Genetic instability in intestinal metaplasia is a frequent event leading to well-differentiated early adenocarcinoma of the stomach

European Journal of Cancer 36 (2000) 1113±1119

www.ejconline.com

Genetic instability in intestinal metaplasia is a frequent event leading to well-di€erentiated early adenocarcinoma of the stomach K. Kobayashi a,b,*, T. Okamoto a, S. Takayama a, M. Akiyama b,c, T. Ohno d, H. Yamada b a Department of Surgery II, Jikei University School of Medicine, 3-25-8 Nishi-Shinbashi, Minato-ku, Tokyo 105-8461, Japan Department of Molecular Genetics, Institute of DNA Medicine, Jikei University School of Medicine, 3-25-8 Nishi-Shinbashi, Minato-ku, Tokyo 105-8461, Japan c Department of Pediatrics, Jikei University School of Medicine, 3-25-8 Nishi-Shinbashi, Minato-ku, Tokyo 105-8461, Japan d Department of Oncology, Institute of DNA Medicine, Jikei University School of Medicine, 3-25-8 Nishi-Shinbashi, Minato-ku, Tokyo 105-8461, Japan b

Received 27 July 1999; received in revised form 21 December 1999; accepted 2 March 2000

Abstract To understand the development of well-di€erentiated adenocarcinoma in the stomach, we examined genetic instability in 31 patients with stage Ia gastric cancer. Triplets of tissue specimens (normal/metaplasia/tumour) from 33 lesions were examined for microsatellite instability (MSI) and loss of heterozygosity (LOH), using nine microsatellite loci. Frameshift mutations in the transforming growth factor b receptor type II (TGF- RII) (A)10, Bcl-2-associated X protein (BAX) (G)8, hMSH3 (A)8 and hMSH6 (C)8 genes were also studied. In this study, a high incidence of MSI (MSI-H) was de®ned as samples containing 30% or more MSI positive loci, and a low incidence of MSI (MSI-L) as samples which had less than 30% MSI. MSI-L was observed in 19 cancerous lesions (58%), and MSI-H in three (9%). Eleven intestinal metaplasia lesions (33%) showed MSI-L, but no metaplasia lesions exhibited MSI-H. Frameshift mutation was observed in only one cancerous lesion (3%) at the (A)10 tract of TGF- RII. In contrast, LOH was observed in 24 cancerous lesions (73%), and in 15 (45%) of intestinal metaplasia lesions. Intriguingly, these alterations tend to be coincident between metaplasia and cancerous lesions in the same sets of specimens, and there was no case that showed alterations in metaplasia, but not in cancerous lesions. These ®ndings suggest that metaplasia and well-di€erentiated adenocarcinoma in the stomach may have the same molecular backgrounds, and that these two lesions may be chronologically connected. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Microsatellite instability; Replication error; Loss of heterozygosity; Gastric cancer; TGF- RII; BAX; hMSH3; hMSH6

1. Introduction Gastric cancer is the second most common form of cancer in the world [1]. There are two distinctive histopathological subtypes, an intestinal or well-di€erentiated type, and a di€use or poorly di€erentiated type [2]. These two subtypes show quite di€erent clinical manifestations and histopathological characteristics. In particular, many epidemiological studies have shown that the well-di€erentiated type of gastric cancer originates from intestinal metaplasia [3]. Pathological studies * Corresponding author at present address: Mammalian DNA Repair Lab, Imperial Cancer Research Fund, Clare Hall Laboratories, Blanche Lane, South Mimms, Herts EN6 3LD. Tel.: +44-207-2693883; fax: +44-207-269-3801. E-mail address: [email protected] (K. Kobayashi).

also supported a close association between these conditions, because the cancers are always located adjacent to intestinal metaplasia [3,4]. However, there have been few studies of the pathological linkage from the perspective of molecular events. The development and evolution of malignant tumours is generally accelerated by accumulation of genetic mutations. The stepwise accumulation of somatic mutations is widely accepted as being involved in the process of colorectal carcinogenesis [5]. Well-di€erentiated gastric cancers also demonstrate similar genetic abnormalities [3]. Moreover, mutations of the p53 and adenomatous polyposis coli (APC) genes were observed in some intestinal metaplasia lesions [6,7]. These observations suggested that intestinal metaplasia is a precursor lesion of well-di€erentiated gastric adenocarcinoma.

0959-8049/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0959-8049(00)00066-6

1114

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119

Recently, it has been reported that the development of some gastric cancers is associated with genetic instability, which leads to multistep accumulation of somatic mutations [8±12]. Linkage analysis of hereditary nonpolyposis colorectal cancer (HNPCC) syndrome identi®ed four genes involved in DNA mismatch repair [13± 15]. Mutations of any of these genes result in a failure to correct errors that occur during DNA-replication (RER; replication error). Microsatellites, which consist of short repeated nucleotide sequences interspersed throughout the genome, are useful markers for detecting RER, because repetitive sequences are prone to nucleotide deletion or insertion during replication. In fact, microsatellite instability (MSI) has been shown to be informative for the investigation of carcinogenesis [16,17]. Many studies of advanced gastric cancers have indicated that MSI is associated with mutations in genes responsible for tumorigenesis. Genes encoding transforming growth factor b receptor type II (TGF- RII) and Bcl-2 associated X protein (BAX), which contain repeated mononucleotide sequences in coding exons, show a high prevalence of frameshift mutations in MSI (+) cancers [18±22]. Somatic or germ line mutations of hMSH3 and hMSH6 have also been reported to be associated with MSI (+) cancers [22±24]. These alterations in repetitive sequences may be useful to investigate the relationship between intestinal metaplasia and well-di€erentiated adenocarcinoma in the stomach. Therefore, we examined MSI, LOH and frameshift mutations in early well-di€erentiated gastric cancers and intestinal metaplasia lesions. Results obtained in this study suggest a strong connection between these two lesions. 2. Patients and methods 2.1. Patients 31 patients with early gastric carcinoma of the welldi€erentiated type were studied. All of these patients were treated surgically between 1994 and 1997 at the Department of Surgery II, Jikei University School of Medicine, Japan. They consisted of 26 men (84%) and 5 women (16%) aged from 45 to 79 years old at the time of surgery. None of the patients matched the Amsterdam criteria for HNPCC [25]. Other familial cancer syndromes were also negative based on their family histories. All tumours were reviewed pathologically and were classi®ed according to the criteria of the Japanese Research Society for Gastric Cancer [26]. Clinical staging based on the criteria of both Japanese Classi®cation of Gastric Carcinoma (2nd English ed.) and TNM classi®cation (5th ed.) revealed that all patients were stage Ia [27].

2.2. Tissue samples The gastric carcinomas, adjacent intestinal metaplasia and musclaris propria for corresponding normal controls were obtained from formaldehyde-®xed, paranembedded tissues. As 2 patients had two distinct cancerous lesions, a total of 33 normal/metaplasia/tumour triplet samples from the 31 patients were examined. For DNA extraction, one consecutive slice was ®rst stained with haematoxylin±eosin, and the margin of each lesion was carefully determined. After scraping each lesion, DNA was extracted according to the method of Goeltz and colleagues [28]. 2.3. Analysis of microsatellite instability (MSI) and loss of heterozygosity (LOH) Nine ¯uorescent labelled microsatellite primers, containing (CA)n repeats (D1S103, D2S123, D3S1067, D5S107, D13S153, D17S261, D18S34) were purchased from Research Genetics (Huntsville, AL, USA). Two (A)n repeats (BAT25, BAT40) were obtained from TaKaRa Shuzo Co, Ltd (Shiga, Japan) [29]. In particular, D2S123, D5S107 and D17S261 are thought to be associated with hMSH2, APC and p53, respectively [16,17]. The (A)n repeats of BAT25 and BAT40 are located in an intron of the c-kit gene and the 3-bhydroxysteroid dehydrogenase gene, respectively, and are known to be highly susceptible loci for MSI [29]. MSI was analysed using the above ¯uorescently labelled primers. Polymerase chain reaction (PCR) was carried out in 50 ml reaction mixtures containing 25±100 ng of template DNA, 10 pmol of each oligonucleotide primer, 5 units of Taq DNA polymerase (TaKaRa Shuzo Co, Ltd), 10 mM Tris±HCl (pH 8.3), 50 mM KCl and 15 mM MgCl2 and 0.2 mM each of deoxynucleotide triphosphate. PCR ampli®cation was performed for 40 cycles consisting of denaturation for 30 sec at 94 C, annealing for 1 min from 51 to 55 C, and extension for 1 min at 72 C, followed by a ®nal extension for 7 min at 72 C, using a GeneAmp PCR System 9600 thermal cycler (Perkin-Elmer Cetus, Norwalk, CT, USA). The ¯uorescent PCR products were studied by capillary electrophoresis (model 310 automated ¯uorescent DNA sequencer: Applied Biosystems, Foster City, CA, USA). GeneScan Analysis software (Applied Biosystems) was used for assessment of MSI and LOH. MSI was de®ned as positive when at least one variant allele from those of muscular DNA was detected in the tumour or metaplasia DNA [30,31]. LOH was de®ned as positive when the allelic ratio (AR=(T1/T2)/(N1/ N2)) was less than 0.7 as described by Goessl and colleagues [31]. Brie¯y, in the calculation of AR, T1 and N1 were de®ned as the highest peak areas of the shorter allele for the tumour and the control, respectively, and T2 and N2 were also de®ned as the peak areas of the

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119

longer allele in the same manner. In cases in which AR was greater than 1.0, we converted the ratio to obtain a result ranging from 0.00 to 1.00. 2.4. Mutation analysis of the mononucleotide tracts within cancer-related genes and mismatch repair genes The TGF- RII (A)10 tract, BAX (G)8 tract, hMSH3 (A)8 tract and hMSH6 (C)8 tract were also analysed by capillary electrophoresis as mentioned above. These repetitive mononucleotide tracts were ampli®ed using the following primer sets; 50 -CTACCATGACTTTATTCTGGA-30 (forward) and 50 -CAGGAACACATGAAGAAAGTC-30 (reverse) for TGF- RII, 50 -ATCCAGGATCGAGCAGGGCGA-30 (forward) and 50 -CAGTTCGTCCCCGATGCGCT-30 (reverse) for BAX, 50 AGATGTGAATCCCCTAATCAAGC-30 (forward) and 50 -ACTCCCACAATGCCAATAAAAAT-30 (reverse) for hMSH3 and 50 -GGGTGATGGTCCTATGTGTC-

1115

30 (forward) and 50 -CGTAATGCAAGGATGGCGT-30 (reverse) for hMSH6 [22]. 2.5. Statistical analysis Fisher's exact test was used to determine the statistical signi®cance. A level of P<0.05 was considered signi®cant. 3. Results 3.1. Microsatellite instability and loss of heterozygosity in intestinal metaplasia and well-di€erentiated adenocarcinoma of the stomach Since all samples were analysed at ®ve or more of the nine loci, we considered that all 33 sets of specimens were reliable for the assessment of MSI and LOH.

Fig. 1. Microsatellite instability and loss of heterozygosity detected in intestinal metaplasia and well-di€erentiated early gastric cancer. N, normal (muscularis propria); IM, intestinal metaplasia; T, tumour; MSI, microsatellite instability; LOH, loss of heterozygosity. Arrows indicate MSI or LOH. (a) Normal microsatellite assay in patient 15 (homozygous); (b) patient 29 with LOH of tumour DNA; (c) patient 4 with MSI of intestinal metaplasia DNA and tumour DNA; (d) and (e) patient 2. Loci assayed included BAT25 and BAT40. At the BAT25 locus, tumour DNA exhibited MSI. At the BAT40 locus, MSI was observed in tumour DNA, accompanied by LOH in both intestinal metaplasia DNA and tumour DNA.

1116

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119

Electrophoretic patterns are shown in Fig. 1. Triplet DNA samples from the same patients were electrophoresed simultaneously. As the ¯uorescence intensity was slightly di€erent in each sample, we calculated the AR for assessment of LOH. Fig. 1(a) shows the normal pattern, whilst the tumour shown in Fig. 1(b) was LOH positive. Fig. 1(c) shows MSI in both the intestinal metaplasia and the cancerous tissue. The peak length shifted from 207 bp for normal tissue to 203 bp for intestinal metaplasia and cancerous tissue. Fig. 1(d) shows MSI in cancerous tissue. In some cases, both LOH and MSI occurred simultaneously. Fig. 1(e) shows a case in which LOH ®rst occurred in intestinal metaplasia, and additional MSI occurred in another allele of the cancerous tissue. Table 1 summarises the occurrence of MSI and LOH in each lesion. Although the frequency of MSI-L was signi®cantly higher in cancer than in intestinal metaplasia (P=0.013), MSI-L was noted in several of the metaplasia lesions.

All MSI observed in intestinal metaplasia was also observed in the cancerous tissues. In our analysis of MSI, the numbers of loci that could be analysed di€ered markedly between samples. Therefore, we assessed the MSI status of each sample according to the positive percentage of markers. MSI-H was de®ned as samples containing 30% or more MSI-positive loci, and MSI-L was de®ned as less than 30% MSI [32]. Table 2 shows the results obtained according to the above criteria. No intestinal metaplasia exhibited MSI-H, whilst 11 (33%) showed MSI-L. In cancerous tissues, 19 of 33 tumour specimens (58%) were MSI-L and three (9%) exhibited MSI-H. One (patient 1) of the three MSI-H tumours had MSI-L intestinal metaplasia adjacent to it, but the other two MSI-H tumours did not have any MSI in their neighbouring intestinal metaplasia. LOH was observed in 24 of the tumour specimens (73%), and 15 of the intestinal metaplasia lesions (45%) (Table 2). More than half of the tumours with LOH

Table 1 Results of microsatellite assay in intestinal metaplasia and well-di€erentiated early gastric cancers Sample No.

1 2 3 4 5a 6a 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30a 31a 32 33

Age/Sex

58/m 56/m 75/m 51/m 49/m 49/m 70/f 56/m 61/m 48/m 59/m 74/m 81/f 81/m 52/m 79/m 58/m 75/m 71/m 45/m 57/m 51/m 45/m 73/f 67/m 78/m 61/m 56/m 71/m 78/f 78/f 67/m 81/m

Markers D1S103

D2S123

D3S1067

D5S107

D13S153

D17S261

D18S34

BAT25

BAT40

ÿ/ÿ ÿ/n ÿ/ÿ ÿ/ÿ ÿ/n ÿ/L ÿ/ÿ L/L ÿ/n n/n n/n ÿ/ÿ n/n ÿ/ÿ L/L ÿ/n ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n n/n n/n ÿ/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

n/n ÿ/ÿ ÿ/ÿ +/+ L/L ÿ/ÿ ÿ/L ÿ/ÿ L/L ÿ/n +/+ n/n +/+ n/n n/n L/+,L n/n n/n ÿ/L ÿ/+ ÿ/ÿ ÿ/+ L/L ÿ/ÿ L/L ÿ/ÿ L/L ÿ/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

ÿ/+ ÿ/+ ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/L L/L n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

L/+,L L/+,L ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/L L/L ÿ/ÿ ÿ/ÿ L/L ÿ/n ÿ/ÿ n/n ÿ/ÿ L/L L/L ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

n/n n/n ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ +/+ ÿ/ÿ ÿ/ÿ n/n n/n +/+ ÿ/ÿ n/n n/n ÿ/ÿ n/n n/n ÿ/ÿ ÿ/ÿ n/n n/n n/n L/L n/n n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/n ÿ/ÿ ÿ/ÿ n/n

ÿ/L n/+ ÿ/ÿ ÿ/L +/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/L L/L ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ L/L ÿ/ÿ ÿ/ÿ ÿ/L ÿ/L ÿ/ÿ ÿ/L ÿ/ÿ ÿ/ÿ L/L ÿ/ÿ ÿ/L ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

+/+ ÿ/+ n/n ÿ/ÿ n/+ +/+ ÿ/ÿ +/+ +/+ +/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/+ ÿ/+ ÿ/ÿ ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/n ÿ/ÿ ÿ/L ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/n

ÿ/+ ÿ/+ ÿ/ÿ ÿ/+ n/n n/n ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

ÿ/+ L/+,L ÿ/+ ÿ/L ÿ/ÿ ÿ/+ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/n ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/+ ÿ/ÿ ÿ/+ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ n/n ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ ÿ/ÿ

Left side of the slash (/) indicates microsatellite alterations in intestinal metaplasia and the right side shows the result in cancer tissue. +, positive for microsatellite instability;-, negative for microsatellite instability; L, positive for loss of heterozygosity; n, not determined. a Samples 5, 6 and 30, 31 are the distinctive tissues from same patients, respectively.

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119 Table 2 Frequency of microsatellite alterations in 32 patients

MSI-H MSI-L LOH

Intestinal metaplasia (%)

Cancer (%)

0/33 11/33 (33) 15/33 (45)

3/33 (9) 19/33 (58) 24/33 (73)

* y

MSI-H: high-level microsatellite instability in 530% of the markers; MSI-L: low-level microsatellite instability in <30% of the markers; LOH, loss of heterozygosity. *P=0.013; yP=0.028.

already possessed it at the stage of intestinal metaplasia. The frequency of LOH was also increased signi®cantly in cancerous tissues compared with intestinal metaplasia (P=0.028). LOH preferentially occurred at D2S123, D5S107 and D17S261 loci. D2S123 and D5S107 loci were especially vulnerable to LOH in intestinal metaplasia, whilst LOH at D17S261 was most common in cancerous tissues. Moreover, cancer specimens positive for alterations had a tendency to be accompanied by positive metaplasia specimens, and vice versa, which was con®rmed to be statistically signi®cant (P=0.005 for MSI and P=0.001 for LOH) (Table 3). 3.2. Mutation of the mononucleotide tracts within the cancer-related genes and mismatch-repair genes

1117

connection between intestinal metaplasia and well-differentiated adenocarcinoma in the stomach have been rare. To approach this problem, we focused on cancer in an early stage. In an analysis of 33 early cancer specimens paired with corresponding metaplasia lesions, we found that MSI was relatively common in intestinal metaplasia, and that the molecular changes noted, either MSI or LOH, tend to be coincident between metaplasia and cancerous lesions in the same sets of specimens (Table 3). Moreover, there was no case that showed alterations in metaplasia, but not in cancerous lesions. This ®nding implies that in many cases, cancer conserves the status in metaplasia lesions, suggesting the chronological connection between these two di€erent lesions. These results indicate that metaplasia and welldi€erentiated adenocarcinoma in the stomach may have the same molecular backgrounds, and supports the argument that intestinal metaplasia in the stomach may be a precursor for well-di€erentiated adenocarcinoma. Many previous investigations of advanced gastric cancers indicated that MSI plays an important role with a frequency of 18-38% as an early event in stomach carcinogenesis [8±12,19,20]. However, our observations in the early stage of cancers indicated that the frequency of MSI-H was lower than that reported elsewhere and that MSI-L was more commonly observed. Recently,

We also studied the frameshift mutation of microsatellites within cancer-related genes (TGF- RII and BAX) and mismatch repair genes (hMSH3, hMSH6). None of the cancer specimens showed frameshift mutations in these genes with the exception of patient 2, in which a frameshift mutation of the TGF- RII (A)10 tract was observed. A tumour specimen from patient 2 clearly exhibited a single base deletion in the (A)10 tract of TGF- RII (Fig. 2). 4. Discussion Although the role of intestinal metaplasia as a precancerous lesion is widely accepted, reports showing the

Table 3 Patterns in occurrence of microsatellite instability (MSI) or loss of heterozygosity (LOH) in paired specimens of intestinal metaplasia and cancers Caÿ

Ca+

IMÿ

MSI/LOH 11/9

MSI/LOH 11/9

IM+

0/0

11/15

IM, Intestinal metaplasia; Ca, Cancer.

P=0.005/0.001

Fig. 2. Frameshift mutation of the TGF- RII gene in tumour tissues of patient 2. N, normal; IM, intestinal metaplasia; T, tumour DNA. Results of mutation analysis of the TGF- RII gene. PCR product from tumour DNA exhibited a peak shift from 90 to 89 bp. The arrow indicates a one-base deletion within the polyadenine repeat.

1118

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119

Boland and associates referred to the distinction between MSI-H and MSI-L in colorectal cancers [32]. Gleeson and colleagues and Hamamoto and coworkers also reported that these low-level microsatellite alterations might not result from the impairment of the mismatch repair system, but might re¯ect a relaxation in genetic stability, which increases susceptibility to mutation throughout the cell genome [33,34]. These and our results suggested that MSI-L frequency depends on the proliferation status of the relevant cells. The MSI-L phenotype results in a vulnerability of the cells to mutagens, a subsequent accumulation of DNA damage and eventual malignant transformation with exposure to various exogenous mutagens. In the present study, 24 tumour tissues and 15 intestinal metaplasia lesions were LOH-positive, and these frequencies were higher than those of MSI. Surprisingly, more than half of all LOH observed in cancers were seen at the stage of intestinal metaplasia, and the frequency of LOH increased from 45% in intestinal metaplasia to 73% in the cancerous tissues. Colorectal cancers acquire genetic instability via two major pathways: MSI which induces subtle changes in DNA sequences; and chromosomal instability. LOH is most commonly observed on the latter pathway [35]. Thus, our observation indicated that chromosomal instability is more closely linked to the initiation of carcinogenesis than MSI in well-di€erentiated adenocarcinoma of the stomach. LOH in intestinal metaplasia occurred at D2S123 and D5S107. Since D2S123 and D5S107 are hMSH2- and APC- linked loci, respectively [16,17], we speculate that LOH of these loci might be related to the development of carcinogenesis from intestinal metaplasia. In contrast, most LOH at D17S261 was detected in the cancerous tissues. D17S261 is a p53-linked locus, and LOH in this locus may be involved in the progression of cancer. Frameshift mutations within the coding regions of TGF- RII and BAX have been reported in gastric cancer [22]. Mutations in hMSH3 and hMSH6 will enhance MSI resulting in tumour cells with a mutator phenotype. In our study, only one cancerous specimen (patient 2), which was categorised as MSI-H, exhibited a frameshift mutation in the (A)10 tract within TGF RII and no frameshift mutations in the other three genes was observed in all the tumour specimens examined. Although several studies have demonstrated a high frequency of mutations in these genes [18±22], a frameshift mutation in these genes was rare in this study. This may derive, at least, partly from the small number of the tumour specimens that were examined, or from the low frequency of MSI-H tumours in the set of samples. Thus, alterations in these intragenic mononucleotide repeats could not be used as a molecular marker to discuss the connection between intestinal metaplasia and well-di€erentiated early gastric cancer.

Acknowledgements We are most grateful to Professor Teruaki Aoki (Director, Department of Surgery II, Jikei University School of Medicine) for helpful advice and critical reading of the manuscript. We also thank Miss Ritsuko Nakayama for her excellent assistance in sequencing, and Dr Takaaki Ohashi and Dr Yoshio Inoue for comments on the manuscript. This work was supported in part by the special Coordination Fund for ``Research for the future'' from the Science and Technology Agency and by a grant in aid to a ``High-Tech Research Center (for Institute of DNA Medicine)'' from the Ministry of Education of Japan. References 1. Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer 1993, 54, 594±606. 2. Lauren P. The two histological main types of gastric carcinoma: di€use and so-called intestinal-type carcinoma. Acta Pathol Microbiol Scand 1965, 64, 31±39. 3. Correa P, Shiao Y-H. Phenotypic and genotypic events in gastric carcinogenesis. Cancer Res 1994, 54(Suppl.), 1941s±1943s. 4. Stemmermann GN. Intestinal metaplasia of the stomach: a status report. Cancer 1994, 74, 556±564. 5. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990, 61, 759±767. 6. Ochiai A, Yamauchi Y, Hirohashi S. p53 mutations in the nonneoplastic mucosa of the human stomach showing intestinal metaplasia. Int J Cancer 1996, 69, 28±33. 7. Nishimura K, Yokozaki H, Haruma K, Kajiyama G, Tahara E. Alterations of the APC gene in carcinoma cell lines and precancerous lesions of the stomach. Int J Oncol 1995, 77, 587± 592. 8. Rhyu M-G, Park W-S, Meltzer SJ. Microsatellite instability occurs frequently in human gastric carcinoma. Oncogene 1994, 9, 29±32. 9. Chong J-M, Fukayama M, Hayashi Y, et al. Microsatellite instability in the progression of gastric carcinoma. Cancer Res 1994, 54, 4595±4597. 10. Lin J-T, Wu M-S, Shun C-T, et al. Microsatellite instability in gastric carcinoma with special references to histopathology and cancer stages. Eur J Cancer 1995, 31A, 1879±1882. 11. Chung Y-J, Song J-M, Lee J-Y, et al. Microsatellite instabilityassociated mutations associate preferentially with the intestinal type of primary gastric carcinomas in a high-risk population. Cancer Res 1996, 56, 4662±4665. 12. Dos Santos NR, Seruca R, Constancia M, Seixas M, SobrinhoSimoes M. Microsatellite instability at multiple loci in gastric carcinoma: clinicopathologic implications and prognosis. Gastroenterology 1996, 110, 38±44. 13. Fishel R, Lescoe MK, Rao MRS, et al. The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 1993, 75, 1027±1038. 14. Bronner CE, Baker SM, Morrison PT, et al. Mutation in the DNA mismatch repair gene homologue hMLH1 is associated with hereditary non-polyposis colon cancer. Nature 1994, 368, 258±261. 15. Nicolaides NC, Papadopoulos N, Liu B, et al. Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 1994, 371, 75±80.

K. Kobayashi et al. / European Journal of Cancer 36 (2000) 1113±1119 16. Thibodeau SN, Bren G, Schaid D. Microsatellite instability in cancer of the proximal colon. Science 1993, 260, 816±819. 17. PeltomaÈki P, Aaltonen LA, Sistonen P, et al. Genetic mapping of a locus predisposing to human colorectal cancer. Science 1993, 260, 810±812. 18. Markowitz S, Wang J, Myero€ L, et al. Inactivation of the type II TGF-b receptor in colon cancer cells with microsatellite instability. Science 1995, 268, 1336±1338. 19. Ohue M, Tomita N, Monden T, et al. Mutations of the transforming growth factor b type II receptor gene and microsatellite instability in gastric cancer. Int J Cancer 1996, 68, 203±206. 20. Renault B, Calistri D, Buonsanti G, Nanni O, Amadori D, Ranzani GN. Microsatellite instability and mutations of p53 and TGF-b RII genes in gastric cancer. Hum Genet 1996, 98, 601±607. 21. Rampino N, Yamamoto H, Ionov Y, et al. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 1997, 275, 967±969. 22. Chung Y-J, Park S-W, Song J-M, et al. Evidence of genetic progression in human gastric carcinomas with microsatellite instability. Oncogene 1997, 15, 1719±1726. 23. Malkhosyan S, Rampino N, Yamamoto H, Perucho M. Frameshift mutator mutations. Nature 1996, 382, 499±500. 24. Yamamoto H, Sawai H, Perucho M. Frameshift somatic mutations in gastrointestinal cancer of the microsatellite mutator phenotype. Cancer Res 1997, 57, 4420±4426. 25. Vasen HFA, Mecklin J-P, Khan PM, Lynch HT. Hereditary nonpolyposis colorectal cancer. Lancet 1991, 338, 877. 26. Japanese Research Society for Gastric Cancer. Japanese classi®cation of gastric carcinoma. Second English Edition. Gastric Cancer 1998, 1, 10±24.

1119

27. Sobin LH, Wittekind Ch. TNM Classi®cation of Malignant Tumours. 5th ed. New York, Wiley-Liss, Inc, 1997. 28. Goeltz SE, Hamilton SR, Vogelstein B. Puri®cation of DNA from formaldehyde ®xed and paran embedded human tissue. Biochem Biophys Res Comm 1985, 130, 118±126. 29. Dietmaier W, Wallinger S, Bocker T, Kullmann F, Fishel R, RuÈscho€ J. Diagnostic microsatellite instability: de®nition and correlation with mismatch repair protein expression. Cancer Res 1997, 57, 4749±4756. 30. Toh Y, Oki E, Oda S, et al. An integrated microsatellite length analysis using an automated ¯uorescent DNA sequencer. Cancer Res 1996, 56, 2688±2691. 31. Goessl C, Heicappell R, MuÈnker R, et al. Microsatellite analysis of plasma DNA from patients with clear cell renal carcinoma. Cancer Res 1998, 58, 4728±4732. 32. Boland CR, Thibodeau SN, Hamilton SR, et al. A national cancer institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998, 58, 5248±5257. 33. Gleeson CM, Sloan JM, McGuigan JA, Ritchie AJ, Weber JL, Russell SEH. Ubiquitous somatic alterations at microsatellite alleles occur infrequently in Barrett's-associated esophageal adenocarcinoma. Cancer Res 1996, 56, 259±263. 34. Hamamoto T, Yokozaki H, Semba S, et al. Altered microsatellite in incomplete-type intestinal metaplasia adjacent to primary gastric cancers. J Clin Pathol 1997, 50, 841±846. 35. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998, 396, 643±649.