The genetic causes of cancer result in dysregulation of a wide variety of cellular pathways. Owen Witte, MD, and his collaborators were among the first to characterize oncogenic kinase activation in leukemia with the discoveries of the BCR-ABL fusion and BTK genes, leading to the development of several early targeted therapies. “The idea that you could be so selective to hit the single causative agent or genetic change in a cancer really opened a whole new world,” he said. Witte's discoveries have been recognized by this year's AACR Award for Outstanding Achievement in Blood Cancer Research. As a University Professor at the University of California, Los Angeles, and the founding Director Emeritus of the Eli & Edythe Broad Center of Regenerative Medicine & Stem Cell Research, Witte's current work focuses on the identification of targets and development of immunotherapies for solid epithelial cancers. He spoke with Blood Cancer Discovery's David Gennert, PhD, about the history of therapeutic target discovery and the guiding principles of his research.
Much of your early work was on murine leukemia viruses. Was it expected that this mouse viral biology would be important for understanding human cancer?
More than 50 years ago, there wasn't the human genome sequence. People were looking for model systems to study to understand mechanisms of cancer. The best models were viral models like the Rous sarcoma virus and the Abelson murine leukemia virus—there were literally hundreds of viruses that had been described. The whole field of virology was completely transformed by the discovery that this group of retroviruses was highly concentrated for biological effects that gave cancer phenotypes.
At the time, it was really a dramatic emergence of this idea that the viral genes captured from the cell were a way to understand cancer pathogenesis. (Michael) Bishop and (Harold) Varmus won the Nobel Prize for showing that this was true.
Was there a driving question that pushed you to research cancer biology?
I started medical school at Stanford and began working with the now-very-famous scientist Irv Weissman. He gave me one piece of advice, which is to find a really big question and stick with it–you'll get lots of answers along the way. For me, the big question started with a mouse sarcoma virus. We were interested in studying it because it produced a tumor very quickly, but in certain strains of mice, it would regress 10 days later. This was an amazing example of aggressive pathology and immune rejection. I've been working on the same question for 50 years—to understand what sort of genetic influence could cause cancer so rapidly, and what kind of immune response could cause the rejection of the tumor.
How did that lead you to BCR-ABL?
I ended up going to David Baltimore's lab for postdoctoral training, where the Abelson virus was being studied as a model for growth stimulation and control of B cells. My job was to figure out the proteins expressed by the virus, because such a small genome had this profound biological effect. David's approach is to not study a problem at the top level of what happened. Instead, study the problem to the deeper level of how it happened. From the biological phenomenon to a mechanistic insight.
Dr. Witte says he has spent the past 50 years trying “to understand what sort of genetic influence could cause cancer… and what kind of immune response could cause the rejection of the tumor.”
The critical finding was that viral oncogenic transformation is due to a single protein (v-abl, or viral abl). The v-abl gene is a result of recombination between a viral structural protein, gag, with a cellular gene, c-ABL (cellular abl, ABL1). It was unexpected that it would be so straightforward, but now there are dozens of examples of viruses in which hijacking a single cellular gene has such an effect—myc or ras, for example.
Work that I did with Asim Dasgupta showed that the v-abl protein was a kinase and that its end product was phosphotyrosine. Many viruses are now known to harbor genes and gene control pathways that are intracellular or receptor tyrosine kinases. That was really an epic discovery.
How was the connection made between kinase activity in the virus and human pathophysiology?
The best thing that can happen is to do an experiment for one reason but discover something important for a completely different reason. At that point it was still just an interesting enzyme that caused cancer in animals—mice or chickens. When I moved to my own lab at UCLA, one of my early graduate students, Jamie Konopka, was searching for cell lines that produce higher amounts of c-ABL protein, because it is present in vanishingly low amounts per cell. He extracted c-ABL from a human leukemia cell line, K562, and there was a giant protein, one we now call BCR-ABL.
We could show that this big protein was immunoreactive with multiple c-abl antibodies and was kinase-active by our various technologies. And then the whole thing fell into place.
Other groups had studied the genomic rearrangement of the Philadelphia chromosome and BCR-ABL in chronic myelogenous leukemia (CML), other groups found a very large chimeric RNA in the cells, and we had the protein and enzymatic activity. It was as strong a correlative case as you could ever make that a protein was related to a type of cancer very specifically.
But the absolute most critical experiment done in my lab was that we cloned the entire BCR-ABL cDNA. My wife, Jami McLaughlin, and a post-doc, Ann Marie Mes-Masson, transferred it into a retroviral vector and showed that it had biological activity just like viral abl and could transform early B cells. That pretty much fulfilled Koch's postulates, that an agent you isolate from a disease can go back and cause the disease.
It was clear as a bell from these studies that this protein was the best single target for making a drug to treat cancer. Brian Druker and Nick Lydon later came up with a compound that turned into imatinib, which was the first and best targeted therapy for cancer one could ever find.
“The best thing that can happen is to do an experiment for one reason but discover something important for a completely different reason.”
What was known about oncogenes at the time? Was it thought that these viral factors were the answer to curing cancer?
It was thought by those of us who worked on viral oncogenes that they would lead to that kind of information, but it was not commonly accepted in the broader cancer field. Cancer was still viewed as a genetic disease caused by mutations—there were problems with DNA repair or there were problems with a change in metabolism; the work of Warburg was quite dominant in people's minds. The therapy for cancer then was basically chemotherapy—cellular toxins and poisons that inhibited DNA replication. All perfectly logical and critical advances in the field for patient benefit. But the idea that you could be so selective to hit the single causative agent or genetic change in a cancer really opened a whole new world. v-abl didn't get you there, but BCR-ABL certainly did. That really changed the outlook and the whole design of trying to find new therapies.
Another highly effective targeted therapy you helped bring to patients is ibrutinib. What was the process for discovering a second target—that is, BTK?
BCR-ABL is a fusion gene. BTK represents a very different kind of target.
We thought there must be a cellular gene inside of early B lymphocytes that is doing a job similar to BCR-ABL or v-abl, and it's just accentuating that function. That hypothesis was completely wrong. BTK and BCR-ABL have essentially nothing to do with each other in terms of pathway function, yet both are important for the life of a lymphocyte.
We sequenced upstream from the kinase domain in clones that we captured from B lymphocytes. One of these clones became important when we mapped it to two disease gene loci: in man, X-linked agammaglobulinemia, the failure to make a proper number of functional B cells. In mouse, it landed on a locus associated with X-linked immune deficiency. It turns out both of those diseases are caused by a deficiency of either expression or function of the gene, now called Bruton's tyrosine kinase, or BTK.
What was important functionally was this gene is required for the sustenance and expansion of B lymphocytes. As a cancer biologist, I thought this could have something to do with the requirement for malignant B cells to stay alive and receive signals to grow.
We now know that BTK is part of the B-cell receptor signaling complex, any one of the components of which is essential for B-cell survival and growth. This became a target for the development of kinase inhibitors, notably ibrutinib.
When we found BTK, it seemed important as the first cytoplasmic tyrosine kinase with loss of function associated with human disease. But its real impact was on what it can do (as a therapeutic target) to treat cancer and autoimmunity.
Why are the BCR-ABL fusion and BTK genes drivers in primarily hematologic malignancies?
That's not completely understood. The study of hematologic malignancies, as a field, was so far ahead of the study of other cancers because hematologists could easily get human blood for analysis. It was much harder to study solid tumors because getting repeated biopsies or analysis material is harder to do.
Maybe another reason is that the number of genetic requirements to get a deranged blood formation system may not be as many genetic hits as required in some of the solid cancers. BCR-ABL as a single gene defect can give you a discernible malignant phenotype of chronic-phase CML. Clearly, other genetic influences come to bear as the disease progresses—loss of function of TP53, activation and expression of genes like Ikaros (IKZF1)—but you recognize the malignancy even at the early stages with one dominant gene. Many lymphomas also have a single dominant gene change.
What has changed in the study of genetic drivers of cancer during your career?
The advent of sequencing technologies and the sequencing of the human genome allowed us to see what's expressed and how the chromatin is organized down to a single-cell level. The discovery tools have now enabled as much discovery from fresh human cancer material as studying some model systems.
In the 1970s and early 1980s, there was no human genome sequence. There was no browser you could look things up on, so things were done in a very different manner. These oncogenic viruses had basically captured the essential information you needed to know. By studying the viruses, we got these first mechanistic insights into what was causing cancers, like the connection of the Abelson virus to the BCR-ABL gene.
But now, all the genes are there. It's just a question of how they work.
What importance do you place on basic science advances in understanding disease biology compared with the translatability of the research?
There's always the need to remind everyone that this all comes from basic science. It all comes from curiosity, asking a good question, and trying to understand how things work. And then there's a group of people who can turn these ideas that might have been studied in an animal model into a therapeutic. I now appreciate how hard that is.
The take home message here is that it's basic science, basic science, basic science. But you have to be smart enough to hand it off to people who know how to do those other jobs.
“Whenever I'm consulting for a new cancer therapy, I say, ‘What's the genetic evidence? What's the test of function?’”
You've branched out into research on other malignancies, such as prostate and lung cancer. Were you going where the research led you, or was this a planned shift based on your own interests?
I didn't feel like I was done studying blood cancers particularly, but I got interested in studying prostate cancer for personal, family reasons.
At the time, I realized that not a single new thing had been done since the work of Charles Huggins in the 1960s, which defined the effect of castration on the growth of prostate cancer. Instead of dropping everything and starting to work on prostate cancer, I had the good fortune of having Rob Reiter come to work with me, a urological surgeon still here at UCLA. And I had Charles Sawyers in my lab at the time working on BCR-ABL doing fantastic work on resistance mechanisms.
Charles was interested in working on epithelial cancers as well, so we formed a kind of triumvirate. Rob was a urologist who knew a lot about prostate cancer; Charles, an oncologist; and me, a general cancer biologist. We pooled our resources and connected with the urology department and other people like Arie Belldegrun and Jean DeKernion.
What parallels can you draw between your work in the hematopoietic system/hematologic cancers and your work on solid tumors?
In my lab over the last 20 years, we've built a model in which we could study prostate cancer with a combined in vitro/in vivo transformation system. Jung Wook Park published a model system to make super-aggressive forms of prostate cancer. It followed the same game-plan that we used for BCR-ABL and other transformation events. It was the application of the same principles. So maybe I don't need any new tricks.
In the realm of immunotherapy, where much of your current research is focused, what translational and clinical advances are most promising for patients?
When a new technology becomes available, everybody tries that on everything. For example, there was an explosion of trying antibodies in different settings when it became clear that they could have therapeutic potential. The same is happening now with CAR T cells. There will be attempts to make CAR T cells functional for one or another type of cancer, and some will work and some won't work. Same thing with checkpoint inhibitors. Just from the PD-1 inhibitor class alone, there have been probably 500 clinical trials looking across diseases.
We need a better understanding of these reagents and what cancers are doing in each individual case. Lung, breast, and kidney cancers are treated by different groups of people. Well, disease mechanisms are not really relevant to the organ. They're relevant to the cancer, and if the cancers share an antigen, you might develop a target and a therapeutic attack on that target. We break down the barriers of traditional oncology because the therapy is breaking down the barrier, and I think that's what's going to be happening next. You see it already with the checkpoint inhibitors.
Some of your current work involves mapping and targeting the epigenome, which is often highly dysregulated in cancer. How do you think about epigenetic mechanisms in cancer as emerging targets and their role in therapeutic resistance?
You need to define elements of a control system which have demonstrated that they are genetically penetrant when their function is either enhanced or blocked. Whenever I'm consulting for a new cancer therapy, I pound the table and I say, “What's the genetic evidence? What's the test of function?” I ask what genetic tests of function have been done or presented in the literature that say a particular protein, when changed in intensity of function, changes biology. Until you have that, you don't really have a good target.
The whole Polycomb complex is a very good example of targetable entities within a big biological function. Chromatin writers, erasers, etc., are all really important to think about.
Science changes its opinion based on data, and the data shows that chromatin modification, chromatin structure and function are critical for cell growth, maintenance, and disease states. The question now is: how, where, and why?