De Bortoli Massimiliano, Castellino Robert C, Lu Xin-Yan, Deyo Jeffrey, Sturla Lisa Marie, Adesina Adekunle M, Perlaky Laszlo, Pomeroy Scott L, Lau Ching C, Man Tsz-Kwong, Rao Pulivarthi H, Kim John Y H
Texas Children's Cancer Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA.
BMC Cancer. 2006 Sep 12;6:223. doi: 10.1186/1471-2407-6-223.
Medulloblastoma is the most common malignant brain tumor of childhood. Improvements in clinical outcome require a better understanding of the genetic alterations to identify clinically significant biological factors and to stratify patients accordingly. In the present study, we applied cytogenetic characterization to guide the identification of biologically significant genes from gene expression microarray profiles of medulloblastoma.
We analyzed 71 primary medulloblastomas for chromosomal copy number aberrations (CNAs) using comparative genomic hybridization (CGH). Among 64 tumors that we previously analyzed by gene expression microarrays, 27 were included in our CGH series. We analyzed clinical outcome with respect to CNAs and microarray results. We filtered microarray data using specific CNAs to detect differentially expressed candidate genes associated with survival.
The most frequent lesions detected in our series involved chromosome 17; loss of 16q, 10q, or 8p; and gain of 7q or 2p. Recurrent amplifications at 2p23-p24, 2q14, 7q34, and 12p13 were also observed. Gain of 8q is associated with worse overall survival (p = 0.0141), which is not entirely attributable to MYC amplification or overexpression. By applying CGH results to gene expression analysis of medulloblastoma, we identified three 8q-mapped genes that are associated with overall survival in the larger group of 64 patients (p < 0.05): eukaryotic translation elongation factor 1D (EEF1D), ribosomal protein L30 (RPL30), and ribosomal protein S20 (RPS20).
The complementary use of CGH and expression profiles can facilitate the identification of clinically significant candidate genes involved in medulloblastoma growth. We demonstrate that gain of 8q and expression levels of three 8q-mapped candidate genes (EEF1D, RPL30, RPS20) are associated with adverse outcome in medulloblastoma.
髓母细胞瘤是儿童期最常见的恶性脑肿瘤。临床疗效的改善需要更好地了解基因改变，以识别具有临床意义的生物学因素，并据此对患者进行分层。在本研究中，我们应用细胞遗传学特征来指导从髓母细胞瘤基因表达微阵列谱中识别具有生物学意义的基因。
我们使用比较基因组杂交（CGH）分析了71例原发性髓母细胞瘤的染色体拷贝数畸变（CNA）。在我们之前通过基因表达微阵列分析的64例肿瘤中，27例被纳入我们的CGH系列。我们分析了CNA和微阵列结果与临床疗效的关系。我们使用特定的CNA过滤微阵列数据，以检测与生存相关的差异表达候选基因。
在我们的系列中检测到的最常见病变涉及17号染色体；16q、10q或8p缺失；以及7q或2p增益。还观察到2p23-p24、2q14、7q34和12p13的反复扩增。8q增益与较差的总生存率相关（p = 0.0141），这并不完全归因于MYC扩增或过表达。通过将CGH结果应用于髓母细胞瘤的基因表达分析，我们在64例更大的患者群体中鉴定出三个位于8q的基因，它们与总生存率相关（p < 0.05）：真核翻译延伸因子1D（EEF1D）、核糖体蛋白L30（RPL30）和核糖体蛋白S20（RPS20）。
CGH和表达谱的互补使用可以促进识别参与髓母细胞瘤生长的具有临床意义的候选基因。我们证明8q增益和三个位于8q的候选基因（EEF1D、RPL30、RPS20）的表达水平与髓母细胞瘤的不良预后相关。
