New Insights Into Testicular Germ Cell Tumorigenesis from Gene Expression Profiling1

[CANCER RESEARCH 62, 2359–2364, April 15, 2002] New Insights into Testicular Germ Cell Tumorigenesis from Gene Expression Profiling1

Rolf I. Skotheim, Outi Monni, Spyro Mousses, Sophie D. Fosså, Olli-P. Kallioniemi, Ragnhild A. Lothe,2 and Anne Kallioniemi Department of Genetics, Institute for Cancer Research [R. I. S., R. A. L.] and Department for Oncology and Radiotherapy [S. D. F.], The Norwegian Radium Hospital, N-0310 Oslo, Norway; Biomedicum Biochip Center, Biomedicum Helsinki, FIN-00290 Helsinki, Finland [O. M.]; Cancer Genetics Branch, National Human Genome Research Institute, NIH, Bethesda, Maryland 20892 [O. M., S. M., O-P. K., A. K.]; and Laboratory of Cancer Genetics, Institute of Medical Technology, University of Tampere and Tampere University Hospital, FIN-33520 Tampere, Finland [A. K.]

ABSTRACT teristic of virtually all TGCTs (4, 5). In addition, specific gains and losses from several other chromosomal regions have been described. 3 We have shown recently that about half of the human TGCTs reveal Although molecular studies have shown some genes to be altered at DNA copy number increases affecting two distinct regions on chromosome the DNA and/or expression levels in a limited number of TGCTs, the arm 17q. To identify potential target genes with elevated expressions attributable to the extra copies, we constructed a cDNA microarray target genes reflecting the nonrandom chromosomal aberrations re- containing 636 genes and expressed sequence tags from chromosome 17. main unknown (for a review of TGCT genetics, see Refs. 6 and 7). The expression patterns of 14 TGCTs, 1 carcinoma in situ, and 3 normal We have demonstrated recently by a genome-wide copy number testis samples were examined, all with known chromosome 17 copy num- analysis using CGH that sequences on chromosome arm 17q are bers. The growth factor receptor-bound protein 7 (GRB7) and junction frequently overrepresented in TGCTs, and two common regions of plakoglobin (JUP) were the two most highly overexpressed genes in the copy number increase were identified (8). Gain of the proximal TGCTs. GRB7 is tightly linked to ERBB2 and is coamplified and coex- region, 17q11–q21, is preferentially observed in nonseminomas, pressed with this gene in several cancer types. Interestingly, the expression whereas gain of the distal region, 17q24–qter, is common to all levels of ERBB2 were not elevated in the TGCTs, suggesting that GRB7 TGCTs (8). Nonrandom gain at 17q has also been reported in several might be the target for the increased DNA copy number in TGCTs. other cancer types (9–13). Because of the limited knowledge of altered gene expression in the development of TGCTs, we also examined the expression levels of 512 addi- To identify differentially expressed genes on chromosome 17 in tional genes located throughout the genome. Several genes novel to tes- TGCTs, we analyzed a series of TGCTs and normal testicular samples ticular tumorigenesis were consistently up- or down-regulated, including using a custom-made cDNA microarray with a comprehensive chro- POV1, MYCL1, MYBL2, MXI1, and DNMT2. Additionally, overexpression mosome 17 coverage (14). All analyzed tumors had been studied of the proto-oncogenes CCND2 and MYCN were confirmed from the previously by CGH, and thus, the expression profiles could be related literature. The overexpressions were for some of the target genes closely to the DNA copy numbers. The expression levels of 512 additional associated to either seminoma or nonseminoma TGCTs, and hierarchical genes mapping elsewhere in the genome, including many cancer- cluster analysis of the gene expression data effectively distinguished related genes, were also analyzed in the same set of TGCT samples. among the known histological subtypes. In summary, this focused functional genomic characterization of TGCTs has lead to the identification of new gene targets associated with a common genomic rearrangement as MATERIALS AND METHODS well as other genes with potential importance to testicular tumorigenesis. Tumor and Cell Line Samples. Eighteen testicular tissue samples were analyzed, including 8 pure seminomas, 6 nonseminomas, 1 carcinoma in situ INTRODUCTION (from the vicinity of a nonseminoma), and 3 normal samples. The tumors were selected from a series of primary TGCTs analyzed previously by CGH (8). TGCT3 is the most common malignancy among adolescent and Four of the 8 seminomas had gains at distal 17q, including the 17q24–qter young adult Caucasian males, and the incidence has been steadily region (Fig. 1A). The 6 nonseminomas (4 embryonal carcinomas and 2 im- increasing over the past 50 years (1, 2). TGCTs are classified into two mature teratomas) had large gains at chromosome 17, all including the 17q11– main histological subtypes, seminomas and nonseminomas, and there q21 region. The use of these samples in cDNA microarray experiments was are two models describing their development from carcinomas in situ approved by the Regional Committee for Medical Research Ethics in Norway and the NIH Office of Human Subjects Research. (3). Either both subtypes develop independently from carcinomas in A pool of two breast cancer cell lines, HBL100 and MDA-436 (American situ, or they develop as a continuum where seminomas may progress Type Culture Collection, Manassas, VA), was used as a common reference in further into nonseminomas. the cDNA microarray experiments. These cell lines were selected based on the TGCTs are generally in the triploid range, and isochromosome 12p facts that they show no increase in copy number at 17q and express most genes or gain of DNA sequences from chromosome arm 12p is a charac- on the cDNA microarray to some extent (10, 14). cDNA Microarray Experiments. The construction of the cDNA microar- Received 12/27/01; accepted 2/14/02. ray with comprehensive chromosome 17 coverage has been described previ- The costs of publication of this article were defrayed in part by the payment of page ously by Monni et al. (14). The microarray consisted of printed PCR products charges. This article must therefore be hereby marked advertisement in accordance with from 636 sequence-verified IMAGE cDNA clones (Research Genetics, Hunts- 18 U.S.C. Section 1734 solely to indicate this fact. ville, AL), including 201 known genes from the entire chromosome 17 and 435 1 Supported by the Research Council of Norway (to R. I. S.) and the Norwegian Cancer Society (to R. A. L.). ESTs from the 17q arm. An additional 512 sequence-verified IMAGE cDNA 2 To whom requests for reprints should be addressed, at Department of Genetics, clones were placed on the array, representing transcribed sequences located Institute for Cancer Research, The Norwegian Radium Hospital, N-0310 Oslo, Norway. elsewhere in the genome. Eighty-eight of these were housekeeping genes and Phone: 47-22934415; Fax: 47-22934440; E-mail: [email protected]. were used for calibration among the different experiments (15), 162 were a 3 The abbreviations used are: TGCT, testicular germ cell tumor; CCND2, cyclin D2; CGH, comparative genomic hybridization; DNMT2, DNA (cytosine-5)-methyltransferase 2; selection of known or putative cancer-related genes, and 262 were a collection ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2; EST, expressed of genes and ESTs from chromosome 10. sequence tag; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GRB7, growth factor Preparation and printing of the cDNA clones on glass slides, probe prepa- receptor-bound protein 7; JUP, junction plakoglobin; LLGL2, lethal giant larvae (Drosophila) rations, hybridizations, and image generation and analyses were performed as homolog 2; MYBL2, B-Myb; MYCL1, L-Myc; MYCN, N-Myc; MXI1, MAX-interacting protein 1; PDE6G, phosphodiesterase 6G, cGMP-specific, rod, gamma; POV1, prostate cancer described (16). Briefly, mRNA was isolated from the test samples by using the overexpressed gene 1; RT-PCR, reverse transcription-PCR. Trizol reagent (Life Technologies, Inc., Rockville, MD) and oligo(dT)25 dyna- 2359

Fig. 1. Genetic changes in testicular germ cell tumors. A, genomic copy number gains on chromosome 17 as seen by comparative genomic hybridization. Each colored bar represents gain from the corresponding chromosome segment in the tumor indicated below. B, sample tree (dendrogram) from the hierarchical cluster analysis of 18 testicular samples. Letters below the dendrogram represent the tissue sources: N, normal testis; C, carcinoma in situ; I, immature teratoma; E, embryonal carcinoma, and S, seminoma. The vertical distances on the dendrogram reflect the relatedness of neighbor- ing samples. A gene expression map with pseudo- colors coding for the normalized ratios of up- and down-regulated genes in TGCTs located on chromosome 17 (C) and elsewhere in the genome (D)is shown. Transcribed sequences are presented in rows, and each experiment (cDNA sample) is shown in columns. Thus, each cell in the matrix represents the expression level of a single transcript in a single sample. Numbers in parentheses behind the ESTs provide the IMAGE clone ids. E, color coding for the normalized expression ratios (expression relative to the average expressions in the three normal samples).

beads (Dynal Biotech, Oslo, Norway) according to the manufacturers’ speci- data were hierarchically clustered by both gene and sample sides (501 clones fications. From the reference cell lines, mRNA was isolated directly by using and 18 experiments). The average-linkage clustering method was used with FastTrack 2.0 mRNA isolation kit (Invitrogen, Carlsbad, CA). Labeled cDNA Pearson’s correlation similarity measure. Before calculation of the correlation was synthesized from 1–3or5␮g mRNA (test or reference, respectively) in between two genes or samples, the original ratio was log transformed, followed an oligo(dT)-primed polymerization with SuperScript II reverse transcriptase by subtracting the mean from the ratio. The sample tree (dendrogram) is drawn (Life Technologies) in the presence of either Cy3 (test) or Cy5 (reference) with “real” instead of fixed distances (in-house cluster analysis software at the labeled dUTP (Amersham Pharmacia, Piscataway, NJ). The Cy3-labeled test National Human Genome Research Institute, NIH). cDNA from the various testicular samples and Cy5-labeled reference cDNA Validation by Real-Time RT-PCR. We used real time RT-PCR (TaqMan were mixed and simultaneously hybridized to the cDNA microarray. system; Applied Biosystems, Foster City, CA) to validate the mRNA expres- The fluorescence intensities at the targets were detected by a laser confocal sion levels of three genes (GRB7, JUP, and POV1) in 10 testicular samples (3 scanner (Agilent Technologies, Palo Alto, CA). For each array element, a ratio normal testicular tissues and 7 TGCTs). In this quantitative RT-PCR, a between the relative fluorescence intensities of the test and reference was dual-labeled fluorescent probe is degraded concomitant with PCR amplifica- calculated. This ratio was divided by the average expressions of the 88 tion. Input target mRNA levels are calculated from the time (measured in PCR housekeeping genes, giving a calibrated ratio. The calibrated ratio was then cycles) at which the reporter fluorescent emission increases beyond a threshold normalized by dividing it by the average calibrated ratio of the three normal level, as measured by an ABI PRISM 7700 Sequence Detector (Applied testicular samples. Thus, this normalized ratio reflects relative up- or down- Biosystems). regulated gene expression from normal to neoplastic testicular tissues. Primers and probes targeting the mRNA sequences (Table 1) were designed Statistics. A two-tailed t test for equality of means was used to calculate the using the Primer Express software (Applied Biosystems). cDNA synthesized statistical significance of differences in expression levels between different from 50 ng of mRNA was used as PCR template in a total volume of 25 ␮l groups of samples. The hierarchical cluster analysis was done on successful containing 200 nM of each oligonucleotide primer and probe (MedProbe, Oslo, gene elements (i.e., clones where all experiments had spot sizes Ͼ100 area Norway), 0.2 mM of each deoxynucleotide triphosphate, 1 ϫ TaqMan buffer, units and fluorescence intensities stronger than 200 fluorescence units) with 6mM MgCl2, 1.25 units of AmpliTaq Gold, 0.25 units of AmpErase UNG (all more than 4-fold differential expression within the sample set. The resulting Applied Biosystems), and 0.8% glycerol. The PCR program was initiated by 2 2360

Table 1 Primers and probes used for real time RT-PCR The probes were 3Ј labeled with TAMRA and 5Ј labeled with 6-FAM (GRB7, JUP, and POV1) or JOE (GAPDH). Gene Forward primer Reverse primer Probe GRB7 TGG CCT CTC GGT CTG TAC AAA GGC AGG GAA TTA TGG GAG CGT GAA ACC GCC TGG GCT GC JUP CCA AAA ACA TAA AGC GAT AAT AAT AAA ACA C CCC CAT TTC CCG CAC AT CTG CTT GGA CCT CCC CCA GCC POV1 AAC CCC TAA CCC AGG ACA CAG AGA GAC ACA GCC CTC CTT TCA G TGG CAC CTC AGG CCC CTT TCC T GAPDH GAA GGT GAA GGT CGG AGT C GAA GAT GGT GAT GGG ATT TC CAA GCT TCC CGT TCT CAG CC min at 50°C and 10 min at 95°C before 40 thermal cycles, each of 15 s at 95°C 124915) were the most up-regulated transcripts (Fig. 1C). Among the and 1 min at 60°C. overexpressed genes on 17q, GRB7 was significantly more expressed Primers and probe targeting GAPDH were multiplexed together with prim- in nonseminomas and JUP in seminomas (both P Ͻ 0.01). ers and probes targeting each gene of interest. For both the test genes and For the clones mapping elsewhere in the genome, the most up- GAPDH, standard curves were made from which relative expression values regulated transcribed sequences in TGCTs, i.e., the highest average were calculated. The expression levels of the genes of interest were then normalized ratios across all tumor samples, were in decreasing order calibrated by dividing by the expression of GAPDH. Again, division by the average values of the three normal testicular samples normalized all calibrated MYBL2, CCND2, MYCN, POV1, EST (272938), and MYCL1. The expression values, and thus, these values were comparable with the normalized average expression levels of POV1 and MYCL1 were significantly ratios from the microarray experiments. higher in seminomas than in nonseminomas (P Ͻ 0.01). The expression data also revealed several genes, such as MXI1 and DNMT2, that Ͼ RESULTS were down-regulated in the TGCTs (on average 3-fold down- regulated and P Ͻ 0.01 for differential expression between normal The expression levels of 636 chromosome 17-specific transcripts testis and TGCTs; Fig. 1, C and D). as well as 512 transcripts located elsewhere in the genome were The expression levels of GRB7, JUP, and POV1 were validated in determined in 18 testicular tissue samples by cDNA microarrays. 10 samples by real time RT-PCR, and overexpression in tumors Hierarchical cluster analysis with a set of 501 differentially ex- (compared with normal testicular samples) were seen by both methods pressed genes separated the TGCT samples according to their (Fig. 3). known histological subgroups (Fig. 1B). The single carcinoma in situ sample, representing a precursor stage, was most closely DISCUSSION related to the normal testis specimens. The seminomas formed a single cluster, whereas within the nonseminomas, immature tera- Increased DNA copy number is a common mechanism for overex- tomas and embryonal carcinomas clustered into separate groups. A pression of genes promoting neoplastic and malignant cell behavior. comprehensive gene expression map for the 51 genes that were In TGCTs, frequent DNA copy number increase of several chromo- differentially expressed at a 0.01 significance level and had on somal regions has been observed (17, 18). However, not much is average Ͼ3-fold up- or down-regulation across all tumors, or known with regard to the specific genes that are targeted for overex- within a histological subgroup, is shown in Fig. 1, C and D. pression. Recently, we identified two novel regions on chromosome To identify up-regulated genes from the two regions with frequent arm 17q with common copy number increase in TGCT (8). In the copy number increase on chromosome 17, the normalized ratios of the present study, we used gene expression analysis by cDNA microar- genes were plotted as a function of their physical map positions (Fig. rays as a high throughput method to identify potential target genes in 2). This visualization indicated that not all transcripts located in a these two regions. The comprehensive coverage of the microarray region with increased copy number show increased expression. At the enabled us to determine which genes were overexpressed in TGCT, as proximal region (17q11–q21), GRB7 and JUP were consistently the compared with normal testicular tissue, and therefore most likely to be most overexpressed transcripts in the TGCTs (Fig. 1C). At the distal involved in driving the genomic alteration. Furthermore, the microar- region (17q24–qter), LLGL2, PDE6G, and EST (IMAGE clone ray was constructed to include several additional genes with known or

Fig. 2. Expression levels of transcripts on chromosome 17. Normalized expression values of genes localized on chromosome 17 were plotted as a function of their physical map positions (mega- basepairs from p-telomere, obtained from the Uni- versity of California, Santa Cruz database, http:// genome.ucsc.edu/). Individual data points were connected with a line. The chromosome ideogram is shown only for approximate visual comparison. Examples of an embryonal carcinoma with gain from the proximal gained region in TGCT (17q12- 21; A), an immature teratoma with extra copies of the whole chromosome 17 (B), a seminoma with gain at the distal region (17q24–qter; C) and a seminoma with no copy number changes on chromosome 17 (D) are shown.

2361

involvement in survival, proliferation, and migration of primordial germ cells (30). Additional knowledge about GRB7 that strengthens its potential importance in TGCT development is the RAS-associating-like domain (31) and its role in cell migration (32, 33). Interest- ingly, the expression of GRB7 in esophageal carcinomas is related to metastatic progression (34). JUP is also located within the proximal 17q region gained in TGCTs. It was up-regulated in tumors, with and without genomic gain by CGH. JUP belongs to the catenin family and encodes a submem- branous junctional plaque protein in both desmosomes and interme- diate junctions. It may have oncogenic potential through its function in the Wnt signaling pathway (35, 36), although the importance of this pathway in TGCT remains to be elucidated. At the distal 17q region frequently gained in TGCTs, 17q24–qter, Fig. 3. Validation of cDNA microarray results by real time RT-PCR. The expression the cDNA microarray analyses implicated the LLGL2 and PDE6G levels of GRB7, JUP, and POV1 were analyzed by real-time RT-PCR in 10 of the same mRNA samples used for cDNA microarray analyses. Relative mRNA levels in TGCTs genes, as well as an uncharacterized EST (124915), as consistently and normal testicular tissues are indicated by f and ⅜ respectively. The results from both up-regulated in the TGCTs. The Drosophila orthologue to LLGL2, methods are shown as normalized ratios (i.e., expression levels relative to the average of the three normal testicular samples). 1(2)gl, functions as a tumor suppressor (37). However, the function may be different in germ cells, because 1(2)gl is required for survival of germ-line cells in Drosophila (38), and our results show that putative cancer-related functions, which makes the present study the LLGL2 is up-regulated in human TGCT. Furthermore, LLGL2 are most extensive expression analysis of potentially cancer-promoting abundantly represented in some cDNA libraries derived from human 5 genes in TGCT. lung and prostate tumors. PDE6G encodes an effector protein in- The cDNA microarray-based expression survey in TGCTs and volved in phototransduction in the eye (39), and to our knowledge, no normal testis samples revealed several novel results: cancer-related function has been linked to this gene. (a) the hierarchical cluster analysis of the cDNA microarray data From the genes located elsewhere in the genome, three human grouped the samples according to their correct histological subtypes. homologues of avian retroviral oncogenes, MYCN at 2p24.1, MYCL1 This is rather surprising, taken into account the limited number and at 1p34.3, and MYBL2 at 20q13.1, were among the most overex- highly selected nature of the transcripts included in this analysis, and pressed genes in the TGCTs. The chromosomal locations of MYCN might indicate that genes located on chromosome 17 are fundamental and MYCL1 are both within regions that are gained in approximately for the biological characteristics of these tumors. one-third of all TGCTs, whereas the locus of MYBL2 is rarely in- (b) Comparison of the microarray expression data and the DNA volved in copy number changes (8). All three gene products are copy number increases along chromosome 17 as determined by CGH localized to the nucleus and function as transcriptional transactivators. showed that most genes were not transcriptionally up-regulated be- The MYCN overexpression has been detected previously in TGCT cause of extra DNA copies. These results are in line with our previous (40). Remarkably, studies of neuroblastomas give evidence for both data on breast cancer (14) and indicate that increased gene copy statistical and structural associations between MYCN overexpression number does not always lead to increased gene expression. and gain of 17q21–ter (12, 41). Furthermore, several E-boxes (the In the present study, we have identified overexpressed genes lo- common DNA binding site of the MYC family proteins) are found in cated in the two common regions of copy number increase on chro- the promoter region of CCND2, and MYC overexpression has been mosome 17 in TGCTs. At the proximal region (17q11–q21), the shown to induce chromosomal and extrachromosomal instability of cDNA microarray survey showed consistent overexpression of the the CCND2 gene at 12p13 (42). GRB7 and JUP genes. GRB7 is closely linked to the ERBB2 oncogene Gain of chromosome arm 12p, often through the presence of (20 kb apart4), and has been shown frequently coamplified and coex- isochromosome 12p (4), is the most common genetic aberration in pressed with ERBB2 in breast, esophageal, and gastric cancers (19– TGCTs. Two smallest regions of overlapping amplifications on 12p 22). Interestingly, the expression of ERBB2 was not elevated in any of have been suggested, one at 12p13 (8) and one more proximal region the TGCT samples. To our knowledge, this represents the first exam- (43), harboring the candidate genes CCND2 and KRAS2, respectively. ple where increased copy number at the ERBB2 locus does not lead to transcriptional activation of ERBB2. Furthermore, this indicates that We showed that both genes were transcriptionally up-regulated in other genes at this locus, such as GRB7, are critical for the develop- TGCTs, but CCND2 significantly more than KRAS2. Activating mu- ment of TGCTs and possibly to other tumor types with ERBB2 tations of KRAS2 have only rarely been detected in TGCTs (44, 45), amplification. further reducing the importance of this proto-oncogene in TGCTs. GRB7 encodes an adaptor protein that through its Src homology 2 The observed overexpression of CCND2 is in keeping with a study by domain interacts with the cytoplasmic domain of the growth factor Houldsworth et al. (46), where CCND2 had the highest increased receptor ERBB2 (19). Thus, increased expression of one of these expression among a set of six candidate genes on 12p, and with a proteins may be sufficient to promote tumor development. GRB7 also study by Bartkova et al. (47), finding CCND2 protein expression binds to several other tyrosine kinase growth factor receptors, includ- related to early stages of TGCTs. ing KIT, platelet-derived growth factor receptor, RET, and INSR The POV1 gene at 11q12 was highly expressed in all seminomas (23–26), as well as to cytoplasmic tyrosine kinases (27, 28). The KIT and in the carcinoma in situ, but neither in the normal samples nor in proto-oncogene has previously been suggested to play a role in TGCT the nonseminomas. Thus, independently of developmental model, this development, both attributable to increased expression (29), and by its gene may be an early-onset gene in the development of seminoma

4 Internet address: http://genome.ucsc.edu/ (Aug. 6, 2001 freeze). 5 Internet address: http://www.ncbi.nlm.nih.gov/UniGene/. 2362

TGCTs. In analogy, precursor lesions of prostate cancer have shown 10. Forozan, F., Mahlamaki, E. H., Monni, O., Chen, Y., Veldman, R., Jiang, Y., Gooden, increased expression of POV1 (48). G. C., Ethier, S. P., Kallioniemi, A., and Kallioniemi, O. P. Comparative genomic hybridization analysis of 38 breast cancer cell lines: a basis for interpreting comple- In addition to DNA copy number, other means of transcriptional mentary DNA microarray data. Cancer Res., 60: 4519–4525, 2000. control are obviously important to TGCT development. These may 11. Kokkola, A., Monni, O., Puolakkainen, P., Larramendy, M. L., Victorzon, M., include regulation of transcription factors and disruption of the DNA Nordling, S., Haapiainen, R., Kivilaakso, E., and Knuutila, S. 17q12–21 amplicon, a novel recurrent genetic change in intestinal type of gastric carcinoma: a comparative methylation pattern. Hence, the gene products of MXI1 and DNMT2 genomic hybridization study. Genes Chromosomes Cancer, 20: 38–43, 1997. are potential candidates because of their consistently down-regulated 12. Lastowska, M., Nacheva, E., McGuckin, A., Curtis, A., Grace, C., Pearson, A., and mRNA levels demonstrated by our microarray survey. The MXI1 Bown, N. Comparative genomic hybridization study of primary neuroblastoma tumors. United Kingdom Children’s Cancer Study Group. Genes Chromosomes Can- protein is a transcriptional repressor through its binding to MAX, a cer, 18: 162–169, 1997. MYC heterodimerization partner (49). Thus, by its competition for 13. Pack, S. D., Karkera, J. D., Zhuang, Z., Pak, E. D., Balan, K. V., Hwu, P., Park, W. S., MAX, MXI1 antagonizes the MYC transcription factors, of which we Pham, T., Ault, D. O., Glaser, M., Liotta, L., S. D., and Wadleigh, R. G. Molecular cytogenetic fingerprinting of esophageal squamous cell carcinoma by comparative have shown increased mRNA levels in TGCTs for both MYCN and genomic hybridization reveals a consistent pattern of chromosomal alterations. Genes MYCL1. Interestingly, the MXI1 mouse homologue is mapped within Chromosomes Cancer, 25: 160–168, 1999. the region with highest score in a genome-wide linkage analysis 14. Monni, O., Ba¨rlund, M., Mousses, S., Kononen, J., Sauter, G., Heiskanen, M., Paavola, P., Avela, K., Chen, Y., Bittner, M. L., and Kallioniemi, A. Comprehensive targeting TGCT susceptibility loci in mice (50, 51). Furthermore, the copy number and gene expression profiling of the 17q23 amplicon in human breast MXI1 gene is commonly mutated or deleted in prostate carcinomas cancer. Proc. Natl. Acad. Sci. USA, 98: 5711–5716, 2001. (52). The DNMT2 gene has strong sequence homology to the DNA- 15. Chen, Y., Dougherty, E. R., and Bittner, M. L. Ratio-based decisions and the quantitative analysis of cDNA microarray images. J. Biomed. Optics, 2: 364–374, (cytosine-5)-methyltransferases, although its catalytic activity has yet 1997. to be demonstrated (53). Additionally, DNMT2 is transcriptionally 16. Mousses, S., Bittner, M. L., Chen, Y., Dougherty, E. R., Baxevanis, A., Meltzer, P. S., down-regulated in colorectal, stomach, and hepatocellular cancers and Trent, J. M. Gene expression analysis by cDNA microarrays. In: F. J. Liversey and S. P. Hunt (eds.), Functional Genomics, pp. 113–137. Oxford: Oxford University (54, 55). Press, 2000. In summary, the present study has identified altered expression of 17. Chaganti, R. S., Rodriguez, E., and Bosl, G. J. Cytogenetics of male germ-cell tumors. several genes that are novel to testicular tumorigenesis. The increased Urol. Clin. North Am., 20: 55–66, 1993. 18. Rosenberg, C., Schut, T. B., Mostert, M., Tanke, H., Raap, A., Oosterhuis, J. W., and copy numbers observed at 17q11–q21 in TGCTs are associated with Looijenga, L. Chromosomal gains and losses in testicular germ cell tumors of overexpression of GRB7, and in contrast to other tumor types, not with adolescents and adults investigated by a modified comparative genomic hybridization approach. Lab. Investig., 79: 1447–1451, 1999. overexpression of the ERBB2 oncogene. In addition, JUP, MYCN, 19. Stein, D., Wu, J., Fuqua, S. A., Roonprapunt, C., Yajnik, V., D’Eustachio, P., MYCL1, MYBL2, CCND2, and POV1 are consistently overexpressed in Moskow, J. J., Buchberg, A. M., Osborne, C. K., and Margolis, B. The SH2 domain TGCTs compared with their expressions in nonneoplastic testicular tis- protein GRB-7 is co-amplified, overexpressed and in a tight complex with HER2 in breast cancer. EMBO J., 13: 1331–1340, 1994. sue. Furthermore, our data show clear gene expression differences be- 20. Kauraniemi, P., Ba¨rlund, M., Monni, O., and Kallioniemi, A. New amplified and tween seminoma and nonseminoma TGCTs. The average expression highly expressed genes discovered in the ERBB2 amplicon in breast cancer by cDNA level of GRB7 was significantly higher in nonseminomas than in semi- microarrays. Cancer Res., 61: 8235–8240, 2001. 21. Tanaka, S., Mori, M., Akiyoshi, T., Tanaka, Y., Mafune, K., Wands, J. R., and nomas, whereas the expressions of JUP, MYCL1, and POV1 were highest Sugimachi, K. Coexpression of Grb7 with epidermal growth factor receptor or in seminomas. The putative cancer-related functions of all these genes, in Her2/erbB2 in human advanced esophageal carcinoma. Cancer Res., 57: 28–31, addition to the previously reported overexpression of MYCN and CCND2 1997. 22. Akiyama, N., Sasaki, H., Ishizuka, T., Kishi, T., Sakamoto, H., Onda, M., Hirai, H., in TGCT (40, 46), suggest that the applied cDNA microarrays are Yazaki, Y., Sugimura, T., and Terada, M. Isolation of a candidate gene, CAB1, for sensitive and specific enough to discover oncogenic gene expression cholesterol transport to mitochondria from the c-ERBB-2 amplicon by a modified changes in TGCT. Thus, the consistently overexpressed ESTs may also cDNA selection method. Cancer Res., 57: 3548–3553, 1997. 23. Tho¨mmes, K., Lennartsson, J., Carlberg, M., and Ro¨nnstrand, L. Identification of reflect genes playing important roles in TGCT oncogenesis. In conclu- Tyr-703 and Tyr-936 as the primary association sites for Grb2 and Grb7 in the sion, this focused functional genomic characterization has lead to the c-Kit/stem cell factor receptor. Biochem. J., 341: 211–216, 1999. identification of new gene targets associated with a common genomic 24. Yokote, K., Margolis, B., Heldin, C. H., and Claesson-Welsh, L. Grb7 is a down- stream signaling component of platelet-derived growth factor ␣- and ␤-receptors. rearrangement in TGCT and of genes with potential clinical impact J. Biol. Chem., 271: 30942–30949, 1996. worthy further investigation. 25. Pandey, A., Liu, X., Dixon, J. E., Di Fiore, P. P., and Dixit, V. M. Direct association between the Ret receptor tyrosine kinase and the Src homology 2-containing adapter protein Grb7. J. Biol. Chem., 271: 10607–10610, 1996. REFERENCES 26. Kasus-Jacobi, A., Bereziat, V., Perdereau, D., Girard, J., and Burnol, A. F. Evidence for an interaction between the insulin receptor and Grb7. A role for two of its binding 1. Bergstro¨m, R., Adami, H. O., Mohner, M., Zatonski, W., Storm, H., Ekbom, A., domains, PIR and SH2. Oncogene, 19: 2052–2059, 2000. Tretli, S., Teppo, L., Akre, O., and Hakulinen, T. Increase in testicular cancer 27. Han, D. C., and Guan, J. L. Association of focal adhesion kinase with Grb7 and its incidence in six European countries: a birth cohort phenomenon. J. Natl. Cancer Inst., role in cell migration. J. Biol. Chem., 274: 24425–24430, 1999. 88: 727–733, 1996. 28. Keegan, K., and Cooper, J. A. Use of the two hybrid system to detect the association 2. Devesa, S. S., Blot, W. J., Stone, B. J., Miller, B. A., Tarone, R. E., and Fraumeni, of the protein-tyrosine-phosphatase, SHPTP2, with another SH2-containing protein, J. F. J. Recent cancer trends in the United States. J. Natl. Cancer Inst., 87: 175–182, Grb7. Oncogene, 12: 1537–1544, 1996. 1995. 29. Strohmeyer, T., Reese, D., Press, M., Ackermann, R., Hartmann, M., and Slamon, D. 3. Dieckmann, K. P., and Skakkebæk, N. E. Carcinoma in situ of the testis: review of Expression of the c-kit proto-oncogene and its ligand stem cell factor (SCF) in normal biological and clinical features. Int. J. Cancer, 83: 815–822, 1999. 4. Atkin, N. B., and Baker, M. C. Specific chromosome change, i(12p), in testicular and malignant human testicular tissue. J. Urol., 153: 511–515, 1995. tumours? Lancet, 2: 1349, 1982. 30. Mauduit, C., Hamamah, S., and Benahmed, M. Stem cell factor/c-kit system in 5. Rodriguez, E., Houldsworth, J., Reuter, V. E., Meltzer, P., Zhang, J., Trent, J. M., spermatogenesis. Hum. Reprod. Update, 5: 535–545, 1999. Bosl, G. J., and Chaganti, R. S. Molecular cytogenetic analysis of i(12p)-negative 31. Wojcik, J., Girault, J. A., Labesse, G., Chomilier, J., Mornon, J. P., and Callebaut, I. human male germ cell tumors. Genes Chromosomes Cancer, 8: 230–236, 1993. Sequence analysis identifies a ras-associating (RA)-like domain in the N-termini of 6. Chaganti, R. S., and Houldsworth, J. Genetics and biology of adult human male germ band 4.1/JEF domains and in the Grb7/10/14 adapter family. Biochem. Biophys. Res. cell tumors. Cancer Res., 60: 1475–1482, 2000. Commun., 259: 113–120, 1999. 7. Looijenga, L. H., and Oosterhuis, J. W. Pathogenesis of testicular germ cell tumours. 32. Han, D. C., Shen, T. L., and Guan, J. L. Role of Grb7 targeting to focal contacts and Rev. Reprod., 4: 90–100, 1999. its phosphorylation by focal adhesion kinase in regulation of cell migration. J. Biol. 8. Kraggerud, S. M., Skotheim, R. I., Szymanska, J., Eknæs, M., Fosså, S. D., Stenwig, Chem., 275: 28911–28917, 2000. A. E., Peltoma¨ki, P., and Lothe, R. A. Genome profiles of familial/bilateral and 33. Manser, J., Roonprapunt, C., and Margolis, B. C. elegans cell migration gene mig-10 sporadic testicular germ cell tumors. Genes Chromosomes Cancer, in press, 2002. shares similarities with a family of SH2 domain proteins and acts cell nonautono- 9. Lothe, R. A., Karhu, R., Mandahl, N., Mertens, F., Sæter, G., Heim, S., Børresen- mously in excretory canal development. Dev. Biol., 184: 150–164, 1997. Dale, A. L., and Kallioniemi, O. P. Gain of 17q24–qter detected by comparative 34. Tanaka, S., Sugimachi, K., Kawaguchi, H., Saeki, H., Ohno, S., and Wands, J. R. genomic hybridization in malignant tumors from patients with von Recklinghausen’s Grb7 signal transduction protein mediates metastatic progression of esophageal neurofibromatosis. Cancer Res., 56: 4778–4781, 1996. carcinoma. J. Cell Physiol., 183: 411–415, 2000. 2363

35. Kolligs, F. T., Kolligs, B., Hajra, K. M., Hu, G., Tani, M., Cho, K. R., and Fearon, 45. Ridanpa¨a¨, M., Lothe, R. A., O¨ nfelt, A., Fosså, S. D., Børresen, A. L., and Husgafvel- E. R. ␥-catenin is regulated by the APC tumor suppressor and its oncogenic activity Pursiainen, K. K-ras oncogene codon 12 point mutations in testicular cancer. Environ. is distinct from that of ␤-catenin. Genes Dev., 14: 1319–1331, 2000. Health Perspect., 101 (Suppl. 3): 185–187, 1993. 36. Barker, N., and Clevers, H. Catenins, Wnt signaling and cancer. Bioessays, 22: 46. Houldsworth, J., Reuter, V., Bosl, G. J., and Chaganti, R. S. Aberrant expression of 961–965, 2000. cyclin D2 is an early event in human male germ cell tumorigenesis. Cell Growth 37. To¨ro¨k, I., Hartenstein, K., Kalmes, A., Schmitt, R., Strand, D., and Mechler, B. M. Differ., 8: 293–299, 1997. The 1(2)gl homologue of Drosophila pseudoobscura suppresses tumorigenicity in 47. Bartkova, J., Rajpert-De Meyts, E., Skakkebæk, N. E., and Bartek, J. D-type cyclins transgenic Drosophila melanogaster. Oncogene, 8: 1537–1549, 1993. in adult human testis and testicular cancer: relation to cell type, proliferation, differentiation, and malignancy. J. Pathol., 187: 573–581, 1999. 38. De Lorenzo, C., Strand, D., and Mechler, B. M. Requirement of Drosophila I(2)gl 48. Cole, K. A., Chuaqui, R. F., Katz, K., Pack, S., Zhuang, Z., Cole, C. E., Lyne, J. C., function for survival of the germline cells and organization of the follicle cells in a Linehan, W. M., Liotta, L. A., and Emmert-Buck, M. R. cDNA sequencing and columnar epithelium during oogenesis. Int. J. Dev. Biol., 43: 207–217, 1999. analysis of POV1 (PB39): a novel gene up-regulated in prostate cancer. Genomics, 39. Pittler, S. J., Baehr, W., Wasmuth, J. J., McConnell, D. G., Champagne, M. S., van 51: 282–287, 1998. Tuinen, P., Ledbetter, D., and Davis, R. L. Molecular characterization of human and 49. Zervos, A. S., Gyuris, J., and Brent, R. Mxi1, a protein that specifically interacts with bovine rod photoreceptor cGMP phosphodiesterase ␣-subunit and chromosomal Max to bind Myc-Max recognition sites. Cell, 72: 223–232, 1993. localization of the human gene. Genomics, 6: 272–283, 1990. 50. Collin, G. B., Asada, Y., Varnum, D. S., and Nadeau, J. H. DNA pooling as a quick 40. Shuin, T., Misaki, H., Kubota, Y., Yao, M., and Hosaka, M. Differential expression method for finding candidate linkages in multigenic trait analysis: an example of protooncogenes in human germ cell tumors of the testis. Cancer, (Phila.), 73: involving susceptibility to germ cell tumors. Mamm. Genome, 7: 68–70, 1996. 1721–1727, 1994. 51. Matin, A., Collin, G. B., Asada, Y., Varnum, D., and Nadeau, J. H. Susceptibility to 41. O’Neill, S., Ekstrom, L., Lastowska, M., Roberts, P., Brodeur, G. M., Kees, U. R., testicular germ-cell tumours in a 129.MOLF-Chr 19 chromosome substitution strain. Schwab, M., and Bown, N. MYCN amplification and 17q in neuroblastoma: evidence Nat. Genet., 23: 237–240, 1999. for structural association. Genes Chromosomes Cancer, 30: 87–90, 2001. 52. Prochownik, E. V., Eagle, G. L., Deubler, D., Zhu, X. L., Stephenson, R. A., Rohr, 42. Mai, S., Hanley-Hyde, J., Rainey, G. J., Kuschak, T. I., Paul, J. T., Littlewood, T. D., L. R., Yin, X., and Brothman, A. R. Commonly occurring loss and mutation of the Mischak, H., Stevens, L. M., Henderson, D. W., and Mushinski, J. F. Chromosomal MXI1 gene in prostate cancer. Genes Chromosomes Cancer, 22: 295–304, 1998. and extrachromosomal instability of the cyclin D2 gene is induced by Myc overex- 53. Robertson, K. D. DNA methylation, methyltransferases, and cancer. Oncogene, 20: 3139–3155, 2001. pression. Neoplasia, 1: 241–252, 1999. 54. Kanai, Y., Ushijima, S., Kondo, Y., Nakanishi, Y., and Hirohashi, S. DNA methyl- 43. Roelofs, H., Mostert, M. C., Pompe, K., Zafarana, G., van Oorschot, M., van Gurp, transferase expression and DNA methylation of CpG islands and peri-centromeric R. J., Gillis, A. J., Stoop, H., Beverloo, B., Oosterhuis, J. W., Bokemeyer, C., and satellite regions in human colorectal and stomach cancers. Int. J. Cancer, 91: 205– Looijenga, L. H. Restricted 12p amplification and RAS mutation in human germ cell 212, 2001. tumors of the adult testis. Am. J. Pathol., 157: 1155–1166, 2000. 55. Saito, Y., Kanai, Y., Sakamoto, M., Saito, H., Ishii, H., and Hirohashi, S. Expression 44. Mulder, M. P., Keijzer, W., Verkerk, A., Boot, A. J., Prins, M. E., Splinter, T. A., and of mRNA for DNA methyltransferases and methyl-CpG-binding proteins and DNA Bos, J. L. Activated ras genes in human seminoma: evidence for tumor heterogeneity. methylation status on CpG islands and pericentromeric satellite regions during human Oncogene, 4: 1345–1351, 1989. hepatocarcinogenesis. Hepatology, 33: 561–568, 2001.

2364

Rolf I. Skotheim, Outi Monni, Spyro Mousses, et al.

Cancer Res 2002;62:2359-2364.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/62/8/2359

Cited articles This article cites 51 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/62/8/2359.full#ref-list-1

Citing articles This article has been cited by 16 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/62/8/2359.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/62/8/2359. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.