of TRPM-2 and TGF01 mRNAs were observed in in vitro or in vivo grown tumor cells treated with 5-10 \i.M toremifene. Elevated levels of. TRPM-2, but not TGF01 ...
Apoptosis in ToremifeneInduced Growth Inhibition of Human Breast Cancer Cells In Vivo and In Vitro Anni M. Warri, Riikka L. Huovinen, Aire M. Laine, Paula M. Martikainen, Pirkko L. Hdrkonen*
1412
REPORTS
Antiestrogens block the stimulatory effects of estrogens on breast cancer growth. They function by competing with estrogen for binding of estrogen receptors, subsequently inhibiting or modifying the interaction of estrogen receptors with DNA. Antiestrogens have been reported to counteract estrogen effects, e.g., secretion of several growth factors and growth controlling enzymes [reviewed in (7)] and entering of cells to S phase and G2/M phases of cell cycle (2-5). Moreover, it has been suggested that antiestrogens are also able to inhibit growth by additional mechanisms, such as stimulation of tumor growth factor beta-1 (TGFpi) secretion (6). The mechanisms by which anticancer compounds induce tumor regression are largely unknown and can involve both
an enhanced cell death and arrested cell proliferation. Cell death is caused either by necrosis or by an active process as a response to a specific stimulus (or lack of the stimulus) that leads to elimination of a cell population. This process of programmed cell death, called apoptosis, is characterized by a marked reduction in cell volume, nuclear condensation, cleavage of nuclear DNA to oligonucleosomal length fragments, and an increase in buoyant density of the cells (7-13). Necrosis, in contrast, is characterized by clumping of chromatin into ill-defined masses. DNA is randomly degraded and appears as a smear when analyzed in agarose gel. Gross swelling of organelles and formation of large vacuoles and, finally, loss of membrane integrity are the main features of necrosis. Several anticancer compounds, as well as other external stimuli, have been reported to induce apoptotic changes in different normal and malignant cells (11,12). It is still unknown if all cells are able to undergo apoptosis under certain circumstances or if it is characteristic of only the cells of some tissues that are regulated by specific hormone(s) or growth factor(s) (9,11). Elevated expression of the TGF(3 1 gene (17) and also a regression-specific gene, testosterone-repressed prostatic message-2 (TRPM-2) (14-16), have been connected to apoptosis. We have studied the mechanisms by which antiestrogens induce tumor regression by studying the effect of a new antiestrogen, toremifene, on two steroid-sensitive human breast cancer cell lines. The effect of estrogen withdrawal was determined in parallel to evaluate the contribution of the antihormone effect.
Materials and Methods Cell Lines and Tissue Culture Estrogen-sensitive MCF-7 and ZR-75-1 human breast tumor cell lines were given by Dr. P. Darbre, Imperial Cancer Research Fund (London, England). MCF-7 cells originate from the laboratory of Dr. C. K. Osborne (University of Texas Health Science Center, San Antonio) (.IS). All
*See "Notes" section following "References."
Journal of the National Cancer Institute, Vol. 85, No. 17, September 1, 1993
Downloaded from jnci.oxfordjournals.org at Georgetown University Library on June 9, 2011
Background: Antiestrogens inhibit the stimulative effects of estrogens on breast cancer growth, but the mechanism(s) by which they trigger tumor regression are not completely understood. Growth retardation and tumor regression can be achieved by enhanced cell death and/or arrested cell proliferation. Purpose: Our aim was to investigate the effect of a new antiestrogen, toremifene, on human breast cancer cells grown either in culture or as tumors in nude mice. Methods: The growth and morphology of in vitro cultured cells of the human breast cancer cell line MCF-7 were monitored by time-lapse video. MCF-7 cells and ZR-75-1 human breast cancer cells were grown as tumors in nude mice and subsequently examined by electron microscopy. The integrity of DNA isolated from these cells was determined by standard gel electrophoretic techniques. Northern blot hybridization analysis was used to determine the steady-state levels of the mRNAs for testosterone-repressed prostatic message-2 (TRPM-2), tumor growth factor beta-1 (TGF31), and pS2 (a small, cysteine-rich protein of unknown function). Results: Time-lapse video microscopy of the cell cultures indicated that treatment with 7.5 \LM toremifene for 3 days caused approximately 60% of the cells to exhibit morphologic characteristics typical of cells undergoing programmed death, or apoptosis. The number of mitoses gradually decreased to zero over a 3to 4-day period. Estrogen withdrawal for the same length of time resulted in an approximately equal number of
apoptoses and mitoses. These changes were not associated with the pattern of DNA fragmentation, detectable as ladders in agarose gels, that is characteristic of the DNA of cells undergoing apoptosis. Elevated levels of TRPM-2 and TGF01 mRNAs were observed in in vitro or in vivo grown tumor cells treated with 5-10 \i.M toremifene. Elevated levels of TRPM-2, but not TGF01, mRNA were observed in the tumor cells after estrogen withdrawal. The steady-state level of pS2 mRNA in the tumor cells dropped in response to either toremifene treatment or estrogen withdrawal. Conclusion: Toremifene causes growth inhibition of estrogen-sensitive breast cancer cells by inducing some cells to undergo apoptosis and by inhibiting other cells from entering mitosis. The higher than normal amounts of TRPM-2 and TGFpi protein that would likely result from the elevated levels of TRPM-2 and TGFpl mRNAs measured in these cells after toremifene treatment may have an important role in the growth inhibition process. Implication: Apoptosis as an active, targeted process provides a potential new therapeutic approach for treating breast cancer. [J Natl Cancer Inst 85:1412-1418, 1993]
the cells were cultured as monolayers in plastic tissue culture dishes (Nunc, Roskilde, Denmark) in RPMI-1640 culture medium without phenol red, supplemented with 5% heat-inactivated fetal calf serum or 2X dextran-charcoal stripped fetal calf serum (iFCS and DC-FCS, respectively), 2 mM L-glutamine, 1 nM 17B-estradiol (E,), and 10 u.g/mL insulin. The cells were cultured in a humidified atmosphere of 95% air and 5% CO 2 at 37 °C. The RPMI-1640 culture medium, FCS, and L-glutamine were purchased from G1BCO BRL (Paisley, Scotland); E2 and insulin were from Sigma Chemical Co. (St. Louis, Mo.). E 2 and/or the antiestrogens toremifene or tamoxifen (obtained from Orion Corporation Medipolar, Oulu, Finland) were dissolved in 70% ethanol and added to the culture medium as indicated in the "Results" section. The cell numbers were determined as described earlier (19) by counting the cell nuclei in a Coulter counter (Coulter Corp., Harpenden, England).
Time-Lapse Video Microscopy For the experiments, 2 X 105 MCF-7 cells were plated in 5 mL hormone-free (E-) medium in 25-cm2 flasks (Nunc). One day after plating (day 1), the medium was changed and 7.5 u.M toremifene ± I nM E, was added. Before videotaping, the medium was equilibrated with 5% CO 2 at 37 °C in an incubator for 10 minutes. The flask was then capped and transferred to a 37 "C heated stage of an inverted microscope (Nikon Diaphot, Nikon Corp., Tokyo, Japan). Single cells were viewed using phase-contrast optics at 20X objective magnification coupled to a JVC 3CCD KY-F30 video camera (Victor Company, Tokyo, Japan). The field to be analyzed was selected so that, in the beginning of filming, it contained approximately 60 cells. Recording was performed so that five successive pictures during 5 seconds were taken at I-minute intervals. Filming was continued for 4-5 days starting either on day I immediately after changing the experimental medium (toremifene treatment without E2) or on day 3 (toremifene + E 2 and the control E- experiment). After these time points, the increase in cell numbers slowed. In every case, the media were changed again on day 3, and the culture was similarly equilibrated with CO 2 before videotaping was continued. The change in pH of the medium with time was not significant. The film was viewed on a video monitor. The cumulative number of mitoses and apoptoses per field was counted at 24-hour intervals.
Animal Experiments Nude athymic mice (nu/nu-BALB/cABom) were purchased from Bomholtgard (Rye, Den-
ethanol (20) or by using a DNA isolation kit (ImBedKit; New England Biolabs, Beverly, Mass.), which involves embedding the DNA in agarose according to the manufacturer's protocol. Aliquots of DNA (10 u.g), prepared by both methods, were analyzed by electrophoresis on 1.8% agarose gels. The gels were run 17 hours at 35 V, then stained with ethidium bromide, and photographed under UV illumination. Fragments of lambda DNA digested with EcoRI and tfmdIII or pBR322 DNA digested with Hae III were run as molecular weight standards on all gels.
Northern Blot Analysis
For the electron microscopic studies, tumors were removed from nude mice and fixed with cacodylate-buffered glutaraldehyde, washed in the same buffer, dehydrated and postfixed with osmium tetroxide, and embedded in Epon (EPON 812; Merck and Co., Inc., Darmstadt, Federal Republic of Germany). Semilhin 1-u.m sections were stained with toluidine blue for light microscopy, and representative areas of the tumors were selected for electron microscopic analysis. Electron microscopic 60-nm sections were stained with uranyl acetate and lead citrate. Electron micrographs were taken with a JeolJEM I00SX transmission electron microscope (Japan Electronic Optical Ltd., Tokyo, Japan).
Total RNA was extracted from the cells and tumors into a solution of guanidine isothiocyanate and purified by ultracentrifugation in a CsCl 2 gradient as described earlier (2/). Ten micrograms of total RNA was run on denaturing 1% agarose-formaldehyde gels (20), stained with ethidium bromide, photographed under UV light, and transferred onto GeneScreen Plus nylon membrane (Du Pont NEN Research Products, Boston, Mass.) according to the manufacturer's instructions. The filters were hybridized and washed as suggested by the manufacturer. The following probes were used: a 1.7-kilobase (kb) insert of pG17H complementary DNA (cDNA) clone of rat TRPM-2 (22) (a gift of Dr. M. Tenniswood, Department of Biochemistry, University of Ottawa, Ontario, Canada); 350-base pair (bp) and 215-bp Psl I fragments of pS2 cDNA clone (23,24) provided by Dr. P. Chambon (INSERM, Strasbourg, France); and an £coRI insert of human TGFB1 cDNA (25) obtained from Genentech, Inc. (South San Francisco, Calif). The inserts were [32P]dCTP labeled by random oligonucleotide primer extension (26) (random primer labeling kit; Boehringer Mannheim Corp., Indianapolis, Ind.) to a specific activity of approximately 109 cpm/u.g DNA. After hybridization and washing, the filters were exposed to autoradiographic films (X-Omat; Eastman Kodak Co., Rochester, N.Y.) at -80 °C. For quantitation, the signal intensities of the autoradiographic films and the ribosomal RNAs (rRNAs) in the photographs of the corresponding ethidium bromide-stained gels were scanned by Ultroscan laser densitometer by using the GelScanXL software program (Pharmacia, Uppsala, Sweden). The hybridization intensities were then normalized against the amount of rRNA detected by ethidium bromide fluorescence in the corresponding samples (27).
Analysis of DNA Fragmentation
Results
The MCF-7 cells were cultured as described above, except that unattached cells were collected by centrifugation of the culture media every 24 hours, pooled together, and analyzed separately from the cells growing as monolayers. The cells were plated in E2-free medium. On the day following plating (day I), the media were changed and E2 (1 nM) or toremifene was added. The cells were cultured for 7 days. DNA was extracted from the cells as well as from the nude mouse tumors by treating the cells and tumor samples with proteinase K, extracting with phenol and chloroform, and precipitating with
Effect of Toremifene on Growth of MCF-7 Cells In Vitro
Transmission Electron Microscopic Analysis
Journal of the National Cancer Institute, Vol. 85, No. 17, September 1, 1993
Toremifene had a concentrationdependent and duration-dependent inhibiting effect on the growth of MCF-7 cells in culture [as also observed earlier in these cells (28-30) and in the human ZR-75-1 breast cancer cells (19) grown in cell culture and as tumors in nude mice]. The effect of 5 [iM toremifene REPORTS 1413
Downloaded from jnci.oxfordjournals.org at Georgetown University Library on June 9, 2011
For the experiments, 2 X 105 MCF-7 cells were plated in 5 mL of 5% DC-FCSRPMI-1640, without E2 and insulin (day 0) in 50-mm diameter petri dishes. The next day (day I) the culture medium was removed, and new medium containing toremifene or tamoxifen with or without E2 was added. The media were changed every 2nd or 3rd day. All the experiments were repeated at least twice.
mark). In the experiments, we used 8- to 12week-old intact female mice, implanted with E,pellets (19). Silastic capsules of 5 mm (0.058-in outer diameter X 0.077-in internal diameter; Dow Corning Corp., Midland, Mich.), filled with 17B-estradiol, were implanted subcutaneously in the flanks of the animals under light ether anesthesia 1 day before injection of cells. Cells (107 MCF-7 and ZR-75-1) in exponential growth phase were trypsinized and washed. Under light anesthesia, each animal was given a subcutaneous injection of these cells in 0.2 mL of medium into one shoulder (23 mice for each cell line). The occurrence of tumor was monitored twice a week; tumor sizes were measured in two dimensions in millimeters, and the volumes were calculated according to a formula: V = (ir/6)(rf, X
