TWIST1 promoter methylation is associated with prognosis in tonsillar squamous cell carcinoma

Human Pathology (2013) 44, 1722–1729

www.elsevier.com/locate/humpath

Original contribution

TWIST1 promoter methylation is associated with prognosis in tonsillar squamous cell carcinoma☆ Mi Jung Kwon MD, PhD a,1 , Ji Hyun Kwon MD b,1 , Eun Sook Nam MD, PhD b , Hyung Sik Shin MD, PhD b , Dong Jin Lee MD, PhD d , Jin Hwan Kim MD, PhD d , Young Soo Rho MD, PhD c , Chang Ohk Sung MD e , Won Jae Lee MD f , Seong Jin Cho MD, PhD b,⁎ a

Department of Pathology, Hallym University Sacred Heart Hospital, Hallym University College of Medicine, Anyang, Gyeonggi-do, 431-070, Republic of Korea b Department of Pathology, Kangdong Sacred Heart Hospital, Hallym University College of Medicine, Gangdong-gu, Seoul 134-701, Republic of Korea c Department of Otorhinolaryngology–Head and Neck Surgery, Ilsong Memorial Institute Head and Neck Cancer, Kangdong Sacred Heart Hospital, Hallym University College of Medicine, Seoul, 134-701, Republic of Korea d Department of Otorhinolaryngology–Head and Neck Surgery, Kangnam Sacred Heart Hospital, Hallym University College of Medicine, Seoul 150-950, Republic of Korea e Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, 138-736, Republic of Korea f Department of Physical Medicine and Rehabilitation, Veterans Medical Center, Seoul, 134-791, Republic of Korea Received 21 January 2013; revised 26 February 2013; accepted 14 March 2013

Keywords: Tonsil; Squamous cell carcinoma; TWIST1; Promoter methylation

Summary Tonsillar squamous cell carcinomas (TSCC) frequently present with locally advanced diseases and cervical metastases, which are associated with poor prognoses. Epithelial-mesenchymal transition (EMT) is critical for tumor invasiveness and metastatic potential. Recent studies have shown that TWIST1inducing EMT is overexpressed and hypermethylated in several cancers, indicating disease progression. The aim of the present study was to determine the clinical and prognostic significance of TWIST1 hypermethylation and EMT-related protein expression in TSCC. Methylation levels of TWIST1 promoter were analyzed by quantitative real-time methylation-specific polymerase chain reaction. Immunohistochemical analyses of TWIST1, Snail, and SMAD nuclear interacting protein-1 (SNIP1) were performed in 65 formalin-fixed, paraffin-embedded blocks of surgically resected specimens. TWIST1 promoter hypermethylation was found in 27.7% (18/65) of TSCCs. TWIST1 promoter hypermethylation was associated with poor differentiation (P = .012). Contralateral cervical lymph node metastasis was more frequently observed in TWIST1-methylated tumors (P = .029). High protein expressions of TWIST1, Snail, and SNIP1 were observed in 14 TSCC specimens (21.5%), 21 TSCC specimens (32.3%), and 38 TSCC specimens (58.5%), respectively. SNIP1 expression correlated significantly with TWIST1 methylation (P = .001), whereas TWIST1 protein expression did not. Contralateral cervical lymph node metastasis was an independent risk factor of the decreased overall survival rate (P = .002). TWIST1 methylation (P = .031) and pN stage (P = .037) were independent factors of poor prognoses affecting

☆ Conflict of interest statement: The authors declare no conflict of interest. ⁎ Corresponding author. E-mail address: [email protected] (S. J. Cho). 1 These authors contributed equally to this work.

0046-8177/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.03.004

TWIST1 methylation in tonsillar cancers

1723

disease-free survival. TWIST1 promoter hypermethylation may be a useful molecular marker for predicting prognoses and contralateral cervical lymph node metastases in patients with TSCC. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Tonsillar squamous cell carcinomas (TSCCs) are the most common oropharyngeal cancer, accounting for approximately 70% to 80% of this region [1]. Most TSCCs are diagnosed as locally advanced diseases, and even early-stage TSCCs present with clinically apparent cervical metastases [1,2]. This highly invasive and metastatic potential is associated with the poor prognoses of TSCC [1]. Epithelial-mesenchymal transition (EMT) is critical for the development of invasiveness and metastatic potential in tumor progression [3]. EMT is a complex process during which cellular phenotype and function change into a migrating and invasive form [4]. This process is triggered by a set of diverse stimuli to growth factor signaling and hypoxia as well as by tumor– stromal cell interactions with transcription factors such as TWIST1, Snail, and SMAD nuclear interacting protein-1 (SNIP1) [5]. Recent studies have reported that the protein expressions of TWIST1, Snail, and SNIP1 are associated with tumor progression and poor prognosis in several types of cancers including head and neck cancers [6–8]. Genetic and epigenetic changes possibly governed by the transcription factor TWIST1 have been thought to accumulate in primary tumors over the years before cancer cells are mature to spread and become fully malignant [9]. Aberrant promoter methylation is often responsible for gene silencing in various malignant tumors, and several genes are epigenetically down-regulated in head and neck squamous cell carcinomas (SCCs) [10]. TWIST1 may be a predictor of metastasis in head and neck cancers [7,8,11]. However, TWIST1 methylation status and its prognostic significance in patients with TSCC remain to be determined. The aim of the present study was to investigate the TWIST1 promoter methylation and the protein expression of TWIST1, Snail, and SNIP1 in a cohort of 65 patients with TSCC to evaluate the clinicopathologic relevance of TWIST1 promoter methylation and the EMT-related protein expression and to assess their prognostic significance.

2. Materials and methods 2.1. Patients and histologic evaluation This study was conducted using the formalin-fixed, paraffin-embedded tissues obtained from 65 patients with primary TSCC who underwent surgery at Ilsong Memorial Institute of Head and Neck Cancer, Kangdong Sacred Heart Hospital between 1997 and 2010. Inclusion criteria were

primary resection, no prior treatment, complete medical records, and the availability of all histopathologic slides of the resected specimens. Clinical information including age, sex, treatment modality, and survival or recurrence was analyzed using medical records and radiologic study results. All the glass slides from 65 TSCCs were reviewed by 2 pathologists for diagnosis confirmation and selection of a representative section for immunohistochemical and molecular studies. The diagnosis and histologic differentiation were evaluated according to the World Health Organization classification [12], and the tumor staging was based on the American Joint Committee on Cancer TNM staging system [13]. Surgical treatment was followed by postoperative radiotherapy in 13 patients and by chemoradiation in 27 patients. Twenty-five patients were treated with surgery alone. All patients underwent neck dissection on at least 1 side. This study was approved by the Institutional Ethnics Committee of Kangdong Sacred Heart Hospital (Seoul, Korea).

2.2. Quantitative real-time methylation-specific polymerase chain reaction Genomic DNA was extracted from 10-μm-thick sections of 10% neutral formalin-fixed, paraffin-embedded tumor tissue blocks using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). Formalin-fixed, paraffin-embedded DNA was treated with sodium bisulfite using an EZ DNA methylation kit (Zymo Research, Orange, CA). Methylation-specific real-time polymerase chain reaction (PCR) assays were performed in the ABI 7900HT Fast Real-time PCR system (Applied Biosystems, Foster City, CA). Primer pairs used were as follows: TWIST1 forward, TTTCGGATGGGGTTGTTATC and reverse, AAACGACCTAACCCGAACG, with the TaqMan probe 6FAM-AACGACCCAAACACTCACCAAATCGC-BBQ; ACTB forward, TGGTGATGGAGGAGGTTTAGTAAGT and reverse, AACCAATAAAACCTACTCCTCCCTTAA, with the TaqMan probe 6FAM-ACCACCACCCAACACACAATAACAAACACA-BBQ. The housekeeping βactin gene (ACTB) was used to normalize a methylationindependent control reaction. For relative quantification, amounts of the methylated DNA (percentage of methylated reference) at a TWIST1 promoter region were normalized to the methylation value of the calibrator, which was defined as 100%. Universal methylated DNA (QIAGEN) was used as the calibrator. The percentage of methylated reference was defined as 100 × 2(sampleACTB(ct) − sampleTWIST1(ct))/ 2(calibratorACTB(ct) − calibratorTWIST1(ct)). To discriminate between individual methylation levels, a cutoff value of 4% or greater

1724 was defined as “methylated,” and a cutoff value of less than 4% as “unmethylated” based on a previous study [14].

2.3. Immunohistochemistry Formalin-fixed, paraffin-embedded tumor tissue was cut into 5 μm in thickness and was mounted on positively charged slides. These sections were deparaffinized with xylene and rehydrated through a gradient alcohol series. The sections were subsequently immersed in 10 mmol/L citrate buffer (pH 6.0) and heated in an autoclave at 121°C for 5 minutes to retrieve nonenzymatic antigen. A dual endogenous enzyme block solution (Dako Cytomation Inc, Carpinteria, CA) was applied to the tissue sections for 10 minutes to block the endogenous peroxidase activity. The primary antibodies were mouse monoclonal antibodies against TWIST1 (1:20, ab50887; Abcam Ltd, Cambridge, UK), and polyclonal antibodies against Snail (1:100, ab85931; Abcam Co Ltd) and SNIP1 (1:100, ab25837; Abcam Ltd,). The sections were incubated for 90 minutes at room temperature, followed by incubation with Dako Envision-labeled polymer (Dako Cytomation Inc) for 30 minutes. Stains were developed by reaction with diaminobenzidine chromogen for 5 to 10 minutes and then counterstained with hematoxylin for 5 minutes. Both the intensity of immunohistochemical staining and the proportion of stained tumor cells were evaluated as described previously [15]. The intensity of staining was scored as follows: 0, negative; 1+, weak; 2+, moderate; and 3+, strong. Staining proportion was rated according to the percentage of positive cells and expressed on a 5-point scale as follows: 0, less than 5%; 1, 5% to 25%; 2, 26% to 50%; 3, 50% to 75%; and 4, more than 75%. The sum score of 4 or higher was classified as a high expression, and the score of lower than 4 as a low expression. Two pathologists blinded to patients' clinical data interpreted all immunostained slides, and cases with discrepant scores were reevaluated to achieve a consensus score.

M. J. Kwon et al. defined as the interval from the first day of surgery until death or the end of follow-up. Survival differences between individual groups were calculated using the Kaplan-Meier method with the log-rank test. We used the Cox proportional hazards model for the multivariate analysis of recurrencefree and overall survival rates. Recurrence-free and overall survival rates were analyzed up until January 2011. SPSS version 18 (SPSS Inc, Chicago, IL) and Stata/IC statistical software version 12 (StataCrop LT, College Station, TX) were used for all statistical analyses. A P value of less than .05 was considered statistically significant.

3. Results 3.1. Correlations between TWIST1 promoter methylation and clinicopathologic factors The promoter methylation levels of TWIST1 were quantitatively determined by real-time methylation-specific PCR in 65 TSCC specimens. The median methylation value was 0.8% (mean ± SD, 7.30% ± 21.26%; range, 0%-100%). TWIST1 hypermethylation was found in 18 (27.7%) of the 65 TSCC specimens, using a hypermethylation value of greater than 4% as a cutoff. Clinical and pathologic characteristics of patients with TSCC were analyzed according to their methylation status (Table 1). TWIST1 hypermethylation was more frequently observed in poorly differentiated tumors than in well- or moderately differentiated tumors (P = .012). Contralateral cervical lymph node metastasis was more frequently found in TWIST1-methylated tumors (P = .029). No significant differences were seen in TWIST1 hypermethylation according to sex, age, tumor location, pT stage, ipsilateral cervical lymph node metastatic status, and pTNM stage (P = .612, P = .264, P = .965, P = .746, P = .564, and P = .859, respectively).

2.4. Statistical analysis

3.2. Protein expressions of TWIST1, Snail and SNIP1 and the correlations with TWIST1 promoter methylation and clinicopathologic factors

The χ2 test was used to examine possible associations of TWIST1 methylation status with age, tumor location, T stage, tumor differentiation, contralateral cervical lymph node status, and TNM stage. The associations between the TWIST1 methylation status and sex/ipsilateral cervical lymph node status were determined using 2-tailed Fisher exact test. Correlation analyses between EMT-related marker expressions and clinical/pathologic variables were performed using 2-tailed Fisher exact test. The associations between EMT-related marker expressions and TWIST1 promoter methylation values were examined using the Mann-Whitney U test. Disease-free survival was defined as the interval from the first day of surgery until tumor progression or the end of follow-up. Overall survival was

High expressions of TWIST1, Snail, and SNIP1 were observed in 14 (21.5%), 21 (32.3%), and 38 (58.5%) TSCCs, respectively. TWIST1 expression was frequently observed in stromal cells than epithelial cells, whereas Snail and SNIP1 expressions showed both stromal and epithelial cells (Fig. 1). We compared TWIST1 methylation values according to the high or low expression levels of TWIST1, Snail, and SNIP1. TWIST1 methylation values correlated with low SNIP1 expressions (P = .001; Fig. 2). However, there were no significant correlations between TWIST1 methylation values and high expressions of TWIST1 or Snail. Associations of clinicopathologic factors with expressions of TWIST1, Snail, and SNIP1 are shown in Table 2. There were no significant differences in immunohistochemical

TWIST1 methylation in tonsillar cancers

1725

Table 1

Association between TWIST1 methylation and clinicopathologic features

Variable

Total (n = 65), n (%)

TWIST1 methylation status Methylated

Unmethylated

57 (87.7) 8 (12.3)

16 (88.9) 2 (11.1)

41 (87.2) 6 (12.8)

43 (66.2) 22 (33.8)

10 (55.6) 8 (44.4)

33 (70.2) 14 (29.8)

40 (61.5) 25 (38.5)

11 (61.1) 7 (38.9)

29 (61.7) 18 (38.3)

59 (90.8) 6 (9.2)

16 (88.9) 2 (11.1)

43 (91.5) 4 (8.5)

41 (63.1) 24 (36.9)

7 (38.9) 11 (61.1)

34 (72.3) 13 (27.7)

53 (81.5) 12 (18.5)

15 (83.3) 3 (16.7)

38 (80.9) 9 (19.1)

11 (16.9) 54 (83.1)

6 (33.3) 12 (66.7)

5 (10.6) 42 (89.4)

10 (15.4) 55 (84.6)

3 (16.7) 15 (83.3)

7 (14.9) 40 (85.1)

Sex Male Female Age (y) ≤60 N60 Tumor location Right side Left side T classification pT1-3 pT4 Tumor differentiation Well/Moderate Poorly Ipsilateral neck LN Metastasis No metastasis Contralateral neck LN Metastasis No metastasis pTNM stage I-II III-IV

P

.612

.264

.965

.746 .012 ⁎

.564 .029 ⁎

.859

Abbreviation: LN, lymph node. ⁎ Statistically significant, P b .05.

expressions of TWIST1, Snail, and SNIP1 according to sex, age, tumor location, pT stage, tumor differentiation, ipsilateral or contralateral cervical lymph node metastatic status, and pTNM stage.

3.3. Correlations between TWIST1 methylation and clinical outcomes Kaplan-Meier survival curves were constructed to assess the prognostic significances of the TWIST1 methylation status, pT stage, tumor differentiation, and contralateral

cervical lymph node metastatic status. The overall survival time was shorter in patients with the TWIST1-methylated tumors than in those with the TWIST1-unmethylated tumors (mean, 32 months versus 71 months; P = .039) (Fig. 3A). In addition, the pT4 stage and poor differentiation were associated with worse overall survival than the pT1 to pT1-3 stages and well/moderate differentiation (P = .017 and P = .045, respectively). The overall survival time was significantly shorter in patients with the contralateral cervical lymph node metastasis than that in those without (mean, 23 months versus 72 months; P = .001).

Fig. 1 Representative images of high expressions of TWIST1 (A), Snail (B), and SNIP1 (C) in TSCC specimens. TWIST1 expression was more frequently observed in stromal than in epithelial cells, whereas expressions of Snail and SNIP1 were observed equally in both stromal and epithelial cells. Magnification ×200.

1726

M. J. Kwon et al.

Fig. 2 Correlation of TWIST1 promoter methylation values with protein expressions of TWIST1, Snail, and SNIP1 in TSCC. TWIST1 promoter methylation values did not correlate with TWIST1 (A) or Snail (B) expression. C, However, TWIST1 promoter methylation values correlated with negative SNIP1 expression.

Table 2

Associations of clinicopathologic features with immunohistochemical expressions of TWIST1, Snail, and SNIP1

Variable

Sex Male Female Age (y) ≤60 N60 Tumor location Right side Left side T classification pT1-3 pT4 Differentiation Well/Moderate Poorly Ipsilateral LN Metastasis No metastasis Contralateral LN Metastasis No metastasis pTNM stage I-II III-IV TWIST1 status Methylation Unmethylation

TWIST1

P

High

Low

12 2

45 6

8 6

35 16

7 7

33 18

11 3

48 3

8 6

33 18

11 3

42 9

2 12

9 42

3 11

7 44

Snail

P

High

Low

17 4

40 4

15 6

28 16

14 7

26 18

19 2

40 4

15 6

26 18

19 2

34 10

4 17

7 37

3 18

7 37

.799

Low

33 5

24 3

28 10

15 12

21 17

19 8

34 4

25 2

25 13

16 11

31 7

22 5

6 32

5 22

6 32

4 23

5 33

13 14

.805

.128

.557

.075

.302

.955

.603

.511

.416

.747

.591

.309

.562

.992

.752

.479

.772

.590

.316

Abbreviation: LN, lymph node. ⁎ Statistically significant, P b .05.

High

.587

.316

16 35

P

.253

.421

2 12

SNIP1

.915

.095 3 18

15 29

.002 ⁎

TWIST1 methylation in tonsillar cancers

1727

Fig. 3 Kaplan-Meier curves for overall survival (A) and disease-free survival (B) depending on TWIST1 methylation status. Patients who had TSCCs with TWIST1 methylation showed lower overall and disease-free survival rates than did those with TWIST1 unmethylation.

TWIST1 methylation (Fig. 3B), pN1-3 stage, and the presence of ipsilateral and contralateral cervical lymph node metastasis were associated with shorter disease-free survival than TWIST1 unmethylation (P = .027), pN0 stage (P = .028), the absence of ipsilateral (P = .044) and contralateral (P = .009) cervical lymph node metastasis. The aforementioned clinicopathologic risk factors and TWIST1 methylation status that correlated significantly with overall or disease-free survival rates by univariate analysis were further analyzed using a Cox regression. As shown in Table 3, contralateral cervical lymph node metastasis was found to be an independent risk factor of death (P = .002; hazard ratio, 3.377; 95% confidence interval [95% CI], 1.553-7.344). TWIST1 methylation (P = .031; hazard ratio, 2.182; 95% CI, 1.074-4.434) and pN stage (P = .037; hazard ratio, 3.060; 95% CI, 1.071-8.746) were confirmed to be independent prognostic factors for disease-free survival.

4. Discussion In the present study, we first investigated TWIST1 promoter methylation in patients with TSCC as a predictor of clinical outcomes. TWIST1 promoter methylation has scarcely been investigated in malignant tumors other than breast and colorectal cancers that exhibit TWIST1 promoter methylation at frequencies of 16% to 77% and 52% to 56%, Table 3

respectively [16–19]. In our study, TWIST1 hypermethylation was observed in 27.7% of the TSCCs. The TWIST1 hypermethylation was strongly associated with tumor poor differentiation and the presence of lymph node metastasis in the contralateral neck. The correlation between TWIST1 methylation and lymph node metastasis has also been reported in breast cancer; hypermethylation of TWIST1 is more frequently seen in lymph node–positive breast tumors [20]. TWIST1 has several properties that facilitate tumor progression, including the triggering of EMT, the inhibition of apoptosis, and the enhancement of angiogenesis [21]. TWIST1 has been shown to induce EMT through chromatin remodeling [22]. These biological properties are prerequisite for metastatic growth [22]. TWIST1 methylation may be involved in the pathway through which development, progression, and metastasis of TSCC are regulated. Okada et al [17] have demonstrated that TWIST1 methylation levels are higher in colorectal adenomas and cancers than in normal colorectal tissues. They suggested that TWIST1 methylation may be useful for screening colorectal tumors. Our results also suggested that TWIST1 methylation may represent the histopathologic features and contralateral cervical lymph node metastasis of aggressive tumors. Previous studies have concluded that there is no significant correlation between TWIST1 methylation and TWIST1 protein expression [17,18,23]. Although DNA methylation within the promoter region of a gene can frequently inhibit or down-regulate gene transcription [10],

Multivariate analyses of overall and disease-free survival rates in patients with tonsillar cancer Overall survival

Contralateral LN metastasis TWIST1 methylation pN(+) vs pN0

HR

95% CI

3.377 1.421 2.211

1.553-7.344 0.598-3.375 0.760-6.435

Abbreviations: HR, hazard ratio; LN, lymph node; pN(+), pN1-3. ⁎ Statistically significant, P b .05.

P

.002* .426 .145

Disease-free survival HR

95% CI

1.716 2.182 3.060

0.694-4.247 1.074-4.434 1.071-8.746

P

.243 .031 ⁎ .037 ⁎

1728 not all promoter gene methylation processes are associated with silenced or decreased gene expression [24]. In our study, TWIST1 promoter hypermethylation did not result in the loss of TWIST1 protein expression in tumor samples with methylation, which is consistent with the result of a previous study [17]. Interestingly, TWIST1 methylation exhibited an inverse correlation with SNIP1 protein expression, but not with that of Snail or TWIST1. This implies that TWIST1 promoter hypermethylation may not be involved in TWIST1 gene silencing in TSCC and, rather, may be influenced by alternative compensatory molecular pathways. Okada et al [17] explained the discordance between TWIST1 protein expression and methylation in 2 aspects. One plausible explanation is that HIF-1 regulates the TWIST expression by binding directly to the hypoxia-response element in the TWIST proximal promoter, which is a common feature in solid cancers [25,26]. The other explanation is that TWIST1 promoter methylation is an early event that precedes compensatory TWIST1 overexpression [18]. Notably, TWIST1 promoter methylation is associated with the occurrence of promoter region methylation of other genes [27]. That some gene methylation occurs secondary to the upstream effect may partly explain the occurrence of gene hypermethylation in the absence of changes in gene expression [23]. In the present study, TWIST1 expression was more frequently observed in stromal than in epithelial cells, whereas expressions of Snail and SNIP1 were observed equally in both stromal and epithelial cells. Protein expressions did not correlate with clinical or pathologic parameters. Jouppila-Matto et al [11] have shown similar results in patients with pharyngeal SCC. They described that the stromal expression of TWIST1 is more frequently observed in hypopharyngeal than in oropharyngeal tumors [11]. In contrast, strong and abundant TWIST1 expression is described in patients with cervical and esophageal SCCs [28,29]. EMT regulators are likely to show various expressions and to play an essential role in the pathophysiology of several carcinomas [11]. Genes undergoing methylation during malignant progression are potentially prognostic markers [30]. Furthermore, gene hypermethylation in the absence of changes in expression may be useful as a biomarker if it consistently predicts clinical outcomes such as recurrence, metastasis, and survival [23]. Because few studies have investigated the relationships between TWIST1 methylation status and the prognosis of tumors, the prognostic impact of TWIST1 hypermethylation on survival is controversial. Cho et al [14] have reported that TWIST1 promoter hypermethylation in breast cancer is associated with cancer-specific mortality rates. On the other hand, Okada et al [17] and Ruppenthal et al [16] have demonstrated that TWIST1 hypermethylation is not related to survival of cancer patients. In our study, TWIST1 hypermethylation was an independent prognostic factor predicting the low overall survival rate in patients with TSCC. Mechanisms by which TWIST1 methylation exerts effects on survival have not yet been elucidated. TWIST1 hypermethylation in patients with

M. J. Kwon et al. TSCC appears to be related to metastatic potential and ultimately influences patients' clinical outcomes. Recently, human papillomavirus (HPV)–related oropharyngeal carcinoma has become an important subtype of head and neck SCC. HPV-associated oropharyngeal SCCs show a better prognosis than do HPV-negative tumors and may better respond to radiochemotherapy [10,31]. The absence of data regarding the prevalence of HPV within our cohort of patients with TSCC represents a limitation of our study. However, key messages emerged from our results. TWIST1 methylation may indicate the histopathologic features of aggressive tumors and the presence of contralateral cervical lymph node metastases, which treatment protocols must take into consideration. It is worthwhile to mention that TWIST1 methylation analysis can detect contralateral cervical lymph node metastases at an early stage of TSCC. TWIST1 methylation appears to be a predictor of poor clinical outcomes in patients with TSCC. TWIST1, a key EMT molecule, may constitute the primary epigenetic target for the treatment for patients with TSCC.

References [1] Cohan DM, Popat S, Kaplan SE, Rigual N, Loree T, Hicks Jr WL. Oropharyngeal cancer: current understanding and management. Curr Opin Otolaryngol Head Neck Surg 2009;17:88-94. [2] Chung EJ, Oh JI, Choi KY, et al. Pattern of cervical lymph node metastasis in tonsil cancer: predictive factor analysis of contralateral and retropharyngeal lymph node metastasis. Oral Oncol 2011;47: 758-62. [3] Yang J, Weinberg RA. Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis. Dev Cell 2008;14: 818-29. [4] Thiery JP, Sleeman JP. Complex networks orchestrate epithelialmesenchymal transitions. Nat Rev Mol Cell Biol 2006;7:131-42. [5] Polyak K, Weinberg RA. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat Rev Cancer 2009;9:265-73. [6] Yokoyama K, Kamata N, Hayashi E, et al. Reverse correlation of Ecadherin and snail expression in oral squamous cell carcinoma cells in vitro. Oral Oncol 2001;37:65-71. [7] Liang X, Zheng M, Jiang J, Zhu G, Yang J, Tang Y. Hypoxiainducible factor-1 alpha, in association with TWIST2 and SNIP1, is a critical prognostic factor in patients with tongue squamous cell carcinoma. Oral Oncol 2011;47:92-7. [8] Kojc N, Zidar N, Gale N, et al. Transcription factors Snail, Slug, Twist, and SIP1 in spindle cell carcinoma of the head and neck. Virchows Arch 2009;454:549-55. [9] Klein CA. Cancer. The metastasis cascade. Science 2008;321:1785-7. [10] Marsit CJ, McClean MD, Furniss CS, Kelsey KT. Epigenetic inactivation of the SFRP genes is associated with drinking, smoking and HPV in head and neck squamous cell carcinoma. Int J Cancer 2006;119:1761-6. [11] Jouppila-Matto A, Narkio-Makela M, Soini Y, et al. Twist and snai1 expression in pharyngeal squamous cell carcinoma stroma is related to cancer progression. BMC Cancer 2011;11:350. [12] Slootweg PJ, Eveson JW. Tumours of the oral cavity and oropharynx. In: Barnes L, Eveson JW, Reichart P, et al, editors. World Health Organization Classification of Tumours Pathology & Genetics Head and Neck Tumours. Lyon, France: IARC Press; 2005. p. 166-75.

TWIST1 methylation in tonsillar cancers [13] Fleming ID, Cooper JS, Henson DE, et al. AJCC cancer staging manual. Philadelphia: Lippincott-Raven; 2002. [14] Cho YH, Shen J, Gammon MD, et al. Prognostic significance of genespecific promoter hypermethylation in breast cancer patients. Breast Cancer Res Treat 2012;131:197-205. [15] Kim SH, Lee SH, Choi YL, Wang LH, Park CK, Shin YK. Extensive alteration in the expression profiles of TGFB pathway signaling components and TP53 is observed along the gastric dysplasiacarcinoma sequence. Histol Histopathol 2008;23:1439-52. [16] Ruppenthal RD, Nicolini C, Filho AF, et al. TWIST1 promoter methylation in primary colorectal carcinoma. Pathol Oncol Res 2011; 17:867-72. [17] Okada T, Suehiro Y, Ueno K, et al. TWIST1 hypermethylation is observed frequently in colorectal tumors and its overexpression is associated with unfavorable outcomes in patients with colorectal cancer. Genes Chromosomes Cancer 2010;49:452-62. [18] Gort EH, Suijkerbuijk KP, Roothaan SM, et al. Methylation of the TWIST1 promoter, TWIST1 mRNA levels, and immunohistochemical expression of TWIST1 in breast cancer. Cancer Epidemiol Biomarkers Prev 2008;17:3325-30. [19] Lee JS, Fackler MJ, Lee JH, et al. Basal-like breast cancer displays distinct patterns of promoter methylation. Cancer Biol Ther 2010;9: 1017-24. [20] Van der Auwera I, Bovie C, Svensson C, et al. Quantitative assessment of DNA hypermethylation in the inflammatory and non-inflammatory breast cancer phenotypes. Cancer Biol Ther 2009;8:2252-9. [21] Maestro R, Dei Tos AP, Hamamori Y, et al. Twist is a potential oncogene that inhibits apoptosis. Genes Dev 1999;13:2207-17. [22] Mironchik Y, Winnard Jr PT, Vesuna F, et al. Twist overexpression induces in vivo angiogenesis and correlates with chromosomal instability in breast cancer. Cancer Res 2005;65:10801-9.

1729 [23] Swift-Scanlan T, Vang R, Blackford A, et al. Methylated genes in breast cancer: associations with clinical and histopathological features in a familial breast cancer cohort. Cancer Biol Ther 2011;11:853-65. [24] Morey Kinney SR, Smiraglia DJ, James SR, Moser MT, Foster BA, Karpf AR. Stage-specific alterations of DNA methyltransferase expression, DNA hypermethylation, and DNA hypomethylation during prostate cancer progression in the transgenic adenocarcinoma of mouse prostate model. Mol Cancer Res 2008;6:1365-74. [25] Gort EH, van Haaften G, Verlaan I, et al. The TWIST1 oncogene is a direct target of hypoxia-inducible factor-2alpha. Oncogene 2008;27: 1501-10. [26] Yang MH, Wu MZ, Chiou SH, et al. Direct regulation of TWIST by HIF-1alpha promotes metastasis. Nat Cell Biol 2008;10:295-305. [27] Sunami E, Shinozaki M, Sim MS, et al. Estrogen receptor and HER2/neu status affect epigenetic differences of tumor-related genes in primary breast tumors. Breast Cancer Res 2008;10:R46. [28] Sasaki K, Natsugoe S, Ishigami S, et al. Significance of Twist expression and its association with E-cadherin in esophageal squamous cell carcinoma. J Exp Clin Cancer Res 2009;28:158. [29] Shibata K, Kajiyama H, Ino K, et al. Twist expression in patients with cervical cancer is associated with poor disease outcome. Ann Oncol 2008;19:81-5. [30] Shen L, Catalano PJ, Benson III AB, O'Dwyer P, Hamilton SR, Issa JP. Association between DNA methylation and shortened survival in patients with advanced colorectal cancer treated with 5-fluorouracil based chemotherapy. Clin Cancer Res 2007;13:6093-8. [31] Semrau R, Duerbaum H, Temming S, et al. Prognostic impact of human papillomavirus status, survivin, and epidermal growth-factor receptor expression on survival in patients treated with radiochemotherapy for very advanced nonresectable oropharyngeal cancer. Head Neck 2012 [Epub ahead of print]. http://dx.doi.org/10.1016/hed.23126.