INTRODUCTION

The mdm2 gene has been mapped to chromosome 12q13–14 and encodes a 90 kDa cellular oncoprotein. The gene structure on the human chromosome was identified in 1992 (1). Like viral proteins, mdm2 protein is also capable of binding to p53 (2, 3, 4). The complex that develops in this process inactivates the tumor suppressive function of p53 and prevents p53 from intervening into the cell cycle. The functional effect of inactivation through the formation of a p53-mdm2-complex thus corresponds to a complex formation between wild-type p53 and mutated p53 (5). Moreover, mdm2 is capable of forming complexes with both wild-type p53 and mutated p53. The interaction and relative ratios of p53 and mdm2 protein seem to play an important role in regulating cell division (3). Simultaneously, mdm2 inhibits the p53 function by binding to its transcriptional activation domain (6).

Furthermore, it was shown that p53 activates the expression of the mdm2 gene in an autoregulatory feedback loop (7, 8). Otherwise, the overexpression of mdm2 can reduce the amount of endogenous p53 through enhanced proteasome-dependent degradation (9). Recent research has suggested that mdm2 also promotes the rapid degradation of p53 under conditions in which p53 is otherwise stabilized (10).

At the level of amplification or protein overexpression, mdm2 plays a role in a variety of tumors such as sarcomas (11, 12) and epithelial neoplasms, such as breast and lung tumors (13, 14, 15). In the early gastric carcinogenesis, however, mdm2 protein overexpression seems to play no role (16). To the best of our knowledge, this study is the first to investigate mdm2 expression and amplification in advanced-stage gastric cancer.

MATERIALS AND METHODS

Tissue

We investigated a series of 43 gastric carcinomas received in the Department of General Surgery, Magdeburg University, between 1995 and 1996. The resected stomachs were sent to the Department of Pathology for rapid section.

In addition to routine preparation, further representative tumor material was prepared, snap-frozen in liquid nitrogen, and stored at −70° C. All tumors showed a deep invasion of the gastric wall and were classified as advanced-stage cancers according to the pTNM system (pT 2–4). The number, origin, and involvement of the regional lymph nodes excised were also determined according to the International Union Against Cancer TNM system (17). Clinical data on the presence of metastases were included in this classification. For the histologic tumor typing, Lauren's classification system was used (18). There were 18 carcinomas of the diffuse type (mean age of patients, 60.9; range, 33 to 76; sex: f/m, 9/9), 20 intestinal cases (mean age of patients, 66.2; range, 34 to 83; sex: f/m, 5/16), and four mixed-type tumors (mean age of patients, 67.5; range, 64 to 71; sex: f/m, 2/2). One tumor was classified as a primary small cell carcinoma of the stomach.

Immunohistology

Mdm2 and p53 protein

Immunohistochemical analyses of p53 and mdm2 protein were performed on formalin-fixed, paraffin-embedded material. For the reactions, standardized and automated immunohistochemistry were applied (Nexes; Ventana Medical Systems, Tucson, AZ). Sections (4 μm thick) were deparaffinized in a series of graded alcohols and microwaved in EDTA buffer for 2 × 10 min at 450 W.

After the sections were cooled to room temperature and rinsed with Tris buffer, incubation was performed with anti-p53 (Do-1; Calbiochem, Cambridge, MA) or anti-mdm2 (SMP-14; DAKO, Hamburg, Germany) in an immunostainer. Before incubation of the primary antibodies, endogenous peroxidase was blocked by preincubation of the slides with 0.3% H2O2. The p53 antibody reacted specifically with both the wild type of protein p53 and its mutant form, recognizing an epitope between the amino acids 37 and 45 (19). Anti-p53 was diluted 1:30 and incubated at 37° C for 26 min; anti-mdm2 was diluted 1:50 and incubated at 37° C for 30 min. Bound primary antibodies were detected by the avidin-biotin-complex method (Ventana Medical Systems) labeled with horseradish peroxidase and diaminobenzidine as substrates. All detection steps were performed at 37° C in the immunostainer.

Assessment of both antigens was made by counting 1,000 cells per section at high magnification (high-power field, 40×), and the percentage of positive tumor cells was determined. According to Kawai et al. (20), a tumor was considered positive for p53 when more than 10% of the tumor cells were immunopositive. With regard to positive mdm2 antibody reaction, the cutoff point was defined as 10%. For both antibodies p53 and mdm2, only the positive nuclear staining was evaluated. Sections from a breast carcinoma were used as a positive control for mdm2 and p53, respectively. The immunohistochemistry slides were evaluated independently by two pathologists (TG and CH).

Molecular Genetic Methods

DNA preparation

DNA was prepared from 2- to 10-mg fresh tissue through a phenol-chloroform extraction (21).

To estimate the percentage of tumor cells in hematoxylin and eosin slides, tumor tissue corresponding to all snap-frozen tumor samples was obtained. We used only those cases in which at least 70% tumor cells were seen. A dilutional effect caused by nontumor cells on the DNA analysis should thus be excluded.

Mdm2 amplification

Southern blot analysis was performed by 10 μg of DNA with EcoRI (AGS, Heidelberg, Germany), separating the fragments by electrophoresis on 0.8% agarose gels, and by transferring them to nylon filters (Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany). The DNA on the filters was then sequentially hybridized with probes for mdm2 (a cDNA fragment containing −300 to 1600 nt of the mdm2 gene [i.e., almost the entire coding region, provided by Dr. B. Vogelstein]) and for β-actin used as a control (1) for unequal amounts of DNA loaded. Probe labeling was done with the use of random oligonucleotide primers (Gene Images; CDP Star; Amersham Pharmacia Biotech Europe).

Hybridization signals were quantified by densitometry (VDS Amersham Pharmacia Biotech Europe; Dorval, Quebec, Canada; Biotech, Bethesda, MD). Signals that had an intensity at least three times higher than those from samples with a normal copy number of the gene (placenta) were scored as amplification. The osteosarcoma cell line OsA-CL (ATCC, Rockville, MD), in which mdm2 is amplified, was used as a positive control (1).

Polymerase chain reaction–single-strand conformation polymorphism technique

The conserved regions of the p53 gene (exons 4 to 8) were amplified and sequenced. Primers have recently been described by Günther et al. (13).

A total volume of 50 μl polymerase chain reaction (PCR) mixture contained 100 ng DNA, 250 μm dNTPs, 25 pmol primer each, 1.5 mm MgCl2, 1 × PCR buffer: 20 mm Tris-HCl (pH 8.4), 50 mm KCl, and 2.5 u Taq-DNA-polymerase (Life Technologies, Karlsruhe, Germany). PCR consisted of 35 cycles (92° C for 1 min, 52 to 62° C for 5 min, 72° C for 1 min) and was performed in an automated thermocycler (PTC, Watertown, MA).

Routine evidence of the PCR products was given on ultrathin polyacrylamide gels (PAA, 0.3- to 0.45-mm thick, 8 to 15%, depending on the fragment length) at 15° C for approximately 2.5 h in a horizontal electrophoresis system (Multiphor, Pharmacia/Biotech). DNA fragments were demonstrated using a modified silver staining protocol according to Budowle et al. (22). The single-strand conformation polymorphism (SSCP) technique (23) served as a prescreening method for the detection of mutations. Each PCR product that showed aberrantly migrating single strands was always amplified twice to exclude amplification artifacts. In brief, 4.5 μl of the PCR product and 4.5 μl of 100% formamide buffer (0.05% bromphenol, 0.05% xylencyanol) were denatured at 98° C for 5 min, subsequently chilled on ice, and applied to a 0.5 × mutation-detection-enhancement gel (AT Biochem). Gels were stained according to a protocol described by Goldman and Merril (24).

PCR products that showed mobility shifts of their single strands were directly sequenced on an automated fluorescence sequencer (ALF-Express, Pharmacia, Biotech).

Statistical Evaluation

The significance of associations between genetic and immunohistologic findings and clinicopathologic factors was evaluated statistically by Fisher's exact test (two tail) using SAS statistical software system (SAS Institute, Inc., Cary, NC). A probability of P <.05 was considered significant.

RESULTS

Mdm2

Of 43 gastric carcinomas investigated, 18 (41.8%) showed an amplification of the mdm2 oncogene (Fig. 1). The gene was upregulated in 9 of 18 (50%) diffuse-type carcinomas, in the small cell carcinoma, and in 3 of the 4 mixed-type tumors. Only 5 of the 20 (25%) intestinal-type carcinomas were found to show mdm2 gene amplification.

FIGURE 1
figure 1

Representative Southern blot of Eco-RI-digested genomic DNA demonstrating the level of mdm2 amplification in gastric carcinomas. Signal ratios of mdm2: β-actin (estimated densitometrically), which were three times higher than those of placenta, were scored as amplification (+). Samples without mdm2 amplification are scored (−).

Full size image

The expression of the mdm2 protein could also be demonstrated in 18 tumors. In 9 of 23 (39.1%) carcinomas with a diffuse growth pattern (Fig. 2), including 2 of the mixed type, the reaction was positive in more than 10% of the tumor cells. Nine of 20 carcinomas of the intestinal type (45%) were mdm2 positive.

FIGURE 2
figure 2

Mdm2-positive immunostaining in a diffuse-type gastric carcinoma, localized in the corpus (×400).

Full size image

Nontumor cells, such as connective tissue cells and smooth muscle cells of the stomach wall or epithelial cells of the mucosa, reacted only occasionally to the mdm2 antibody. Only the parietal cells of the corpus mucosa showed a nonspecific positive immunoreaction.

p53

The screening of carcinomas for genetic p53 alterations by SSCP analysis and direct sequencing revealed mutations in 9 of 43 (20.9%) tumors (Figs. 3 and 4). Only one carcinoma of the diffuse type showed a p53 mutation (5.5%); a second mutation was verified in the small cell carcinoma. Mutations were also seen in 7 of 20 intestinal-type tumors (35%). The four carcinomas with mixed differentiation showed no p53 mutations. The mutation spectrum is presented in Table 1.

FIGURE 3
figure 3

Polyacrylamide control gel of polymerase chain reaction of exon 5 of the p53 gene. c, negative control without DNA; m, DNA ladder; 1–9, gastric carcinomas. In lane 6 (tumor 2 in Table 1), a heteroduplex formation (arrows) can be seen, reflecting the heterozygous TA insertion confirmed by sequence analysis.

Full size image
FIGURE 4
figure 4

Single-strand conformation polymorphism analysis of exon 8 of the p53 gene in gastric tumor samples. In lane 4, an aberrantly migrating single strand can be seen, which was confirmed as a heterozygous missense mutation in codon 272 (tumor 6 in Table 1) in the sequence analysis.

Full size image
TABLE 1 p53 Mutation Spectrum in Gastric Cancer
Full size table

p53 overexpression was demonstrated in 14 of 20 (70%) tumors with a glandular growth pattern (Fig. 5). Twelve of them were of the intestinal type, and 2 were carcinomas of the mixed type, which expressed p53 in their glandular component. In contrast, the p53 protein was found by immunohistology in only four carcinomas with a diffuse growth pattern. No p53 overexpression was seen in the small cell tumor.

FIGURE 5
figure 5

p53-positive immunoreaction in an intestinal-type gastric carcinoma. Intestinal metaplasia on the left (×400).

Full size image

As summarized in Table 2, tumors with a diffuse growth pattern, including those classified as mixed type and the small cell carcinoma, tended to show a higher frequency of mdm2 amplification than carcinomas of the intestinal type (P =.062). There was also a tendency for p53 gene mutations to occur more frequently in the intestinal type (P =.059).

TABLE 2 Distribution of mdm2 and p53 Alterations in Intestinal-Type Carcinomas and Tumors with a Diffuse Growth Pattern, Including Mixed-Type Carcinomas and Small Cell Carcinoma
Full size table

The mdm2 protein expression showed no differences concerning the tumor type (P =.763), whereas a significantly higher frequency of the p53 protein overexpression could be found in carcinomas of the intestinal type (P =.001).

Altogether, approximately 80% of the gastric carcinomas investigated showed mdm2 and/or p53 alterations (Table 3), either at the level of genetic changes or at the protein level.

TABLE 3 mdm2/p53 Alterations in Gastric Cancer (n = 43)
Full size table

DISCUSSION

Numerous investigations have addressed alterations of the p53 tumor suppressor gene as well as the accumulation of the p53-coded protein in the genesis of different tumors (25). Several studies have demonstrated that p53 alterations, including protein overexpression, also play a role in the pathogenesis of gastric carcinoma (25, 26). On the one hand, the functions of the p53 gene, regulating the cell cycle (27) and apoptosis, are disturbed by gene mutations and deletions. On the other hand, both viral proteins and oncoproteins, such as mdm2, may manipulate the p53 function and thus, alternatively, exert an influence on the cell cycle via this pathway (2, 3, 4). The frequency of the mdm2 amplification in gastric carcinoma has not yet been investigated. Although in our study the mdm2 gene was amplified in 9 of 18 diffuse-type carcinomas, in the primary gastric small cell carcinoma and in 3 of 4 cases classified as mixed-type tumor, only 5 of the 20 intestinal-type tumors showed an amplification. Of the total of 43 gastric carcinomas investigated, the mdm2 amplification rate amounted to 41.8%. Marchetti et al. (15) determined an amplification rate of only 6% in non–small cell carcinomas of the lung. Investigating 52 breast carcinomas, the same group demonstrated mdm2 amplification in 23% (14). Higher amplification rates were reported in some soft tissue tumors. Leach et al. (11) observed mdm2 amplification in approximately 33% of the liposarcomas and malignant fibrous histiocytomas investigated. Another study of leiomyosarcomas found mdm2 amplification in only 2 of 29 tumors (6.9%) (12). In osteosarcomas, mdm2 upregulation was found in 14% (28). In contrast, as in Ewing's sarcoma (29), mdm2 seems to play no role in the pathogenesis of carcinomas of the cervix uteri (30). Blok and co-workers (16, 31) investigated the mdm2 protein expression in 45 early gastric cancers. At this stage of tumorigenesis, there was no evidence of an mdm2-linked dysfunction of p53. In our study, an immunohistologic proof of the mdm2 gene product could be given in 18 advanced gastric carcinomas. We found no significant difference between the carcinomas of the diffuse type (38.9%) and the intestinal type (45%). A statistically significant association between mdm2 gene amplification and mdm2 protein expression could not be demonstrated. Mdm2 protein was expressed in eight tumors, without a gene amplification being present. Such a constellation was also partly observed in a series of breast carcinomas investigated by Marchetti et al. (14). Cordon-Cardo et al. (32) also described similar phenomena in soft tissue tumors. This could be ascribed to an autoregulative activation of mdm2 expression by p53 (7). In seven cases, the mdm2 gene was amplified, but there was no expression of the mdm2 protein. Cordon-Cardo and co-workers (32) found no protein in 6 of 11 soft tissue tumors with mdm2 amplification. This could be explained by an alteration in the amplified gene that prevents transcription. In this connection, Marchetti et al. (14) reported that the antibodies used may recognize only different epitopes so that immunohistologic evidence is not always possible.

In nine tumors (20.9%), p53 gene mutations were detected by the SSCP technique, followed by direct sequencing. Concerning the tumor type, the distribution of p53 alterations is in contrast to that of mdm2 amplification. Only one diffuse-type carcinoma (5.5%) and the small cell carcinoma showed p53 gene mutations. In contrast, 7 of 20 (35%) carcinomas of the intestinal type showed gene alterations. Ranzani et al. (33) also found considerably higher mutation rates in intestinal-type tumors.

p53 protein overexpression was seen in 12 of the 20 intestinal-type tumors. In two carcinomas of the mixed type, p53 was demonstrated in their glandular component. Only 22.2% of the diffusely growing carcinoma showed protein expression. The markedly higher expression of the p53 protein in the intestinal type is in accordance with the results obtained by Fléjou et al. (26).

Because only approximately 20% of the gastric carcinomas show mdm2/p53 alterations neither at the genetic nor at the posttranslational level, we can draw the following conclusions: (1) Alterations in the mdm2/p53 pathway are a part of the pathogenesis of gastric carcinoma. (2) Tumors with a diffuse growth pattern in particular are frequently accompanied by an amplification of the mdm2 oncogene during genesis, whereas p53 alterations are expected to occur in intestinal-type carcinomas. The question as to which stage of tumorigenesis allows mdm2 alterations to gain importance remains to be answered in further investigations.