Smart and Dumb Tumors

Although there were many interesting presentations at ASCO on which to blog(and I will try to cover some of them in a later post), I thought that Dr. George Sledge’s Presidential Address [click here to listen to it] was one of the very best talks I have heard in a long while. In the talk, which is just now available to the public, George discussed some of the items I brought up in my blog on controlling cancer care costs, but did so from the perspective of the cancer cell itself.

Let’s suppose that the DNA in your cells has been around for a very very long time. Leaving aside the discussions of creation vs. intelligent design vs. evolution, the cells which make up “you” have been through an exceptional amount of engineering which we are beginning to understand. As such, they have “learned” to specialize – for example the epithelial cells in your prostate gland, while containing all the code to do the same thing as the epithelial cells in your skin or stomach, instead do only the subroutine of making PSA and other proteins required for their function to assist in reproduction. But they also have the ability to fend off chemical insults, repair damage from radiation, and to replace themselves via stem cells that rarely divide and simply sit in the prostate (and other organs) as a warehouse from which to repopulate cells that die as a part of the natural process of organ renewal.

Now suppose there is a mutation that causes this complex machinery to become unregulated. A prime example is the TMPRSS-ERG/ETS mutation which seems to be an early example of things that go wrong in prostate cancer. In this case, a piece of a gene that is sensitive to testosterone stimulation becomes aberrantly hooked to another gene that codes a signal protein called a transcription factor. The result is an abnormal protein called a fusion protein that will stimulate cancer development in the presence of testosterone. Pretty simple to work on – just get rid of testosterone…. If this was the end of the story, we would have a “dumb” tumor. We could attack it with one approach, removing testosterone, and cure most patients. This is part of the story of why most patients will achieve some sort of remission when receiving a drug that reduces testosterone like leuprolide or goserelin.

However, as with most complex problems, that is the “simple answer” that doesn’t work as well as we would like. In prostate cancer, by the time we detect it, there are many many more mutations. Many of these are being discovered and we can develop drugs to attack them. However, by Dr. Sledge’s definition, these are therefore “smart” tumors – ones that will require, perhaps a different drug for each mutation. The “hot drugs” I have covered elsewhere in this blog, like abiraterone or MDV-3100 are really just better drugs to attack the same “dumb” switch (testosterone), and thus they do not lead to cure, although the improvement in survival is welcome.

An example of a next target is the PTEN mutation, or loss. This, too, is quite common in prostate cancer and means that the cell has lost a protein that functions as a tumor suppressor. It is really really hard to replace a protein in a cell that is lost due to mutation. However, ongoing research suggests that cells with PTEN loss may be sensitive to PARP inhibitors. Thus, if we want to look at how to take the next “big step” in treating prostate cancer, we would first evaluate a large group of patients and find those with TMPRSS-ERG fusions, then those with PTEN loss, and use all of the new hormone switch inhibitors, THEN add a PARP inhibitor. This is what is meant by “individualized medicine” that is the current buzz. Only those patients with each of the mutations that lead ultimately to sensitivity to an approach would be tested, and the FDA would be asked to approve a drug for that subset of a subset of patients.

Returning to the Sledge lecture, George did some fancy math (which I didn’t understand) to calculate the number of patients you might have to screen just to get the study ready for testing. Then, of course, the pharmaceutical company has to calculate how much it costs to develop a drug for many fewer patients than just “all prostate cancer patients”, and factor that into the cost, since the FDA approval would be for many fewer individuals. As you can surmise, the cost for “individualized medicine” suddenly becomes daunting – and in the example I have proposed, we have dealt with only two types of mutations of the many that we already know about. (for example cells can get signals from all sorts of growth factors beyond testosterone)

So the picture is encouraging from a basic science point of view, but pretty overwhelming from the standpoint of “when will we have a cure?” I think the most realistic hope is that some combination of these approaches, for some patients, will convert prostate cancer into a chronic illness like diabetes. The question is whether the new agents will ever be as cheap as insulin. It’s too bad that prostate cancer falls into the category of being a “smart tumor”.