Purpose: Resistance to tamoxifen (TAM) represents a significant challenge to the management of breast cancer. We previously reported that the estrogen receptor (ER)-negative hormone-independent T47D:C42 cell line has both elevated protein kinase Cα (PKCα) protein expression and basal activator protein-1 activity compared with the parental ER+ (hormone-dependent) T47D:A18 cell line. Stable transfection of PKCα to the T47D:A18 breast cancer cell line results in increased basal activator protein-1 activity, reduced ER function, increased proliferation rate, and hormone-independent growth (Tonetti et al., Br. J. Cancer, 83: 782–791, 2000). In this report, we further characterize the role of PKCα overexpression in vivo to elucidate a possible molecular mechanism of tamoxifen resistance.
Experimental Design: To determine whether the T47D:A18/PKCα cell line would produce hormone-independent tumors in athymic mice, we injected T47D:A18, T47D:A18/neo, or the T47D:A18/PKCα20 cell clones bilaterally into the mammary fat pads of athymic mice. Tumor growth was evaluated following treatment with estradiol (E2), TAM, and the pure antiestrogen, ICI 182,780.
Results: Mice receiving either T47D:A18 or T47D:A18/neo cells produced tumors that grew in response to E2 treatment, whereas the untreated control and TAM-treated groups showed no tumor growth. Interestingly, mice receiving the T47D:A18/PKCα20 clone produced tumors in both the control and TAM groups, whereas tumor growth was inhibited in mice treated with E2. PKCα was also overexpressed in an MCF-7 tumor model that also exhibited TAM-stimulated and E2-induced regression.
Conclusions: These results suggest that overexpression of PKCα in breast tumors results in hormone-independent tumor growth that cannot be inhibited by TAM treatment. Furthermore, the finding that E2 has an antitumor effect on breast tumors overexpressing PKCα is a novel observation that may have important therapeutic implications.
TAM3 is an effective endocrine therapy in ∼60% of all ER-positive/progesterone receptor-positive breast cancer cases. The remaining 40% are classified as de novo resistant, and it is unknown why these ER/progesterone receptor-positive tumors do not respond to TAM therapy. Acquired TAM resistance will occur in almost all advanced cases and will often respond to a second-line hormonal therapy, such as an aromatase inhibitor or pure antiestrogen. Some of these cases can be attributed to lack of a functional ER; however, in the majority of cases the mechanism of resistance is unknown (1) . It would be useful to identify a molecular marker that could predict a priori those patients who are likely to respond initially to TAM therapy. Furthermore, identification of a therapeutic target upon TAM failure would be an important advance.
An inverse relationship between ER and PKC expression has been well described in the literature (2) , and increased AP-1 activity has been reported to occur in hormone-independent breast cancer cell lines and tumors (3, 4, 5) . PKC is a family of serine-threonine protein kinases that consists of 12 isozymes, α, βI, βII, γ, δ, ε, ξ, η, θ, ι, μ, and ν, that have been described to date (6) . PKC isozymes exhibit subcellular and tissue-specific localization and, by interacting with distinct substrates, elicit a variety of biological processes including, but not limited to cell cycle progression, differentiation, migration, and apoptosis (6 , 7) . The PKC isozymes are key regulators that occupy a central point in the cell effecting signaling through a multitude of biochemical pathways. Activation of PKCs by the tumor promoter TPA results in increased transcriptional activity of genes containing a TPA-responsive element (8) . A Fos-Jun heterodimer or a Jun-Jun homodimer, known as the AP-1 complex, specifically binds to the TPA-responsive element and induces gene transcription (9 , 10) . PKC isozymes are known to have differing abilities to activate the AP-1 pathway (11 , 12) .
We have previously shown that the ER− hormone-independent human breast cancer cell line T47D:C42 has both elevated PKCα protein expression and basal AP-1 activity compared with the parental ER+ T47D:A18 cell line (13) . Stable transfection of PKCα to the T47D:A18 cell line results in increased basal AP-1 activity, reduced ER function, an increased proliferation rate, and hormone-independent growth (13) . Two previous reports have described the establishment of MCF-7/PKCα stable transfectants, but the authors report conflicting results (14 , 15) . Ways et al. (15) described reduced ER expression, decreased estrogen-dependent gene expression, and a more aggressive neoplastic phenotype in vivo, whereas the MCF-7/PKCα transfectants described by Manni et al. (14) exhibited diminished tumor formation in nude mice.
To determine whether the T47D:A18/PKCα cell line would produce hormone-independent tumors, we established T47D:A18, T47D:A18/neo or T47D:A18/PKCα tumors in ovariectomized athymic mice. We found that the tumors derived from T47D:A18/PKCα exhibited a unique response to E2 administration. These tumors failed to grow in the presence of E2, but tumors could be established in the absence of hormone and by TAM administration. Preexisting tumors rapidly regressed after E2 treatment. We recently reported that MCF-7 tumors transplanted over 5 years in athymic mice progressed through different stages of hormonal response (16) . In the present report, we characterize the T47D:A18/PKCα tumor model and examine the PKC expression profile of the MCF-7 tumor model that also regresses in response to E2.
MATERIALS AND METHODS
Cell Lines and Culture Conditions.
The human breast cancer cell line T47D:A18, a hormone-responsive clone that has been described previously (17) , was maintained in phenol red-containing RPMI 1640 supplemented with 10% FBS. Stable transfectant cell lines T47D:A18/neo and T47D:A18/PKCα20 (13) were maintained in RPMI 1640 supplemented with 10% FBS containing G418 (500 μg/ml). When indicated, prior to proliferation assays, cell lines were placed in phenol red-free RPMI 1640 supplemented with 10% 3× dextran-coated charcoal-treated FBS (E2-free medium) for 3 days.
T47D:A18/neo and T47D:A18/PKCα20 were seeded into T25 tissue culture flasks at 3 × 104 cells/ml in E2-free RPMI supplemented with 500 μg/ml G418. Medium containing compound (10−9 m E2, 10−6 or 10−7 m 4-OHT, or 10−6 or 10−7 m ICI 182,780) was added the following day, and the cells were counted on days 2–10.
Establishment of T47D:A18/PKCα Tumors in Athymic Mice.
T47D:A18, T47D:A18/neo, and T47D:A18/PKCα20 cells were injected s.c. (1 × 107 cells/site) into the axillary mammary fat pads of ovariectomized 4–6-week-old BALB/c athymic mice (Harlan Sprague Dawley, Madison, WI). Mice were divided into three treatment groups consisting of 10 mice/group: control (no treatment); E2 (E2 capsule); and TAM (tamoxifen). E2 was administered via silastic capsules (1.0 cm) implanted s.c. between the scapules. The capsules were replaced every 10 weeks. TAM was administered p.o. at a dose of 1.5 mg/animal daily for 5 days each week in a suspension containing 90% 10 g/l carboxymethylcellulose in double-distilled water and 10% 995 ml/l polyethylene glycol 400–5 ml/l Tween 80. ICI 182,780 (a generous gift from AstraZeneca Pharmaceuticals, Macclesfield, United Kingdom) was injected s.c. at a dose of 5 mg (in 0.1 ml of peanut oil) per animal each week. Tumor cross-sectional area was determined weekly by Vernier calipers and calculated using the formula: length/2 × width/2 × π. Mean tumor area was plotted against time in weeks to monitor tumor growth. Differences in mean tumor area between groups were measured by ANOVA followed by an unpaired, two-sided Student’s t test. The mice were sacrificed by CO2 inhalation and cervical dislocation; tumors were excised and serially transplanted, immediately fixed in 10% buffered formalin for immunohistochemistry, or snap frozen in liquid nitrogen. Frozen tumor specimens were stored at −80°C.
The MCF-7 tumors MCF-7E (E2-stimulated, TAM-inhibited), MCF-7 ST (short-term passage of tumors in TAM, E2-stimulated, TAM-stimulated), MCF-7 TAM LT (long-term passage of tumors in TAM, TAM-stimulated, E2-inhibited) and MCF-7 TAME (E2-stimulated, TAM-inhibited) were passaged in athymic mice as described previously (16) . The mice were sacrificed by CO2 inhalation and cervical dislocation; tumors were excised and either serially transplanted or immediately fixed in 10% buffered formalin for immunohistochemistry.
Nuclear Extract Isolation and Western Blot.
Nuclear extracts were prepared from the T47D:A18 cell line and tumors by a freeze-thaw method (18) . ERα expression was assessed by Western blot analysis using the ERα polyclonal antibody G-20 (Santa Cruz Biotechnology, Santa Cruz, CA). The G-20 antibody was diluted 1:200 in Tris-buffered saline-Tween 20 [20 mm Tris (pH 7.6), 137 mm NaCl, 0.1% Tween 20] containing 5% dry milk. The Supersignal West Dura Western detection system (Pierce, Rockford, IL) was used to visualize the ERα band. Verification of the ERα-immunoreactive band was achieved with a blocking peptide obtained from Santa Cruz Biotechnology. Equal loading of total protein per lane was assessed by staining the gel after electrotransfer with Gel-Code (Pierce).
Excised tumor samples were immediately fixed with 10% buffered formalin and embedded in paraffin. Five-μm-thick sections were deparaffinized by incubation for 1 h at 60°C and rehydrated with 100% CitriSolv dipped in 100% ethanol 10 times. Endogenous peroxidase activity was blocked with 0.3% H2O2 in methanol for 30 min, followed by immersion in graded alcohols. After rinsing with distilled water, antigen retrieval was accomplished by incubating the sections with Protease XXIV (BioGenex, San Ramon, CA) for 10 min and then rinsing in PBS (pH 7.4). Immunostaining was carried out with a HistoMark Biotin Streptavidin Kit (KPL, Gaithersburg, MD). After rinsing with PBS, the sections were blocked for 1 h with normal goat serum. The sections were incubated overnight at 4°C with the primary PKC α (C-20), δ (C-17), or β1 (C-16) rabbit polyclonal antibodies (1:20; Santa Cruz Biotechnology). The sections were then incubated for 30 min with biotinylated goat anti-rabbit IgG (H+L), followed by incubation for 30 min with horseradish peroxidase-streptavidin. Sections were rinsed with PBS for 5 min between each reaction. 3,3′-Diaminobenzidine (DAB+) solution (DAKO, Carpinteria, CA) was used for coloration. Finally, the sections were counterstained with hematoxylin. The specificity of PKCα was verified by two negative controls, substitution of normal rabbit immunoglobulin (DAKO) in place of the PKCα antibody and preincubation with a PKCα blocking peptide (Santa Cruz Biotechnology). Stained sections were photographed at ×40 magnification by an Olympus BX40 microscope attached to a SONY DP10 digital camera.
The identification of apoptotic cells in situ was carried out via the TUNEL-mediated method, which incorporates labeled nucleotides on DNA strand breaks. Tissue sections of PKCα-overexpressing T47D tumors were 5-μm thick and spread on silanized slides. Paraffin-embedded tissue sections were deparaffinized and progressively rehydrated. Endogenous peroxidase activity was blocked with 0.3% H2O2. Sections were then pretreated with proteinase K from DAKO (20 μg/ml for 10 min). The In Situ Cell Death Detection, POD System (Roche Molecular Biochemicals) was then used according to the manufacturer’s instructions. Apoptotic cells exhibiting fragmented DNA were detected by fluorescein-labeled nucleotides, POD-conjugated antifluorescein antibody (Roche Molecular Biochemicals), and the POD substrate DAB+ (DAKO). Staining of nuclei was expressed as the percentage of TUNEL-positive apoptotic cells among 1000 cells in each tumor. TUNEL-positive cells were counted in ×40 high-power fields in 10 different sections of each of two tumors, 100 cells/section. Means were compared between E2 and control tumors by nested repeated-measures ANOVA (19) . Individual tumors were nested within the E2 and control groups. Repeated measures were obtained at 1, 2, 4, 5, 7, 8, 9, and 15 days post-E2 capsule implantation. Differences in variability between E2 and control tumors were taken into account in the analysis. For each day, a t test was done to compare means between E2 and control tumors. To account for multiple significance testing, the Bonferroni significance criterion of P < 0.00625 was used for these eight tests. Statistical analyses were performed using SAS statistical software (SAS/STAT Software, 6.12 edition; SAS Institute Inc., Cary, NC).
T47D:A18/PKCα Cells Form Hormone-independent Tumors in Athymic Mice.
We have determined that overexpression of PKCα in T47D:A18 produces a hormone-independent phenotype in cell culture (13) . To determine whether this phenotype would be expressed in vivo, we injected T47D:A18, T47D:A18/neo, and T47D:A18/PKCα20 cells into the mammary fat pads of ovariectomized athymic mice. The mice were divided into three treatment groups; no treatment (controls), E2 (1.0-cm capsule), and TAM (1.5 mg/day, five times per week). Eight weeks after the injection of T47D:A18 and T47D:A18/neo cells, tumor growth was observed in the E2 group (Fig. 1, A and B)⇓ , whereas no tumors were detected in any of the mice in the control or TAM groups. Remarkably, in the mice receiving T47D:A18/PKCα cells, tumors appeared in both the control and TAM groups (Fig. 1C)⇓ . These results suggest that T47D:A18 tumors that overexpress PKCα appear to be hormone independent. However, none of the mice in the E2 group developed tumors. An independent clone, T47D:A18/PKCα5, exhibited similar growth inhibition in the presence of E2. Injection of this clone resulted in only 1 tumor from a total of 20 injection sites (10 mice) in the presence of E2 (results not shown). This observed lack of response to E2 in vivo is distinct from our previous observations in cell culture (13) .
Contrasting Actions of Estradiol in Vitro and in Vivo.
To compare the in vivo and in vitro response of our cell lines, we performed proliferation assays in the presence of E2, 4-OHT, or the pure antiestrogen, ICI 182,780. In contrast to the E2 effect in vivo, E2 acts as an agonist stimulating the growth of the T47D:A18/PKCα cells (Fig. 2B)⇓ as reported previously (13) . However, similar to the stimulatory activity of TAM in vivo, 4-OHT acts as a partial agonist alone and is unable to completely block the action of E2 in T47D:A18/neo cells (Fig. 2)⇓ . Similar results were obtained with E2 at 10−9 m combined with 4-OHT at both 10−7 and 10−6 m. The pure antiestrogen, ICI 182,780, alone completely inhibited the growth of both the neo- and PKCα-transfected cell lines, whereas it was not as effective at blocking E2 in the T47D:A18/PKCα cells.
Analysis of PKC Isozyme and ERα Expression in Tumors.
To verify that the tumors produced by injection of T47D:A18/PKCα cells into the athymic mice overexpress PKCα, we performed immunohistochemical analysis. Tumors were excised from the T47D:A18/neo E2 group and the T47D:A18/PKCα control and TAM groups. When we used a PKCα-specific polyclonal antibody, overexpression of PKCα was apparent in the T47D:A18/PKCα20 tumors grown either in the presence of TAM (Fig. 3D)⇓ or untreated (Fig. 3F)⇓ relative to the T47D:A18/neo tumor grown in the presence of E2 (Fig, 3B)⇓ . Because we previously reported that stable transfection of PKCα into T47D:A18 cells concomitantly results in up-regulation of PKC β1 and δ in cell culture (13) , we examined the expression of these PKC isozymes in the tumors as well. We find that both PKCβI and δ are up-regulated in the T47D:A18/PKCα tumors relative to the T47D:A18/neo tumors (results not shown).
To determine whether PKCα overexpression would result in down-regulation of ERα, Western blot analysis was performed with nuclear extracts from T47D:A18, T47D:A18/neo, and T47D:A18/PKCα20 tumors following various treatments (Fig. 4)⇓ . ERα expression in both T47D:A18 and T47D:A18/neo tumors grown by E2 stimulation was similar. The T47D:A18/PKCα tumors, either untreated or treated with TAM, appeared to exhibit elevated ERα expression relative to the T47D:A18 and T47D:A18/neo tumors. However, because the PKCα-overexpressing tumors did not grow in the presence of E2, it was not possible to compare E2 and TAM effects in these tumors. Therefore, we can only conclude that ERα protein expression is maintained in PKCα-overexpressing tumors. It is interesting to note that the migration of the tumor immunoreactive ERα appeared to be faster than recombinant ERα protein and the band from the T47D:A18 cell line. The reason for this is unknown; however, we are confident in the specificity of the bands because a blocking peptide was able to demonstrate competition (results not shown). Our results indicate that the T47D:A18/PKCα20 tumors were similarly overexpressing PKC isozymes and that ERα protein expression was maintained, as we reported previously in cell culture (13) .
E2 Inhibits the Growth of Established Tumors Overexpressing PKCα.
We determined whether, in addition to preventing formation of T47D:A18/PKCα tumors, E2 inhibits the growth of established tumors. In this experiment, T47D:A18/PKCα cells were injected into 20 athymic mice and were left untreated for 7 weeks, at which time the mean tumor size was ∼0.6 cm2. At 7 weeks, the mice were randomized to either continue on the untreated control arm (10 mice) or received implants of an E2 capsule (10 mice; Fig. 5A⇓ ). One week after the E2 capsules were implanted, the tumors continued to grow. However, after 2 weeks, the tumors began to regress, and at 14 weeks the mean tumor size was ∼0.2 cm2. We conclude that E2 both prevents the establishment of T47D:A18/PKCα tumors and induces regression of preexisting T47D:A18/PKCα tumors.
The Pure Antiestrogen ICI 182,780 Inhibits TAM-stimulated T47D:A18/PKCα Tumors.
The pure antiestrogen, ICI 182,780 (Faslodex) is, at present, in Phase III clinical trials as a second-line endocrine therapy following TAM failure in the clinic (20) . To determine whether ICI 182,780 was capable of controlling TAM-stimulated T47D:A18/PKCα tumors, we performed a crossover experiment. Twenty athymic mice received bilateral injections of T47D:A18/PKCα cells, and oral TAM was administered for 7 weeks until the mean tumor cross-sectional area reached 0.3 cm2. The mice were then randomized to either continue TAM treatment or were given weekly injections of ICI 182,780 (Fig. 5B)⇓ . Three weeks after randomization, mice in the ICI 182,780 treatment arm exhibited tumor stabilization, whereas the tumors in the mice in the TAM treatment arm continued to increase in size.
PKCα Is Overexpressed in TAM-stimulated and E2-inhibited MCF-7 Tumors.
To determine whether PKCα is similarly overexpressed in another model of E2-induced tumor regression, we examined PKCα expression in MCF-7 tumors displaying a cyclical model of TAM sensitivity (16) . In this model, we found that E2 had an antitumor effect on MCF-7 tumors that were serially transplanted in TAM-treated athymic mice for >5 years, and tumors exhibiting the following growth characteristics were described. Initially, MCF-7 tumors were stimulated to grow by E2 and were growth inhibited by TAM (MCF-7 E tumors). Treatment of the athymic mice with TAM for 1 year resulted in tumors that were stimulated by both E2 and TAM (MCF-7 ST tumors). After 5 years of TAM treatment, E2 caused tumor regression (MCF-7 LT tumors). Approximately 50% of the MCF-7 LT tumors regrew with E2 and reverted back to the original E2-stimulated phenotype (MCF-7 TAME tumors). The PKC isozyme expression profiles of MCF-7 E, MCF-7 ST, MCF-7 LT, and MCF-7 TAME tumors were determined by immunohistochemical analysis. Immunoreactive PKCα appeared to be more abundant in the MCF-7 ST and MCF-7 LT tumors (Fig. 6, B and C)⇓ relative to MCF-7 E, and MCF-7 TAME tumors (Fig. 6, A and D)⇓ . These results suggest that PKCα up-regulation is associated with TAM-stimulated tumor growth and E2-induced regression in both the T47D:A18 and MCF-7 tumor models.
E2-induced Apoptosis in T47D:A18/PKCα Tumors.
The TUNEL assay was used to determine whether apoptosis contributes to the E2-induced regression of the T47D:A18/PKCα tumors. Forty athymic mice received injections of T47D:A18/PKCα cells and were left untreated, which allowed the mean tumor volume to reach 0.4 cm2 at 6 weeks postinjection. At that time, a 1.0-cm E2 capsule was implanted in 20 of the mice, and the other 20 mice were left untreated. Tumors were excised from each treatment group on days 1, 2, 4, 5, 7, 8, 9, and 15 after capsule implantation, and the TUNEL assay was performed to assess apoptosis. We observed a statistically significant increase in apoptosis 5 days after E2 capsule implantation compared with control tumors (P = 0.0062), and apoptosis remained elevated through day 15 (Fig. 7)⇓ . The time frame of apoptosis corresponded with tumor regression beginning on day 5 (Fig. 7C⇓ , inset). These results indicate that apoptosis contributes to E2-induced tumor regression in vivo.
TAM-resistant breast cancer is a considerable clinical obstacle for which there are few therapeutic options. Identification of the key factors involved in the molecular mechanism of TAM resistance will undoubtedly lead to the development of logical therapeutic targets. Accumulating evidence points to the activation of the AP-1 pathway as a possible mechanism of resistance (4 , 5 , 21, 22, 23) . We have linked this observation with other pieces of information—the inverse relationship of ER and PKC (2) , the elevated PKC activity observed in malignant versus normal breast tissue (24 , 25) , and the fact that PKC is an upstream activator of AP-1 (9, 10, 11, 12 , 26)— to formulate our hypothesis that overexpression of PKC is sufficient to cause TAM-resistant (or -stimulated) breast cancer.
We have shown that overexpression of PKCα in T47D:A18 cells is sufficient to produce a hormone-independent phenotype in vitro (13) . T47D cells have previously been demonstrated to form TAM-stimulated tumors in athymic mice at an accelerated rate compared with the MCF-7 TAM-stimulated tumor model (27) . However, the hormone-responsive phenotype of the T47D:A18/PKCα20 tumors cells is distinct. T47D:A18/PKCα cells form tumors in athymic mice in the absence of E2 supplementation and form TAM-stimulated tumors immediately (within 6 weeks, Fig. 1C⇓ ) after inoculation of the cells. PKC activity and ERα expression are inversely related in breast cancer cell lines (2) . We know that ER function is impaired in the T47D:A18/PKCα20 cells in culture, as assessed by the activation of an estrogen response element-luciferase reporter plasmid (13) , but ERα protein levels are maintained in both cell culture and tumors (Fig. 4)⇓ . Because E2 treatment results in an initial down-regulation of ERα in T47D:A18 cells in culture, with a gradual return to control levels thereafter, whereas 4-OHT increases steady-state ERα levels (28) , we cannot make a direct comparison of ERα levels between T47D:A18 or T47D:A18/neo and T47D:A18/PKCα20 tumors. Evaluation of ERβ protein expression was attempted by Western blot with several commercially available ERβ antibodies, but because of inconsistency in detection, conclusions regarding ERβ expression in T47D:A18 transfectants could not be made (results not shown). It has previously been reported that antiestrogens, including TAM, are capable of activating the AP-1 pathway through ERβ (23 , 29) . T47D cells are reported to express ERβ as assessed by reverse transcription-PCR (30 , 31) ; however, the effect of PKCα overexpression on ERβ function and expression is unknown and is the focus of ongoing studies in our laboratory.
In contrast to the growth-stimulatory response of E2 in cell culture, E2 supplementation prevented T47D:A18/PKCα tumor formation and caused regression of established tumors. These results suggest that some factor present in the tumor microenvironment is responsible for the differences in hormone responsiveness of T47D:A18/PKCα observed in vitro and in vivo. Breast tumors are comprised of not only epithelial cells, but also endothelial cells, fibroblasts, macrophages, and lymphocytes. It is possible that E2 generates a growth-inhibitory response from these surrounding stromal cells that acts on the tumor cells. Because we have found that only 20% of the tumor cells are undergoing apoptosis (Fig. 7)⇓ , other mechanisms are likely to contribute to tumor regression. Therefore, at present we are examining alterations to the receptors of both apoptotic and angiogenic signals in the PKCα-overexpressing T47D cells. There are two other reports in the literature that describe E2-induced regression in a breast cancer tumor model. The first was the T61 tumor, which was derived from a primary breast cancer, initially described by Brunner et al. (32) . Similar to the T47D:A18/PKCα tumor, the T61 tumor is ovarian independent, and E2 is inhibitory. Unlike the T47D:A18/PKCα tumor, T61 is growth inhibited by TAM. Yao et al. (16) recently described a cyclical model of hormonal response in MCF-7 tumors that may have important clinical implications for the treatment of TAM-resistant tumors. Our results show that two tumors representing different exposures to TAM (1 year and 5 years) appear to overexpress PKCα, similar to the T47D:A18/PKCα20 tumor. However, the distinction between the T47D and MCF-7 tumors overexpressing PKCα is that the TAM-stimulated and E2 regression phenotypes exist in a single T47D tumor (T47D:A18/PKCα20), whereas these growth characteristics are separately exhibited in the two MCF-7 tumor models, MCF-7 ST and MCF-7 LT. These results suggest that PKCα overexpression is not sufficient to cause E2 regression in MCF-7 tumors, and therefore, during the transition from ST and LT TAM treatment, an additional molecular lesion must occur. Alternatively, perhaps the level of PKCα overexpression achieved in the T47D tumor is greater.
There are several examples of E2-induced regression described in breast cancer cell lines. Sonnenschein et al. (33) reported that the MCF-7-derived E8CASS cell variant selected after 9 months of continuous culture in charcoal-dextran stripped human serum is growth inhibited by 30 and 300 pm E2. To ascertain the mechanism of the E2-induced regression in these cells, Szelei et al. (34) prepared a subtractive library to identify genes that may be involved. Two candidate genes were identified: E9, a putative zinc-finger protein, and E43, which is homologous to the human actin-related protein 3. These proteins are thought to be involved in apoptosis and actin nucleation, respectively. Song et al. (35) also reported E2-induced apoptosis in MCF-7 long-term estrogen deprived (LTED) cells that can be abrogated by ICI 182,780, suggesting that apoptosis is mediated by the ER. T47D L(hE) cells described by Fernandez et al. (36) were cultured in E2-deficient medium long term and also exhibit E2-induced growth inhibition. However, these cells were found to express elevated ER levels with a C→A transversion resulting in a H513N amino acid change in the ligand binding domain.
Although the PKCα overexpressing tumors are TAM-stimulated, ICI 182,780 causes tumoristasis (Fig. 5B)⇓ . It is interesting to note that, based on data from Paech et al. (23) , the pure antiestrogen ICI 182,780 is expected to activate AP-1 through ERβ, yet Faslodex appears to be as effective in the clinic as a second-line endocrine therapy after TAM failure as anastrozole (37) . Therefore, it is possible that the effects of ICI 182,780 obtained with reporter constructs in vitro may not translate to the in vivo situation. Additionally, the TAM-stimulated phenotype of T47D:A18/PKCα is likely to be ER dependent because the mechanism of action of pure antiestrogens is destruction of the ER (38) . PKCα expression may be a useful predictor for the selection of patients who may benefit from a pure antiestrogen after TAM treatment failure.
PKCα overexpression results in cross-regulation of both βII and δ isoforms (results not shown; Ref. 13 ). Therefore, it is difficult to attribute the death response, or any other response, to PKCα alone in stably transfected cell lines. We are in the process of using PKCα antisense oligonucleotides to specifically down-regulate only PKCα in an attempt to determine the importance of this isozyme in the unique hormone-responsive phenotype exhibited by T47D:A18/PKCα tumors. Our finding that E2 has an antitumor effect on breast tumors overexpressing PKCα is a novel observation that may have important therapeutic implications. Diethylstilbestrol was considered to be the preferred endocrine treatment for advanced breast cancer prior to the introduction of TAM (39) and has been shown to be effective after disease progression subsequent to TAM (40) . Therefore, it may be possible to preselect patients who would benefit from diethylstilbestrol therapy after TAM treatment failure, based on high PKCα tumor expression. Alternatively, if PKCα is highly expressed in the primary tumor at diagnosis, an estrogen-like compound may be superior to TAM, based on our PKCα-overexpressing model that resulted in TAM-stimulated tumor growth. As a first step toward investigating this hypothesis, we are examining PKCα expression in biopsies from patients treated with TAM who experienced disease recurrence compared with patients who have remained disease-free with at least 5 years follow-up.
We are grateful to Dr. Sunil Badve for helpful discussions and assistance with the TUNEL assay and immunohistochemical staining, and to Dr. Alfred Rademaker for assistance with the statistical analyses.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by ACS, Illinois Division Grant 97-27 (to D. A. T.), NIH Grant RO1 CA79847 (to D. A. T.), and the Avon Products Foundation Breast Cancer Research and Care Program.
↵2 To whom requests for reprints should be addressed at Department of Pharmaceutics and Pharmacodynamics (MC865), College of Pharmacy, University of Illinois at Chicago, 833 S. Wood Street, Chicago, IL 60612-7231. Phone: (312) 996-0888; Fax: (312) 996-0098; E-mail:
↵3 The abbreviations used are: TAM, tamoxifen; ER, estrogen receptor; PKC, protein kinase C; AP-1, activator protein-1; TPA, 12-O-tetradecanoylphorbol-13-acetate; E2, 17β-estradiol; FBS, fetal bovine serum; 4-OHT, 4-hydroxytamoxifen; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; POD, peroxidase.
- Received May 7, 2001.
- Revision received June 22, 2001.
- Accepted June 22, 2001.