You can’t throw a DNA sequencer these days without hitting a geneticist who’s found some disease-causing mutation. But mutations-- changes in the protein-encoding sequence of DNA bases--aren’t the only things that can go wrong with a gene. Sometimes a gene with a perfectly normal base sequence doesn’t produce the protein it ought to--and sometimes cancer is the result. Now researchers think they know why such seemingly good genes go bad. The reason is an error in a system called methylation, which normally helps determine whether genes are switched on or off.
Although all our cells have the same genes, they do not all make the same proteins; each type of tissue makes its own distinctive set. Methylation is one of the cell’s ways of turning off DNA it doesn’t need. When a cell first develops in the embryo, and again each time it divides, an enzyme called DNA methyltransferase cruises along the newly formed DNA and adds a chemical cap--a methyl group--at specific sites: sites where the base cytosine (C) is followed by the base guanine (G). For some reason, regions of DNA that are heavily methylated coil up tightly, like an unstretched Slinky. That seems to be how the cell packs away, out of the reach of gene-activating proteins, the large stretches of DNA that aren’t genes at all but ancient, noncoding junk.
Actual genes are blocked in a more direct way: methyl groups bind to the gene’s promoter region, at the beginning of the gene, thus preventing proteins from binding to that region and turning the gene on. Consider as an example the gene for hemoglobin. Every cell has it, but only red blood cells need it. Thus, in all cells except red blood cells, the promoter region of the gene is methylated and the gene is turned off.
No one knows how the cell knows to leave some CG groups unmethylated and some genes on. Indeed, about half our genes have promoters that are littered with CG groups and yet are shielded from methylation: these are the housekeeping genes that every cell needs. The tumor- suppressor genes are one group of housekeeping genes. They make proteins that prevent cells from dividing too rapidly. A few years ago it occurred to tumor biologist Stephen Baylin at the Johns Hopkins Medical Institutions that if a tumor-suppressor gene were accidentally methylated, a cell might become cancerous. Methylation effectively takes that gene out, says Baylin. It’s the same effect as a mutation.
Baylin and his colleagues first tested their hypothesis on the Von Hippel-Lindau gene, a tumor-suppressor gene that has been linked to kidney cancer. In 20 percent of the kidney tumor samples they looked at, the Hopkins researchers found abnormal methylation. Instead of finding two mutated VHL genes, as the traditional cancer model predicts--every cell has two copies of the gene, and only if both are damaged does the cell lack the tumor suppressor--they found only one. The other VHL gene appeared perfectly normal--except that its promoter was methylated.
Next Baylin and his colleagues looked at p16, a tumor-suppressor gene linked to many cancers. Examining tumor samples from the lung, head, and neck, they found abnormal methylation of the p16 promoter in roughly 20 percent. Most recently they found abnormal methylation in 40 percent of colon tumor samples and in 30 percent of breast cancer samples.
In some cancer cells the patterns of DNA methylation are being grossly altered, Baylin concludes. Instead of having mutations in the coding region of the gene that would alter the protein, what they actually have is this methylation of the promoter. So they’re not even making the protein. More evidence of a link between abnormal methylation and cancer comes from researchers at MIT: they found that mice with low levels of the enzyme that methylates DNA also had fewer precancerous colon polyps.
Since no one knows what controls the normal methylation pattern, no one knows what causes abnormal methylation either. But Baylin sees a silver lining in his discovery of a new form of cancer-causing genetic havoc. This is a potentially reversible process, he says. The reason we know that methylation is a really important part of gene silencing is that you can turn the genes back on--at least in cell culture--by treating them with drugs. Several people and companies are beginning to explore this approach.