Cancer Cell Biology
The E-cadherin gene is silenced by CpG methylation in human oral squamous cell carcinomas
Version of Record online: 30 MAY 2001
Copyright © 2001 Wiley-Liss, Inc.
International Journal of Cancer
Volume 93, Issue 5, pages 667–673, 1 September 2001
How to Cite
Nakayama, S., Sasaki, A., Mese, H., Alcalde, R. E., Tsuji, T. and Matsumura, T. (2001), The E-cadherin gene is silenced by CpG methylation in human oral squamous cell carcinomas. Int. J. Cancer, 93: 667–673. doi: 10.1002/ijc.1386
- Issue online: 22 JUL 2001
- Version of Record online: 30 MAY 2001
- invasion and metastasis;
- human tongue SCC
Reduction of E-cadherin strongly relates to invasiveness and metastasis in vitro. To clarify CpG methylation around the promoter region of the E-cadherin gene in oral squamous cell carcinoma (SCC), we examined the DNA samples of various human SCC cell lines and primary oral SCC tissues by methylation-specific polymerase chain reaction (MSP). CpG methylation of the E-cadherin gene markedly correlated to the reduction of E-cadherin expression in human oral SCC cell lines. In primary oral SCC tissues, only 1 of 5 preserved E-cadherin-expressing tissues was methylated, whereas methylation was found in 17 (94.4%) of 18 E-cadherin-reduced tissues. Our results suggest that reduction of E-cadherin expression is associated with CpG methylation of the E-cadherin gene promoter. We recently established two cell lines with high and low metastatic potential, UM1 and UM2, from SCC primary tongue tissue of a patient. E-cadherin expression of high-metastatic UM1 was clearly lower than that of low-metastatic UM2, and MSP results showed CpG methylation in the UM1 but not the UM2 cell line. To investigate whether demethylation of CpG methylation of the E-cadherin gene could restore expression and function of E-cadherin, we treated UM1 with the demethylating agent 5-azacytidine (5-aza) and found that E-cadherin expression was indeed restored by demethylation. Moreover, in the demethylated UM1, invasion of the collagen gel was clearly suppressed compared with the untreated UM1. These results suggested that inactivation of E-cadherin expression resulted from CpG methylation of the gene promoter; a correlation between CpG methylation of the E-cadherin gene promoter and invasive potential was also suggested. © 2001 Wiley-Liss, Inc.
The results of treatment for oral squamous cell carcinomas (SCCs) have been gradually improving due to advances in cancer therapies. However, this tumor is characterized by a high degree of local invasiveness to surrounding tissues as well as a high incidence of metastasis to cervical lymph nodes1 and consequently causes local recurrence or distant organ metastases. Therefore, the prevention and control of such invasion and metastasis are important aspects of cancer therapies for oral SCCs.
The pathogenesis of cancer metastasis consists of multiple sequential steps.2–6 To accomplish the process of metastasis to other organs, cancer cells first detach from the primary tumor and invade the extracellular matrix. In this step, it has been widely accepted that the reduction of E-cadherin frequently causes cell release from the primary tumor.7, 8 E-cadherin, a 120 kD cell surface glycoprotein involved in calcium-dependent epithelial cell specific cell adhesion, is well known to mediate cell-cell communication and cell-cell adhesion between epithelial cells.9, 10 The reduction in E-cadherin, which acts as an invasion suppressor in human cancer, strongly relates to invasiveness and metastasis in vitro.8, 11–13 In the clinicopathological evaluation of human oral SCCs, the reduction of E-cadherin expression has been also reported to correlate closely with tumor invasion and metastasis.14–16
The expression of some genes can be frequently inactivated by reversible epigenetic events rather than genetic events.17–21 Recently, methylation of a CpG island that has been identified within the 5′ proximal promoter region, as an alternative pathway of inactivating the E-cadherin gene, has been suggested.22 In human breast, prostate and hepatocellular carcinomas, hypermethylation of the CpG island around the 5′ regulatory areas of the E-cadherin gene has been correlated with reduction in E-cadherin expression.23–27 More recently, hypermethylation of the E-cadherin promoter has also been described in primary human gastric carcinomas28 and in leukemia.29 On the other hand, it has been reported that loss of E-cadherin expression has been associated with mutation of the E-cadherin gene.26, 27 However, the mechanisms responsible for the reduction or absence of E-cadherin in human oral SCCs are still controversial.
In the present study, we analyzed CpG methylation around the E-cadherin promoter region in various human oral SCC cell lines and oral SCC tissue samples obtained from therapy-free patients. Recently, we established the human oral SCC cell lines UM1 and UM2, which exhibited high and low invasive and metastatic activity, respectively, derived from a primary lesion of a patient with tongue carcinoma.30 The expression of E-cadherin in UM1 was clearly lower than that in UM2. To investigate the relationship between E-cadherin expression and invasiveness, we examined the effect of the demethylating agent 5-azacytidine (5-aza) using UM1 and UM2. Treatment of E-cadherin low-level UM1 with 5-aza resulted in a reversal of function as well as re-expression of E-cadherin.
MATERIAL AND METHODS
Cell lines and primary tissues
Human oral SCC cell lines, HSC-2, HSC-3, HSC-4, KB, KOSC-3, NO-1-N1, NO-1-U1 and SAS, were obtained from the Health Science Research Resources Bank (Osaka, Japan). IMC-2, IMC-3, Sa3 and T3M-1 clone 2 were obtained from the RIKEN GENE BANK (Tsukuba, Japan). UM1 and UM2 were previously established from a primary lesion of a single tongue carcinoma of a patient who had not received any treatment.30 UM1 exhibited a higher motility and invasive and metastatic activity than UM2 by in vivo and in vitro assays.30 A431 was used as positive control for E-cadherin expression,16 and MDA231 was used as negative control for E-cadherin expression31 and methylated control of the E-cadherin gene.18
We studied 23 samples of primary oral SCCs, which were resected in the Department of Oral and Maxillofacial Surgery II, Okayama University Dental Hospital, Okayama, Japan. The tumors were treated by biopsy and surgery without radiation and chemotherapy.
Methylation analyses of E-cadherin gene by methylation-specific PCR
Genomic DNAs were extracted from these oral SCC cell lines and tissue samples using a Sepa Gene device (Sanko Jyunyaku, Tokyo, Japan). The DNAs were modified with sodium bisulfite and analyzed according to the modification of the method described by Herman et al.32 The sets of primers were E-cadherin M-sense (5′ TTAGGTTAGAGGGTTATCGCGT-3′) and E-cadherin M-antisense (5′-TAACTAAAAATTCACCTACCGAC-3′) for the methylated sequence, and E-cadherin U-sense (5′ TAATTTTAGGTTAGAGGGTTATTGT-3′) and E-cadherin U-antisense (5′-CACAACCAATCAACAACACA-3′) for the unmethylated sequence. Briefly, 1 μm of genomic DNA was denatured by treatment with NaOH and modified by sodium bisulfite. DNA samples were then purified using Wizard DNA purification resin (Promega, Madison, WI, USA), treated with NaOH, precipitated with ethanol and resuspended in 30 μl water. Modified DNAs were amplified in a total volume of 20 μl 1 × GeneAmp PCR Gold Buffer (PE Applied Biosystems, Foster City, CA, USA) containing 1.0 mM MgCl2, 1 μM each primer, 0.2 mM dNTPs and one unit Taq polymerase (AmpliTaq Gold DNA polymerase, PE Applied Biosystems). After activation of the Taq polymerase at 95°C for 10 min, PCR was performed in a thermal cycler (TaKaRa PCR Thermal Cycler MP, TaKaRa, Tokyo, Japan) for 40 cycles consisting of denaturation at 95°C for 30 sec, annealing at 57°C (E-cadherin M-primer sets) or 53°C (E-cadherin U-primer sets) for 30 sed and extension at 72°C for 30 sec, followed by a final 10 min extension at 72°C. PCR products were then loaded onto non-denaturing 7.5% polyacrylamide gels, stained with ethidium bromide and visualized under UV illumination.
Analysis of E-cadherin protein expression
The expression of E-cadherin in oral SCC cell lines was determined by Western blot analysis using a mouse monoclonal antibody against human E-cadherin (TaKaRa) according to the method previously described.30 The amount of E-cadherin protein was measured using a TaKaRa E-cadherin enzyme immunoassay (EIA) Kit (TaKaRa), a commercially available enzyme-linked immunosorbent assay, according to the manufacturer's instruction.
Immunohistochemical staining for E-cadherin in the oral SCC tissues was performed on 5 μm paraffin sections. After blocking of endogenous peroxidase activity, the sections were incubated in an anti-human E-cadherin monoclonal antibody (Transduction Laboratories, Lexington, KY; diluted 1:200 in TBS-T) for 16 hr at 4°C, following pre-treatment with DAKO Labeled Polymer (DAKO, Tokyo, Japan) for 1 hr. E-cadherin expression was visualized with the diaminobenzidine (DAB) Chromogen System (Shandon/Lipshow, Pittsburgh, PA, USA). Histological evaluation of E-cadherin expression was done according to Shiozaki's classification.11
Analysis of E-cadherin function in UM1 and UM2 after treatment with 5-aza
UM1 exhibited loose cell-cell adhesion and exhibited high invasive and metastatic activity, whereas UM2 conversely exhibited firm cell-cell attachment and exhibited low invasive and metastatic activity.30 To examine the effects of demethylation, we treated UM1 and UM2 with 10 μM 5-aza, a demethylating agent.33 Analyses of E-cadherin expression and methylation of the 5′ CpG island after treatment with 10 μM 5-aza were performed using Western blot and methylation-specific PCR (MSP), respectively.
In addition, the invasiveness of these cell lines in vitro was evaluated with organotypic raft culture according to the method described by Matsumoto et al.34 In brief, cancer cells (5 × 105 cells suspended in 2 ml medium with or without 10 μM 5-aza) were seeded on the collagen gel including cultured fibroblasts. At the first and second week after inoculation of the cancer cells, samples of both cell lines were processed for histological analysis. The number of cells invading the gel was quantitatively counted to measure invasiveness.
Statistical differences in the amounts of E-cadherin protein produced by the three groups classified as methylated and/or unmethylated DNA were analyzed by the Kruskal-Wallis test. Statistical differences in the number of invasive cells between the 5-aza-treated and untreated UM1 cell lines were determined by Mann-Whitney's U-test for nonparametric samples.
CpG methylation and expression of E-cadherin protein in human oral SCC cell lines
The expression of E-cadherin in 14 oral SCC cell lines was examined by Western blot analysis (Fig. 1a). Three cell lines (KB, IMC-2 and IMC-3) lacked detectable E-cadherin protein, and the other 11 cell lines expressed various amounts of E-cadherin protein. MSP analysis showed that 7 of the 11 cell lines retained E-cadherin expression (HSC-4, NO-1-N1, NO-1-U1, SAS, Sa3, T3M-1 clone2 and UM2) and presented the unmethylation. Four of the other cell lines (KOSC-3, HSC-2, HSC-3 and UM1) exhibited bands for both methylation and unmethylation (Fig. 1b). In the unmethylated group, higher E-cadherin expression was detected, compared with the group exhibiting bands for both methylation and unmethylation. On the other hand, the CpG island in the three cell lines lacking E-cadherin expression (KB, IMC-2 and IMC-3) were markedly methylated. There was a statistically significant difference in amount of E-cadherin protein among the three groups classified as methylated and/or unmethylated DNA (Fig. 1c).
CpG methylation and expression of E-cadherin protein in human oral SCC tissues
We investigated the relationship between the expression of E-cadherin and CpG methylation in 23 tissue specimens of oral SCC patients (Table I). In the preserved type, E-cadherin protein was expressed in normal epithelial tissue, and only one (20%) of five cases exhibited methylation. In contrast, 17 (94.4%) of 18 cases classified as reduced were methylated. A heterogenous case with different degrees of E-cadherin expression varying from cell to cell did not exhibit methylation.
|Preserved (n = 5)||Reduced|
|Heterogenous (n = 8)||Weak and homogenous (n = 6)||Lost (n = 4)|
E-cadherin function of UM1 and UM2 after treatment with 5-aza
Although UM1 and UM2 were derived from a primary lesion of a patient with tongue carcinoma, the amount of E-cadherin produced by UM1 was markedly lower than that in UM2 (Fig. 1a). We assessed the effect of demethylation induced by 5-aza on the expression of E-cadherin and the functional E-cadherin of these cell lines by MSP analysis. Treatment with 5-aza did not affect expression of E-cadherin protein and the unmethylated condition of UM2 (Fig. 2a,b). However, UM1 had restored E-cadherin expression (Fig. 2a) and demethylated the CpG methylation of E-cadherin (Fig. 2b). Moreover, untreated UM1 exhibited a scattered growth pattern, but treated UM1 exhibited firm cell-cell adhesion and grew in a colony pattern as well as the growth pattern of UM2 (Fig. 2c).
We examined the effects of 5-aza treatment on the invasiveness of each of the cell lines using organotypic raft culture assay. Untreated UM1 actively invaded the collagen gel matrix and exhibited a sporadic invasion pattern. In contrast, 5-aza-treated UM1 exhibited some stratified cell layers over the matrix, and invasiveness into the gel matrix was suppressed compared with the untreated group (Fig. 3a). Treatment with 5-aza significantly decreased the number of cells that invaded the gel matrix (Fig. 3b). However, UM2 exhibited no remarkable change (Fig. 3a,2 and 4).
An essential factor for the success of clinical therapy of SCCs of the head and neck is to suppress to lymph node metastasis; therefore it is important to predict the metastatic ability of primary tumors. Detachment of the cancer cells from the primary tumor is the inititial step in the metastatic process.2–6 E-cadherin plays a major role in this step,7, 8 since impaired expression of E-cadherin is frequently observed in tumors having a high degree of invasion and lymph node metastasis.8, 9, 11, 13 It has been suggested that the E-cadherin system could be inactivated by multiple mechanisms, including both genetic and epigenetic events.20
The mechanism of reduction and loss of E-cadherin expression has not been fully understood. Allelic loss of the E-cadherin gene locus (16q22) has been reported in 30% to 50% of breast, prostate and hepatocellular carcinomas.23, 35, 36 Extensive analyses reveal that mutations within the E-cadherin coding sequence in gastric and breast cancers are rare,37–40 but Saito et al.,41 demonstrated that there was no E-cadherin gene mutation in human oral SCCs by SSCP analysis. Recently, several studies have reported that CpG methylation around the promoter region may be associated with a possible mechanism of inactivation of the E-cadherin gene in various human cancers.23–29, 41 It has been reported that in E-cadherin expression-negative carcinoma cell lines established from human lung, stomach, bladder, liver, breast and prostate carcinoma, the methylation state was seen as one of the epigenetic events.24, 26
In the present study, E-cadherin expression-negative oral SCC cell lines were methylated, whereas E-cadherin expression-positive cell lines exhibited unmethylation or both methylation and unmethylation. There was a significant difference in E-cadherin expression among the three groups regardless of classification: methylated and/or unmethylated DNA. In oral SCC primary tissues, only one of the five samples preserving E-cadherin expression was methylated, whereas 94.4% of the reduced cases were methylated. Saito et al.41 demonstrated that CpG methylation was found in 9 of 52 primary oral SCCs, and 89% of the 9 methylated cases exhibited reduction in E-cadherin expression and a histologically diffuse invasive type of tumor. Their and our results suggested that 5′ CpG island methylation of the E-cadherin gene promoter could cause reduction in E-cadherin expression in oral SCCs. However, our results showed a high frequency (78%) of methylation in oral SCC primary tissues compared with the results of Saito et al.41 (17%). The high frequency of methylation may be related to the fact that half of our cases were graded as T4, according to the International Histological Classification of Tumors,42 because of loss of E-cadherin expression correlates with advanced T43 and we used MSP, which has a high sensitivity to methylation detection.32 Therefore, the difference in methylation frequency seen may be caused by the different methods used.
Differences in metastatic potential have been used to investigate the functional and molecular mechanisms of metastasis.2, 5, 6 We recently established the high and low invasion and metastasis cell lines, UM1 and UM2, respectively, in order to investigate the mechanism of metastasis.30 Although these cell lines were derived from the same patient's primary tumor, UM1 (with high invasiveness and metastatic ability) exhibited low expression of E-cadherin protein and the methylation of E-cadherin gene, whereas UM2 (with low invasiveness and metastatic ability) exhibited high expression of E-cadherin and unmethylation. 5-Aza, a demethylating agent, demethylated the CpG methylation of UM1 and restored E-cadherin expression, but this agent did not affect UM2. Moreover, UM1 restored E-cadherin-mediated cell-cell adhesion, like UM2, by a demethylating agent. These findings strongly suggest that inactivation of E-cadherin expression results from CpG methylation of the gene promoter.
It is well known that reduction in E-cadherin protein stimulates the cancer cell invasion caused by the loss of cell-cell adhesion in primary tumor. We demonstrated that demethylation of UM1 suppressed invasion into the collagen gel. It seems that E-cadherin re-expression treated with 5-aza may lead to functional E-cadherin expression in UM1. However, 5-aza may affect the methylated sites of other locations like p16 (CDK2/MTS1/p16INK4a) and other genes44, 45 by nonspecific activity. Therefore, the morphological change of UM1 from the scattered type to colony type may be dependent on the demethylataion of other genes. However, Yoshiura et al.26 have suggested that treatment with 5-aza morphologically changed the colony-type cells from the scattered epithelial ones cells, through demethylation of the E-cadherin promoter and the release from gene silencing. Immunohistochemical examination revealed that the 5-aza-treated UM1 exhibited some stratified cell layers over the collagen matrix and strong E-cadherin expression, whereas the untreated UM1 exhibited very weak E-cadherin expression (data not shown). It is suggested that the E-cadherin re-expression caused by the demethylating agent inhibited invasion. Other studies have proposed that the silencing of E-cadherin promoter activity in carcinomas is due to the loss of factors binding to the regulatory region.46, 47 Our results indicated that the reduction in E-cadherin expression was associated with CpG methylation around the promoter rather than the loss of factor binding on the E-cadherin promoter, because of E-cadherin re-expression in UM1 after treatment with the demethylating agent.
In conclusion, the present study demonstrated that there is a close link between 5′ CpG island methylation and reduction in E-cadherin in human oral SCC cell lines. Thus it is possible that new therapeutic strategies can now be developed based on re-expression of gene products by a targeted gene delivery approach.
- 4Metastatic cell attachment to and invasion of vascular endothelium and its underlying basal lamina using endothelial monolayer. In: NicolsonGL, MilasL. Cancer invasion and metastasis: biological and therapeutic aspects. New York: Raven Press, 1984. p 145–67., , , .
- 42HermanekP, SobinLH, eds. TNM classification of malignant tumours, 4th ed. Berlin: Springer-Verlag, 1987.