Loss of phosphatase and tensin homologue deleted on chromosome 10 and phosphorylation of mammalian target of rapamycin are associated with progesterone refractory endometrial hyperplasia

Authors


  • This work was presented at the 42nd Annual Meeting of the American Society of Clinical Oncology, Orlando, Florida, May 13–17, 2005.

Karen H. Lu, MD, Department of Gynecologic Oncology, University of Texas MD Anderson Cancer Center, 1155 Herman P. Pressler Street, CPB 6.3244, Unit 1362, Houston, TX 77030, USA. Email: khlu@mdanderson.org

Abstract

The objective of our study was to evaluate the phosphatase and tensin homologue deleted on chromosome 10 (PTEN), p27, and mammalian target of rapamycin (mTOR) expressions in women with progesterone-responsive and refractory endometrial hyperplasia (EH) samples and to determine if these markers could be associated with response or used as potential targets for treatment. Thirty-eight matched pre- and posttreatment pairs of paraffin-embedded endometrial biopsies were obtained from patients with EH. Immunohistochemical analysis for PTEN, p27, and phospho-mTOR were performed on all samples. Median age at diagnosis was 49 years (20–79 years). Median treatment interval was 3 months (1–12 months). Sixteen patients (42.1%) had complete resolution of their hyperplasia (responders), and 22 (57.9%) had persistent hyperplasia (nonresponders) after treatment with progesterone. In the pretreatment samples, no markers were found to predict nonresponders. In posttreatment samples, loss of PTEN expression with phospho-mTOR expression was observed in more nonresponders than responders (40.9% vs 6.3%; P= 0.03). Phospho-mTOR overexpression was found in 63.6% of nonresponders. We found that persistent hyperplasia refractory to progesterone therapy was associated both with the loss of PTEN and with the loss of phosphorylation of mTOR. In select cases of non–responsive progesterone refractory EH, a rational target for treatment may involve the mTOR pathway.

Endometrial hyperplasia (EH) and specifically complex atypical EH is a premalignant lesion that can progress to endometrial cancer (EC)(1). In a study examining the natural history of EH, Kurman et al.(1) found that 1% of women with simple hyperplasia and 29% of women with complex atypical EH progressed to EC(1).

The American College of Obstetrics and Gynecology has recently published a practice bulletin regarding the management of EC, which includes management of EH patients with progesterone. These recommendations include oral, parental, or intrauterine progesterone agents and have stressed serial endometrial biopsies every 3 months while on progesterone therapy(2). Response rates have varied but underscore the fact that we are currently unable to clinically determine which patients will respond to progesterone therapy based only on histologic diagnosis of EH.

The phosphatase and tensin homologue deleted on chromosome 10 (PTEN) is a tumor suppressor protein that opposes the pathway of phosphoinositide 3-kinase/Akt/mammalian target of rapamycin (mTOR) through dephosphorylation and subsequent decreased downstream activation of mTOR(3). PTEN mutations are the most common abnormality associated with both EH and EC. PTEN mutations have been shown in 34–83% of human ECs and in up to 55% of EH, a precursor of EC(4–6).

Loss of PTEN leads to activation of Akt, which in turn phosphorylates the downstream protein mTOR resulting in cell proliferation(7). mTOR phosphorylation activates downstream S6K, which is involved in synthesis of proteins and subsequent cell growth(7). mTOR inhibitors are currently being studied for treatment of EC and other solid tumors(3).

PTEN may also act as a suppressant of tumor activity through activation and increased expression of p27. Decreased p27 expression has been associated with alteration in PTEN in human ECs(8). p27 is a cyclin-dependent kinase inhibitor, which may act as a tumor suppressor and may also be involved with EH. PTEN and p27 expressions have recently been shown to be decreased in unmatched EH specimens compared with controls of proliferative endometrium(9).

The objective of our study was to evaluate patients who had successful response to progesterone treatment vs progesterone refractory cases, and to determine whether there are markers associated with response to therapy. In addition, we were interested in determining if mTOR inhibition is a rational target for the treatment of progesterone refractory EH.

Material and methods

After institutional review board approval from the participating institutions, 38 matched preprogesterone- and postprogesterone-treated pairs of paraffin-embedded endometrial biopsies were obtained from patients with EH from MD Anderson Cancer Center, Houston, Texas, and Kaiser Permanente Medical Center, San Francisco, California. Hematoxylin and eosin–stained sections of endometrium were reviewed by a gynecological pathologist (R.R.B.) blinded to the previous diagnosis. Inclusion in this EH specimen study included availability of both pre- and posttreatment hyperplasia blocks. Exclusion criteria included disagreement of hyperplasia diagnosis from primary pathology reports when compared with central pathology review and limited material for immunohistochemistry evaluation.

Biopsy specimens from both pre- and posttreated specimens were classified as normal endometrium, simple hyperplasia, simple hyperplasia with atypia, complex hyperplasia, and complex atypical hyperplasia (AEH). The criteria for evaluation of hyperplasia followed the World Health Organization’s classification of EH. Patients were divided into study groups based on response to progesterone therapy. Response was defined as complete resolution of hyperplasia.

Immunohistochemical analysis for PTEN (Cascade Bioscience, Winchester, MA) at 1:300 concentration, p27 (BD Biosciences Pharmingen, San Diego, CA) at 1:200 concentration, and phospho-mTOR (Cell Signaling Technology, Beverly, MA) at 1:50 concentration was performed on all samples of formalin-fixed, paraffin-embedded endometrium using commercially available monoclonal antibodies. The involved sections were deparaffinized in xylene, rehydrated through graded alcohols, and transferred to phosphate-buffered saline. All samples were blocked using 3% of H2O2 in methanol for 15 min and rinsed with phosphate-buffered saline. Slides were then blocked using 1% normal goat serum for 20 min. Slides were incubated overnight at 4°C at 1:300 concentration of PTEN, 1:200 concentration of p27, and 1:50 concentration of phospho-mTOR. This was followed by a standard avidin–biotin peroxidase complex assay using Vector ABC Elite Kit (Vector Laboratories, Burlingame, CA). Slides were developed with diaminobenzidene (DAB) (Zymed Laboratories, San Francisco, CA) and counter stained with 10% hematoxylin. Loss of PTEN expression was determined if more than 10% of the glands did not express PTEN and was modified from previous work from Zheng et al.(10). Positive phospho-mTOR was defined as moderate (2+) to strong (3+) staining in at least 20% of the glands(11). Positive p27 was defined as moderate (2+) to strong (3+) staining in at least 50% of the glands and was adapted from previous work that described evaluation of positive p27 status by immunohistochemistry(9).

For statistical analysis, Chi-square or Fisher exact test was used to compare the proportions between groups, and t tests were used for continuous variables. A P value <0.05 was considered statistically significant. All analyses were performed using SPSS 12.0 (Chicago, IL). No adjustment was made for multiple comparisons.

Results

Only 38 patients met inclusion criteria and were included in the analysis. Median age at diagnosis was 49 years (20–79 years). Progesterone treatment varied and included megestrol, norethindrone, and medroxyprogesterone (Table 1). Median treatment interval was 3 months (1–12 months). Sixteen patients (42.1%) had complete resolution of their hyperplasia (responders), and 22 (57.9%) had persistent hyperplasia despite treatment with progesterone (nonresponders). Although limited by sample size and incomplete clinical data, neither age (mean of 45.7 years for responders vs 48.5 years for nonresponders; P= 0.25) nor treatment duration (mean of 3.36 months for responders vs 5 months for nonresponders; P= 0.08) were associated with response of EH to progesterone.

Table 1.  Clinical characteristics, treatment, and response including age, progesterone therapy, and defined response
PatientAge, yearsInitial diagnosisHormone (months)Response
  • SH, simple hyperplasia; SAH, simple hyperplasia with atypia; CH, complex hyperplasia; n/a, not available.

  • a

    Endometrial cancer.

128AEHMegestrol (5)AEH/grade 1a
261CHMegestrol (3)Normal
331CHMegestrol (3)Normal
468SHMegestrol (11)Normal
574CHMedroxyprogesterone (n/a)Normal
653AEHMegestrol (4)CH/grade 1a
769CHMedroxyprogesterone (n/a)SH
853SHMedroxyprogesterone (n/a)Normal
979CHMegestrol (12)SH
1047CHMedroxyprogesterone (4)CH
1174SAHMegestrol (3)SH
1260CHMedroxyprogesterone (3)Normal
1320CHMegestrol (3)Normal
1430AEHMegestrol (3)Grade 1a
1565CHMedroxyprogesterone (6)Normal
1626AEHMegestrol (3)Grade 1a
1723AEHMegestrol (3)AEH/grade 1a
1846AEHMedroxyprogesterone (3)Normal
1944AEHNorethindrone (3)Normal
2049AEHMedroxyprogesterone (3)CH
2140AEHMedroxyprogesterone (4)CH
2273CHMedroxyprogesterone (12)CH
2345AEHMegestrol (2)AEH
2454AEHMedroxyprogesterone (6)AEH
2542AEHMegestrol (3)Normal
2649AEHMedroxyprogesterone (3)Normal
2748AEHMegestrol (3)Normal
2842AEHMedroxyprogesterone (3)Normal
2949AEHMegestrol (n/a)CH
3046CHNorethindrone (3)AEH
3139AEHMedroxyprogesterone (6)AEH
3245AEHMegestrol (3)CH
3351AEHMegestrol (1)Normal
3453AEHMegestrol (3)AEH
3561AEHMedroxyprogesterone (6)Normal
3636AEHNorethindrone (6)AEH
3741AEHNorethindrone (n/a)AEH
3847AEHNorethindrone + medroxyprogesterone (3)CH

Response to progesterone was noted to be increased in patients without AEH compared with patients with AEH (8/14 or 57.1% vs 8/24 or 33.3%; P= 0.19), but this result did not reach statistical significance (Table 2). Five cases of progesterone refractory EH were also noted to have a posttreatment diagnosis of EC (Table 1).

Table 2.  Pretreatment pathology diagnosis and response to progesterone therapy
Pretreatment diagnosisRespondersNonrespondersPercentage response
  1. SAH, simple hyperplasia with atypia; CH, complex hyperplasia; SH, simple hyperplasia.

Complex atypical EH81633
SAH010
CH6555
SH20100
Total162242

Pretreatment expression of PTEN, p27, and phospho-mTOR was similar between responders and nonresponders (Table 3). Positive phospho-mTOR expression was observed in 25 of 38 (65.8%) of pretreatment and 23 of 38 (60.5%) of posttreatment samples (Table 3). Table 3 shows that none of the biologic markers investigated, including PTEN, phospho-mTOR, or p27, were predictive of response to progesterone therapy or of a subsequent diagnosis of EC.

Table 3.  PTEN, p27, and mTOR expression in pre- and posttreatment progesterone specimensa
PTEN, p27, or mTOR expressionResponders, N= 16, n (%)Nonresponders, N= 22, n (%)P value
  1. aValues in table are numbers of patients (percentages) unless otherwise indicated.

  2. bSignificance at P < 0.05. No adjustment for multiple comparisons has been made.

Pretreatment loss of PTEN expression8 (50)9 (40.9)0.74
Pretreatment loss of p27 expression6 (37.5)10 (45.5)0.74
Pretreatment phospho-mTOR expression12 (75)13 (59.1)0.49
Pretreatment loss of PTEN expression and phospho-mTOR overexpression7 (43.8)6 (27.3)0.32
Pretreatment loss of PTEN and p27 expression5 (31.3)4 (18.2)0.45
Posttreatment loss of PTEN expression4 (25)12 (54.5)0.10
Posttreatment loss of p27 expression10 (62.5)10 (45.5)0.34
Posttreatment phospho-mTOR expression9 (56.3)14 (63.6)0.74
Posttreatment loss of PTEN expression and phospho-mTOR overexpression1 (6.3)9 (40.9)0.03b
Posttreatment loss of PTEN and p27 expression4 (25)6 (27.3)1.00

In posttreatment samples, it was observed that a loss of PTEN expression (Fig. 1) was greater in nonresponders compared with responders (54.5% vs 25%; P= 0.10), but this result, however, did not reach statistical significance. In addition, increased or persistent positive phospho-mTOR expression was observed in a greater percentage of nonresponders than responders in posttreatment samples, but this result did not reach statistical significance (Fig. 2). Expression of p27 in the posttreatment samples was greater in the responders compared with the nonresponders, but this was also not statistically significant. Table 3 shows that the combination of loss in PTEN expression and overexpression of phospho-mTOR was more likely to be in the nonresponder group compared with the responder group (40.9% vs 6.3%; P= 0.03).

Figure 1.

Involution of the PTEN-negative glands after successful treatment with progesterone therapy (hematoxylin and eosin stained, magnification ×400 and ×200).

Figure 2.

Increase in mTOR expression following treatment with progesterone in a patient with persistent EH (hematoxylin and eosin stained, magnification ×200 and ×200).

An analysis of the paired specimens was undertaken with comparison of the change of PTEN, p27, or phospho-mTOR expression between pre- and posttreatment (Table 4). Although not statistically significant, those patients that were observed to have an increased PTEN expression in the posttreatment compared with pretreatment specimens were somewhat more likely to respond to progesterone (31.3% vs 4.5%; P= 0.07). The paired analysis of persistent or increased phospho-mTOR overexpression (56.3% vs 63.6%; P= 0.74) was higher in the progesterone refractory group, but this result did not reach statistical significance. Table 4 shows that increased p27 expression and combined loss of PTEN with persistent or increased phospho-mTOR overexpression were not associated with response to progesterone.

Table 4.  Associations between PTEN, p27, and mTOR expression in pre- and posttreatment progesterone specimensa
PTEN, p27, or mTOR expressionResponders, N= 16, n (%)Nonresponders, N= 16, n (%)P value
  • a

    Values in table are numbers of patients (percentages) unless otherwise indicated.

Increased PTEN expression after progesterone therapy5 (31.3)1 (4.5)0.07
Increased phospho-mTOR overexpression after progesterone therapy9 (56.3)14 (63.6)0.74
Decreased PTEN expression and persistent/increased phospho-mTOR overexpression1 (6.3)4 (18.2)0.37
Increased p27 expression after progesterone therapy3 (18.8)6 (27.3)0.71

Comment

Progesterone therapy is recommended for the management of complex atypical EH and early EC in patients who are high-risk operative candidates or who wish to maintain their fertility(2). Progesterone is believed to cause involution of the endometrial glands through pathways involved in apoptosis(12). The rate of response to progesterone therapy may be influenced by dose, length of treatment, and presence of cytologic atypia.

Ferenczy and Gelfand demonstrated that cytologic atypia in EH was a prognostic factor for progestin therapy response. The rate of response to progestin was significantly greater in patients without atypia compared with those who had EH with atypia (52/65 or 80% vs 5/20 or 25%; P < 0.01)(13). Randall and Kurman evaluated 17 paired patient specimens with a diagnosis of EH with atypia and reported a 94.1% (16/17) response rate with progesterone therapy; however, this response rate also included three cases of hyperplasia without atypia(14). Horn et al., in a recent study from Germany, evaluated 560 progesterone-treated patients with EH, which included a subset of patients with complex hyperplasia with atypia. This study demonstrated that among the seven cases with AEH treated with progestin, there were no responses and three patients progressed to EC(15).

There was clearly a broad range of response rates, but consistently, AEH with atypia was associated with a decreased response. We found that following treatment with progesterone, 16 patients (42.1%) had complete resolution of their hyperplasia (responders) and 22 (57.9%) had persistent hyperplasia or progression (nonresponders). Although not statistically significant in our study, response to progesterone was noted to be less common in patients with AEH compared with patients without atypical hyperplasia.

We hypothesize that the different molecular pathways exist in responders vs nonresponders to progesterone therapy for EH. Identifying a molecular marker to predict response or nonresponse would be clinically useful. Other markers that have been evaluated in paired EH specimens include Bcl-2 expression (a protein marker involved in resistance to apoptosis) and Fas/FasL expression (involved in the progression to apoptosis and linked to the menstrual cycle) and found significant associations with apoptosis in EH and the response to progesterone therapy(12,16,17). These studies underscore the potential of a molecular marker in EH that would help identify response to progesterone therapy.

Erkanli et al. studied the expression of PTEN, p27, and survivin in unmatched 10 normal endometrium cases and 38 EH cases. They demonstrated that compared with proliferative endometrium, there was a statistically significant loss of both PTEN and p27 in EH(9). These findings suggest that loss of PTEN and p27 may be important first steps toward carcinogenesis. Although not statistically significant, our findings did demonstrate a difference in PTEN expression in those patients who responded to progesterone compared with those who did not respond to treatment. We did not, however, demonstrate a difference with p27 expression, and further study will be needed to elucidate the association between p27 and EH.

In our study, the combination of loss of PTEN and positive phospho-mTOR expression was significantly associated with progesterone refractory hyperplasia. Although not statistically significant, we also found an increase in PTEN expression in those patients who responded to progesterone. This change in expression in pre- and posttreatment samples will not be useful to predict pretreatment response to progesterone, but this “signature” may be helpful in determining alternative treatments. Further study is needed to determine if protein marker expression in combination with histologic subtype can be used to predict response to progesterone treatment.

Targeted inhibition of mTOR with an intravenous agent CCI-779 (Wyeth, Madison, NJ) in the Pten heterozygote animal model has been shown to decrease endometrial hyperplastic lesions and decrease proliferation compared with vehicle and untreated groups(18). We recently studied an oral mTOR inhibitor RAD001 (Novartis, East Hanover, NJ) and demonstrated decreased EH in the Pten heterozygote murine model by increased apoptosis and decreased cell proliferation(19).

We found that after progesterone treatment, persistent hyperplasia refractory to progesterone therapy was associated both with the loss of PTEN and phosphorylation of mTOR. A larger sample size will be needed to validate the combination of loss of PTEN and overexpression of mTOR as a signature of progesterone refractory EH. In those cases of non–responsive progesterone refractory EH, continued clinical treatment with progesterone may not be effective. Our findings of decreased PTEN and phosphorylation of mTOR in progesterone refractory EH may also suggest a clinical role for mTOR inhibitors in this patient population.

Acknowledgment

This study was funded by the Specialized Programs of Research Excellence grant in uterine cancer.

Ancillary