Loss of E-cadherin expression is associated with aberrant 5′ CpG island methylation in various tumors.
Loss of E-cadherin expression is associated with aberrant 5′ CpG island methylation in various tumors.
The authors analyzed the methylation status and immunohistochemical expression of E-cadherin in 142 endometrial tissues, consisting of 21 normal endometria, 17 endometrial hyperplasias, and 104 endometrial carcinomas.
All normal endometria and endometrial hyperplasias showed positive staining of E-cadherin, and methylation of the E-cadherin gene was not detected in any samples. In endometrial carcinoma, the positive ratio of methylation was higher and was associated with tumor dedifferention and myometrial invasion. In G1 endometrial adenocarcinomas, 66.7% showed positive staining and 33.3% showed heterogeneous staining. Methylation of the E-cadherin gene was detected in 15.6%. In G2 tumors, 19.0% showed positive staining, 69.0% showed heterogeneous staining and 11.9% showed negative staining. Methylation of the E-cadherin gene was found in 50.0%. In G3 tumors, 9.1% showed positive staining, 54.5% showed heterogeneous staining and 36.3% showed negative staining. Methylation of the E-cadherin gene was found in 81.8% of the tumors. Of the samples with no-myometrial invasion, 23.1% had methylation. In those with invasion in less than half of the myometrium, 28.6% did and in those with invasion of half or more of the myometrium, 55.6% had methylation. Of samples that did not have lymph node metastasis, 33.7% had methylation, whereas of samples that had lymph node metastasis, 60.0% had methylation.
This is the first report to analyze methylation of the E-cadherin gene promoter of endometrial carcinoma and the evidence suggests that methylation of the E-cadherin gene occurs in association with the acquisition of invasive capacity. Cancer 2003;97:1002–9. © 2003 American Cancer Society.
Cadherins are a family of cell-cell adhesion molecules essential for tight connection between cells.1 E-cadherin is the major cadherin molecule expressed in epithelial cells. The cadherin-mediated cell adhesion system is known to act as an invasion suppressor system in cancer cells, since noninvasive cells can be transformed into invasive ones when treated with antibodies to block cadherin's function or with cadherin-specific antisense RNA;2, 3 transfection of human cancer cell lines with E-cadherin cDNA can reduce their invasiveness.4 In fact, immunohistochemical examination has revealed that decreased E-cadherin expression is associated with tumor dedifferentiation and progression in endometrial carcinoma5 and many other tumors.6–10
DNA methylation in the promoter regions of many genes is associated with the regulation of gene expression. It results in transcriptional silencing of the gene, either through a direct effect or via a change in the chromatin conformation that inhibits transcription.11 The transformation of normal mammary epithelial cells into carcinoma and the subsequent progression to invasion and metastasis involve the accumulation of numerous genetic hits, including the activation or amplification of dominant oncogenes and the deletion or inactivating mutation of key tumor suppressor genes.12 It has recently become evident that tumor suppressor genes may also be transcriptionally silenced in association with aberrant promoter-region CpG island methylation.13, 14 Loss of E-cadherin expression is also associated with aberrant 5′ CpG island methylation in various tumors.15–17 In endometrial carcinoma, aberrant promoter-region CpG island methylation has been reported in some genes such as estrogen receptor α18, 19 and progesterone receptor B.20 Though it is thought that hypermethylation of the E-cadherin gene is associated with immunohistochemical distribution and tumor progression, tumor grade, tumor invasion, and lymph node metastasis, this has not been examined in endometrial carcinoma. In the current study, we analyzed the promoter-region CpG island methylation of the E-cadherin gene and compared it with the histopathologic features of lesions in normal endometria, endometrial hyperplasia, and endometrial carcinomas.
Samples of endometrial tissues were obtained from 142 women who had undergone hysterectomy or curettage at the Sapporo Medical University Hospital. Biopsy samples were obtained according to institutional guidelines (university hospital), and informed consent was obtained from patients. Normal endometrial tissue (n = 21) was taken from the unaffected endometrium of normally menstruating females with myoma or adenomyosis of the uterus. Endometrial hyperplasia tissues (n = 17) were taken from the endometrial curettage and diagnosed according to the system of the World Health Organization. As a result, three cases were found to be simple hyperplasia, nine were endometrial hyperplasia complex, and five were atypical endometrial hyperplasia. Endometrial carcinoma tissues (n = 104) were taken from modified radical hysterectomy, salpingo-oophorectomy, or selective pelvic lymphadenectomy, with para-aortic lymphadenectomy. The endometrial carcinomas were also graded according to the WHO system. This resulted in 45 cases of tumor grade G1 (well differentiated adenocarcinoma), 42 cases of G2 (moderately differentiated adenocarcinoma), 11 cases of G3 (poorly differentiated adenocarcinoma), 2 cases of serous adenocarcinoma, and 4 cases carcinosarcoma. Surgical stage was classified according to the International Federation of Gynecology and Obstetrics (FIGO) system. This result in 13 cases IA, 36 cases IB, 19 cases IC, 2 cases IIA, 3 cases IIB, 7 cases IIIA, 1 case IIIB, 19 cases IIIC, and 3 cases IVB. The samples were fixed by 10% buffered formalin for immunohistochemistry, and parts of the samples were frozen and kept at −80 °C until analysis.
Tissues were fixed overnight in 10% buffered formalin, dehydrated, and embedded in paraffin. Five micrometer serial sections of each sample were used. Sections were cut, floated onto albumin coated slides, dried at 56 °C, deparaffinized in xylene, rehydrated, and washed with phosphate-buffered saline (PBS) for 15 minutes at room temperature. Specimens were treated in a microwave oven in 0.01 mol/L citrate buffer (pH 6.0) for 30 minutes at 100 °C, slowly cooled to room temperature, and then washed with PBS for 5 minutes at room temperature. After quenching endogenous peroxidase with 3% hydrogen peroxide in PBS for 10 minutes at room temperature, the sections were incubated with a blocking solution (PBS containing 5% skimmed milk) for 60 minutes at room temperature. Then the slides were incubated overnight at 4 °C with a 1:1000 dilution of anti-E-cadherin antibody (Takara, Tokyo, Japan). After several washes with PBS, they were incubated with a second antibody, a 1:200 dilution of peroxidase conjugated anti-mouse immunoglobulin (Dakopatts, Glostrup, Denmark), for two hours. The color reaction was developed by the silver intensification procedure described previously.21 For the negative control, the same dilution of nonimmunized mouse immunogloblin was used as the first antibody.
All of the normal endometrium examined in the current study showed uniform, glandular staining for E-cadherin on the cell-cell borders regardless of the time in the menstrual cycle. The pattern of staining of the endometrial tumors was categorized into three types according to a previous study performed by Sakuragi et al.,5 and staining evaluation was performed by two independent observers (T.S. and M.N.) without knowledge of clinical outcome. The positive E-cadherin expression was uniform and showed a homogeneous, fine glandular pattern (Fig. 1A) as in normal endometrium. In the heterogeneous staining pattern, only some part of the tumor showed positive staining or the cell-cell boundaries had a coarse, irregular, granular staining (Fig. 1B). In the negative staining pattern, E-cadherin expression was seen on only a few cells or it was not detected at all (Fig. 1C). The classification of the E-cadherin staining pattern was evaluated at the section close to the myometrium. Statistical analyses were performed using the Mann-Whitney test for clinocopathologic values (grade, depth of invasion, lymph node metastasis, and surgical stage), and significance was set at P < 0.05.
DNA was extracted from frozen samples and kept at −80 °C; 1 μg of the DNA was denatured using NaOH and treated with sodium bidulfite for 16 hours according to Herman et al.22 Bisulfite modified DNA was amplified with specific primer pairs for the methylated DNA (E-cad-M, forward: 5'-ttaggttagagggttatcgcgt-3' and reverse: 5'-taactaaaaattcacctaccgac-3') and the unmethylated DNA (E-cad-u, forward: 5'-taattttaggttagagggttattgt-3' and reverse, 5'-cacaaccaatcaacaacaca-3'). One hundred ng of modified DNA were applied to 25 μL of polymerase chain reaction (PCR) mixture containing 2.5U AmpliTaq DNA polymerase (Takara), 1.5 mmol/L MgCl, 1 × Taq buffer, and 0.2 mmol/L four deoxynucleotide triphosphates. Thirty-eight cycles of PCR were carried out with the program of 30 seconds at 94 °C, 1 minute at 57 °C for E-cad-m or 53 °C for E-cad-u, and 1 minute at 72 °C. Parts of the PCR products were electrophoresed on 2.5% agarose gel. There were three patterns for methylation status (Fig. 2): only unmethylated, only methylated, and a mixture of methylated and unmethylated forms. Since these samples contain not only tumor cells but also surrounding connective tissue cells that originally do not express E-cadherin, we decided to group these results as unmethylated and methylated (which included both unmethylated and mixed results). Statistical analyses were performed using the Mann-Whitney test for clinocopathologic values (grade, depth of invasion, lymph node metastasis, and surgical stage) and significance was set at P < 0.05.
In the current study, we analyzed methylation of the E-cadherin gene and its expression by IHC using 142 endometrial tissues and their DNAs. These tissues consisted of 21 normal endometria, 17 endometrial hyperplasias, 98 endometrioid adenocarcinomas, 4 sarcomas of the endometrium, and 2 other endometrial carcinomas (Table 1). In the normal endometria, all 21 samples showed positive staining of E-cadherin by IHC. Methylation of the E-cadherin gene was not detected in any of these samples. All 17 endometrial hyperplasias, 3 simple hyperplasias, 9 adenomatous hyperplasias, and 5 atypical hyperplasias showed positive expression by IHC, but methylation of the E-cadherin gene was not detected in any of these samples.
|Pathologic diagnosis||No. of samples||Methylation||IHC of E-cadherin|
In the 98 endometrioid adenocarcinomas, there were 45 G1 tumors, 42 G2 tumors and 11 G3 tumors. Of the 45 G1 tumor samples, 30 (66.7%) showed positive staining and 15 (33.3%) showed heterogeneous staining by IHC. Methylation of the E-cadherin gene was detected in 7 of the 45 samples (15.6%). Of the 42 G2 tumors, 8 (19.0%) had positive staining, 29 (69.0%) heterogeneous staining, and 5 (11.9%) negative staining (G1 vs. G2, P < 0.05). Methylation of the E-cadherin gene was found in 27 G2 tumor samples (50.0%; G1 vs. G2, P < 0.05). Of the 11 G3 tumors, 1 (9.1%) had positive staining, 6 (54.5%) had heterogeneous staining, and 4 (36.3%) showed negative staining (G1 vs. G3, P < 0.05). Methylation of the E-cadherin gene was found in nine G3 tumors (81.8%; G1 vs. G3, P < 0.05).
The correlations between methylation of the E-cadherin gene and the immunohistochemical findings for E-cadherin in endometrial adenocarcinomas are shown in Table 2. The ratio of samples that had methylation of the E-cadherin gene became higher with the decrease of E-cadherin expression by IHC (positive > heterogeneous > negative; positive vs. heterogeneous and heterogeneous vs. negative, P < 0.05). Only 2 of 39 samples (5.1%) with positive staining of E-cadherin showed methylation of the E-cadherin gene, whereas 26 of 50 samples (52.0%) with heterogeneous staining showed methylation, as did 9 of 9 (100%) with negative staining.
|No. of samples||Methylation of E-cadherin gene|
These results are shown in Table 3. The ratio of hypermethylation of the E-cadherin gene was higher in advanced stages; however, there were no significant differences among them. Immunohistologic findings for E-cadherin were similar to the results of hypermethylation, and there were no significant differences in immunohistologic findings among the surgical stages.
|Surgical stage||No. of samples||Methylation||IHC of E-cadherin|
There was also a significant correlation between the depth of myometrial invasion and hypermethylation of the E-cadherin gene (Table 4). Of the 13 samples that did not have myometrial invasion, all showed positive staining of E-cadherin by IHC and 3 of the 13 (23.1%) had methylation of the E-cadherin gene. Of the 49 samples that had invasion in less than half the myometrium, 19 (38.8%) had positive staining, whereas 28 of the 49 (57.1%) had heterogeneous staining and 2 (4.1%) were negative (negative vs. < 0.5, P < 0.05). Of these 49 samples, 14 (28.6%; negative vs. < 0.5, P < 0.05) had methylation of the E-cadherin gene. Of the 36 samples that had invasion of half or more of the myometrium, 7 (19.4%) showed positive staining, 17 (47.2%) had heterogeneous staining and 12 (33.3%) were negative (negative vs. ≥ 0.5, P < 0.05). Twenty of the 36 samples (55.6%; < 0.5 vs. ≥ 0.5, P < 0.05) had methylation of the E-cadherin gene.
|No. of samples||Methylation||IHC of E-cadherin|
|Lymph node metastasis|
|Positive lymph node||12||—||1||4||7|
In the current study, the correlation between lymph node metastasis and hypermethylation of the E-cadherin gene was analyzed in 98 endometrioid adenocarcinomas. As shown in Table 4, 83 of the 98 endometrioid adenocarcinomas did not have lymph node metastasis, and the other 15 cases did. Of the 83 samples that did not have lymph node metastasis, 37 (44.6%) showed positive staining of E-cadherin, 41 (49.4%) heterogeneous staining, and 5 (6.0%) negative staining, all by IHC. Twenty-eight samples (33.7%) had methylation of the E-cadherin gene. Of the 15 samples that had lymph node metastasis, 2 (13.3%) showed positive staining of E-cadherin, 9 (60.0%) heterogeneous staining, and 4 (26.7%) negative staining, all by IHC. Nine samples (60.0%) had methylation of the E-cadherin gene. No significant differences were found between samples that were positive and negative for methylation and IHC.
Additionally, we analyzed E-cadherin expression in lymph nodes that had metastasis by IHC. Of these 12 samples, one sample showed positive staining, 4 samples heterogeneous staining and 7 samples were negative. Of these seven samples that showed negative staining (Fig. 3A), three had shown heterogeneous staining in situ (Fig. 3B).
There were six cases without endometrioid adenocarcinoma; two were serous adenocarcinoma, and four were carcinosarcoma. The results for these samples are shown in Table 5. The two cases of serous adenocarcinoma, both of which were Stage IIIC and had the lymph node metastasis, had hypermethylation of E-cadherin, and the immunohistochemical finding was heterogeneous in both cases. In the four cases of carcinosarcoma, one case was surgical Stage IC, two Stage IIIC, and one Stage IVB. Hypermethylation was found in one sample, which was Stage IIIC, and all cases showed negative staining of E-cadherin by IHC.
|Sample||Surgical stage||Myometrial invasion||Lymph node metastasis||IHC of E-cadherin||Methylation status|
Cell-cell adhesion participates in histogenesis and plays a critical role in the establishment and maintenance of cell polarity and cell society. Reduced cell-cell adhesiveness allows cancer cells to disobey the social order, resulting in destruction of the histologic structure, the morphologic hallmark of malignant tumors.17 In cancers in vivo, particularly the diffuse type, tumor cells are dissociated throughout the entire tumor mass, lose their cell polarity, and infiltrate the stroma in a scattered manner.17 Consistent with this concept, immunohistochemical studies have revealed that decreased E-cadherin expression is associated with tumor dedifferentiation and progression in endometrial carcinoma5 and other tumors.6–10 In the current study, decreased expression of E-cadherin in endometrioid adenocarcinoma was associated with tumor dedifferentiation (G1 > G2 > G3) and myometrial invasion because samples that had invasion in less than half of the myometrium showed a higher ratio of positive staining of E-cadherin than samples that had invasion of half or more of the myometrium. Though the expression of E-cadherin showed a tendency to decrease in advanced surgical stages and when there was lymph node metastasis, there were no significant differences, probably because of the small number of samples. However, of seven samples that showed negative staining in lymph nodes, three had shown heterogeneous staining in situ. The evidence suggests that E-cadherin negative cells selectively separated from the tumor and metastasized to the lymph nodes.
A recent study showed CpG methylation around the promoter region of the E-cadherin gene and induction of E-cadherin expression following treatment with the DNA methyltransferase inhibitor 5-azacytidine in human cancer cell lines lacking E-cadherin expression.23 It was discovered that some tumor suppressor genes, including RB, VHL, p15, and p16, were inactivated as a result of reduced expression due to CpG methylation.24 As observed with these tumor suppressor genes, the E-cadherin invasion suppressor gene in human cancers is silenced by an epigenetic mechanism, DNA hypermethylation.17 In the current study, hypermethylation in the promoter region of the E-cadherin gene was correlated with tumor progression, tumor dedifferentiation, and the depth of myometrial invasion. The number of samples that had methylation increased with tumor dedifferentiation, G1 (15.6%) < G2 (50.0) < G3 (81.8%). Of the samples that had invasion in less than half the myometrium, 28.1% had methylation, whereas 55.6% of samples that had invasion of half or more of the myometrium had methylation. Methylation of the E-cadherin gene correlated quite well with the immunohistochemical findings. Recently, there have been reports showing a correlation between tumor progression and methylation of the E-cadherin gene promoter in human tumors.16, 17, 25–30 However, to our knowledge, this is first report that analyzed the methylation of the E-cadherin gene promoter in endometrial carcinoma.
In the current study, we analyzed six cases without endometrioid adenocarcinoma. Two cases were serous adenocarcinoma and four were carcinosarcoma. Serous adenocarcinoma31 and carcinosarcoma32 are known to be more aggressive, with a lower survival rate and higher rate of death from disease than the usual type of endometrial adenocarcinoma. The two cases of serous adenocarcinoma, both of which were Stage IIIC and had lymph node metastasis, had hypermethylation of E-cadherin, and immunohistochemistry was heterogeneous in both cases. The results revealed that the existence of hypermethylation and immunohistochemical findings for E-cadherin in the clinical stage, tumor invasion, and lymph node metastasis were quite similar to those of endometrioid adenocarcinomas. All four cases of carcinosarcoma, though, had negative staining of E-cadherin by IHC. Hypermethylation was found only in one sample. This may have been because the carcinosarcomas consisted mostly of sarcoma components, which originally do not express E-cadherin, and hypermethylation could not detected in them.
In the current study, methylation of the E-cadherin gene was rare in normal endometrium and endometrial hyperplasia. Endometrioid adenocarcinoma, which accounts for the majority of endometrial cancers, typifies the group of endometrial carcinomas that develop from atypical endometrial hyperplasia in the setting of excess estrogenic stimulation.33 The evidence suggested that methylation of the E-cadherin gene might not contribute to the early events of endometrial carcinogenesis. Conversely, methylation of the E-cadherin gene was frequently found in undifferentiated tumors (81.8% of G3 tumors and 50.0% of G2 tumors). Furthermore, it was also correlated with the depth of myometrial invasion. The evidence suggested that methylation of the E-cadherin gene occurred in association with the acquisition of invasive capacity.
E-cadherin is a Ca2+-dependent adhesion molecule that, in association with alpha-, beta-, and gamma-catenin, constitutes the major component of adherens junctions in vertebrates.34 Both beta- and gamma-catenin bind directly to the cytoplasmic domain of E-cadherin, whereas alpha-catenin links the bound beta- and gamma-catenin to the microfilament network of the cytoskeleton.35 Several reports have indicated that E-cadherin, an epithelial-specific cadherin, is a key molecule for the maintenance of epithelial integrity and of polarized states in association with alpha-, beta-, and gamma-catenin, and that the reduction of E-cadherin-mediated cell-cell adhesion favors the dispersion of cancer cells.35 In our previous study, moderate or strong staining of beta-catenin in the nuclei was observed in 60.0% of endometrial hyperplasia samples and 30.0% of endometrial cancer samples, though the beta-catenin gene and adenomatous polyposis coli (APC) protein have neither mutation nor deletion.36–38 This evidence implies that E-cadherin mediated cell adhesion is reduced in endometrial hyperplasias, though reduced expression of E-cadherin by hypermethylation was not found in endometrial hyperplasia.
The authors thank M. Kim Barrymore for editing the article.