Gilliland: These three diseases are the most common of all these myeloproliferative syndromes. Tyrosine kinases had been identified as playing critical roles in related blood cancers, but nobody knew what causes these three. They had remained mysteries.

In the past we had used what are called translocations to find the relevant mutations. These are chromosomal abnormalities. You can see them from 50,000 feet looking down at the cells. You can literally see them in a microscope. They tell you exactly where the gene is that causes the disease. They point to the exact location among the 20,000 genes and three billion nucleotides in each of the cells. They’re like a large red flag waving in the microscope.

But there is no known chromosomal translocation in these three diseases. We know there’s a mutation involved, but we had no clue as to where they might be or even which of the 20,000 genes in our genome were involved. Nor did we know that there would be one mutation key to all three. That was a surprise.

Levine: Not in this way.

Gilliland: There had been epidemiologic types of surveys, where you send questionnaires out and people fill them out and send them back. As far as I know, this was the first example of soliciting a specific patient population for blood samples and for DNA from their germ line. We were asking them to swab their mouths. This is the first time this was done. I’m sure it’s not going to be the last.

Levine: We did confirm the diagnoses by reviewing charts from almost all the patients.

Levine: First we extract DNA from the patients’ blood cells and then we use these platforms first to increase the amount of DNA, using PCR amplification, and then to sequence all the kinases to look for candidate mutations. Once the candidates are identified, we then went back to the patient’s germline DNA—their non-blood DNA—and asked whether the mutations we found were inherited or acquired. Based on previous work by many people, everything we knew suggested these mutations would be acquired.

Levine: There is evidence that they run in families, but in most patients, that is not the case.

Levine: We found a single mutation in one kinase: a single spot in the JAK2 kinase. The result is a single substitution of one amino acid for another. And, to our surprise, we found it in the vast majority of the patients with these diseases. We then demonstrated that it’s an acquired event in these patients—the same mutation over and over. We were fairly surprised by that observation. It wasn’t a spectrum of different mutations, but a single mutation that was recurrent in three different diseases. It is also important to note that several independent groups from the U.K., from Switzerland, and from France also identified the JAK2 mutation at about the same time, using different approaches.

Gilliland: Ross’s analysis involved 120,000 different DNA sequencing runs and he had to sift through tens of millions of base pairs of DNA to find this one mutation. The platform he used to do this simply didn’t exist a few years ago. This was an enabling technology.

Levine: That’s a very interesting question. Again it’s surprising that most of these patients have a single mutation, but a significant minority do not. So when we made the first observation, that was the unanswered question. Subsequent to our initial report, a number of studies have identified alternative mutations in some of these patients who don’t have the classical JAK2 mutation. A group out of Cambridge University, led by Anthony Green and with whom we were honored to collaborate, identified other mutations in JAK2 in a small number of patients with polycythemia vera.

We now hypothesize that mutations in the genes that interact with JAK2, not kinases, but genes that interact with this tyrosine kinase, might be the cause of the disease in the absence of the JAK2 mutation. In our laboratory, we found a mutation in a thrombopoietin receptor called MPL. That’s like JAK2 in that it’s an acquired mutation that activates the MPL gene. The important thing is that all the mutations identified to date activate this JAK2 kinase activity, suggesting that, regardless of the specific mutation, ultimately the disease is caused by constitutive activation of JAK2.

Levine: I would suspect that it’s mostly the discovery, although I hope there are people interested in the paper for the internet component. The diseases we’re talking about have gone a century since they were first identified without anyone understanding the genetic cause. So our research is of significant interest. It’s not the ultimate answer of what causes these diseases, but it allows us to begin to understand these diseases on a molecular basis.

Gilliland: I agree with that. There is some interest in the internet protocol, but you still have to be in a circumstance where you can enable that approach. Not every group can do that. It requires that you have the necessary clinical interfaces, and human consent committees, etc.

Another reason, by the way, that this paper is cited so frequently is that it identifies an outstanding target for intervention. So the citations are not just coming from the academic side—understanding what’s behind these diseases. They’re coming from the pharmaceutical community, which is extremely interested in trying to develop the next Gleevec-like drug to treat these diseases.

Levine: A lot. First, as we mentioned, some alternate mutations have been described to begin to account for patients without the classical JAK2 mutation. Then a number of groups, including Gary’s, have been working hard to understand what the function of this gene is, in terms of how it’s activated—what happens when it’s expressed in mice? And the answer is you get diseases that look a lot like human polycythemia vera, which is very interesting and important. However, much work remains to be done, particularly, as Gary said, in designing drugs to inhibit these genes. There is also quite a substantial amount of work by many groups trying to understand the clinical relevance of this mutation.

Gilliland: For me, watching Ross do this research, I would have to say that it just looked like ass-busting work to me: sequencing every single tyrosine kinase in the entire human genome, and running through the amount of sequences he did. It would have been nice if there had been a little bit more luck involved, but in the end it was a lot of hard work.

Gilliland: We know these diseases are basically cancers associated with overproduction of blood cells—white cells, red cells, and platelets. Normally, production responds to our need. If we’re anemic, for example, we make a hormone called erythropoietin, which stimulates production of red blood cells. If we have bacteria in our blood stream, we make a hormone called GCSF that produces high levels of white blood cells. When the need for these cells passes, everything returns to normal. Now the hormones are no longer generated in response to the stimulus. That’s how this normally works. The beauty of this is that it really does make sense physiologically.

The way the signal is transmitted from the hormone to the cell that will grow up to become a red cell is through this JAK tyrosine kinase. When erythropoietin binds to the red cell precursor, it activates JAK2, and that in turn sends all the appropriate growth signals into a red cell. This mutation seems to bypass the need for the hormone. Now JAK2 is always on, and it’s continuously sending those signals down into the nucleus to tell these cells to grow. That’s basically the disease phenotype we see in humans: overproduction of one or more blood cell lineages, be it white cells, red cells, or platelets.

Part of the proof that this is right is that if you put this mutated JAK2 gene into a mouse, it develops incredibly high white and red blood cell counts. That says this mutation in and of itself is sufficient to cause disease. The mutation seems to simply hijack the cells’ normal growth control. It turns the switch on and there’s no way to turn it back off. And that’s where the therapeutic potential lies. If we have a key that will fit into the lock of this tyrosine kinase and switch it back off again, these cells ought to go back to behaving normally.

Levine: The important thing is that this mutation allows us to begin to understand these diseases. There’s still a lot we don’t understand. Not everyone has the mutation, and this same mutation causes similar but distinguishable diseases in different people. There have to be other genetic events involved in these diseases, and I think the explosion of genomic platforms available, which are just expanding by the month, will empower resources worldwide to understand what other mutations are involved in these and other blood cancers and other cancers, in general. That’s really empowering and that’s the goal now.

We also have to bring this back to the patient. So the other goal is to come out with these drugs, as Gary said, that target this mutation so we can actually provide treatments for these diseases.

Another thing I would add is that this interest in tyrosine kinases has led to them being evaluated in almost every human cancer now. It’s being done by groups all over the world. And in some cases this research has been incredibly fruitful.

Gilliland: If I had unlimited resources, I would pursue the notion that Ross mentioned, that this paradigm applies not just to these rare blood diseases, but to every human cancer; that they all, to some extent, follow this path. As it turns out, these kinases also play important roles in some types of gastrointestinal tumors, in lung cancers, in breast cancers, etc. They are very good targets for intervention. The drugs that exist now are effective and have very mild side effects compared with conventional therapies. So I think the potential is now there to really put a dent in the armor of all cancers.

But we need the resources to expand this program into all tumors. These platforms are not cheap. It’s very expensive to do DNA sequencing analysis, and that’s just one of about five or six technology platforms emerging that we can and should use. These are very expensive and they are hypothesis-generating experiments, not hypothesis-testing, which means they’re very valuable but it’s not easy to get federal support for this kind of research. And we’re very dependent on that. Because of conflict-of-interest issues, we don’t accept resources from the pharmaceutical industry at all.

Gilliland: It’s actually mandated by the Howard Hughes Medical Institute (HHMI), of which I am technically an employee. I think it’s appropriate. We have to be very careful about conflict of interest when it comes to testing for mutations in human DNA and looking for treatments. We shouldn’t have any financial interest in looking for these outcomes. That’s why in an ideal world, which this isn’t, the entire budget for the National Cancer Institute per year—$5 billion, which is not chump change—would not be less than one month’s worth of the cost of the war in Iraq. We have to understand that as a society, if our priorities are to support these initiatives in cancer biology and treatment, we must adequately support them. We now have the tools in hand to begin to tackle cancers. We can see the light at the end of the tunnel, but we’re severely resource-limited, just when we need it the most.

Levine: I don’t know what Gary would say about this, but the important thing is that, although we now understand a lot about these cancers, we do not understand enough. And one thing we don’t understand is why they occur. In a lot of cases, it might be the errors resulting from the enormous number of times that a cell has to replicate its DNA for its daughter cells. Even with an incredibly accurate DNA replication system, the error rate is large enough that this is going to happen. What we don’t understand is why some mutations happen over and over—this JAK2 mutation, for example. In some cases, there might be environmental or genetic or inherited cues.

Gilliland: I agree. I think that’s probably where Ross’s research would be going if he had unlimited resources. We don’t understand where the mutations come from or the genetic contributions, and we don’t understand well how the environment interacts with the host, if you will. That’s very important, like it is for any disease. If you can prevent the disease, it’s always better than trying to treat it after the fact.

D. Gary Gilliland, Ph.D., M.D.
Brigham and Women’s Hospital
Harvard Medical School
Boston, MA, USA

Ross L. Levine, M.D.
Brigham and Women’s Hospital
Harvard Medical School
Boston, MA, USA