This article is mentioned in:

Introduction

Given that pancreatic cancer is the fourth most common cause of cancer-related mortality and has the lowest patient survival rate of any solid cancer type (1,2), early diagnosis and therapy remain a major challenge.

Pancreatic cancer has been shown to be epigenetic, since it is a genetic disease characterized by widespread and profound alterations in DNA methylation. TWIST1 is a highly conserved transcription characteristic that belongs to the family of basic helixloop-helix proteins and is involved in embryonic development through the regulation of the migration-invasion program [termed the epithelial-mesenchymal transition (EMT)] during neural crest migration, while regulating mesodermal determination, myogenesis and morphogenesis (3–5). Although growing evidence demonstrates that TWIST1 methylation is a notable tumor biomarker in various tumors (6–9), little is known concerning the clinical significance of TWIST1 methylation in human primary pancreatic cancer. These correlations were investigated in the present study. Results showed that TWIST1 is methylated more frequently in pancreatic cancer compared to non-neoplastic pancreatic tissue, providing a new diagnostic marker for pancreatic cancer.

Materials and methods

Materials

Formalin-fixed, paraffin-embedded primary pancreatic cancer tissues and corresponding non-neoplastic pancreatic tissues from 33 patients who underwent surgical resection between 2004 and 2009 were evaluated. DNA was prepared from cells in microdissected, 5-μm histopathological sections, as described previously (9). Clinicopathological characteristics were available for the patients. The mean age of the patients (17 males) was 64.4 years. Eighteen tumors were located in the pancreatic head, 10 in the pancreatic body and 3 in the pancreatic tail. One tumor was stage I, 4 were stage II, 15 were stage III, 7 were stage IVa and 6 were stage IVb (classified according to the classification of pancreatic carcinoma of the Japan Pancreas Society). The study protocol was approved by the Institutional Review Board of Yamaguchi University Graduate School of Medicine. Informed consent was obtained from each patient.

Sodium bisulfite modification of DNA

Bisulfite treatment was performed as reported previously (9). Two micrograms of genomic DNA in 50 μl of water were denatured with 5.5 μl of 2 M NaOH at 37°C for 10 min, followed by incubation with 30 μl of 10 mM hydroquinone and 520 μl of 3 M sodium bisulfite (pH 5.0) at 50°C for 16 h in the dark. DNA was then purified with 50 μl of water and a DNA Cleanup kit (Promega Corporation, Madison, WI, USA), according to the manufacturer’s instructions, incubated with 5.5 μl of 3 M NaOH at room temperature for 5 min, precipitated with 1 μl of 20 mg/ml glycogen, 33 μl of 10 M ammonium acetate and 260 μl of 100% ethanol, washed with 70% ethanol and finally re-suspended in distilled water. DNAs used as positive controls for methylated and unmethylated alleles were SssI methyltransferase-treated placental DNA (New England Biolabs, Ipswich, MA, USA) and lymphocyte DNA, respectively.

KRAS mutations

DNA sequencing was used to evaluate mutations in exon 2 of KRAS, as described previously (10).

Methylation assay

The TWIST1 Combined Bisulfite Restriction Analysis (COBRA) primers were F: 5′-TGTGTA GAAGTTGTTGTTATT-3′ and R: 5′-CRAAAAAAACTAT CCTAAC-3′ (9). PCR amplification was performed for a total of 40 cycles with an annealing temperature of 55°C. The PCR product was digested with BstUI (New England Biolabs). The digested PCR products were separated by electrophoresis on 4% agarose gels. Digested fragments, which represent methylated DNA, were quantified by densitometry.

Results

The TWIST1 methylation lever was higher in pancreatic cancer compared to corresponding non-neoplastic pancreatic tissue (Fig. 1). The mean TWIST1 methylation was 66.7% for pancreatic cancer tissue and 15.0% for corresponding non-neoplastic pancreatic tissue (P= 0.0004). After setting the cut-off point at 15.0%, which was the mean level of TWIST1 methylation in nonneoplastic pancreatic tissue, the correlations between TWIST1 methylation status and clinicopathological characteristics were studied. No correlations were detected between TWIST1 methylation status in cancer tissue and clinicopathological characteristics (Table I).

Figure 1. Distribution of TWIST1 methylation in primary pancreatic cancer and non-neoplastic tissues is shown. Each sample is indicated by an open circle. The horizontal lines are the mean level in each group.

Table I. Correlations between TWIST1 methylation status and clinicopathological findings in pancreatic cancer patients.

Table I.

Correlations between TWIST1 methylation status and clinicopathological findings in pancreatic cancer patients.

	TWIST1 methylation status
Characteristics	Methylated	Unmethylated	P-valuea
Age (years)
>60	12	7	0.719
≤60	10	4
Gender
Male	13	4	0.282
Female	9	7
Histological typeb
Well	7	4	1.000
Mod-poor	15	7
TNM stagec
I, II, III	14	7	1.000
IVa, IVb	8	4
Perineural invasiond
+	19	8	0.375
−	3	3
Venous invasiond
+	16	7	0.696
−	6	4
Lymphatic invasiond
+	20	9	0.586
−	2	2
KRAS mutation
+	14	8	1.000
−	4	3

a Analyzed by the Fisher’s test.

b Well, well-differentiated adenocarcinoma; mod-poor, moderately or poorly differentiated adenocarcinoma.

c Classified based on the International Union Against Cancer tumor-node-metastasis classification.

d Classified based on the classification of pancreatic carcinoma of the Japan Pancreas Society.

Discussion

The TWIST1 methylation level was higher in pancreatic cancer compared to non-neoplastic pancreatic tissue. To the best of our knowledge, this is the first study on the distinct difference in TWIST1 methylation levels in healthy pancreatic tissues and pancreatic cancer, suggesting a potential for TWIST1 as a biomarker for the early detection of pancreatic cancers using pancreatic juice DNA-based assays.

Hypermethylation of DNA in promoter CpG islands results in the transcriptional silencing of cancer-related genes (11). In the present study, hypermethylation of TWIST1 was observed frequently in pancreatic cancer. However, upregulation of TWIST mRNA in pancreatic cancer has been reported (12). This discrepancy may be due to the lack of a direct correlation between TWIST1 methylation and TWIST1 expression in primary colorectal (9) and breast cancers (13). Characteristics other than TWIST1 methylation may also affect TWIST1 expression in pancreatic cancer. TWIST1 expression is regulated by hypoxia, a common characteristic in solid cancers, in an HIF-1α- and HIF-2α-dependent manner (14,15). Alternately, TWIST1 promoter methylation might be an early event, preceding compensatory TWIST1 overexpression (13).

In conclusion, increasing evidence has demonstrated that TWIST1 exhibits a unique characteristic as a tumor marker. Since TWIST1 methylation levels are higher in pancreatic cancers compared to corresponding healthy pancreatic tissues, TWIST1 methylation may be a feasible epigenetic marker for the detection of pancreatic cancer, using a pancreatic juice DNA test. Confirmatory studies using independent data sets are required to support these findings.

References