Signaling Pathways of Apoptosis Activated by Aromatase Inhibitors and Antiestrogens

INTRODUCTION

For many years, research in the field of endocrine-mediated breast cancer has focused on the proliferative effects of estrogen. However, recent work has also demonstrated a role for this steroid hormone on the suppression of apoptosis (1 , 2) . Apoptosis is the highly orchestrated process of cell death. Apoptosis can be initiated by a wide variety of stimuli, including developmental signals, cellular stress, and disruption of cell cycle (3) . However, the execution of apoptosis is remarkably uniform, involving characteristic morphological and biochemical changes (4) . A number of the key factors involved in the regulation, coordination, and execution of apoptosis have been identified. Among them, the Bcl-2 family of proteins is a critical regulator of apoptosis, acting to either inhibit or promote cell death. In addition, most of the morphological changes that occur during apoptosis are caused by a set of cysteine proteases named caspases (5 , 6) .

Unlike most types of cancer, breast cancer is known to be under hormonal control. Estrogen has been reported to stimulate the growth of breast cancer expressing functional ERs 1 by affecting the cell cycle machinery (7 , 8) and by inducing specific growth factors and their receptors (9) . In addition to inducing breast cancer cell proliferation, estrogen prevents the induction of apoptosis, at least in part, by altering the expression of the Bcl-2 family of proteins. Treatment of MCF-7 breast cancer cells with E₂ results in a decrease in proapoptosis Bax (2) and an increase in antiapoptosis Bcl-2 mRNA and protein (1) . It has been reported that E₂ deprivation or antiestrogen treatment induces apoptotic cell death in MCF-7 tumors (10 , 11) . Recent studies have shown that the induction of apoptosis and inhibition of proliferation associated with E₂ deprivation of MCF-7 tumors was associated with increased expression of ER, p27 protein levels, p21 protein levels, and a decrease in Bcl-2, cyclin D1, and Rb protein expression (12 , 13) .

Hormonal therapies have an important role in treating ER-positive breast cancer patients. Currently, two pharmacological strategies are used to ameliorate the proliferative effects of estrogens. One is the inhibition of estrogen action using antiestrogens that interact with the ER. The other is the inhibition of estrogen synthesis by inhibitors of aromatase (14) . The antiestrogen Tam has been shown to produce a significantly better response rate than chemotherapy in postmenopausal patients with ER-positive breast cancer. However, resistance to Tam inevitably develops and is mediated, in part, by its partial agonist properties. Aromatase inhibitors are a new group of drugs that are clinically useful in advanced breast cancer and Tam-resistant patients. These compounds block the synthesis of estrogens and do not have estrogenic activity (15) . Compared with Tam, treatment with the aromatase inhibitor Let has been shown to be superior to Tam in our preclinical model (16) and recently in breast cancer patients also (17) . Let treatment causes regression in MCF-7 human breast cancer cells transfected with the aromatase gene (MCF-7Ca) and grown as tumors in nude mice. This suggested that the net decrease in cell number is being induced by apoptosis. To date, there have been no studies on the molecular mechanisms of apoptosis induced by aromatase inhibitor.

In this study, we have investigated the effects of the aromatase inhibitors: Let, Anas, and 4-OHA compared with the antiestrogens Tam and Fas and E2W condition on cell cycle progression and induction of apoptosis. The estrogen-dependent human breast cancer cell line (MCF-7) stably transfected with the human aromatase gene (Ref. 18 ; MCF-7Ca) was used to study the molecular pharmacology associated with treatment with aromatase inhibitors. We have identified the molecular mechanisms and signaling pathways activated during induction of apoptosis by focusing on the involvement of antiapoptotic proteins, proapoptotic proteins, and caspases. In addition, we demonstrated the in vivo effects of Let compared with Tam, Fas, and E2W on inducing apoptosis of MCF-7Ca grown as tumors in nude mice. Finally, we have correlated protein expression with mRNA expression to additionally characterize the molecular mechanisms associated with treating MCF-7Ca breast tumor xenografts with aromatase inhibitor and antiestrogen.

MATERIALS AND METHODS

Materials

DMEM, P/S, trypsin/EDTA solution, DPBS, and G418 were purchased from Life Technologies, Inc. (Grand Island, NY). Phenol red-free improved minimum essential medium (without partial remission) and phenol red-free trypsin/EDTA were obtained from Biofluids, Inc. (Rockville, MD). FBS and dextran-coated charcoal-treated serum were from Hyclone (Logan, UT). E2, AD, 4-OHTam, Tam, PI, RNase (DNase free), 10% neutral-buffered formalin, poly-l-lysine, RNA later, and paraformaldehyde were purchased from Sigma-Aldrich Chemical Company (St. Louis, MO). Let was generously provides by Dr. Dean Evans (Novartis Pharmaceuticals, Basel, Switzerland). Anas and Fas were kindly supplied by Dr. Alan Wakeling (Astra Zeneca Pharmaceuticals, Macclesfield, United Kingdom). 4-OHA was synthesized in our laboratory (15) .

Cell Line

The MCF-7 human breast cancer cell line stably transfected with the human aromatase gene, designated MCF-7Ca (18) , were kindly provided by Dr. Shiuan Chen (City of Hope, Duarte, CA). MCF-7Ca cells were maintained in a standard medium of DMEM supplemented with 5% FBS, 1% P/S, and 700 μg/ml G418. Cells were cultured at 37°C in a humidified incubator with 5% CO₂ and 95% air.

Cell Synchronization

MCF-7Ca cells (1.5 × 10⁶ cells) were plated into a T-75 flask and were cultured in standard medium for 4 days. Then to circumvent interference from the steroids present in FBS and the estrogenic effect of phenol red, cells were switched to E₂-free medium, which consisted of improved MEM without phenol red with 5% dextran-coated charcoal-treated serum, 1% P/S, and 700 μg/ml G418 4 days before the start of the experiments (day −4). At day −1, cells were harvested by trypsinization. Cells were synchronized at G₁ phase of cell cycle by this procedure before each experiment.

Cell Growth Assay

At day −1, flasks of 80% confluent synchronized MCF-7 Ca cells were harvested by trypsinization. Cells were seeded in triplicate into 24-well plates at 1 × 10⁴ cells/well and were allowed to attach overnight. On day 0 of the experiment, the medium was replaced with fresh medium containing the aromatase substrate, AD at a concentration of 25 nm. The three aromatase inhibitors, Let, Anas, and 4-OHA at concentrations of 1–1000 nm, and the two antiestrogens, 4-OHTam and Fas at a concentration of 1000 nm, were added to medium to test their effects on cell growth. Cells cultured without AD designated E2W were treated with ethanol vehicle. The medium and drugs were refreshed every 3 days. On day 8 of treatment, the cells were harvested, and total cell number was determined using a Coulter Counter.

Cell Cycle Analysis

On day −1 of the experiments, synchronized MCF-7 Ca cells were plated into T-75 flasks at 5 × 10⁵ cells and allowed to adhere overnight. On day 0, the cells were treated with 1 μm Let, Anas, 4-OHA, 4-OHTam, or Fas in the presence of AD (25 nm). The medium containing the appropriate treatments was changed on day 3 and day 6. Cells were harvested on days 3, 6, and 8 of treatment, washed twice with DPBS, and were fixed with 100% cold ethanol. Cells were stained with PI solution (50 μg/ml PI in DPBS, 0.5 μg/ml RNase-DNase free), and the fluorescence of individual nuclei (10,000 events) was analyzed by flow cytometry (Becton Dickinson FACScan). The percentage of cells in the G₀-G₁, S, and the G₂-M phases of the cell cycle was determined using ModFitLT V3.1 (Verity Software House, San Jose, CA).

Caspase Activity Assays

Caspase activation was determined using a caspase-3 cellular activity assay kit, a caspase-6 colorimetric assay kit, and a caspase-8 assay kit from Calbiochem-Novabiochem (San Diego, CA) as recommended by the manufacturer’s instructions. Briefly, 100 μg of whole cell lysates were used for the assays. The volume of cell lysates was adjusted to 50 μl with lysis buffer, and 50 μl of 2× reaction buffer were added to each sample. Five μl of 4 mm substrate DEVD-pNA (caspase-3), VEID-pNA (caspase-6), or Ac-IETD-pNA (caspase-8) were added, and the samples were incubated at 37°C for 1 h. Absorbance of the samples was read at 405 nm in a microtiter plate reader. The changes in activity were determined as a percentage of the vehicle-treated cells.

Animals

Ovariectomized female BALB/c athymic nude mice 4–6 weeks of age were obtained from National Cancer Institute (Frederick, MD). The animals were housed in a pathogen-free environment under controlled conditions of light and humidity and received food and water ad libitum. All animal studies were performed according to the guidelines and approval of the Animal Care Committee of the University of Maryland School of Medicine.

MCF-7Ca Xenograft Tumor Growth in Female Athymic Nude Mice

Subconfluent MCF-7Ca cells cultured in DMEM with 5% FBS, 1% P/S, and 700 μg/ml G418 were harvested by scraping. Cells were collected by centrifugation, washed in DPBS and resuspended in Matrigel (10 mg/ml) at 2.5 × 10⁷ cells/ml. Each ovariectomized mouse received s.c. injections at two sites in each flank with 0.1 ml of the cell suspension (2.5 × 10⁶ cells/site). AD (100 μg/day, s.c.) was injected 1 day after cell inoculation and, unless noted, was continued for the duration of the experiment. Tumor volumes were measured weekly with calipers and were calculated using the formula 4/3π × r₁² × r₂ (r₁ < r₂). Treatments began when the tumor volumes reached 300 mm³. Animals were grouped so that total tumor volume of each group was approximately the same. Animals then received s.c. injections of Let (10 or 100 μg/day), Tam (100 μg/day), or (Fas 1 mg/day) in addition to the AD supplement. Control mice received s.c. injections of AD only. E2W mice received s.c. injections of vehicle (0.3% hydroxypropyl cellulose, 0.1 ml/day) without AD supplement. Animals were treated for the indicated time periods, after which, they were sacrificed by decapitation, and the blood was collected. Tumors and uteri were excised, cleaned, and weighed. Harvested tumors were placed on ice and rapidly fixed in 10% neutral-buffered formalin (Sigma-Aldrich Chemical Co.) overnight before being embedded in paraffin for immunohistochemistry or stored in RNAlater solution at −80°C for protein and gene expression analysis.

Detection of Apoptosis

In Vitro Assay.

Synchronized MCF-7Ca cells were treated with Let, Anas, 4-OHA, or Fas at concentrations 100 and 1000 nm in the presence of AD (25 nm). E2W cells were cultured in steroid-free medium in the absence of AD. On day 8 of treatment, adherent cells were harvested by trypsinization and combined with cells floating in the media. Cells were washed with ice-cold DPBS, and 1 × 10⁶ cells were suspended in 0.5 ml of DPBS. The cell suspension was fixed by adding 5 ml of 1% (w/v) paraformaldehyde in DPBS and incubated on ice for 15 min. Cells were washed twice with ice-cold DPBS before being fixed overnight at −20°C in 70% (v/v) ethanol. Apoptosis was determined by the TUNEL method using the APO-Br dUTP kit (Phoenix Flow Systems, Inc., San Diego, CA) according to the manufacturer’s directions. Apoptotic cells were detected by flow cytometry analysis. Data were analyzed by using CELLQuest-BrdUrd software from Phoenix Flow Systems, Inc.

In Situ TUNEL Assay.

Paraffin fixed tissue sections 4-μm thick were spread on poly-l-lysine coated slides. The slides were pretreated with proteinase K (20 μg/ml, 15 min) before TUNEL staining using ApopTag Plus Peroxidase In Situ Apoptosis Detection kit for immunoperoxidase staining (Intergen, Purchase, NY) according to the manufacturer’s instructions.

Preparation of Total Cell Lysates

Adherent and floating cells were harvested after 1, 2, 3, and 4 days of treatment. The floating cells were collected by centrifugation at 500 × g (4°C) for 20 min, and the cells that remained attached to the culture flask were detached by scraping. Both the floating and attached cells were mixed together, centrifuged, and washed in 20 ml of ice-cold DPBS. The cell pellet was resuspended in chilled cell lysis buffer [0.1 m Tris HCl, 0.5% Triton X-100, protease inhibitors mixture (Complete TM; Boehringer, Indianapolis, IN)] and sonicated for 20 s. The homogenates were transferred to new Eppendorf tubes and were incubated on ice for 30 min. The homogenates were spun at 10,000 × g for 30 min, and supernatants were separated and stored at −80°C.

Tumors were thawed on ice, washed, and homogenized in ice-cold T-PER Tissue Protein Extraction Reagent (Pierce, Rockford, IL) plus protease inhibitors mixture. The homogenates were centrifuged at 5000 × g at 4°C for 30 min. Supernatants were stored at −80°C. Protein concentrations were determined using a Bio-Rad kit (Bio-Rad, Hercules, CA).

Gel Electrophoresis and Western Blotting

Equal amounts of total protein (50–100 μg) were subjected to SDS-PAGE (60 V, 3 h), using the mini-Protean3 electrophoresis module assembly (Bio-Rad) and transferred (90 V, 1 h) to nitrocellulose membranes (Hybond ECL; Amersham, Buckinghamshire, UK). Immunodetections were performed using mouse monoclonal antibodies against human Bcl-2, human caspase-9, actin, rabbit polyclonal antibodies against human Bax (Oncogene Research Products, Boston, MA), rabbit polyclonal antibodies against human caspase-7 (Cell Signaling Technology, Beverly, MA), mouse monoclonal antibodies against human p21, mouse monoclonal antibodies against human p53 (Upstate Biotechnology, Lake Placid, NY), or purified mouse antihuman PARP monoclonal antibody (BD PharMingen, San Diego, CA). Immunoreactive bands were visualized using the enhanced chemiluminescence detection reagents (Amersham Corp., Arlington Heights, IL) according to the manufacturer’s instructions and quantitated by densitometry using Bio-Rad software (Quantity One).

RNA Isolation and RPA

Total RNA, free from DNA, was isolated from MCF-7Ca tumors using the Trizol reagent (Invitrogen, Carlsbad, CA). The purity and integrity of the RNA isolated were determined by UV absorbance spectrophotometer and by agarose gel electrophoresis, respectively.

The RiboQuant Multiprobe protection assay system (BD PharMingen) was used for simultaneous detection and quantification of multiple mRNA species. We used customized templates (Custom human template set; BD PharMingen) to detect levels of bcl-2, bax, PARP, caspase-3, caspase-6, caspase-7, caspase-8, caspase-9, p53, p21, cyclin D1, PCNA, and c-myc mRNAs. The template also included two housekeeping genes, L32 and GADPH, as internal standards. The assays were preformed as described in the manufacturer’s protocol. Briefly, ³²P-labled antisense RNA was transcribed in vitro using T7 RNA polymerase (PharMingen) and [α-³²P]UTP (Amersham Life Science, Piscataway, NJ). Twenty μg of total RNA were hybridized with excess of ³²P-labeled probes at 56°C for 16 h. Single-stranded and unhybridized excess mRNA was digested by RNase-DNase free (PharMingen). The protected hybridized double-stranded RNA was precipitated, dried, and resuspended in 5 μl of 1× loading buffer (PharMingen). Samples were separated by electrophoresis in a standard sequencing gel of 5% polyacrylamide-8 m urea gel, 0.5% Tris-boric acid-EDTA at 50 W. The gel was dried on blotting paper for 1 h at 80°C. The dried gel was placed on BioMax MR film (Eastman Kodak Company, Rochester, NY) and was exposed at −70°C. For comparative analysis, the mRNA levels of each gene were quantitated as a percentage of the ribosomal protein L32 mRNA levels.

Statistical Analysis

The statistical differences were analyzed using the Student’s t test on SigmaPlot 2000. All statistical tests were two sided, and differences were considered to be statistically significant when P < 0.05.

RESULTS

Aromatase Inhibitor and Antiestrogen Inhibit Growth of MCF-7Ca Cells in Culture.

To address the direct effect of aromatase inhibitors on the proliferation of MCF-7Ca cells in culture, cells were cultured in E₂-free medium 4 days before the experiment. The aromatase substrate, AD, was then added as the source of E₂ synthesis. The optimum concentration of AD that induced cell proliferation to a level comparable with E₂ (1 nm) in cells grown in normal medium after 8 days of treatment was 25 nm (Fig. 1A) ⇓ . This concentration of AD was added on day 0 of all future experiments of MCF-7Ca cells in culture. The three aromatase inhibitors Let, Anas, and 4-OHA at concentrations 1–1000 nm dramatically reduced AD-induced cell proliferation in a dose-dependent manner (Fig. 1B) ⇓ . The antiproliferative effects of aromatase inhibitors were comparable with that induced by E2W. At the same concentration (1000 nm), all three aromatase inhibitors showed greater growth inhibitory effects than antiestrogens 4-OHTam and Fas (P < 0.005 for 4-OHTam and P < 0.05 for Fas versus the aromatase inhibitors).

Aromatase Inhibitors and Antiestrogens Induce Disruption of Cell Cycle Progression and Activation of Apoptosis.

To investigate the causes of the antiproliferative effects of aromatase inhibitors and antiestrogens, cells were stained with PI, and cell cycle analysis was performed by flow cytometry. The cell cycle profile of MCF-7Ca cells treated with aromatase inhibitors, antiestrogens, and E2W were similar as each treatment increased in the percentage of cells in the G₀-G₁ phase of the cell cycle and reduced the fraction of cells in the S and G₂-M phases (Table 1) ⇓ . Among all of the treatments, Let caused the greatest accumulation of cells (83.42%) in the G₀-G₁ phase. This compared with 49.70% in the control. In response to treatment with Let, there was also a significant decrease (P < 0.005) in the number of cells in S phase (10.56%) and G₂-M phase (6.01%). This compared with 35.22 and 15.07%, respectively, in the control-treated cells. These effects of aromatase inhibitors and antiestrogens on cell cycle progression were detected starting on day 3 of treatment (Table 1) ⇓ and were continued to day 6 and day 8 (data not shown).

In addition to inhibiting cell cycle progression, we also examined whether treatment with aromatase inhibitors and antiestrogens was reducing cell number by inducing cell death. PI staining for DNA content showed an increase in the percentage of cells in the sub-G₁ phase (data not shown), which raised the possibility that cells were undergoing apoptosis. To investigate this possibility, TUNEL assays, which detect DNA fragmentation (a characteristic of apoptosis), were performed using flow cytometry analysis. After 8 days of treatment, Let (100 and 1000 nm) induced apoptosis by 14.38 and 13.51%, respectively, compared with 1.82% of the control cells (Fig. 2) ⇓ . Treatment with Anas (1000 nm) induced apoptosis in 13.07% of cells and 4-OHA (100 and 1000 nm) in 9.46 and 10.77% of cells, respectively. The induction of apoptosis by Let and Anas was comparable with that induced by E2W. Interestingly, the level of apoptosis induced by pure antiestrogen Fas (9.44%) was lower than that induced by Let and Anas.

Taken together, these results indicated that the antiproliferative effects of aromatase inhibitors and antiestrogens on MCF-7Ca cells are mediated by disrupting cell cycle progression by causing growth arrest at the G₁ phase of cell cycle and by inducing cell death by apoptosis.

Apoptosis Induced by Let, 4-OHTam, Fas, and E2W Involves Down-Regulation of Bcl-2 and Up-Regulation of Bax Protein Expression.

After observing induction of apoptosis by aromatase inhibitors and antiestrogens, we set out to examine the mechanisms involved. Because Let was the most effective of the three inhibitors, we tested this compound in comparison to 4-OHTam, Fas, and E2W.

We first investigated the effect of Let, 4-OHTam, Fas, and E2W on the expression of Bcl-2 protein. All treatments induced a decrease in the level of Bcl-2 protein when compared with the control. This reduction was detectable on day 1 of treatment, and there was an additional decrease on day 2 of treatment (Fig. 3A) ⇓ . Let induced down-regulation of Bcl-2 protein in dose- and time-dependent manner. Quantitative analysis showed a reduction to 0.46, 0.27, and 0.25-fold of control by Let concentrations of 10, 100, and 1000 nm, respectively, on day 2 of treatment. Treatment with the antiestrogens 4-OHTam and Fas decreased Bcl-2 protein expression more than the aromatase inhibitors to 0.17- and 0.08-fold when compared with control. On day 2, E2W decreased the Bcl-2 levels to a less extent than all other treatments to 0.46-fold of control. On day 3 and day 4, the reduction in Bcl-2 protein levels after treatment was less than the levels on day 2 (data not shown).

We next examined the effects of Let, 4-OHTam, Fas, and E2W on expression of the proapoptotic protein Bax. MCF-7Ca cells showed increased expression of Bax in response to all treatments when compared with control. Obvious quantitative changes were seen on day 1 with the peak on day 2 of treatment (Fig. 3B) ⇓ . These increases were maintained through day 3 and day 4 of treatment (data not shown). On day 2, 4-OHTam and Fas increased Bax protein levels to 2.9- and 2.5-fold of control similar to the increase by E2W of 2.7-fold. Let (10, 100, and 1000 nm) up-regulated Bax protein expression to a greater extent with increase of 5.2-, 3.6-, and 4.0-fold, respectively.

Effects of Let, 4-OHTam, and Fas on the Activation of Caspases in MCF-7Ca Cells.

The involvement of Bcl-2 and Bax implied that induction of apoptosis by these treatments may be via the mitochondrial pathway. We next examined whether caspase-9, a molecule that plays a major role as an initiator caspase in this pathway, was activated. To address this question, we used an antibody to caspase-9 that can detect both the pro- and active forms of the protein. In comparison to control, Let (100 and 1000 nm) increased expression of the active-form (M_r 35,000) of caspase-9 by 1.2- and 1.5-fold, respectively, after 3 days of treatment. The greatest increase was seen on day 4 of treatment with 1.6-, 2.2-, and 2.0-fold increases by Let concentrations of 10, 100, and 1000 nm, respectively (Fig. 3C) ⇓ . E2W cells showed significantly less increase in expression of the active form by 1.6- and 1.4-fold on day 3 and day 4, respectively. Fas treatment increased expression of the active-form of caspase-9 to about the same extent as E2W. Interestingly, 4-OHTam had only a modest effect on expression of the active-form of caspase-9, which increased by 1.2-fold.

To identify whether other caspases downstream of caspase-9 were affected, caspase activity assays were performed. Caspase-3 is believed to be the major effector caspase that functions downstream of caspase-9. It has been reported that caspase-3 in MCF-7 cells is nonfunctional (19 , 20) . Using a caspase-3 substrate (DEVD-pNA), no induction of caspase-3 was detected after treatment with Let, 4-OHTam, Fas, and E2W in MCF-7Ca cells (data not shown). This suggests the involvement of other effector caspases in apoptosis induced in MCF-7Ca cells after treatments with aromatase inhibitors and antiestrogens.

Another caspase activity assay using a caspase-6-specific substrate (VEID-pNA) was performed. No induction in caspase-6 activity was detected on day 1 and day 2 after treatment (data not shown). The activity was slightly increased on day 3 of treatment (Fig. 4) ⇓ . On day 4, the caspase-6 activity in Let-treated cells was increased in the range of 160–190% of control in a concentration-dependent manner. The activity was comparable with that induced by E2W. 4-OHTam and Fas-treatment caused a smaller increase in caspase-6 activity than Let-treatment with activity of 150 and 130% of control, respectively.

We also analyzed whether caspase-8, the initiator caspase in death receptor-mediated pathway, is involved in apoptotic activation in our test system. Examination of caspase-8 activity using caspase-8-specific substrate Ac-IETD-pNA revealed no increase in activation of caspase-8 activity (data not shown). Therefore, the apoptotic-signaling pathway activated by aromatase inhibitors and antiestrogens is likely to be mediated via the mitochondrial pathway.

The Effect of Let, Tam, Fas, and E2W on MCF-7Ca Tumors Growth in Female Ovariectomized Nude Mice.

After data were obtained from MCF-7Ca grown in culture, we use the MCF-7 Ca tumor xenograft model to examine the effects of therapies on tumor growth in vivo. An in vivo intratumoral aromatase nude mouse model that simulates, to some extent, postmenopausal, hormone-dependent breast cancer has been developed and established in our laboratory (21) . Using this model, the effects on MCF-7Ca tumor growth of Let, Tam, Fas, and E2W compared with control were studied. Daily s.c. injections of AD (100 μg) were sufficient to induce MCF-7Ca cells to form tumors in ovariectomized nude mice. Mean tumor volumes reached 300 mm³ after 6 weeks of AD administration. As shown in Fig. 5A ⇓ , Let (10 and 100 μg) induced regression in tumor growth. This effect was evident as early as day 3 of treatment, and tumor volumes continued to reduce over the duration of experiment. On day 14 of treatment, Let (10 μg/day) and Let (100 μg/day) had reduced tumor volumes to 54 and 62% of the initial volume, respectively. Tumor volumes in the control animals were increased by 135% on day 14 of treatment. Although the antiestrogens Tam and Fas reduced the rate of tumor growth (105 and 98%, respectively, at day 14) when compared with control, neither treatment caused a regression in tumor volume. In animals receiving only the vehicle, which mimics the estrogen withdrawal conditions, tumor volume also decreased (80% of the initial volume at day 14 of treatment) but to a lesser extent than that observed in the Let-treated groups. On day 14 of treatment, mean tumor volume in Let 10- and 100-μg groups were significantly lower than in E2W group (P = 0.001 and 0.016, respectively).

At the indicated time periods of treatment, animals were sacrificed, and the tumors were removed, cleaned, weighed, and stored at −80°C. The uterus was also removed from each animal and weighed. It has been reported that the weight of uterus is correlated with the levels of circulating estradiol in this model (16 , 22) . The uterine weights of the Let-treated animals were significantly less than those of the control animals (Fig. 5B) ⇓ . The reduction in uterine weight by Let treatment was evident as early as day 3 of treatment (P < 0.001 versus control) and was prolonged through day 14 of treatment (P < 0.001 on day 7 and P < 0.05 on day 14 versus control). Uterine weights in E2W animals were also decreased, but the reduction occurred more slowly than in the Let-treated animals (P < 0.05 on day 7 and day 14 versus control). Animals treated with Fas also showed a significant decrease in uterine weights (P < 0.05 on days 3, 7, and 14 versus control). In contrast, the uteri of the Tam-treated animals were hyperplasic and weighed significantly more than those of the control animals (P < 0.001 and P < 0.05 on day 3 and day 7, respectively, versus control). This is consistent with the fact that Tam has an agonistic effect on the endometrium in mice.

Detection of Apoptosis in MCF-7Ca Tumors Grown in Female Ovariectomized Nude Mice.

To determine whether the apoptosis induced by aromatase inhibitors and antiestrogens in MCF-7Ca cells in vitro also occurred in MCF-7Ca tumors grown in vivo, we used two approaches to examine their effects. The first approach used in situ TUNEL staining to detect fragmented DNA indicative of cells undergoing apoptosis. TUNEL staining of control tumors revealed a low incidence of apoptosis with the apoptotic index range from 0.4 to 2.1% apoptotic cells (Fig. 6, A and B) ⇓ . The induction of apoptosis detected in the Let-treated tumors was dose- and time-dependent manner, which peaks at 14.5% (P = 0.023 versus control) and 15.8% (P = 0.039 versus control) on day 3 of treatment for Let 10 and 100 μg/day, respectively. Although the induction of apoptosis by Let was sustained for the duration of the experiment, the percentage of apoptotic cells declined with the longer treatment (5.5 and 9.9% apoptotic cells on day 14 for Let 10 and 100, respectively). E2W tumors showed a similar pattern as Let-treated tumors but with lower percentage of apoptosis. Although we did not observed significant reduction in tumor volumes, Tam and Fas also induced apoptosis of 5.0% (P = 0.03 versus control) and 5.5% (P = 0.024 versus control) on day 3 of treatment but decreased to 2.1 and 4.9% on day 14 of treatment.

The second approach to detect the early events of apoptosis was by observing the proteolysis of PARP, the DNA repair enzyme that can be cleaved by effector caspases. Western immunoblotting for the PARP protein revealed a reduction in the levels of the intact form (M_r 116,000) along with an elevation in the levels of the proteolytic form (M_r 85,000) in tumors treated with Let, Tam, Fas, and E2W (Fig. 7A) ⇓ . On Day 3, levels of the proteolytic form of PARP in tumors treated with Let (10 and 100 μg/day) and E2W were slightly increased to 1.25-, 1.29-, and 1.12-fold of control, respectively. The maximum level of proteolysis was achieved on day 7 of treatment when compared with control tumors, the proteolytic form of PARP was increased significantly (P < 0.05) by 4.82- and 29.25-fold in the Let 10 and Let 100, respectively. Levels were also significantly higher than in E2W tumors. On day 14, the levels of cleaved PARP were 4.11-, 9.54-, and 1.59-fold higher compared with control. In the Tam and Fas-treated tumors, levels of the proteolytic form of PARP were increased 2.68- and 4.49-fold on day 7 and 1.59- and 2.01-fold but not significantly compared with control on day 14 of treatment, respectively. This result was in agreement with the apoptotic index data obtained from the TUNEL assays in vitro and in vivo.

Activation of Caspase-9 and Caspase-7 in MCF-7Ca Tumors Grown in Nude Mice.

Tumors were analyzed for expression of the active forms of caspase-9 and caspase-7. In agreement with the in vitro study, caspase-9 was also activated in MCF-7Ca tumors treated with Let, Tam, Fas, and E2W but not in control tumors (Fig. 7B) ⇓ . Maximal activation of caspase-9 occurred at day 3 of treatment. Proteolysis of PARP in MCF-7Ca tumors treated with Let, Tam, Fas, and E2W indicated the involvement of another effector caspase than caspase-6, which does not cleave PARP. Nuclear lamins has been reported to be the major substrate of caspase-6 (23) . Caspase-3 is believed to be the major protease that functions in cleavage of PARP. Having found that caspase-3 is not activated in MCF-7Ca cells treated with aromatase inhibitors and antiestrogens, we assayed for expression of caspase-7, which has also been reported to cleave PARP (23 , 24) . Fig. 7C ⇓ shows a marked significant activation of caspase-7 in Let-treated MCF-7Ca tumors. The protein expression of the cleaved active-form (M_r 20,000) of caspase-7 was increased in the Let-treated tumors in a dose- and time-dependent manner. The protein level of the proenzyme form (M_r 35,000) was gradually decreased in correlation with an increase in the cleaved form of caspase-7. Tumors from animals treated by estrogen withdrawal also showed activation of caspase-7. The protein level of the active-form in the E2W tumors was also increased in a time-dependent manner but to a significantly lesser extent than in the Let-treated tumors. The active form of caspase-7 was also increased in the Tam-treated tumors to about the same extent as E2W tumors, although we did not detect a significant number of apoptotic cells in this group. Interestingly, no active form of caspase-7 was detected in the Fas-treated tumors, although we did find PARP proteolysis and apoptotic cells in this group.

Up-Regulation of p53 and p21 Protein Levels in MCF-7Ca Cells and Tumors.

After seeing that aromatase inhibitors and antiestrogens induced cell cycle arrest at G₀-G₁ phase, we additionally investigated the molecular mechanisms involved by focusing on the molecules that regulate at the G₁-S-phase transition. One critical regulator of both apoptosis and cell cycle is p53. The tumor suppressor protein p53 blocks cell cycle progression at the G₁-S transition (25) . Levels of p53 protein expression in MCF-7Ca cells were not altered after treatments in vitro (Fig. 8A) ⇓ . However, we did observe significant up-regulation in expression of p53 protein in MCF-7Ca tumor xenografts (Fig. 8C) ⇓ . Compared with control tumors, the high dose of Let (100 μg/day) induced a 3.25-fold increase in p53 protein levels on day 3 of treatment. On day 7 and day 14 of treatment, the induction declined to 1.56- and 1.23-fold of control, respectively. Up-regulation of p53 protein expression by the low dose of Let (10 μg/day) occurred at a later time than was observed with Let (100 μg/day). Significantly increased p53 protein levels were detected on day 7 (2.56-fold) and remain high on day 14 of treatment (2.06-fold). Significant up-regulation of p53 by Tam peaked on day 3 (2.91-fold), whereas Fas induced p53 expression by 2.45-fold of control on day 7 of treatment. In the E2W tumors, we observed a slight and transient increase in p53 levels (1.69-fold of control) on day 3 of treatment.

In addition, we also examined the effects of aromatase inhibitors and antiestrogens on expression of the p21 protein both in vitro and in vivo. Levels of p21 protein of MCF-7Ca cells were increased in all treatments compared with control. Let induced up-regulation in expression of p21 protein in a dose- and time-dependent manner (Fig. 8B) ⇓ . Although higher levels of p21 protein expression were induced by Let (10 μm), at equivalent concentrations of 1 μm, there was no significant difference between p21 protein expression in Let-, 4-OHTam-, and Fas-treated cells. In agreement with in vitro results, expression of p21 protein is also significantly up-regulate in MCF-7Ca tumors after treatment of Let compared with control, E2W, and antiestrogens (Fig. 8D) ⇓ . On day 7 of treatment, Let (10 and 100 μg/day) increased p21 protein expression by 9.36- and 3.48-fold of control, respectively. This effect was maintained through day 14 of treatments with increases of 3.73- and 3.34-fold of control. Up-regulation of p21 protein expression was also observed on day 7 in Fas-treated tumors when increase in p21 protein levels were 4.07-fold of control but returned to the control level on day 14 of treatments. Only a slight increase of p21 protein expression was evident in E2W and Tam-treated tumors (1.31- and 1.35-fold of control, respectively, on day 7 of treatments).

Modulation in the Expression of Genes Involved in Apoptosis and Cell Cycle Regulation.

To further characterize the molecular mechanisms associated with the antiproliferative effects and induction of apoptosis induced by aromatase inhibitors and antiestrogens, we examined changes in the expression of genes involved in these processes. We used the RPA method using the RiboQuant Multiprobe protection assay system to simultaneously detect and quantify levels of expression of multiple mRNA species. On the basis of our preliminary in vitro results, we selected to examine expression of bcl-2, bax, PARP, caspase-3, caspase-6, caspase-7, caspase-8, caspase-9, p53, p21, cyclin D1, PCNA, and c-myc genes in MCF-7Ca breast tumor xenografts. In MCF-7Ca tumors treated with Let 10, Let 100, Tam, Fas, and E2W for 3 days, bax mRNA levels were increased by 1.13-, 1.23-, 1.34-, 1.36-, and 1.56-fold of the control, respectively. The increase in the levels of bax mRNA was transient because levels similar to those in the control tumors had returned on day 7 and day 14 of treatment. For bcl-2, a significant decrease in mRNA levels was found in Let 10 and Let 100 tumors with the peak of reduction on day 7 of treatment. On day 7, the levels of bcl-2 mRNA in Let 10 and 100 were 0.73- and 0.65-fold of control and were at the same level on day 14 of treatment. No alteration of bcl-2 mRNA level was detected on day 3 and day 7 for E2W-, Tam-, or Fas-treated tumors. However, on day 14, the level of bcl-2 mRNA in E2W tumors had dropped to 0.64-fold of control. On day 14, a slight decrease in expression of bcl-2 mRNA to 0.9-fold of control was detected in Tam- and Fas-treated tumors (data not shown).

Quantitative analysis of caspase-6, caspase-7, caspase-8, and caspase-9 mRNA expression showed no significant change in the levels of these genes in the MCF-7Ca-treated tumors. There was also no significant increase in the expression of PARP mRNA after treatment with Let, Tam, Fas, or E2W. In contrast, there was a significant reduction in the level of c-myc mRNA. Let treatment of MCF-7Ca tumors resulted in a dramatic decrease in levels of c-myc mRNA. The reduction in c-myc mRNA expression started on day 3 and was prolonged through day 7 and day 14 of treatment (Fig. 9) ⇓ . Expression of c-myc mRNA also gradually decreased in the Tam-, Fas-, and E2W-treated tumors, but the time course was later than in the Let-treated tumors. On day 14 of treatment, the levels of c-myc mRNA in all treatment groups were almost undetectable.

Reduction in cyclin D1 and PCNA mRNA levels were also observed in MCF-7Ca tumors treated with Let, Tam, Fas, and E2W, and these reductions occurred in a time-dependent manner. Compared with the control, levels of cyclin D1 mRNA were reduced by 0.73-, 0.66-, 0.57-, 0.72-, and 0.86-fold, and PCNA mRNA levels were reduced by 0.61, 0.48, 0.46, 0.60, and 0.75-fold on day 14 of treatment with Let 10, Let 100, Tam, Fas, and E2W, respectively (Fig. 9) ⇓ . The levels of p53 mRNA were increased in MCF-7Ca tumors treated with Let, Tam, Fas, and E2W for 3 days before gradually returning to the basal level on day 14 of treatment. A similar profile was also seen for p21 mRNA (Fig. 9) ⇓ .

DISCUSSION

The aim of this study was to explore the molecular mechanisms associated with the antiproliferative effect of aromatase inhibitors in comparison with antiestrogens for the treatment of breast cancer. By culturing aromatase-transfected MCF-7Ca cells in steroid-free medium supplemented with AD, we were able to test the in vitro effects of both the aromatase inhibitors and antiestrogens on cell growth, cell cycle progression, and induction of apoptosis. Our results indicated that aromatase inhibitors, antiestrogens, and E2W treatment induced inhibition of MCF-7Ca cell proliferation. However, the growth inhibitory effect of aromatase inhibitors was superior compared with antiestrogens (Tam and Fas) or E2W. This has also been demonstrated in recent clinical trials where Let and Anas were more effective in comparison with Tam (17 , 26) . In addition, using our in vivo intratumoral aromatase nude mouse model, we found that Let induced regression of MCF-7Ca tumors, whereas tumor growth arrest was observed in the Tam- and Fas-treated tumors. These results suggest that at least in our tested system, the inhibition of estrogen synthesis by an aromatase inhibitor is more effective than inhibiting the action of estrogen with an antiestrogen. These findings are consistent with our previous studies using this model (27 , 28) . The xenograft model is similar to the postmenopausal breast cancer patient in several respects. Aromatase is expressed in nonovarian tissue and is not regulated by gonadotropins. Although circulating levels of estrogen are low in postmenopausal women, the concentrations of estrogen in the breast are similar to premenopausal levels (29, 30, 31) , in part, because of steroid uptake and, in part, because of local synthesis via aromatase expressed in stromal cells, tumor epithelial cells, and adipocytes (32 , 33) . Although expression of aromatase is limited to the tumor cells in the xenograft model and estrogen concentrations may be relatively high compared with patients, systemic aromatase inhibitor treatment will inhibit estrogen synthesis in all sites and reduce the amount taken up as well as that produced locally. It would seem logical that wherever estrogen is produced, the determining factor would be the amount of estrogen available to the tumor cell’s ERs. An advantage of this xenograft model is that because tumor cells are ER positive and express aromatase, both antiestrogens and aromatase inhibitors can be studied. Although androstenedione is provided as a supplement because the adrenal hormones are deficient in the nude mice (34) , there are no direct effects on tumor growth (16 , 21) .

Estrogens exert their proliferative effect on hormone-dependent breast cancer cells by stimulating cell cycle progression. Studies have been shown that antiestrogen treatment or estrogen deprivation prevents MCF-7 cells entering into S phase of the cell cycle (12 , 35) . We demonstrate here that both aromatase inhibitors and antiestrogens induce growth arrest of MCF-7Ca cells by blocking the G₁-S-phase transition of cell cycle. Preventing entry into S phase subsequently led to a decrease in the percentage of cells in S and M phase. Tumor growth rate is dependent on both proliferation rate and apoptotic rate. In addition to inducing cell cycle progression, estrogens are known to protect against cell death by apoptosis. Several studies both in vitro and in vivo have reported that Tam, Fas, and E2W induce apoptosis (10 , 12 , 36) . Using MCF-7Ca cells culture in vitro and in vivo growth in nude mouse, this is the first study to compare the efficacy of aromatase inhibition by Let, Anas, and 4-OHA to 4-OHTam, Fas, and E2W to induce apoptosis. Our results indicate that aromatase inhibitors have higher efficiency in inducing apoptosis than antiestrogens and E2W. Among all of the treatments, Let was shown to be the most effective at inducing apoptosis. Moreover, the order of potency of the aromatase inhibitors at inhibiting cell proliferation, cell cycle arrest, and apoptotic induction was Let > Anas >4-OHA. It is possible that the differences in efficacies of the three aromatase inhibitors resulted from differences in their abilities to inhibit aromatase activity. We have previously reported that in several aromatase-expressing cell lines, including MCF-7Ca, Let is more effective than Anas and 4-OHA at inhibiting aromatase activity (37) . It should be noted that apoptosis observed in MCF-7Ca tumors treated with Let, Tam, Fas, and E2W was maintained through 14 days of treatment but the maximal induction was at day 3 of treatment. It was unexpected that Let was more effective than E2W. Whether it reflects other actions of the compound is unknown and will require additional investigation.

Previous studies have indicated that E₂ protects against breast cancer cell death by apoptosis is associated with increased levels of the antiapoptotic Bcl-2 protein. Our results here demonstrate that the induction of apoptosis in MCF-7Ca cells by Let, 4-OHT, Fas, and E2W is associated with a reduction in the levels of Bcl-2. The mechanisms responsible for up-regulation of Bcl-2 by E2 are not fully understood but likely involved the interaction of E₂ with ER (38) . In breast carcinomas, Bcl-2 expression has been shown to correlate with ER status (39) . It has been reported that E₂ up-regulates the bcl-2 mRNA level in MCF-7 cells and that the mechanism is via ER binding to two estrogen response elements located within the bcl-2 coding region (40) . It has also been shown that Tam-induced apoptosis in MCF-7 cells involves down-regulation of Bcl-2 but not Bax (41) . In this study, we found up-regulation of proapoptotic protein Bax accompanied with down-regulation of antiapoptotic protein Bcl-2 after all treatments. The modulation in levels of Bcl-2 and Bax was concomitant and were observed as early as day 1 of treatments. Collectively, down-regulation of an antiapoptotic protein together with up-regulation of a proapoptotic protein would be expected to produce a more powerful response and a greater induction of apoptosis. It is well established that Bcl-2 and Bax acts upstream of apoptotic mitochondrial pathway (42) . Decreasing of Bcl-2/Bax ratio leads to opening of mitochondrial permeability transition pores and the release of cytochrome c (43) . Cytochrome c interacts with Apaf-1 and dATP (or ATP) and promotes the proteolytic activation of the caspase-9 zymogen (44) . Activated caspase-9 then initiates a cascade activation of downstream caspases-3, -7, -2, -6, -8, and -10 (45) . We detected the activation of caspase-9 on day 3 and day 4 of treatments with hormonal therapies, which is after the peak of modulation in Bcl-2 and Bax expression. Our results suggest that Bcl-2 and Bax are critical in the early events of apoptosis induced by Let, Tam, and Fas. MCF-7 cells, the parent cell line of MCF-7Ca cells, are known to have lost functional caspase-3 expression because of a 47-bp deletion within exon 3 of the caspase-3 gene (19 , 20) . Caspase-3 is believed to be the major executioner caspase downstream of caspase-9. In agreement with the published data using MCF-7 cells, we did not see activation of caspase-3 in MCF-7Ca cells with any of the treatments. This is most likely because MCF-7Ca cells also lack functional caspase-3.

Initiator caspases such as caspase-9 or caspase-8 can activate the apoptotic pathway not only by activating caspase-3 but also through activation of caspase-7 (46, 47, 48) . Moreover, in vitro studies have shown that caspase-3 and caspase-7 have very similar substrate specificity, one of which is PARP (24) . Additional studies using a cell-free systems and caspase-3-deficient cells from CASP-3-knockout animals have reported that in absence of caspase-3, PARP is cleaved by caspase-7 (49 , 50) . In the present study, we observed the activation of caspase-7 in MCF-7Ca tumors treated with Let, Tam, and E2W but not in tumors treated with Fas. Additionally, levels of the active-cleaved form of caspase-7 were correlated with levels of PARP cleavage and an increased in apoptotic index. We also demonstrated that caspase-6 is activated after treatment with Let and E2W and to a lesser extent by Tam and Fas. Activation of caspase-9 and caspase-7 and the inactivation of PARP correlated with a reduction in expression of the proforms of these enzymes. Analysis of the RPAs indicated that there was no induction in the mRNA levels of caspase-8, caspase-9, caspase-6, caspase-7, and PARP. These results suggested that the activation of caspases observed in our studies is likely to be regulated at the protein level rather than at the transcription level.

E₂-induced cell cycle G₁-phase progression is mediated by activation of Cdk4 and Cdk2 via increased expression of its regulatory subunit, cyclin D1, and decreased association to Cdk inhibitor p21 (7) . It has been reported that Fas induces expression of p21 protein and mRNA (51) and that E₂ deprivation up-regulates expression of p21 but not cyclin D1 (12) . The tumor suppressor protein p53 is a key regulator of cell cycle progression and apoptosis (52) . As a transcription factor, p53 increases transcription expression of several genes, including p21 and bax (53) , and inhibits the expression of bcl-2 by binding to the silencer element in the bcl-2 gene (54) . Here, we demonstrated that the inhibition of cell cycle progression in associated with the hormonal treatment of MCF-7Ca cells correlates with a decrease in cyclin D1 mRNA levels and an increase in the expression of p53 and p21 protein and mRNA levels. We also observed a reduction in c-myc mRNA levels in the tumors treated with Let, antiestrogens, or E2W. Increased c-myc transcription and protein levels are an early response to treatment with E₂. E₂ activates c-myc expression via the interaction of the E₂-ER complex with a 116-bp estrogen response element in the c-myc promoter (55) . The oncoprotein c-myc induces cell proliferation by promoting a G₁-S-phase transition. One mechanism of which is by increasing the transcription of the phosphatase Cdc25A, which is required for activation of cyclin-Cdk complexes in the cell cycle machinery (56) . Studies by others showed that reduction of c-myc expression is followed by activation of cyclin-dependent kinase inhibitor p27, which induces growth arrest and apoptosis (57) . Interestingly, it has been reported that cyclin D1-deficient mice remain fully sensitive to the oncogenic pathway driven by c-myc (58) . This implied different mechanisms of cyclin D1 and c-myc in regulation of cell growth. Therefore, the marked and rapid decrease in expression of both cyclin D1 and c-myc induced by Let should give some advantage for breast cancer treatment.

To the best of our knowledge, this is the first study documenting the mechanisms and signaling pathways regulating cell cycle and induction of apoptosis upon treating breast cancer cells with aromatase inhibitors in comparison with antiestrogens. Our results indicate that aromatase inhibitors and antiestrogens inhibit the proliferation of MCF-7Ca breast cancer cells by inducing cell cycle arrest in the G₀-G₁ phase, and this is coupled with increased apoptosis. The apoptotic signaling pathways induced involves down-regulation of Bcl-2, up-regulation of Bax, and the activation of caspase-9, caspase-6, and caspase-7. The mechanisms responsible for cell cycle arrest are associated with up-regulation of p53 and p21 protein and mRNA levels concomitant with down-regulation of cyclin D1 and c-myc mRNA expression. Although the mechanisms involved appeared to be similar for antiestrogens and aromatase inhibitors, the most significant effects occurred with Let, which were greater than with E2W and consistent with marked effects of Let on tumor and cell growth.