Almost half a century since President Nixon started the ‘war on cancer’, President Obama promised ‘a new moonshot’ so that America may become ‘the country that cures cancer once and for all‘. Yet serious problems have a knack for persisting despite high-minded -- if overly optimistic -- political speeches. Despite decades of research and billions of dollars in public and private funding, cancer remains one of the world’s likeliest killers. Why?
Broadly, cancer is caused by uncontrolled cell division, when the mutation of a small set of genes causes cells to multiply indefinitely and invade the space of healthy neighbouring tissue. But cancer is best understood as a chemical process. Each cell contains the full human DNA - around 24000 genes - and each cell division makes a full copy. A gene is a sequence of chemical bases that instructs cells to turn glucose and oxygen into energy and which amino acids and proteins to manufacture, encoding information such as when to multiply and when to self-terminate - the programmed cell death known as apoptosis. Carcinogens are anything that corrupt this chemical code: radiation that breaks ionic bonds, and substances and viruses that disrupt cell metabolism or bind to DNA directly and change its structure. Still, perhaps the most insidious carcinogen is sheer bad luck: chance mutations over a lifetime of cell divisions. While DNA can self-repair, it won’t if the mending instructions themselves are damaged. Therefore apoptosis is key in preventing cells with gene mutations from replicating. However, if the signal for self-termination too is altered, cells will both multiply indefinitely and continue to replicate the initial mutations, making further DNA corruption more likely. This is why in tumours mutations tend to accumulate. Fixing all this is hard for many reasons.
First, every cancer is a different disease. Not only is leukemia different from melanoma but every tumour is caused by a different set of genetic mutations such that no two cancers are ever the same. Every tumour follows a unique genetic path so that one person may live and another die having the ‘same’ cancer and taking the same medication. Second, surgery is not always feasible, radiation is difficult to target accurately and the drugs known collectively as chemotherapy cannot distinguish between rapidly dividing cells in tumours and cells whose rapid division is legitimate and vital, such as those in hair and stomach lining. Finally, the accumulation of mutations makes it hard to identify the genes that started the process. Without knowing those it is hard to develop personalised medication or predict the effectiveness of various drugs on individual people. This is a problem big data can solve. Alas, because genome sequencing has only been possible for little over a decade, some of the world’s biggest genome databases still have only thousands of samples. Many millions may be required.
Yet change is under way. In 2008, whole genome sequencing could be purchased for $350,000; today that cost is under $1000, with results in two and a half months (the Human Genome Project took 13 years). Further, people who use DNA testing companies can chose to share their genomes with researchers, helping solve the data problem. Together with better computing, this can make the development of personalised treatments feasible. In 2000, Bill Clinton declared it ‘conceivable that our children’s children will know the term cancer only as a constellation of stars’. Overly optimistic, perhaps. Yet replace ‘children’ with ‘grandchildren’ and it might just come to pass.
