Localized alterations in copy number are found widely across the genomes of malignancies, almost all of them acquired during the lifetime of the host. These somatic copy-number alterations (SCNA) indicate regions of gain or loss, which may play a role in pathogenesis; although, given their high frequency and broad distribution, it may be hard to distinguish those that are causal from secondary phenomena. This paper, from a large collaborative group led by Matthew Meyerson and others in Boston, used an Affymetrix technology to catalogue patterns of SCNA in 3,131 specimens across a range of tumor types. These came from primary tumor material in most cases, with a minority derived from cell lines or short-term cultures.
A mean of 24 gains and 18 losses were found in each sample, although higher frequencies were seen in some solid tumors. The alterations could be separated into focal changes (median of 1.8 Mbases long) or those representing gain or loss of the whole arm of a chromosome. In a typical malignancy, some 25 percent of the genome was affected by arm-length SCNAs and 10 percent by focal changes, with around 2 percent overlap. The amplitude of change was generally a single copy, although some showed much higher amplifications. By determining which focal areas of SCNA appeared at a frequency higher than predicted by their size, the investigators identified 158 regions of apparently significant change: 76 amplifications and 82 deletions. The amplifications contained a median of 6.5 genes (ranging from 0-143), with 25 containing previously identified oncogenes. The deletions contained a median of seven genes with a smaller proportion (11 percent) containing known tumor suppressors, perhaps because deletions as secondary events are more likely to occur in gene-poor regions. Grouping involved genes by family revealed that several known pathogenic targets, including kinases, cell-cycle regulators, and MYC family members, were frequently affected. Apoptosis-modulating genes stood out as particularly often involved; both MCL1 and BCL2L1 (BCL-XL) were amplified in a relatively large proportion of samples. NF-κB pathway genes were also frequently found. To test the functional significance of MCL1 and BCL2L1 amplifications they used shRNA to knock down their expression in cell lines and demonstrated growth retardation and induction of apoptosis, which were more pronounced in lines with the relevant amplification, supporting the hypothesis that these are pathogenetic events.
In Brief
Genome-based array technologies can now generate massive amounts of data on copy-number changes in malignant cells, and the substantial overlap among different tumor types is quite revealing, suggesting common pathways of transformation even in quite disparate tissues. Genes such as MCL1 and BCL2L1, long known to be abnormal in lymphomas, are more widely implicated in a variety of cancers. The fact that this mapping has brought up genes known to play a role in malignant transformation is evidence for its accuracy, and the functional studies on the two anti-apoptotic genes confirm the apparent role of amplification in maintaining growth potential in the cells where this occurs, even at relatively low copy number. The real interest in this approach is in the potential to identify new pathogenic events in malignancy. More than three-quarters of the 158 peak regions of SCNA did not contain known targets, suggesting that much more information may be forthcoming, even if one discounts the deletions that are simply evidence of “noise” and the loss of regions that can be readily tolerated. This approach will be particularly powerful when combined with deep sequence analysis of somatic mutations in tumors, as is now being undertaken in the International Cancer Genome Consortium. The convergence of these techniques will soon yield an immensely detailed map of the events driving malignant transformation.
Competing Interests
Dr. Johnson indicated no relevant conflicts of interest.
