Genetic mutations predict which cancers will respond to treatment
THE International Cancer Genome Consortium, an alliance of laboratories that is trying to produce a definitive list of the genetic mutations that cause cancer, is accumulating data at an astonishing rate. About 3,000 individual breast tumours, for example, have now had their genotypes published. But these data will not, by themselves, help patients. For that, they have to be collected in the context of a drug trial. And this is just what Matthew Ellis and his colleagues at Washington University in St Louis have done for women suffering from breast cancer. Their methods, if they prove to work for other cancers too, may revolutionise treatment.
国际癌症基因组协作组（THE International Cancer Genome Consortium）是试图建立一份会引起癌症的基因突变完整清单的实验室联盟，它积累数据的速度让人吃惊。例如，它已经发表了大约3000种不同的乳房肿瘤的基因型。但光凭这些数据本身无法帮助患者。要医治病人，人们必须结合药物试验采集数据。而这正是在圣路易斯市的华盛顿大学（Washington University in St Louis）工作的马修?埃利斯（Matthew Ellis）及其同事们为罹患乳腺癌的妇女们所作的工作。如果事实证明他们的方法对其他癌症也有用的话，这可能会是癌症治疗的一次革命。
Dr Ellis and his team sequenced the whole genomes of both cancerous and normal tissue from 46 women with tumours of a type called oestrogen-receptor-positive breast cancer. They also sequenced just the gene-containing regions of the genome-about 1% of total DNA-from an additional 31 women, and parts of the sequences of 240 more. They then compared the healthy and tumorous genomes of each patient, in order to discover which genes had mutated in the cancer.
In this, they were following the normal protocol of the cancer genome consortium. The novelty of their approach was that the women in question had each been involved in one of two clinical trials of a drug called letrozole. These trials established letrozole as a standard treatment for people with this type of breast cancer, but not all patients benefit equally from the drug. Dr Ellis hoped to find out why.
As they report in Nature, he and his team discovered 18 genes that were often mutated. Some were the usual suspects of cancer genetics. These included p53, a gene that, when working properly, suppresses cancer by regulating DNA repair, cell division and cellular suicide, and MAP3K1 and MAP2K4, which both promote cell growth. Others, though, were a surprise. At the top of that list were five which had previously been linked to leukaemia, but were not thought to affect solid tumours.
By combining their newly acquired genetic data with clinical data from the participants, Dr Ellis and his colleagues showed that those whose tumours carried mutations in p53 (16% of the total) were less likely to have responded to letrozole than women whose tumours had normal p53. Conversely, those whose tumours had changes in either MAP3K1 or MAP2K4 (another 16%) had better than average responses to the drug.
This sort of information has obvious implications for treatment. And the cheapness of modern gene-sequencing methods, particularly those that are looking for specific mutations suspected in advance, means that a tumour's mutational complement can be worked out easily in an appropriately equipped pathology laboratory. In the case of oestrogen-receptor-positive breast cancer, the genetic analysis has not yet gone so far as to be able to say with certainty which drug will produce the best result for a given individual, but Dr Ellis's result lays a foundation on which such an edifice might be built for breast cancer and perhaps for other types of tumour, too.