Cancer Biology, Medicine and Translational Research
Dana-Farber Cancer Institute
Dr. Kaelin is also a professor of medicine at Dana-Farber Cancer Institute and at Brigham and Women's Hospital, Harvard Medical School; senior physician at Brigham and Women's Hospital; and associate director, Basic Science, of the Dana-Farber/Harvard Cancer Center.
Although William Kaelin once decided he was unsuited to laboratory work, he has already won several awards that recognize young leaders in cancer research. His studies of rare forms of cancer have uncovered molecular pathways that are important in kidney cancer and have been targeted in several clinical trials.
Kaelin majored in math and chemistry at Duke University and then stayed at Duke for medical training. But his first stint in a lab was "horrible." "So I thought I was making the right decision to become a practicing physician," he says.
After an internship and residencies at Johns Hopkins Hospital, Kaelin moved to the Dana-Farber Cancer Institute to train in clinical oncology. For the first year, he treated patients; then he embarked on 2 years in a research laboratory to fulfill his board certification requirements and to give himself one last taste of laboratory life. But the lab closed down 4 months after he moved in. "I was getting sign after sign that laboratory research was not for me," he says.
By a stroke of luck, he landed in the lab of David Livingston, who was working on the molecular basis of retinoblastoma, the most common pediatric eye cancer. "Having just spent a year on the front line, taking care of patients with cancer, it was pretty clear that the only real hope for these people was to eventually have a more precise molecular understanding of cancer and to try to translate that knowledge into more effective therapeutics," he says.
In Livingston's lab, Kaelin isolated a DNA-binding protein, E2F, that promotes cell proliferation. E2F is normally kept in check by retinoblastoma tumor suppressor protein, pRB, which tightly regulates cell division. But when mutations prevent pRB from performing this function, cells divide unrestrainedly, triggering retinoblastoma.
Throughout his career, Kaelin has maintained this interest in the regulation of cell division. "But the field is relatively mature now, so we try to contribute where we can rather than getting sucked into just doing the next obvious experiment," he says.
In 1992, Kaelin set up his own lab down the hall from Livingston. Looking for a project, he read about the identification of the gene for von Hippel-Lindau (VHL) disease. Patients with this hereditary cancer may develop tumors in the kidneys, adrenal glands, or pancreas. And they often acquire tumors in the central nervous system that look like nests of blood vessels. Whenever Kaelin had encountered this rare condition, he had wondered if it involves an abnormal response to normal oxygen levels. As well as growing lots of blood vessels, the tumors sometimes produce erythropoietin, which stimulates red cell production. Thus, they behave like tissue that is short of oxygen.
Kaelin discovered that when oxygen levels are normal, the VHL protein helps mark another cellular protein, called HIF, for destruction. When oxygen levels fall, HIF is allowed to persist, so it can restore a healthy supply of oxygen by promoting blood vessel growth and stimulating erythropoietin production. But how does VHL know that oxygen is scarce and it should leave HIF alone?
For many years, scientists had been trying to understand how cells sense and adapt to changing oxygen levels. Kaelin's group discovered that when oxygen is present, HIF acquires a hydroxyl group (-OH). When there isn't enough oxygen to provide the oxygen atom in this group, HIF is not hydroxylated. Moreover, oxygen availability determines the efficiency of the hydroxylating enzyme. So HIF remains unadorned when oxygen is scarce, escapes being tagged by VHL (because the hydroxyl group serves as the binding signal), and survives to initiate blood vessel growth.
This discovery was surprising because hydroxylation had never been identified as a cellular signaling mechanism. But since Kaelin's finding was published in the journal Science in 2001, National Cancer Institute investigators have identified several genes in the human genome that are predicted to encode protein hydroxylases. "A major thrust of our laboratory now is to understand their functions and what their targets are," Kaelin says.
In 2002, one of Kaelin's postdocs noted that another hereditary syndrome, tuberous sclerosis complex, which produces widespread but benign tumors, shared some features with VHL disease. For example, he had observed that rodents with a mutated tuberous sclerosis gene are more like humans with VHL disease than are rodents with VHL mutations. The postdoc determined that mutations in the tuberous sclerosis gene, such as those that occur in the human disease, cause too much HIF to be made for various reasons. "That was a satisfying molecular explanation for the phenotypic overlap," Kaelin says.
Kaelin's work is already being translated into clinical medicine. His group determined that kidney cancer often involves nonhereditary VHL mutations that have the same effect as hereditary VHL mutations: overproduction of VEGF (vascular endothelial growth factor), which both promotes blood vessel growth and stimulates erythropoietin production. Recently, the FDA approved the use of two VEGF inhibitors for treating kidney cancer.
The work on tuberous sclerosis led the group to demonstrate that, when kidney tumors lose functional VHL, they overproduce HIF. One solution was to inhibit a protein called mTOR (mammalian target of rapamycin), which regulates HIF synthesis. Now treatments for kidney cancer are combining mTOR inhibitors with inhibitors of VEGF and of HIF synthesis. "So I think we are starting to make a little dent in kidney cancer," Kaelin says.
Kaelin's group hopes to identify genes that complement or cooperate with VHL mutations to promote kidney cancer. They are also looking for genes that become essential when VHL function is lost, because inactivating the products of such genes would be another therapeutic strategy. "I'm hopeful that by having an even more complete molecular understanding of kidney cancer, we can help guide the next generation of clinical trials, which almost certainly will involve combinations of drugs that affect different molecular pathways that are important in kidney cancer," Kaelin says.