Abstract
To better understand the pathogenesis of glioblastoma multiforme (GBM) and to increase the accuracy of predicting outcomes for patients with this disease, we performed genome‐wide screening for DNA copy‐number aberrations in 22 glioma‐derived cell lines using a custom‐made comparative genomic hybridization‐array. Copy‐number gains were frequently detected at 3q, 7p, 7q, 20q, Xp and Xq, and losses at 4q, 9p, 10p, 10q, 11q, 13q, 14q, 18q, and 22q. Among several non‐random chromosomal aberrations, the gain/amplification of DNA at 5p, which has never been reported before in GBM, was detected with a relatively high ratio (log2 ratio = 0.41–1.19) in four cell lines. Amplification and subsequent overexpression of SKP2, a possible target of amplification within 5p, were detected in four of the 22 cell lines. We also investigated the expression of the gene product in primary GBM by immunohistochemistry, which revealed increased levels of Skp2 in 11 of the 35 tumors examined (31.4%). Heightened expression of Skp2 was associated with shorter overall survival (P = 0.001, logrank test), especially in patients younger than 65 years. These results suggest that overexpression of Skp2 through gene amplification may contribute to the pathogenesis of GBM, and that overabundance of the protein might be a useful prognostic tool in patients with this disease.(Cancer Sci 2005; 96: 676 – 683)
Glioblastoma multiforme (GBM) is the most common malignant and aggressive form of glioma. Despite advances in surgical and clinical neuro‐oncology, prognosis remains poor for patients with GBM: in a recent study, the observed survival rates were 42.4% at 6 months, 17.7% at 1 year, and 3.3% at 2 years.( 1 ) Because various genetic alterations lead to the development of malignant phenotypes of GBM tumors,( 2 ) an increased understanding of the genetics of this disease may lead to the development of better protocols for its clinical management.
According to the multistep model of carcinogenesis in solid tumors, various genetic alterations occur sequentially and accumulate in a cell lineage, at the nucleotide level, as well as at the chromosome level.( 3 ) The first genetic abnormality detected in gliomas was amplification of EGFR (7p12); this occurs in just under 35% of GBM.( 4 ) Amplification and overexpression of PDGFRA (4q12) has been reported in a small subset of GBM.( 5 ) A relatively small region at 12q13‐15, harboring CDK4 and MDM2, is amplified in approximately 13% of GBM. CDK4 is almost always included in the amplicon and is invariably overexpressed when amplified; MDM2 is included in only approximately 8% of such amplicons, but is always overexpressed when amplified.( 6 ) The chromosomal region containing CDKN2A/p16 at 9p21 is homozygously deleted in at least 30–40% of primary (de novo) GBM.( 7 ) Mutation or homozygous deletion of RB1 at 13q14 is found in approximately 13% of GBM,( 7 ) and hypermethylation of its promoter is found in 25% of GBM.( 8 ) Mutation of the TP53 gene has been found in all malignancy grades of astrocytic tumors; approximately 40–70% of diffuse astrocytomas and anaplastic astrocytomas contain p53 mutations, although the frequency is lower (30–40%) in GBM.( 7 ) More than 90% of GBM lose alleles from 10q, especially 10q23, which consistently includes the PTEN gene.( 9 ) In some studies PTEN was mutated in 44% of GBM,( 10 ) and homozygously deleted in 10% of other GBM.( 11 ) However, the relationships between those aberrations, and their prognostic importance, remain uncertain.
A recent technical development, the comparative genomic hybridization‐based array (CGH array), permits reliable detection and direct mapping of copy‐number alterations at high resolution throughout the genome. CGH arrays are in common use now for genome‐wide screening of DNA copy‐numbers in a variety of human cancers.( 12 , 13 , 14 , 15 ) Because the variability of outcomes among patients with GBM might reflect diversity among the genetic changes acquired by the tumor cells, the identification of specific alterations and the discovery of their relationships to clinicopathological characteristics may provide novel insights and lead to effective strategies for identifying GBM that have heightened potential for malignancy.
To explore the genetic alterations that might affect initiation and/or progression of this type of tumor, we examined 22 cell lines derived from malignant gliomas using an in‐house CGH array, the ‘MCG Cancer Array‐800’.( 12 , 13 , 14 , 15 ) Among the losses and gains involving several chromosomal regions, we noted a gain at 5p; this region of amplification had previously been detected in low‐grade astrocytomas, but never before in GBM.( 16 , 17 ) We recently reported that SKP2, the gene encoding S‐phase kinase‐associated protein 2 (Skp2), was a possible target for 5p amplification in lung cancers and biliary tract cancers.( 18 , 19 , 20 ) Skp2 promotes the ubiquitin‐mediated proteolysis of target molecules including p27Kip1, and positively regulates the cell cycle. We considered that overabundance of Skp2 might play an important role in the pathogenesis of GBM as well, and in the work reported here we assessed the significance of its expression as a predictive marker for the prognosis of primary cases of this disease. This approach revealed that overexpression of Skp2 is an independent indicator of poor prognosis in patients with GBM.
Materials and Methods
Cell lines and primary tumors
The 22 human glioma cell lines used in the present study are listed in Table 1. All cell lines except U‐87 MG and U‐373 MG were obtained from the Japanese Collection of Research Bioresources (Osaka, Japan); U‐87 MG and U‐373 MG were obtained from the American Type Culture Collection (Rockville, MD, USA). All lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS) and 100 U/mL penicillin/100 µg/mL streptomycin.
Table 1.
Summary of 22 human glioma cell lines
| No. | Cell line | Histology* | Description |
|---|---|---|---|
| 1 | A‐172 | Glioblastoma | |
| 2 | AM‐38 | Glioblastoma | GFAP and S‐100‐positive, ACNU (a cancer chemotherapy drug)‐sensitive |
| 3 | Becker | Astrocytoma | GFAP‐negative |
| 4 | GB‐1 | Glioma | MDR1 and P‐glycoprotein‐expressing. GFAP, vimentin and fibronectin‐positive |
| 5 | KALS‐1 | Glioma | GFAP, vimentin and CD 13‐positive |
| 6 | KINGS‐1 | Anaplastic astrocytoma | GFAP, S‐100, vimentin, CD 13 and HNK‐1‐positive |
| 7 | KNS‐42 | Glioma | GFAP‐positive, S‐100 and NSE‐negative |
| 8 | KNS‐60 | Glioma | GFAP, S‐100 and NSE‐negative |
| 9 | KNS‐81 | Glioma | GFAP and S‐100‐positive, NSE‐negative |
| 10 | KNS‐89 | Gliosarcoma | GFAP and NSE‐positive, S‐100‐negative |
| 11 | KS‐1 | Gliosarcoma | |
| 12 | Marcus | Astrocytoma | GFAP‐negative |
| 13 | NMC‐G1 | Glioma | FGF‐9‐producing |
| 14 | No. 10 | Anaplastic glioma | GFAP‐positive |
| 15 | No. 11 | Anaplastic glioma | GFAP‐positive |
| 16 | SF126 | Glioblastoma | GFAP‐negative |
| 17 | T98G | Glioblastoma | Anchorage‐independent |
| 18 | U‐251 MG | Astrocytoma | GFAP‐positive |
| 19 | U‐373 MG | Glioblastoma‐astrocytoma | |
| 20 | U‐87 MG | Glioblastoma‐astrocytoma | |
| 21 | YH‐13 | Gliosarcoma | GFAP and S‐100‐positive, ACNU (a cancer chemotherapy drug)‐resistant |
| 22 | YKG‐1 | Gliosarcoma | GFAP and S‐100 protein identified. HSR on marker chromosome identified. α and β‐tumor growth factors expressed |
Histological subtype of the primary tumor from which each cell was derived.
Paraffin‐embedded specimens of primary GBM were obtained from 35 unrelated patients treated at the Department of Neurosurgery, Osaka University Medical School (Osaka, Japan), with written consent from each patient and after approval by the local ethics committee. The mean age of patients was 52.9 years (range 11–81 years). Hematoxylin and eosin staining was performed in the routine way, and the tumors were classified according to World Health Organization criteria. The duration of overall survival was calculated for each patient from the date of primary surgery to the date of the last follow‐up visit or death.
CGH array analysis
We prepared our custom‐made CGH array (MCG Cancer Array‐800) using 800 BAC/PAC clones carrying genes or Sequence‐Tagged Site (STS) markers for elements that we judged might be of potential importance in cancer genesis or progression( 12 , 13 , 14 , 15 ). Hybridizations were carried out as described elsewhere.( 13 , 14 , 15 ) Briefly, test DNA from each GBM cell line and reference DNA from healthy male volunteers were labeled, respectively, with Cy3‐ or Cy5‐dCTP, precipitated together with ethanol in the presence of Cot‐1 DNA, redissolved in a hybridization mix (50% formamide, 10% dextran sulfate, 2× standard saline‐citrate [SSC], 4% sodium dodecylsulfate [SDS], pH 7), and denatured at 75°C for 10 min. After preincubation at 37°C for 30 min, each mixture was applied to array slides and incubated at 42°C for 48–72 h. After hybridization, the slides were washed once in a solution of 50% formamide, 2× SSC (pH 7.0) for 15 min at 50°C, once in 2× SSC, 0.1% SDS for 15 min at 50°C, and once in a 0.1 mol/L sodium phosphate buffer containing 0.1% Nonidet P‐40 (pH 8.0) for 15 min at room temperature, then scanned with a GenePix 4000B (Axon Instruments, Foster City, CA, USA). Acquired images were analyzed with GenePix Pro 4.1 imaging software (Axon Instruments). Fluorescence ratios were normalized so that the mean of the middle third of the log2 ratios across the array was zero. Average ratios that deviated significantly (> 2 SD) from zero were considered abnormal.
Fluorescence in situ hybridization
Metaphase chromosome slides were prepared as described previously.( 21 ) A BAC containing the SKP2 gene (RP11–36A10) was labeled with biotin‐16‐dUTP by nick‐translation, denatured with Cot‐1 DNA, and then hybridized to the chrmosome slides. Fluorescent detection of hybridization signals was carried out as described elsewhere.( 21 ) The cells were counter‐stained with 4′,6‐diamidino‐2‐phenylindole (DAPI).
Quantitative real‐time reverse transcription‐polymerase chain reaction
Levels of SKP2 mRNA were measured by means of a real‐time fluorescence detection method.( 14 ) Single‐stranded cDNA were generated from total RNA using the SuperScript First‐Strand Synthesis System (Invitrogen, Carlsbad, CA, USA). Quantitative real‐time reverse transcription–polymerase chain reaction (RT‐PCR) was performed with an ABI PRISM 7900HT (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol, using CYBR Green and primers specific for SKP2 (SKP2 forward, 5′‐CGCTGCCCACGATCATTTAT‐3′ and SKP2 reverse, 5′‐TGCAACTTGGAACACTGAGACA‐3′). The glyceraldehyde‐3‐phosphate dehydrogenase gene (GAPDH) served as an endogenous control (Applied Biosystems); the expression level of SKP2 mRNA in each sample was normalized on the basis of the respective GAPDH content and recorded as a relative expression level. PCR amplification was performed in duplicate for each sample.
Immunohistochemistry
Indirect immunohistochemistry was performed on formalin‐fixed, paraffin‐embedded tissue sections as described elsewhere,( 20 , 22 ) with minor modification. De‐waxed sections were re‐hydrated in a series of ethanol washes. Antigens were retrieved by autoclave pretreatment in a high‐pH target retrieval solution (Dakocytomation, Carpinteria, CA, USA), and endogenous peroxidase was blocked using 3% hydrogen peroxide. Sections were incubated overnight at 4°C with antihuman Skp2 monoclonal antibody (Zymed Laboratories, diluted 1 : 200) or normal mouse serum. Staining was completed with EnVision+ peroxidase (Dakocytomation). Antigen–antibody reactions were visualized with 3,3′‐diaminobenzidine, and sections were counterstained with hematoxylin. Nuclear staining of Skp2 was scored by evaluating the average of percentages of stained nuclei within three representative areas of each tumor. Expression of Skp2 was graded as either high (≥ 10% of tumor‐cell nuclei showing immunopositivity), or low (< 10% of tumor‐cell nuclei showing immunopositivity, or no staining). Each section was examined at ×200 magnification. The observer who assessed all staining results was blinded to the clinical outcomes of the patients.
Statistical analysis
The Mann–Whitney U‐test was used to compare the level of SKP2 mRNA expression with DNA copy‐number status. Probabilities of survival were calculated by using the Kaplan–Meier method, and statistical differences between groups were evaluated by logrank tests. P‐values of less than 0.05 were considered significant.
Results
CGH array analysis of glioma cell lines
We assessed copy‐number alterations among the 22 glioma cell lines using the same batch of MCG Cancer Array‐800 slides for all of them. Figure 1 documents the frequencies of copy‐number gains and losses across the entire genomes of all 22 cell lines. Table 2 lists the clones showing the most frequent gains or losses in this series, and genes/markers with amplifications or homozygous deletions. Some degree of gain and/or loss was seen in every cell line. Our CGH array predicted frequent copy‐number gains for 3q, 7p, 7q, 20q, Xp and Xq, and frequent losses for 4q, 9p, 10p, 10q, 11q, 13q, 14q, 18q, and 22q. Amplifications were detected in 11 glioma cell lines, and 29 genes (clones) were represented (Table 3). Six of those genes, PDGFRA (4q12), IGFBP7 (4q12), CDH12 (5p14.3), CDH10 (5p14.2), DAB2 (5p13), and EGFR (7p12.3–12.1), were each amplified in more than two cell lines. Not all of the lines contained high‐level amplifications (log2 ratio ≥ 2). However, homozygous deletions (log2 ratio ≤ −2) had occurred in nine of the 22 lines; in particular, MTAP and CDKN2A/p16 at 9p21.3 were homozygously deleted in seven and nine lines, respectively. The copy‐number aberrations revealed through CGH array analysis were mostly consistent with those of our earlier conventional CGH analysis of the same glioma cell lines (data not shown), and with results of other published reports using primary glioma samples, as represented by PDGFRA, EGFR, GLI, and CDK4. ( 23 , 24 ) However, our CGH array analysis disclosed additional regions that had never been revealed in glioma cell lines by conventional CGH, such as small gains and losses. In the U‐251 MG cell line, for example, the CGH array identified independent amplifications of PDGFRA and IGFBP7 at 4q12. In the Marcus cell line, the CGH array identified a cryptic homozygous loss of NF1 at 17q11.2, a feature that had never been detected by any other molecular genetic method, including conventional CGH.
Figure 1.
Genome‐wide frequencies of copy‐number gains (above zero, green) and losses (below zero, red) in 22 glioma‐derived cell lines. Clones are ordered as chromosomes 1–22, X, and Y, and within each chromosome on the basis of the University of California, Santa Cruz mapping position.
Table 2.
Most frequently gained and/or lost clones
| Alteration | Gene | Locus | Frequency (%)* |
|---|---|---|---|
| Gain | EIF4G | 3q27 | 50.0 |
| ETV5 | 3q28 | 47.4 | |
| EGFR | 7p12.3–12.1 | 52.3 | |
| CDC10 | 7p14.2 | 54.5 | |
| IGFBP1 | 7p14–p12 | 56.8 | |
| MYCLK1 | 7p15 | 50.0 | |
| TAX1BP1 | 7p15.2 | 47.7 | |
| TCRG | 7p15–p14 | 54.5 | |
| IL6 | 7p21 | 47.7 | |
| PMS2 | 7p22 | 56.8 | |
| MUC3 | 7q22 | 47.7 | |
| MET | 7q31 | 45.5 | |
| SMOH | 7q31–q32 | 45.5 | |
| BRAF | 7q34 | 45.5 | |
| CDK5 | 7q36 | 52.3 | |
| AR | Xq12 | 47.7 | |
| CuI4B | Xq24 | 47.7 | |
| MCF2 | Xq27 | 45.5 | |
| MAGEA2 | Xq28 | 45.5 | |
| CTAG | Xq28 | 61.4 | |
| Loss | AFP | 4q11–q13 | 40.9 |
| FGF5 | 4q21 | 50.0 | |
| ABCG2 | 4q22 | 52.3 | |
| NFKB | 4q24 | 43.2 | |
| p16 | 9p21 | 63.6 | |
| MTAP | 9p21.3 | 47.7 | |
| BMI1 | 10p13 | 43.2 | |
| PCDH15 | 10q21.1 | 40.9 | |
| PGR | 11q22 | 50.0 | |
| FGF9 | 13q11–q12 | 45.5 | |
| ZNF198 | 13q11–q12 | 47.7 | |
| FLT1 | 13q12 | 45.5 | |
| FLT3 | 13q12 | 54.5 | |
| BRCA2 | 13q12–13 | 54.5 | |
| RB1 | 13q14 | 47.7 | |
| PIBF1 | 13q22.1 | 43.2 | |
| KLF12 | 13q22.1 | 52.3 | |
| HNF3A | 14q12 | 47.7 | |
| MBIP | 14q12 | 52.3 | |
| FKHL1 | 14q13 | 43.2 |
Alterations were defined by log2 ratio thresholds of 0.4 and −0.4 for copy‐number gain and loss, respectively. Using this threshold, we generated a frequency table. In this table, the 20 most frequently gained and lost clones are shown, ordered according to chromosomal positions.
Table 3.
Genes showing amplification and homozygous deletion among 22 glioma cell lines
| Alteration | Genes | Locus | No. | Average ratio (log2 ratio) |
|---|---|---|---|---|
| High‐level amplification(2.0 ≤ log2) | None | |||
| Amplification(1.0 < log2 < 2.0) | ALX | 1p13.3 | 1 (4.5%) | 1.11 |
| MUC1 | 1q21 | 1 (4.5%) | 1.02 | |
| ARHGEF2 | 1q21 | 1 (4.5%) | 1.13 | |
| PMF1 | 1q23.1 | 1 (4.5%) | 1.04 | |
| NTRK1 | 1q23–24 | 1 (4.5%) | 1.30 | |
| ETV5 | 3q28 | 1 (4.5%) | 1.04 | |
| MUC4 | 3q29 | 1 (4.5%) | 1.40 | |
| PDGFRA | 4q12 | 2 (9.1%) | 1.76 | |
| IGFBP7 | 4q12 | 2 (9.1%) | 1.19 | |
| TERT | 5p15 | 1 (4.5%) | 1.20 | |
| CDH12 | 5p14.3 | 2 (9.1%) | 1.14 | |
| CDH10 | 5p14.2 | 2 (9.1%) | 1.04 | |
| PC4 | 5p13 | 1 (4.5%) | 1.16 | |
| SKP2 | 5p13 | 1 (4.5%) | 1.19 | |
| DAB2 | 5p13 | 2 (9.1%) | 1.23 | |
| RREB1 | 6p25 | 1 (4.5%) | 1.04 | |
| EEF1E1 | 6p24.3 | 1 (4.5%) | 1.16 | |
| TFAP2A | 6p24 | 1 (4.5%) | 1.09 | |
| TPMT | 6p22.3 | 1 (4.5%) | 1.20 | |
| E2F3 | 6p22 | 1 (4.5%) | 1.10 | |
| PMS2 | 7p22 | 1 (4.5%) | 1.14 | |
| EGFR | 7p12.3–12.1 | 2 (9.1%) | 1.19 | |
| CDK6 | 7q21–q22 | 1 (4.5%) | 1.11 | |
| PRIM1 | 12q13 | 1 (4.5%) | 1.31 | |
| GL1 | 12q14 | 1 (4.5%) | 1.45 | |
| CDK4 | 12q14 | 1 (4.5%) | 1.48 | |
| FUS | 16p11.2 | 1 (4.5%) | 1.44 | |
| CYLD | 16q12–q13 | 1 (4.5%) | 1.09 | |
| GRB2 | 17q24–25 | 1 (4.5%) | 1.15 | |
| Homozygous deletion(log2 ≤−2.0) | MTAP | 9p21.3 | 7 (31.8%) | −2.57 |
| CDKN2A(P16) | 9p21.3 | 7 (40.9%) | −2.70 |
Amplification and subsequent overexpression of SKP2 in glioma cell lines
Our CGH array revealed gains of DNA at 5p in four cell lines (KNS‐42, KNS‐60, Marcus, SF126). Because we had recently identified SKP2 (5p13.2) as a possible target for amplification in other tumors,( 18 , 19 , 20 ) we focused further examination on the copy‐number and expression status of this gene. Copy‐numbers of SKP2 were determined by fluorescence in situ hybridization (FISH) in the two cell lines (KNS‐60 and SF126) that had shown the highest ratios during CGH array analysis of a BAC (RP11‐36A10) containing SKP2 sequence (Fig. 2a). KNS‐60 cells showed 10 FISH signals and SF126 cells showed 11 (Fig. 2b).
Figure 2.
(a) Genetic changes observed on chromosome 5 in the KNS‐60 and SF126 cell lines. Comparative genomic hybridization‐based (CGH) array analysis identified amplification of almost the entire short arm of chromosome 5, but no amplification in the long arm. A red arrow indicates the clone containing the SKP2 gene. (b) Representative images of fluorescence in situ hybridization (FISH) experiments using BAC RP11–36A10 (green) on metaphase chromosomes from the KNS‐60 and SF126 cell lines. This BAC generated 10 and 11 signals, respectively, in the KNS‐60 and SF126 cells.
Next we examined the expression of SKP2 by using real‐time quantitative RT‐PCR experiments. The pattern of SKP2 expression (Fig. 3) accorded with the gene's copy‐number gain pattern determined by CGH array analysis. The correlation between copy‐number gain and expression status was significant (Mann–Whitney U‐test, P = 0.041).
Figure 3.
Relative expression levels of SKP2 mRNA determined by real‐time reverse transcription–polymerase chain reaction. Results are presented as the ratio between SKP2 and a reference gene (GAPDH). Significant correlation (P = 0.0411) between copy‐number gains and relative expression levels of SKP2 was observed in 22 glioma cell lines. Copy number ratios (log2) of four cell lines with gain of SKP2 are shown in parentheses.
Correlation between Skp2 expression in primary GBM and patient survival
We examined the expression of Skp2 protein in primary GBM using immunohistochemistry (Fig. 4a,b), and compared the results with clinicopathological features, including survival, among 35 patients with GBM. Figure 4c–e shows Kaplan–Meier survival curves for all 35 cases. The mean age of 11 patients whose tumors showed positive Skp2 staining in ≥ 10% of the tumor cells examined (Skp2‐high group) was 53.6 (range, 23–81 years), and that of 24 patients with staining in < 10% of the tumor cells (Skp2‐low group) was 52.6 (range, 11–81 years). There was no significant difference between the mean age of patients in the Skp2‐high group and the mean age of the Skp2‐low group. Overall, the Skp2‐high group showed significantly shorter survival times than Skp2‐low group (Fig 4c, P = 0.0010). Furthermore, the Skp2‐high group had even more significantly shorter overall survival times than the Skp2‐low group when only patients younger than 65 years were considered (Fig 4d, P = 0.0004), but no statistical significance was found between the Skp2‐high group and the Skp2‐low group for patients older than 65 years (Fig. 4e, P = 0.6521).
Figure 4.
Immunohistochemical presentation of Skp2 expression in (a) a glioblastoma multiforme (GBM) tumor with high expression of Skp2, where nuclear Skp2 staining was positive in 32.3% of tumor cells on average; and (b) a GBM tumor with low Skp2 expression, where nuclear staining for Skp2 was positive in an average of only 7.3% of tumor cells. (c) Kaplan–Meier curves for overall survival of all patients (n = 35) with GBM after surgery, stratified according to the nuclear level of Skp2 (P = 0.0010). (d) Kaplan–Meier curves for overall survival of patients younger than 65 years (n = 25); patients with high Skp2 expression showed significantly poorer prognosis in overall survival than patients with low Skp2 expression (P = 0.0004). (e) Kaplan–Meier curves for overall survival of patients older than 65 years (n = 10); Skp2 expression showed no correlation with prognosis in this group (P = 0.6521).
Discussion
Results from the analysis of 22 glioma cell lines using our MCG Cancer Array‐800 were mostly in accord with conventional CGH analyses.( 23 , 24 ) The MCG Cancer Array‐800 sensitively detected gain of EGFR (52.3%), loss of CDKN2A/p16 (63.6%) and RB1 (47.7%), and all other known genetic alterations characteristic of gliomas. The results suggest that these 22 cell lines are representative of gliomas in terms of aberrant genes. However, our analysis also detected amplification of 5p, which has never been reported in glioma cell lines. Gain (including amplification) of 5p was not frequent in our panel of cell lines (4/22, 18.2%), but for those four the ratios in the CGH array analysis were relatively high. Amplification of 5p has been reported in non‐small‐cell lung cancers, small‐cell lung cancers, squamous cell carcinomas of the head and neck, carcinomas of the uterine cervix, osteosarcomas, and malignant fibrous histiocytomas.( 18 , 25 ) Thus we focused on 5p as a region likely to harbor target gene(s) for amplification events involved in the tumorigenesis of GBM.
Recently we reported that SKP2 (5p13.2) was a possible target for amplifications seen in small‐cell lung cancers, non‐small‐cell lung cancers, and biliary tract cancers.( 18 , 19 , 20 ) Its product, Skp2, is an F‐box substrate‐recognition subunit of the SCF ubiquitin‐protein ligase complex, which regulates progression of the cell cycle by targeting regulators such as p27Kip1 for ubiquitin‐mediated degradation. Decreased levels of p27Kip1 seem to be associated with high aggressiveness and poor prognosis in a variety of cancers.( 26 , 27 , 28 , 29 , 30 , 31 ) Those findings prompted us to assess the possibility that amplified and subsequently overexpressed SKP2 might exert an oncogenic role in GBM as well, and that overexpression of Skp2 protein might be a useful prognostic tool for patients with this disease. In the present study we successfully identified increased Skp2 expression through its amplification in cell lines, and found a positive correlation between Skp2 overexpression in primary GBM and poorer outcomes among patients with the disease. That correlation was the most striking finding in our study. The molecular mechanisms underlying increased levels of Skp2 protein in many types of cancers remain unclear. The abundance of Skp2 protein is regulated by two independent mechanisms mediated at the transcriptional and post‐transcriptional levels; the predominant mechanism may vary with cell type.( 32 , 33 , 34 ) Our results, in accord with previous reports,( 18 , 19 , 20 ) indicate that an increase in SKP2 genomic copy‐number is an important mechanism for Skp2 overexpression. However, we previously reported that some biliary tract carcinoma cell lines showed discrepancies between SKP2 DNA copy‐number and the level of Skp2 protein,( 20 ) an indication that a post‐transcriptional regulatory pathway of Skp2 expression also functions in tumors. Mamillapalli et al. reported that PTEN‐deficiency in mouse embryonic stem cells causes a decrease of p27 levels and a concomitant increase of Skp2, suggesting that an increase in Skp2 protein without any gain in copy‐number of the gene might be caused by altered PTEN.( 35 ) Because alteration of PTEN is a major event in GBM,( 10 , 11 ) expression of Skp2 protein could be a more accurate marker for the prognosis of patients with GBM tumors than the copy‐number of the gene itself.
Attempts have been made to identify more reliable predictors of prognosis and response to therapy of malignant gliomas. Raffel et al. correlated the status of the PTEN, CDK4, CDKN2A/p14 ARF , TP53, MDM2, and EGFR genes with patient survival in a series of pediatric malignant astrocytomas, and determined that PTEN mutation was significantly associated with shorter survival.( 36 ) Backlund et al. found that abnormalities in any of four genes encoding components of the Rb1 pathway (CDKN2A, CDKN2B, RB1, and CDK4) were associated with shorter survival of patients with GBM.( 37 ) However, most of those findings remain controversial or inconclusive. Although our observations will have to be confirmed in a larger number of patients with GBM, the present study demonstrated that Skp2 might play an important role in the progression of this type of tumor.
Recently Lee and McCormick reported that downregulation of Skp2 by means of small interfering RNA induced apoptosis in glioma cell lines, suggesting that Skp2 inhibitors and/or Skp2‐regulatory sequences might become features of a useful therapeutic protocol for the treatment of GBM.( 38 ) Although several compounds inhibiting proteasome degradation are already being tested as therapeutic reagents for malignancies,( 39 ) further study will be needed to determine whether reagents directed at the Skp2 molecule itself can provide a more selective target for diagnostic and therapeutic approaches for GBM.
Acknowledgments
We thank Professor Yusuke Nakamura (Human Genome Center, The Institute of Medical Science, The University of Tokyo) for his encouragement. We are also grateful to Ai Watanabe for technical assistance. This work was supported by Grants‐in‐Aid for Scientific Research (B) and Scientific Research on Priority Areas (C) and a Center of Excellence Program for Research on Molecular Destruction and Reconstruction of Tooth and Bone from the Ministry of Education, Culture, Sports, Science, and Technology of Japan; and from Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Corporation (JST).
References
- 1. Ohgaki H, Dessen P, Jourde B et al. Genetic pathways to glioblastoma: a population‐based study. Cancer Res 2004; 64: 6892–9. [DOI] [PubMed] [Google Scholar]
- 2. Kleihues P, Ohgaki H. Primary and secondary glioblastomas: from concept to clinical diagnosis. Neuro-oncology 1999; 1: 44–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Vogelstein B, Kinzler KW. The multistep nature of cancer. Trends Genet 1993; 9: 138–41. [DOI] [PubMed] [Google Scholar]
- 4. Wong AJ, Bigner SH, Bigner DD et al. Increased expression of the epidermal growth factor receptor gene in malignant gliomas is invariably associated with gene amplification. Proc Natl Acad Sci USA 1987; 84: 6899–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Fleming TP, Sexena A, Clark WC et al. Amplification and/or overexpression of platelet‐derived growth factor receptors and epidermal growth factor receptor in human glial tumors. Cancer Res 1992; 52: 4550–3. [PubMed] [Google Scholar]
- 6. Reifenberger G, Ichimura K, Reifenberger J et al. Refined mapping of 12q13‐q15 amplicons in human malignant gliomas suggests CDK4/SAS and MDM2 as independent amplification targets. Cancer Res 1996; 56: 5141–5. [PubMed] [Google Scholar]
- 7. Ichimura K, Bolin MB, Goike HM et al. Deregulation of the p14ARF/MDM2/p53 pathway is a prerequisite for human astrocytic gliomas with G1‐S transition control gene abnormalities. Cancer Res 2000; 60: 417–24. [PubMed] [Google Scholar]
- 8. Nakamura M, Yonekawa Y, Kleihues P et al. Promoter hypermethylation of the RB1 gene in glioblastomas. Lab Invest 2001; 81: 77–82. [DOI] [PubMed] [Google Scholar]
- 9. Steck PA, Pershouse MA, Jasser SA et al. Identification of a candidate tumour suppressor gene, MMAC1, at chromosome 10q23.3 that is mutated in multiple advanced cancers. Nat Genet 1997; 15: 356–62. [DOI] [PubMed] [Google Scholar]
- 10. Wang SI, Puc J, Li J et al. Somatic mutations of PTEN in glioblastoma multiforme. Cancer Res 1997; 57: 4183–6. [PubMed] [Google Scholar]
- 11. Smith JS, Tachibana I, Passe SM et al. PTEN mutation, EGFR amplification, and outcome in patients with anaplastic astrocytoma and glioblastoma multiforme. J Natl Cancer Inst 2001; 93: 1246–56. [DOI] [PubMed] [Google Scholar]
- 12. Inazawa J, Inoue J, Imoto I. Comparative genomic hybridization (CGH)‐arrays pave the way for identification of novel cancer‐related genes. Cancer Sci 2004; 95: 559–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Sonoda I, Imoto I, Inoue J et al. Frequent silencing of low density lipoprotein receptor‐related protein 1B (LRP1B) expression by genetic and epigenetic mechanisms in esophageal squamous‐cell carcinoma. Cancer Res 2004; 64: 3741–7. [DOI] [PubMed] [Google Scholar]
- 14. Takada H, Imoto I, Tsuda H et al. Screening of DNA copy‐number aberrations in gastric cancer cell lines by array‐based comparative genomic hybridization. Cancer Sci 2005; 96: 100–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Izumi H, Inoue J, Yokoi S et al. Frequent silencing of DBC1 is by genetic or epigenetic mechanisms in non‐small cell lung cancers. Hum Mol Genet 2005; 14: 997–1007. [DOI] [PubMed] [Google Scholar]
- 16. Hirose Y, Aldape KD, Chang S et al. Grade II astrocytomas are subgrouped by chromosome aberrations. Cancer Genet Cytogenet 2003; 142: 1–7. [DOI] [PubMed] [Google Scholar]
- 17. Cowell JK, Matsui S, Wang YD et al. Application of bacterial artificial chromosome array‐based comparative genomic hybridization and spectral karyotyping to the analysis of glioblastoma multiforme. Cancer Genet Cytogenet 2004; 151: 36–51. [DOI] [PubMed] [Google Scholar]
- 18. Yokoi S, Yasui K, Saito‐Ohara F et al. A novel target gene, SKP2, within the 5p13 amplicon that is frequently detected in small cell lung cancers. Am J Pathol 2002; 161: 207–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Yokoi S, Yasui K, Mori M et al. Amplification and overexpression of SKP2 are associated with metastasis of non‐small‐cell lung cancers to lymph nodes. Am J Pathol 2004; 165: 175–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sanada T, Yokoi S, Arii S et al. Skp2 overexpression is a p27Kip1‐independent predictor of poor prognosis in patients with biliary tract cancers. Cancer Sci 2004; 95: 969–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ariyama Y, Sakabe T, Shinomiya T et al. Identification of amplified DNA sequences on double minute chromosomes in a leukemic cell line KY821 by means of spectral karyotyping and comparative genomic hybridization. J Hum Genet 1998; 43: 187–90. [DOI] [PubMed] [Google Scholar]
- 22. Imoto I, Yang ZQ, Pimkhaokham A et al. Identification of cIAP1 as a candidate target gene within an amplicon at 11q22 in esophageal squamous cell carcinomas. Cancer Res 2001; 61: 6629–34. [PubMed] [Google Scholar]
- 23. Koschny R, Koschny T, Froster UG et al. Comparative genomic hybridization in glioma: a meta‐analysis of 509 cases. Cancer Genet Cytogenet 2002; 135: 147–59. [DOI] [PubMed] [Google Scholar]
- 24. Suzuki T, Maruno M, Wada K et al. Genetic analysis of human glioblastomas using a genomic microarray system. Brain Tumor Pathol 2004; 21: 27–34. [DOI] [PubMed] [Google Scholar]
- 25. Knuutila S, Bjorkqvist AM, Autio K et al. DNA copy number amplifications in human neoplasms: review of comparative genomic hybridization studies. Am J Pathol 1998; 152: 1107–23. [PMC free article] [PubMed] [Google Scholar]
- 26. Gstaiger M, Jordan R, Lim M et al. Skp2 is oncogenic and overexpressed in human cancers. Proc Natl Acad Sci USA 2001; 98: 5043–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kudo Y, Kitajima S, Sato S et al. High expression of S‐phase kinase‐interacting protein 2, human F‐box protein, correlates with poor prognosis in oral squamous cell carcinomas. Cancer Res 2001; 61: 7044–7. [PubMed] [Google Scholar]
- 28. Masuda TA, Inoue H, Sonoda H et al. Clinical and biological significance of S‐phase kinase‐associated protein 2 (Skp2) gene expression in gastric carcinoma: modulation of malignant phenotype by Skp2 overexpression, possibly via p27 proteolysis. Cancer Res 2002; 62: 3819–25. [PubMed] [Google Scholar]
- 29. Chiarle R, Fan Y, Piva R et al. S‐phase kinase‐associated protein 2 expression in non‐Hodgkin's lymphoma inversely correlates with p27 expression and defines cells in S phase. Am J Pathol 2002; 160: 1457–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Hui AM, Cui X, Makuuchi M et al. Decreased p27 (Kip1) expression and cyclin D1 overexpression, alone and in combination, influence recurrence and survival of patients with resectable extrahepatic bile duct carcinoma. Hepatology 1999; 30: 1167–73. [DOI] [PubMed] [Google Scholar]
- 31. Hui AM, Li X, Shi YZ et al. p27(Kip1) expression in normal epithelia, precancerous lesions, and carcinomas of the gallbladder: association with cancer progression and prognosis. Hepatology 2000; 31: 1068–72. [DOI] [PubMed] [Google Scholar]
- 32. Wirbelauer C, Sutterluty H, Blondel M et al. The F‐box protein Skp2 is a ubiquitylation target of a Cul1‐based core ubiquitin ligase complex: evidence for a role of Cul1 in the suppression of Skp2 expression in quiescent fibroblasts. EMBO J 2000; 19: 5362–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Imaki H, Nakayama K, Delehouzee S et al. Cell cycle‐dependent regulation of the Skp2 promoter by GA‐binding protein. Cancer Res 2003; 63: 4607–13. [PubMed] [Google Scholar]
- 34. Shigemasa K, Gu L, O'Brien TJ et al. Skp2 overexpression is a prognostic factor in patients with ovarian adenocarcinoma. Clin Cancer Res 2003; 9: 1756–63. [PubMed] [Google Scholar]
- 35. Mamillapalli R, Gavrilova N, Mihaylova VT et al. PTEN regulates the ubiquitin‐dependent degradation of the CDK inhibitor p27 (KIP1) through the ubiquitin E3 ligase SCF (SKP2). Curr Biol 2001; 11: 263–7. [DOI] [PubMed] [Google Scholar]
- 36. Raffel C, Frederick L, O'Fallon JR et al. Analysis of oncogene and tumor suppressor gene alterations in pediatric malignant astrocytomas reveals reduced survival for patients with PTEN mutations. Clin Cancer Res 1999; 5: 4085–90. [PubMed] [Google Scholar]
- 37. Backlund LM, Nilsson BR, Goike HM et al. Short postoperative survival for glioblastoma patients with a dysfunctional Rb1 pathway in combination with no wild‐type PTEN. Clin Cancer Res 2003; 9: 4151–8. [PubMed] [Google Scholar]
- 38. Lee SH, McCormick F. Downregulation of Skp2 and p27/Kip1 synergistically induces apoptosis in T98G glioblastoma cells. J Mol Med 2005; 83: 296–307. [DOI] [PubMed] [Google Scholar]
- 39. Mitchell BS. The proteasome: an emerging therapeutic target in cancer. N Engl J Med 2003; 348: 2597–8. [DOI] [PubMed] [Google Scholar]