So yes, the fact that we've been able to sequence the genomes of tumors has made us realize they're much more complex than we ever knew. But we are also slowly starting to develop more genetically targeted therapies – the trouble is, so far these are only targeted at cancers that don't affect a huge number of people.
According to Weinstock, the Dana Farber Cancer Institute has a program called "Profile," which involves looking for a list of commonly mutated genes, to see if those genes are mutated in someone's specific cancer. Eventually, he'd like to see more of the common mutations sequenced and identified, so "we'll be able to say exactly what genes are defective in someone's tumor." Image via Andres Perez.
And once you know what mutations someone has, you can probably figure out which signaling pathways are affected by those mutations, and target those, says Weinstock. For example, if you know someone has a problem with a particular gene, you may know that you can use a specific type of kinase-inhibitor on that patient, to block the enzymes that are involved.
For example, says White, we now have drugs that specifically address mutations in the EGFR receptors. If you happen to have that specific type of cancer and you take the drugs that block those receptors, "the tumors just melt away." Also, there's herceptin for some breast cancers, TKI inhibitors for renal cancer, or MDV3100 for prostate cancer.
"These [drugs] are driven against molecular events," says White. There's also been a shift in the mindset of drug companies, which used to prefer cancer drugs that could be aimed at a wide swathe of patients, to sell lots of drugs. Now, companies are learning that drugs that are very specifically targeted at one relatively tiny population of cancer patients can still make plenty of money.
"I would anticipate the number of drugs we have 50 years from now will be astronomically larger than today, and those drugs will be used in a much more targeted fashion," says Sartor. Instead of a single treatment for breast cancer, we'll differentiate it into "100 different types," each with its own therapy.
"It wasn't very long ago [that] Steve Jobs spent $60,000 having his cancer sequenced," says White. But by next year, there will be a machine available that will do the same thing for $1000 a pop.
No comments:
Post a Comment