Int. J. Cancer: 93, 832– 838 (2001) © 2001 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
EXPRESSION PROFILING OF GASTRIC ADENOCARCINOMA USING cDNA ARRAY Wa’el EL-RIFAI,1,3 Henry F. FRIERSON, JR.,2 Jeffrey C. HARPER,1 Steven M. POWELL1* and Sakari KNUUTILA3 1 Department of Medicine, University of Virginia Health Systems, Charlottesville, VA, USA 2 Department of Pathology, University of Virginia Health Systems, Charlottesville, VA, USA 3 Department of Medical Genetics, Haartman Institute and Helsinki University Central Hospital, University of Helsinki, Helsinki, Finland To investigate the expression profile of gastric adenocarcinoma, cDNA array experiments were performed using Atlas Human Cancer 1.2 K Array (Clontech Laboratories, Palo Alto, CA) on nine xenografted and two primary gastric cancer samples. The expression of the tumor samples was compared to that of two normal gastric epithelial tissues. The expression pattern of the primary tumors was similar to that of xenografted tumors. The up-regulated genes had expression ratios ranging from 2.5 to 16, whereas the down-regulated genes had a range from ⴚ2.5 to ⴚ16. No variation in gene expression was detected in the analysis of the xenografted tumors versus the primary tumors, indicating that the xenografts represented the primary tumors well. Thirtyeight genes showed altered gene expression in 5 or more samples (>45%). Thirty-one genes were up-regulated and seven genes were down-regulated. The most abundantly upregulated genes (ratio >5) included genes such as S100A4, CDK4, MMP14 and beta catenin. The GIF was markedly downregulated (ratio < ⴚ10). To confirm our findings, six genes (three up- and three down-regulated) were selected for semiquantitative RT-PCR analysis. The RT-PCR results were consistent with the array findings. Our approach revealed that several genes are abnormally expressed in gastric cancer and found that genes known to interact in several common molecular pathway(s) were consistently altered. © 2001 Wiley-Liss, Inc. Key words: cDNA arrays; gene expression; xenografts; and gastric adenocarcinomas
Gastric carcinoma (GC) is one of the most common malignancies worldwide and is the second most common cause of cancerrelated death.1 Overall relative 5-year survival rates are currently less than 20%. Cytogenetic studies of gastric adenocarcinomas are few in number and have failed to identify any consistent or noteworthy chromosomal abnormalities.2 Comprehensive studies of DNA copy number changes using comparative genomic hybridization (CGH) have revealed alterations in several chromosomal regions.3–5 Frequent high-level amplifications were noted at 17q and 20q, whereas DNA copy number losses were frequently seen in 4q, 5q, and 9p. However, there are few data on gene expression in gastric cancer. The recent development of cDNA array technology allows the study of gene expression level and gene activation in thousands of genes and sequences.6,7 The data obtained from the array analysis are expected to uncover novel genes related to cancer development and its clinical behavior. However, tumor sample contamination with normal cells makes analysis of hundreds of genes more difficult and less sensitive. Xenografting of human tumors has been used to produce optimal samples that are enriched for neoplastic cells for subsequent molecular analyses. Studies of xenografted tumors generated from human colon adenocarcinomas have led to the discovery of important genetic alterations underlying these malignancies.8 To explore the gene expression pattern in gastric adenocarcinoma, we have applied cDNA array technology to study alterations in the expression profiles of xenografted gastric cancer samples and validated the results on two primary gastric cancers. MATERIAL AND METHODS
Samples Nine xenografted and two primary gastric cancer samples were tested for gene expression using cDNA array hybridization anal-
yses. For a normal reference, two normal gastric epithelial tissue samples, obtained from the same patients that had the primary tumors, were used to evaluate the normal expression profile of gastric epithelium. The reference samples were enriched for the epithelial layer of the stomach through microdissection and mucosal scraping of freshly resected gastric tissue. Xenografts A consecutive series of resection specimens was processed to generate the xenografts according to IRB approved protocols. Fresh tissue from surgically resected adenocarcinomas of the stomach or gastroesophageal junction resected at two academic institutions, Indiana University Medical Center and University of Virginia Health Sciences Center, were collected in RPMI media containing penicillin G (Sigma, St. Louis, MO, 180 g/ml) and streptomycin (Sigma, 300 g/ml) and stored temporarily on ice until further processing for xenografting as previously described.9 In brief, three small pieces of tumor tissue were soaked in Matrigel and then implanted subcutaneously in immunodeficient mice. These mice were then observed for tumor growth and harvested after reaching approximately 1 to 1.5 cm in size. An adjacent piece of tissue from the resected primary tumor specimen was embedded and used to confirm the histology of the xenografted sample. Histologic confirmation of the xenograft tumors was performed by cryostat sectioning of the tumor and staining the slides with hematoxylin and eosin. Our gastrointestinal pathologist (H.F.F.) did all histologic classifications. cDNA array hybridization Total RNA was extracted from all samples using a Qiagen RNA maxiprep kit (Qiagen, Hilden, Germany) following the procedure recommended by manufacturer. The total RNA was treated using a Qiagen RNase-Free DNase set to eliminate any contaminating DNA. The integrity of the RNA was checked on 1% agarose gel electrophoresis. Atlas Human Cancer 1.2 K filters (Clontech, Palo Alto, CA), each containing 1174 genes (for information about genes, see www.clontech.com) were used. The total RNA (3 to 4 g) was converted into cDNA and labeled with 33PdATP using Clontech cDNA array labeling kit. Hybridizations and washes were performed following the recommended manufacturer protocol with minor modifications. Following the washes, the filters were exposed to imaging plates (BAS-MP 2040S; Fuji, Kanagawa, Japan) and scanned with a phosphoimager (Bio-Imaging Analyzer, BAS-2500; Fuji) to obtain high-resolution (16-bit) tiff-format imGrant sponsor: Helsinki University Central Hospital Research Funds; Grant sponsor: Finnish Cancer Institute Foundation: Grant sponsor: NIH; Grant number: CA67900. *Correspondence to: University of Virginia Health Systems, Box 800708, Charlottesville, VA 22908-0708, USA. Fax: ⫹804-243-6169. E-mail:
[email protected] Received 9 October 2000; Revised 15 December 2000; Accepted 29 January 2001 Published online 2 April 2001
833
cDNA ARRAY IN GASTRIC CANCER
ages. The images were imported and analyzed using Atlas image analysis software (Clontech). Validation, thresholds and controls for the cDNA array analysis During printing of arrays, inconsistencies in printing can occur from one lot to another. These variations can be in the form of size, regularity or stickiness of the spot and can affect the analysis outcome. Therefore, filters included in our experiments were obtained from the same (printing) lot to minimize printing variations between different lots. To ensure optimum quality of the hybridizations, none of the filters was used more than three times. Four tumor samples were hybridized twice, each in a different filter in a different experiment to establish the cut-off values and to confirm that the results were repeatable. The two independent hybridization images for the same sample were then compared to each other using Atlas Image software (Clontech). To obtain an expression reference value, we performed five independent hybridizations of the two reference samples. Duplicate hybridizations were done for each reference sample using two identical filters. A fifth hybridization was performed using an RNA pool of the two reference samples. The five hybridization images that were obtained from our reference samples were then pooled into a single composite array image and the intensities of the spots were averaged using an Atlas image. This averaged composite array was used for further comparison for the xenografts. The primary tumor samples were each compared with the corresponding reference sample. To avoid misinterpretation of the results that could result from variation in the hybridizations, any two compared filters were normalized using nine housekeeping genes, and a normalization coefficient calculated for each comparison was used to correct the signal intensities. Our analysis of any two hybridizations obtained from a given sample showed no differences using a ratio of 1.5 for over-expression and ⫺1.5 for under-expression. However, the threshold values we applied were stricter: 2.5 for over-expression and ⫺2.5 for under-expression. In addition, if the intensity of any signal failed to be at least three times above the background value, it was discarded from our analysis. RT-PCR To confirm the cDNA array results, semi-quantitative RT-PCR was performed for six genes displaying expression alterations: three genes that showed up-regulation (S100A4, GRB2 and JUP) and three that showed down-regulation (TIMP-3, DR-nm23 and EGR1). Primer sequences were obtained from Clontech. Reverse transcription was accomplished using 0.8 g of total RNA and reagents from an Advantage RT-for-PCR Kit (Clontech). In each PCR reaction, primers for the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene were used as an internal reference. GAPDH, which is considered a housekeeping gene, showed similar expression levels by cDNA array analysis for the reference and tumor samples. The PCR reactions were performed using a standard protocol. After 20 cycles of amplification, only 2 l of each PCR product was subjected to electrophoresis in 6% denaturing polyacrylamide sequencing gel (National Diagnostics, Atlanta, GA) and was visualized by silver staining. The gels were scanned using UMAX Power Look 2000 (UMAX Data Systems, Hsinchu, Taiwan). The ISIS image analysis system (Metasystems, Altlussheim, Germany) was used for quantification of the results and determination of the signal intensity for each band. The intensity of each tumor sample was compared to the reference sample intensity. Normalization of the results was done using the GAPDH intensities in the reference and tumor samples, and the normalization coefficient was calculated. After normalization, the adjusted intensities were calculated for the amplified gene products, and the ratios were calculated. We used cutoff values of ⫺2.0 and ⫹2.0 for under- and over-expression, respectively. We confirmed the reproducibility of the method by repeating the RT-PCR twice for three genes and the results were consistent.
RESULTS
We explored altered gene expression patterns in two primary gastric cancers, one intestinal type and one diffuse type, and nine additional xenografted human stomach cancers. These cancers included all stages of development (TNM stages I to III), histopathology (intestinal and diffuse types), grade (well-differentiated [WD] to poorly differentiated [PD]) and location (gastroesophageal junction to antrum; Table I). Consistent patterns of altered gene expression were observed in these gastric carcinomas. However, the diffuse gastric cancer sample had fewer alterations than any of the intestinal carcinomas. Tumors arising more proximally at the gastroesophageal junction showed fewer changes than more distal (body and antral) tumors (Fig. 1). We were not able to assess differences in expression pattern between tumors of different histopathology because of the small number of cases. Additional samples and studies are required to demonstrate consistent alterations between subgroups of gastric cancers. Thirty-eight genes had altered expression changes in five or more of the tumors (⬎45%). Compared with the expression of the normal references pool, the up-regulated genes had ratios ranging from 2.5 to 16, whereas the down-regulated genes had a range from ⫺2.5 to ⫺16. Thirty-one genes were up-regulated and seven genes were down-regulated (Table II, Fig. 1). No variation in the expression pattern was detected in the analysis of the pooled primary tumors versus the pooled xenograft tumors, indicating that xenografts are ideal samples for expression studies as they represent the primary tumors well. Several genes were, for the first time, shown in gastric cancer (Table II, Fig. 1). The most abundantly up-regulated genes (ratio ⬎5) included genes such as S100A4, CDK4, MMP14 and beta-catenin. The GIF was markedly downregulated (ratio ⬍ ⫺10). Furthermore, we were able to show expression alterations in genes that have been suggested to interact in common molecular pathways. The semi-quantitative RT-PCR results, for the six genes we tested, were consistent with the array findings as shown in Figure 2. DISCUSSION
We have explored the gene expression profile in nine xenografted gastric cancer samples using a 1.2 K array containing genes that were classified according to their function to be relevant for cancer. We validated the expression results of the xenografted tumors using cDNA arrays on two primary gastric cancer samples and confirmed the expression results using RT-PCR on all samples. Genes that were found to be in more than five tumors (⬎45%) are presented, as they may contribute to gastric cancer tumorigenesis. Over-expression of several genes involved in cell cycle control (Cyclin D1, CDK4, CKS1 and CKS2) and others often used as proliferation markers (nuclear CENP-F, IMPDH2, PCNA, KI-67, CAF-1 and TOP2A) additionally support our results. TABLE I – HISTOPATHOLOGY OF GASTRIC ADENOCARCINOMAS1 Sample no./ code
Age/sex
Site
1/X-12 2/X-25 3/X-26 4/X-27 5/X-62 6/G208 7/X-30 8/X-32 9/X-37 10/X65 11/G204
41/F 69/F 67/M 50/M 69/M 60/M 70/M 57/M 67/F 45/M 72/F
GEJ GEJ GEJ GEJ GEJ GEJ Body Body Body Antrum Body
Histopathology
Grade
TNM stage
Intestinal Intestinal Intestinal Intestinal Intestinal Intestinal Intestinal Intestinal Intestinal Intestinal Diffuse/ Mixed
WD PD MD PD MD WD PD MD MD PD PD
IIIb, T3N2M0 IIIb, T3N2M0 IIIa, T3N1M0 IIIa, T3N1M0 III, T3N1M0 II, T2N1M0 Ib, T2N0M0 Ia, T1N0M0 Ib, T1N0M0 IIIa, T3N1M0 Ib, T2N0M0
1 Samples 1–5 and 7–10 are xenografted tumors. Samples 6 and 11 are primary tumors. M, male; F, female; GEJ, gastroesophageal junction; WD, well-differentiated; PD, poorly differentiated; MD, moderately differentiated.
834
EL-RIFAI ET AL.
FIGURE 1 – A summary image displaying frequent expression alterations, (⬎5 samples), shown in xenografted gastric cancer samples (X), primary gastric tumors (P), pool of xenografted tumors, and pool of primary tumors. Up-regulated genes are shown in red and down-regulated genes in green. The brightness of the color correlated with the degree of expression as shown at the bottom of the image. The gray color (0) indicates genes that had expression values within the thresholds (⫺2.5 to 2.5). Names of the genes are shown on the right-hand side. This image was created using Tree view software written by Michael Eisen, © 1998 –9, Stanford University.
Genes related to cell adhesion and invasion and growth factors Several genes related to cell adhesion and invasion and growth factors were abundantly over-expressed in our samples. S100A4 expression alters the adhesive properties of cells, possibly by remodeling the extracellular matrix. S100A4 is expressed in several tumors including colon cancer.10 However, there are no reports related to S100A4 expression in gastric cancer. S100A4 may exert its effect on metastasis formation not only by stimulating the motility of tumor cells but also by affecting their invasive properties through deregulation of matrix metalloproteinases (MMP)
and tissue inhibitors of matrix metalloproteinases (TIMPs).11 The family of MMPs has been shown to be involved in proteolytic degradation of the extracellular matrix. We demonstrated over-expression of S100A4 and MMP-14 and under-expression of TIMP-3. Furthermore, we demonstrated overexpression of pM5, which is a metalloproteinase-related collagenase. Although several members of MMPs have been shown to be over-expressed in gastric cancer,12 expression of MMP14, TIMP3, pM5 or S100A4, to our knowledge, has not been reported before in gastric cancer.
835
cDNA ARRAY IN GASTRIC CANCER TABLE II – EXPRESSION CHANGES DETECTED IN FIVE OR MORE GASTRIC CANCER SAMPLES (⬎45%) USING cDNA ARRAY Gene name
GenBank accession nos.
Growth inhibitory factor (GIF); D13365; metallothionein-III (MT-III) M93311 Tissue inhibitor of Z30183 metalloproteinases-3 (TIMP3) Placental calcium-binding M80563 protein; S100A4 Hepatoma-derived growth D16431 factor (HDGF) Low-density lipoprotein X13916 receptor-related protein 1 precursor (LRP) Growth factor receptor-bound L29511; protein 2 (GRB2) isoform M96995 DR-nm23 U29656 Matrix metalloproteinase-14 D26512; precursor (MMP-14) X83535 EB1 protein U24166 Junction plakoglobin (JUP); M23410; ␥-cetenin Z68228 Bikunin; hepatocyte growth U78095 factor activator inhibitor 2 Antigen KI-67 (MKI67) X65550 CENP-F kinetochore protein U19769 High-mobility group protein X62534 HMG2 Hint protein; protein kinase C U51004 inhibitor 1 (PKCI1) Cyclin-dependent kinase X54941 regulatory subunit 1 (CKS1) BENE U17077 Proliferating cyclic nuclear M15796; antigen (PCNA) J04718 PM5 protein X57398 Chromatin assembly factor 1 X74262 p48 subunit (CAF1 p48 subunit) Integrin alpha 3 (ITGA3) M59911 BTG protein precursor U72649 Insulin-like growth factor M35410 binding protein 2 (IGFBP2) Cyclin-dependent kinase 4 M14505 (CDK4) Ribonucleoside-diphosphate X59618 reductase M2 subunit -catenin (CTNNB) X87838; Z19054 MCM4 DNA replication X74794 licensing factor; CDC21 homolog Inosine-5⬘-monophosphate L33842 dehydrogenase 2 (IMP dehydrogenase 2; IMPD2) Early growth response protein X52541; 1 (EGR1) M62829 Mesothelin precursor; CAK1 U40434 antigen DNA topoisomerase II alpha J04088 (TOP2A) Cyclin-dependent kinase X54942 regulatory subunit (CKS2) G1/S-specific cyclin D1 X59798; (CCND1); cyclin PRAD1; M64349 bcl-1 oncogene Paxillin U14588 Mitogen-activated protein L35253; kinase p38 (MAP kinase L35263 p38) ERBB-3 receptor proteinM29366; tyrosine kinase precursor M34309 Vascular endothelial growth M32977; factor precursor (VEGF) M27281 Checkpoint suppressor 1 U68723 1
Gene expression ratios1 X1
X2
X3
X4
X5
P6
X7
X8
X9
X10
P11
⫺14.2 ⫺12.5 ⫺16.6 ⫺12.5⫺14.2 ⫺7.4 ⫺4.3 ⫺16.6 ⫺16.6 ⫺5.1 ⫺7.1 ⫺9.1 ⫺8.3 ⫺3.1 23.3
26.4
15.9
4.9
6.1
5.6
4.8
7.9
3.9
4.1
5.1
6.5
4.6
3.9
⫺6.6 ⫺7.6 ⫺8.3 5.6
0.25 ⫺2.5 ⫺3.4 ⫺10.1 ⫺6.2 ⫺5.6
18.8 19.1 4.8
2.9 5.9
5.9
7.2
7.6
20.7
4.3
5.6
11.4
7.1
4.9
5.9
7.4
7.3
4.7
4.1
5.3
4.5
⫺4.5 ⫺4.0 7.8 6.9
9.5
⫺10.3
10
⫺5.1
10
16
9 9
4.6
4.9
9
3.8
4.5
9
⫺5.2 5.9
9 8
4.3 3.8
8 8
2.5
8
10 9.7 8.5
7 7 7
7.6
7
7.5
7
4.7 3.7
3.7 4.9
3.3
3.9 4.1
3.3
4.1 3.7
4.7 3.5
2.5
2.2
2.6
2.5
2.7
2.1
2.2
8.8 9.8 5.2
10.4 12.8 7.1
11.4 10.9 9.9
8.7 10.8 8.5 11.1 8.7 8.8
7.2
6.1
10.9
6.3
7.2
9.3
6.7
10.6
5.9
6.9 5.5
8.8 7.5
6.7
6.5 6.5
7.3 6.3
7 7
6.2 5.3
4.9 4.7
6.8 5.1
6.2 4.3
4.5
5.5 5.1
7 7
6.1
8.1 ⫺3.2 ⫺2.9
7.4 ⫺2.5 ⫺5.3
⫺3.7 ⫺2.5 ⫺4.2
4.9 ⫺3.0 ⫺5.0
7 7 7
10.8
10.4
13.3
5.1
9.8
6
8.1
10.3
12.6
9.3
6
5.6
9.4
9.8
8.4
6
6.4
5.6
6.5
6.3
6
3.1
2.9
6
⫺10.1 ⫺2.8 ⫺4.3
⫺3.3
6
11
5
10
5
13.5 8.9 11.6 10.2
6.7
8.7
7.6
10.9 7.1
6.8 5.1
4.7 6.2
5.2
6.1
5.4 4.7
5.7 4.9
9.7 7.8 8.9 7.1 ⫺3.7 ⫺2.9 ⫺2.5 ⫺3.4 ⫺4.0 ⫺3.6 ⫺5.6 8.1 9.1
9.3
8.7
8.7
7.5
6.1
2.7
2.3
⫺6.6 9.9
5.6 4.7 ⫺3.3
10.1
4.4
3.2 8.2 5.6 8.4
2.3
11.1 6.7
2.4
⫺2.5
4.1 3.6
No. of abnormal cases
6.1
⫺6.6 ⫺5.2 ⫺6.6 ⫺2.6 6.3 6.8 2.6 5.1
Average ratio
8.1
2.5
11.1
10.6
14.2
16.2
7.1
11.7
7.8
8.1
10.1
7.4
11.6
8.3
7.9
6.2 7.1 4.9
5.6
3.6 3.2
2.8
6.9 3.1 ⫺3.1 ⫺4.3 ⫺3.2
9.1
5
6.1
6.6
5.8
5.3
5.9
5
4.5 4.1
5.7 5.4
5.1 7.1
5.7
5.6 5.4
5 5
5.6
3.5
4.4
4.8
5
2.7
3.8
3.7
5
⫺4.1
5
⫺4.0 ⫺6.6
Empty entries indicate that the expression was below analysis thresholds (⫺2.5 to 2.5); X, xenografted tumor; P, primary tumor.
836
EL-RIFAI ET AL.
FIGURE 2 – RT-PCR results for the six tested genes in the 11 tumor samples (1 to 11) and the reference samples (R). The left panel shows three under-expressed genes (TIMP3, EGR1 and DRnm23); over-expressed genes (S100A4, JUP and GRB2) are in the right panel. The upper PCR bands for each gene represent the control gene (GADPH) expression and the lower bands represent the test gene expression. Status of gene expression: ⫺, under-expression; ⫹, over-expression; N, normal expression. Comparison of Array/PCR expression is presented below each sample.
LRP, which is over-expressed in our tumors, is considered essential in cellular migration and invasive processes through mediation of cellular uptake and clearance of inactivated proteaseinhibitor complexes, which regulates proteinase activity at the cell surface. LRP over-expression has been reported in glioblastomas and neuroblastomas.13 To our knowledge, this is the first report to demonstrate the over-expression of LRP in gastric cancer. In addition to the possible role of the aforementioned genes in tumor metastasis and progression, DRnm23, a metastasis suppressor gene frequently under-expressed in cancer, was down-regulated in our tumors. Genes related to extracellular matrix Integrins are major adhesion- and signaling-receptor proteins that mediate cell migration and invasion. Focal adhesion kinase has recently been established as a key component of the multiple signal transduction pathways that are mediated by GRB2 (triggered by integrin stimulation) and subsequent activation of paxillin and MAP kinase.14 Our tumors showed up-regulation of integrin alpha-3, GRB2, paxillin, p38 MAP kinase, c-ERBB3 and VEGF. GRB2 is a widely expressed protein that plays a crucial role in activation of several other growth factors. GRB2 may mediate transmission of ERBB2 oncogenic signals, which in turn activate the mitogen-activated protein (MAP) kinase pathway.15 We demonstrated over-expression of c-ERBB3 (closely related to ERBB2), which belongs to the type 1 family of growth factor receptors with intrinsic tyrosine kinase activity. ERBB3 has been observed to be over-expressed in several tumors including gastric cancer cell lines.16 The p38 MAP kinase-dependent signal transduction pathway(s) plays a role in serine phosphorylation and disassembly of paxillin.17 Expression of VEGF, a multifunctional cytokine, is crucial in angiogenesis during the development of cancer and correlates with disease recurrence in patients with early gastric carcinoma.18 p38 MAP kinase activation by VEGF mediates actin reorganization and cell migration in human endothelial cells.19 By modulating cell migration, p38 may thus be an important regulator of angiogenesis. Genes related to cadherin-catenin complex Over-expression of two components of the cadherin-catenin complex, beta-catenin and gamma-catenin, were detected in our
tumors. Beta-catenin is a multifunctional protein that is both an integral component of adherens junctions and a pivotal member of the Wnt signal transduction pathway. Mutations and over-expression of beta-catenin disrupt the function of E-cadherins. The increased level of beta-catenin through mutations in either betacatenin or adenomatous polyposis coli (APC) has been suggested to be an important oncogenic step in the genesis of a number of malignancies, and it is an early event in colorectal carcinogenesis.20 APC mutations are more frequent in colorectal carcinomas than in gastric carcinomas, which explains the absence of APC expression alterations in our tumors. The over-expression of betacatenin that we demonstrated may be generated by either activating mutations of beta-catenin or, alternatively, by other regulatory factors (e.g., transregulated). Our tumors displayed over-expression of cyclin D1, which is known to be regulated by betacatenin.21 In addition, EB1, a component of the microtubule cytoskeleton that may play a physiologic role connecting APC to cellular division and coordinating the control of normal growth and differentiation processes in the colonic epithelium,22 was over-expressed in our tumors. EB1 over-expression has not been reported in gastric cancer until now. The biological mechanisms of the cadherin-catenin complex and EB1 and their connection to APC in the development of gastric cancer require further investigation. Genes related to growth regulation/apoptosis Among the seven down-regulated genes, several genes were novel for gastric cancer. GIF, which is a hydrophilic metallothionein (MT)-like protein, has been shown to be drastically downregulated in Alzheimer’s disease brains23 and was markedly downregulated in our tumors (ratio ⬍ ⫺10). However, GIF expression has not been studied in human cancer before, and the molecular mechanisms by which GIF may act as a tumor suppressor gene remain to be elucidated. In addition to GIF, EGR-1 functions in growth regulation and suppression of transformation and stimulates apoptosis by transactivation of the p53 gene.24 Loss of Egr-1 expression is a novel finding for gastric cancer. BTG2, which was down-regulated in our tumors, is another gene with p53-dependent expression that may be relevant to cell cycle control and cellular response to DNA damage.25 CHES1 is a member of the fork head/Winged Helix family of transcription factors, which sup-
837
cDNA ARRAY IN GASTRIC CANCER
presses a number of DNA damage-activated checkpoint mutations in Saccharomyces cerevisiae.26 Recently, mutations in checkpoint kinase 2 (CHK2), whose activation in response to DNA damage prevents cellular entry to mitosis, has been suggested as a tumor suppressor gene conferring predisposition to sarcomas, breast cancer and brain tumors in Li-Fraumeni syndrome.27 The downregulation of CHES1 in our tumors contributes to the accumulating data of altered checkpoint genes in cancers. CHES1 may be a novel mutated gene for human gastric cancer that requires further analysis. Genes with other functions Despite the potential carcinogenic role of over-expression of HDGF, mesothelin, RRM2 and bikunin (KOP); these genes have only been studied in human cancers in a limited fashion. HDGF is an endothelial cell growth factor secreted by hepatoma cell lines.28 The mesothelin is a differentiation antigen strongly present on mesotheliomas and ovarian cancers and may affect cellular adhesion.29 The RRM2 supplies deoxyribonucleotide precursors to intranuclear replication enzymes, and its over-expression seems to play an important role in tumor invasiveness.30 The KOP gene encodes a novel putative transmembrane protein with two Kunitztype serine protease inhibitor domains and may participate in tumor cell invasion and metastasis in human pancreatic cancer tissues.31 The biochemical function of these genes in gastric cancer remains to be investigated. HMG2, PKCI-1, BENE and MCM4 are genes that were overexpressed and have not been previously demonstrated in gastric cancer. HMG2, a member of the high-mobility group protein
family, is a non-sequence-specific DNA binding protein that recognizes DNA structure and acts as a co-regulator that increases the DNA binding and transcriptional activity of the steroid hormone class of receptors.32 PKCI-1, a member of the HIT family of proteins, is expressed at relatively high levels in several murine tissues and in a variety of human cell lines prepared from normal tissues or tumors. In vivo studies suggest that the ubiquitously expressed PKCI protein does not function as an inhibitor of PKC but rather acts as an enzyme in a yet to be identified pathway.33 BENE belongs to the MAL family, which is widely expressed in various tissues, suggesting a common role of these proteins in cell biology.34 MCM4, a DNA-replication-initiation factor, is an essential component of a protein complex that prevents DNA from being replicated more than once per cell cycle.35 To our knowledge, neither BENE nor MCM4 genes have been previously studied in cancer. CONCLUSIONS
cDNA array technology is a powerful tool to explore gene expression in cancer. Based on the literature and our findings, the interactions of these genes, altered in our samples, seem to be important events for the development and/or progression of gastric cancer. Further studies should clarify the mechanism through which these different molecular pathways lead to gastric cancer when they are altered. Establishment of the diagnostic and/or prognostic value of these alterations and exploration of the biologic mechanisms underlying their functions in tumor growth are important issues to be addressed in combating this lethal cancer.
REFERENCES
1. 2. 3. 4.
5.
6. 7. 8. 9. 10. 11.
12. 13.
14. 15.
Whelan S, Parkin D, Masuyer E. Trends in cancer incidence and mortality, IARC Scientific Publ., No. 102. Lyon, France: IARC, 1993. Mitelman F. Catalog of chromosome aberrations in cancer, 4th ed. New York: Wiley-Liss, 1991. p 1987. El-Rifai W, Harper JC, Cummings OW, et al. Consistent genetic alterations in xenografts of proximal stomach and gastro-esophageal junction adenocarcinomas. Cancer Res 1998;58:34 –7. Kokkola A, Monni O, Puolakkainen P, et al. 17q12-21 amplicon, a novel recurrent genetic change in intestinal type of gastric carcinoma: a comparative genomic hybridization study. Genes Chromosomes Cancer 1997;20:38 – 43. Larramendy L, El-Rifai W, Kokkola A, et al. Comparative genomic hybridization reveals differences in DNA copy number changes between sporadic gastric carcinoma and gastric carcinomas from patients with hereditary nonpolyposis colorectal cancer. Cancer Genet Cytogenet 1998;106:62–5. DeRisi J, Penland L, Brown PO, et al. Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat Genet 1996; 14:457– 60. Duggan DJ, Bittner M, Chen Y, Meltzer P, Trent JM. Expression profiling using cDNA microarrays. Nat Genet 1998;2110 – 4. Thiagalingam S, Lengauer C, Leach F, et al. Evaluation of candidate tumour suppressor genes on chromosome 18 in colorectal cancers. Nat Genet 1996;13:343– 6. Hahn SA, Seymour AB, Hoque ATMS, et al. Allelotype of pancreatic adenocarcinoma using xengraft model. Cancer Res 1995;55:4670 –5. Velculescu VE, Madden SL, Zhang L, et al. Analysis of human transcriptomes. Nat Genet 1999;23:387– 8. Bjornland K, Winberg JO, Odegaard OT, et al. S100A4 involvement in metastasis: deregulation of matrix metalloproteinases and tissue inhibitors of matrix metalloproteinases in osteosarcoma cells transfected with an anti-S100A4 ribozyme. Cancer Res 1999;59:4702– 8. Allgayer H, Babic R, Beyer BC, et al. Prognostic relevance of MMP-2 (72-kD collagenase IV) in gastric cancer. Oncology 1998;55:152– 60. Yamamoto M, Ikeda K, Ohshima K, Tsugu H, Kimura H, Tomonaga M. Increased expression of low density lipoprotein receptor-related protein/alpha2-macroglobulin receptor in human malignant astrocytomas. Cancer Res 1997;57:2799 – 805. Guan JL. Role of focal adhesion kinase in integrin signaling. Int J Biochem Cell Biol 1997;29:1085–96. Tari AM, Hung MC, Li K, Lopez-Berestein G. Growth inhibition of breast cancer cells by Grb2 downregulation is correlated with inactivation of mitogen-activated protein kinase in EGFR, but not in ErbB2, cells. Oncogene 1999;18:1325–32.
16. Noguchi H, Sakamoto C, Wada K, et al. Expression of heregulin alpha, erbB2, and erbB3 and their influences on proliferation of gastric epithelial cells. Gastroenterology 1999;117:1119 –27. 17. Vadlamudi R, Adam L, Talukder A, Mendelsohn J, Kumar R. Serine phosphorylation of paxillin by heregulin-beta1: role of p38 mitogen activated protein kinase. Oncogene 1999;18:7253– 64. 18. Maeda K, Kang SM, Onoda N, et al. Vascular endothelial growth factor expression in preoperative biopsy specimens correlates with disease recurrence in patients with early gastric carcinoma. Cancer 1999;86:566 –71. 19. Rousseau S, Houle F, Landry J, Huot J. p38 MAP kinase activation by vascular endothelial growth factor mediates actin reorganization and cell migration in human endothelial cells. Oncogene 1997;15:2169 – 77. 20. Damalas A, Ben-Ze’ev A, Simcha I, et al. Excess beta-catenin promotes accumulation of transcriptionally active p53. EMBO J 1999; 18:3054 – 63. 21. Tetsu O, McCormick F. Beta-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 1999;398:422– 6. 22. Berrueta L, Kraeft SK, Tirnauer JS, et al. The adenomatous polyposis coli-binding protein EB1 is associated with cytoplasmic and spindle microtubules. Proc Natl Acad Sci USA 1998;95:10596 – 601. 23. Hozumi I, Inuzuka T, Tsuji S. Brain injury and growth inhibitory factor (GIF)—a minireview. Neurochem Res 1998;23:319 –28. 24. Liu C, Rangnekar VM, Adamson E, Mercola D. Suppression of growth and transformation and induction of apoptosis by EGR-1. Cancer Gene Ther 1998;5:3–28. 25. Rouault JP, Falette N, Guehenneux F, et al. Identification of BTG2, an antiproliferative p53-dependent component of the DNA damage cellular response pathway. Nat Genet 1996;14:482– 6. 26. Pati D, Keller C, Groudine M, Plon S. Reconstitution of a MEC1independent checkpoint in yeast by expression of a novel human fork head cDNA. Mol Cell Biol 1997;17:3037– 46. 27. Bell DW, Varley JM, Szydlo TE, et al. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 1999; 286:2528 –31. 28. Nakamura H, Izumoto Y, Kambe H, et al. Molecular cloning of complementary DNA for a novel human hepatoma- derived growth factor. Its homology with high mobility group-1 protein. J Biol Chem 1994;269:25143–9. 29. Chang K, Pastan I. Molecular cloning of mesothelin, a differentiation antigen present on mesothelium, mesotheliomas, and ovarian cancers. Proc Natl Acad Sci USA 1996;93:136 – 40. 30. Zhou BS, Tsai P, Ker R, et al. Overexpression of transfected human ribonucleotide reductase M2 subunit in human cancer cells enhances their invasive potential. Clin Exp Metastasis 1998;16:43–9.
838
EL-RIFAI ET AL.
31. Muller-Pillasch F, Wallrapp C, Bartels K, et al. Cloning of a new Kunitz-type protease inhibitor with a putative transmembrane domain overexpressed in pancreatic cancer. Biochim Biophys Acta 1998; 1395:88 –95. 32. Boonyaratanakornkit V, Melvin V, Prendergast P, et al. High-mobility group chromatin proteins 1 and 2 functionally interact with steroid hormone receptors to enhance their DNA binding in vitro and transcriptional activity in mammalian cells. Mol Cell Biol 1998;18:4471– 87. 33. Klein MG, Yao Y, Slosberg ED, Lima CD, Doki Y, Weinstein IB.
Characterization of PKCI and comparative studies with FHIT, related members of the HIT protein family. Exp Cell Res 1998; 244:26 –32. 34. Magyar JP, Ebensperger C, Schaeren-Wiemers N, Suter U. Myelin and lymphocyte protein (MAL/MVP17/VIP17) and plasmolipin are members of an extended gene family. Gene 1997;189:269 –75. 35. Fujita M, Yamada C, Tsurumi T, Hanaoka F, Matsuzawa K, Inagaki M. Cell cycle- and chromatin binding state-dependent phosphorylation of human MCM heterohexameric complexes. A role for cdc2 kinase. J Biol Chem 1998;273:17095–101.
