The Cancer Genome Project intends to search every human gene for cancer-related mutations. Its first success is the discovery of such mutations in the BRAF gene.
The growth and proliferation of our cells are controlled, in part, by signals that pass from the outside of the cell to its nucleus via a cascade of molecules known as the mitogen-activated protein kinase (MAPK) pathway. Alterations in the DNA sequence of genes encoding key components of this pathway contribute to the development of many human cancers. On page 949 of this issue, Davies and colleagues1 report that the BRAF gene, which codes for a protein in the MAPK cascade, is activated by mutation in some cancers — most notably in melanomas. This is the first important result to emerge from the Sanger Institute's Cancer Genome Project.
Cancer is a disease of the genome, triggered by the accumulation of genetic errors that eventually transform a normal cell into a tumour cell2. Such mutations might inactivate genes that normally oppose tumour development, or activate genes that drive cell growth or interfere with cell differentiation or death. Identifying the genes that are altered in the stepwise progression to malignancy has become one of the central goals of cancer research. But finding a single disease-associated gene within the genome has been aptly compared to finding a needle in a haystack, and cancer geneticists have generally tried to simplify the task by narrowing their search to part of the haystack before checking candidate genes for mutations.
Although the availability of the human genome sequence has not changed these unfavourable proportions, the haystack has at least been put in order, vastly accelerating the pace of disease-gene discovery. Sequencing the genome required the highly efficient automation of data generation (robotics) and computer analysis (bioinformatics)3,4. Beyond producing the sequence itself, this biotechnological revolution has resulted in a significant culture shift for genome biologists, who now have these finely honed tools at their disposal.
The Cancer Genome Project has taken advantage of this technological prowess, and the human genome sequence itself, to develop a radical approach to discovering cancer-linked genes. Rather than using strategies directed at individual tumour types or chromosomal regions, Davies and colleagues have taken the bold tack of analysing the entire genome in DNA samples representing a wide spectrum of human cancers. As part of this project, they intend to examine every human gene for cancer-related (oncogenic) mutations.
Signal-transduction genes such as those of the RAS family, which encode components of the MAPK pathway, are frequently mutated in cancers. So Davies et al.1 began their screen with other genes in this pathway, looking first at cultured cell lines derived from human tumours. They hit pay dirt with BRAF, which mediates growth signalling at a level just below RAS (Fig. 1). This gene encodes a serine/threonine kinase — an enzyme that modulates the function of other proteins by transferring phosphate groups to them. It is itself activated by such phosphorylation.
Most (80%) of the BRAF mutations detected by Davies et al. resulted in a single amino-acid change in the region of the enzyme that catalyses phosphorylation. This alteration involved the replacement of a neutral amino acid (valine at position 599) by a negatively charged one (usually glutamic acid). The mutation probably leads to constitutive activation of the MAPK pathway, by mimicking the transient phosphorylation of threonine 598 and serine 601 that occurs during normal signalling5. With this growth switch stuck in the 'on' position, tumour development is favoured.
Davies et al. found this mutation in several cancer types, mainly those known to have RAS mutations. For most cancers the rate of BRAF mutation was modest, occurring in fewer than 20% of the cell lines studied. But in melanoma cell lines, the authors found BRAF mutations at the staggering rate of 59%. Davies et al. strengthened these data by finding BRAF mutations in 80% of short-term melanoma cell cultures and 66% of uncultured melanomas, strongly suggesting that the high mutation frequency in cell lines was not an artefact of cell culture. This represents a clear step forward in understanding the biology of melanoma, which has so far yielded few of its genetic secrets.
Melanoma is an aggressive skin cancer that is derived from pigment cells (melanocytes). The disease can be cured surgically if caught early, but once melanoma cells have spread to other parts of the body they are resistant to most current treatments. Until now, CDKN2A — which encodes p16, an inhibitor of the cell-division cycle — was the most commonly mutated gene in melanoma, being inactivated by a variety of mechanisms in some 25% of sporadic melanomas6. Less frequent molecular alterations have been reported in NRAS, a member of the RAS family, and the tumour-suppressor gene PTEN.
BRAF is part of key melanocyte-specific MAPK pathways, such as that involving melanocyte-stimulating hormone, which activates the melanocortin-1 receptor on the cell surface. This may account for the high frequency of BRAF mutations in melanomas relative to other cancers7. Could BRAF be a breakthrough target for future treatments for melanoma? This could be so, and the intense attention that will now be directed at this gene will certainly lead to important insights. However, the identification of RAS mutations 20 years ago has not yet led to clinically useful RAS-targeted therapies.
The discovery of BRAF mutations1 is a convincing validation of the Cancer Genome Project's high-throughput strategy. The mutations might not have been discovered for years without this type of approach. Now the project's investigators face the challenge of going through the rest of the genomic haystack. For their success to continue, they must be certain that their screen is sensitive, thorough and efficient.
Challenges remain; for instance, the process of locating all genes in the human genome (annotation) still continues. Indeed, the precise number of genes is still a matter for debate, somewhat obscuring the goal of testing them all. Also, this approach depends on having DNA from cultured tumour cells; that may be difficult to achieve for some important tumours, such as prostate and pancreatic cancer, which are not easy to culture.
So where does this gargantuan effort leave the many researchers studying the genetics of cancer progression? Although it is too soon to declare focused, smaller-scale research obsolete, the Cancer Genome Project — as yet unrivalled — is likely to produce a steady flow of discoveries. Nonetheless, the identification of a cancer-specific mutation is but the first step in a lengthy process. Further studies are needed to determine when BRAF mutations occur during tumour evolution, what their biochemical effects are, and the possible implications for treating cancer.
Davies, H. et al. Nature 417, 949–954 (2002); advance online publication, 9 June 2002 (doi:10.1038/nature00766).
Balmain, A. Nature Rev. Cancer 1, 77–82 (2001).
Lander, E. S. et al. Nature 409, 860–921 (2001).
Venter, J. C. et al. Science 291, 1304–1351 (2001).
Zhang, B. H. & Guan, K. L. EMBO J. 19, 5429–5439 (2000).
Pollock, P. M., Welch, J. & Hayward, N. K. Cancer Res. 61, 1154–1161 (2001).
Busca, R. et al. EMBO J. 19, 2900–2910 (2000).
About this article
Nature Reviews Cancer (2004)
Nature Reviews Drug Discovery (2003)