Transforming Growth Factor-β, Estrogen, and Progesterone Converge on the Regulation of p27^Kip1 in the Normal and Malignant Endometrium

Introduction

Transforming growth factor-β (TGF-β), a potent inhibitor of cell growth, functions paradoxically in many human cancers ( 1, 2). Whereas disruption of TGF-β signaling is a means by which cancer cells subvert growth inhibition, similar TGF-β–mediated signaling pathways have been shown to mediate metastasis later in malignant progression. Thus, although inhibitors of TGF-β signaling may have potential therapeutic value to prevent metastasis, the initial loss of TGF-β–mediated growth inhibition is related to the development of many human cancers and thus, components of the TGF-β signaling pathway are considered as tumor suppressors ( 3). Related to this important function of TGF-β, we showed in a previous report that disruption of TGF-β signaling is an early event leading to loss of TGF-β–mediated growth inhibition in primary cultures of carcinoma cells isolated from type I endometrial cancer tissue, suggesting a role in the pathogenesis of endometrial carcinoma (ECA; ref. 4). ECA is the most common gynecologic malignancy. Endometrioid type I occurs at a rate of 85% and is caused by unopposed estrogen, which induces hyperplasia, often a precursor to carcinoma in this cancer ( 5– 7). The growth of endometrial glandular epithelium is responsive to many growth factors and to gonadal steroids in a cyclical manner. Normally, within the menstrual cycle, estrogen stimulates proliferation of endometrial epithelium in the (follicular) proliferative phase, and progesterone, derived from the corpus luteum, inhibits epithelial growth and induces glandular differentiation in the (luteal) secretory phase. Accordingly, progestins are used to treat both endometrial hyperplasia and carcinoma ( 8).

Cellular proliferation is accomplished by different cell cycle phase–dependent cyclins complexed to specific cyclin-dependent kinases (cdk) that progressively and additively phosphorylate Rb (and other pocket proteins, p130 and p107) for cell cycle progression ( 9), dictated by a complex network of interactions of pocket proteins with E2F transcription factors and specific gene activation ( 10, 11). TGF-β decreases specific cdks and increases certain cdk inhibitors to maintain Rb in a hypophosphorylated state to block cell cycle progression in late G₁ phase. The regulation of TGF-β activity is cell type and context dependent and classically involves the cooperation of two serine/threonine kinase receptors, TβRI and TβRII, which directly phosphorylate Smad2 and Smad3, which are transcription factors for activation of genes involved in growth inhibition and other TGF-β–related functions ( 3, 12). p27, originally isolated from TGF-β–treated growth arrested cells ( 13), has a dual role in regulating the cell cycle as it is critical to the formation of cyclin D/cdk2 complexes for phosphorylation of Rb but prevents further phosphorylation of Rb by binding to cyclin E/cdk2 and cyclin A/cdk2 for cell cycle arrest in G₁ ( 14). Intracellular levels of p27 are highly regulated by posttranslational mechanisms that control its ubiquitin-mediated degradation and nuclear exclusion ( 15– 20). The significance of maintaining sufficient nuclear levels of p27 to block cdk2 for normal growth inhibition is underscored by the complexity of cellular mechanisms and pathways affecting p27 levels that can go awry to promote the growth of cancer cells ( 16, 17). Whereas there is no compelling evidence for p27 as a classic tumor suppressor, genetically null mice develop pituitary hyperplasia and are predisposed to tumor formation by DNA-damaging agents ( 21). Importantly, mutations in the p27 gene are rare in human cancers ( 22). Instead, decreased nuclear levels of p27 by proteasomal degradation and cytoplasmic mislocalization of p27 are found in many human cancers ( 16, 17, 20, 23, 24).

Both 17β-estradiol (E₂) and progestins, as well as their receptor agonists and antagonists, regulate the growth of female hormone-regulated organs. Furthermore, in vivo and in vitro data show that E₂ plays a critical role in the development of breast cancer and unopposed E₂ is touted as the sole etiologic agent in type I ECA ( 6, 7). Both estrogens and progestins have been shown to affect the levels, cellular localization, and cell cycle binding partners of p27 ( 25, 26) and the association of cdk inhibitors with cyclin/cdk complexes in breast cancer cells ( 27). However, little is known about the mechanisms involved in hormone effects on p27 in the endometrium.

Because p27 blocks cell cycle progression and is isolated from TGF-β–treated cells, we sought to determine whether TGF-β signaling could directly affect the levels of p27 in normal endometrial cells and whether the lack of p27 induction resulting from disrupted TGF-β signaling may be a mechanism involved in endometrial carcinogenesis. We show herein that whereas TGF-β does not affect mRNA levels in both endometrial tissue–derived primary cultures of normal endometrial epithelial (EECs) and carcinoma (ECA) cells, it induces a dose-dependent accumulation of p27 protein in the normal cells in both the cytoplasm and nucleus, with a 2-fold greater increase in the nuclear fraction. However, p27 protein is negligible and not induced by TGF-β in the carcinoma cells lacking TGF-β signaling. In addition, malignant endometrial tissue lysates rapidly degrade exogenously added p27 via the ubiquitin-proteasome pathway. These results suggest that both loss of TGF-β signaling and increased proteasomal degradation of p27 are important molecular events involved in pathologic growth of the endometrium. We further show that addition of E₂ to normal EECs induces mitogen-activated protein kinase (MAPK)–driven ubiquitin-mediated proteasomal degradation of p27 and that progesterone increases p27 protein levels in both normal and malignant cells. Taken together, our results suggest that the effects of TGF-β and ovarian hormones converge on the regulation of the levels of the cdk inhibitor p27 to control endometrial cell growth.

Materials and Methods

Isolation and culture of EECs and ECA cells. Fresh endometrial tissues from type I endometrioid tumors derived from hysterectomies (women ages 35–55 years) at New York University-Tisch and Bellevue Hospitals were placed in iced phenol red–free McCoy's 5A medium (Sigma-Aldrich, Inc., St. Louis, MO) and primary cell cultures initiated within 4 h. Endometrial hyperplasia and carcinoma (ECA) samples were identified and graded by at least two surgical pathologists according to the WHO system [grades I–III, for decreased glandular architecture (Fédération Internationale des Gynaecologistes et Obstetristes system: upper number) and for nuclear atypia (lower number); ref. 28]. Normal endometrial tissue [proliferative phase (PE) and secretory phase (SE)] was obtained from non-cancer-related hysterectomies, generally leiomyomas. Tissue lysates were prepared by freezing and thawing or by disruption with a nitrogen bomb (Parr, Moline, IL), and tissue samples were fixed in 10% buffered formalin for 24 h and embedded in paraffin for immunohistochemistry. All studies involving human subjects were approved by the New York University School of Medicine Institutional Review Board (Assurance of Compliance no. M1177-01). To prepare primary cell cultures, EECs and ECA cells were isolated as described ( 4). The EECs and ECA cells, seeded at 3 × 10⁶ to 4 × 10⁶ per 60-mm Primaria tissue culture plate (Falcon, Lincoln Park, NJ), were grown in McCoy's 5A medium for 24 h or to ∼70% confluence and switched to serum-free medium followed by addition of the respective treatments described below.

Analysis of p27^Kip1 mRNA by reverse transcription-PCR. To serum-free primary cell cultures, recombinant TGF-β1 (R&D Systems, Minneapolis, MN) at 0, 10, and 100 pmol/L was added every other day for 4 days; the cells were washed with HBSS (Mediatech, Inc., Hernden, VA); and total RNA was isolated using RNAZol (Tel-Test, Inc., Friendswood, TX). cDNA was synthesized from 2.0 μg of total RNA using the SuperScript II Reverse Transcriptase Kit (Invitrogen, Carlsbad, CA). To amplify human p27 mRNA, the cDNA corresponding to 20 and 200 ng of total RNA was used with p27 primers (accession no. U10906). The integrity of the RNA was compared with reverse transcription-PCR (RT-PCR) for β-actin (accession no. NM 001101) and the PCR products were confirmed by sequencing.

Cell treatments and Western blot analysis. The serum-free primary cell cultures were treated in individual experiments with TGF-β (0, 10, and 100 pmol/L), 100 nmol/L E₂ (Sigma), 20 μmol/L PD98059 (Calbiochem, La Jolla, CA), and 100 nmol/L medroxyprogesterone 17-acetate [progesterone (Pg); Sigma], 100 nmol/L ICI 182,780 (Zeneca Limited, Wilmington, DE) and 10 μmol/L mifepristone (RU-486; #8046, Sigma) every other day for 4 days. In certain experiments, the cells were treated with TGF-β for 24 h and 10 μg/mL cycloheximide (Sigma) was added 1 h before harvesting the cells. In specified E₂-treated cell cultures, the proteasome inhibitor lactacystin (1–10 μmol/L; Calbiochem) was added 2.5 h before ending the experiment. Experiments were terminated with radioimmunoprecipitation assay (RIPA) lysis buffer containing a cocktail of protease inhibitors (Sigma) and proteins extracted as described in the legend to Fig. 1B . Where noted, nuclear and cytoplasmic fractions were prepared from TGF-β–treated cells using the NE-PER kit (Pierce Biotechnology, Inc., Rockford, IL). Protein concentrations of all cellular supernatant preparations were determined using the DC Protein Assay kit (Bio-Rad Laboratories, Hercules, CA) and equal amounts of protein (5–40 μg) were reduced in Laemmli sample buffer with 0.1 mol/L DTT (Bio-Rad), separated by 5% to 20% SDS-PAGE, and then transferred to Hybond polyvinylidene difluoride (PVDF) membranes (Amersham Biosciences, Piscataway, NJ). The procedures used for individual Western blots are described in the figure legends. The blots were exposed to BioMax X-ray film (Eastman Kodak, Rochester, NY) for protein detection. To quantify the protein bands, densitometry was done using an EDAS 290 and Kodak 1D Image Analysis software (Eastman Kodak).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Analysis of p27 mRNA, protein levels, and subcellular distribution following TGF-β treatment of EECs and ECA cells. Primary EECs and ECA cells were isolated from fresh endometrial tissue (PE, SE, inactive endometrium, hyperplasia, and ECA) and seeded onto Primaria 60-mm plates. Twenty-four hours later or when the cells reached 60% to 70% confluency, the cells were incubated in McCoy's 5A medium containing 2% charcoal-stripped FBS or serum-free media and treated with recombinant TGF-β1 (0–100 pmol/L) every other day for 4 d. A, TGF-β did not alter basal expression of p27 mRNA levels in EECs or ECA cells. Total RNA was isolated from the cells as described in Materials and Methods and p27 mRNA levels were analyzed by semiquantitative RT-PCR. Amplification of p27 mRNA was done with the following primers and method: p27 primers (accession no. U10906), 5′-CTTGCCCGAGTTCTACTACAGAC-3′ (forward) and 5′-CAAATGCGTGTCCTCAGAGTTAG-3′ (reverse). The integrity of the RNA was confirmed by amplification of human β-actin mRNA (accession no. NM 001101) with the following oligonucleotides: 5′-ATCATGTTTGAGACCTTCAA-3′ (forward) and 5′-CATCTCTTGCTCGAAGTCCA-3′ (reverse). After denaturation for 2 min at 95°C, the cycle profile consisted of 30 s at 94°C, 1 min at 60°C, and 2 min at 72°C for 35 cycles. The primers yield a 142-bp PCR product. p27 mRNA levels are compared with β-actin as the housekeeping gene. The primary cells were derived from PE (n = 2), SE (n = 2), and ECA (n = 2 grade I/II and n = 2 grade II/II). B, TGF-β induces a dose-dependent increase in intracellular p27 protein in normal EECs but not in ECA cells. Proteins were extracted from cell cultures with RIPA lysis buffer containing 20 mmol/L Tris-HCl (pH 8.0), 0.137 mol/L NaCl, 10% glycerol, 0.5% NP40, 0.1% sodium deoxycholate, and 3.0 mmol/L EDTA. A mixture of protease inhibitors was added to final concentrations of 1.0 mmol/L PMSF, 0.15 μg/mL aprotinin, 10 μg/mL leupeptin, 10 μg/mL pepstatin, and 5 mmol/L sodium orthovanadate (all from Sigma). The cell cultures were rocked on ice for 10 min followed by centrifugation at 14,000 × g for 15 min at 4°C to clarify the supernatants. Equal protein concentrations (30 μg) in Laemmli sample buffer containing 0.1 mol/L DTT were loaded onto SDS-PAGE. The proteins were electrotransferred to Hybond PVDF membranes for 1 h and the membranes stained with 0.2% Ponceau S (Sigma) to re-ensure equal loading and transfer of proteins. The blots were treated with the Qentix Western Blot Signal Enhancer to increase the sensitivity of immunoblotting, blocked with 5% nonfat dry milk in PBS for 1 h, and incubated with a mouse monoclonal antibody to p27 (1:400; BD Biosciences) in blocking buffer overnight at 4°C. Finally, the blots were incubated with horseradish peroxidase (HRP)–conjugated goat anti-mouse secondary antibody in PBS containing 0.1% Tween 20 (Fisher Scientific, Pittsburgh, PA). The SuperSignal WestDura Extended Duration Substrate kit (Pierce) was used, followed by exposure to X-ray film to detect the protein bands. To ensure equal loading and normalize sample loads for the estimation of protein levels from densitometry, the blots were stripped with 0.5 mmol/L Tris-HCl (pH 6.7), 100 mmol/L DTT, and 2% SDS at 50°C for 30 min; incubated in blocking buffer composed of 5% nonfat dry milk in TBS containing 0.1% Tween 20 (TBST) for 1 h; and then reprobed with a mouse anti–β-actin antibody in 2.5% nonfat dry milk containing TBST (1:12,000; clone AC15, Sigma) followed by horseradish peroxidase–conjugated goat antimouse secondary antibody (1:2,000; Pierce) in TBST containing 2.5% nonfat dry milk for 1 h. Densitometry was done on each β-actin and p27 band and the levels of p27 were quantified following normalizing the intensity of the p27 signal to the β-actin signal. The graphs (right) show the protein levels of each blot, represented as fold change from the control at 0 (Y axis). Two representative samples of normal endometrium from SE (Normal #1) and from PE (Normal #2), one from hyperplastic tissue, and two representative samples from ECA (ECA#1 and ECA#2), both grade I/III, are shown. The graph (right) shows the average fold induction from the blots shown of the two normal samples. The primary EECs used were PE (n = 4) and SE (n = 6). One sample of EECs from inactive endometrium showed similar results. All primary cultures derived from normal endometrium show a dose-dependent response to TGF-β. In contrast, the one primary culture from hyperplasia and all the ECA primary cells from patients representing different grades of cancer [i.e., ECA grade I/II (n = 2), ECA grade II/II (n = 2), and ECA grade I/III (n = 2)] have scant to negligible basal levels of p27 and are not induced by TGF-β. C, TGF-β induces p27 in both the cytoplasm and nucleus with a greater increase in the nuclear fraction compared at 100 pmol/L TGF-β. Primary cultures of normal EECs were treated with TGF-β as described above and nuclear cytoplasmic fractions prepared using the NE-PER kit. Following scraping of the cells from plates on ice, the cells were centrifuged at 300 × g for 5 min to obtain cell pellets to which the kit reagents were added for subcellular fractionation. The extracted proteins (5.0 μg of nuclear fractions and 15 μg of cytoplasmic fractions) were probed for p27 levels by immunoblotting and the blots stripped and reprobed for β-actin, as described in (B). To ensure purity of the nuclear and cytoplasmic fractions, the blots were stripped again and reprobed with a mouse anti–α-tubulin antibody (1:12,000; clone B-5-1-2, Sigma) in blocking buffer overnight 4°C, as a marker of the cytoplasmic fraction, followed by streptavidin horseradish peroxidase (HRP)–conjugated goat antimouse secondary antibody (Pierce) for 1 h. The blots were also probed with a rabbit anti–Sp-1 polyclonal antibody (1:12,000; H-225, Santa Cruz Biotechnology, Santa Cruz, CA) followed by HRP-conjugated goat antirabbit antiserum as a marker of the nuclear fraction (data not shown). The levels of p27, β-actin, and α-tubulin proteins were determined as described in (B). The levels of p27 were normalized to the values obtained for β-actin. The lack of α-tubulin in the nuclear fraction indicates the purity of the subcellular fractionation. The graph (right) depicts the densitometric quantification of the p27 protein bands normalized to β-actin, represented as fold induction; dotted line, comparison of the nuclear and cytoplasmic fractions at 100 pmol/L TGF-β. TGF-β induction of p27 was observed in both nuclear and cytoplasmic fractions but 2-fold higher in the nuclear fraction, at 100 pmol/L, in the primary EEC culture from SE shown here. Normal EECs used for these experiments were SE (n = 2) and PE (n = 1). D, novel protein synthesis is required for up-regulation of p27 in response to TGF-β. EECs were treated with 100 pmol/L TGF-β for 24 h and cycloheximide (CHX) was added 1 h before termination of the experiment. Cell lysates were prepared and 20 μg of protein were analyzed by Western blotting as described in (B). A representative sample of EECs from SE. Similar results were obtained from EECs derived from SE (n = 2) and inactive endometrium (n = 1). The levels of p27 were determined as described in (B) and are depicted in the graph (right) as fold induction.

Immunohistochemical analysis for p27^Kip1. Tissue sections (4 μm), cut from newly embedded or archival paraffin blocks of normal endometrium (PE and SE), hyperplastic endometrium, and different grades of ECA were obtained from the New York University and Bellevue Tumor Registry. Immunohistochemistry was done with a monoclonal antibody to p27 (0.5 μg/mL; BD Biosciences, San Jose, CA) as previously described ( 4) and according to instructions from the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA), except antigen retrieval was required to expose p27 epitopes for immunoreactivity. The intensity of nuclear and/or cytoplasmic immunostaining for the glandular epithelium and stromal cells were assessed by three pathologists and scored (immunostaining levels: 0, none; 1, weak; 2, moderate; 3, intense). The values for intensity of immunostaining were determined and outcome was measured as the product of average staining intensity and the percentage of cells stained at the predominant intensity (usually 100%). The values for intensity of immunostaining for each category were determined using the Kruskal-Wallis nonparametric ANOVA test and P values were estimated using Dunn's multiple comparison test.

Proteasome degradation assay. Endometrial tissue lysates were prepared from 50 to 100 mg of pulverized frozen tissue by freezing in liquid nitrogen and subsequent thawing on ice in 50 mmol/L Tris (pH 8.3), 5 mmol/L MgCl₂, 1 mmol/L DTT containing 100 μg/mL aprotinin, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF), and 10 μg/mL leupeptin, three separate times. The lysates were clarified at 10,000 × g for 10 min; equal amounts of protein (75–100 μg) were incubated with 1 μmol/L histidine–tagged recombinant p27 (rp27; kindly provided by M. Pagano, New York University School of Medicine, NY; ref. 29) in a degradation buffer mixture; and the experiments were done as described in the legend to Fig. 3A. The proteasomes were removed from ECA tissue lysates and hemin (Sigma) was used to block proteasomes in certain experiments.

Ubiquitylation of p27^Kip1. Tissue preparation and ubiquitylation of p27 were done as described ( 29, 30). Briefly, endometrial tissue (50–75 mg) was disrupted in lysis buffer, containing 20 mmol/L Tris-HCl (pH 7.2), 2 mmol/L DTT, 0.25 mmol/L EDTA, and 10 μg/mL each of leupeptin and pepstatin, using a prechilled Parr cell nitrogen bomb apparatus at 1,000 psi on ice for 30 min; lysates were clarified by centrifugation at 10,000 × g for 10 min. rp27 was added to the reaction mixture as the substrate for p27 ubiquitylation and the experiments were done as described in the legend to Fig. 3D.

Results

TGF-β induces nuclear and cytoplasmic p27^Kip1 protein accumulation in normal EECs but not in ECA cells. The effect of TGF-β on p27 mRNA and protein levels in primary cultures of cells derived from normal, hyperplastic, and ECA tissues was determined using semiquantitative RT-PCR and Western blotting, respectively. Normal EECs from both PE and SE and carcinoma cells from all grades of ECA contain basal levels of p27 mRNA, and none show induction of p27 by 100 pmol/L TGF-β ( Fig. 1A). Whereas the basal levels of p27 protein were variable among the primary cultures and, generally, p27 was higher in SE, both the SE (#1) and PE (#2) primary cultures shown here contain moderate basal levels of p27 protein ( Fig. 1B). TGF-β at 10 and 100 pmol/L induced an average 3- and 3.6-fold increase, respectively, in p27 protein levels in EECs derived from normal SE (#1) and normal PE (#2) compared with an untreated sample ( Fig. 1B). A similar response was obtained in the presence of 2% charcoal-stripped fetal bovine serum (FBS; data not shown). In addition, EECs from one inactive (postmenopausal) endometrial sample responded to TGF-β (data not shown). By contrast, TGF-β did not increase p27 protein in all grades of ECA assessed, which generally had none to very low basal levels, as shown here in two grade I/III carcinomas ( Fig. 1B, ECA #1 and #2). Similarly, p27 was not increased by TGF-β in EECs from one sample of endometrial hyperplasia ( Fig. 1B).

Following subcellular fractionation of normal EECs from both PE and SE, 78% of basal p27 levels is localized within the nuclear fraction ( Fig. 1C, graph). However, the TGF-β–induced increase in p27 occurs in both the cytoplasmic and nuclear fractions (7.5- and 3.5-fold over basal levels, respectively) with a 2-fold greater increase in the nucleus at 100 pmol/L TGF-β ( Fig. 1C, graph). The accumulation of p27 partly requires novel protein synthesis because a 67% reduction (3.8- to 1.26-fold) in p27 protein levels is obtained when cycloheximide is added 1 h before the termination of a 24-h incubation with TGF-β ( Fig. 1D).

p27^Kip1 expression is decreased in ECA cells in vivo. Consistent with the negligible levels of p27 in primary cultures of ECA cells ( Fig. 1B), we previously showed a paucity of p27 by immunohistochemistry in the malignant glands in ECA tissue compared with normal ( 31). To determine whether p27 expression differed among various grades of ECA, in premalignant hyperplasia, and in normal PE compared with SE, tissue samples were analyzed by immunohistochemistry. As illustrated ( Fig. 2 ), immunoreactivity ranged from none to slight cytoplasmic staining with some nuclei staining in the glands of PE ( Fig. 2A-a). In contrast, the glands of SE show strong predominantly nuclear p27 immunoreactivity ( Fig. 2A-b, arrows). The pathology of endometrial cancer is characterized by proliferation of the epithelium with loss of glandular architecture, consequential crowding, and eventual obliteration of the stromal cells with increasing grade. Slight cytoplasmic immunostaining for p27 was shown in hyperplasia (simple without atypia shown here; Fig. 2A-c). In marked contrast to the intense p27 immunoreactivity observed in normal SE ( Fig. 2A-b), there is complete absence of p27 expression in the crowded malignant glands of ECA tissues grade I/I and grade III/II ( Fig. 2A-d), showing complete loss of glandular structure and near absence of stromal cells. In all hyperplasias examined (n = 7; including simple, complex without atypia, and complex with atypia histology), p27 was localized to the cytoplasm; the two lowest-grade hyperplasias (i.e., simple without atypia) showed nuclear staining as well. The intensity of p27 immunostaining and number of immunoreactive stromal cells varied among the normal and malignant tissue samples. The tissue samples were ranked by intensity of p27 immunostaining, shown in the box plots ( Fig. 2B). A statistically significant decrease in p27 immunoreactivity was shown between ECA and SE (P ≤ 0.001) and between PE and SE (P ≤ 0.02). Importantly, a Spearman rank correlation value of 0.54 (P ≤ 0.0002) was obtained between p27 and Smad2P immunoreactivity (the latter values were obtained from the same patients in a previous study; ref. 4), providing a statistically significant correlation between loss of TGF-β/Smad2 signaling and p27 expression in ECA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Immunohistochemical analysis of p27 expression in normal endometrium and type I ECA. A, proliferative phase endometrial glands and ECA tissues show little to no immunoreactivity with anti-p27. Tissues on slides from normal endometrium, endometrial hyperplasia, and ECA were subjected to antigen retrieval and immunostained with a monoclonal antibody to p27 using the Vectastain Elite ABC Kit, as described in Materials and Methods. Briefly, the slides were microwaved (Sharp Electronics, Co., Mahwah, NJ) in Citra solution (BioGenex, San Ramon, CA) and boiled gently for 20 min. After cooling, the slides were placed in blocking buffer containing goat serum for 20 min and then incubated with anti-p27 (0.5 μg/mL; BD Biosciences) in blocking buffer overnight at 4°C. Preimmune mouse serum (0.5 μg/mL; Lipshaw Immunon, Pittsburgh, PA) was used as a negative control. Goat antimouse secondary antibody and avidin-biotin complex were applied, as described, and the brown color reaction was developed using the substrate 3,3-diaminobenzidine tetrahydrochloride (0.5%; Acros Organics, Morris Plains, NJ). The slides were counterstained with Gill's #2 hematoxylin (Fisher). a, PE; b, SE; c, simple hyperplasia without atypia; d, ECA grade I/I and grade III/II. Original magnification, ×100 (a–c); ×40 (d). G, gland; S, stroma; Ca, cancer tissue; b, arrows, nucleus in SE glands. B, box plot analyses reflecting comparative intensity of p27 immunostaining in normal PE and SE, hyperplasia, and ECA. The samples were scored as described in Materials and Methods. Because no difference was found in the intensity of immunostaining among all grades of ECA, they were grouped together. The values for intensity of immunostaining for each category (PE, SE, hyperplasia, and ECA) were subjected to the Kruskal-Wallis nonparametric ANOVA test to determine the mean of the ranks. Dunn's multiple comparison test (GraphPad Software, San Diego, CA) was used to estimate P values for statistical significance of the difference in rank sum between each category. The means were considered statistically significant at P ≤ 0.05. Red bars, median value for each sample category. The number of samples for each category is shown in the graph. **, P ≤ 0.001, ECA versus SE; *, P ≤ 0.02, PE versus SE. A nonparametric Spearman rank correlation analysis between the mean values of the ranks, obtained for Smad2P and p27 intensity of immunostaining, was done and a two-tailed P value was calculated from the magnitude of correlation (GraphPad Software). The values for Smad2P [antibody to phosphorylated Smad2 (transcription factor for TGF-β signaling direct phosphorylation/activation by the TGF-β receptor type I) indicating TGF-β signaling and bioactivity through the Smad pathway] were derived from a previous study using identical patients as used here for p27 immunohistochemistry ( 4).

ECA tissue lysates ubiquitylate and degrade exogenous rp27. p27 levels are subject to regulation via the ubiquitin-proteasome pathway ( 16). Exogenous rp27 was previously shown to be ubiquitylated and degraded by malignant cell and tissue lysates using specific assays ( 29, 30, 32). We used these assays to determine whether the absence of p27 in the malignant glands in vivo ( Fig. 2) and in primary ECA cells in vitro ( Fig. 1B) was associated with increased degradation of p27 through the ubiquitin-proteasome pathway. Following incubation of rp27 with tissue lysates from SE, 89% of the rp27 added remains after 20 h, whereas 58% and only 4% of rp27 remain in the ECA tissue lysates after 30 min and 3 h, respectively ( Fig. 3A ). A moderate decrease in rp27 is obtained with tissue lysates from PE (data not shown). Of note, the multiple upper bands above M_r 27 kDa suggested that ubiquitylated rp27 is being detected by anti-p27 and that rp27 is most likely degraded via the ubiquitin-proteasome pathway. Following depletion of the proteasome fraction, 72% of rp27 remains intact at 6 h compared with only 6% remaining after 3-h incubation with an ECA lysate (grade I/II) containing proteasomes ( Fig. 3B). Similarly, replacing ECA proteasomes with those from SE slowed p27 degradation rates compared with that of normal SE (data not shown). The bands above 27 kDa (ECA proteasome depleted; Fig. 3B) most likely represent ubiquitylated rp27 that is not degraded because the proteasomes have been removed. This is further substantiated by the use of the proteasome inhibitor hemin, which partially blocked rp27 degradation by the ECA lysates (grade I/II shown), such that 83% of rp27 is retained in the tumor cell lysate at 3 h compared with 37% in the untreated sample ( Fig. 3C). Although these experiments were done in the presence of protease inhibitors, degradation of rp27 is still evident in the hemin-treated sample. Thus, it seems that other proteases also degraded rp27 in these tissue lysates. The addition of MG132, another inhibitor of proteasomes, similarly blocked degradation of rp27, albeit also not completely (data not shown). rp27 remained intact in two normal SE and one slightly degraded rp27. Moderate degradation of rp27 occurred in two of two lysates from PE. These results may reflect patient variability or rates of rp27 degradation by tissues derived from different days within the phases of the menstrual cycle. In contrast, six of seven ECA tissue lysates, representing all grades of ECA, were able to degrade rp27. The tumor (grade III/III) that did not degrade rp27 concordantly showed moderate expression of p27 by immunohistochemistry (data not shown). Importantly, there was complete correlation between the rate of degradation of rp27 and the levels of p27 observed in the glands in vivo by immunohistochemistry in all tissue samples of both normal and malignant tissue (n = 12). Our data suggest that the loss of p27 in vivo and the degradation of exogenously added rp27 are via ubiquitin-mediated proteasomal degradation. We interpret the differences in ubiquitylated rp27 (slower migrating protein bands detected by anti-p27) among the ECA tissues used in Fig. 3A and B to reflect patient variability and different rates of proteasome degradation of ubiquitylated rp27.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Effect of endometrial tissue extracts on the degradation and ubiquitylation of exogenously added rp27. Tissue lysates of normal and malignant endometrium were incubated with purified histidine-tagged rp27, using a previously described assay designed to detect the capacity for degradation of p27 by cancer tissues ( 32). Briefly, clarified cell lysates were prepared as described in Materials and Methods and equal amounts of protein (75–100 μg) were incubated with 1 μmol/L histidine–tagged rp27 in 30 μL degradation buffer [50 mmol/L Tris-HCl (pH 8.3), 5 mmol/L MgCl_2, 1 mmol/L DTT, 2 mmol/L ATP] containing a cocktail of protease inhibitors (Sigma), as described by Pagano ( 30, 32). The reaction was terminated by removing 5.0 μL (12–16 μg) aliquots into sample buffer at the times indicated and the samples were either flash-frozen in liquid N₂ or subjected to SDS-PAGE followed by electrotransfer onto Hybond PVDF membranes for 1 h. The membranes were stained with Ponceau S and then blocked with 5% nonfat dry milk in PBS overnight at 4°C, followed by incubation with a rabbit polyclonal antibody to human p27 (1:200; Santa Cruz Biotechnology) in blocking buffer for 2 h and then with HRP-conjugated donkey antirabbit secondary antibody (1:2,000) in PBS containing 0.1% Tween 20. The graphs (right) derived from densitometry of the blots represent percent of rp27 remaining based on relative intensity (Y axis) of each band over time compared with the sample relative intensity at 1 min following addition of recombinant rp27. A, tissue lysates from ECA degrade rp27 at a more rapid rate than normal endometrial tissue. The rate of rp27 degradation in a tissue lysate from ECA (right) is compared with that in a tissue lysate from SE (left). The patient samples used were as follows: SE (n = 3), PE (n = 2), and ECA (n = 7). The bands migrating slower than p27 are presumably p27 ligated with different numbers of ubiquitin molecules; this varied among the patient tissue samples. B, removal of the proteasome fraction from ECA lysates increased the amount of exogenous rp27 that remained intact. The proteasome fraction was removed from ECA samples by ultracentrifugation at 100,000 × g for 6 h in the presence of 5 mmol/L MgCl₂, before adding rp27 to the reaction mixtures. The patient samples used were as follows: ECA grade I/II (shown) and ECA grade II/II (n = 2). C, hemin partially blocks the ability of ECA lysates to reduce the levels of exogenously added rp27. Hemin (100 μmol/L), a physical blocker of entry of protein into the proteasome for subsequent degradation, was added to the tissue lysates before adding rp27 and the levels of rp27 remaining were analyzed by densitometry of the blots. An ECA grade I/II tissue lysate is shown. Tissue lysates from different ECA grades were used (n = 5). As a control, an equal amount of rp27 was added to the reaction mixtures at the onset of the experiments (B and C, right lane). D, tissue lysates from different grades of ECA ubiquitylate exogenous rp27 whereas lysates from SE lack this ability. The in vitro ubiquitylation assay using tissue lysates was previously described ( 29, 32). Normal and ECA tissue lysates (40 μg) in a final volume of 10 μL were incubated at 30°C in a reaction mixture containing the following reagents: 1 μmol/L rp27, 40 mmol/L Tris-HCl (pH 7.6), 5 mmol/L MgCl₂, 1 mmol/L DTT, 0.5 mmol/L ATP, 10 mmol/L creatinine phosphate, 0.1 μg/mL creatinine kinase, 1 μmol/L ubiquitin aldehyde (inhibits isopeptidase activity), 1.0 mg/mL methyl ubiquitin (terminates ubiquitin chains for distinct band visualization), 1 μmol/L lactacystin (proteasome blocker; used to block the degradation of ubiquitylated p27), 1 μmol/L okadaic acid (to maintain phosphorylated p27 for ubiquitylation), 10% glycerol, and 5 μmol/L biotinylated ubiquitin, prepared using EZ-link Sulfo-NHS-LC-biotin (Pierce; to provide an excess of ubiquitin). Aliquots were removed at time 0 and 75 min to an equal volume of 2× sample buffer and the ubiquitylated rp27 was analyzed by immunoblotting, as described in (A). The protein constituents that enable the ubiquitylation of p27 are provided by the tissue lysates used in the in vitro ubiquitylation assay ( 29) and the reagents of the reaction mixture stabilize the intermediate ubiquitylated molecules to enable observation of p27 by immunoblotting, which are shown as a ladder of protein bands that migrate slower than native rp27 protein. HeLa cell lysate (20 μg), previously shown to ubiquitylate p27 ( 29), was used as a positive control. The different grades of ECA lysates (left) are shown above the blot. ECA tissue lysates of various grades yielded similar results (n = 12). From the molecular mass increase of one band of slower mobility than M_r 27 kDa, it seems that only one ubiquitin molecule (M_r ∼9 kDa) was conjugated to rp27 (upper arrow) in the reaction mixture. Similarly, lysates from normal PE ubiquitylated rp27 (data not shown; n = 3). In contrast, lysates from normal SE (right) fail to ubiquitylate rp27, as shown by the blot of the two SE tissue lysates (n = 3). As shown here, other proteases degraded rp27 at 75 min in these two SE tissue lysates as protease inhibitors were not included in this assay.

By using an assay/reaction mixture devised to preserve ubiquitylated proteins, we show that p27 is indeed ubiquitylated by endometrial tissue lysates shown to degrade p27 ( 29, 30). As illustrated ( Fig. 3D), tissue lysates from patients representing different grades of ECA consistently ubiquitylate rp27 in vitro after 75-min incubation. In contrast, SE tissue lysates do not ubiquitylate rp27. As protease inhibitors were not included in this reaction mixture, it seems that certain proteases may have reduced the quantity of intact protein after 75 min. Furthermore, in agreement with the lower mean expression level of p27 in vivo in PE compared with SE ( Fig. 2), PE lysates ubiquitylate rp27 (data not shown).

Estrogen induces MAPK-dependent proteasomal degradation of p27^Kip1 in primary cultures of normal EECs. Estrogens have been shown to induce genes involved in cell growth through activation of the MAPK pathway ( 33). As MAPK has been shown to phosphorylate p27 at Thr¹⁸⁷ ( 26, 34), we hypothesized that estrogen might decrease p27 levels to enable proliferation of normal EECs by activating MAPK to phosphorylate p27 for subsequent ubiquitin-mediated degradation. As illustrated ( Fig. 4A ), addition of E₂ to EECs, which contain abundant levels of endogenous p27, reduces endogenous p27 levels by 2.4-fold in normal tissue #1 and by 1.4-fold in normal tissue #2, which are restored and enhanced by 1.5- and 1.2-fold, respectively, over the untreated control, in the presence of the MAPK inhibitor PD98059. In addition, as shown in the blot of normal tissue #2, intracellular levels of p27 were restored to untreated control levels by the estrogen receptor antagonist ICI 182,780. In contrast, cultures of ECA cells with negligible levels of p27 were unaffected by E₂; blocking the estrogen receptor increased p27 levels by 2.4-fold ( Fig. 4B) in this sample. Furthermore, intracellular p27 is preserved by increasing doses of lactacystin in the presence of E₂ (100 nmol/L), suggesting that levels of p27 are attenuated by E₂ via proteasomal degradation ( Fig. 4C). Similar results were obtained with MG132 (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Analysis of intracellular p27 protein levels following treatment of primary normal EECs and ECA cells with E₂. EECs, isolated from normal endometrium containing abundant p27 in vitro and in vivo, as shown in Figs. 1 and 2, respectively, were incubated with E₂ (100 nmol/L) or E₂ plus the MEK inhibitor PD98059 (20 μmol/L; all reagents were added every other day) and the experiments terminated with RIPA buffer. The cell lysates (30–40 μg) were immunoblotted to determine endogenous levels of p27, as described in Fig. 1B. The blots were subjected to densitometry to quantitate the amount of p27 in the treated sample versus the untreated control and the fold change was determined following normalization of the sample load to β-actin. A, estrogen decreases the level of endogenous p27 in EECs from SE (Normal #1) and PE (Normal #2), shown above. Both the MEK inhibitor PD98059 and the estrogen receptor antagonist ICI 182,780 completely block the E₂-induced decrease in p27. B, conversely, one primary culture of ECA (grade III/III) did not contain basal levels of p27 and was unaffected by E₂ treatment. The graphs (right) derived from densitometry of the blots represent the fold change of p27 levels compared with the untreated controls. Similar results were obtained with additional primary cell cultures isolated from separate normal endometrial tissue samples: PE (n = 3), SE (n = 6). C, estrogen decreases endogenous p27 in primary normal EECs by ubiquitin-mediated degradation. To determine whether the decrease in p27 in E₂-treated EECs is due to degradation of p27 via the ubiquitin-proteasome pathway, the cultures were incubated with E₂ (100 nmol/L), and 2.5 h before harvesting the cells, the proteasome inhibitor lactacystin (1 and 10 μmol/L) was added to the cells. Lactacystin (1.0 μmol/L)–treated primary EEC cultures had similar levels of endogenous p27 as the untreated controls (data not shown). The following patient samples were used in these experiments: SE (n = 4), PE (n = 1).

Progesterone increases p27^Kip1 levels in primary cultures of both EECs and ECA cells. Because progesterone inhibits the growth of endometrial glandular cells, we reasoned that this hormone may have an opposite effect than E₂ on p27 and thus increase p27 levels. As estrogen levels increase with temporal progression of the proliferative phase of the menstrual cycle, progesterone receptors are progressively up-regulated ( 35) to enhance the progesterone effect. Therefore, E₂ and progesterone were added to the primary EECs simultaneously to obtain a more robust progesterone-induced response, as previously described ( 36). Whereas progesterone alone increases endogenous p27 levels, E₂ plus progesterone markedly increases p27 by 6.7-fold, as shown in normal tissue #1, and by 2-fold in normal tissue #2 ( Fig. 5A ). Blocking the progesterone receptor with the antagonist mifepristone (RU-486) restored p27 to above the untreated control level. Interestingly, E₂ plus progesterone consistently induced an increase in p27 in ECA cell cultures, as shown in Fig. 5B, by 1.8-fold in a grade ECA III/III that was blocked by RU-486. In fact, blocking the progesterone receptor seemed to reduce p27 levels below untreated control. Therefore, progesterone treatment of both normal and malignant endometrial cells effectively increases intracellular levels of p27.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Analysis of p27 levels following treatment of primary cultures of normal EECs and ECA cells with medroxyprogesterone 17-acetate (Pg). Primary cultures of normal EECs and ECA cells were separately treated with 100 nmol/L E₂, 100 nmol/L progesterone (Pg), and E₂ plus progesterone every other day and the experiments terminated at 4 d with RIPA buffer. The levels of endogenous p27 in the cell lysates (30 μg) were determined by immunoblotting as described in Fig. 1B. Fold induction of p27 is illustrated in the graphs (right). E₂ was added together with progesterone to increase the level of progesterone receptors for a more robust response ( 35). A, estrogen plus progesterone increases the levels of endogenous p27 in primary cultures of normal EECs. The primary cultures of two normal tissues, Normal #1 (SE) and Normal #2 (PE), show the patient variability of basal levels of p27 with different levels of induction of p27 by progesterone. The addition of 10 μmol/L mifepristone (RU-486) to the E₂- and progesterone-treated cell cultures, shown in the cells of Normal #2, blocked the p27 increase induced by E₂ plus progesterone. The patient samples used are as follows: PE (n = 3), SE (n = 4). B, estrogen plus progesterone increases the levels of endogenous p27 in primary cultures of ECA cells. The ECA cell cultures were treated as described in (A). E₂ and progesterone increased the levels of p27 compared with untreated controls and RU-486 inhibited the response (a representative sample, ECA grade III/III, is shown; n = 6).

Discussion

The putative relationship between TGF-β–mediated growth inhibition and p27 has been considered to be via the direct effect of TGF-β on both p15 gene transcription and protein stability leading to increased binding of p15 to cyclin D/cdk4 and cdk6 (and inhibition of kinase activity; ref. 14, 37). This causes p27 to become displaced from cyclin D/cdk4,6 and subsequently accumulate on cyclin E/cdk2 complexes for the inhibition of cdk2 activity, thus blocking S-phase progression. However, cells containing nonfunctional p15 are able to accumulate p27 and inhibit cyclin E/cdk2 in response to TGF-β and, similarly, p15 null cells can still undergo TGF-β–mediated growth inhibition ( 38, 39), suggesting that compensatory or alternative mechanisms exist. As we have previously shown that disrupted TGF-β signaling is an early event in the development of ECA concomitant with loss of growth inhibition ( 4), the present study was undertaken to elucidate growth regulatory mechanisms potentially inactivated by disrupted TGF-β signaling. Therefore, we determined whether TGF-β could directly influence the levels of p27 as an effector of TGF-β–mediated growth regulation in normal primary cultures of EECs. Our studies herein show that whereas TGF-β dose-dependently increases the intracellular levels of p27 protein in cultures of EECs, ECA cells do not respond ( Fig. 1B). Following subcellular fractionation of EECs, the increase in p27 was observed in both the cytoplasmic and nuclear fractions. However, there was a 2-fold greater increase in the nucleus ( Fig. 1C). In untreated EECs, >78% of the p27 was in the nuclear fraction. In addition, the negligible amounts of p27 shown in ECA cells were confined to the cytoplasm (three of seven ECA primary cultures). This is of interest because, more recently, cytoplasmic p27 has been associated with cell motility and metastasis ( 40). Notably, the dose of TGF-β required to induce maximal levels of p27 (100 pmol/L) parallels those previously obtained for TGF-β–mediated growth inhibition of EECs ( 4). However, a direct effect of TGF-β on increasing nuclear levels of p27 for the putative inhibition of G₁ cyclin/cdks has not been previously shown in epithelial cells.

Considering the integral role for p27 in blocking cell cycle progression, decreased levels of p27 have been shown in human cancers ( 16, 23, 24), which often correlate with poor survival ( 23, 41). Our studies show that, irrespective of grade, ECA had a statistically significant decrease in the levels of p27 compared with normal endometrial glands of SE, containing abundant p27 in the nucleus [ Fig. 2A, d compared with b (arrows)]. The negligible levels of p27 in the PE glands ( Fig. 2A-a) and nuclear abundance of p27 in SE glands are expected and consistent with the status of growth and differentiation of glands of PE and SE, respectively. Similarly, other studies have shown decreased expression of p27 in ECA compared with normal endometrium, which did not vary with grade or stage ( 24, 42). In addition, we show that the disruption of TGF-β signaling in ECA has a statistically significant correlation with low to absent levels of p27 in the same patients, including grade I carcinomas ( 4). Therefore, we propose that loss of TGF-β signaling, which otherwise would increase intracellular levels of p27, may be an early event that contributes to the pathogenesis of ECA. Interestingly, because p27 is localized in the cytoplasm in hyperplastic glands ( Fig. 2A-c), we further hypothesize that ECA may commence with p27 being driven from the nucleus, where it is essential for inhibition of cdk2, into the cytoplasm, as an initial event marking neoplastic development of the endometrium and, subsequently, loss of this protein in carcinoma. Therefore, the absence or mislocalization of p27 may be a signal for the onset of ECA.

A plethora of studies show that the fate of p27 and resulting functional consequences are tightly regulated by cell cycle–dependent phosphorylation of the molecule by different kinases at specific sites that dictate both its degradation and nuclear retention ( 15, 16, 19, 20, 43– 45). The most prominent means for posttranslational regulation of nuclear levels of p27, particularly in cancer, is through ubiquitin-mediated degradation of p27 by the 26S proteasome, which is signaled by phosphorylation of p27 on Thr¹⁸⁷ by cyclin E/cdk2 and other kinases during S-G₂ phase of the cell cycle ( 16, 19, 46). In addition, phosphorylation of p27 on Ser¹⁰, Thr¹⁵⁷, and Thr¹⁹⁸ by certain kinases, including Akt, causes cytoplasmic mislocalization of p27 and subsequent ubiquitin-mediated degradation in G₀-G₁ ( 20, 26, 43– 45, 47). In all cases, lower levels of nuclear p27 ostensibly foster high nuclear activity of cdk2 for cell cycle progression. Consistent with the lack of p27 in the glands of ECA tissue in vivo ( Fig. 2), we show that lysates from all grades of ECA tissue ubiquitylate and rapidly degrade exogenously added rp27 compared with normal SE lysates ( Fig. 3A and D). Inhibition of p27 degradation by removal of proteasomes and separately by proteasome inhibitors in these in vitro assays indicates that the degradation of p27 is largely ubiquitin-proteasome mediated ( Fig. 3B and C). In addition, it was evident that other proteolytic systems degraded p27 in the presence of protease inhibitors. Studies using a similar assay showed that lysates from colon carcinoma tissues with low or absent levels of p27 displayed proteasome-dependent degradation of p27 ( 32). We did not determine the extent of nuclear or cytoplasmic degradation of p27. Furthermore, the kinase(s) involved in the phosphorylation of p27 to trigger ubiquitylation present in the ECA tissue lysates is not known. However, MAPK, constitutively active in many human cancers, has been shown to phosphorylate p27 at Thr¹⁸⁷ in breast tumor cells ( 16, 46). Moreover, blocking MAPK decreased epidermal growth factor receptor kinase activity, stabilizing p27 and delaying tumor formation in a mouse mammary model of tumorigenesis ( 34). With the use of specific kinase inhibitors and deletion mutants of p27, we should be able to determine both the kinases and the phosphorylation sites involved in p27 degradation in ECA. Similar to the ECA tissue lysates, PE lysates lacked p27 expression in vivo and ubiquitylated and degraded p27. As TGF-β signaling is intact in these cells, p27 can ostensibly be induced for growth inhibition in the secretory phase of the menstrual cycle when TGF-β levels are higher and signaling is most active ( 4, 48). Importantly, patients' tissue lysates showed 100% correlation between the ability to degrade p27 in vitro and complete absence of p27 from the malignant glands in vivo. Therefore, we postulate that the absence of p27 from the malignant glands and those of PE in vivo might result from a high rate of proteasomal degradation. In this light, proteasome inhibitors have been shown to be a favorable therapeutic route for certain human cancers, including multiple myeloma ( 49). The importance of maintaining sufficient levels of p27 for growth inhibition is underscored by studies showing that overexpression of p27 in human breast epithelial and cancer cell lines could lengthen G₁ phase and suppress tumorigenicity in vivo ( 50). Taken together, the data suggest that the pathogenesis of ECA may be related to a high rate of p27 degradation and simultaneous lack of TGF-β signaling that would otherwise enable TGF-β–mediated accumulation of this protein.

Our findings implicate a dynamic regulation of p27 levels in both normal endometrial growth and pathogenesis of ECA. The hormones, E₂ and progesterone, have mitogenic and antimitogenic effects on endometrial growth, respectively. Whereas the basal levels of p27 protein are variable in tissue culture as well as the degree of response of specific patients, E₂ treatment of normal EECs derived from both PE and SE decreases the levels of endogenous p27 ( Fig. 4) and, conversely, progesterone treatment of EECs from both PE and SE markedly increased the levels of endogenous p27 levels ( Fig. 5). Antagonists of estrogen receptor (ICI 182,780) and the progesterone receptor (RU-486) restored p27 to near basal levels in the normal EEC cultures, indicating that p27 is specifically regulated by hormone receptor–mediated responses.

We show that inhibitors of both MAPK/extracellular signal–regulated kinase kinase (MEK) and ubiquitin-proteasome pathways block the E₂-dependent decrease in cellular levels of p27 in primary cultures of normal EECs, suggesting that the observed E₂-induced proteasomal degradation may be MAPK driven. Thus, at least one mechanism of E₂-induced proliferation in the normal endometrium and in estrogen-induced type I endometrioid ECA may be by lowering p27 levels through this pathway to enable S-phase progression. In breast cancer cells, E₂ activates the MAPK pathway downstream from Ras, initiating phosphorylation of p27 at Ser¹⁰, which causes p27 to bind to the nucleopore protein, CRM1, followed by export of p27 from the nucleus into the cytoplasm for degradation ( 26). Therefore, it is possible that Ras may be involved in E₂-induced activation of MAPK in the endometrium. It has been shown that both the pure estrogen receptor antagonist ICI 182,780 and tamoxifen, successfully used to treat breast cancer, increase p27 and p21 intracellular levels and mediate their binding to cyclin E/cdk complexes to block S-phase entry ( 25, 27). However, our studies implicate the involvement of the ubiquitin-proteasome pathway in E₂-induced p27 degradation in the endometrium ( Fig. 4C).

Progesterone has been shown to induce growth inhibition concomitant with increased p27 levels in an ECA cell line and primary normal EECs by promoting the binding of p27 to cdk2 and decreasing the rate of p27 degradation ( 51). We are currently investigating the mechanism involved in the progesterone-induced increase in p27 observed herein including its effect on subcellular distribution of p27, which presumably would be nuclear to inhibit cdk2 activity, as described above. Importantly, we show that progesterone increases p27 levels in both normal EECs and ECA cells. Therefore, we raise the possibility that the therapeutic action of progesterone on endometrial hyperplasia and carcinoma may be directly related to the increase of nuclear levels of p27 in these cells. Moreover, the response to progesterone is not surprising as 75% of ECA patients, particularly well-differentiated cancers, retain progesterone receptors ( 52).

In primary endometrial normal and malignant cells, we show that the increase in p27 levels induced by TGF-β is not at the transcriptional level but that novel protein synthesis is involved. In further studies, we show that TGF-β may accumulate p27 intracellular levels by preventing its ubiquitin-mediated degradation by down-regulating Skp2 and Cks1 ( 53), ¹ the requisite and rate-limiting proteins of the SCF complex that targets p27 to the proteasome ( 16). This might occur by induction of a corepressor acting on the Cks1 and/or Skp2 promoter. Interestingly, p27 levels are elevated in both Skp2 and Cks1 null mice that have reduced growth rates ( 54, 55). Further to these findings, Skp2 and Cks1 are overexpressed in a number of human cancers ( 16), often with a concomitant decrease in p27. Still, there may be complementary means for TGF-β to increase intracellular levels of p27 for growth inhibition of EECs to prevent its degradation and/or through translational control. Our studies provide evidence that dysregulated TGF-β signaling results in lack of p27 accumulation in ECA. We propose that lack of TGF-β– and/or progesterone-induced accumulation of p27 and increased phosphorylation of p27 by an E₂-activated MAPK, causing heightened proteasome degradation, lead to endometrial hyperplasia, the precursor to estrogen-induced ECA. Importantly, the opposing effects of E₂ and progesterone on the regulation of p27 suggest that this cdk inhibitor is a mediator of both induction of E₂-induced hyperplasia and ECA and progesterone-mediated amelioration of this disease. As depicted ( Fig. 6 ), our studies clearly show that TGF-β, E₂, and progesterone converge on the regulation of p27 levels, thus providing a potential therapeutic target for restoration of normal growth of the endometrium.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

TGF-β, E₂, and medroxyprogesterone 17-acetate (Pg) actions converge on the levels of p27 in primary cultures of normal EECs and ECA cells. The data indicate that p27 levels vary throughout the normal menstrual cycle to control endometrial cell growth. E₂ decreases p27 levels by ubiquitin-mediated degradation in the 26S proteasomes in the proliferative phase (PE) of the menstrual cycle. Conversely, progesterone causes an accumulation of p27 levels in the secretory phase (SE) of the menstrual cycle. In addition, normal TGF-β signaling increases p27 levels in SE. As such, p27 levels are low to absent in endometrial glands in PE and abundant in SE in vivo, as shown ( Fig. 2). p27 is absent in ECA due to a high rate of ubiquitin-mediated proteasomal degradation and disrupted TGF-β receptor signaling, which renders the carcinoma cells unable to be induced by TGF-β for p27 accumulation. Type I endometrioid ECA has been shown to occur due to unopposed E₂, as in anovulation, menopause, and tamoxifen therapy for breast cancer ( 6, 7). We propose that the high rate of p27 degradation coupled with lack of progesterone and the inability of TGF-β to induce p27 accumulation (when TGF-β receptors are down-regulated or signaling is disabled) lead to uncontrolled proliferation of EECs thereby contributing to the pathogenesis of ECA. However, ECA cells that possess progesterone receptors can still accumulate p27 in response to progesterone in vitro (large circle). Because progesterone is therapeutic for ECA ( 8), p27 may be a target for successful therapy by progesterone for endometrial hyperplasia and carcinoma (in inhibition of endometrial cell growth). Double arrows, status of p27 levels in the normal cycling endometrium and in ECA.