) or perforated as observed in the necrotic process. Figures 5A and 5B indicate the relative percentages of annexin+/PI− MCF-7 cells and annexin+/PI+ apoptotic MCF-7 cells, evaluated after different time points during incubation with 10−6 M bisphosphonates. The maximal induction of apoptosis was found after 24-h incubation for all four bisphosphonates (Fig. 5A) and then decreased to be replaced by cell necrosis after 24–48 h (Fig. 5B). Ibandronate and zoledronate were more potent bisphosphonates for the induction of apoptosis (up to 6-fold increase vs. control MCF-7 cells).
Caspase protease activity
A synthetic fluorescent substrate specific for the caspase-3 subfamily was used to detect the enzymatic activity in breast cancer cells treated with bisphosphonates. We found that caspase activity was enhanced in the MCF-7 cell line by all four bisphosphonates after variable durations of incubation (Fig. 6A). Zoledronate again increased caspase activity most rapidly (2-fold increase after 3 h). Pamidronate and ibandronate increased caspase activity by about 2-fold after 8 h or 15 h, respectively. As for the effects on cell growth, clodronate was the least potent compound (1.7-fold increase after 8 h). The stimulatory effects were transient because the enzymatic activity almost returned to basal levels after 48 h.
In the T47D cell line, caspase activity was enhanced little by bisphosphonates (data not shown). Pamidronate and clodronate stimulated caspase activity by a maximum of 25% or 34% after 24 h, respectively. Ibandronate increased caspase activity by 20% after 15 h but zoledronate was again the most rapidly active compound by inducing a 30% increase after 3 h.
To confirm the implication of caspases in the inhibition of cell growth, an inhibitor of caspase activity (z-VAD-fmk) was added to culture media containing 10−6 M bisphosphonates. z-VAD-fmk (20 μM) alone did not exert any significant effects on cell proliferation (Fig. 7). The inhibition of MCF-7 cell survival induced by bisphosphonates could be suppressed by coaddition of z-VAD-fmk (Fig. 7). This compound completely reverted the inhibitory effects of each of the four bisphosphonates on MCF-7 cells. In line with the previous results, the inhibitory effects of bisphosphonates on cell growth were not reversed by z-VAD-fmk in the T47D cell line (data not shown).
Bisphosphonates act by inhibiting osteoclast-mediated bone resorption but some recent data suggest that they also could influence tumor cell growth and/or survival. Several bisphosphonates have been reported to inhibit in vitro human myeloma cell proliferation by inducing their apoptosis(25,26) and preliminary data suggest that this could occur in vivo as well.(27) For breast cancer cells, the only published data concern the inhibition by bisphosphonates of adhesion of MDA.MD.231 cells on bone matrix.(28,29) Here, we report the antiproliferative activities of four structurally different bisphosphonates using established breast cancer cell lines, namely, MCF-7, T47D, and MDA.MB.231.
In clinical practice, these structurally different compounds exhibit different potencies. For example, for the treatment of tumor-induced hypercalcemia, clodronate is usually administered at the dose of 1500 mg, pamidronate at 90 mg, ibandronate at 4–6 mg, and zoledronate at 2–3 mg.(3,8,30) In our in vitro study, clodronate also was the least potent compound because it was, for example, the only one unable to block the cell growth of either of the three cell lines. In contrast, ibandronate and zoledronate were able to reduce cell survival after only 6 h of treatment by about 10–20%, and to almost completely block cell growth at the dose of 10−4 M after 4 days. We can conclude from the growth curves (Figs. 1–2) that the relative scale of bisphosphonate potency in vivo was almost respected with clodronate ≤ pamidronate < ibandronate ≈ zoledronate. Interestingly, zoledronate was always the fastest compound to inhibit cancer cell growth rate (at 3 h), but ibandronate was more potent after 24 h and in long-term cultures.
The inhibitory effects of bisphosphonates on breast cancer cell survival were detected either in the presence or in the absence of FCS and thus they cannot be attributed to the lack of specific factors present in FCS. Removal of bisphosphonates after only 24 h of exposure did not allow the return to basal proliferation rate. This nonreversibility suggested that the growth inhibitory effects of bisphosphonates were caused by induction of cell death rather than by a cell cycle arrest. Thus, our first conclusion is that bisphosphonates can reduce breast cancer cell growth in vitro, resulting from the induction of cell death.
A cytochemical staining of nuclei allowed us to confirm the induction of cell death by bisphosphonates in agreement with the variations observed in cell survival (MTT test, cell counting, or DNA content). Thus, ibandronate and zoledronate, which induced the highest or earliest inhibitory effects on cell growth, also were the most efficient compounds to decrease the number of viable cells in MCF-7 and T47D cell lines. T47D cells exhibited red nuclei without signs of condensation or blebbing, suggesting that reduced cell growth was caused by a direct cell necrosis. On the contrary, MCF-7 cells exhibited clear signs of apoptosis. We performed flow-cytometric analyses (annexin V/PI) to confirm and quantify the induction of apoptosis. Maximal apoptosis was reached after 24 h of incubation and, as expected, was followed by necrosis (after 24–48 h of incubation). Ibandronate and zoledronate were again the most potent compounds by inducing a 4- to 5-fold increase in the number of apoptotic MCF-7 cells. Our second conclusion is that even if MCF-7 and T47D exhibited about the same proportion of dead cells (20-30%), MCF-7 cells preferentially died by induction of apoptosis whereas T47D cells rather exhibited primary cell necrosis in addition to a weak apoptotic process. The very low percentages of apoptotic and necrotic cells observed in the MDA.MB.231 cell line were in perfect agreement with the negligible inhibitory effects that we observed on cell growth.
To further characterize the induction of apoptosis in breast cancer cells under bisphosphonate treatment, we investigated the variations in caspase activity. We found that all four bisphosphonates induced a progressive cleavage of a fluorogenic substrate preferentially recognized by the caspase-3 subfamily of caspases.(31) However, the MCF-7 cell line is defective for the caspase-3 protein(32) (personally verified by Western blot), which suggests that the observed increases in caspase activity could be caused by a caspase-3-like activity or a DEVD-sensitive caspase other than CPP-32. This (these) caspase-3-like protease(s) was (were) activated at different time points according to the bisphosphonate used, in agreement with the inhibitory effects on cell growth.
These increases in caspase activity were relatively small but they represent the activity calculated for the whole cell population. However, our first set of data indicates that only 10–12% of the cell population enters the apoptotic process at the bisphosphonate concentration tested (10−6 M). Thus, a 2-fold increase in total caspase activity may correspond to a 10- to 20-fold increase in the concerned cell population.
Bisphosphonate-induced inhibition of cell proliferation in the MCF-7 cell line was blocked by cotreatment with an inhibitor of caspases. This compound almost completely reverted the inhibition of cell survival induced by each of the four bisphosphonates we tested, confirming the involvement of the caspase enzyme family in MCF-7 cell death. Further experiments are necessary to explain why only part of the cell population was sensitive to bisphosphonates at the concentration of 10−6 M.
The apoptotic process induced in MCF-7 cells resulting from the induction of caspase-3-like protease(s) seemed to be independent from the bisphosphonate structure. At first sight, this is surprising because it has been reported that clodronate acts by a mechanism different from the aminobisphosphonates. This compound can be metabolized to a nonhydrolizable adenosine triphosphate (ATP) analog and causes toxic effects with both necrotic and apoptotic cell death in mammalian cells.(33,34) In contrast, aminobisphosphonates, which are not metabolized, essentially appear to act through inhibition of enzymes of the mevalonate pathway,(35) which also will induce cell death. However, Benford and colleagues(36) reported very recently that both aminobisphosphonates and nonaminobisphosphonates cause activation of caspase-3-like proteases in J774 cells, whenever they act through inhibition of protein isoprenylation or through intracellular accumulation of a cytotoxic analog of ATP.
In contrast to the MCF-7 cell line, the T47D cell line incubated in presence of bisphosphonates underwent primary cell necrosis. In addition to the low percentage of apoptotic cells scored by BET/acridine orange, no marked variation in caspase activity was detected throughout the 48-h study. Furthermore, the addition of the general caspase inhibitor z-VAD-fmk did not revert the inhibition of cell growth at all. Taken together, these data confirmed that an apoptotic phenomenon was not primarily responsible for bisphosphonate inhibitory effects on T47D cell growth.
Three of the four bisphosphonates we tested were already shown to be able to inhibit the proliferation and to induce apoptosis in a human myeloma cell line.(25,26) However, the bisphosphonate concentrations used in these experiments (≥10−4 M) represent quite high concentrations compared with those able to affect significantly breast cancer cell growth in our system. This could indicate that bisphosphonates act differently on tumor cells according to cell origin. The concentrations active in our experimental model cannot be reached in the circulation but will be reached in the resorption lacunae where bisphosphonate concentrations are around 10−4−10−3M.(37) Thus, it is possible that breast cancer cells can be exposed to concentrations of 10−6 M when they are in bone.
In summary, we found that four structurally different bisphosphonates reduced breast cancer cell growth in vitro by inducing cell death. The three breast cancer cell lines tested reacted differently to bisphosphonate treatment. The MDA.MB.231 cell line was not very sensitive, whereas MCF-7 and T47D cells exhibited a high degree of inhibition of cell survival. However, these two cell lines also differed in the cell death pathway. MCF-7 cells died primarily by apoptosis and T47D cells died primarily by necrosis. The amplitude of the inhibitory effects as well as the delay before the observation of a significant effect was dependent on bisphosphonate structure. Clodronate was the least potent whereas ibandronate and zoledronate were the most potent compounds in our in vitro system.
Our present data give further support to the theory that bisphosphonates can directly affect human breast cancer cells. It has already been shown that bisphosphonates can inhibit the adhesion of tumor cells (breast and prostate carcinoma cells) on extracellular bone matrix in vitro(28) as well as the invasive potential of breast cancer cells.(29) Here, we report that bisphosphonates can inhibit breast cancer cell growth by inducing apoptosis or primary necrosis depending on the cell type and the bisphosphonate concentration. The clinical relevance of in vitro observations remains to be shown but a direct inhibitory effect of bisphosphonates on cancer cell growth evidently could contribute to their beneficial effects in cancer patients.
Mrs. N. Kheddoumi is acknowledged for her technical assistance. This study has been supported by grants from Foundation Medic, Fonds de la Recherche Scientifique Médicale (contract 3.4577.96), Fonds National de la Recherche Scientifique (contracts 9.4503.98 and 7.4547.99), Foundation Lambeau-Marteau, and F. Hoffmann (La Roche Ltd, Switzerland).