The revolution in immunotherapy is incredibly exciting. Cancers are being treated, and occasionally, apparently, cured, by tweaking a patient's own immune system. It is a thrilling new development, and a hope to millions of people.
But we need to show some sobriety.
To understand why we haven't found "the cure for cancer", we need to understand why immunotherapy has worked so well in the places it has been tried.
The first is CART therapy. The CART acronym stands for "Chimeric Antigen Receptor T-cell". This method is completely brilliant, in that it uses the killing power of cells that your body already has in abundance, T-cells, and adds just a few key pieces. A T-cell, as it turns out, fights diseases by identifying cells that look infected, and mercilessly killing them. It can do that because the virus is often sitting around on the surface of the cell, or the cell "presents" little bits of the virus that the T-cell can identify are foreign.
A cancer cell, though, often doesn't have bits of virus to present. It just has the usual proteins of every other cell in the body, with a few very small exceptions. T-cells are carefully trained, in a process called "negative selection" in the thymus, to avoid attacking the body's usual proteins. Imagine the consequence if this didn't happen: you'd have your T-cells trying to kill you every minute of every day. Cancer starts off as being human, and then accumulates mutations that make it corrupt. But that starting point, a normal human cell, is invisible to your immune system (as it should be).
However even normal human proteins can occasionally be useful targets. Some cancers, particularly certain cancers of the blood, have very specific proteins that differ from the rest of the body. These are called CD19 and CD20. These proteins basically appear in only two places: the cancer and the normal white blood cells that gave rise to the cancer. Since the immune system is usually completely out of whack anyway, killing off all those few normal white blood cells along with the cancer is acceptable collateral damage. If only they had a way to target those proteins...
Then comes the "one weird trick". Thanks to a whole lot of people, people like Carl June at UPenn, a plot was hatched to give the T-cell a specially tailored gene to make it attack those white blood cell proteins instead of viral proteins. The gene could look a lot like the genes a T-cell already has, but with some special changes slapped together to make that T-cell absolutely ravenous for the sweet taste of CD19 or CD20. When they see their prey they turn into killing machines, and they multiply to kill even more.
It turns out this totally works! Just this year the initial approvals for treatment of childhood Acute Lymphoblastic Leukemia (one of those blood cancers I mentioned), got approval by the FDA to be delivered to patients as a potentially life-saving cure. It's a tremendous accomplishment, and is just the beginning of a flood gate of new therapies. It's a very exciting time to be working on these methods, as we explore how to bring other cancers into the target sights of these CART therapies.
Unfortunately, there is a catch. For one, because the treatment is so expertly honed on a specific protein, the cancer has quite a few ways to dodge the attack. For one, it can just stop making the protein. It turns out that's not so hard for a cancer cell to do, since it's often busy accumulating a rather tremendous number mutations over time, and some cancer cell will get lucky and hit, for example, CD19. Having erased the target on its back, the cell can go on multiplying unencumbered by the astonishingly intricate therapy we have devised.
But additionally, it turns out that ALL is sort of rare among tumors for having such a uniquely defining protein just humming around on the cell surface waiting to be targeted. While cancer cells are certainly very different from non-cancer cells, in many cases a patient's individual tumor may be entirely different from another patient's individual tumor even if they have the same kind of cancer. It's an unfortunate state of affairs that means that finding a target that specifically identifies a patient's cancer, and not all the bystander normal cells that make us who we are, one might have to resort to a specific cure for a specific patient.
A truly "personalized medicine" is a tall order. We might have to develop a vast arsenal of weapons, only to pull out a single, specialized stake to drive into the heart of a given patient's disease.
The immune system, however, is dazzlingly complex, but I wonder if it intrinsically contains the key to the puzzle. The immune system does something very specific in response to an infection that I find staggering: it undergoes evolution. Your body, in service of your survival as an organism, actively uses evolution to identify a receptor that matches a given enemy in the field. To oversimplify the system enormously I'll explain as follows.
When a virus attacks, the immune system throws a bunch of cards in the air and lets them come down scattershot. Those that land closest to a match for the invader start to tickle the cell into reproducing, activating a program that increases the cell's tendency to mutate as well as create new copies. Effectively, this means that the cards that landed nearest to the virus in the space of possible matches send out their own little scattershots again, this time in a tighter area, trying to move closer to a perfect match for their enemy. This process continues until a good match is formed, and the cells are effective enough at their job to eliminate the enemy.
As I mentioned at the top, though, a cancer cell looks a whole lot like a human cell. And there are specific systems designed to make sure that, even as it's evolving towards a better weapon to target, for example, a virus, anything that matches too well with a normal human cell is treated as off-limits. That is great, usually, because it keeps us from developing autoimmune diseases. But if I have a tumor, I desperately want a very certain kind of autoimmune disease, an anti-cancer kind.
We actually have the tools to remove some of those "off-limits" signs. A whole other type of immunotherapy, called "checkpoint inhibitors" allow the body to wander into those danger areas and explore the space of autoimmune disease. That's fantastic if it ends up giving you a cancer targeting T-cell, but often there are complications where a set of T-cells go after the wrong target, like the normal skin, or the gut, or the liver. That can be exceedingly bad news, and it puts a limit on just how far we can go with those therapies.
The benefit, though, is that these checkpoint inhibitors allow the immune system to devise its own attack plan. Within the patient, a process of evolution occurs to counter the tumor's own evolving approach. That is a fantastically powerful process. No pill cooked up in a lab somewhere can shift and feint with the tumor, and that fact has too frequently doomed some of our recent targeted therapies.
I wonder, though, if we can somehow get the best of both worlds. The incredible power of building a T-cell from scratch, in the lab, combined with the power of evolution inherent in the immune system. I can imagine a world where, in the contained safety of a dish, a host of a patient's own T-cells are modified and reacted against the patient's tumor to develop that "personalized" book of armaments. Those could then be vetted not in the living chaos of the patient's own body, but again in a dish, to weed out dangerous autoimmune functions. Thus processed, if these cells could then be expanded and put back into the blood to attack a tumor, they would be a hailstorm of guided missiles.
This might be a fantasy, but there's some beauty in it. Attack somatic evolution with somatic evolution. If only Darwin could see us now.
