Loss of PTEN expression is associated with metastatic disease in patients with endometrial carcinoma

Abstract

BACKGROUND

The PTEN tumor suppressor gene frequently is involved in endometrial carcinoma. Loss of heterozygosity and mutations reportedly are common, although the biologic importance of these changes remain largely unknown. The objective of this study was to assess the pattern of PTEN expression by immunohistochemistry in a large series of patients with endometrial carcinoma.

METHODS

A population-based series of 316 patients with endometrial carcinoma who had long and complete follow-up was investigated for PTEN expression and its correlation with clinicopathologic variables, tumor markers, and survival.

RESULTS

PTEN protein expression was mainly cytoplasmic in tumor cells, with no expression seen in 56 of 279 patients (20%) who had evaluable results. A heterogeneous staining pattern was found in 70 tumors (25%). A significant association between the loss of PTEN expression and metastatic disease was identified (P = 0.05). However, PTEN staining did not influence survival significantly.

CONCLUSIONS

The loss of PTEN expression is relatively frequent in endometrial carcinoma and is associated significantly with metastatic disease. This indicates that the PTEN system plays an important role in some endometrial carcinomas, but further studies of PTEN protein expression related to various genetic alterations are necessary. Cancer 2002;94:2185–91. © 2002 American Cancer Society.

DOI 10.1002/cncr.10434

The tumor suppressor gene PTEN (MMAC1), which has been cloned and mapped to chromosome 10q23,1, 2 reportedly is involved in the regulation of focal adhesion, cellular migration,3 and tumor cell proliferation.4 Germline PTEN mutations are found in autosomal-dominant Cowden disease, which is characterized by hamartomas and increased risk of breast and thyroid carcinoma,5 and somatic mutations of the PTEN gene have been identified in a large number of other human malignancies.1, 2, 6 In sporadic endometrial carcinomas, loss of heterozygosity of chromosome 10q23 was found initially in about 40% of tumors.7 Subsequently, PTEN mutations were identified in 30–50% of these tumors,8–10 and PTEN alterations may represent an early event in tumorigenesis.11 The objective of the current study was to investigate whether PTEN staining may represent a marker of specific clinicopathologic characteristics or biologic features of endometrial carcinomas.

MATERIALS AND METHODS

Patient Sample

All 316 patients who were diagnosed with endometrial carcinoma in Hordaland County, Norway, during 1981–1990 have been studied, as described previously.12 Briefly, the patients were staged retrospectively according to the 1988 International Federation of Gynecology and Obstetrics (FIGO) criteria, and all microscopic slides were reclassified and graded by one pathologist (L.A.A.) according to the 1994 World Health Organization criteria.12–14 In five patients, the diagnoses were based on cytologic examination only, with no histologic specimens available. Twelve patients were excluded due to changed diagnoses at reclassification. Of the remaining 299 patients, paraffin blocks were available for further investigations from 284 patients (95%).

DNA Ploidy, S-Phase Fraction, and Steroid Receptor Analyses

In a subset of the patient population, DNA ploidy, S-phase fraction, and steroid receptor concentrations had been estimated in fresh tumor tissue that was collected prospectively. Technical details and findings have been reported previously,13, 15 and data were available for comparison.

Immunoblotting

To compare the specificity for two different PTEN antibodies studied, cell lysates in sample buffer with known PTEN status (generously provided by Dr. Charis Eng16) were boiled for 10 minutes, and the proteins were separated in a 10% acrylamide sodium dodecyl sulfate-polyacrylamide electrophoresis gel using 5 μg per lane. The proteins were transferred to a nitrocellulose membrane (Bio-Rad, Cambridge, MA) at 30 mA overnight in Tris-glycine buffer. The membranes were then blocked in 5% dry milk with 0.05% Tween-20 and incubated with two different specific human anti-PTEN antibodies; A2B1 (SC7974; Santa Cruz Biotechnology, Santa Cruz, CA; 1:250 dilution) or clone 6H2.1 (Cascade BioScience, Winchester, MA; 1:400 dilution). Horseradish peroxidase-conjugated goat antimouse immunoglobulin G (IgG; H&L; Bio-Rad) were added at a dilution of 1:3000. The bands were visualized by enhanced chemiluminescence detection (ECL kit; Amersham, Braunschweig, Germany).

Immunohistochemistry

Hematoxylin and eosin-stained sections were used to select the most cellular part of the tumor and the area with lowest architectural differentiation in sections with heterogeneity. Thin sections (5 μm) of the selected formalin fixed and paraffin embedded specimens were then used for immunohistochemistry. Two different antibodies were studied. The sections were deparaffinized through toluol, to alcohol, to phosphate-buffered saline (PBS) followed by microwave epitope retrieval in citrate buffer, pH 6.0 (A2B1: 750 W for 7.5 minutes and 500 W three times for 5 minutes each; 6H2.1: 750 W for 7.5 minutes and 350 W four times for 5 minutes each). The sections were incubated for 1 hour at room temperature with the A2B1 PTEN antibody diluted 1:25 in PBS or with the 6H2.1 PTEN antibody diluted 1:300 in PBS. The staining procedure was performed by the Dako TechMate™ 500 slide processing equipment, and the sections were stained according to the standard avidin-biotin method recommended by the manufacturer (Dako, Copenhagen, Denmark) and counterstained with Harris hematoxylin. Sections from colon carcinoma that repeatedly had been found to express PTEN were included as positive controls. In addition, nuclei in stromal cells were used as internal positive controls. Negative controls, which were included in every run, were obtained by omitting the primary antibody and by using an irrelevant monoclonal mouse antibody (IgG1; code no. X0931; Dako). There were 279 samples that had evaluable PTEN staining for the A2B1 antibody with the presence of positively stained nuclei in stromal cells. In addition, a subset of samples from 138 patients also was stained with the 6H2.1 PTEN antibody. For these 138 patients, additional tumor tissue had been collected prospectively and frozen at primary surgery, and this subgroup has been examined for potential selection bias.13, 15, 17

Blinded for patient characteristics and outcome, the slides were examined with a standard light microscope for immunohistochemical staining by three of the authors (H.S., I.S., and L.A.A.). It is important to note that regions of atypical endometrial hyperplasia close to the tumors were not included when staining was recorded. The staining was located mainly in the tumor cell cytoplasm, and cytoplasmic PTEN staining was recorded by a semiquantitative and subjective grading system that considered both the intensity of staining and the proportion of tumor cells in the selected section that showed a positive reaction, as described previously for other markers.12 Briefly, intensity was recorded as 0 (no staining) to 3 (strong staining); and the percentage of cytoplasmic staining area was recorded as 0 (no tumor cells positive), 1 (positive staining in < 10% of tumor cells), 2 (positive staining in 10–50% of tumor cells), or 3 (positive staining in > 50% of tumor cells). A staining index was calculated as the product of staining intensity and staining area. On the basis of previous studies,16, 18 we particularly wanted to determine whether the loss of PTEN expression was associated with other markers of prognostic importance. Therefore, we compared the group with no cytoplasmic staining (index = 0) with the rest (index > 0).

We also examined the correlation between PTEN expression and other tumor markers for cell proliferation (Ki-67), cell cycle regulation (p53, p21, and p16), and angiogenesis (Factor VIII) that have been studied previously by immunohistochemistry for these tumors. The procedures and results for these markers have been described elsewhere.12–14

Statistics

Associations between categoric variables were examined by the Pearson chi-square test. Comparisons of median values of continuous variables between groups were evaluated by nonparametric tests (the Mann–Whitney test for two groups or the Kruskall–Wallis test for more than two groups). The median follow-up for survivors was 9 years (range, 4–15 years). None of the patients was lost due to insufficient follow-up data.12–14 Univariate analyses of time to recurrence (recurrence free survival) or time to death due to endometrial carcinoma (cause specific death) were performed using the product-limit procedure (Kaplan–Meier method), with the time of primary operation as the entry date, as reported previously.19 The log-rank test was used to compare the survival curves for groups of patients defined by categories of each variable.

Prior to performing the studies, it was calculated that, with a reported overall 5-year survival rate of 70% for patients with endometrial carcinoma, 300 tumors would need to be analyzed to achieve 80% power to detect a 20% change in survival due to altered PTEN expression if 10% of the tumors showed loss of PTEN expression.

Data were analyzed using the SPSS software package (SPSS, Inc., Chicago, IL). The research was approved by the Norwegian Data Inspectorate and the Institutional Review Board at the University of Chicago (Protocol 9457).

RESULTS

Although it does not guarantee that an antibody will behave with similar specificity in immunohistochemical studies, a Western blot analysis was performed to compare the specificity for two different commercially available human anti-PTEN antibodies (A2B1 and 6H2.1). Western blot analysis was performed on one PTEN positive cell line (T47D) and one PTEN negative cell line (M468) (Fig. 1).16 The results showed that both antibodies were specifically reactive against PTEN (Fig. 1, lanes 1 and 2). There were no bands in the PTEN negative cell line M468 for either antibody (Fig. 1, lanes 4 and 5).

Figure 1. Western blot analysis of PTEN positive (T47D) and PTEN negative (M468) cell lines: lane 1, T47D/clone 6H2.1 PTEN antibody; lane 2, T47D/A2B1 PTEN antibody; lane 3, T47D/no primary antibody; lane 4, M468/clone 6H2.1 PTEN antibody; lane 5, M468/A2B1 PTEN antibody. MW: molecular weight.

PTEN protein expression in tumor cells was mainly cytoplasmic, with no expression seen in 56 of 279 evaluable tumors (20%) using the A2B1 antibody or in 27 of 138 evaluable tumors (20%) using the 6H2.1 antibody. Loss of PTEN expression for the two different antibodies was correlated highly (P = 0.001). Ninety-five tumors (34%) had a cytoplasmic staining index of 1–2, 68 tumors (24%) had a staining index of 3–4, and 60 tumors (22%) had a staining index > 4 (A2B1 antibody). Examples of a tumor with strong cytoplasmic staining in the majority of tumor cells is shown in Figure 2A, and tumors with no PTEN expression in tumor cells are shown in Figure 2B,C. A heterogeneous staining pattern, defined as variation in staining that follows different morphologic patterns, or in adjacent glands with similar morphology was seen in 70 tumors (25%), as shown in Figure 2D.

Figure 2. PTEN (A2B1 antibody) expression in endometrial carcinoma. (A) Tumor with strong cytoplasmic staining in the majority of the tumor cells (original magnification, ×250). (B) Tumor with positively stained stromal cells but no PTEN expression in the cytoplasm of the tumor cells (original magnification, ×250). (C) Tumor with positively stained stromal cells but no PTEN expression in the cytoplasm of the tumor cells (original magnification, ×250). (D) Heterogeneous staining pattern showing heterogeneity in adjacent glands with similar morphology (original magnification, ×250).

Loss of PTEN staining was correlated significantly with higher FIGO stage, as shown in Table 1. The same pattern persisted when endometrioid tumors were analyzed separately in a subgroup analysis. There tended to be an association between loss of PTEN protein expression and serous papillary/clear cell histologic subtype in the largest series of tumors stained with the A2B1 antibody (P = 0.1); however, in the smaller sample set stained with the 6H2.1 antibody, this correlation was not observed (Table 1).

Table 1. PTEN Expression in Tumor Cells Related to International Federation of Gynecology and Obstetrics Stage and Histologic Type in a Population-Based Study of Patients with Endometrial Carcinoma

Variable	No. (%) positive for A2B1 PTEN antibodya			No. (%) positive for 6H2.1 PTEN antibodyb
	n (%) Index = 0	n (%) Index > 0	P valuec	n (%) Index = 0	n (%) Index > 0	P valuec
FIGO: International Federation of Gynecology and Obstetrics; n.s.: not significant. a Data were available from 279 patients. b Data were available from 138 patients. c Chi-square test. d Data were not available on FIGO stage from 1 patient.
FIGO staged
I/II	40 (18)	184 (82)	0.05	20 (17)	98 (83)	0.06
III/IV	16 (30)	38 (70)	—	7 (35)	13 (65)	—
Histologic type
Endometrioid, adenoacanthoma, adenosquamous	47 (19)	203 (81)	0.1	23 (19)	100 (81)	n.s.
Serous papillary, clear cell	9 (31)	20 (69)	—	4 (27)	11 (73)	—

No significant correlations were seen between PTEN expression (A2B1 and 6H2.1 antibody) and age, histologic grade, DNA index, S-phase fraction, estrogen receptor and progesterone receptor concentration, microvessel density or expression of Ki-67, p53, p21, or p16 proteins in these tumors. The presence of a heterogeneous staining pattern was found less frequently in tumors with loss of p16 expression in the largest series stained with the A2B1 antibody (P = 0.02), and the same tendency was seen in the smaller sample set stained with the 6H2.1 antibody, even if this correlation did not reach statistical significance (P = 0.1). None of the clinicopathologic variables or any of the other investigated tumor biomarkers was related to heterogeneity in the PTEN staining.

PTEN expression did not influence survival significantly in univariate survival analysis. The 5-year survival rate for patients with tumors that expressed PTEN was 77%, compared with 71% for patients with tumors that did not express PTEN, as illustrated for A2B1 in Figure 3 (P = 0.42). Among the 232 patients who underwent surgery with curative intent, the 5-year recurrence free survival rate was 83% for patients with tumors that expressed PTEN, compared with 81% for patients with tumors that did not express PTEN (P = 0.88). The same pattern persisted when the endometrioid tumors were analyzed separately in a subgroup analysis, with no significant influence on survival or recurrence free survival for any of the two antibodies studied. Patients with FIGO Stage I disease also were investigated separately in subgroup analyses without showing any survival difference. Because there have been very few PTEN immunostaining studies, we also performed additional survival analyses with other cut-off points of cytoplasmic staining. However, these analyses showed similar survival rates for all the subgroups. Heterogeneity in PTEN staining did not influence survival. Equivalent survival analyses were performed for the series of tumors that were stained with the 6H2.1 PTEN antibody, and that series also showed no significant influence on survival. The Kaplan–Meier survival curves are shown in Figure 3.

Figure 3. Cytoplasmic PTEN expression did not influence survival significantly (Mantel–Cox test) in this population-based study of patients with endometrial carcinoma. Survival curves were estimated according to the Kaplan–Meier method with death due to endometrial carcinoma as the end point.

DISCUSSION

For several types of malignancies, PTEN gene alterations have been linked to advanced disease,2 and other studies have shown that PTEN may be important for the regulation of tumor cell invasion and metastasis.3, 5 Among patients with sporadic endometrial carcinoma, loss of heterozygosity of chromosome 10q23 was found initially in about 40% of tumors,7 and, subsequently, PTEN mutations were identified in 30–50% of these tumors.8–10 In the studies of patients with endometrial carcinoma, PTEN mutations were associated with the endometrioid subtype20–22 and favorable prognosis22 rather than with increased metastatic potential, which has been noted in some other types of malignancies.1, 2, 23 One recent study on primary prostate carcinoma found that loss of PTEN expression was correlated significantly with high Gleason score and advanced disease stage.18 This appears to be in line with our current finding.

To our knowledge, this is the first report on PTEN protein expression in a large and population-based series of patients with endometrial carcinoma. Our finding that 20% of the tumors showed complete loss of PTEN expression supports the suggestion that PTEN may play a significant role as a tumor suppressor gene in this disease. This also is in line with recent reports showing that inactivation of PTEN in endometrial carcinoma, through either deletions, promoter methylation, or point mutations, appears to be the most frequent genetic alteration in this tumor type reported to date.7–10, 24 The relative importance of various genetic alterations, however, is not known.

One recent report on PTEN expression in patients with endometrial hyperplasia and carcinoma25 identified loss of expression in 20 of 33 endometrioid carcinomas studied (61%), which is greater than what was found in the current study; however, the mutation rate of 85% observed in their series also was very high compared with previous studies, which reported mutation rates of 30–50%.8–10 However, the selection of patients with endometrioid endometrial adenocarcinoma and premalignant lesions in the study by Mutter et al.25 may explain some of the differences in the findings. This approach is in contrast to the approach used for the current population-based study, which focused only on the tumors and did not include regions of atypical endometrial hyperplasia close to the tumors when the staining was recorded. In the same study, the authors found loss of PTEN expression in 2 of 8 nonendometrioid tumors (25%),25 but the correlation between PTEN expression and histologic subtype did not reach statistical significance. In contrast, we observed a tendency, although it was not statistically significant, toward a greater proportion of nonendometrioid tumors showing loss of PTEN expression for only one of the antibodies studied; thus, the correlation between PTEN expression and histologic subtype is not clear.

In a recent report on 33 patients with primary ductal adenocarcinoma of the breast, a correlation between the loss of PTEN expression and estrogen/progesterone receptor negativity was indicated; however, those authors await further studies to elucidate the significance of this observation.16 We found no associations between hormone receptors and PTEN staining in the current series.

In conclusion, we found that the loss of PTEN protein staining was relatively frequent for both antibodies studied among the patients in our large, population-based series on endometrial carcinoma. Although no prognostic influence was seen, the significant association between the loss of PTEN expression and metastatic disease indicates that the PTEN system may play an important role in a subgroup of endometrial carcinomas. However, further studies of PTEN protein expression related to histologic subgroups and various mechanisms of gene inactivation, such as mutations, deletions, and promoter methylation, will be necessary to elucidate further the functional role of PTEN in endometrial carcinogenesis.

Acknowledgements

The authors are grateful to Dr. Charis Eng (Human Cancer Genetics Program, the Ohio State University Comprehensive Cancer Center, Columbus, OH), who kindly sent cell lysates from PTEN positive and PTEN negative cell lines. The authors thank Mrs. Gerd Lillian Hallseth and Mr. Bendik Nordanger for excellent technical assistance and their colleagues for returning details of patients under their care. They also thank the Cancer Registry of Norway for information.