Mechanisms involved in the evolution of progestin resistance in human endometrial hyperplasia—precursor of endometrial cancer

Available online at www.sciencedirect.com R

Gynecologic Oncology 88 (2003) 108 –117

www.elsevier.com/locate/ygyno

Mechanisms involved in the evolution of progestin resistance in human endometrial hyperplasia—precursor of endometrial cancer Sa Wang,a Jeffery Pudney,b Joon Song,c Gil Mor,c Peter E. Schwartz,c and Wenxin Zhenga,c,* b

a Department of Pathology, Yale University School of Medicine, New Haven, CT, USA Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, CT, USA c Brigham and Women’s Hospital, Harvard University, Boston, MA, USA

Received 21 February 2002

Abstract Background. Successful treatment of endometrial hyperplasia with progestins is commonly accompanied by the finding of an inactive or suppressed endometrium after therapy. However, approximately 30% of the endometrial hyperplasia cases do not respond to progestins and hyperplastic glands persist. The Fas/FasL system is known to play a role in tissue remodeling as a result of changes in menstrual hormone levels. The aims of this study are to examine Fas/FasL expression in endometrial hyperplasia of pre- and postprogestin treatment samples and to study the Fas/FasL regulation in vitro with Ishikawa cells after progestin stimulation. Design. Pre- and posttreatment paraffin-embedded endometrial hyperplasia tissue samples from 26 women were examined by immunohistochemistry for changes in Fas/FasL expression related to the administration of progestins. Among 26 patients, 18 were successfully treated with progestins and 8 failed treatment. Fas/ FasL positivity was defined by the presence of 10% or more immunoreactive epithelial cells in each specimen. In positive cases, a percentage or an immunoscore of immunoreactive cells was given by counting 500 cells. Cell viability was evaluated by the MTT assay. The in vitro effects of progesterone on Fas/FasL expression and apoptosis in Ishikawa cells were examined by using Western blot and TUNEL assays, respectively. Results. Fas immunoreactivity was present in 4/26 (15%) preprogestin cases with an average of 16% of the epithelial cells expressing Fas. FasL was expressed in 21/26 (80%) pretreatment cases with an average of 42% of the hyperplastic glandular cells being positive. In postprogestin cases, an increase of Fas expression (14/18, 77%) with an average of 47% stained cells was seen in responders (P ⬍ 0.001), while FasL was found in 16/18 (89%) responders with an average of 65% of cells positive (P ⫽ 0.587). In nonresponders, no significant changes in Fas/FasL expression were detected compared to pretreatment samples. With in vitro Ishikawa cells, a slight increase (10 –20%) of Fas and FasL protein expression was detected after 24 h of progesterone treatment, but a more significant increase (220 –343%) of both Fas and FasL expression was found after 48 h of withdrawing progesterone, which parallels apoptotic activity. Conclusions. The Fas/FasL system may be involved in the development of endometrial hyperplasia. Part of the molecular mechanisms of progestin therapy for endometrial hyperplasia is through upregulation of Fas/FasL expression. Dysregulation of Fas/FasL expression in hyperplastic endometrium may be part of the molecular mechanisms for nonresponders to progestin treatment. Intermittent, rather than continuous, progestin treatment may be more effective clinically for the treatment of endometrial hyperplasia. © 2003 Elsevier Science (USA). All rights reserved. Keywords: Endometrial hyperplasia; Progestin; Apoptosis; Fas/FasL

Introduction Endometrial hyperplasia is related to prolonged estrogen stimulation in association with a diminished to absent pro* Corresponding author. Department of Pathology, Yale University School of Medicine, 20 York Street, EP 2-608, New Haven, CT 06520-8070. E-mail address: [email protected] (W. Zheng).

gestational activity. It is well documented that these conditions have the potential to progress to endometrial cancer when left untreated, particularly when “atypia” is present in hyperplastic endometrium [1]. High-dose progestins are effective agents, and their use in the treatment of endometrial hyperplasia has been well documented [2– 4], especially in young women who wish to retain fertility or in older women who are poor surgical

0090-8258/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0090-8258(02)00008-2

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

candidates. Successful treatment of endometrial hyperplasia with progestins is commonly accompanied by the histologic finding of an inactive or suppressed endometrium after therapy [5,6]. Hyperplastic glands may retain some of their architectural abnormalities but the glandular cells show attenuation. Mitotic activity is deceased, the stroma becomes decidulized, and spiral arteries become inconspicuous. However, up to 30% of patients fail to respond to progestin therapy, especially when atypical endometrium is present [1,6 – 8]. Histologically, in atypical glandular hyperplasia, progestin therapy fails to produce any effect or may produce only partial changes: part of the endometrium shows secretory changes in glands that remain crowded and irregular in shape. Previous explanations on this phenomenon of progestin resistance have included decreased availability of progestin receptors (PR) prior to treatment and alterations in PR regulatory function [9 –14]. Subsequently, molecular mechanisms involving progestin resistance were shown to be related to dysregulation of TGF-alpha and EGF receptors and apoptosis of the endometrial glandular cells [15]. We have shown in recent studies that bcl-2 expression is decreased following successful progestin treatment of endometrial hyperplasia, whereas the bcl-2 expression is persistent in foci of hyperplastic endometrium not responding to progestin therapy [16]. Furthermore, successful progestin treatment may be directly related to progestin-induced apoptosis, which occurs in the first few days after progestin treatment [17]. It seems that the responding versus nonresponding endometrium is largely related to alterations of molecules related to cellular apoptotic signaling pathways. In addition to bcl-2, no other molecules in the apoptotic pathway involving endometrial epithelial cell apoptosis have also been found [18,19]. Fas, also called APO-1 or CD95, is a member of the tumor necrosis factor (TNF)/nerve growth factor (NGF) family and is a type-I membrane protein [20]. Fas ligand (FasL), a type II membrane protein, also belongs to the TNF superfamily [21]. The Fas-FasL interaction is one of the essential events in the induction of apoptosis including that of the process of cyclic menstrual shedding [19,22–24]. When Fas is activated, the intracellular death domain in the Fas molecule binds to the Fas-associated death domaincontaining protein (FADD), which then recruits the FADDlike interlukin-1␤-converting enzyme (FLICE) [25–27]. The activation of FLICE leads to the formation of the death-inducing signaling complex that initiates downstream activation of caspase 3, 6 and 7 and 8, which cause massive degradation of the endometrium. Fas and FasL antigens localized on the Golgi apparatus and vesicles during the late proliferative phase are incorporated into the cell membranes during the secretory phase, and are coexpressed on the cell membranes of endometrial glands throughout the menstrual cycle [19]. However, FasL exhibits peak expression during the secretory phase [18]. Expression of Fas and FasL appears to be controlled by ovarian steroids including progesterone. Under the influence of sex steroids such as estrogen,

109

endometrial cells proliferate as a result of estrogen-induced activation of antiapoptotic genes and inhibition of proapoptotic genes. On the other hand, progesterone acts as a differentiation factor, blocking mitosis and inducing terminal differentiation [28]. In addition, it is well known that synchronous endometrial shedding occurs typically when serum levels of progestin are withdrawn rather than maintained. This process is characterized by the activation of numerous genes, including proapoptotic factors [29]. However, the current clinical management of endometrial hyperplasia uses continuous large doses of progestins rather than cyclic applications. This study had two aims. The first aim was to study Fas/ FasL regulation in pre- and postprogestin-treated samples in order to further understand the molecular mechanisms involved in nonresponse of hyperplastic endometrium to progestin therapy. The second aim was to use the Ishikawa cell line as an in vitro model to measure Fas/FasL-related apoptotic activities in continuous progesterone treatment and progesterone withdrawal experiments in order to broaden understanding of the molecular mechanisms of progestin effects. These studies may ultimately help to develop a better modality for progestin therapy in endometrial hyperplasia, a precursor lesion of endometrial cancer.

Materials and methods Selection of matched cases Twenty-six paired pre- and postprogestin-treated endometrial hyperplasia samples were studied. Among them, 15 cases were derived from our previous studies on bcl-2 and apoptosis [16,17], while 11 pairs of samples were collected from the Department of Pathology at Yale-New Haven Hospital from 2000 to 2001. The method of collection has been previously described [16,17]. The clinical information is summarized as follows. The 26 cases comprised 18 progestin responders (successfully treated) and 8 nonresponders (failure to respond to progestin treatment). The responders were defined by histologically complete attenuation of the endometrial glandular cells, decidualized stroma, and no residual hyperplasia on subsequent endometrial samplings. The nonresponders were defined by the presence of residual endometrial hyperplasia on tissue samplings performed after a trial of progestin therapy. The patients’ age, the treatment regimen, and the pretreatment findings are summarized in Table 1. The ages of the responders ranged from 31 to 51 years (median of 42), and ages of those who failed treatment ranged from 23 to 49 years (median of 39). Patients from both groups were treated with one of the following progestins: megestrol acetate, medroxyprogesterone acetate, norethindrone, or depomedroxyprogesterone acetate. Duration of therapy ranged from 2 to 11 months among the group of successfully treated cases, and 2 to 12 months among the patients who

110

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

Table 1 Summary of patients treated with progestins Age (years) Responders 25 33 39 42 45 25 31 32 40 42 45 48 50 51 41 44 45 47 Nonresponders 37 39 41 49 24 27 39 23

Progestin dose

Tx duration (months)

Dx

Post-tx findings

MA 40 mg qd MA 40 mg qd MPA 20 mg qd Norethindrone acetate 1 mg qd MPA 20 mg qd MA 40 mg qd MPA 10 mg qd MPA 10 mg qd Days 1–10 Norethindrone acetate 1 mg qd MPA 10 mg qd Days 1–10 MPA 20 mg qd MPA 10 mg qd Days 1–10 MPA 20 mg qd MPA 20 mg qd MA 40 mg qd MPA 20 mg qd MPA 10 mg qd Days 1–10 MPA 10 mg qd Days 1–10

5 6 3 5 6 3 3 9 2 6 4 3 3 2 3 2 3 11

ACH ACH ASH ACH ACH CH CH CH CH CH CH CH CH CH SH SH SH SH

ProgE ProgE ProgE ProgE, SM PE, Prog E ProgE, SM ProgE, SM PE, ProgE, SM ProgE PE, ProgE ProgE ProgE ProgE ProgE ProgE ProgE, SM ProgE PE

MPA 20 mg qd MPA 20 mg qd MPA 20 mg qd MPA 10 mg qd Norethindrone 1 mg qd Progesterone 200 mg q4–6 months MPA 10 mg qd MPA 10 mg qd

6 3 2 3 3 12 6 6

ACH ACH ACH ACH CH, CH, CH SH

Focal ACH, ProgE ACH, SM ACH, SM focal CH, ProgE focal CH, ProgE focal CH, ProgE, SM CH SH, PE

Note. ACH, complex hyperplasia with atypia; CH, complex hyperplasia; SH, simple hyperplasia; ASH, simple hyperplasia with atypia; ProgE, progestin effect; PE, proliferative endometrium; SM, squamous metaplasia; MA, megestrol acetate; MPA, medroxyprogesterone acetate; qd, once a day; Days 1–10, taking the progestin from Day 1 to Day 10 in menstrual cycle.

experienced treatment failure. Among the responders, there were 4 with an initial diagnosis of simple hyperplasia without atypia, 1 simple hyperplasia with atypia, 9 with complex hyperplasia without atypia, and 1 with complex hyperplasia with atypia. The nonresponders included 1 case of simple hyperplasia with atypia, 3 cases of complex hyperplasia without atypia, and 4 cases of complex hyperplasia with atypia prior to the initiation of progestin treatment. Pathological diagnoses of the endometrial lesions were made by one of our gynecologic pathologists (W.Z.). Tissue processing The pre- and posttreatment tissue samples of all the cases were fixed in 10% buffered formalin and processed routinely for paraffin embedding. Tissue obtained prior to treatment included 12 endometrial biopsies and 14 suction curettage specimens. Tissue samples obtained after therapy included 8 endometrial biopsies, 13 suction curettage specimens, and 5 hysterectomy specimens. Five-micrometer sections for IHC were cut and placed on treated slides. A section on each case was stained with hematoxylin and eosin (H&E) and examined microscopically to confirm the diagnosis (S.W. and W.Z.).

Histologic evaluation of posttreatment samples H&E sections of endometrial samples obtained 2 to 11 months after a trial of progestin administration were assessed for therapeutic response to treatment. Patients were considered to have complete regression of hyperplasia (responders) if posttreatment endometrial samplings showed no histologic evidence of hyperplasia and there were findings of benign proliferative or secretory endometrium and/or progestin effect with stromal decidualization and cuboidal to attenuated changes in the glands. If more than 10% of the entire endometrial sample contained endometrial hyperplasia after progestin treatment, the patient was considered to have persistent disease (nonresponder). Immunohistochemistry for Fas and FasL detection Detection of Fas/FasL expression was performed using a rabbit polyclonal IgG containing antihuman FasL or mouse monoclonal IgG containing antihuman Fas (N-20 or B-10, respectively; Santa Cruz Biotechnology, Inc., Irvine, CA). The appropriate dilutions were determined in preliminary experiments. Immunohistochemistry (IHC) was performed as described by Zheng et al. [30,31]. Briefly, parallel sec-

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

tions for demonstration of Fas and FasL were initially deparaffinized and rehydrated. The slides were placed in 3% hydrogen peroxide-methanol mixture for 10 min and then thoroughly rinsed in distilled water. Antigen retrieval was performed through microwave heating in 10 mM citrate buffer at pH 6.0 (BioGenex, San Ramon, CA). Parallel sections were incubated with anti-Fas (1:200 dilution) or anti-FasL (1:200 dilutions) (R&D, Minneapolis, MN) in a moisture chamber for 12 h and then rinsed with PBS. This was followed by a 45-min incubation with biotinylated secondary antibody (Vector, Burlingame, CA) and subsequent PBS wash for 10 min. Slides were bathed a second time in hydrogen peroxide using a 3% hydrogen peroxidePBS mixture for 10 min and then washed thoroughly. A strepavidin-biotin-alkaline phosphatase conjugate (Vector) was then applied for 30 min followed by four PBS washes. Sections were then developed with the substrate red kit (Vector) and counterstained with hematoxylin, dehydrated, cleared, and mounted. Expression of Fas and FasL was assessed using a graded scale on glandular immunoreactivity. Evaluation was performed without referenced knowledge of the state of treatment or the response to therapy. Cellular reactivity for Fas/FasL was based on the presence of distinct membrane/ cytoplasmic staining. Glandular reactivity for Fas/FasL was graded on a scale of 0 to 3⫹ with a score of 0 representing negative expression. A score of 0 was assigned when ⬍10% of the glandular cells showed positive reactivity. A score of 1⫹ was assigned when 10 –33% of the glandular cells expressed positive reactivity. A score of 2⫹ was assigned when 33– 66% of the glandular cell expressed positive reactivity. Finally, a score of 3⫹ was assigned when ⬎66% of the endometrial glandular cells showed positive reactivity. A total of 500 cells were evaluated in the determination of glandular immunoreactivity. All IHC slides were reviewed independently by at least two of our investigators and confirmed by a third investigator. Human lymph node tissues and normal proliferative endometrium were utilized as positive controls for Fas/FasL immunostaining. In negative controls, the primary antibodies were omitted or preabsorbed with recombinant Fas or FasL, as described previously [23]. Cell culture and hormone treatment The Ishikawa cell line, which was derived from a welldifferentiated endometrial adenocarcinoma, was generously provided by Dr. Masato Nishida, Tsukuba University, Tsukuba City, Japan. The culture conditions were described previously [17]. Cell culture experiments are described briefly as follows. The cells were plated in 96-well microtest plates in 200 ␮l of culture medium per well at a cell density of 1 ⫻ 105 cells/ml. At 80% confluence, the media was changed to 1% FBS, and incubation was initiated with 10 mM progesterone or with cell media only. Dose-response analysis of the Ishikawa cells was performed with progesterone treatment prior to the progester-one-induced apopto-

111

sis experiments. The doses tested were 0.01, 0.1, 1, and 10 ␮M. We failed to obtain 100 ␮M progesterone due to its solubility limitation in ethanol. We selected 10 ␮M progesterone as the testing concentration because this dose showed maximal inhibition of Ishikawa cell growth in vitro. Each treatment condition was performed in triplicate. The progesterone standard was provided by Dr. Frank Stanczyk, Reproductive Endocrine Research Laboratory, Department of Obstetrics and Gynecology, University of Southern California Keck School of Medicine. The hormone was dissolved in ethanol and prepared in culture media with a final ethanol concentration not exceeding 0.1%. Progesterone was added to the cultures for 48 h and then removed for another 24 or 48 h prior to Fas/Fas ligand Western analysis. Cell viability and Fas-mediated apoptosis Cell viability was measured by the MTT assay as previously described [32,33]. Single-cell suspensions were prepared by passing the cells through a 21-gauge needle, preventing the formation of aggregates. Each experiment was done in triplicate and repeated at least three times. We did not find differences in the proliferation rate between each individual well in the same group. A suspension in 200 ␮l of culture media with 5 ⫻ 103 ⫺ 5 ⫻ 104 cells in MEM supplemented with 10% FBS for Ishikawa cells was distributed into each well of a 96-well flat-bottomed microplate (Becton-Dickinson, Lincoln Park, NJ) and incubated for 48 h. Then the medium was removed and cells were incubated with experimental medium containing various amounts of progesterone described above. To determine the sensitivity of our system to Fas-mediated apoptosis, variable concentrations of mouse monoclonal IgG anti-Fas Ab (R&D, Minneapolis, MN) was added to 200 ␮l of medium for 24 h after each experimental condition. Thereafter, 20 ␮l 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma, St. Louis, MO) dye was added to each well 4 h before the end of the incubation. The nonreacted dye and medium in the wells were poured off, and 100 ␮l of acidified isopropyl alcohol was added to solubilize the reactive crystals. The absorbency at 540 nm was measured by an automatic microplate reader (Model 550, Bio-Rad, Hurcules, CA). All assays were done in triplicate. The percentage of antiprolifertive effect was calculated using the following formula: 100 (1 – absorbance of experimental well/absorbance of positive control well). Detection of apoptotic cells with terminal Deoxy(d)-UTP Nick-end labeling (TUNEL) The presence of apoptotic cells was determined by propidium iodide staining and the TUNEL assay. Endometrial Ishikawa cells (1 ⫻ 105) were incubated in chamber slides and treated as described above. Propidium iodide (1 ␮g/ml) was applied to the chambers after each hormonal treatment, and the slides were fixed with 4% paraformaldehyde and

112

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

examined under a fluorescence microscope. To further confirm the presence of apoptotic cells following progesterone treatment and withdrawal we carried out the TUNEL assay according to the method we previously used [17] with a few modifications. Briefly, cells were incubated with permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate) for 2 min on ice. Following three washes with PBS and one rinse with distilled deionized water, the sections were covered with terminal deoxynucleotidyl transferase (TdT) buffer [30 mM Trizma base (pH 7.2), 140 mM sodium cacodylate, and 1 mM cobalt chloride]. Then, the TdT buffer containing 0.2 U/␮l TdT (Boehringer Mannheim, Gaithersburg, MD) and 10 ␮M fluorescent 16-dUTP (Boehringer Mannheim) was added to the cells, and the slides were incubated in a humidified chamber at 37°C for 60 min. The reaction was terminated by transferring the slides to 50 mM Tris-HCl, pH 7, for 15 min at room temperature. The slides were then washed with PBS. Endogenous peroxidase was inactivated with 0.3% H2O2 in methanol for 15 min at room temperature followed by 3 ⫻ 10-min washings with PBS. To block the nonspecific binding of the antibodies, sections were incubated with 5% normal goat serum in PBS-5% BSA for 30 min at room temperature and washed 3 ⫻ 5 min with PBS. The sections were then incubated for 2 h at room temperature with an antifluorescein antibody, A Fab fragment from sheep, and conjugated with alkaline phosphatase (AP) (Boehringer Mannheim, Geithersburg, MD) and the AP substrate Blue Kit (Vector) (blue color). The presence of positively stained Ishikawa cells was evaluated using a Zeiss fluorescent microscope and Image Analysis software Openlab 2 (Improvision, Lexiton, MA). Fas and Fas ligand western blot analysis Proteins were extracted from Ishikawa cell in lysis buffer (TRIzol reagent, GIBCO BRL, Gaithersburg, MD) according to the manufacturer’s instructions. The protease inhibitors in the lysis buffer included 1 mM Na3VO4. 10 ␮g/ml leupeptin, 10 ␮g/ml aprotinin, and 4 mM PMSF. The protein concentration was determined by a detergentcompatible protein assay (Bio-Rad Laboratories, Hercules, CA). Samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 10% polyacrylamide gels and transferred to nitrocellulose membranes as described previously [34]. Equal amounts of protein in each lane were confirmed by staining the membrane with Ponceau Red. Immunoblotting was performed after blocking the membranes with 5% powdered milk in water. The blots were incubated first with antibody for FasL (monoclonal antibody, clone 33, Transduction laboratory, Lexington, KY, at 1:1000 dilution) or for Fas (monoclonal antibody, B-10, Santa Cruz Biotechnology, at 1:500 dilution) for 2 h. After washing with TBS-T for several times, the secondary antibody (peroxidaselabeled horse antimouse, Vector) was added and incubated for 30 min, and then developed with

TMB peroxidase substrate kit (Vector). All experiments were repeated at least three times and the intensity of the signal was analyzed using a digital imaging analysis system (1D Image Analysis Software, Scientific Imaging Kodak Co.). Statistical analysis Pre- and posttreatment measurements for Fas/FasL expression were each analyzed with respect to treatment outcome using paired and unpaired Student’s t tests. Student’s t test was also used for assessing the change in Fas/FasL regulations in the Ishikawa cell line after the designated treatment periods with progesterone. All P values were twosided and values less than 0.05 was considered significant.

Results Clinical responses after progestin therapy In 18 of 26 (69%) patients total regression of hyperplasia was attained after a mean treatment duration of 4.4 months (ranging from 2 to 11 months) (Table 1). The patients median age was 42 years. Of these 18 cases, 12 showed endometrial changes consistent with progestin effect, 1 showed only proliferative endometrium, and 5 had squamous metaplasia associated with progestin effect. In 8 of 26 (31%) patients, persistent hyperplasia was present at posttreatment evaluation. In this group of patients, progestin administration ranged from 2 to 12 months with a mean treatment duration of 5.1 months. The median age of these patients was 39 years. Among the 8 patients, 4 had evidence of progestin effect adjacent to hyperplastic endometrium and 3 had changes of squamous metaplasia accompanying persistent disease. Fas/FasL expression in pre- and postprogestin-treated endometrial samples Results of Fas and FasL expression in the endometrium, pre- and posttreatment, are summarized in Table 2. Fas immunoreactivity was present in 4/26 (15%) preprogestin cases with an average of 16% positive epithelial cells. In postprogestin cases, an increase of Fas detection (14/18, 77%) was seen with an average of 47% stained cells in responders (P ⬍ 0.001), while Fas immunoreactivity was only found in 3/8 (38%) nonresponders with an average of 28% positive cells. Comparing the responders and nonresponders in the postprogestin treatment samples, the percentage of Fas-positive cases was significantly increased (P ⬍ 0.05). Expression of Fas and FasL is illustrated in Fig. 1. When we scored the immunoreactivity as described under Material and Method and used in our previous bcl-2 study [16], the mean Fas score among the responders was 0.27 before treatment and 2.36 after treatment (P ⬍ 0.001). In the group of nonresponders, the mean Fas score seen prior to

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117 Table 2 Comparison of Fas and FasL expression between pre- and postprogestin treatment in relation to treatment outcome

Cases of Fas (%) Cases of FasL (%)

Preprogestin (n ⫽ 26)

Postprogestin Responders (n ⫽ 18)

Nonresponders (n ⫽ 8)

Pos

Neg

Pos

Neg

Pos

Neg

4 (15) 21 (81)

22 (85) 5 (19)

14 (78) 16 (89)

4 (22) 2 (11)

3 (37) 6 (75)

5 (63) 2 (25)

Note. Compared to the preprogestin treatment, the percentage of positive Fas expression cases was significantly increased after progestin treatment (P ⬍ 0.001). Comparing responders and nonresponders within postprogestin treatment subjects, responders have a significantly more Fas-positive cases than nonresponders (P ⬍ 0.05). No significant changes of FasL were detected after progestin treatment. Fisher’s exact test was used for the analysis.

treatment was 0.29, while a mean score of 0.83 was observed after progestin therapy. The difference of Fas immunoreactivity expressed as IHC scores between responders and nonresponders is also statistically significant (P ⬍ 0.05, Table 3). Increased Fas immunoreactivity was seen in the majority of the responders. A negative Fas immunoreactivity score was observed in 4 of the 18 responders. Fas expression was seen in 3 of 8 nonresponders. In these 3 cases, the number of positive cells was also less than that for responders. FasL was expressed in 21/26 (80%) pretreatment cases with an average of 42% positive hyperplastic glandular cells. In contrast to Fas immunoreactivity, there was no

113

Table 3 Comparisons of Fas and FasL expression with average immunoscores between responders and nonresponders

Fas score FasL score

Preprogestin (n ⫽ 26)

Postprogestin Responders (n ⫽ 18)

Nonresponders (n ⫽ 8)

0.27 (⫹/⫺0.08) 2.39 (⫹/⫺0.12)

2.36 (⫹/⫺0.09) 2.48 (⫹/⫺0.10)

0.83 (⫹/⫺0.11) 2.19 (⫹/⫺0.13)

Note. Compared to the preprogestin treatment, the IHC score of Fas was significantly increased after prosgestin, treatment (P ⬍ 0.001). Comparing responders and nonresponders within postprogestin treatment subjects, responders have a significantly higher Fas immunoscores than nonresponders (P ⬍ 0.05). No significant changes of FasL were detected after progestin treatment. Numbers represent mean immunoreactivity scores (standard deviation); P values were based on Fisher’s exact test.

significant change in FasL expression in either responders or nonresponders. FasL expression was seen in 16/18 (89%) of responders with an average of 65% positive cells, while FasL was seen in 6/8 (75%) nonresponders with an average of 61% positive cells. Although the FasL expression increased mildly in responders and decreased slightly in nonresponders, the changes did not reach statistical significance. Representative pictures are shown in Fig. 1. FasL expression in these cases is summarized in Tables 2 and 3. The patterns of Fas and FasL expression in endometrial glands were mainly cytoplasmic. Some cases showed a tendency for luminal or cell apical staining. Apoptotic cells in endometrial glandular lumens usually showed strong intensity for both Fas and FasL staining. This phenomenon

Fig. 1. Representative expression of Fas and FasL by immunohistochemistry in pre- and postprogestin-treated endometrial samples. The upper panels represent Fas expression, while the lower panels show FasL immunoreactivity. Weak and focal Fas immunoreactivity was present mainly in the cytoplasm of the hyperplastic endometrial epithelium prior to progestin treatment (A). Endometrial epithelium from a responder showed strong diffuse Fas immunoreactivity with a tendency for luminal or cell apical staining (B). Glandular epithelium from a nonresponder shows no increase of Fas expression compared to the preprogestin-treated sample. However, strong Fas staining is seen in the degenerating and shedding cells/debris within the lumen (C). In contrast to Fas expression, FasL immunoreactivity shows no significant differences between preprogestin- (D) and postprogestin-treated samples (E, F) nor between responder (E) and nonresponder (F). Overall FasL immunoreactivity is moderate to strong in the cytoplasm of endometrial glandular cells. Sporadic immunoreactivity for both Fas and FasL is present in endometrial stromal cells. Original magnifications: 200⫻.

114

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

was observed in both pre- and postprogestin treatment groups. Endometrial stromal cells stained sporadically for both Fas and FasL. There was no significant difference between pre- and postprogestin-treated samples and no difference was seen between responders and nonresponders. Vascular and myometrial smooth muscle cells exhibited largely no Fas or FasL expression in the cases studied. Fas/FasL regulation and endometrial cell apoptosis after progesterone treatment and withdrawal Hormonal responses in vitro are different from those in vivo. However, experiments in vitro usually are better controlled particularly when the response may be time dependent and frequent sampling is impractical from patients. We used the Ishikawa cell line as an in vitro model to address two following issues. Does Fas and FasL regulation by progesterone parallel each other and does it directly correlated to cellular apoptotic activity? If the results are as expected, does progesterone removal enhance this Fas/ FasL-related apoptosis. Ishikawa cells were treated with progesterone as described above. Ishikawa cells after 24 hours of progesterone treatment showed 10 –20% increased levels of Fas protein expression compared to Fas levels prior to progesterone treatment (Fig. 2). More impressively, Fas expression increased dramatically after progesterone removal. The levels of Fas up-regulation were time-dependent. There was a 41% increase after 24 h and a peak 220% increase 48 h after the removal of progesterone from the culture media. FasL showed increments parallel to those of Fas when progesterone was present for 24 h. The level of FasL expression further increased by 343% at 24 h and 280% at 48 h after progesterone removal. The data are illustrated in Fig. 2 (Western blot). The increases in both Fas and FasL expression in the cells after removal of progesterone were statistically significant (P ⬍ 0.001). The data on continuous progesterone treatment up to 96 h have been reported previously [17] (see Discussion). Cellular apoptosis is expected when both Fas and FasL are expressed in the same cells. In order to determine whether increased Fas and FasL expression in the treatment and withdrawal of progesterone are correlated to cellular apoptosis, we stained Ishikawa cells using a TUNEL assay following progesterone manipulations. The number of apoptotic cells increased 20% after 24 h of progesterone treatment compared to nonhormone-treated samples. A further dramatic increase in the number of apoptotic cells was observed 48 h after progesterone removal. Representative pictures are shown in Fig. 3.

Discussion Cyclic menstruation involves a complex set of events requiring interactions of many genes with stimulatory and

Fig. 2. Fas and FasL protein expression by Western blot in Ishikawa cells following progesterone treatment and withdrawal. Top panel: Fas protein is present in low levels in Ishikawa cells prior to progestin treatment (lane C). A slight increase of Fas expression is seen after progesterone treatment for 24 h (PR). Significant increase in Fas expression is seen in 24 h (41%) and in 48 h (220%) following removal of progesterone. Middle panel: FasL expression is also slightly induced after 24 h of progesterone treatment, while a more dramatic increase of FasL expression is demonstrated 24 h (343%) and 48 h (280%) after progesterone removal from the media. The peak of the FasL expression is observed 24 h following removal of progesterone. Lane 1, control prior to progesterone treatment; lane 2, cultured Ishikawa cells in the presence of progesterone for 24 h; lane 3, 24 h after withdrawal of progesterone; lane 4, 48 h after progesterone withdrawal. Bottom panel: Beta-actin levels in corresponding samples show no significant changes in the presence or absence of progesterone.

inhibitory effects on cell apoptosis within the endometrium and stroma [35]. Fas-FasL interaction in the same endometrial glandular cell is suggested to be a possible autocrine mechanism of cell death in this process [24]. Endometrial hyperplasia represents an imbalance between stimulatory factors (mainly estrogen) and inhibitory factors (mainly progesterone) in the endometrium. However, it is still unclear what exactly happens in the process of endometrial hyperplasia at the molecular level. Previous studies on Fas/ FasL expression in different phases of the menstrual cycle revealed a synchronous expression of Fas and FasL in glandular cells throughout the menstrual cycle [19,36]. This diffuse Fas/FasL expression pattern suggests that endometrial glandular apoptosis might be mediated by the Fas/FasL system [19,22,37]. In the current study, we found Fas expression in only 4 (15%) of 26 cases, while FasL was expressed in the majority cases of preprogestin-treated endometrial hyperplasia samples. From this perspective, reduced Fas expression in endometrial epithelium may represent one of the early steps in the development of endometrial hyperplasia. It is believed that there are also interactions between the antiapoptotic gene bcl-2 and the apoptotic Fas/FasL systems. Previous studies by others have shown that Fas/FasL expression and

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

115

Fig. 3. Effect of progestin treatment and withdrawal on Fas/FasL-mediated apoptosis in Ishikawa cells. The same experimental conditions as described for Western blot were used for detection of apoptotic cells with the TUNEL assay. No apoptotic cells are seen in control IgG-treated cells (A). The number of apoptotic cells is slightly increased when progesterone is present for 24 h (B). A dramatic increase in apoptotic cells is seen after progesterone withdrawal for 24 h.

activation are associated with cyclic changes of bcl-2 in the menstrual cycles [19,36]. Our previous endometrial hyperplasia study has shown that persistent bcl-2 expression is present in hyperplastic endometrium [16]. It is known that the functional site of bcl-2 is upstream of interleukin-converting enzyme (ICE)-like protease, and bcl-2 inhibits Fasmediated apoptosis by inactivating the ICE-like protease which is located downstream of Fas/FasL in the apoptotic pathway [38,39]. Therefore, in conjunction with these previous findings, the current results suggest that both persistent bcl-2 expression and an imbalance in expression of the Fas and FasL system play a role in the responsiveness to progestin. Furthermore, persistent bcl-2 expression in hyperplastic endometrial glands may be caused by reduced Fas expression. This assumption is further supported by Maruo’s study demonstrating increased Fas expression and decreased bcl-2 expression in the endometrium when a progestin-releasing device is inserted into endometrium (levo-orgestrel-releasing intrauterine system) [40]. Administration of progestational agents is a well-established mode of therapy in the treatment of endometrial hyperplasia. The induction of apoptosis has been found to be an important component for the therapeutic effect of progestins in our previous studies [16,17]. In our established paired pre- and postprogestin treatment model, we have found that endometrial cell apoptosis may be an early event after progestin treatment [17]. To further broaden our understanding of the progestin effect and the mechanisms of progestin resistance, we compared the Fas/FasL expression in this paired model system, in which many confounding factors are well-controlled, since pre- and postprogestintreated samples were from the same patients. Our data showed that Fas expression was significantly increased in responders compared to nonresponders (Tables 2 and 3). In contrast to Fas expression, FasL showed no significant difference between pre- and postprogestin samples or between responders and nonresponders. This imbalanced expression of Fas and FasL system after administration of progestin may represent one of the intrinsic abnormalities in the process of endometrial hyperplasia development. Undefined

alterations in factors that participate in the control of Fas/ FasL could exist which dissociate the expression of these apoptosisrelated membrane proteins. Imbalanced expression of Fas/FasL system in hyperplastic endometrium could contribute to depressed cellular response to the effects of progetins. In addition to this autocrine-related apoptosis in endometrial glandular cells, dysregulation of Fas/FasL may also be involved in emergence of progestinresistant cells under the pressure of progestin selection. The induction of apoptosis is likely a mechanism that would account for the marked decrease in glandular volume seen with progestin therapy. However, our finding of insignificant difference of FasL expression between pre- and postprogestin-treated samples is not readily explained. It is generally recognized that expression of Fas alone does not guarantee activation of Fas-mediated apoptosis. The level of synchronous FasL expression is also important in activating the Fas/FasL apoptotic pathway [42]. In this study, the FasL expression remained unchanged in postprogestin-treated samples. Nevertheless, in cell culture Ishikawa cells showed an increase of both FasL and Fas simultaneously during the time of progesterone treatment. There are three plausible reasons to explain the discrepancy between the in vivo and the in vitro findings. First of all, endometrial hyperplasia is mainly due to prolonged estrogen stimulation. It is known that FasL is up-regulated by estrogen. This is consistent with our finding that high levels of FasL were present in the pretreated hyperplastic samples compared to the posttreatment samples. When FasL expression is high, additional augmentation of FasL expression by progestin may be diminished. In addition, the active form of soluble FasL in human endometrium must be cleaved and released from the cell surfaces by matrix metalloproteinase [41]. Therefore, the surface expression of FasL may not parallel the biologic function of FasL in human endometrium. Second, no additional increment of FasL may represent an intrinsic impairment of the hyperplastic endometrium. Accordingly, this imbalance of Fas and FasL expression in hyperplastic endometrium may make it easier for the hyperplastic cells to survive and grow in spite of various antitumor mechanisms

116

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

present in vivo [42]. This process may ultimately contribute to the development of endometrial carcinoma. Third, dramatic up-regulation of FasL after progestin treatment in vivo may not be detected simply because none of the endometrial samples were studied in the first week or first few days after progestin therapy. Our endometrial samples were examined at least 1 month after the therapy. The significance of Fas and FasL in endometrial stromal cells is unclear. Many previous studies did not show stromal Fas and FasL expression in cyclic endometrium except for a study by Selam and associates [19,24,43]. In the later study, FasL expression in benign endometrial stromal cells increased gradually through the mid- and late-proliferative and secretory phases of the endometrium [43]. We did not observe significant changes of stromal Fas and FasL expression in pre- and post-progestin-treated samples. The determination of the role of stromal cell apoptosis in endometrial hyperplasia and their response to progestin treatment will require further study. Epithelial cells entrapped in the glandular lumens usually show apoptotic appearance. These cells typically show strong diffuse Fas and FasL expression. Since this phenomenon is sporadically present in both pre- and postprogestintreated samples, we speculate that there may have been a certain cell turnover rate in hyperplastic endometrium. The significance of this cell turnover in the pathogenesis of endometrial hyperplasia remains to be defined. Probably the most intriguing finding of this study is that the induction of Fas and FasL is significantly increased when progesterone is withdrawn compared to the levels of Fas and FasL expression when progesterone is present in the medium of cultured Ishikawa cells. Our in vitro Ishikawa cell culture studies showed a slight increase (10 –20%) in Fas and FasL protein expression after 24 h of progesterone treatment, which is comparable to our previous finding [17]. In our previous study we also found that apoptotic cells increase by 45 and 35% when progesterone is present for 48 and 72 h, respectively [17]. However, more significant increases in both Fas and FasL levels were found 24 and 48 h after withdrawal of the progesterone. The levels of Fas and FasL induction after progesterone withdraw parallels that of the apoptotic activity of Ishikawa cells. It is a well-observed clinical fact that menstruation or uterine bleeding occurs when progesterone levels decrease sharply. Treatment with progestins causes inhibition of normal menstruation, but breakthrough bleeding can occur. Based on many previous studies, including ours, we believe that progesterone does directly induce apoptosis in endometrial cells [16, 17,19,36]. It seems, however, that this apoptotic activity is not enough to cause uterine bleeding when progesterone is continuously present. If all of the above observations are true, the progestin-induced endometrial apoptosis seems much more significant with progesterone withdrawal than when progesterone is continuously present. This finding raises concerns over the efficacy of the clinical management of endometrial hyperplasia or lowgrade endometrial cancer

with progestins. Currently, continuous large dose of highpotency progestin (medroxyprogesterone acetate, 10 mg/ day) for 3 to 6 months is standard for endometrial hyperplasia [44]. The nonresponse rate as high as 30% for progestin treatment may be partially due to the inefficiency of apoptosis induction by the continuous progestin application. Cyclic withdrawal of progestin, in a monthly or biweekly fashion, may substantially increase the efficacy of the treatment. Since this is a simple alteration of clinical management, changing from continuous to cyclic application of progestins, if the clinical outcome dramatically improves, more studies to confirm our findings and clinical trials toward this direction are urgently needed. In conclusion, the above results have the following implications. An imbalanced expression of the Fas/FasL system may play a role in the responsiveness to progestins. Up-regulation of Fas and FasL expression may represent one of the main molecular mechanisms mediating tissue remodeling of the therapeutic effect of progestins. Elevated tissue levels of Fas expression may serve as a progestin treatment efficacy marker when the progestin-treated endometrium samples are evaluated. Fas/FasL-induced apoptotic process may be impaired through asynchronous expression of Fas and FasL on endometrial epithelial cells. This impaired endometrial glandular apoptosis in response to progestin treatment may be part of the molecular mechanisms for the emergence of progestin-resistant cells, and may ultimately contribute to the development of endometrial carcinoma. More importantly, cyclic application of progestins in the clinical management of endometrial hyperplasia or lowgrade cancers may significantly improve the outcome of progestin therapy.

References [1] Kurman RJ, Kaminski PF, Norris HJ. The behavior of endometrial hyperplasia. A long-term study of “untreated” hyperplasia in 170 patients. Cancer 1985;56:403–12. [2] Wentz WB. Progestin therapy in endometrial hyperplasia. Gynecol Oncol 1974;2:362–7. [3] Gal D, Edman CD, Vellios F, Forney JP. Long-term effect of megestrol acetate in the treatment of endometrial hyperplasia. Am J Obstet Gynecol 1983;146:316 –21. [4] Heikkinen J, Vaheri R, Kainulainen P, Timonen U. Long-term continuous combined hormone replacement therapy in the prevention of postmenopausal bone loss: a comparison of high- and low-dose estrogen-progestin regimens. Osteoporosis Int 2000;11:929 –37. [5] Kistner RW. The effects of progestational agents on hyperplasia and carcinoma in situ of the endometrium. Int J Gynaecol Obstet 1970; 8:561–572. [6] Deligdisch L. Effects of hormone therapy on the endometrium. Mod Pathol 1993;6:94 –106. [7] Ferenczy A, Gelnd M. The biological significance of cytologic atypia in progestogen-treated endometrial hyperplasia. Am J Obstet Gynecol 1989;160:126 –31. [8] Deligdisch L. Hormonal pathology of the endometrium. Modern Pathol 2000;13:285–94. [9] Benraad TJ, Friberg LG, Koenders AJ, Kullander S. Do estrogen and progesterone receptors (E2R and PR) in metastasizing endometrial

S. Wang et al. / Gynecologic Oncology 88 (2003) 108 –117

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

cancers predict the response to gestagen therapy? Acta Obstet Gynecol Scand 1980;59:155–9. Ehrlich CE, Young PCM, Cleary RE. Cytoplasmic progesterone and estradiol receptors in normal, hyperplastic, and carcinomatous endometria: therapeutic implications. Am J Obstet Gynecol 1981;141: 539 – 46. Martin PM, Rolland PH, Gammerre M, Serment H, Toga M. Estradiol and progesterone receptors in normal and neoplastic endometrium: correlations between receptors, histopathological examinations and clinical responses under progestin therapy. Int J Cancer 1979;23: 321–9. Metzger DA, Lessey BA, Soper JT, McCarty KS, Haney AF. Hormone-resistant endometriosis following total abdominal hysterectomy and bilateral salpingo-oophorectomy: correlation with histology and steroid receptor content. Obstet Gynecol 1991;78:946 –50. Murphy LC, Dotzlaw H, Wong MS, Miller T, Murphy LJ. Mechanisms involved in the evolution of progestin resistance in human breast cancer cells. Cancer Res 1991;51:2051–7. Satyaswaroop PG, Carke CL, Zaino RJ, Mortel R. Apparent resistance in human endometrial carcinoma during combination treatment with tamoxifen and progestin may result from desensitization following downregulation of tumor progesterone receptor. Cancer Lett 1992;62:107–14. Zupi E, Sbracia M, Marconi D, Alo P, Atzei G, Brazolin T, et al. TNF alpha expression in hyperplastic endometrium. Am J Reprod Immunol (Copenhagen) 2000;44:153–9. Amezcua CA, Muderspach LI, Felix JC, Zheng W. Down-regulation of bcl-2 is a potential marker of the efficacy of progestin therapy in the treatment of endometrial hyperplasia. Gynecol Oncol 1999;73: 126 –36. Amezcua CA, Lu JJ, Stanczyk FZ, Zheng W. Apoptosis may be an early event of progestin therapy for endometrial hyperplasia. Gynecol Oncol 2000;79:169 –76. Peng X, Maruo T, Matsuo H, Takekida S, Deguchi J. Serum deprivation-induced apoptosis in cultured porcine granulosa cells is characterized by increased expression of p53 protein, Fas antigen and Fas ligand and by decreased expression of PCNA. Endocrine J 1998;45: 247–53. Yamashita H, Otsuki Y, Matsumoto K, Ueki K, Ueki M. Fas ligand, Fas antigen and Bcl-2 expression in human endometrium during the menstrual cycle. Mol Human Reprod 1999;5:358 – 64. Itoh N, Yonehara S, Ishii A, Yonehara M, Mizushima S, Sameshima M, et al. The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 1991;66:233– 43. Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family. Cell 1993;75:1169 –78. Watanabe H, Kanzaki H, Narukawa S, Inoue T, Katsuragawa H, Kaneko Y, et al. Bcl-2 and fas expression in eutopic and ectopic human endometrium during the menstrual cycle in relation to endometrial cell apoptosis. Am J Obstet Gynecol 1997;176:360 – 8. Song J, Sapi E, Brown W, Nilsen J, Tartaro K, Kacinski BM, et al. Roles of Fas and Fas ligand during mammary gland remodeling. J Clin Invest 2000;106:1209 –20. Nagata S. Fas-induced apoptosis and diseases caused by its abnormality. Genes Cells 1996;1:873–9. Boldin MP, Varfolomeev EE, Pancer Z, Mett IL, Camonis JH, Wallach D. A novel protein that interacts with the death domain of Fas/APO1 contains a sequence motif related to the death domain. J Biol Chem 1995;270:7795– 8. Muzio M, Chinnaiyan AM, Kischkel FC, O’Rourke K, Shevchenko A, Ni J, et al. FLICE, a novel FADD-homologous ICE/CED-3-like protease, is recruited to the CD95 (Fas/APO-1) death-inducing signaling complex. Cell 1996;85:817–27.

117

[27] Kuwana T, Smith JJ, Muzio M, Dixit V, Newmeyer DD, Kornbluth S. Apoptosis induction by caspase-8 is amplified through the mitochondrial release of cytochrome c. J Biol Chem 1998;273:16589 –94. [28] Slayden OD, Rubin JS, Lacey DL, Brenner RM. Effects of keratinocyte growth factor in the endometrium of rhesus macaques during the lutealfollicular transition. J Clin Endocrinol Metab 2000;85:275– 85. [29] Rudolph-Owen LA, Slayden OD, Matrisian LM, Brenner RM. Matrix metalloproteinase expression in Macaca mulatta endometrium: evidence for zone-specific regulatory tissue gradients. Biol Reprod 1998; 59:1349 –59. [30] Zheng W, Feng YJ, Gandhi M, Siu S, Hom E, Caputo T, et al. Persistent expression of bcl-2 onco-protein in endometrial carcinoma correlates with hormone receptor positivity. Int J Gynecol Cancer 1996;6:235– 40. [31] Zheng W, Cao P, Zheng M, Kramer EE, Godwin T. p53 and bcl-2 expression in endometrial adenocarcinoma. Comparison of serous and endometrioid subtypes (see comments). Gynecol Oncol 1996;61: 167–74. [32] Song J, Sapi E, Brown W, Nilsen J, Tartaro K, Kacinski B, et al. Expression of Fas and Fas ligand during pregnancy, lactation and involution and its potential role during mammary gland remodeling. J Clin Invest 2000;106:1209 –24. [33] Mor G, Kohen F, Garcia-Velasco J, Nilsen J, Brown W, Song J, et al. Regulation of Fas ligand expression in breast cancer cells by estrogen: functional differences between estradiol and tamoxifen. J Steroid Biochem Mol Biol 2000;73:185–94. [34] Zheng W, Butwell T, Hecket L, Griswold M, Bellve AR. Pleiotypic actions of the seminiferous growth factor on two testis cell lines: comparisons with acidic and basic fibroblast growth factors. Growth Factor 1990;3:73– 82. [35] Kokawa K, Shikone T, Nakano R. Apoptosis in the human uterine endometrium during the menstrual cycle. J Clin Endocrinol Metab 1996;81:4144 –7. [36] Rogers PA, Lederman F, Plunkett D, Affandi B. Bcl-2, Fas and caspase 3 expression in endometrium from levonorgestrel implant users with and without breakthrough bleeding. Human Reprod 2000; 15(Suppl 3):152– 61. [37] Tabibzadeh S, Zupi E, Babaknia A, Liu R, Marconi D, Romanini C. Site and menstrual cycle-dependent expression of proteins of the tumour necrosis factor (TNF) receptor family, and BCL-2 oncoprotein and phasespecific production of TNF alpha in human endometrium. Human Reprod 1995;10:277– 86. [38] Shimizu S, Eguchi Y, Kamiike W, Matsuda H, Tsujimoto Y. Bcl-2 expression prevents activation of the ICE protease cascade. Oncogene 1996;12:2251–7. [39] Boise LH, Thompson CB. Bcl-x(L) can inhibit apoptosis in cells that have undergone Fas-induced protease activation. Proc Natl Acad Sci USA 1997;94:3759 – 64. [40] Maruo T, Laoag-Fernandez JB, Pakarinen P, Murakoshi H, Spitz IM. Effects of the levonorgestrel-releasing intrauterine system on proliferation and apoptosis in the endometrium. Human Reprod 2001;16: 2103– 8. [41] Kayagaki N, Kawasaki A, Ebata T, Ohmoto H, Ikeda S, Inoue S, et al. Metalloproteinase-mediated release of human Fas ligand. J Exp Med 1995;182:1777– 81. [42] Zornig M, Grzeschiczek A, Kowaslski K-B, Hartmann K-U, Moroy T. Loss of Fas/APO-1 receptor accelerates lymphomagenesis in Eu L-myc transgenic mice but not in animals infected with MoMuLV. Oncogene 1995;10:2397– 401. [43] Selam B, Kayishi UA, Mulayim N, Arici A. Regulation of Fas ligand expression by estradiol and progesterone in human endometrium. Biol Reprod 2001;65:979 – 85. [44] Tierney LM, Mcphee SJ, Papadakis MA. Endometrial hyperplasia. In: Current medical diagnosis and treatment, 41st ed. 747. 2002.