Characterization of p53 gene mutations in esophageal squamous cell carcinoma cell lines: increased frequency and different spectrum of mutations from primary tumors.

We screened 29 human esophageal squamous cell carcinoma (ESC) cell lines for mutations of the p53 gene through all coding exons and exon-intron junctions. Mutations were found in 22 cell lines (76%), consisting of 20 single-base substitutions, 2 small deletions and 1 single-base insertion. Out of 20 single-base substitution, 5 were located at the exon-intron junctions and mRNAs with abnormal splicing were detected by RT-PCR in 4 of them. A G:C to T:A transversion, which occurred rather frequently in resected tumors of ESC, was observed in only 1 cell line, and, instead, frequent transitions at CpG sites were detected. We also examined 65 fresh tumor materials, from all of which we tried to establish cell lines, and detected mutations in 26 samples (40%). Compared with the results in these fresh tumor materials, the mutation incidence in cell lines was significantly high and the mutation spectrum was also different. From these 65 tumors, 10 cell lines were established, including 3 cell lines from 26 tumors with p53 mutations and 7 cell lines from 39 without mutations, which indicates that there was no significant correlation between the status of the p53 gene in each fresh tumor and its establishment as a cell line. In 7 cell lines established from mutation-free tumors, newly acquired mutations were detected in 5, which suggests that mutations might occur during the process of establishing cell lines.

p53 mutation, murine double minute 2 amplification, and human papillomavirus infection are frequently involved but not associated with each other in esophageal squamous cell carcinoma.

We examined the relationship between p53 mutation, murine double minute 2 (MDM2) gene amplification, and human papillomavirus (HPV) infection in 72 esophageal squamous cell carcinomas. We identified p53 mutations in 29 tumors (40.3%) by PCR-single-strand conformation polymorphism analysis and direct sequencing. Amplification of the MDM2 gene was detected by Southern blot hybridization in 13 (18.1%) of 72 tumor tissues and in 4 (33.3%) of 12 cultured esophageal squamous cell lines. All four cell lines with MDM2 amplifications showed overexpression of the MDM2 mRNA in Northern blotting. We observed HPV infection in 15 (20.8%) of 72 tumor tissues by specific PCR amplification and Southern blot hybridization. In most tumors, amplification of the MDM2 gene or infection of HPV was not associated with p53 mutations, except in four cases with p53 mutation and MDM2 amplification, and three cases with p53 mutation and HPV infection. Since p53 mutations, MDM2 overexpression, and HPV infection are all considered to abrogate the normal function of p53 protein, each of these genetic changes may be equally important in tumorigenesis. In addition, we found that patients with MDM2 amplification exhibited a significantly shorter survival period (P = 0.0053).

Allelotype analysis of esophageal squamous cell carcinoma.

In previous studies, we have shown that allelic loss on chromosome 17p, on which the p53 gene is located, is very frequent, and loss-of-function mutations of the p53 gene are closely associated with the tumorigenesis of esophageal cancer. In this study, we performed allelotype analysis to investigate whether other tumor suppressor genes are also involved in esophageal cancer. Using 55 polymorphic DNA markers covering every autosomal arm except 13p, 21p, and 22p, restriction fragment length polymorphism analysis was performed on 36 esophageal squamous cell carcinomas (ESCs) and their adjacent normal tissue samples. Frequent loss of heterozygosity (LOH) of > 30% of the informative cases was observed on chromosomes 3p (41.1%), 5q (52.6%), 6p (30.4%), 8p (33.3%), 9p (35.7%), 9q (30.8%), 11p (32.4%), 13q (52.7%), 17p (55.2%), 17q (33.3%), 18q (45.7%), and 19q (30.4%). Among these, LOH on 5q, 13q, 17p, and 18q was previously reported in ESC and is considered to involve the APC, RB, p53, and DCC genes, respectively. However, our deletion analysis of chromosome 18q revealed that the region commonly lost did not include the DCC locus, suggesting that a possible tumor suppressor gene on 18q other than the DCC gene is involved in ESC. We screened 60 primary ESC tumors and 20 cultured ESC cell lines for the mutation of the APC gene within a mutation cluster region in exon 15, where the "hot spot" of somatic mutation for colorectal and pancreatic cancers is thought to be. We could not find any mutation despite the high frequency of LOH on chromosome 5q. We also analyzed the relationship between the clinicopathological data and the allelic loss and found that LOH on chromosomes 6p and 13q was associated with poor prognosis.

Loss of 17p, mutation of the p53 gene, and overexpression of p53 protein in esophageal squamous cell carcinomas.

We analyzed mutations of the p53 gene by single-strand conformation polymorphism analysis of polymerase chain reaction products and direct sequencing through all coding exons and exon-intron junctions in 32 cases with esophageal squamous cell carcinoma. Mutations were detected in 15 of 32 (47%) tumor samples, in which G:C to T:A transversions were rather frequent (33%). Previously, we reported deletion of chromosome 17p where the p53 gene is located in 45% of Japanese esophageal squamous cell carcinoma, and here the relationship between mutation of the p53 gene and loss of 17p was analyzed. Mutations were observed in 12 of 16 patients with loss of 17p, whereas only 2 of 11 without loss were positive for mutations, suggesting that mutations of the p53 gene were closely associated with a 17p deletion. Furthermore, we immunohistochemically analyzed the expression of p53 protein in esophageal squamous cell carcinoma tumor tissues using a monoclonal antibody. Five of 6 tumors with missense mutations of the p53 gene were positively stained, while in tumors with nonsense mutations or without mutations of the p53 gene staining was very weak or negative. These results suggest a good correlation between mutations and abnormal expression of the p53 gene.

Clinicopathological characteristics of duodenal microgastrinomas.

Duodenal gastrinomas do not seem to behave as malignantly as sporadic pancreatic gastrinomas. Statistical analysis of 49 patients with sporadic pancreatic gastrinoma and 21 patients with sporadic duodenal gastrinoma reported since 1980 in Japan revealed that the incidence of hepatic metastasis was 57% in patients with sporadic pancreatic gastrinoma and only 9% in patients with sporadic duodenal gastrinoma (p less than 0.01). These findings suggest that there is an essential biological differences between duodenal and pancreatic gastrinoma. Five patients with sporadic duodenal microgastrinoma (tumor diameter less than 5mm) in our hospital had no hepatic metastases; however, 4 patients had lymph node metastases. Immunohistochemical study of 5 sporadic duodenal microgastrinomas and 6 sporadic pancreatic gastrinomas revealed that the sporadic duodenal gastrinomas contained significantly fewer insulin-producing or glucagon-producing cells than sporadic pancreatic gastrinomas. The cellular composition of the metastatic lymph nodes from duodenal microgastrinomas was similar to that of the primary tumor. This difference in cellular composition between the duodenal microgastrinomas and the pancreatic gastrinomas suggests that the process of development and differentiation of gastrinoma cells is different.

Characterization of 21 newly established esophageal cancer cell lines.

Twenty-one esophageal cancer cell lines (KYSE series) have been established from the resected specimens of patients with esophageal cancer. Three lines, KYSE-30, KYSE-50, and KYSE-70, were derived from the implanted tumor of nude mice (initial passage); others were derived from resected specimens. Each cell line was morphologically distinct. Detailed cytogenetic analysis indicated that each cell line was karyotypically unique, and DNA fingerprint analysis showed no cross-contamination among cells. Doubling time ranged from 13.7 to 75.5 hours, and modal chromosome numbers ranged from 46 to 120. Most cell lines grew in monolayer, but two cell lines (KYSE-50 and KYSE-360) grew as floating cell aggregates. No correlation was demonstrated between the establishment of cell lines and cell differentiation. These cell lines are the first reported to be homogeneous and individually unique and may provide a useful model for the study of human esophageal cancer.

[Protein-free culture of esophageal cancer cell lines].

We have established 13 esophageal cancer cell lines capable of growing in a protein-free environment. The growth of these cells was not affected by conditioned medium, but the growth of NIH3T3 cells and human fibroblasts was stimulated by conditioned medium. On the other hand, conditioned medium inhibited the growth of human endothelial cells. Amplified int-2 oncogene correlated well with the growth of cells in a protein-free environment but the number of EGF receptors and growth effect of EGF did not relate to such growth. Esophageal cancer cells grow automatically, possibly involving mesenchymal cells via the paracrine system. This results in a poor prognosis in patients.

Deletion of 17p and amplification of the int-2 gene in esophageal carcinomas.

We have analyzed allelic deletion at 23 loci on 18 different chromosomes in 35 esophageal squamous cell carcinoma tissues by using restriction fragment length polymorphism markers. Loss of heterozygosity was detected on chromosomes 2, 3, 6, 7, 11-14, 16-18, 21, and 22, while no loss was detected on chromosomes 1, 4, and 8-10. Only the loss of chromosome 17p was detected with high frequency (45%), and losses on other chromosomes had frequencies of less than 22%. These losses with low frequencies might be random losses caused by chromosomal rearrangement during the course of tumor development and progression. On the contrary, the loss of 17p might play an important role in the development of esophageal squamous cell carcinoma, such as inactivation of a tumor suppressor gene. Amplification of the int-2 gene was observed in 39% of the tumors. However, no significant relationship between int-2 amplification and the deletion of any chromosome was detected.

Unresponsiveness of insulinoma cells to secretin: significance of the secretin test in patients with insulinoma.

It is well known that B cells in the pancreas release insulin when stimulated by secretin, but there have been few reports on the response of insulinoma cells to secretin. In five patients with insulinoma, changes in serum immunoreactive insulin (IRI) concentration were measured after the intravenous injection of secretin into the peripheral vein before and after extirpation of the insulinoma. The extirpated insulinomas were cultured and tested for their response to secretin. The rise in serum IRI in response to secretin in patients with insulinoma was significantly slower and smaller than in normal volunteers. After removal of the insulinoma, the response to secretin became prompt and increased with time. Cultured insulinoma cells did not release insulin when stimulated by secretin. Therefore, it is concluded that the response of insulinoma cells to secretin is quite different from that of normal beta cells, and that the function of beta cells in the insulinoma-bearing pancreas is suppressed by the autonomous hypersecretion of insulin by the insulinoma. The extent of the decrease in function of the beta cells in patients with insulinoma can be estimated by the intravenous secretin test. Thus, the secretin test is sometimes useful in the differentiation of hypoglycemia due to insulinoma from that due to beta cell hyperplasia or alimentary hyperinsulinemia.