Research Articles

Authors' Summary:
The Genomic Landscapes of Human Breast and Colorectal Cancers

Laura D. Wood et al.

How many genes are mutated in a human tumor? Answering this question would have seemed like science fiction just a decade ago. However, as a result of advances in technology, we have been able to answer this question in breast and colorectal cancers: There are ~80 DNA mutations that alter amino acids in a typical cancer. Examining the overall distribution of these mutations in different cancers of the same type leads to a new view of cancer genome landscapes: They are composed of a handful of commonly mutated gene “mountains” but are dominated by a much larger number of infrequently mutated gene “hills.”

The current study expands upon previous work (1) and includes analysis of the sequences of 20,857 transcripts from 18,191 human genes, including the great majority of those that encode proteins. The genes were sequenced in 11 breast and 11 colorectal cancers. Any gene that was mutated in the tumor but not in normal tissue from the same patient was analyzed in 24 additional tumors. Selected genes were further analyzed in another 96 colorectal cancers to better define their mutation frequency and aid subsequent bioinformatic analyses.

Statistical analyses suggested that most of the ~80 mutations in an individual tumor were harmless and that <15 were likely to be responsible for driving the initiation, progression, or maintenance of the tumor. Though the numbers of mutant genes in breast and colorectal cancers were similar, the particular genes that were mutated were quite different, as were the type of mutations found. For example, mutations converting 5′-CpG to 5′-TpG were much more frequent in colorectal than in breast cancers, indicating differences in mutagen exposure or DNA-repair processes.

Resize Image

A two-dimensional map of genes mutated in colorectal cancers, in which a few gene “mountains” are mutated in a large proportion of tumors while most “hills” are mutated infrequently. The mutations in two individual tumors are indicated on the lower map. Note that most mutations are outside hills or mountains and may be harmless.

The mutational landscapes of cancers can be shown on a map on which each gene is represented at a single point (see figure for the landscape for colorectal cancers). The heights of the peaks reflect the mutation frequency of each gene. A few gene “mountains” are mutated in a large proportion of tumors; most genes are mutated in <5% of tumors and are represented as “hills” in the figure. In the lower panel, the mutated genes in two colorectal tumors are indicated by differently colored dots. The mutated genes in the two tumors overlap to only a small extent. These differences are likely to be the basis for the wide variations in tumor behavior and responsiveness to therapy.

Historically, the focus of cancer research has been on the gene mountains, in part because they were the only alterations that could be identified with available technologies. However, our data show that the vast majority of mutations in cancers do not occur in such mountains. This new view of cancer is consistent with the idea that a large number of mutations, each associated with a small fitness advantage, drive tumor progression (2). It is the “hills” and not the “mountains” that dominate the cancer genome landscape.

Are these landscapes hopelessly complex? The large number of "hills" actually reflects alterations in a much smaller number of cell signaling pathways. Indeed, pathways rather than individual genes appear to govern the course of tumorigenesis (3). Accordingly, we devised methods to classify mutant genes into commonly altered pathways. Disruption of a pathway by mutation in any one of its genetic components would presumably lead to similar changes in growth. The <15 driver mutations in an individual tumor likely reflect alterations in a similar number of pathways.

Sequencing alone cannot definitively determine whether a specific gene “hill” actually contributes to tumor formation. We therefore used various bioinformatic and structural analyses to help determine which were pathogenic. Integration with functional studies will also be essential; indeed, several of the candidate cancer genes identified in our study have been independently implicated in tumorigenesis through functional studies reported by others.

In sum, our results make it clear that it is now “easy” to identify the genetic alterations in cancers on a genome-wide scale. It is much more difficult to elucidate the precise role of these alterations in tumorigenesis. The compendium of genetic changes in individual tumors provides new opportunities for individualized diagnosis and treatment of cancer. Taking advantage of these opportunities is the major challenge for the future.

Summary References

1. T. Sjöblom et al., Science 314, 268 (2006).
2. N. Beerenwinkel et al., PLoS Comput. Biol. 3, e225 (2007).
3. B. Vogelstein, K. W. Kinzler, Nat. Med. 10, 789 (2004).

Did you find this summary useful? Let us know your thoughts by taking a brief survey.

Read the Full-Text Research Article