SEARCH

SEARCH BY CITATION

Keywords:

  • p16INK4a;
  • hypermethylation;
  • chronic inflammation;
  • glandular atrophy;
  • gastric cancer

Abstract

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The p16INK4a tumor suppressor gene can be inactivated by promoter region hypermethylation in many tumor types including gastric cancers. However, p16INK4a promoter hypermethylation in the surrounding non-tumorous tissues of gastric cancers has not been studied in detail. We therefore examined 46 gastric cancers, corresponding adjacent non-tumorous tissue samples and 8 gastric tissue samples of chronic gastritis by performing methylation-specific polymerase chain reaction, and we analyzed p16INK4a protein expression using immunohistochemistry and Western blot. p16INK4a promoter hypermethylation was observed in 43% of gastric cancers and 59% of adjacent non-tumorous tissues; however, none of the samples retrieved from the chronic gastritis patients displayed p16INK4a promoter hypermethylation. Gastric cancers showed an inverse correlation between vascular invasion and p16INK4a promoter hypermethylation, and adjacent non-tumorous tissues displayed a close association among the grade of chronic inflammation, presence of glandular atrophy and p16INK4a promoter hypermethylation. p16INK4a expression was markedly decreased in samples with p16INK4a promoter hypermethylation when compared with samples without p16INK4a promoter hypermethylation. These results suggest that p16INK4a promoter hypermethylation is an early and frequent event in gastric carcinogenesis and may serve as a new prognostic biomarker for the risk of gastric cancers. © 2001 Wiley-Liss, Inc.

The p16INK4a protein binds to cyclin-dependent kinase 4 (CDK4) and CDK6 and induces a G1-phase arrest in the molecular machinery of the cell cycle by interfering with binary cyclin D-CDK4 complexes.1 Loss of p16INK4a or its transcripts breaks down the regulatory mechanism of the cell cycle. Inactivation of p16INK4a is one of the most commonly observed abnormalities in human cancers.2 Similar to well-known mechanisms of inactivation of other tumor suppressor genes, it has been reported that functional loss of p16INK4a frequently occurs as a consequence of loss of heterozygozity (LOH) with somatic mutations of the remaining allele or homozygous deletions.3–5 However, in many primary tumors, p16INK4amutations or homozygous deletions are relatively rare,6–8 suggesting that p16INK4a is inactivated by an alternative mechanism. p16INK4a CpG island promoter hypermethylation correlated with transcriptional silencing, whereas treatment with the demethylating agent 5-deoxyazacytidine reactivated transcription.9–11

Gastric cancer remains a common disease world-wide, with a dismal prognosis. Primary gastric cancers revealed few intragenic alterations in p16INK4a,12–15 whereas recent studies have reported p16INK4a promoter hypermethylation.16–19 However, p16INK4a promoter hypermethylation in the surrounding non-tumorous tissues of gastric cancers has not been studied in detail. Therefore, we investigated p16INK4a promoter hypermethylation in gastric cancers and adjacent non-tumorous tissue and also analyzed hypermethylation in the gastric tissues of chronic gastritis.

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Patients, tissue samples and cell line

Forty-six gastric cancers, corresponding adjacent non-tumorous tissues and 8 gastric tissues of chronic gastritis were obtained from the Pathology Department of Kosin University Medical College, Pusan, and Dongguk University College of Medicine, Kyongju, South Korea. Gastric cancers and adjacent non-tumorous tissue as well as gastric tissue samples of chronic gastritis were obtained immediately after surgical resection and endoscopic procedures. Half of the non-tumorous tissues were fixed in 10% neutral buffered formalin for histological study and immunohistochemistry and the other half of the non-tumorous tissue and tumor samples were immediately frozen in liquid nitrogen and stored at −70°C until DNA and protein extraction. Each tissue was microdissected in cryostat to obtain appropriate tissue regions. Genomic DNA obtained from the SNU398 hepatoma cell line was a gift from Dr. S. I. Suh (Keimyung University, Daegu, South Korea).

Histopathological analysis

Pathological features were analyzed by a pathologist who was unaware of the results of methylation-specific PCR (MSP) and p16INK4a expression. Depth of tumor invasion, tumor differentiation, vascular invasion and nodal metastasis were examined in gastric cancers. We classified tumor differentiation by grade of glandular differentiation. Acute and chronic inflammation, intestinal metaplasia and glandular atrophy were graded in each adjacent non-tumorous tissue sample and the gastric tissue samples of chronic gastritis as suggested by the Updated Sydney System.20

DNA extraction and MSP

Using a DNA Mini kit (Qiagen, Hilden, Germany), genomic DNA was isolated from previously frozen tissues according to the manufacturer's instructions. DNA samples were modified by sodium bisulfite as reported previously.21 Modified DNA was subjected to MSP.21 Briefly, genomic DNA is modified by treatment with sodium bisulfite, which converts all unmethylated cytosines to uracils, while 5-methylcytosines remain unaltered. Primers are designed to amplify each of the sequences specifically based on these chemically induced differences. The primers were as follows: unmethylated p16INK4a (5′-TTA TTA GAG GGT GGG GTG GAT TGT-3′, 5′-CAA CCC CAA ACC ACA ACC ATA A-3′) and methylated p16INK4a (5′-TTA TTA GAG GGT GGG GCG GAT CGC-3′, 5′-GAC CCC GAA CCG CGA CCG TAA-3′). PCR was performed in a final volume of 50 μl using 100 ng of modified genomic DNA, 1.5 mM MgCl2, 20 pmol primer set, 0.2 mM deoxynucleotide triphosphate, PCR buffer and 1.25 U AmpliTaq Gold polymerase (Perkin-Elmer, Foster City, CA). After denaturation at 95°C for 7 min, DNA amplification was carried out in 30 cycles consisting of denaturation at 95°C for 45 sec, primer annealing at 65°C for 45 sec and elongation at 72°C for 1 min, followed by a final 7 min extension at 72°C. Each PCR reaction mixture was loaded onto 2% agarose gel, electrophoresed, stained with ethidium bromide and visualized under UV illumination. p16INK4a methylation status, methylated or unmethylated, was determined by noting the presence or absence of corresponding PCR products in an agarose gel, respectively.

Immunohistochemistry

Sections of 4 μm thickness were made and spread on poly-L-lysine-coated slides. Paraffin sections were immersed in three changes of xylene and hydrated using a graded series of alcohols. Antigen retrieval was performed routinely by immersing the sections in 0.01 M citrate buffer (pH 6.0) in a pressure cooker and autoclaving for 15 min. Sections were blocked with normal horse serum at 37°C for 30 min and then incubated with a primary antibody for 1 hr at room temperature. Primary antibody was monoclonal anti-mouse p16INK4a (clone F-12; Santa Cruz Biotechnology, Santa Cruz, CA) diluted at 1:100. Staining was achieved with a DAKO LSAB+kit (DAKO, Santa Babara, CA) and developed with 3, 3′-diaminobenzidine tetrahydrochloride (Zymed, San Francisco, CA). Sections were counterstained for 5 min with Meyer's hematoxylin and then mounted. Inflammatory cells and reactive stromal cells served as positive internal control. As negative control, the primary antibodies were omitted.

Quantitation of immunohistochemistry

Cells were considered positive when the clear staining in nucleus could be identified. Ten fields were randomly selected and the positive epithelial cells were counted. Labeling index (LI) was calculated as percentage values taking the total number of examined cells into account. At least 1,000 epithelial cells were examined throughout the entire tumor.

Western blot analysis

Frozen tissues were sonicated according to Sgambato et al.22 The supernatants were assayed for protein content by the Bradford method and stored at −70°C until used. Fifty micrograms of protein extracts were run in a NuPAGE 4-12% Bis-Tris polyacrylamide gel (NOVEX, San Diego, CA) and electro-blotted onto a nitrocelluose membrane (BioRad, Richmond, CA). The blots were then blocked overnight at 4°C and incubated for 2 hr at room temperature with monoclonal anti-mouse p16INK4a (clone F-12; Santa Cruz Biotechnology) diluted at 1:500 and anti-human low molecular weight cytokeratin (clone AE-1; Lab Vision, Fremont, CA) diluted at 1:1,000. After washing with TBS-0.05% Tween 20, the blots were incubated for 45 min at room temperature with the secondary antibody (DAKO) at 1:1,000 dilution, washed again and developed with Western blotting chemiluminescence luminol reagent (Santa Cruz Biotechnology) for X-ray film examination. Low molecular weight cytokeratin was used for control of the epithelial component of extracts.

Statistical analysis

The significance of differences between methylated and unmethylated groups was evaluated by Chi-square test, Fisher's exact test, and t-test. The significance of frequency of methylation status in gastric cancers, corresponding non-tumorous tissues and the gastric samples of chronic gastritis was examined by Chi-square test and Fisher's exact test. Correct p < 0.05 was considered significant

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Methylation status of examined tissues and relation between the methylation status of gastric cancers and clinicopathological features

The SNU398 hepatoma cell line served as a positive control for p16INK4a promoter hypermethylation. We confirmed that the SNU398 cell line showed a clear band when amplified with methylated-specific primers and did not display a band when treated with unmethylated primers (Fig. 1). p16INK4a promoter hypermethylation was observed in 43% (20/46) of gastric cancers, 59% (27/46) of adjacent non-tumorous tissues and 0% (0/8) of the chronic gastritis samples (Fig. 1). Fifteen samples exhibited p16INK4a promoter hypermethylation in carcinomas and adjacent tissues, 5 samples displayed hypermethylation in only carcinomas, 12 samples exhibited hypermethylation in only adjacent tissues, and 14 samples did not reveal hypermethylation in carcinomas and adjacent tissues.

thumbnail image

Figure 1. Methylation-specific PCR with methylated p16INK4a primer sets (a) and unmethylated p16INK4a primer sets (b), and Western blot analysis using p16INK4a (c) and low molecular weight cytokeratin (d). A PCR mixture omitting DNA template (Water) was used as negative control, and the SNU398 hepatoma cell line (Control) is completely methylated. Gastric cancers (T) and corresponding adjacent non-tumorous gastric tissue samples (N) are shown. Tumors in cases 18 and 38 exhibit p16INK4a promoter hypermethylation, as do adjacent non-tumorous tissue samples from cases 35 and 38, but the chronic gastritis (CG) samples do not. All the samples except for the SNU398 hepatoma cell line and negative control exhibit bands when amplified with unmethylated primers. p16INK4a expression is markedly decreased in samples with p16INK4a promoter hypermethylation (M) when compared with the p16INK4a promoter unmethylated group (U). Low molecular weight cytokeratin serves as an epithelial loading control.

Download figure to PowerPoint

Clinicopathological characteristics and the methylation status of p16INK4a promoter in 46 gastric cancers were compared and analyzed (Table I). The parameters depth and differentiation of tumor, nodal metastasis, age and sex did not correlate with p16INK4a promoter hypermethylation, but vascular invasion, interestingly, showed an inverse correlation. Thirteen of 19 samples without vascular invasion had p16INK4a promoter hypermethylation, and only 7 of 27 samples with vascular invasion showed hypermethylation (p < 0.05, Chi-square test). p16INK4a LI was 0.41 ± 0.15 in samples with p16INK4a promoter hypermethylation and 0.85 ± 0.08 in samples without hypermethylation (Fig. 2) (p < 0.05, t-test).

Table I. Clinicopathological Features of 46 Gastric Cancers According to the Methylation Status of p16INK4a Promoter
 p16INK4a unmethylated group (n = 26)p16INK4a methylated group (n = 20)p-value1
  • 1

    Statistical significance was determined using the Chi-square test, Fisher's exact test and t-test.

Age (yr)56.31 ± 6.8555.45 ± 10.64>0.05
Sex>0.05
 Male2016
 Female64
Tumor depth>0.05
 T111
 T222
 T31715
 T462
Differentiation>0.05
 Well54
 Moderate73
 Poor1413
Nodal metastasis0.30 ± 0.260.28 ± 0.23>0.05
Vascular invasion<0.05
 Absence613
 Presence207
thumbnail image

Figure 2. Immunohistochemical stain of p16INK4a in gastric cancers (a and b) and adjacent non-tumorous gastric tissue samples (c and d). Neutrophils and lymphoplasma cells show positive immunoreactivity for p16INK4a and serve as positive internal control (arrowhead). Samples with p16INK4a promoter hypermethylation (a and c) show decreased p16INK4a labeling index compared with samples without hypermethylation (b and d). Scale bar = 21.2 μm.

Download figure to PowerPoint

Relation between the methylation status of the adjacent non-tumorous gastric tissues of gastric cancers and clinicopathological features

Eight chronic gastritis samples showed no glandular atrophy and intestinal metaplasia and displayed mild chronic inflammatory cell infiltrate. Polymorphonuclear neutrophil activity was absent in two cases, mild in four cases and moderate in two cases. Helicobacter pylori density was absent in one case, mild in four cases and moderate to marked in three cases. The presence of glandular atrophy correlated with p16INK4a promoter hypermethylation, since 23 of 32 samples exhibiting glandular atrophy had p16INK4a promoter hypermethylation (Table II). In contrast, only 4 of 14 samples without glandular atrophy exhibited hypermethylation (p < 0.05, Chi-square test). Twenty-four of 36 samples showing moderate to severe chronic inflammation had p16INK4a promoter hypermethylation, and 3 of 10 samples with mild chronic inflammation exhibited hypermethylation (p < 0.05, Fisher's exact test). p16INK4a promoter hypermethylation was more frequently found in antral tissues (16 of 22 samples [73%]) than in fundic tissue (11 of 24 [46%]) (p = 0.06, Chi-square test). The parameters grade of acute inflammation and intestinal metaplasia, age and sex did not correlate with p16INK4a promoter hypermethylation (p < 0.05, Chi-square and t-tests). p16INK4a LI was 0.31 ± 0.18 in samples with p16INK4a promoter hypermethylation and 0.57 ± 0.19 in samples without hypermethylation (Fig. 2) (p < 0.05, t-test).

Table II. Clinicopathological Features of 46 Adjacent Non-Tumorous Gastric Tissues According to the Methylation Status of p16INK4a Promoter
 p16INK4a unmethylated group (n = 19)p16INK4a methylated group (n = 27)p-value1
  • 1

    Statistical significance was determined using the Chi-Square test, Fisher's exact test and t-test.

Age (yr)56.89 ± 8.5654.74 ± 8.47>0.05
Sex>0.05
 Male1323
 Female64
Acute inflammation>0.05
 Mild713
 Moderate to marked1214
Chronic inflammation<0.05
 Mild73
 Moderate to marked1224
Intestinal metaplasia>0.05
 Absence1414
 Presence513
Glandular atrophy<0.05
 Absence104
 Presence923
Helicobacter pylori>0.05
 Absence00
 Presence1927
Site>0.05
 Fundus1311
 Antrum616

Western blot

To confirm the specificity of the immunohistochemical results, we analyzed two gastric cancers with p16INK4a promoter hypermethylation and two cancers without hypermethylation. As shown in Figure 1, p16INK4a expression was markedly decreased in samples with p16INK4a promoter hypermethylation compared with the unmethylated group.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We found p16INK4a promoter hypermethylation in 20 of 46 gastric cancers (43%). This frequency is in accordance with previous studies.16, 18, 19 The fact that p16INK4a expression was lower in samples with p16INK4a promoter hypermethylation suggests that the functional inactivation of p16INK4a occurs through promoter hypermethylation. However, in our study, the complete loss of p16INK4a expression was not observed in samples showing p16INK4a promoter hypermethylation. This is, however, not so surprising since recent studies have revealed that p16INK4a expression is highly affected by the frequency, but not the existence itself, of methylation of the p16INK4a promoter in human bladder, colon and melanoma cell lines as well as hepatocellular carcinomas.23, 24 We found that all 46 gastric cancers exhibited bands for the unmethylated p16INK4agene, which could be caused by accompanying normal cells such as inflammatory cells, fibroblasts and vascular endothelial cells as well as by unmethylated regions of p16INK4a in cancer cells. The clinicopathologic survey of gastric cancers showed that methylated p16INK4a promoter cases correlated with the absence of vascular invasion but not with depth of tumor, differentiation and nodal metastasis. This result is quite interesting and might indicate a relatively good prognosis for gastric cancers exhibiting p16INK4a promoter hypermethylation. Larger number of samples would be required to describe the clinicopathologic features of methylated p16INK4a promoter cases.

Our study demonstrates for the first time, to our knowledge, that p16INK4a promoter hypermethylation is an early and probably important event in the development of gastric cancers. None of the samples retrieved from the chronic gastritis patients displayed p16INK4a promoter hypermethylation, whereas hypermethylation was observed in 59% of the non-tumorous tissue samples adjacent to gastric cancers. Moreover, the methylation of adjacent non-tumorous tissues correlated with the presence of glandular atrophy and the grade of chronic inflammation.

Klump et al.25 examined p16INK4a promoter hypermethylation in the progression of Barrett's esophagus and concluded that it emerged at a very early stage in the multistep process of neoplastic progression in Barrett's esophagus. Hsieh et al.26 suggested that p16INK4a promoter hypermethylation was a frequent and early occurring event during the process of neoplastic progression in ulcerative colitis. However, in gastric cancers, the methylation status of adjacent non-tumorous tissues and its relationship with clinicopathological features have not been studied in detail. Suzuki et al.16 found p16INK4a promoter hypermethylation in adjacent non-tumorous tissues of gastric cancers but did not investigate the frequency or detailed clinicopathologic features of adjacent tissue methylation.

As shown in our study, the close association among p16INK4a promoter hypermethylation, glandular atrophy and grade of chronic inflammation is a remarkable result. In Korea, H. pylori infection, a group I carcinogen of gastric cancer, begins in infancy; at 5 years of age, the infection rate is up to 50% and continues at about 80–90% after 8 years of age.27H. pylori infection damages gastric mucosa through bacterial toxicity and the effects of inflammation. Thereafter, gastric mucosa may either regenerate and return to normal or undergo an adaptive reparative process that leads to glandular atrophy and intestinal metaplasia and is known as precursor lesions of gastric cancer.28, 29 Chronic H. pylori infection could also disrupt the equilibrium of cellular turnover of gastric mucosa.30 The p16INK4a promoter hypermethylation observed in samples with glandular atrophy caused by chronic inflammation may be related to the disturbance of mucosal maturation and disequilibrium of cell kinetics in gastric mucosa. It will be important to determine whether p16INK4a promoter hypermethylation is involved in glandular atrophy. Each tissue was microdissected in a cryostat to obtain appropriate tissue regions. Therefore, we excluded the possibility that p16INK4a promoter hypermethylation of adjacent non-tumorous tissue may occur through contamination of cancer cells.

The 12 carcinomas in our study did not exhibit p16INK4a promoter hypermethylation, whereas their adjacent non-tumorous tissue samples did. This singular result seems to suggest that genetic lesions other than p16INK4a promoter hypermethylation are involved and possibly provide an even more decisive clonal growth advantage. However, methodologic shortcomings in the detection of hypermethylation could also be responsible for this finding. It has been shown that within any one tumor, detectable p16INK4a promoter hypermethylation is not necessarily a homogenous phenomenon.31 Thus this finding might be caused by the random selection of a small methylated area. Klump et al.25 have reported that some specimens classified as hypermethylated showed weak amplification product only when silver staining was performed, whereas staining with ethidium bromide did not detect an amplification product. Thus, the sensitivity of ethidium bromide may explain this result.

To evaluate the potential of methylation status for identifing the risk of gastric cancers, we examined p16INK4a promoter hypermethylation in chronic gastritis samples. Interestingly, such hypermethylation was not found, and thus we thought that p16INK4a promoter hypermethylation might constitute a new prognostic biomarker for the risk of gastric cancers. However, a prospective study and larger numbers of patients will be needed to evaluate the significance of this marker.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank Dr. Suh S. I. for the gift of the cell line and Shim Y.H. for many helpful discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES