SEARCH

SEARCH BY CITATION

Keywords:

  • DNA methyltransferase 3b;
  • esophageal squamous cell carcinoma;
  • prognosis

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

BACKGROUND:

The identification of potential tumor markers can improve therapeutic planning and patient management. The objective of this study was to highlight the role of DNA methyltransferase 3b (DNMT3b) in esophageal squamous cell carcinoma (SCC).

METHODS:

One hundred seventy-three esophageal SCC samples were analyzed using immunohistochemical staining to correlate the expression of DNMT3b with clinical outcome. Furthermore, a human esophageal SCC cell line, CE81T, was selected for cellular and animal experiments to investigate changes in tumor behavior and treatment response after the manipulation of DNMT3b expression.

RESULTS:

The incidence of nuclear DNMT3b immunoreactivity in esophageal cancer specimens was significantly higher than in nonmalignant epithelium, and this incidence was linked positively to developing distant metastasis (56% in localized disease vs 80% in distant metastasis; P = .002). Furthermore, increased expression of DNMT3b was linked significantly to lower treatment response rates (P = .002) and reduced survival rates (P = .000). Inhibition of DNMT3b expression resulted in slower cellular proliferation, increased cell death, a less invasive capacity, and less epithelial-mesenchymal-transition changes. Moreover, DNMT3b silencing vectors sensitized esophageal cancer cells to irradiation and cisplatin treatment. The current results also indicated that constitutional activation of signal transducer and activator of transcription 3 (STAT3) signaling associated with inhibited expression of suppressor of cytokine signaling 3 (SOCS3) may be the mechanism underlying more aggressive tumor growth in DNMT3b-positive esophageal cancer.

CONCLUSIONS:

DNMT3b was linked significantly to a poor prognosis for patients with esophageal cancer. Moreover, the current results indicated that targeting this enzyme may be a promising strategy for treating esophageal cancer, as evidenced by inhibited aggressive tumor behavior and treatment resistance. Cancer 2012. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Esophageal cancer, an aggressive upper gastrointestinal malignancy, is the sixth most common cause of cancer-related deaths. It normally presents as a locally advanced disease and is associated with a poor outcome.1 In terms of histology, esophageal cancer can be divided into squamous cell carcinoma (SCC) and adenocarcinoma, and there are marked differences in incidence, natural history and treatment outcomes between these carcinomas.2 The histologic progression of esophageal SCC, which includes transition from normal squamous epithelium to dysplasia, carcinoma in situ, and, ultimately, invasive SCC,3 has been relatively established, but knowledge concerning molecular progression is limited. Furthermore, the lack of adequate therapeutic options for patients with esophageal cancer indicates that identification of new molecular markers could be important for the effective management of this cancer.

Multiple genetic and epigenetic alternations are involved in the growth of esophageal cancer.4, 5 Aberrant DNA methylation plays a key role in carcinogenesis, leading to epigenetic silencing of tumor-suppressor genes involved in cell cycle regulation, apoptosis, and DNA repair.6, 7 DNA hypermethylation frequently occurs in precancerous tissue in addition to cancer tissue. Hypermethylation events within the promoter regions of p14ARF (the alternate reading frame [ARF] product of cyclin-dependent kinase 2A [CDKN2A]), the CDKN2A tumor suppressor protein p16INK4a, and E-cadherin have been related to the development of esophageal SCC.8, 9 DNA methylation is typically mediated by DNA methyltransferases (DNMT); and, in mammalian genomes, there are 3 DNMT genes: DNMT3a and DNMT3b are responsible for de novo methylation and modify unmethylated DNA, and it is believed that DNMT1 is responsible for maintaining methylation patterns.10, 11 DNMT activity reportedly is increased in cancer cells and may be related to tumor aggressiveness and a poor prognosis.12-14 Epigenetic silencing reported in gastric cancer and esophageal cancer is associated with overexpression of the DNMT family.5, 15, 16 Overexpression of DNMT3b, but not DNMT1 or DNMT3a, has been observed in several cancer types, suggesting that DNMT3b plays an important role during tumorigenesis.17, 18 DNMT3b is overexpressed in esophageal and gastrointestinal adenocarcinomas,15, 19, 20 but its role in esophageal SCC and tumor progression remains to be elucidated. Therefore, in this study, we investigated the role of DNMT3b in esophageal SCC in vitro and in vivo and examined the correlation between its expression level and the clinical outcomes of patients with esophageal SCC using immunochemical staining analysis.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Characteristics of Patients

This study was approved by the Institutional Review Board of our Chang Gung Memorial Hospital. Patients who disrupted the treatment regimen and those who underwent surgery alone for early stage esophageal cancer were excluded from the study. One hundred seventy-three patients with esophageal SCC who completed curative treatment were enrolled in the study. The curative treatment for esophageal cancer included neoadjuvant, concurrent chemoradiation (CCRT) combined with surgery when the tumor was deemed resectable or definitive CCRT for inoperable disease according to the guidelines proposed by the oncology team at our hospital. On completion of neoadjuvant CCRT, patients underwent a repeat computed tomography scan and endoscopic examination to determine their response to treatment. If the tumor was considered resectable, then patients underwent surgery for the residual tumor. A pathologic complete response was defined as no residual invasive tumor in the surgical specimen for patients who underwent surgical intervention. For patients who refused surgery or for whom surgery was contraindicated, a second course of CCRT was administered with 2 additional courses of chemotherapy and radiotherapy reaching a total radiotherapy dose from 60 to 63 grays Gy. We retrospectively collected the specimens from these 173 patients at diagnosis for immunochemical analysis. The main study endpoints were overall survival, disease-free survival, and response to neoadjuvant therapy. Statistically, survival probabilities were analyzed using the Kaplan-Meier method. The significance of between-group differences was assessed using the log-rank test. Multivariate analyses were performed using a Cox regression model for overall survival.

Immunohistochemical Staining

Formalin-fixed, paraffin-embedded tissues from 173 patients with esophageal cancer were used for immunochemical staining. Among these patients, 20 dissected esophageal cancer specimens also were constructed into tissue microarray (TMA) blocks using AutoTiss 1000 (EverBio Technology, Inc., Richmond, British Columbia, Canada) for immunochemical analysis. The TMA block contained esophageal SCCs and the adjacent nonmalignant epithelia. The quality of TMA slides was confirmed by the pathologist using hematoxylin and eosin-stained slides. Formalin-fixed, paraffin-embedded tissues were cut into 4-μm sections, mounted on slides, deparaffinized with xylene, and dehydrated using a graded ethanol series. They were incubated overnight with antibodies against DNMT3b and phosphorylated signal transducer and activator of transcription 3 (p-STAT3) (1:100 dilution) at 4°C. Immunoreactivity for DNMT3b and p-STAT3 proteins was scored using a semiquantitative method by evaluating the number of positive tumor cells over the total tumor cells. The samples of adjacent nonmalignant epithelia contained in slides were used as an internal reference. Scores were assigned by using 5% increments. The staining was scored independently by 2 observers who were blind to the clinical outcome of patients. Discordant scores were reviewed, and consensus was reached. In the current study, we considered >10% positive staining of tumor cells a positive immunohistochemical (IHC) score, which was defined in receiver operating characteristic (ROC) curve analysis. The 72 patients who received treatment between 2002 and 2006 were used as the testing cohort, and 101 patients who received treatment in subsequent years were used as the validation cohort.

Cell Culture and Reagents

The human esophageal cancer cell line CE81T, derived from well differentiated SCC of the esophagus, was obtained from the Bioresource Collection and Research Center (Hsinchu, Taiwan). The human normal intestine cell line CCL-241 was obtained from the American Type Culture Collection (Rockville, Md). The DNMT inhibitor 5-aza-2′-deoxycytidine and the Janus kinase (JAK) inhibitor AG490 were obtained from Sigma Chemical Company (St. Louis, Mo) and Calbiochem (La Jolla, Calif), respectively. STAT3 small interfering RNA (siRNA) was obtained from R&D (Minneapolis, Minn), and the DNMT3b-green fluorescent protein (GFP) silencing vector and GFP-control vector were purchased from IvivoGen (San Diego, Calif).

Immunoblotting

For Western blot analyses of whole cells and esophageal tissue extracts, specimens were homogenized and/or treated with lysis buffer (Calbiochem). The esophageal tissue specimens comprised 6 cancer tissue specimens and 6 nonmalignant epithelium specimens from the esophagus. An NE-PER kit (Pierce, Rockford, Ill) was used to separate nuclear and cytoplasmic proteins. To determine the in vitro effects of STAT3 siRNA, the JAK inhibitor, and a DNMT inhibitor, proteins were extracted from cells in the presence or absence of STAT3 siRNA for 72 hours, in the presence or absence of either 50 μM AG490 for 24 hours or 5 μM 5-aza-2′-deoxycytidine for 36 hours, respectively. Proteins related to DNA repair were examined 30 minutes after 9 grays of irradiation.

Real-Time Reverse Transcriptase-Polymerase Chain Reaction Analysis

Real-time reverse transcriptase polymerase chain reaction (RT-PCR) analysis was performed on RNA extracted from cells and tissue specimens (6 cancer tissue specimens and 6 nonmalignant esophageal tissues; 2 specimens were run in each lane). RNA (2 μg) was reverse-transcribed with a random primer to obtain the first combinational DNA strand. The sequences of primers for DNMT3b and for suppressor of cytokine signaling 3 (SOCS3) were 5′-GACTCGAAGACGCACAGCTG −3′/5′-CTCGGTCTTTGCCGTTGTTATAG′, and 5′-TGATCCGCGACAGCTCG −3′/5′-TCCCAGACTGGGTCTTGACG′, respectively. To control for loading differences, a β-actin primer was used as a control. The optimized PCR was performed on an iCycler Iq multicolor real-time PCR detection system (Bio-Rad, Hercules, Calif). Significant fluorescent PCR signals from carcinoma tissues were normalized to the mean value of the signals obtained from the nonmalignant tissues and cells under control conditions.

Cell Migration and Cell Invasion Assay

Capacities for cell invasion were determined using a cell invasion assay (Trevigen, Gaithersburg, Md). Cells were starved by incubating them with serum-free medium for 24 hours before performing the assay. The top chambers were precoated with basement membrane extract. After a 24-hour incubation, the number of cells in the bottom chamber was determined by measuring the fluorescent anion calcein released from intracellular calcein acetoxymethylester. To validate the cell migration experiments, scratch assays were performed. A 2-mm-wide scratch was drawn across each cell layer using a pipette tip, and the plates were photographed at the times indicated in the figures.

Immunofluorescent Staining

Cells that demonstrated exponential growth were seeded on to coverslips for immunofluorescent staining with or without treatment. At the stated times after treatment, the cells were fixed, permeabilized with 2% paraformaldehyde for 5 minutes, and washed in phosphate-buffered saline with Tween-20 (PBST). The slides were incubated for 1 hour at room temperature with antibody against E-cadherin and with a fluorescein isothiocyanate-conjugated secondary antibody and then counterstained with 4′,6-diamidino-2-phenylindole (DAPI). For DNMT3b, p-STAT3, SOCS3 and E-cadherin, slides were incubated with Texas Red-conjugated secondary antibody for 1 hour and then counterstained with DAPI to observe nuclei. After 2 washes in PBST, specific target proteins were observed using a fluorescence microscope.

Tumor Xenografts

Eight-week-old male athymic nude mice were used, and all animal experiments conformed to the protocols approved by the Experimental Animal Committee of our hospital. Cells (1 × 106 cells per implantation, 5 animals per group) were implanted subcutaneously on the dorsal gluteal region. Tumor size was measured every 3 days after implantation (day 0), and tumor volume was calculated assuming an ellipsoid shape. The effects of DNMT3b on tumor growth were investigated in vivo and were determined by analyzing the tumor growth curves for each cell type.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Level of DNA Methyltransferase 3b in Esophageal Squamous Cell Carcinoma

The expression of DNMT3b in the tissue specimens (malignant vs adjacent nonmalignant) was examined by Western blot analysis, and messenger RNA was demonstrated observed with real-time RT-PCR (Fig. 1a,b). According to our data, cancer specimens had a significantly higher levels of DNMT3b than nonmalignant specimens. The IHC data from TMA slides confirmed the finding that DNMT3b was overexpressed in tumor tissues compared with adjacent, nonmalignant epithelial tissues (Fig. 1c). Of the 173 esophageal cancer tissues that we assayed for DNMT3b using IHC analysis, 65% (112 tissues) had positive immunoreactivity (60% [66 of 109 tissues] in tumors that were classified as ≤T3 vs 72% [46 of 64 tissues] in T4 tumors; P = .13) (Fig. 1d). Furthermore, there was a positive correlation between DNMT3b overexpression and cancers that developed distant metastasis. Eighty percent (51 of 64 patients) of individuals who developed distant metastases had positive staining for DNMT3b, and only 56% (61 of 109 patients) of individuals without distant metastases expressed DNMT3b (P = .002) (Table 1).

thumbnail image

Figure 1. Levels of DNA methyltransferase 3b (DNMT3b) in were examined in patients with esophageal squamous cell carcinoma (SCC). (a) Levels of DNMT3b in 6 paired cancer (C) and adjacent nonmalignant (N) tissue specimens were examined by reverse transcriptase-polymerase chain reaction (RT-PCR) analysis (with 2 specimens per lane) and by real-time RT-PCR. For real-time RT-PCR analysis, the y-axis on the chart indicates the ratio of DNMT3b (messenger RNA [mRNA]) to β-actin in cancer and nonmalignant tissue specimens; columns, means of 3 separate experiments; bars, standard deviations; asterisks, P < .05. CE81T is a human esophageal SCC cell line, and CCL-241 is a human normal intestine cell line. (b) Levels of DNMT3b in 6 paired cancer and adjacent nonmalignant tissue specimens were examined by Western blot analysis (2 specimens per lane). (c) These representative slides reveal immunohistochemical staining with DNMT3b antibody in esophageal cancer and adjacent nonmalignant epithelium from tissue microarray blocks. (d) Immunohistochemical staining with anti-DNMT3b antibody is observed in human esophageal cancer specimens (original magnification, ×100 magnification on the left; ×400 on the right).

Download figure to PowerPoint

Table 1. Clinicopathologic Characteristics of Patients With Esophageal Cancer: Immunohistochemical Investigation
 IHC-DNMT3b Status: No. of Patients 
CharacteristicNegativePositiveP
  • Abbreviations: DNMT3b, DNA methyltransferase 3b; IHC, immunohistochemistry; LN, lymph node; neoadjuvant Tx, neoadjuvant chemoradiotherapy; s/p, status post.

  • a

    P < .05.

Total no.61112 
Age, y   
 ≥563359.858
 <562853 
Tumor location   
 Upper third1325.805
 Middle third3466 
 Lower third1421 
Tumor stage  .138
 I-II2025 
 III-IV4187 
LN metastasis  .274
 Negative2028 
 Positive4184 
Response to neoadjuvant Tx  .002a
 Partial response or better5066 
 Poor response1146 
Surgery s/p neoadjuvant Tx  .634
 Yes1837 
 No4375 
Locoregional recurrence/persistent  .000a
 No4241 
 Yes1971 
Distant metastasis  .002a
 Negative4861 
 Positive1351 

The Role of DNA Methyltransferase 3b in Tumor Growth

To investigate whether alternating DNMT3b expression plays a role in aggressive tumor growth, CE81T cancer cells were transfected with a DNMT3b-GFP silencing vector. The DNMT3b silencing vector significantly inhibited DNMT3b expression in cancer cells, as demonstrated in Figure 2a. The effect of DNMT3b on tumor cell growth was determined by viable cell counting over 6 days, and by colony formation and observation in xenograft tumors. The data indicated that DNMT3b silencing vectors significantly inhibited tumor growth in vitro (Fig. 2b) and in vivo (Fig. 2c). Changes in the apoptotic rate and in cell cycle distribution also were measured. The cell death rate measured by flow cytometry analysis was increased from 5.9% ± 1.4% to 13.1% ± 2.2% in CE81T cells after treatment with a DNMT inhibitor (Fig. 2d). The DNMT3b silencing vector also obviously increased cell death as measured by propidium iodide staining. Furthermore, the DNMT3b silencing vector resulted in cell cycle arrest (Fig. 2e) and in significant increases in autophagy in immunofluorescence studies with an antibody against LC3 (Fig. 2f). We also investigated whether the vector affected the expression of apoptosis-related and cell cycle-related proteins. Figure 2g demonstrates that suppressed DNMT3b was associated with increased levels of p53 and decreased levels of cyclin D1, bcl-2, and survivin.

thumbnail image

Figure 2. The role of DNA methyltransferase 3B (DNMT3b) in tumor growth is illustrated. (a) The effect of the DNMT3b-green fluorescent protein (GFP) silencing vector on the level of DNMT3b in human esophageal squamous cell carcinoma cells (CE81T cells) is revealed by with (Top) immunofluorescence and (Bottom) Western blot analysis. The results are shown in representative slides. DNMT3b levels were decreased significantly by DNMT3b-GFP silencing vector (DM) compared with the control-GFP vector (V). CCL-241 are cells from a human normal intestine cell line. DAPI indicates 4′,6-diamidino-2-phenylindole; W, wild type; V, cells transfected with control vector alone; DM−, cells transfected with DMNT3b silencing vector. (b) The effect of DNMT3b on the proliferation rate of CE81T cancer cells was evaluated by colony formation and viable cell counting. The y-axis on the chart represents the viable cell number, and the x-axis indicates the day (D) cells were counted. Points indicate the means of separate experiments; bars, standard deviations; asterisk, P < .05. (c) The effect of DNMT3b inhibition on tumor xenograft tumor growth is illustrated. Each point on the chart indicates the mean of 3 separate experiments; bars, standard deviation; asterisk, P < .05; +, positive. The expression of DNMT3b also was evaluated by immunochemical staining and immunofluorescent staining (IF) for xenografts (original magnification, ×400). (d) The effect of DNMT3b inhibition on cell death was demonstrated by fluorescence-activated cell sorting (FACS) and immunofluorescence staining. The results are shown in representative slides. Q indicates quartile; PI, propidium iodide. (e) Cells were analyzed by FACS for DNA content (Top) under control conditions and (Bottom) with DNMT3b inhibition. The data are illustrated in representative graphs. (f) The effects of DNMT3b silencing vector on autophagy are demonstrated by immunofluorescence. The results are shown from representative slides. The level of autophagosomal marker LC3 II also was examined by Western blot analysis. W indicates wild type; V, cells transfected with control vector alone; DM−, cells transfected with DMNT3b silencing vector; DI, cells transfected with DNMT inhibitor. (g) The effects of DNMT3b silencing vector on the expression of apoptosis-related and cell cycle-related proteins were evaluated by Western blot analysis.

Download figure to PowerPoint

The Role of DNA Methyltransferase 3b in Tumor Invasion and Underlying Mechanisms

Inhibition of DNMT3b expression using DNMT3b silencing vectors attenuated the invasion capacity of esophageal cancer cells, as demonstrated using migration scratch and invasion assays (Fig. 3a). Epithelial mesenchymal transition (EMT) is a key event in tumor invasiveness. Therefore, we sought to determine whether this was the mechanism underlying the effects of DNMT3b on esophageal cancer invasiveness. The DNMT3b silencing vector induced esophageal cancer cells to increase their epithelial characteristics, as determined by changes in expression of E-cadherin and Snail, which is an E-cadherin repressor (Fig. 3b,c). EMT reportedly is associated with several invasion-related factors, including hypoxia-inducible factor 1α (HIF-1α), vascular endothelial growth factor (VEGF), and matrix metalloproteinase 9 (MMP-9). Our data revealed that inhibiting DNMT3b by the silencing vector resulted in lower expression of HIF-1α, VEGF, and MMP-9 (Fig. 3c). Although the molecular events underlying EMT are complex, the activation of STAT3 signaling reportedly plays important roles in the induction of aggressive tumor behavior and EMT changes in cancer.21, 22 Moreover, we previously reported that activated STAT3 signaling induced by proinflammatory cytokines was critical for cancers in the upper aerodigestive tract.23 Accordingly, this we also examined the relation of STAT3 activation with DNMT3b and EMT. When STAT3 signaling was inhibited by STAT3 siRNA or the JAK inhibitor AG490, the decrease in EMT and apoptosis-related protein was comparable to that induced by the DNMT3b silencing vectors (Fig. 3d). Moreover, the DNMT 3b silencing vector significantly inhibited the activation of STAT3, as determined using immunofluorescence and Western blot analyses (Fig. 3d,e). Therefore, the results suggest that alternation of STAT3 activation and the subsequent EMT changes may be responsible for the aggressive tumor behavior noted in DNMT3b-positive esophageal cancers.

thumbnail image

Figure 3. The role of DNA methyltransferase 3B (DNMT3b) inhibition in aggressive tumor behavior and epithelial mesenchymal transition (EMT) changes are illustrated. (a) The effect of DNMT3b inhibition on the migration and invasion ability of esophageal cancer cells is shown. The plates were photographed at the indicated times (H indicates hour). The results are shown from representative slides and quantitative data (the y-axis indicates the relative ratio normalized to the distance under the control condition). (b) Changes in E-cadherin in cells were evaluated, and the results are shown in representative slides. E-cadherin expression quantified by calculating the numbers of cells that were positive for E-cadherin immunofluorescence divided by the total cell numbers for each condition. The y-axis represents the ratio normalized by the respective wild-type value. DAPI indicates 4′,6-diamidino-2-phenylindole. (c) The changes in EMT-associated proteins were evaluated in transfectants by Western blot analysis. W indicates wild type; V, cells transfected with control vector; DM−, cells transfected with DMNT3b silencing vector; DI, cells transfected with DNMT inhibitor; VEGF, vascular endothelial growth factor; MMP-9, matrix metalloproteinase 9; HIF-1 alpha, hypoxia-inducible factor 1α. (d) The effect of DNMT3b silencing vector on the activation of signal transducer and activator of transcription 3 (STAT3) signaling was examined by Western blot analysis. P-STAT3 indicates phosphorylated STAT3; AG, cells treated with the Janus kinase inhibitor AG490; SI, cells transfected with STAT3 small-interfering RNA. (e) The effect of DNMT3b silencing vector on the activation of STAT3 was examined by using immunofluorescence analysis.

Download figure to PowerPoint

Constitutional Activation of Signal Transducer and Activator of Transcription 3 Induced by DNA Methyltransferase 3b

We further examined the role of STAT3 activation in esophageal SCC. According to IHC data from the TMA blocks, approximately 50% of cancer tissue specimens had positive staining for p-STAT3 compared with adjacent, nonmalignant epithelial tissues (Fig. 4a). Of the 173 cancer specimens, positive staining for p-STAT3 was evident in 47% (82 of 173) of esophageal cancer tissues (Fig. 4b). Moreover, a significant, positive correlation was observed in a cancer specimen that expressed DNMT3b and p-STAT3 (Fig. 4c). We also investigated the mechanism underlying the link between DNMT3b overexpression and constitutional activation of STAT3. It has been reported that SOCS3 is a well characterized mechanism that regulates STAT3 activation24; therefore, we was hypothesized that SOCS3 may mediate the regulation of STAT3 activation by DNMT3b. We examined the level of SOCS3 in esophageal cancer cells using RT-PCR and Western blot analyses. Figure 4d indicates that the DNMT3b silencing vector significantly increased the level of SOCS3 associated with the down-regulation of STAT3 activation. On the basis of these data, we suggest that the constitutional activation of STAT3 may be mediated by the inhibition of SOCS3 evident in DNMT3b-positive esophageal cancers.

thumbnail image

Figure 4. The role of DNA methyltransferase 3B (DNMT3b) inhibition on the activation of signal transducer and activator of transcription 3 (STAT3) is illustrated. (a) These are representative slides of immunohistochemical staining with phosphorylated STAT3 (p-STAT3) antibody in samples of esophageal cancer and adjacent nonmalignant epithelium from tissue microarray blocks. (b) Immunohistochemical staining with anti-p-STAT3 antibody is observed on representative slides of human esophageal cancer specimens (original magnification, ×100 on the left, ×400 on the right). (c) P-STAT3 levels correlate positively with DNMT3b expression in human esophageal cancer specimens (P = .0000). Representative slides from a selected tumor specimen were stained positive for both DNMT3b and p-STAT3, and another tumor specimen was stained negative for both DNMT3b and p-STAT3 (original magnification, ×400). IHC indicates immunohistochemistry. (d) The effect of DNMT3b on the level of suppressor of cytokine signaling-3 (SOCS3) was examined by immunofluorescence, Western blot analysis, and real-time reverse transcriptase-polymerase chain reaction analysis. On the chart to the right, the y-axis indicates the ratio of SOCO3 in cells under different conditions and wild-type (W) cells under the control condition. V indicates cells transfected with control vector; DM−, cells transfected with DMNT3b silencing vector; DI, cells transfected with DNMT inhibitor; mRNA, messenger RNA.

Download figure to PowerPoint

DNA Methyltransferase 3b Correlates With Treatment Response in Esophageal Squamous Cell Carcinoma

Regarding clinical data, the expression of DNMT3b was linked significantly to a lower response rate to curative treatment in 173 patients with esophageal SCC (Table 1) and with a lower pathologic complete response rate in 55 patients who underwent surgical intervention (Table 2). The findings strongly suggest that DNMT3b contributes to treatment resistance in patients with esophageal SCC. Therefore, the role of DNMT3b in treatment resistance and the mechanisms underlying this were investigated in our study. Colony-forming assay data (Fig. 5a), annexin-propidium iodide staining (Fig. 5b), and viable cell counts (Fig. 5c) demonstrated that the DNMT3b silencing vector significantly sensitized esophageal cancer cells to irradiation and cisplatin treatment. To explore the mechanisms underlying the sensitization effect induced by DNMT3b inhibition, changes in STAT3 signaling and DNA damage in irradiated transfectant xenografts were investigated. Figure 5d demonstrates that the DNMT3b silencing vector significantly attenuated the increase in DNMT3b and activation of STAT3 induced by irradiation. Furthermore, the ability to perform DNA repair is the major determinant of radiosensitivity, and inhibiting DNMT3b significantly augmented radiation-induced levels of phosphorylated histone H2AX (p-H2AX), an indicator of DNA damage and oxidative DNA damage.

thumbnail image

Figure 5. The effects of DNA methyltransferase 3b (DNMT3b) on treatment response are illustrated (a) CE81T cells (a human esophageal squamous cell carcinoma cell line) were irradiated with 0 gray (Gy), 2 Gy, 4 Gy, and 6 Gy; and survival curves were generated by using colony-forming assays. Each point is an average of 3 experiments. Cells were radiosensitized by the DNMT3b silencing vector. An asterisk indicates P < .05. (b) Flow cytometry with Annexin V and propidium iodide (PI) staining was used to reveal apoptosis in irradiated cells with or without DNMT inhibitor. RT indicates radiotherapy; DAPI, 4′,6-diamidino-2-phenylindole. (c) The influence of DNMT3b silencing vector on the cytotoxicity induced by cisplatin is illustrated. The numbers of viable cells were counted after incubation in drug-free medium for a further 48 hours after a 2-hour exposure to increasing concentrations of cisplatin. The y-axis represents the relative cell number normalized by the number of untreated cells. Columns indicate the means of 3 separate experiments, and bars indicate standard deviations. (d) The influence of DNMT3b inhibition on phosphorylated signal transducer and activator of transcription 3 (p-STAT3), phosphorylated histone H2AX (p-H2AX), and 8-oxoguanine (8-oxoG) was evaluated by Western blot analysis and immunochemical staining in irradiated tumor xenografts. W indicates wild type; V, cells transfected with control vector; DM−, cells transfected with DMNT3b silencing vector.

Download figure to PowerPoint

Table 2. The Significance of DNA Methyltransferase 3b Expression Stratified by Complete Response to Neoadjuvant Treatment in 55 Patients With Esophageal Cancer Who Underwent Surgical Resection
 DNMT3b: No. of Patients 
Pathologic CRPositiveNegativeP
  1. Abbreviations: CR, complete response; DNMT3b, DNA methyltransferase 3b.

Positive510.001
Negative328 

Correlation Between DNA Methyltransferase 3b Levels and Clinical Outcome

The role of DNMT3b in the clinical outcome of patients with esophageal cancer was evaluated further using IHC staining. The median age of all 173 patients in this analysis was 56 years (range, 33-86 years). Histologically, all primary tumors were SCCs. The patients were classified further as either DNMT3b-positive or DNMT3b-negative on the basis of the IHC results. In addition to treatment resistance, DNMT3b was correlated significantly with a greater probability of developing distant metastasis and a higher recurrence rate after curative treatment (Table 1). In univariate and multivariate analyses, poor treatment response, no tumor resection, and positive staining for DNMT3b and p-STAT3 were linked significantly to shorter survival (Tables 3 and 4; Fig. 6). Furthermore, in the subgroup of 118 patients who received definite CCRT, positive staining for DNMT3b and p-STAT3 had predictive power concerning survival (Table 5). However, these proteins did not have significant predictive value for prognosis in the subgroup of 55 patients who underwent surgical intervention.

thumbnail image

Figure 6. The correlation between DNA methyltransferase 3b (DNMT3b) level and clinical outcome is illustrated. Survival differences are illustrated according to positive DNMT3b staining for (a) all 173 patients, (b) the subgroup that received definite concurrent chemoradiation (CCRT), and (c) the subgroup that received preoperative CCRT and also underwent surgery. These Kaplan-Meier overall survival curves indicate that the DNMT3b-positive group was linked to shorter survival than the respective negative group. Cum survival indicate cumulative survival.

Download figure to PowerPoint

Table 3. Univariate Analysis to Determine the Factors Associated With Prognosis in Patients With Esophageal Cancer
VariableOSDFSP
  • Abbreviations: DFS, disease-free survival; DMT3b, DNA methyltransferase 3b; OS, overall survival; p-STAT3, phosphorylated signal transducer and activator of transcription 3.

  • a

    P < .05.

Clinical stage.067.001a 
Tumor resection.004a.000a 
Positive staining for DNMT3b.000a.000a 
Positive staining for p-STAT3.000a.000a 
Locoregional recurrence.000a.000a 
Distant metastasis.014a.000a 
Treatment response.000a.000a 
Table 4. Multivariate Analysis to Determine Molecular Markers Associated With Prognosis in Patients With Esophageal Cancer: Overall Survival and Disease-Free Survival
VariableOR95% CIP
  • Abbreviations: CCRT, concurrent chemotherapy and radiotherapy; CI, confidence interval; DNMT3b, DNA methyltransferase 3b; OR, odds ratio; p-STAT3, phosphorylated signal transducer and activator of transcription 3.

  • a

    P < .05.

Overall survival   
 DNMT3b staining4.27591.9954-9.1628.0002a
 p-STAT3 staining1.94261.0834-3.4832.0258a
 Tumor resection0.38820.2182-0.6908.0013a
 Clinical stage1.27240.6681-2.4230.4636
 Treatment response2.29781.3587-3.8861.0019a
Disease-free survival   
 DNMT3b staining2.78591.6629-4.6117.0001a
 p-STAT3 staining1.55051.0008-2.4029.0492a
 Tumor resection0.35170.2230-0.5545.0000a
 Clinical stage1.81731.0746-3.0755.0259a
 Treatment response2.54131.6767-3.8519.0001a
Table 5. Multivariate Analysis to Determine the Factors Associated With Prognosis (Overall Survival) in Patients With Esophageal Cancer
VariableOR95% CIP
  • Abbreviations: CCRT, concurrent chemotherapy and radiotherapy; CI, confidence interval; CR, complete response; DNMT3b, DNA methyltransferase 3b; OR, odds ratio; p-STAT3, phosphorylated signal transducer and activator of transcription 3.

  • a

    P < .05.

Definite CCRT group   
 DNMT3b staining4.39581.8523-10.4322.0008a
 p-STAT3 staining2.63101.2344-5.6078.0122a
 Clinical stage1.33710.5722-3.1246.5024
 Treatment response2.29271.2395-4.2410.0082a
Preoperative CCRT and surgery group   
 DNMT3b staining4.66210.8961-24.2549.0673
 p-STAT3 staining1.48690.5346-4.1512.4460
 Clinical stage0.68860.2156-2.1991.5289
 Pathologic CR1.89520.3621-9.9189.4490

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Understanding the molecular mechanisms underlying the development of aggressive tumor growth in esophageal SCC is pivotal to identifying novel targets for pharmacologic intervention. It has been reported that a switch to accumulating DNA hypermethylation may be caused by the overexpression of DNA methyltransferases.25 Epigenetic gene silencing, promoted by DNMTs, has been observed in various malignancies, supporting the claim that DNMT genes are overexpressed in human cancers and during cellular transformation.7, 13, 26, 27 It has been reported that DNMT3b participates in the carcinogenesis of several cancer types,19, 28, 29 but the role of DNMT3b in esophageal SCC requires further investigation. In the current study, DNMT3b expression levels were greater in SCC of the esophagus than in nonmalignant tissues. Furthermore, positive staining for DNMT3b was associated preferentially with higher rates of distant metastasis and recurrence. To further investigate whether DNMT3b is responsible for aggressive tumor growth of esophageal SCC, we suppressed DNMT3b expression in esophageal cancer cells using a silencing vector. Data obtained from cell and xenograft tumor growth experiments revealed that inhibiting DNMT3b resulted in slower tumor growth associated with increased apoptosis, autophagy, and cell cycle arrest. Furthermore, attenuated invasiveness was noted in transfectants with DNMT3b silencing vectors using cellular invasion assays. The molecular and phenotypic changes involved in the transformation of an epithelial cell to a mesenchymal cell type appear to be functionally relevant to the invasive characteristics of epithelial tumors.30 It has been reported that EMT is associated with esophageal SCC progression.31 At the molecular marker level, EMT is characterized by a loss of E-cadherin and increased expression of vimentin and invasion-related factors.32 In the current study, the DNMT3b silencing vectors induced an increase in E-cadherin and decreases in Snail, VEGF, and MMP-9 in esophageal SCC. On the basis of these findings, we believe that EMT may be responsible for the increased invasiveness observed in DNMT3b-positive esophageal cancers.

Constitutive activation of STAT3 signaling reportedly contributes to oncogenesis by promoting proliferation and EMT and by inhibiting apoptosis.21, 33, 34 According to our previous study and other series,23, 27, 34, 35 activated STAT3 signaling induced by proinflammatory cytokines is critical for cancers in the upper aerodigestive tract. Furthermore, DNMT reportedly is associated with the activation of STAT3 signaling in various cancers.13, 27, 36, 37 Accordingly, we examined the link between STAT3 activation, DNMT3b, and EMT. When STAT3 signaling was inhibited, the decreases in EMT and apoptosis-related protein were comparable to those induced by the DNMT3b silencing vector. Moreover, inhibition of DNMT3b attenuated STAT3 activation, as demonstrated in Western blot analyses and immunofluorescence studies. Therefore, it is likely that constitutional activation of STAT3 plays an important role in DNMT3b transmitting to downstream targets that regulate cell growth, EMT, and invasiveness. Furthermore, IHC data obtained from clinical samples demonstrated that DNMT3b expression was correlated significantly with p-STAT3 staining. The regulation of STAT3 signaling activity and crosstalk activation in cancer and nonmalignant cell lines is complex and cell context-dependent. SOCS3 acts as an important regulator of the STAT3 pathway.24, 35 It is generally accepted that SOCS3 acts as a critical negative-feedback regulator of the JAK-STAT signaling pathway. Methylation of SOCS3 has been implicated in head and neck SCC and lung cancer associated with abnormal activation of STAT3.38, 39 Therefore, it is proposed that the status of SOCS3 is important in terms of inhibited STAT3 signaling induced by the DNMT3b silencing vector. To clarify the effects of DNMT3b on SOCS3, we investigated the level of SOCS3 in CE81T transfectants. The data indicated that the DNMT3b silencing vector augmented SOCS3 expression associated with inhibited STAT3 signaling. Therefore, the inhibition of STAT3 activation caused by the DNMT3b silencing vector may result from the up-regulation of SOCS3, although further research is required to confirm this.

Radiotherapy and chemotherapy are crucial for the treatment of esophageal SCC. This study has demonstrated that the level of DNMT3b is related to the aggressiveness of esophageal cancer. Furthermore, enhanced expression of DNMT3b was associated significantly with a lower response rate after treatment, demonstrating a role in predicting clinical outcomes. To clarify the role of DNMT3b in the resistance of esophageal cancer to treatment, cell death after irradiation or cisplatin treatment was analyzed in the presence or absence of the DNMT3b silencing vector. Data obtained from the viable cell number assay and clonogenic assays demonstrated that inhibition of DNMT3b sensitizes esophageal cancer cells to irradiation and cisplatin treatment. Activation of STAT3 signaling reportedly is a mechanism responsible for resistance to CCRT. Furthermore, extensive DNA damage caused by radiation or anticancer agents can result in cell death or sensitivity to clinical treatment if left unrepaired.40 Phosphorylated H2AX is an indicator of induced double-strand breaks, and its expression after irradiation correlates with treatment sensitivity.41 The current data demonstrate that inhibited DNMT3b is associated with enhanced radiation-induced DNA damage and attenuated STAT3 activation. The cell line data and clinical data suggest that the level of DNMT3b is important in terms of treatment response. We also examined the power of DNMT3b to predict clinical outcome. In our univariate and multivariate analyses, poor treatment response, positive staining of DNMT3b, and no tumor resection were linked significantly to shorter survival.

The results from this study suggest that overexpression of DNMT3b is responsible for more aggressive tumor growth and resistance to treatment in esophageal SCC and is linked to activated STAT3 signaling. These data support the emerging notion that DNMT3b and p-STAT3 are clinically significant predictors, and DNMT3b may represent a suitable target for esophageal SCC treatment. Therefore, we suggest that targeting DNMT3b may be a potential treatment strategy for patients with esophageal SCC.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

This study was supported by the National Science Council, Taiwan (grant 98-2314-B-182-038-MY2).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES