Up-Regulation of Cyclin-Dependent Kinase 4/Cyclin D2 Expression but Down- Regulation of Cyclin-Dependent Kinase 2/Cyclin E in Testicular Germ Cell Tumors1

[CANCER RESEARCH 61, 4214–4221, May 15, 2001] Up-Regulation of Cyclin-dependent Kinase 4/Cyclin D2 Expression but Down- Regulation of Cyclin-dependent Kinase 2/Cyclin E in Testicular Germ Cell Tumors1

Bettina A. Schmidt,2 Achim Rose, Christine Steinhoff, T. Strohmeyer, Michael Hartmann, and Rolf Ackermann Department of Urology, Heinrich-Heine-University of Duesseldorf, 40225 Duesseldorf, Germany [B. A. S., A. R., C. S., R. A.], and Department of Urology, Hospital of the Armed Forces, 22099 Hamburg, Germany [M. H.]

ABSTRACT causes death in about 20–30% of patients with highly advanced stages. The present state of research provides only marginal insights Testicular germ cell tumors (GCT) characteristically display two chro- into the molecular pathogenesis of these tumors and the mechanisms mosome 12 abnormalities: the isochromosome i(12p) and concomitant underlying the effectiveness of chemotherapy. deletions of the long arm. Some genes important in the control of the G -S 1 The most frequent cytogenetic observation indicative of testicular cell cycle checkpoint G1-S, i.e., cyclin-dependent kinases 2 and 4, cyclin D2 are located on this chromosomal region. Therefore, testicular GCTs were GCT is the presence of an isochromosome of the short arm of analyzed as to the expression of CDK2, CDK4, CDK6, and the expression chromosome 12, i12(p), in up to 80% of all tumors. Although this of their catalytic partners cyclins D1, D2 and E by semiquantitative finding has been confirmed by various investigators (6–8), its diag- reverse transcription-PCR. Cyclin D2, located on 12p, was overexpressed nostic and prognostic relevance is still under discussion (9, 10). in 69% (31 of 45) of the tumors by a mean factor of 8, including all Additional abnormalities involve deletions of the long arm of chro- histological subtypes. In addition, the cyclin D2 partner CDK4 was in- mosome 12, particularly the 12q13 and 12q22 regions, in 40% of all creased in 41% (21 of 51) of all tumors by a factor of 6, most strongly in histological subtypes (11). These findings suggest the localization of embryonal carcinomas. Sixty-four percent of the seminomas and 23% of the non-seminomas had decreased expression of CDK6 by a mean factor one or more tumor suppressor genes on 12q and the localization of Statistical analysis using configural frequency analysis relevant proto-oncogenes on 12p (12). The acquired data supports the .(0.009 ؍ of5(P and regression analysis revealed that cyclin D2 and CDK4 expression hypothesis of a possible role for chromosome 12 in the development .whereas expression and progression of testicular GCTs ,(0.000052 ؍ P ;0.682 ؍ were strongly correlated (r2 An important molecular change in testicular GCTs involves the loss .(0.00085 ؍ P ;0.382 ؍ of CDK6 did not correlate with either of them (r2 CDK2 and its catalytic partner cyclin E were down-regulated in 40% (19 of mRNA and protein expression of the RB tumor-suppressor gene. of 47) and 42% (19 of 45) of the tumors, respectively, by a factor of 7 each. We found a highly significant suppression or lack of expression of the Western blots and immunohistochemical experiments confirmed cyclin RB gene at both mRNA and protein levels in 95% and 71%, respec- D2 and CDK4 overrepresentation and reduced expression of cyclin E and CDK2 tumors in the few tumors under protein study. Despite its local- tively, of seminomas and non-seminomas (13). The RB protein was ization on 12q13, a hot spot for loss of heterozygosity in testicular GCTs not detected by immunohistochemistry in undifferentiated tumors (>40%), Southern blotting revealed no gross DNA alteration of the CDK2 (seminomas, embryonal carcinoma and chorion carcinomas). Only gene. Because up-regulation of the cyclin D2/CDK4 complex and down- differentiated cells of teratocarcinomas and mixed tumors showed regulation of cyclin E/CDK2 complex were found in seminomas as well as some protein expression. However, using Southern blotting, no alter- in non-seminomas and in all tumor stages, these findings seem to be early ations were found at the DNA level (13). events during tumorigenesis of testicular GCTs. Together with previous The nuclear protein encoded by the RB gene is consistently ex- findings that retinoblastoma mRNA and protein expression is strongly pressed by all normal mammalian cells and tissues. However, a decreased in these tumors, these data suggest an unusual deregulated G -S 1 number of different tumor types (RB, bronchial carcinoma, osteosar- checkpoint as a decisive event for germ cell tumors. coma, glioma, and carcinoma of the prostate and bladder) are associated with inactivation of the RB gene by deletion, mutation and/or INTRODUCTION lack of down-regulation (14–16). The RB protein is an essential protein in cell cycle regulation, and its function is regulated by With an incidence of 6–7 per 100,000 men, GCTs3 are the most phosphorylation. In G0 and the early G1 phase, hypophosphorylated frequent solid malignant tumors in men 20–40 years of age and RB is complexed with the cellular transcription factor E2F (17, 18). In represent the most frequent cause of death from solid tumors in this late G1, a significant hyperphosphorylation of the RB protein by age group (1, 2). The rapid progress of GCTs causes early lymph node cyclin-dependent kinases (CDK2, CDK4, and CDK6) in complex metastases and/or distant metastases. At the time of diagnosis, up to with their catalytic partners (cyclins D1, D2, D3, and E) occurs 50% of the patients suffer from metastatic disease (3). Because of (19–22). As a consequence, E2F is set free and binds to promoters of highly effective combinations of cytostatic drugs, cure rates of 85– genes that are involved in the early steps of the subsequent S phase, 90% can be reached (4, 5). Despite this success, the disease still i.e., DNA polymerase ␣ and ␦, dihydrofolate reductase, thymidine kinase, cyclin E, CDK 2, c-myc, and E2F-1 itself. During the subse- Received 3/6/00; accepted 2/28/01. quent S, G , and mitosis phases, the RB protein remains hyperphos- The costs of publication of this article were defrayed in part by the payment of page 2 charges. This article must therefore be hereby marked advertisement in accordance with phorylated. 18 U.S.C. Section 1734 solely to indicate this fact. Several regulators of RB are located on the regions of chromosome 1 The work was supported by a grant from the Deutsche Forschungsgemeinschaft (Ac 29/1-2). 12 frequently affected by aberrations in GCTs (23–25). While human 2 To whom requests for reprints should be addressed Bettina Schmidt, Heinrich-Heine- cyclin D2 gene (CCND2) is located on 12p13, the human CDK2 gene University of Duesseldorf, Department of Urology; Moorenstr. 5, 40225 Duesseldorf, (CDK2) and CDK4 gene (CDK4) are localized on 12q13. These Germany Phone: ϩ49–211-8118110; Fax: ϩ 49–211-8117804 e-mail: bschmidt@uni- duesseldorf.de. findings, combined with the above-mentioned observations on RB 3 The abbreviations used are: GCT, germ cell tumor; RB, retinoblastoma; RT-PCR, expression in GCTs, give rise to the hypothesis that the cell cycle reverse transcription-PCR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; CFA, configural frequency analysis; ISA, interacting structure analysis; FISH, fluorescence in restriction point G1-S may be entirely disrupted in testicular GCTs. situ hybridization. The aim of this study was to test this hypothesis in primary testicular 4214

GCTs by measuring the mRNA expression of those proteins which facturer’s protocol. Chemiluminescence was detected on enhanced chemilu- directly regulate pRB function: the cyclin-dependent kinases CDK2, minescence-Hyperfilm (Amersham). Signals were quantified by video scan- CDK4, and CDK6 and their catalytic partners (cyclin D1, cyclin D2, ning densitometry (One D-SCAN; Scanalytics, Billerica, Madison, WI) and and cyclin E). standardized for ␤-actin expression in each matched normal testicular tissue. One set of experiments was performed with GAPDH as the internal control. However, because of the localization of the GAPDH gene on chromosome 12p, MATERIALS AND METHODS GAPDH expression was additionally standardized to ␤-actin expression in each tumor/normal pair. Overexpression in tumor was defined as expression Tumor and adjacent normal testicular tissue was obtained by surgery, shock Ͼ2-fold of the corresponding normal tissue; normal as between 0.5-fold and frozen in liquid nitrogen, and stored at Ϫ70°C. The tissue cohort included 58 2.0- fold; and decreased expression as Ͻ0.5-fold, after standardization to the GCTs and the adjacent nonmalignant testicular tissues. Histological tumor ␤-actin expression ratios of tumor and normal tissue. subtyping and staging was performed according to the WHO International Southern Blot Analysis. Genomic DNA (10 ␮g) was digested with 8 units Histological Classification of Tumors (26) and the American Joint Committee of TaqI restriction enzyme (Life Technologies, Inc., Karlsruhe, Germany), on Cancer (27). resolved on an 0.8% agarose gel, and transferred to a nylon membrane Isolation of Total RNA. Total cellular RNA was extracted from frozen (Hybond; Amersham, Buckinghamshire, UK) according to standard protocols tissues using guanidinium thiocyanate, phenol-chloroform, and CsCl density (28). Prehybridization and hybridization were performed in Dig-Hybridization gradient centrifugation as described (28). Purified RNA was quantified by UV buffer (Roche Diagnostics, Mannheim, Germany) at 42°C. Fifty ng of the spectroscopy scans from 340 nm to 210 nm (UviKon 900; Kontron Instru- 778-bp CDK2 probe (pCDK2), representing nucleotide numbers 166–943 ments, Neufarn, Germany). Integrity of RNA was confirmed by denaturing PAGE. (kindly provided by Guido Reifenberger, Department of Neuropathology, Primers. Primers specific for cyclin E, cyclin D1, cyclin D2, cyclin D2 University of Duesseldorf, Duesseldorf, Germany) was digoxigenine-labeled pseudogene, CDK2, CDK4, and CDK6 cDNA were designed using sequences by oligonucleotide random priming (38) and used for hybridization. The final ϫ published previously (24, 25, 29–33) by the primer design program OLIGO, wash was carried out in 0.5 SSC-0.1% SDS at 68°C. Version 4.4. Primers for cyclin D1, cyclin D2, CDK2, and CDK4 have been Western Blot Analysis. Tissues were pulverized using mortar and pestle published previously by our laboratory (34). Primer sequences are summarized on dry ice. Tissue powder was added directly to ice-cold lysis buffer containing in Table 1. The human D-type cyclin genes CCND1 (cyclin D1), CCND2 25 mM Tris-HCl (pH 7.4), 50 mM sodium chloride, 0.5% sodium deoxycholic ␮ (cyclin D2), and CCND3 (cyclin D3) share an average of 57% identity over the acid, 2% NP40, 0.2% SDS, 1 mM phenylmethylsulfonyl fluoride, 50 g/ml ␮ entire coding region and 78% in the cyclin region. The cyclin D2 and its aproptinin, and 50 g/ml leupeptin and incubated for 15 min on ice. Lysates pseudogene are 86% homologous. The downstream primer for cyclin D2 were boiled for 10 min, passed through a 25-gauge needle three times, and ϫ cDNA was therefore chosen to bind in a small region deleted in the pseudo- centrifuged at 14,000 g for 15 min. Protein concentration from the super- gene. All primers for RT-PCR were selected to span one or more introns to natant was quantified by using bicinchoninic acid assay (Pierce Biochemicals, ␮ avoid amplification of residual DNA (Table 1). Rockford, IL). Five to 25 g of total cellular proteins were mixed with ϫ Semiquantitative RT-PCR. Because our extracted RNA was used in for- one-fifth volume of 5 SDS loading buffer [250 mM Tris-HCl (pH 6.8), 10% mer Northern expression studies on the metastasis suppressor genes nm23 and SDS, 50% glycerol, and 0.001% bromphenol blue] and boiled for 5 min. DCC and the c-myc oncogene (36, 37), we had not enough RNA for additional Together with a prestained protein marker, these samples were separated by Northern blottings. Therefore, we performed semiquantitative RT-PCR with 10% SDS-PAGE and transferred by semidry blotting to Immobilon mem- highly sensitive chemiluminescence detection as described previously (36). branes (Millipore, Bedford). Membranes were probed with mouse monoclonal cDNA was synthesized by RT from 2 ␮g total RNA using oligo(dT) primer antibodies against cyclin D1 (clone HD 11; dilution, 1:200; Oncogene Re- (Sigma Chemical Co., St. Louis, MO) and AMV reverse transcriptase (Pro- search Products, Boston, MA), cyclin E (clone HE12; dilution, 1:1000), CDK6 mega, Madison, WI). One-tenth of the cDNA was used for amplification by (clone B10; dilution, 1:1000), goat polyclonal antibodies against cyclin D2 PCR. Amplification of cDNAs was performed using 20-pmol primers each. (dilution, 1:1000), CDK4 (dilution, 1:1000), and CDK2 (dilution: 1:500). All DNA polymerase (1 unit; DynaZyme; Biometra, Go¨ttingen, Germany), 50 ␮M antibodies, with the exception of cyclin D1 antibody, were purchased from dNTP mixture, and 2.5 ␮M digoxigenine-labeled dUTP (Roche Diagnostics, Santa Cruz Biotechnology, Heidelberg, Germany. Detection was performed Mannheim, Germany) were used for each reaction. An initial 5-min denatur- using antimouse and antigoat antibodies conjugated with horseradish peroxi- ation at 96°C and 5 min annealing at the indicated temperature (Table 1) were dase and the enhanced chemiluminescence system (Amersham Life Sciences). carried out. Cycles of polymerization, denaturation, and annealing were then Immunohistochemistry for Cyclin D2, Cyclin E, CDK4, and CDK2. performed for 90 s at 72°C, 30 s at 96°C, and 45 s at the empirically Paraffin-embedded, formalin-fixed tissue sections (5 ␮m) were deparaffinized

determined optimal annealing temperature. The ranges of cycle numbers in xylol and rehydrated. Sections were treated with 0.1% H50O2 in methanol yielding amplification were determined for each gene in preliminary experi- for 30 min. Antigen retrieval was achieved by microwave treatment in 1.8 mM ments. The housekeeping genes GAPDH and/or ␤-actin were coamplified as citrate buffer (pH 6) for 10 min (cyclin E) or treatment with 1 mg/ml Pronase an internal control. To compensate for low expression of some cell cycle E (Merck, Darmstadt, Germany) in PBS for 1 h (cycline D2), or treatment with regulators, addition of control gene primers to the PCR reaction was delayed 0.0125% trypsine solution (Roche, Mannheim, Germany; pH 7.6), for 1 h by a number of cycles (compare Table 1). PCR products were separated on a (CDK 4 and CDK 2). Nonspecific binding was blocked by incubation with 1.6% agarose gel and transferred to a nylon membrane (Hybond; Amersham, rabbit serum (DAKO, Glostrup, Denmark) for cyclin D2, CDK 2, and CDK 4, Buckinghamshire, United Kingdom). Incorporated digoxigenine was detected and horse serum for cyclin E (Vector, Burlingam), for 30 min. Primary using an alkaline phosphatase-labeled anti-digoxigenine antibody and CSPD as antibodies used for immunohistochemistry were the same as those for Western substrate (Roche Diagnostics, Mannheim, Germany) according to the manu- Blotting. Primary antibodies were diluted 1:500, 1:100, 1:500, and 1:200 for

Table 1 RT-PCR primers and experimental conditions used for analyzing mRNA expression of G1-S cell cycle regulators Annealing No. of Product Gene Forward primer Reverse primer temperature cycles length Source CDK 2 5ЈGCTTTCTGCCATTCTCATCG-3Ј 5ЈGTCCCCAGAGTCCGAAAGAT-3Ј 57°C 21 316 bp (34) CDK 4 5ЈACGGGTGTAAGTGCCATCTG-3Ј 5ЈTGGTGTCGGTGCCTATGGGA-3Ј 59°C 18 464 bp (34) CDK 6 5ЈCGAATGCGTGGCGGAGATC-3Ј 5ЈCCACTGAGGTTAGAGCCATC-3Ј 62°C 23 499 bp This paper Cyclin D1 5ЈGAGACCATCCCCCTGACGGC-3Ј 5ЈTCTTCCTCCTCCTCGGCGGC-3Ј 65°C 22 484 bp (34) Cyclin D2 5ЈTGCATGTTCCTGGCCTCC-3Ј 5ЈTTAAAGTCGGTGGCACACA-3Ј 65°C 19 247 bp (34) Cyclin D2 pseudo 5ЈCCAGGAACATGCAGACAG-3Ј 5ЈTCAAGTGCGTGCAGAAGG-3 62°C 19 292 bp This paper Cyclin E 5ЈATACAGACCCACAGAGACAG-3Ј 5ЈTGCCATCCACAGAAATACTT-3Ј 61°C 15 301 bp This paper ␤-actin 5ЈTGACGGGGTCACCCACACTGTGCCCATCTA-3Ј 5ЈCTAGAAGCATTTGCGGTGGACGATGGAGGG-3Ј 63°C 9 661 bp Stratagene, CA GAPDH 5ЈTCCCATCACCATCTTCCA-3Ј 5ЈACTCACGCCACAGTTTCC-3Ј 59°C 15 379 bp (35) 4215

Expression of Cyclins. Cyclin D2 was expressed in normal testicular tissues and was found to be overexpressed in 69% (31 of 45) of all testicular GCTs by a mean factor of 7.9 with various histologies (Fig. 2). This overexpression affected both seminomas and non- seminomas. Strong overexpression of cyclin D2 was found in 80% (17 of 21) of the localized tumors and in 58% (14 of 24) of the tumors Fig. 1. mRNA expression of GAPDH and ␤-actin in testicular GCTs. mRNA expres- with lymph node and/or distant metastases (Table 2). This difference sions in testicular GCTs (A) were evaluated by semiquantitative multiplex RT-PCR and was not statistically significant (␹2, P ϭ 0.102; Fisher, P ϭ 0.121; compared with matched normal testicular tissue (B). Ratios of GAPDH:actin expression in tumor versus normal testicular tissues were determined to 5.4, 11, 8.7, 0.9, 1.4, 1.2, 3.3, odds ratio, 3.036) but allowed this interpretation in form of a trend 5.2, 1.6, and 2.4 (from left to right, patient ID 47 to patient ID 106) by videodensitrometric (Table 3). We excluded expression of the cyclin D2 pseudogene (31) analyses. as the cause for overexpression by a subset of tumors with specific primers resulting in a longer PCR product during amplification of cyclin D2 pseudogene (data not shown). The cyclin D2 pseudogene antihuman cyclin E, antihuman cyclin D2, antihuman CDK 2, and antihuman was expressed neither in tumor nor in normal testicular tissue of ten CDK 4, respectively. Incubation was performed overnight at 4°C. Biotinylated patients. Analysis of an additional 10 normal and tumor tissues of secondary antibodies (Vector) were applied for 1 h with subsequent incubation different histologies revealed only minute cyclin D1 mRNA expres- with horseradish peroxidase-conjugated streptavidin (Vector) for 1 h. Visual- sion, although we used three more amplification cycles than in cyclin ization was achieved by incubation with diaminobenzidine (Sigma Chemical D2 PCR. In contrast, cyclin D1 expression was found with teratocar- Co., St. Louis, MO) and counterstaining with Mayer’s hematoxiline (Merck, cinoma cell lines TERA-1 and TERA-2 (see Fig. 2), urothelial cancer Darmstadt, Germany). cell lines VMCub I, VMCub III, and TCC Sup, and prostate cancer Statistical Analysis. All statistical tests were carried out using the SAS program, Version 6.12, and SPSS, Version 9.0. The program provided in Ref. cell lines DU145 and LNCaP using the same amplification profile 39 was used for CFA. The significance of differences of expression between (published previously by our group; see Ref. 34). different histological subtypes and tumor stages was analyzed using Fisher’s Expression of cyclin E mRNA was decreased in 42% (19 of 45) of exact test and the ␹2 test (including odds ratios and confidence intervals). GCTs by a mean factor of 6.8, predominantly in seminomas (in 58% Statistical evaluation of the correlation between mRNA expression of each pair versus 36.4% in non-seminomas). Interestingly, most of the tumors of the regulators was performed by linear and quadratic regression analysis. with lowered cyclin E expression were localized tumors (11 of 21; Possible relations between the five cell cycle regulators were analyzed using 52%), whereas only 29% (7 of 24) of the tumors with lymph node descending and ascending ISA as a first step. Then, the results were confirmed or rejected using of the asymptotic version of hypergeometric CFA (39, 40).

RESULTS In the first sets of experiments, we used the housekeeping gene GAPDH as internal standard. However, because of chromosomal localization of GAPDH on 12p, GAPDH was found overexpressed in testicular tumors. Therefore, we standardized all expression results Fig. 2. mRNA expression of cyclin D2, cyclin D1, and GAPDH in testicular GCTs. from all tumors, including those with GAPDH, to ␤-actin expression. mRNA expressions in testicular GCTs (A) were evaluated by semiquantitative multiplex GAPDH was increased by a mean factor of 3.3 in 55% of all tumors RT-PCR and compared with matched normal testicular tissue (B). Ratios of cyclin D2/GAPDH expression in tumor versus normal testicular tissues were determined from (Fig. 1). There was no specific correlation of GAPDH expression to this figure to 1.1, 1.0, 1.4, 1.5, 8.0, 6.7, 2.2, 2.1, and 7.8 (from left to right, patient ID 41 tumor stage or histological subtype. to patient ID 114) by videodensitrometric analyses.

Table 2 mRNA expression of G1-S cell cycle regulators—association to histological tumor subtypes and stage (performed according the WHO Histological Classification of Tumors (26) and the American Joint Committee on Cancer (27) Upper listed values of each row indicate percentages, parentheses indicate the number of affected tumor. Average factors of up- and down-regulation are given in the second line of each row in italics. Quantification was done by laser scanning densitometry of RT-PCR products and standardization to actin expression.a Histology/Stage CDK 2 2 CDK 4 1 CDK 6 2 Cyclin D2 1 Cyclin E 2 Seminomas 42.9 (6/14) 33.3 (5/17) 64.3 (9/14) 75.0 (8/12) 58.3 (7/12) 10.9 7.9 5.4 7.5 6.3 Non-seminomas 39.4 (13/33) 47.1 (16/34) 23.3 (7/30) 69.7 (23/33) 36.4 (12/33) 5.5 5.5 4.0 8.1 7.1 Embryonal carcinomas 42.9 (3/7) 71.4 (5/7) 28.6 (2/7) 80.0 (8/10) 44.4 (4/9) 5.1 4.2 3.0 9.1 4.0 Mixed tumors 72.7 (8/11) 53.8 (7/13) 33.3 (3/10) 75.0 (9/12) 36.4 (4/11) 5.6 4.9 4.2 7.8 5.0 Teratocarcinomas 16.6 (2/12) 33.3 (4/12) 9.1 (1/11) 60.0 (6/10) 27.2 (3/11) 7.7 6.6 6.3 7.6 12.9 Choriocarcinomas 00.0 (0/2) 00.0 (0/2) 50.0 (1/2) 00.0 (0/2) 50.0 (1/2) 2.9 2.4 N0, M0* 52.0 (13/25) 35.7 (10/28) 38.1 (8/21) 80.1 (17/21) 52.4 (11/21) 9.2 7.8 4.5 8.4 9.1 N1-3, M0* 30.1 (4/13) 42.9 (6/14) 50.0 (7/14) 53.3 (8/15) 31.2 (5/16) 4.5 4.6 4.9 7.7 3.4 M1* 22.2 (2/9) 55.5 (5/9) 11.1 (1/9) 66.6 (6/9) 25.0 (2/8) 6.0 3.1 5.3 6.8 2.0 Total 40.4 (19/47) 41.2 (21/51) 36.4 (16/44) 68.9 (31/45) 42.2 (19/45) 7.2 5.9 4.7 7.9 6.8 a 2 ϭ down-regulation; 1 ϭ up-regulation; * all histological subtypes. 4216

Table 3 Results of Fisher’s exact test and ␹2 test for significance of mRNA expression cluding the ascending and descending ISA was performed. ISA was data of cell cycle regulators (from Table 2) related to histological subtype (seminomas used in a first calculation for any pair of the regulators CDK2, CDK4, versus non-seminomas) or tumor stage (N0 versus N1 and/or M1) CDK6, cyclin D2, and cycline E and in a second calculation for any P P Odds Confidence (Fisher) (␹2) Ratio Interval higher order of interactions between them. The first calculation revealed only one strict relation, i.e., between CDK4 and cyclin D2 Seminomas vs. non-seminomasa CDK 22 1.0 0.825 1.154 0.325–4.101 (Table 4). However, additional interactions were found for the fol- CDK 41 0.366 0.227 0.469 0.135–1.623 lowing triples: CDK4, CDK2, cycline D2; CDK4, cyclin D2, cyclin E; CDK 62 0.017 0.009 5.914 1.484–23.564 Cyclin D21 1.0 0.846 0.87 0.12–3.566 CDK4, cycline D2, CDK6. Higher combinations (one versus three, Cyclin E2 0.306 0.187 2.45 0.36–9.442 two versus two, and one versus four) showed no significant interac- N0 vs. N1-M1a tion. All results were proven by the asymptotic hypergeometric ver- CDK 22 0.136 0.085 2.889 0.850–9.816 CDK 41 0.408 0.382 0.606 0.197–1.868 sion of CFA (Table 4). Thus a strong correlation between two regu- CDK 62 1.0 0.820 1.154 0.337–3.946 lators is only given for CDK4 and cyclin D2, which is the same result 1 Cyclin D2 0.121 0.102 3.036 0.781–11.807 as that from regression-analysis (Fig. 5a). This finding is valid for Cyclin E2 0.138 0.113 2.671 0.782–9.122 high levels of expression of CDK4 and cyclin D2 and for low levels a 2 ϭ downregulation; 1 ϭ upregulation. of expression of both, too. To get significance in correlated expression, at least three regulators have to be considered for all other relations (Table 4; high expressions in CDK4, CDK2, and cyclin D2; and high expressions of CDK4, cyclin E, and cyclin D2). The interaction between CDK4, cyclin D2, and CDK6 found with ISA could not be confirmed with CFA and thus had to be rejected. Fig. 3. mRNA expression of CDK2 and CDK4 and ␤-actin in testicular GCTs. mRNA As expected from the results of the ISA, there was a linear corre- expressions in testicular GCTs (A) were evaluated by semiquantitative multiplex RT-PCR lation between cyclin D2 and CDK4 (r2 ϭ 0.585; Fig. 5a, left) which and compared with matched normal testicular tissue (B). Ratios of CDK4/actin expression was even improved in a quadratic model (r2 ϭ 0.682; Fig. 5a, right). in tumor versus normal testicular tissues were determined to 0.4, 1.0, 1.9, 0.6, 1.2, 2.1, 2.5, 2 0.3, and 1.3 and ratios of CDK2/actin expression to 0.5, 0.2, 2.0, 0.5, 0.3, 0.3, 0.1, 0.2, and For all other combinations, no strict correlation (r Ͼ 0.5) was 0.2 (from left to right, patient ID 6 to patient ID 114). obtained, in agreement with the result from ISA. At least three factors have to be considered for a strict correlation. A slight correlation and/or distant metastasis showed this change (Table 2). Statistical between CDK4 and CDK6 (Fig. 5b; r2 ϭ 0.2652 in the linear model, analysis confirmed this finding as a trend (Table 3, ␹2: P ϭ 0.112; and r2 ϭ 0.381 in the quadratic model) was found using regression- Fisher, P ϭ 0.138, odds ratio, 2.671). analysis. Expression of CDKs. CDK4 expression was increased in 41% (21 Analysis of Protein Expressions. To investigate whether mRNA of 51) of all tumors. This increase was more frequent (16 of 34; 47%) expression data were mirrored by protein expression data, we per- in non-seminomas than in pure seminomas (5 of 17; 33%) and was formed preliminary Western blotting analysis and immunohistochem- most prevalent in embryonal carcinomas (5 of 7; 71%; Table 2). istry from a subset of tumors (Figs. 6 and 7). However, we had However, the average increase in expression was more pronounced in residual tissues and paraffin-embedded tissue blocks from only a few the seminomas (factor 7.9) than in non-seminomas (factor 5.5). Al- patients, which had been previously analyzed by RT-PCR for mRNA though there was an increase in percentage of CDK4 overexpression expression. By comparing the tumor/normal ratios of each cell cycle with tumor stage, there was no significance in Fisher’s exact test and the ␹2 test (Table 3). In contrast, CDK2 mRNA expression was strongly decreased in 40% (19 of 47) of the GCTs (Fig. 3). This decrease affected both seminomas and non-seminomas, but was much more pronounced (9.2-fold) in the pure seminomas than in the non-seminomas (4.5- fold). Interestingly, only 2 of 12 (17%) teratocarcinomas had decreased CDK2 expression. There was an inverse relation to tumor stage, with the highest frequency (13 of 25; 52%) and the strongest decrease (9.2-fold) in localized tumors (Table 2). This finding is supported by statistical analysis (Table 3), revealing a trend Fig. 4. Analysis of the CDK2 gene by Southern hybridization. DNA was analyzed from (P ϭ 0.085; odds ratio 2.89) for this relationship. different testicular tumors with reduced CDK2 mRNA expression (14, 25-fold; 19, 5-fold; 27, 4-fold; 41, 13-fold; and 96, 5-fold reductions in expression) and from one tumor with Most of the tumors had normal CDK6 expression, but 36% (16 of CDK2 overexpression (13, 3-fold) and compared with matched normal testicular tissue 44) showed a down-regulation of the gene by a mean factor of 4.7. (B). Nine of 14 (64%) seminomas, but only 7 of 30 (23%) of the non- seminomas were affected (Table 2). This difference was highly sta- Table 4 Correlation of mRNA expression of the cell cycle regulators CDK2, CDK4, tistically significant (P ϭ 0.009; odds ratioϭ5.9; Table 3). There was CDK6, cyclin D2, and cyclin E according to ISA and CFA no correlation to tumor stage. The level of mRNA expression is indicated in parentheses. For all, statistics significance levels were adjusted by the Bonferoni method for a starting ␣-level of 0.05. Only Analysis of the CDK2 Gene. To analyze a loss of the gene as a significant results are shown. potential cause for a decreased CDK2 expression, DNA from tissue Method Regulator P Adjusted ␣ pairs of patients with decreased CDK2 expression was studied by ISA CDK 4, cyclin D2 0.000052 0.050000 Southern hybridization (Fig. 4). However, the data revealed no dif- ISA CDK 4 ϫ (CDK 2, cyclin D2) 0.001574 0.001786 ferences in gene copy number or in gross structural alteration of the ISA CDK 4 ϫ (cyclin E, cyclin D2) 0.001050 0.001786 CDK2 gene in these patients, with the exception of tumor 27, which ISA CDK 4 ϫ (CDK 6, cyclin D2) 0.000847 0.001786 CFA CDK 4 (high), cyclin D2 (high) 0.004487 0.050000 seems to have a decreased number of copies or a loss of the gene. CDK 4 (low), cyclin D2 (low) 0.001112 0.050000 Configural and Correlation Analysis. To find possible relations CFA CDK 4 (high), CDK 2 (high), cyclin D2 (high) 0.000000 0.000125 between any combination of the five cell cycle regulators CFA in- CFA CDK 4 (high), cyclin E (high), cyclin D2 (high) 0.000001 0.000125 4217

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2001 American Association for Cancer Research. CYCLINS AND CDKS IN TESTICULAR TUMORS regulator as determined by video scanning densitometry of Western and RT-PCR bands, mRNA expression data for cyclin D2, cyclin E, and CDK6 could be clearly confirmed (Table 5). In addition, CDK4 overexpression as well as CDK2 down-regulation was found in this subset of tumors, although the level of expression ratios (tumor: normal) for proteins and mRNA were not the same for every tumor. Contrary to negligible mRNA expression (see Fig. 2), cyclin D1 protein expression was pretty high in tumor and normal testicular tissues (Fig. 5). To confirm further cyclin D2 and CDK4 overexpression and cyclin E and CDK2 down-regulation in testicular tumors, we performed immunohistochemical analysis for these G1-S cell cycle regulators (Fig. 7). In normal testicular tissues, mainly later stages of spermatogenesis (spermatids) demonstrated strong nuclear cyclin D2 staining, whereas Sertoli cells and the early stages of spermatogenesis were negative (Fig. 7a). In contrast, seminoma cells (Fig. 7b) had strong Fig. 6. Protein expression of in testicular GCTs. Protein expressions of G1-S cell cycle regulators in testicular GCTs (A) were evaluated by Western blotting analysis and nuclear and cytoplasmatic cyclin D2 staining. CDK4 was not detected compared with matched normal testicular tissue (B). Quantification was carried out using in normal testicular tissue, neither in germ cells nor in Leydig or videodensitometry. Sertoli cells. However, nuclear CDK4 staining was detected in smooth muscle cells of blood vessels in the same section (Fig. 7c). In contrast with normal tissues, testicular tumors had heterogeneous CDK 4 expression was found mainly in Sertoli cells of normal testicular expression, with very high levels of CDK4 in tumors with epithelial tissues (Fig. 7g). The seminoma shown in Fig. 7h was completely differentiation and weaker expression in tumor stroma (Fig. 7d). negative, with the exception of endothelial cells of blood vessels Cyclin E protein was strongly expressed in seminiferous tubules of within the tumor. normal testicular tissue in basal layers with spermatogonia (Fig. 7e). Embryonal carcinomas as shown in Fig. 7f displayed completely DISCUSSION negative tumor cells, whereas infiltrating lymphocytes and capillaries

(arrows) stained positive for cyclin E in the same section. CDK2 Because of the chromosomal localization of several G1-S restriction point cell cycles regulators on chromosome 12, which is typically altered in testicular GCTs, we analyzed the expressions of several

G1-S regulator molecules involved in the regulation of RB function. The mRNA of the cyclin D2 gene, located on 12p13, was strongly overexpressed by an average factor of 7.9 in 69% of all tumors throughout all histological subtypes. Although we did not do any studies on cyclin D2 gene amplification, with our tumors used for expression studies, our result on cyclin D2 overexpression is in accordance with the presence of isochromosome i(12p) in Ͼ80% of the testicular GCTs published by others (6–8), and we believe that this overrepresentation of 12p may be the cause of cyclin D2 overexpression. However, the housekeeping gene GAPDH, which is located on the same band on chromosome 12, was overexpressed by a mean factor of only 3.3. One recent triple-color FISH study of 12p amplifications in 49 testicular GCTs using 12p band-specific painting probes demonstrated the amplification of the three bands 12p11.2, p12, and p13 in all tumors with copy numbers varying from 4–11 (41). Eighty-four percent of the tumors displayed one or three iso- chromosomes 12, and many had multiple copies of parts of 12p elsewhere in the genome. The authors concluded that several genes located in different parts on the distal region of 12p (12p12 and/or 12p13) are involved in the pathogenesis of testicular tumors. There- fore, the difference between cyclin D2 and GAPDH mRNA overexpression may be caused by the fact that two or more regions of 12p13 may be amplified to different degrees. The first evidence suggesting cyclin D2 to be involved in the pathogenesis of GCTs came from Sicinski et al. (42), who developed cyclinD2 knockout mice. Although widely expressed during normal mouse embryo development, cyclin D2 knockout animals were phe- Fig. 5. Correlation analysis of cyclin D2 and CDK4 (a) and CDK6 and CDK4 notypically indistinguishable from their wild-type littermates, possi- (b) mRNA expression in testicular GCTs (left, linear model; right, quadratic model). bly because of the expression of other cyclin D types in most tissues. The regression equation was calculated for cyclin D2 and CDK4 to: cyclinD2 ϭ Ϫ4.798 ϩ 5.070 ϫ CDK4 for cyclin D2/CDK4 for the linear model and cy- However, mutant females were infertile because of an incapacity to clinD2 ϭϪ3.041 Ϫ 1.198 ϫ CDK4 ϩ 0.704 ϫ CDK42 for the quadratic model; for release oocytes, whereas mutant males displayed hypoplastic testis. In CDK6 and CDK4 to: CDK6 ϭ 0.572 ϩ 0.194 ϫ CDK4 for the linear model and CDK6 ϭ 1.065 Ϫ 0.192 ϫ CDK4 ϩ 0.044 ϫ CDK42 for the quadratic model. addition, these authors found very high mRNA expression levels in Confidence intervals of 95% are indicated. cell lines from yolk sac tumors and an increase in the copy number of 4218

Table 5 Comparison of protein expression ratios (tumor:normal) of patient samples shown in Western blots of Fig. 5 versus their mRNA expression as determined by RT-PCR Quantification of protein and mRNA expression was done by laser scanning densitometry. mRNA expression of cyclin D1 was too low for quantification and is indicated as n.d. (not detected). Tumor ID 13 Tumor ID 41 Tumor ID 78 Tumor ID 100

Gene Protein mRNA Protein mRNA Protein mRNA Protein mRNA Cyclin D2 2.1 4.1 0.8 1.1 3.2 6.5 8.6 3.1 CDK 4 1.3 9.9 5.0 1.5 2.1 2.6 7.2 4.1 Cyclin E 0.3 0.3 0.8 0.1 0.4 0.4 0.3 0.6 CDK2 0.1 0.4 0.8 0.1 0.4 0.7 0.9 0.5 Cyclin D1 0.7 n.d. 1.0 n.d. 2.7 n.d. 0.6 n.d. CDK 6 0.6 0.2 0.8 1.3 1.2 1.6 1.2 0.8 the cyclin D2 gene in 15 of 19 cell lines tested. Houldsworth et al. (43) related with the specific pathway of differentiation. Although we found a marked variation of the cyclin D2 mRNA expression level found cyclin D2 expression in the later stages of spermatogenesis between six cell lines derived from testicular GCTs, but no correlation (spermatids), these findings were confirmed by our immunohistolog- to the copy number. The protein was overexpressed to different ical study (see Fig. 7, a and b) on these few tumors under study; this extents in the individual cell lines, with inverse correlation to the was not the case in our mRNA expression study, where different status of differentiation. Although normal germ cells contained no tumor types showed similar frequencies and extents of overexpression protein, seminomas with no evidence of differentiation displayed the (Table 3). However, this is in accord with an analysis of the ampli- highest protein expression. The expression patterns of teratomas cor- fication rate of the 12p region by FISH (41), which revealed no

Fig. 7. Immunohistochemical analysis of the expression of cyclin D2 (A and B), CDK4 (C and D), cyclin E (E and F), and CDK2 proteins (G and H) in normal testicular tissues with seminiferous tubules (left; A, C, E, and F) and testicular GCTs (right; B, D, F, and H). Tumors in B and H are seminomas, whereas the tumor in D and F represent the mixed tumor subtype and embryonal carcinomas, rsp. Magnification is ϫ400 in each section with the exception of C, where a magnification of ϫ250 was used. The arrows indicate positive staining of smooth muscle cells in a vein in C, of lymphocytes and capillaries in F, and of blood vessel cells in H.

4219

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2001 American Association for Cancer Research. CYCLINS AND CDKS IN TESTICULAR TUMORS correlation with the histological tumor type. The finding of cyclin D2 percentages and factors in localized tumors, indicating an early event overexpression in seminomas as well as in localized tumors suggests in the tumorigenesis of testicular GCTs. that this occurs early in the tumorigenesis of testicular GCTs. The cofactor for CDK2 action in RB phosphorylation, cyclin E In a subset of nine testicular tumors and adjacent normal testicular recently has been suspected to be involved in tumorigenesis of breast tissue, we did not find any significant cyclin D1 mRNA expression. carcinomas. For these tumors, cyclin E overexpression is suggested as These results fit very well to immunohistochemical data from Bart- a prognostic marker (51). In excellent correlation with the expression kova et al. (44), who showed that cyclin D1 protein expression was data for CDK2, we found a down-regulation of cyclin E in 42% of the absent in normal germ cells, seminomas, and embryonal carcinomas. testicular GCTs in both seminomas and non-seminomas. The trend for They detected moderate to low cyclin D1 expression in some differ- an inverse correlation to tumor stage was found to nearly the same entiated compounds of teratomas. Interestingly, this group found percentages with cyclin E and CDK2. Our preliminary immunohisto- variable positivity for cyclin D3 in invasive testicular tumors as well chemical studies confirm the decreased cyclin E expression results on as a correlation of cyclin D3 with differentiation, whereas normal the protein level, too (Fig. 7, e and f). However, decreased cyclin E germ cells again were negative. However, the role for cyclin D3 in protein expression was observed to higher frequencies in seminomas promoting S-phase entry and initiation or maintenance of differenti- (85%; 23 of 27) than in non-seminomas (39%; 9 of 23). The finding ation has to be figured out in the future. of down-regulation of the cyclin E/CDK2 complex in testicular tu- CDK4 and CDK6 in complex with cyclin D2 act to drive the cell mors is unprecedented in any tumor and is quite surprising, inasmuch cycle. Because the 12q13 region where the CDK4 gene is located is as both proteins were expected to be necessary for rapid cell cycle often deleted in testicular GCT, a decreased expression of this gene progression. The reason for down-regulation of both regulators in would be expected. However, we found a CDK4 overexpression by an testicular tumors is unclear but may be a consequence of the decreased average factor of 6 in 41% of the tumors correlating with cylin D2 expression of RB (13), because cyclin E transcription is positively overexpression. Although this overexpression fits to the postulated regulated by E2F and RB protein (29). function of CDK4 as an oncogene described for glioblastomas (45) In conclusion, up-regulation of the cyclin D2/CDK4 complex and and osteosarcomas (46, 47), our mRNA expression findings are in the down-regulation of cyclin E/CDK2 complex as well as down- contrast with protein expression data of six GCT cell lines, where only regulation of CDK6 in seminomas seem to be frequent and early little variation of CDK4 protein expression was found (43). More events during tumorigenesis of testicular GCTs. In particular, overexpression of CDK4 and cyclin D2, the correlated expression of both consistent with our data, cyclin D2 protein was in complex with mRNAs, and the complex formation of both proteins in GCT cell lines CDK4 protein in these cell lines. ISA, CFA, and regression analysis suggest the concerted oncogenic actions of both in these tumors. of our data revealed a strict correlation between CDK4 and cyclin D2 CyclinD2/CDK4 complexes may act as promoters of the cell cycle in mRNA expression in testicular tumors. testicular tumors despite the lack of RB. It has been shown for cyclin Until now there are only a few tumor subsets with known alter- D1/CDK4 and cyclin E/CDK2 complexes that their overexpression ations of the CDK6 gene. In gliomas, an amplification-associated and can accelerate S-phase entry and promote cell cycle progression an amplification-independent increase of CDK6 protein was identified independently of RB/E2F (52), presumably by acting upon down- (48). According to CDK6 function in phosphorylation of pRB to- stream targets that are rate-limiting for S-phase entry. Interestingly, gether with cyclin D2/CDK4, an overexpression of the gene was this bypass of the G arrest is a feature of the cyclin D1/CDK4 expected. However, CDK6 overexpression was found in only 16% of 1 complex but not of cyclin D1 alone. Alternatively, overexpression of the tumors. Most of the tumors had normal CDK6 expression, and cyclin D1/CDK4 in testicular GCTs may reflect a nonphysiological 33% showed a decreased mRNA expression, particularly seminomas loss of the specificity of the CDK4. The additional down-regulation of ␹2 ϭ with high significance (64%; , P 0.009; odds ratio, 5.9). Houlds- cyclinE/CDK2 in testicular GCTs may not have any influence on the worth et al. (43) have also reported a decreased CDK6 protein oncogenic process, and overexpression of cyclin D2 and CDK4 may expression in two of six cell lines and demonstrated that CDK6 was be sufficient to initiate the transforming action in testicular germ cells. in complex with cyclin D2/CDK4. Although regression analysis of The unusual type of the deregulation of the G1-S checkpoint seems to our data revealed only a weak correlated expression of CDK4 and be the key event for the development of testicular GCTs. CDK6, a highly significant interaction was revealed by ISA for the CDK4, CDK6, and cyclin D2 triple. (Table 4). However, CFA could not confirm the correlated expression of the three cell cycle regulators ACKNOWLEDGMENTS all together. Loss of heterozygosity studies have revealed deletions of the 12q13 We thank Christiane Hader for excellent technical assistance, mathemati- cian Janine Illian from the Department of Psychology, Heinrich-Heine- and 12q22 regions in Ͼ40% of testicular GCTs (8, 12). Several genes University of Duesseldorf, for advice and helpful discussions during statistical with relevance to tumorigenesis, including the CDK2 gene, are lo- analysis, and Dr. Wolfgang Schulz for critical reading of the manuscript and cated in this chromosomal region, although CDK2, like CDK4,is stimulating discussions. already known to act more as an oncogene than as a tumor suppressor gene. In colorectal carcinoma, for example, the CDK2 gene is amplified, and CDK2 activity is increased (49, 50). Surprisingly, with REFERENCES respect to CDK2 function during G1-S cell cycle regulation, but in 1. McKiernan, J. M., Goluboff, E. T., Liberson, G. L., Golden, R., and Fisch, H. Rising concordance with the previous loss of heterozygosity findings in 40% risk of testicular cancer by birth cohort in the United States from 1973–1995. J. Urol., 162: 361–363, 1999. of testicular GCTs (12), we found a lowered CDK2 expression by a 2. Wei␤bach, L., and Bussar-Maatz, R. Pathogenese, Diagnostik und Therapie von mean factor of 7.2 in 40% of seminomas and non-seminomas. We Hodentumoren. Urologe (A), 35: 163–172, 1996. excluded major alterations of the CDK2 gene as the cause for de- 3. Peckham, M. Testicular cancer. Acta Oncol., 27: 439–453, 1988. 4. Schmoll, H. J. Biology and treatment of testicular cancer. Consultant series: the role creased CDK2 expression in these tumors by Southern blots. Smaller of haematopoietic growth factors, 1–13. United Kingdom: Gardiner Caldwell Comu- deletions of the CDK2 gene or a transcriptional regulation of CDK2 nications, Ltd., 1993. 5. Aareleid, T., Sand, M., and Hedelin, G. Improved survival for patients with testicular may be the reason for CDK2 down-regulation. Our data revealed a cancer in Europe since 1978. EUROCARE Working Group. Eur. J. Cancer, 34: trend for an inverse correlation with tumor stage, with the highest 2236–2240, 1998. 4220

6. Delozier-Blanchet, C. D., Engel, E., and Walt, H. Isochromosome 12p in malignant 30. Palmero, I., Holder, A., Sinclair, A. J., Dickson, C., and Peters, G. Cyclins D1, and testicular tumors. Cancer Genet. Cytogenet., 15: 375–376, 1985. D2 are differentially expressed in human B-lymphoid cell lines. Oncogene, 8: 7. Suijkerbuijk, R. F., van de Veen, A. Y., van Echten, J., Buys, C., de Jong, B., 1049–1054, 1993. Oosterhuis, J. W., Warburton, D. A., Cassiman, J. J., Schonk, D., and Geurts van 31. Inaba, T., Matsushime, H., Valentine, M., Roussel, M. F., Sherr, C. J., and Look, A. T. Kessel, A. Demonstration of the genuine iso 12p character of the standard marker Genomic organisation, chromosomal localization and independent expression of chromosome of testicular germ cell tumors and identification of further chromosome human cyclin D genes. Genomics, 13: 565–574, 1992. 12 aberrations by competitive in situ hybridization. Am. J. Hum. Genet., 48: 269– 32. Tsai, L. H., Harlow, E., and Meyerson, M. Isolation of the human CDK2 gene that 273, 1991. encodes the cyclin A- and adenovirus E1A-associated p33 kinase. Nature (Lond.), 8. Rodriequez, E., Mathew, S., Reuter, V., Ilson, D. H., Bosl., G. J., and Chaganti, 353: 174–177, 1991. R. S. K. Cytogenetic analysis of 124 prospectively ascertained male germ tumors. 33. Meyerson, M. L., Enders, G. H., Wu, C. L., Su, L. K., Gorka, C., Nelson, C., Harlow, Cancer Res., 52: 2285–2291, 1992. E., and Tsai, L. H. A family of human Cdc2-related protein kinase. EMBO J., 11: 9. Bosl, G. J., Dmitrovsky, E., Reuter, V. E., Samaniego, F., Rodriguez, E., Geller, 2909–2917, 1992. N. L., and Chaganti, R. S. Isochromosome of the short arm of chromosome 12: 34. Oya, M., Schmidt, B., Schmitz-Dra¨ger, B. J., and Schulz, W. A. Expression of G1/S clinically useful markers for male germ cell tumors. J. Natl. Cancer Inst., 81: transition regulatory molecules in human urothelial cancer. Jpn. J. Cancer Res., 89: 1874–1878, 1989. 719–726, 1998. 10. Bosl, G. J., Ilson, D. H., Rodriguez, E., Motzer, R. J., Reuter, V. E., and Chaganti, 35. Schoefeld, A., Luqman, Y., Shousha, S., Sinnett, H. D., and Coombes, P. Cancer Res., R. S. K. Clinical relevance of the i(12p) marker chromosome in germ cell tumors. 54: 2986–2990, 1994. J. Natl. Cancer Inst., 86: 349–355, 1994. 36. Schmidt, B., Ackermann, R., Hartmann, M., and Strohmeyer, T. Alterations of the 11. Samaniego, F., Rodriguez, E., and Houldsworth, J. Cytogenetic and molecular anal- metastasis suppressor gene nm23 and the proto-oncogene c-myc in human testicular ysis of human male germ tumors: Chromosome 12 abnormalities and gene amplifi- germ cell tumors. J. Urol., 158: 2000–2005, 1997. cation. Genes Chromosomes Cancer, 1: 289–300, 1990. 37. Strohmeyer, D., Langenhof, S., Ackermann, R., Hartmann, M., Strohmeyer, T., and 12. Murty, V. V. V. S., Houldsworth, J., Baldwin, S., Reuter, V., Hunziker, W., Besmer, Schmidt, B. Analysis of the tumor suppressor gene DCC in testicular germ cell P., Bosl, G., and Chaganti, R. S. K. Allelic deletions in the long arm of chromosome tumors: mutations and loss of expression. J. Urol., 157: 1973–1976, 1997. 12 identify sites of candidate tumor suppressor genes in male germ cell tumors. Proc. 38. Feinberg, A. P. and Vogelstein, B. A technique for radiolabelling DNA restriction Natl. Acad. Sci. USA, 89: 11006–11010, 1992. endonuclease fragments to high specific activity. Anal. Biochem., 132: 6–13, 1983. 13. Strohmeyer, T., Reissmann, P., Cordon-Cardo, C., Hartmann, M., Ackermann, R., 39. Krauth, J. Einfu¨hrung in die Konfigurationsfrequenzanalyse (KFA). BELTZ Psy- and Slamon, D. J. Correlation between retinoblastoma gene expression and differen- chologie Verlagsunion, Weinheim, Basel, Germany, 1993. tiation in human testicular tumors. Proc. Natl. Acad. Sci. USA, 88: 6662–6666, 1991. 40. Krauth, J. Principles of configural frequency analysis. Zeit. f. Psychology, 193: 14. Lindardopoulos, S., Gonos, E. S., and Spandidos, D. A. Abnormalities of retinoblas- 363–375, 1995. toma gene structure in human lung tumors. Cancer Lett., 71: 67–74, 1993. 41. Henegariu, O., Vance, G. H., Heiber, D., Pera, M., and Heerema, N. Triple-color 15. Ozaki, T., Ikeda, S., Kawai, A., Inoue, H., and Oda, T. Alterations of the retinoblas- FISH analysis of 12p amplification in testicular germ cell tumors using 12p band toma susceptible gene accompanied by c-myc amplification in human bone and soft specific painting probes. J. Mol. Med., 76: 648–655, 1998. tissue tumors. Cell. Mol. Biol., 39: 235–242, 1993. 42. Sicinski, P., Donaher, J. L., Geng, Y., Parker, S. B., Gardner, H., Park, M. Y., Robker, 16. Lundberg, A. S., and Weinberg, R. A. Control of the cell cycle and apoptosis. Eur. J. R. L., Richards, J. S., McGinnis, L. K., Biggers, J. D., Eppig, J. J. Bronson, R. T., Cancer, 35: 531–539, 1999. Elledge, J., and Weinberg, R. A. Cyclin D2 is an FSH-responsive gene involved in 17. Weinberg, R. A. The retinoblastoma protein and cell cycle control. Cell, 81: 323–330, gonadal cell proliferation and oncogenesis. Nature (Lond.), 384: 470–474, 1996. 1995. 43. Houldsworth, J., Reuter, V., Bosl, G. S., and Chaganti, R. S. K. Aberrant expression 18. Adams, P. D., Li, X., Sellers, W. R., Baker, K. B., Leng, X., Harper, J. W. Taya, Y., of cyclin D2 is an early event in human male germ cell tumorigenesis. Cell Growth and Kaelin, W. R., Jr. Retinoblastoma protein contains a C-terminal motif that targets Differ., 8: 293–299, 1997. it for phosphorylation by cyclin-cdk complexes. Mol. Cell. Biol., 19: 1068–1080; 44. Bartkova, J., Rajpert-De Meyts, E., Skakkebek, N. E., and Bartek, J. D-type cyclins 1999. in adult human testis and testicular cancer: relation to cell type, proliferation, 19. Dulic, V. E., Lees, E., and Reed, S. I. Association of human cyclin E with a periodic differentiation, and malignancy. J. Pathol., 187: 573–581, 1999. G1/S phase protein kinase. Science (Wash DC), 257: 1958–1961, 1992. 45. Reifenberger, G., Ichimura, K., Reifenberger, J., Elkahloun, A. G., Meltzer, P. S., and 20. Koff, A., Giordano, A., Desai, D., Yamashita, J. W., Harper, J. W., Elledge, S., Collins, V. P. Refined mapping of12q13–q15 amplicons in human malignant gliomas Nishimoto, T., Morgan, D. O., Franza, B. R., and Roberts, J. Formation and activation suggests CDK4/SAS and MDM2 as independent amplification targets. Cancer Res., of a cyclin E-cdk2 complex during the G1 phase of the human cell cycle. Science 56: 5141–5145, 1996. (Wash DC), 257: 1689–1694, 1992. 46. Wunder, J. S., Eppert, K., Burrow, S. R., Gogkoz, N., Bell, R. S., and Andrulis, I. L. 21. Sherr, C. J. Cancer cell cycles. Nature (Lond.), 274: 1672–1676, 1996. Co-amplification and overexpression of CDK4, SAS and MDM2 occurs frequently in 22. Parry, D., Mahony, D., Willis, K., and Lees, E. Cyclin D-CDK subunit arrangement human parosteal osteosarcomas. Oncogene, 18: 783–788; 1999. is dependent on the availability of competing INK4 and p21 class inhibitors. Mol. 47. Benassi, M. S., Molendini, L., Gamberi, G., Sollazzo, M. R., Ragazzini, P., Merli, M., Cell. Biol., 19: 1775–1783; 1999. Magagnoli, G., Sangiorgi., L., Bacchini, P., Bertoni, F., and Picci, P. Altered G1 phase 23. Xiong, Y., Menninger, J., Beach, D., and Ward, D. C. Molecular cloning and mapping regulation in osteosarcoma. Int. J. Cancer, 74: 518–522, 1997. of CCND genes encoding human D-type cyclins. Genomics, 13: 575–584, 1992. 48. Costello, J. F., Plass, C., Arap., W., Chapman, V. M., Held, W. A., Berger, M. S., 24. Schiffman, D., Brooks, E. E., Brooks, A. R., Chan, C. S., and Milner, P. G. Su-Huang, H. J., and Cavenee, W. K. Cyclin-dependent kinase 6 (CDK6) amplifi- Characterization of the human cyclin dependent kinase 2 gene. Promoter analysis and cation in human gliomas identified using two-dimensional separation of genomic gene structure. J. Biol. Chem., 271: 12199–12204, 1996. DNA. Cancer Res., 57: 1250–1254, 1997. 25. Zuo, L., Weger, J., Yang, B., Goldstein, M., Tucker, M. A., Walker, G. J., Hayward, 49. Kitahara, K., Yasui, W., Kuniyasu, H., Yokozaki, H., Akama, Y., Yunotani, S., N., and Dracopoli, N. C. Germline mutations in the p16ink4a binding domain of Hisatsugu, S., Hisatsugu, T., and Tahara, E. Concurrent amplification of cyclinEand CDK4 in familial melanoma. Nat. Genet., 12: 97–99, 1996. CDK2 genes in colorectal carcinomas. Int. J. Cancer, 62: 25–28, 1995. 26. Mostofi, F. K., and Sobin, L. H. International Histological Classification of Tumours 50. Kim, J. H., Kang, M. J., Park, C. U., Kwak. H. J., Hwang, Y., and Koh, G. Y. No. 16. Geneva: WHO, 1977. Amplified CDK2 and cdc2 activities in primary colorectal carcinoma. Cancer (Phila.), 27. American Joint Committee on Cancer. Manual for the Staging of Cancer, Ed. 3 85: 546–553, 1999. Philadelphia: J. B. Lippincott, Inc., 1988. 51. Keyomarsi, K., O’Leary, N., Molnar, G., Lees, E., Fingert, H. J., and Pardee, A. B. 28. Sambrook, J., Fritsch, E. F., and Maniatis, T. (eds.) Molecular Cloning, Ed. 2. Cold Cyclin E, a potential prognostic marker for breast cancer. Cancer Res., 54: 380–385, Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1989. 1994. 29. Geng, Y., Eaton, E. N., Picon, M., Roberts, J. M., Lundberg, A. F., Gifford, A., 52. Leng, X., Connell-Crowly, L., Goodrich, D., and Harper, J. W. S-phase entry upon Sardet, C., and Weinberg, R. A. Regulation of cyclin E transcription by E2Fs and ectopic expression of G1 cyclin dependent kinases in the absence of retinoblastoma retinoblastoma protein. Oncogene, 12: 1173–1180, 1996. protein phosphorylation. Curr. Biol., 7: 7212, 1997

4221

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2001 American Association for Cancer Research. Up-Regulation of Cyclin-dependent Kinase 4/Cyclin D2 Expression but Down-Regulation of Cyclin-dependent Kinase 2/Cyclin E in Testicular Germ Cell Tumors

Bettina A. Schmidt, Achim Rose, Christine Steinhoff, et al.

Cancer Res 2001;61:4214-4221.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/61/10/4214

Cited articles This article cites 43 articles, 13 of which you can access for free at: http://cancerres.aacrjournals.org/content/61/10/4214.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/61/10/4214.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/61/10/4214. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.