Mutations in the ST7/RAY1/HELG locus rarely occur in primary colorectal, gastric, and hepatocellular carcinomas

Human cancers frequently show a loss of heterozygosity on chromosome 7q31, which indicates the existence of broad-range tumour-suppressor gene(s) at this locus. Truncating mutations in the ST7 gene at this locus are seen frequently in primary colon cancer and breast cancer cell lines. Therefore, the ST7 gene represents a novel candidate gene for the tumour suppressor at this locus. However, more recent studies have reported that ST7 mutations are infrequent or absent in primary cancer and cell lines. To ascertain the frequency of mutations of the ST7 gene in cancer cells, we examined mutations in the ST7 coding sequence in 48 colorectal, 48 gastric, and 48 hepatocellular carcinomas using polymerase chain reaction–single-strand conformational polymorphism and direct sequencing. We detected somatic mutations, which were located near the exon–intron junction in intron 8, in only three out of 144 cases. We conclude that mutations in the ST7 gene are rare in primary colorectal, gastric, and hepatocellular carcinomas.

The ST7 gene, which in other contexts is designated as RAY1 (Vincent et al, 2000) or HELG (Hughes et al, 2001), maps within this critical region. Recently, frameshift mutations in the ST7 gene have been observed frequently in primary colon cancer and breast cancer cell lines (Zenklusen et al, 2001). The introduction of ST7 cDNA suppressed the tumorigenicity of a prostate cancer cell line in vivo (Zenklusen et al, 2001). These results suggest that the ST7 gene is a candidate tumour-suppressor gene within this critical region. However, there have been reports that somatic mutation in the ST7 gene is extremely rare (Hughes et al, 2001;Thomas et al, 2001;Brown et al, 2002;Dong and Sidransky, 2002). Thus, the previous data on ST7 gene mutations show conflicting results.
In this study, we investigated the true frequency of ST7 gene mutations by examining 48 primary colorectal cancers, 48 primary gastric cancers that frequently show LOH on 7q31 (Nishizuka et al, 1997), and 48 primary hepatocellular carcinomas that show highlevel expression of the ST7/RAY1 gene (Vincent et al, 2000;Zenklusen et al, 2001). We surveyed mutations in the entire ST7 coding sequence using polymerase chain reaction -single-strand conformational polymorphism (PCR -SSCP) analysis and direct DNA sequencing.

Tissue specimens and DNA extraction
Specimens from 48 colorectal, 48 gastric, and 48 hepatocellular carcinomas and corresponding noncancerous tissues were obtained at surgery from Japanese patients. The samples were frozen immediately in liquid nitrogen and stored at À801C until use. High-molecular-weight DNA was extracted using the standard phenol/chloroform procedure.
Polymerase chain reaction -single-strand conformational polymorphism analysis All samples were examined by PCR -SSCP analysis for mutations throughout the entire coding sequence of the ST7 gene (exons 1a -16b). The exon -intron boundaries were identified by comparing the cDNA sequences of ST7 (GenBank accession no. AY009152) and the genomic DNA sequence of chromosome 7q31 (AC002542). Using this information, we designed intronic primers for each genomic region, except for exons 1b and 16b (Table 1). The primers for exons 1b and 16b were prepared as described previously (Thomas et al, 2001).
The genomic DNA template (50 ng) was incubated in a total volume of 10 ml PCR buffer that contained 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl 2 , 100 nM of each primer, 200 mM of each deoxynucleotide triphosphate, 1.5 Ci of alpha-[ 32 P]dCTP (Amersham Pharmacia), and 0.5 U of rTaq DNA polymerase (TaKaRa). The following PCR conditions were employed: 30 s at 951C, 30 s at 581C or 621C, and 90 s at 721C for 35 cycles, followed by 10 min at 721C in a thermal cycler (GeneAmp 9700; Applied Biosystems). Single-strand conformational polymorphism analysis was performed with the low-pH buffer system, which allowed improved separation of fragments of up to 800 bp in length (Kukita et al, 1997). The 32 P-labelled PCR products were denatured, loaded on nondenaturing polyacrylamide gels that contained 10% polyacrylamide (99 : 1 acrylamide to bisacrylamide) and TPE (30 mM Tris (pH 6.8), 20 mM PIPES, and 1 mM Na 2 EDTA), and electrophoresed in TPE buffer at 101C. The gels were dried and analysed with the BAS 2000 system (Fuji Photo Films). To exclude potential PCR artefacts, all positive cases were tested independently at least three times.

Sequencing analysis
PCR fragments that showed different mobilities were purified using the QIAquick PCR Purification Kit (QIAGEN), and directly sequenced in both directions using the BigDye Terminator Kit and the ABI 3100 DNA Sequencing System (Applied Biosystems).

Analysis of microsatellite instability
We assessed microsatellite instability using five reference markers (D2S123, BAT25, BAT26, D5S346, and D17S250) and the criteria recommended by the National Cancer Institute workshop (Boland et al, 1998;Yamada et al, 2002).

Statistical analysis
Statistical analysis was performed using the StatView 5.0. program (SAS Institute Inc.). The w 2 , Fisher's exact, and Mann -Whitney Utests were used for background and clinicopathological data. A Pvalue of less than 0.05 was considered to be statistically significant.

Ethics
This study was carried out with the approval of the ethical committee of Gunma University Faculty of Medicine.

RESULTS
We detected a somatic mutation in the polypyrimidine tract within the splice-acceptor site of the intron 8 -exon 9 junction. Deletions in intron 8 (À32 nucleotides from exon 9) were found in one out of 48 (2.1%) of the colorectal cancers, and in two out of 48 (4.1%) of the gastric cancers ( Table 2). The number of deleted nucleotides in one tumour sample ranged from one to three bases, which demonstrates the highly heterogeneous nature of the tumour. A representative case is shown in Figure 1. All the three patient groups exhibited high-frequency microsatellite instability (microsatellite instability-high; MSI-H).
We also detected a G to A substitution at the first nucleotide of codon 143 (GenBank accession no. AY009152) in exon 5 of one of the colorectal cancer cases (Figure 2, Table 2). This substitution resulted in an amino-acid change from Ala to Thr. The same substitution was found in the corresponding normal tissue from the same patient. Thus, the change represents a germline mutation or rare polymorphism.
In addition, we identified four single-nucleotide polymorphisms (SNPs) in introns 8, 10, 11, and 15 of the ST7 gene (Table 3). There were no correlations between these SNPs and the clinicopathological data.

DISCUSSION
We detected somatic mutations in the polypyrimidine tract within the splice-acceptor site of intron 8, although the frequency of  1a  273  62  gaatcatcccggcagacac  gcgcgagttgcactaacttt  1b  234  58  agcagagaggagcgctgaa  ttgcactaactttccggggc  2  148  62  ccttgttcttctccctttctc  ttaaatgagaaggactccacc  3  230  58  aacagtgaccataaacacgct  aaataatattgcaaactgaagg  4  162  62  gtagtgtcactgaacttacgc  ctgtctttgctctctgaacc  5  369  58  aggtcttgcttttctctctca  gaggggactcatttcaacata  6  220  62  ggattgacttggtgttttctc  atcctccagttcaaatgcagt  7  176  62  gtgactctctctgaatgttcc  tcatttggttagaagtagggc  8  237  62  ggctttgtaattgatggtggc  acaattctgatccccccaatgc  9  342  58  tcaacatcctcactcaaaagc  tctgtaagccactgatcccaa  10  195  58  attccttggtttcttctgccc  gggaaaatacatcaaaagagg  11  191  58  cctgcaaacttatgtgttcct  aacacatctcaattccggtca  12  168  62  ggatggtttttgtctttctgc  atcataacgagttcctgtggg  13  237  62  attaacacaagtgtgtcctgc  ttagcaccttttcatgctctt  14  209  62  cacaaacattggacatctctg  ctggctgaagagaggtgaga  15  351  58  gggtcagatgttggctatgg  cttggctttccccatccatt  16a  200  62  ggtttctgctgacttctgtg  aaggagttggcacagaggag  16b 225 58 aggcgagtgcaatcagaaag gaggaggagcagttttggtg mutation was low. The polypyrimidine tract is essential for efficient branch-point utilisation and splice-site recognition, and deletions within this region affect splicing efficiency (Roscigno et al, 1993). We were unable to examine whether the mutation at the polypyrimidine tract induced insufficient splicing, because the appropriate RNA samples were not available. Therefore, we could not confirm the involvement of this mutation in carcinogenesis and the progression of colorectal and gastric cancers. However, considering the fact that mutations were found only in the tumour samples and not in the corresponding noncancerous samples, we cannot exclude the possibility that this mutation confers advantages upon these cancer cells under selective pressure. All of the cases with the polypyrimidine-tract mutation showed high-frequency microsatellite instability. It is known that a simple mononucleotide repeat can act as a mutational target in tumours that show high-frequency MSI (Perucho, 1999;Yamada et al, 2002). Therefore, the polypyrimidine tract, in which we found mutations, may be a mutational target in MSI-positive tumours, and mutations therein may be involved in carcinogenesis and the progression of MSI-positive tumours. Microsatellite instabilityhigh is found rarely in hepatocellular carcinomas, and we could not detect any mutations in the polypyrimidine tract in the 48 cases of hepatocellular carcinoma (Saeki et al, 2000;Yamamoto et al, 2000;Wang et al, 2001).
We also detected a single-nucleotide substitution with aminoacid change in one patient with colon cancer. This substitution was identical to that identified previously in the breast cancer cell line MDA-MB435 (Thomas et al, 2001). Since the corresponding noncancerous cell line was not available, these investigators could not determine whether the change was somatic or germline specific. In contrast, we found the same substitution in the corresponding normal tissue from the same patient. Therefore, the change is not somatic, but represents a germline mutation or rare polymorphism. Further functional studies are needed to clarify the ramifications of this amino-acid substitution.
In addition, we detected four SNPs in the ST7 gene locus. There were no correlations between these SNPs and the clinicopathological data. The consequences of these SNPs for colorectal, gastric, and hepatocellular carcinomas are unclear.
Contrary to the result of Zenklusen et al, we rarely detected mutations in the ST7 gene of patients with colorectal, gastric, or hepatocellular carcinoma, a finding that has been corroborated by other groups (Hughes et al, 2001;Thomas et al, 2001;Brown et al, 2002;Dong and Sidransky, 2002). In our study, there were no  technical problems in detecting mutations; the frequent detection of SNPs demonstrates the high sensitivity of our procedure. Although the reason for the discrepancy between our results and those of Zenklusen et al is unclear, we (and the aforementioned groups) propose the following possible explanations: the use of selected specimens, the presence of PCR artefacts, and the effects of culture passages (Hughes et al, 2001;Thomas et al, 2001;Brown et al, 2002). We conclude that ST7 gene mutations are rare in colorectal, gastric, and hepatocellular carcinomas. Our results do not exclude the possibility that the ST7 gene is inactivated by other molecular mechanisms, such as aberrant hypermethylation or haplo-insufficiency (Merlo et al, 1995;Largaespada, 2001). Since there have been no reports on the expression of the ST7 gene in cancer cells, further studies are needed to understand the role of this gene in carcinogenesis and the progression of these cancers.