About 4 years ago, I went to a seminar at TSRI that convinced me that cancer would be over in relatively short order. The man speaking (I wish I could remember who it was) showed that his group had been able to target radioactive heavy metals directly to melanoma solid tumors to destroy them. The data were striking; enormous, football-sized tumors shrank to nothing in a matter of weeks, and the therapy worked more than 90% of the time. I couldn’t understand why this wasn’t a bigger deal, why wasn’t this front-page news?
As it turns out, this wasn’t a novel concept – targeting chemotheraputics directly to tumors has been happening for a long time. Since that seminar, I’ve seen news articles talking about numerous studies showing targeted delivery of various drugs and even bee venom to tumor cells, leading to their elimination. Recently, a group at CalTech published a paper in Nature showing nano-particle delivered siRNA can destroy melanoma tumors in a very successful human clinical trial. Is this the future of the fight against cancer?
The trouble with eliminating cancer is that it’s your own cells causing the problem. Anything that can kill a tumor cell can also potentially kill a normal cell. The trick for scientists and clinicians is to focus treatments to the few ways that tumors are different. The most obvious difference is that tumor cells are dividing rapidly (that’s what causes tumors to be dangerous in the first place), so the first (and still most successful) chemotherapies are lethal to dividing cells (this is the same reason that radiation works). The trouble of course, is that tumor cells are not the only cells dividing in the body. Side effects of these traditional treatments include hair loss (hair-follicles need to rapidly divide), nausea (the cells lining the stomach and intestine need to be replenished all the time) and immunodeficiency (most cells of the immune system need to divide to fight infection), and these side-effects are not mild. The drugs are essentially poison, and the goal is to poison the body just enough to kill the cancer without killing the patient.
But division isn’t the only way that tumors are different. Tumor cells are ravenous, and must be constantly supplied with more nutrients, so drugs targeting angiogenesis (the formation of new blood vessels) are useful in slowing or stopping tumor progression. Cancer cells are riddled with mutations and genomic instability, so drugs that inhibit parts of the DNA repair pathways are being researched (this is a bit more complicated, so I’ll save further explanation for another post). But you can probably imagine why interfering with these processes can also be dangerous to the rest of the body, not just the cancer.
One of the best hopes for truly specific cancer treatment is the existence of tumor antigens. These are proteins that are expressed or presented on the surface of tumor cells, but are absent or present at much lower levels on normal cells. There are many reasons this might happen (that wikipedia article explains it pretty well, and feel free to ask if you’re curious), but the key here is specificity. We can use antibodies (and other proteins that bind with high specificity) to send drugs and other therapies directly to the tumors. This is how the researchers I heard at TSRI targeted radiation, it’s how bee-venom was targeted in nano-particles, and it’s how the group at CalTech targeted siRNA. So why melanoma?
Melanoma is one of the easiest cancers to study. Taking a biopsy of the tumor doesn’t require invasive surgery (it’s on the skin), and the originating cells can often be studied well before they’ve actually crossed the line into cancer. So we know a lot about what tumor antigens melanoma expresses. Some other types of cancer also have well-known tumor antigens, but many don’t. In addition, even the well-known tumor antigens of melanoma aren’t expressed on every tumor of this type. Every type of cancer is different, and every individual’s cancer is different. Translating these targeted therapies to be more widely used in the clinic is still a long way off.