DNA methylation and histone modification regulate silencing ... - Nature

Oncogene (2007) 26, 3989–3997

& 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

DNA methylation and histone modification regulate silencing of epithelial cell adhesion molecule for tumor invasion and progression K-Y Tai1,6, S-G Shiah1,6, Y-S Shieh2,3,6, Y-R Kao1, C-Y Chi1, E Huang1, H-S Lee4, L-C Chang3, P-C Yang5 and C-W Wu1 1 Institute of Cancer Research, National Health Research Institutes, Miaoli, Taiwan, ROC; 2School of Dentistry and Cancer Epigenetic Laboratory, National Defense Medical Center, Taipei, Taiwan; 3Department of Oral Diagnosis, Tri-service General Hospital, Taipei, Taiwan; 4Department of Pathology, Tri-service General Hospital, Taipei, Taiwan and 5Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan

Epithelial cell adhesion molecule (Ep-CAM) is believed to have a critical role in carcinogenesis and cell proliferation. However, the association of Ep-CAM with cancer invasion and progression is less clear. We found that EpCAM was highly expressed on low-invasive cells compared with highly invasive cells. Forced expression of Ep-CAM decreased cancer invasiveness, and silencing Ep-CAM expression elevated cancer invasiveness. EpCAM expression was associated with promoter methylation. Treatment with a demethylating agent, and/or the histone deacetylase inhibitor reactivated Ep-CAM expression in Ep-CAM-negative cells and inhibited cancer invasiveness. Using a promoter–reporter construct, we demonstrated methylation of the promoter fragment drive Ep-CAM-silenced transcription. Additionally, silenced Ep-CAM gene in cancer cells was enriched for hypermethylated histone 3 lysine 9. When unmethylated and active, this promoter was associated with acetylated histone 3 lysine 9. Furthermore, we observed an increased association of Ep-CAM promoter with repression components as tumor invasiveness increased. In cancer tissues, Ep-CAM expression significantly correlated with tumor progression and associated with promoter methylation. Our data support the idea that modulation of Ep-CAM plays a pivotal role in tumor invasion and progression. Moreover, aberrant DNA methylation of Ep-CAM is implicated in enhancing invasive/metastatic proclivity of tumors. Oncogene (2007) 26, 3989–3997; doi:10.1038/sj.onc.1210176; published online 8 January 2007 Keywords: Ep-CAM; invasion; promoter methylation

Correspondence: Dr C-W Wu, Institute of Cancer Research, National Health Research Institutes or Dr Y-S Shieh, National Defense Medical Center, 116, Sec 6, Michuan E Road, Taiwan, ROC. E-mail: [email protected] or [email protected] 6 These authors contributed equally to this work. Received 19 June 2006; revised 5 October 2006; accepted 23 October 2006; published online 8 January 2007

Introduction The epithelial cell adhesion molecule (Ep-CAM) is a 40 kDa epithelial transmembrane glycoprotein encoded by the GA733-2 gene (Armstrong and Eck, 2003; Winter et al., 2003). Although the precise biological function of Ep-CAM is not completely defined, it has been suggested that Ep-CAM is involved in the cell cycle, proliferation and differentiation of epithelial cells (Munz et al., 2004). In cancer, Ep-CAM is overexpressed by most human epithelial malignancies (Drapkin et al., 2004; Went et al., 2004, 2005). For these reasons, Ep-CAM has attracted attention as a tumor maker and as a target for cancer therapy (Armstrong and Eck, 2003; Winter et al., 2003). Despite overexpression of Ep-CAM on cancer cells, its role in tumor progression is still a matter of debate. For example, recent studies found reduced Ep-CAM expression in circulating and metastatic tumors compared with their corresponding primary tumors (Takes et al., 2001; Rao et al., 2005). Furthermore, Jojovic et al. (1998) showed that Ep-CAM expression could be transiently lost during the process of metastasis. These results suggest that dynamic change of Ep-CAM expression could involve in cancer progression. However, the mechanism affecting Ep-CAM expression is not clear. Aberrant DNA methylation is the main epigenetic mechanism for regulation of gene expression in human cancers (Robertson, 2001; Jones and Baylin, 2002). A recent study of the methylation status of the Ep-CAM promoter showed few methylated CpG sites in EpCAM-positive cell lines compared with Ep-CAMnegative cell lines (Spizzo et al., 2006). Demethylation of the promoter in Ep-CAM-negative cell lines by treatment with 5-aza-20 -deoxycytidine partially restored Ep-CAM expression, which implicated promoter methylation in the regulation of Ep-CAM expression. However, there has been no investigation of the potential role of promoter methylation in dynamically changing Ep-CAM in tumor progression. DNA methylation, histone modification, chromatin remodeling and gene regulation are interconnected in mammals (Fuks, 2005; Klose and Bird, 2006). The mechanisms underlying these observations are still

Regulation of Ep-CAM in tumor invasion K-Y Tai et al

3990

obscure, but recent findings on how chromatin structure regulates gene expression are paving the way towards their elucidation. The evidence that several methylCpG-binding proteins (e.g., MeCP2 and MBD2) interact with histone deacetylase (HDAC) and recruit the corepressor protein supports a mechanism linking DNA methylation and histone modifications. However, the functional relevance of histone modifications and DNA methylation in the regulation of the Ep-CAM gene expression is unknown. To clarify the role of Ep-CAM in cancer invasion, we examined Ep-CAM expression and cell invasiveness in a variety of tumor cells. By engineering invasive cells to express Ep-CAM or silencing Ep-CAM with short hairpin RNA (shRNA), effect of Ep-CAM on cancer invasiveness was investigated. Our results suggest that high expression of Ep-CAM can decrease in vitro cell invasiveness. Treating invasive cells with a cytosine methylation and histone-acetylated inhibitor, we observed epigenetic silencing of the Ep-CAM gene expression in progressively invasive cancer cells, showing that a general trend of increasing invasiveness is associated with promoter hypermethylation. Therefore, the potential utility of the epigenome of Ep-CAM as an invasiveness and progression marker could be evaluated.

Results

No. Invaded Cells

a

Effect of Ep-CAM expression on invasion If Ep-CAM is involved in cell invasion, changes in Ep-CAM expression should affect cancer invasive capacity. A full-length Ep-CAM-expressing vector was constructed and transfected to the highly invasive, Ep-CAM-negative cells NTUB1. The protein level in the Ep-CAM transfectant, NTUB1/Ep-CAM was examined. Increased Ep-CAM protein expression was observed in NTUB1/Ep-CAM cells (Figure 2a). When cell invasiveness was analysed, invasion of NTUB1/ Ep-CAM cells was inhibited compared with vectoralone and mock cells (Figure 2b). As Ep-CAM was proposed to be involved in cell proliferation, the growth rate of these cell lines was measured. There was no significant cell growth difference between NTUB1 and NTUB1/Ep-CAM cells during our experiment. Next, we performed a gene-silencing experiment using shRNA in the lower invasive, Ep-CAM-positive cells, CL1-0. Western blotting revealed a significant knockdown in Ep-CAM expression in the shRNA-transfected cells, CL1-0/shEp-CAM (Figure 2c) and concomitantly elevated cell invasion (Figure 2d). The results demonstrated that expression of Ep-CAM is associated with tumor invasiveness.

5000

c

4500 4000 3500

Invasion Percentage (%)

Expression of Ep-CAM is inversely correlated with cancer invasiveness To examine Ep-CAM gene expression and its correlation with cancer invasiveness, we examined a set of lung adenocarcinoma cell lines, CL1-0, CL1-3 and CL1-5, which were established in our lab with progressively

increasing invasiveness (Chu et al., 1997) (Figure 1a). Ep-CAM expression progressively decreased as invasive capacity increased (Figure 1b). Further studies were performed on human cancer cell lines that originated from lung, colon, ovary and bladder. Ep-CAM expression was upregulated in lowly invasive cancer cells and downregulated in highly invasive cancer cells (Figure 1c–e). This finding indicates that Ep-CAM may play an important role in cancer invasion.

3000 2500 2000 1500 1000 500 0 CLI-3

140 120 100 80 60 40 20 0

6 Dr 520 T137 Wi I-H H NC

CLI-5

I-5

d

L C

C

L

I-0 L C

b

I-3

CLI-0

160

I-0 CL

I-5 CL

2 ES

1 UB NT

1 PA

520 376 1 1 I-0 LI-5 S2 TUB A1 Dr CI-H CL HT N C E P N Wi Ep-CAM

Ep-CAM

GAPDH α-Tublin

e

520 76 Dr CI-H T13 LI-0 LI-5 S2 E C H Wi N C

1 UB A1 P NT Ep-CAM

α-tubillin

Figure 1 Expression of Ep-CAM and correlation with invasiveness in cancer cell lines. Invasion ability (a) and Ep-CAM expression (b) in cancer cells CL1-0, CL1-3 and CL1-5 derived from the same parent cell. Expression of Ep-CAM in a series of highly and lowly invasive cancer cells (c) was determined by conventional RT–PCR (d) and Western blotting (e). Oncogene


3991

l

tro

a m

o

ck

Ve

cto

lo r-a

ne NT

UB

E 1/

p-

CA

M

ne

g

e sil

n-

p-

hE

/s

0 1-

CL

No

Ve

M CA

in

nc

lo

r-a

o ct

Ep-CAM

Ep-CAM

α-Tubulin

α-Tubulin

b 20000 18000

d 10000

16000 14000 12000 10000 8000 6000 4000 2000 0

No. Invaded Cells

No. Invaded Cells

c

n co

k

m

oc

to

c Ve

r-a

lon

e

CA

NT

1 UB

/E

9000 8000 7000 6000 5000 4000 3000 2000 1000 0

M

p-

r

e

on

-al

tor

c Ve

ol

tr on

gc

cin

len

i n-s

cto

ve

AM p-C hE s / 1-0 CL

No

Figure 2 Over- and knockdown expression of Ep-CAM relevant to cancer invasiveness. The highly invasive cell line NTUB1 transfected with Ep-CAM showed increased Ep-CAM protein expression in the NTUB1/Ep-CAM transfectant compared with control mock and vector alone cells (a) and resulted in inhibiting cell invasiveness (b). The minimally invasive cell line CL1-0 was transfected with shEp-CAM. Cells were collected and lysed after 7 days and the expression of Ep-CAM was analysed by Western blotting. shRNA/Ep-CAM downregulation of Ep-CAM expression in CL1-0/shEp-CAM (c) accompanied the reduction of cell invasion ability (d).

Correlation of methylation status and expression of Ep-CAM To investigate whether promoter methylation is involved in regulating Ep-CAM expression, we used methylation-specific PCR (MSP) to examine the methylation status of the Ep-CAM gene in highly and lowly invasive cells, in which Ep-CAM expression is downand upregulated, respectively. In all the lowly invasive cells, the Ep-CAM gene was unmethylated. In contrast, methylated alleles were detected in highly invasive cells (Figure 3a). Densely methylated DNA associated with transcriptionally repressive chromatin is characterized by the presence of underacetylated histones (Fuks, 2005; Klose and Bird, 2006). Thus, the effect of DNA demethylation and histone acetylation on Ep-CAM gene expression was investigated in Ep-CAM-downregulated cells by treating with 5-aza (DNA methylation inhibitor), trichostatin A (TSA, an inhibitor of histone deacetylase) or both. Treatment with 5-aza (1 mM for upto 3 days) elevated expression of the Ep-CAM gene in highly invasive cells, but only minimal changes were observed with TSA alone (Figure 3b). Synergistic induction of EpCAM gene expression in these cells was achieved by the combined addition of 300 nM TSA and 1 mM 5-aza for 24 h (Figure 3c). Concomitantly, treatment of these cells with 5-aza resulted in significantly decreased invasiveness of highly invasive cells in which Ep-CAM expression had been repressed by promoter methylation (Figure 3d). These findings indicated that DNA methylation and histone modifications are responsible for silencing or reactivation of Ep-CAM gene expression and result in changes in cancer invasiveness. The relevance of Ep-CAM gene expression to the activation of its promoter activities in these cell types

was investigated by reporter gene analyses. The EpCAM promoter fragment was cloned into a luciferase reporter construct (Ep-Pro) and transfected into CL1-0 cells. When the cells were transfected with an unmethylated fragment of the Ep-CAM promoter, high levels of luciferase activity were observed (Figure 3e, bar 2). This fragment also contains a consensus binding site for Sp-1, which has been reported to regulate Ep-CAM transcriptional activity (McLaughlin et al., 2004). Further elevated promoter activity was observed in the presence of Sp-1 transcription factor (Figure 3e, bar 3). Additionally, to mimic the methylation status of the Ep-CAM gene promoter, the Ep-CAM promoter–reporter construct was treated with SssI CpG-methylase or DNA methyltransferase (DNMT)1 and transfected into CL1-0 cells. In the presence of CpG methylase, the promoter activity was inhibited. Luciferase activity was abolished in cells transfected with the methylated construct, even in the presence of Sp-1 (Figure 3e, bar 4–6). These results indicated that Ep-CAM transcription is regulated by methylation of promoter sequences. Histone 3 modification on the silenced or derepressed Ep-CAM gene promoter Covalent histone tail modifications of lysine 9 at histone 3 (H3K9), including acetylation or methylation, regulate these different states of chromatin configuration and gene transcription. The possibility that site-specific histone methylation works coordinately with histone acetylation during silencing or derepression of the Ep-CAM gene in the context of its promoter methylation state was investigated by chromatin immunoprecipitation assays (ChIP) with primers that encompassed the Ep-CAM promoter region. In low-invasive cells Oncogene


3992

M

M

ES-2

No drug

TSA 5-aza dC

No drug

HT1376

CL-1-0

WiDr

CLI-5

U

U

M

U

NTUB1

M

U

No. Invaded Cells

U

10000

PA-1

PA-1

8000 6000 4000

0

Dr 376 520 Wi I-H HT1 NC

EpCAM

e

GAPDH

c

add DMSO add 5-AZA

2000

5-aza dC

U M

5-aza dC

M

NTUB1

No drug

U

ES2

TSA

b

M

5-aza dC

U

NCIH 520

M

NCIH520 CL1-0 HT1376 CLI-5

TSA

WiDr

d 12000

Highly Invasive

TSA

Poorly Invasive

No drug

a

PGL3 EP-Pro SP1 Sssl DNMT1

+ _ _ _ _

_ + _ _ _

2 ES

I-0 LI-5 C CL

_ + + _ _

_ + _ + _

1 UB NT

_ + _ _ +

1 PA

_ + + + _

90000

(TSA + 5-aza dC)

Ep-CAM

GAPDH

luciferase activity

5 ES -2 NT UB 1 PA -1

-1-

CL

-1-

0

76 13

HT

CL

IH NC

Wi

Dr

52

0

80000 70000 60000 50000 40000 30000 20000 10000 0

Figure 3 Relation of Ep-CAM methylation with gene expression and cancer invasiveness. (a) Promoter methylation of Ep-CAM was detected by MSP. U, primer specific for unmethylated DNA. M, primer specific for methylated DNA. 5-Aza and TSA treatment caused reactivation of Ep-CAM gene (b and c) and reduced cancer invasiveness (d). Synergistic derepression of Ep-CAM gene expression by 5-aza and TSA in highly invasive cancer cells (c). Ep-CAM gene activity was repressed as in vitro SssI-methylated or Dnmt1-methylated Ep-CAM promoter construct was transfected into CL1-0 cells (e) . Bar graph summarizing the luciferase activities measured in lysates from CL1-0 cells transfected with methylated or unmethylated Ep-Pro. In total, 1 mg of the plasmid Ep-Pro that had or had not been treated in vitro with CpG methylase was transfected into CL1-0 cells by LipofectAMINE in the presence or absence of the Sp1 expression vector (3 mg). Luciferase activities were measured 48 h post transfection. All experiments were repeated at least three times. Bars (from left to right): (1) PGL3 alone; (2) unmethylated Ep-Pro alone; (3) unmethylated Ep-pro plus Sp1; (4) SssI methylated Ep-pro alone; (5) Dnmt1 methylated Ep-pro alone; (6) SssI methylated Ep-pro plus Sp1.

expressing Ep-CAM, association of the Ep-CAM gene promoter with acetylated H3K9 was significantly increased compared with highly invasive cells (Figure 4a). Methylated H3K9 was present in the silenced Ep-CAM gene promoter of highly invasive cells, but it was largely decreased in the activated Ep-CAM gene promoter in low-invasive cells (Figure 4b). These results suggest that a zone of lessacetylated H3K9 plus methyl-H3K9 surrounds the hypermethylated, silenced Ep-CAM promoter. This same promoter, when unmethylated and active, is embedded in acetylated H3K9 plus dimethyl-H3K9. In this epigenetic process, a DNMT bound to an adaptor molecule such as heterochromatin protein 1 (HP1) would add a methyl group to DNA only on chromatin that is methylated at H3K9. The catalysis of DNA methylation by DNMT would permit binding of methyl-CpG-binding domain (MBD) proteins to DNA. The bound MBD proteins would attract HDAC complexes and H3K9 methyltransferase such as Suv39h1 to prepare H3K9 for methylation. We used CL1-0, CL1-3 and CL1-5 cells (in which the expression of Ep-CAM gene is progressively silenced) to examine the regulatory complex in the Ep-CAM promoter. We repeated the ChIP analyses, using antibodies specific for these molecules. As invasive ability progressively increased from CL1-0 cell, CL1-3 to CL1-5 cells, gradual increases in association of the Ep-CAM gene promoter Oncogene

with HP1, Suv39h1, HDAC1 and DNMT1 and 3b were observed in CL1-3 and CL1-5 cells (Figure 4c and d). Correlation of Ep-CAM expression and methylation status in cancer tissue To further examine whether Ep-CAM expression relates to disease progression of cancer, we used immunostaining to compare Ep-CAM expression in 82 patients with lung adenocarcinoma. Immunoreactivity of Ep-CAM was detected in bronchial epithelial cells (which acted as the positive internal control for the analysis) (Winter et al., 2003), as well as in lung cancer cells (Figure 5a–c). In cancer tissue, the staining pattern was heterogeneous and positive Ep-CAM protein expression was observed in 51 (62%) of 82 cases. Negative or reduced expression of Ep-CAM was observed in nodal metastases compared with the primary tumors (Figure 5d). Correlations between the incidence of positive immunoreactivity for Ep-CAM and the clinicopathologic features of the patients are shown in Table 1. Ep-CAM protein expression was significantly associated with clinical stage (P ¼ 0.01) and lymph node involvement (P ¼ 0.009). To determine whether the methylation in the Ep-CAM promoter is associated with Ep-CAM gene expression, microdissection of foci with known EpCAM immunostaining status was used to correlate


3993 H3K9 Acetylation

a

Widr NCIH520 HT1376 CL1-0 CL1-5 ES-2 NTUB-1 PA-1

Fold Acetylation

15

10

CL1-0 CL1-3 CL1-5

c

20

α-DNMT 1 α-DNMT 3b α-HDAC 1 α-HP 1 α-Suv39H1

5

Input Mouse IgG

0 P2 (-682~-540)

P3 (-356~-140)

b

d 16000 H3K9 methylation

14000

Widr NCIH520 HT1376 CL1-0 CL1-5 ES-2 NTUB-1 PA-1

1.5

1

12000 Relative Pixes

Fold methylation

2

CL1-0 CL1-3 CL1-5

10000 8000 6000 4000

0.5

2000 0

0 P2 (-682~-540)

Dnmt 1

P3 (-356~-140)

Dnmt 3b

HDAC1

HP1

Suv39h1

Figure 4 H3K9 modification and regulatory factors association with Ep-CAM promoter in highly and lowly invasive cells. (a) Quantitative analyses of occupancy of modified H3K9 acetylation and (b) methylation to Ep-CAM promoter in highly and lowly invasive cells. (c) ChIP analyses of the association of various regulatory factors with the Ep-CAM promoter region in CL1-0, CL1-3, and CL1-5 cells with gradual elevated invasiveness and (d) quantification of the intensities of bands of each binding protein.

a

c

b

d

e Case 1 U M

Primary tumor Case 2 U M

Case 3 U M

Lymph node metastases Case 4 U M

Case 3 U M

Case 4 U M

Figure 5 Ep-CAM expression in normal tissue and cancer lesions. Immunostaining pattern of Ep-CAM in normal bronchial epithelium is homogeneous (a) and (b and c) heterogeneous in tumor tissue. Representative cases with positive (b) and reduced expression (c) in cancers and negative staining in metastatic lymph node lesion (d). (original magnification 200 in a–c). Representative MSP of Ep-CAM in cancer tissues. Samples of cases 3 and 4 were same patients in primary tumor and lymph node metastases (e).

changes in methylation status of Ep-CAM with its expression in cancer lesions. DNA extraction, bisulfite modification and methylation status were detected in 51

specimens. Association of Ep-CAM methylation status with clinicopathologic features was examined and no significant correlation among them was noted. In 33 (33/ Oncogene


3994 Table 1 Correlation of Ep-CAM expression with clinicopathologic features of patients with lung adenocarcinoma Clinicopathologic features

Ep-CAM expression

P-value

Table 2

Methylation status of Ep-CAM promoter in case of positive and reduced Ep-CAM expression

IHC of Ep-CAM

Reduced (n ¼ 31) Positive (n ¼ 51) Gender Male (n ¼ 45) Female (n ¼ 37) Size p3.0 cm (n ¼ 27) >3.0 cm (n ¼ 55) LN involvement No (n ¼ 44) Yes (n ¼ 38) Differentiation Well (n ¼ 20) Moderate (n ¼ 40) Poor (n ¼ 22) Metastasis No (n ¼ 62) Yes (n ¼ 20) Staging Local (n ¼ 27) Invasive (n ¼ 55)

17 (20%) 14 (17%)

25 (30%) 26 (32%)

0.389

8 (10%) 23 (28%)

19 (23%) 32 (39%)

0.205

11 (13%) 20 (24%)

33 (40%) 18 (22%)

0.009

8 (10%) 14 (17%) 9 (11%)

12 (15%) 26 (32%) 13 (16%)

0.876

20 (24%) 11 (13%)

42 (51%) 9 (11%)

0.061

5 (6%) 26 (32%)

22 (27%) 29 (35%)

0.010

EP-CAM, epithelial cell adhesion molecule. n denotes the number of cases. Expression of Ep-CAM are classified as ‘Positive’ if the score is X4 and ‘Reduced’ if the score is p3 according to the method described previously (Gastl et al., 2000). In the staging, ‘Local’ includes tumors in stages 0, IA and IB and ‘Invasive’ includes tumors in stages IIA, IIB, IIIA, IIIB and IV. Bold values reflect statistic significance with P-value o0.05, measured by w2 test.

51) samples with Ep-CAM-positive expression, unmethylated promoter was detected in 27 (53%) cases and methylated promoters were detected in six (12%) cases. In 18 (18/51) samples with reduced Ep-CAM expression, only six (12%) samples with unmethylated promoter and methylated promoters were detected in 12 (23%) samples. There was significant association of EpCAM expression and its promoter methylation status (Table 2). Notably, when examining methylation status in seven lymph node lesions with reduced Ep-CAM expression, a high proportion of methylated promoters (7/11) were detected.

Discussion As the role of Ep-CAM in cancer progression is not clearly elucidated and the mechanisms regulating Ep-CAM expression is not clear, we examined the functional relevance of Ep-CAM expression and tumor invasiveness, and investigated DNA methylation and histone modification in the regulation of Ep-CAM gene expression. A direct role for Ep-CAM in cancer invasiveness was demonstrated in our in vitro experiments. In supporting this notion, several lines of evidence could explain our results. Like other adhesion molecules, Ep-CAM provides invasion-suppressor properties to epithelia through cell–cell aggregation – normally nonadhesive cell lines can be induced to aggregate through transfection of Ep-CAM and have Oncogene

Positive (n ¼ 33) Reduced (n ¼ 18) Total (n ¼ 51)

P-value

MSP U

M

27 (53%) 6 (12%) 33

6 (12%) 12 (23%) 18

0.001

EP-CAM, epithelial cell adhesion molecule; U, unmethylation; M, methylation. Bold values reflect statistic significance with P-value o0.05, measured by w2 test.

reduced mobility and invasive behaviors (Litvinov et al., 1994). In a model system, Ep-CAM-mediated adhesion can suppress invasion of tumor cells grafted in mice (Basak et al., 1998). Negative or limited Ep-CAM expression in primary laryngeal carcinoma has been linked to the presence of nodal metastases and in colorectal carcinoma to a poor prognosis (Takes et al., 1997; Basak et al., 1998). Therefore, it seems possible that Ep-CAM-negative cells have greatly reduced cell– cell adhesion, which promotes invasion or metastasis. Our data suggest the possibility that Ep-CAM-targeted immunotherapy may induce tumor invasion and metastasis. This concern is supported by the findings that, in some tumors, reduced Ep-CAM expression in circulating and metastatic tumors is relatively lower than in their corresponding primary tumors (Takes et al., 2001; Rao et al., 2005), which implies dynamic change and plasticity of Ep-CAM expression during cancer progression (Jojovic et al., 1998). This must be of concern in any attempt to use Ep-CAM-targeted immunotherapy. Findings from cancer tissue experiments corroborated our observations in vitro that Ep-CAM is overexpressed in lung adenocarcinoma and its expression correlated with cancer invasiveness and tumor progression. In line with our findings, recent studies showed that advanced stage of ovarian cancer had significant lower Ep-CAM expression than stage I diseases (Kim et al., 2003), and loss of Ep-CAM expression in gastric cancer could imply aggressive disease (Songun et al., 2005). However, these findings differ from some observations of increased Ep-CAM expression associated with cancer malignancy (Kumble et al., 1996; Osta et al., 2004). Controversial results were also reported in prognosis implication by Ep-CAM expression. Ep-CAM overexpression correlated significantly with poor survival in breast cancer (Gastl et al., 2000). In contrast, positive expression of Ep-CAM associated with better prognosis of gastric cancer (Songun et al., 2005) had been demonstrated. Additionally, Ep-CAM expression may be varied in subtype of tumor cell morphology. For example, in renal tumors clear cell type showed little Ep-CAM expression, whereas chromophobe type demonstrated strong positive expression (Seligson et al., 2004). Although Ep-CAM expression is not associated with tumor progression in overall lung cancer, correlation between Ep-CAM expression and metastasis and tumor stage in specific subtype of lung cancer such as adenocarcinoma was not analysed (Went et al., 2006).


3995

Furthermore, significant difference of Ep-CAM expression in various subtypes of lung cancer was observed (Went et al., 2006). Thus, it is possible that the function of Ep-CAM may depend on the biological context of tumor microenvironment and its role in cancer could be tissue- and/or cell- specific. In this study, we found that progressive loss of EpCAM expression is associated with tumor invasiveness, and that Ep-CAM is heterogeneous in cancer tissues. Our results provide evidence that promoter methylation could repress expression of Ep-CAM during tumor progression. First, Ep-CAM is methylated and its expression downregulated in highly invasive cells. In contrast, unmethylated Ep-CAM was observed in lowly invasive cells (which expressed high levels of Ep-CAM). Second, treatment with 5-aza-dC and TSA reactivated Ep-CAM expression and inhibited cancer invasiveness, suggesting that Ep-CAM plays a role in this process and that DNA methylation and histone modification are responsible for silencing Ep-CAM gene expression. These results are further supported by Spizzo et al. (2006) finding that promoter methylation of Ep-CAM is, in part, responsible for Ep-CAM expression in cancer cells. A recent study showed that Ep-CAM overexpression was present in 0% of normal epithelia, 93.9% of intestinal metaplasia, 42.9% of gastric epithelial dysplasia and 34.3% of adenocarcinoma (Joo et al., 2005), suggesting it dynamically changes during malignant transformation and progression. Furthermore, Rao et al. (2005) study detected Ep-CAM expression in primary, circulating and metastatic tumors and showed that Ep-CAM level is low in circulating tumor cells compared with primary and metastatic sites. Collectively, these findings indicate that dynamic change of Ep-CAM is a regulatory event in the process of invasion or metastasis and DNA methylation is one of the major mechanisms in regulating Ep-CAM expression during this process. Recent reports suggest a histone core hypothesis in which combinations of distinct modifications at particular sites on the histone tail compose a ‘histone code’ that affects which proteins are capable of interacting with histone–DNA complexes and, consequently, how a single gene activity can be regulated (Fuks, 2005; Klose and Bird, 2006). In agreement with this notion, the precise association of H3K9 to the Ep-CAM gene repressive or derepressive state as presented in present study indicates that the modification status of H3 K9 (acetylation and methylation) at the Ep-CAM gene promoter region serves as a histone code for Ep-CAM gene expression. Thus, our data provide evidence that aberrant cytosine methylation, histone hypoacetylation and H3K9 methylation act together in a discrete region of the Ep-CAM promoter to inactivate Ep-CAM transcription during tumor invasion and progression. In conclusion, based on our findings and the functions proposed for EP-CAM, we postulate that this adhesive molecule plays distinct and different roles at various stages in the evolution and progression of cancers. Our results show that inactivation of Ep-CAM by hypermethylation is emerging as an important mechanism for

promoting invasion. Moreover, we demonstrated that expression of Ep-CAM in tumor progression is regulated through epigenetic methylation. These results may yield new strategies for the diagnosis, prevention and treatment of cancer.

Materials and methods Cell culture Human lung carcinoma cell lines with different invasive capabilities (CL1-0, CL1-3 and CL1-5) have been described previously (Chu et al., 1997). HT-1376 human bladder carcinoma, human colon carcinoma, ES2, PA1 ovary carcinoma and NCI-H520 human lung carcinoma cells were obtained from the American Type Culture Collection (Manassas, VA, USA). NTUB1 bladder carcinoma cells were from Dr Pan-Chyr Yang. All of the cells were cultured at 371C in 5% CO2 and RPMI-1640, supplemented with 10% bovine serum, 2.98 g/l HEPES, 2 g/l NaHCO3, 100 U/ml penicillin and 100 mg/ml streptomycin. In vitro invasion assay The invasion ability was examined using 24-well insert-based assays (BD Biosciences, Franklin Lakes, NJ, USA) and performed as described previously (Shieh et al., 2005). RNA extraction, reverse transcription polymerase chain reaction (RT–PCR) Total cellular RNA was extracted, purified and converted to cDNA. The PCR primers used were: Ep-CAM sense 50 -TTGCTCAAAGCTGGCTGCCA-30 and antisense 50 -GGA TCCAGTTGATAACGCGT-30 (see Supplementary methods for details). Protein extraction and Western blot analysis Immunohistochemistry was carried out as described (Shieh et al., 2005). Ep-CAM cell transfectants Human Ep-CAM cDNA was amplified by RT–PCR with primers Ep-CAM-HindIII-f, 50 -TATAAAGCTTATGG CGCCCCCGCAGGTCCTCGCGT-30 and Ep-CAM-PstI-r, 50 -TATACTGCAGTTAGTCATTGAGTTCCCTATGCATC-30 . The PCR product was digested with HindIII and PstI restriction enzymes and subcloned into pEGFP-N1 (Clontech, Mountain View, CA, USA) under the control of the cytomegalovirus promoter. NTUB1 cells transfected with pEGFP or pEGFP–Ep-CAM were selected in G418 (Calbiochem, San Diego, CA, USA) to generate NTUB1/pEGFP and NTUB1/Ep-CAM cells, respectively. ShRNA Two clones (V2HS_134160 and V2HS_134161) of shRNA targeting Ep-CAM (NM_002354) was obtained from GENDISCOVERY (Open Biosystem, Drive Huntsville, Australia). Transfection of two clones’ shRNA for the targeting of any endogenous Ep-CAM gene was performed using LipofectAMINE Reagent (Invitrogen, Carlsbad, CA, USA). After 48 h of incubation, cells were transferred to medium containing Puromycin for selection. After 2 weeks selection, specific silencing of the targeted gene was confirmed by Western blot analysis. pSM2c vector and nonsilencing shRNA construct served as the control. Oncogene


3996 Specimens and immunohistochemistry Specimens with the diagnosis of lung adenocarcinoma were identified from the pathology archive of Tri-Service General Hospital. The histological differentiation and clinical stage was determined according to World Health Organization criteria. For further analysis, tumor stage was subdivided into two groups: (1) local tumors and (2) invasive tumors (Shieh et al., 2005). The antibodies included anti-human Ep-CAM antibody (1:500) (Transduction Laboratory, Lexington, KY, USA). Immunodetection was performed with a standard avidin– biotin–peroxidase complex detection kit (Dako Corp, Carpinteria, CA, USA). DNA extraction, bisulfite modification and MSP Genomic DNA was isolated using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA). The genomic DNA was modified by EZ DNA Methylation Kit (Zymo Research, Orange, CA, USA). The primers used for MSP: EpCAM-M1F-50 -TTTAACGTCGTTATGGAGACGA-30 and EpCAM-M1R-50 -GCTAATACTCGTTAATAAATCACCG-30 ; EpCAM-U1F-50 -TTTAATGTTGTTATGGAGATGA-30 and EpCAM-U1R-50 -ACCACTAATACTCATTAATAAATCACC AC-30 were purchased from MDBio (Frederick, MD, USA). 5Aza-dC and TSA treatment Cells were treated with mock or 1 mM 5-aza-dC (Sigma, St Louis, MO, USA) for 48 h or with 300 nM TSA (Sigma) for 24 h, as described previously (Cameron et al., 1999). Luciferase reporter assay The Ep-CAM promoter region (250 to þ 90) was cut from TOPO PCR 2.1 at the KpnI and XhoI sites and then inserted

into a luciferase reporter vector PGL3 (Promega, Madison, WI, USA) and digested with KpnI and XhoI to generate Ep-CAM-luc plasmid. PGL3 basic was used as a negative control (see Supplementary methods for details). In vitro methylation Plasmid DNA, Ep-pro, was methylated by incubating 1 mg of DNA with 2.5 U of SssI methylase or DNMT1 (New England Biolabs, Beverly, MA, USA) and 160 mmol/l S-adenosylmethionine for 2 h at 371C followed by a 20 min incubation at 651C. DNA was precipitated and suspended in 10 ml TE buffer (10 mmol/l Tris–HCl, 1 mmol/l ethylenediaminetetraacetic acid) and cotransfected into cells with a pRL-SV40 as described above. After 48 h, cells were washed with phosphatebuffered saline, harvested and analysed for luciferase and renilla activity using a Luciferase Kit (Promega, Madison, WI, USA). ChIP assays ChIP assay was performed according to the manufacturer’s instruction (ChIP Assay Kit, Upstate C, Lake Placid, NY, USA) (see Supplementary methods for details). Acknowledgements This work was supported by grants from the National Health Research Institute 92A1-PAPP01-1, National Science Council NSC 95-2314-B-016-034, Tri-Service General Hospital TSGHC95-24 and National Defense Medical Center DOD 95-01-04, Taiwan.

References Armstrong A, Eck SL. (2003). EpCAM: a new therapeutic target for an old cancer antigen. Cancer Biol Ther 2: 320–326. Basak S, Speicher D, Eck S, Wunner W, Maul G, Simmons MS et al. (1998). Colorectal carcinoma invasion inhibition by CO17-1A/GA733 antigen and its murine homologue. J Natl Cancer Inst 90: 691–697. Cameron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB. (1999). Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer. Nat Genet 21: 103–107. Chu YW, Yang PC, Yang SC, Shyu YC, Hendrix MJ, Wu CW et al. (1997). Selection of invasive and metastatic subpopulations from a human lung adenocarcinoma cell line. Am J Resp Cell Mol Biol 17: 353–360. Drapkin R, Crum CP, Hecht JL. (2004). Expression of candidate tumor markers in ovarian carcinoma and benign ovary: evidence for a link between epithelial phenotype and neoplasia. Hum Pathol 35: 1014–1021. Fuks F. (2005). DNA methylation and histone modifications: teaming up to silence genes. Curr Opin Genet Dev 15: 490–495. Gastl G, Spizzo G, Obrist P, Dunser M, Mikuz G. (2000). Ep-CAM overexpression in breast cancer as a predictor of survival. Lancet 356: 1981–1982. Jojovic M, Adam E, Zangemeister-Wittke U, Schumacher U. (1998). Epithelial glycoprotein-2 expression is subject to regulatory processes in epithelial–mesenchymal transitions during metastases: an investigation of human cancers transplanted into severe combined immunodeficient mice. Histochem J 30: 723–729. Oncogene

Jones PA, Baylin SB. (2002). The fundamental role of epigenetic events in cancer. Nat Rev Genet 3: 415–428. Joo M, Kim H, Kim MK, Yu HJ, Kim JP. (2005). Expression of Ep-CAM in intestinal metaplasia, gastric epithelial dysplasia and gastric adenocarcinoma. J Gastroenterol Hepatol 20: 1039–1045. Kim JH, Herlyn D, Wong KK, Park DC, Schorge JO, Lu KH et al. (2003). Identification of epithelial cell adhesion molecule autoantibody in patients with ovarian cancer. Clin Cancer Res 9: 4782–4791. Klose RJ, Bird AP. (2006). Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci 31: 89–97. Kumble S, Omary MB, Fajardo LF, Triadafilopoulos G. (1996). Multifocal heterogeneity in villin and Ep-CAM expression in Barrett’s esophagus. Int J Cancer 66: 48–54. Litvinov SV, Velders MP, Bakker HA, Fleuren GJ, Warnaar SO. (1994). Ep-CAM: a human epithelial antigen is a homophilic cell–cell adhesion molecule. J Cell Biol 125: 437–446. McLaughlin PM, Trzpis M, Kroesen BJ, Helfrich W, Terpstra P, Dokter WH et al. (2004). Use of the EGP-2/Ep-CAM promoter for targeted expression of heterologous genes in carcinoma derived cell lines. Cancer Gene Ther 11: 603–612. Munz M, Kieu C, Mack B, Schmitt B, Zeidler R, Gires O. (2004). The carcinoma-associated antigen EpCAM upregulates c-myc and induces cell proliferation. Oncogene 23: 5748–5758. Osta WA, Chen Y, Mikhitarian K, Mitas M, Salem M, Hannun YA et al. (2004). EpCAM is overexpressed in breast cancer and is a potential target for breast cancer gene therapy. Cancer Res 64: 5818–5824.


3997 Rao CG, Chianese D, Doyle GV, Miller MC, Russell T, Sanders RA et al. (2005). Expression of epithelial cell adhesion molecule in carcinoma cells present in blood and primary and metastatic tumors. Int J Oncol 27: 49–57. Robertson KD. (2001). DNA methylation, methyltransferases, and cancer. Oncogene 20: 3139–3155. Seligson DB, Pantuck AJ, Liu X, Huang Y, Horvath S, Bui MH et al. (2004). Epithelial cell adhesion molecule (KSA) expression: pathobiology and its role as an independent predictor of survival in renal cell carcinoma. Clin Cancer Res 10: 2659–2669. Shieh YS, Lai CY, Kao YR, Shiah SG, Chu YW, Lee HS et al. (2005). Expression of axl in lung adenocarcinoma and correlation with tumor progression. Neoplasia 7: 1058–1064. Songun I, Litvinov SV, van de Velde CJ, Pals ST, Hermans J, van Krieken JH. (2005). Loss of Ep-CAM (CO17-1A) expression predicts survival in patients with gastric cancer. Br J Cancer 92: 1767–1772. Spizzo G, Gastl G, Obrist P, Fong D, Haun M, Grunewald K et al. (2006). Methylation status of the Ep-CAM promoter region in human breast cancer cell lines and breast cancer tissue. Cancer Lett [doi:10.1016/j.canlet.2006.03.002].

Takes RP, Baatenburg de Jong RJ, Schuuring E, Hermans J, Vis AA, Litvinov SV et al. (1997). Markers for assessment of nodal metastasis in laryngeal carcinoma. Arch Otolaryngol Head Neck Surg 123: 412–419. Takes RP, Baatenburg de Jong RJ, Wijffels K, Schuuring E, Litvinov SV, Hermans J et al. (2001). Expression of genetic markers in lymph node metastases compared with their primary tumours in head and neck cancer. J Pathol 194: 298–302. Went P, Dirnhofer S, Salvisberg T, Amin MB, Lim SD, Diener PA et al. (2005). Expression of epithelial cell adhesion molecule (EpCam) in renal epithelial tumors. Am J Surg Pathol 29: 83–88. Went P, Vasei M, Bubendorf L, Terracciano L, Tornillo L, Riede U et al. (2006). Frequent high-level expression of the immunotherapeutic target Ep-CAM in colon, stomach, prostate and lung cancers. Br J Cancer 94: 128–135. Went PT, Lugli A, Meier S, Bundi M, Mirlacher M, Sauter G et al. (2004). Frequent EpCam protein expression in human carcinomas. Hum Pathol 35: 122–128. Winter MJ, Nagtegaal ID, van Krieken JH, Litvinov SV. (2003). The epithelial cell adhesion molecule (Ep-CAM) as a morphoregulatory molecule is a tool in surgical pathology. Am J Pathol 163: 2139–2148.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene