Regulation of medulloblastoma by the sterol synthesis pathwayOur goal is to discover properties of brain tumors that make them susceptible to new treatments, particularly treatments more effective and less toxic than current ones. In the past year about 350 people in the U.S., mostly children fewer than ten years old, were diagnosed with medulloblastoma, a devastating tumor arising in the cerebellum and the most common malignant brain tumor in childhood. The tumors are fast-growing, invasive, and must be treated with surgery, radiation, and chemotherapy. Of those afflicted, 60-80% will survive for five years, but treatment often severely impairs children, who are still growing rapidly. Two of ten survivors will exhibit serious neurological symptoms, such as difficulty balancing and speech loss. New ways to stop medulloblastoma are desperately needed.
The frequency of medulloblastoma is higher in people carrying certain types of altered chromosomes. Some of the exact genes involved were identified only relatively recently, and more remain to be found. The first discovery of a specific mutation that causes medulloblastoma came from our work and that of others on a gene called
PATCHED (PTCH). As developmental biologists, we had been studying the gene for many years because it is an important regulator that organizes embryo tissues. We began by working on the gene in the fruit fly
Drosophila, where it was discovered, and then used the fly gene to isolate the corresponding mouse and human genes. The human gene mapped to part of chromosome 9, near the location of an inherited birth defect syndrome called Basal cell nevus syndrome (BCNS) or Gorlin’s syndrome. We proved that mutations in
PTCH are responsible for the syndrome by showing that inheritance of the damaged gene tracks with inheritance of the birth defects. BCNS patients also have an increased incidence of certain tumors, including basal cell carcinoma (BCC) of the skin and medulloblastoma. To test whether the
PTCH mutations are in fact responsible for the increased incidence of these tumors, we engineered a mouse strain that, like the BCNS patients, had only a single working
PTCH gene. These mice developed medulloblastoma, and pathologists confirmed the similarity of the mouse tumors to human tumors. We also made a different mouse model of the skin cancer, and that too was confirmed. In addition the mice had birth defects akin to those sometimes seen in human BCNS patients, like abnormally large body size. BCC is the most common human cancer, and it was exciting to see how basic science--studying the control of cell fates in fly embryos--led directly to discovering the genetic basis of a human skin cancer, but it was even more exciting to think we had a chance to discover a way to prevent or control medulloblastoma which, unlike BCC, is life-threatening.
The initial findings gave us a strikingly powerful new handle on the problem of how medulloblastomas arise. The next step was to investigate whether
PTCH has a normal function in the cerebellum. The Patched1 (Ptc1) protein, which is encoded by the gene, is a component of a signaling system that is used in cell-cell communication in the growth of many of our tissues and organs. The signaling system is called Hedgehog (Hh) signaling, a name derived from the appearance of the mutant flies where it was discovered. The Hh protein, called Shh in us, is secreted by certain cells, moves through tissue, and is received by other cells. The receiving cells respond to the signal by changing their properties, in some cases by growing. We showed that Shh signal in the cerebellum is a powerful stimulant of growth, particularly of the type of neuron precursor cell from which medulloblastomas arise. Shh stimulates cell growth by binding to Ptc1, its receptor, and inactivating it; the ligand and receptor are antagonists. The findings made sense in light of the tumors. In people with BCNS, there is evidently not enough Ptc1 to adequately restrain the Shh signal, and some of the neuron precursor cells continue to grow beyond the proper time, or resume growth later. That simple idea is not the whole story. Other genes interact with
PTCH to control the frequency of tumors. For example PTCH mutant mice with mutations in the tumor suppressor gene
p53 have a substantially increased tumor frequency, while mutations in the gene for the signaling protein
Igf2 dramatically reduce the frequency.
It is not yet clear how many cases of medulloblastoma are due to damaged Shh signaling. At least a quarter of cases studied so far carry mutations in known components of the pathway, and others beyond those show signs of the pathway becoming altered through unknown mechanisms. History demonstrates clearly that studying powerful genetic models of disease, even if they reflect only a particular group of patients, often leads to key insights about more common forms of the disease. This was true, for example, in studying rare cholesterol metabolism diseases and discovering an important cause of heart disease and stroke. With this in mind, we are continuing to study the mouse model to discover how to block tumor growth. We study the roles of Shh and Ptc1 in normal cerebellum development as well as in tumors, since it is correct to view medulloblastoma as a developmental growth process gone awry.
We have provided our mouse model of BCNS to dozens of laboratories all over the world, and have been pleased at how useful it has been. We now know many of the genes regulated by Shh and Ptc1 in the cerebellum, including the cell cycle gene
Nmyc, which we and others identified as a direct target of Shh regulation. Another exciting development was the discovery of a naturally occurring chemical called cyclopamine, found in a mountain plant called the Corn Lily, which specifically binds to and inhibits a component of Hh signaling. Cyclopamine and its derivatives are being explored as possible anti-cancer drugs, but that has the difficulty that inhibiting the Hh pathway may also cause developmental defects, particularly in children. Several university labs and companies have engaged in screens of libraries of small molecules to find those that influence Hh signaling, and some promising lead molecules have been found. However most of the initial clinical trials of these drugs have, sadly, failed. New approaches are needed.
In recent years we have used genomics methods to learn about gene regulation by Shh and Ptc1. We also developed a method to mark and purify early tumor cells, a remarkable opportunity to follow the genesis of a brain tumor, and changes in gene activity, from a stage when it has only a thousand or so cells to a mature invasive tumor. Most recently we have been identifying direct targets of Shh by determining where a transcription factor controlled by Shh binds to chromosomes, thus finding regulated genes. From these experiments came a surprise, which was that medulloblastoma tumors massively increase the production of enzymes involved in sterol synthesis, compared to normal cerebellum cells. This led us to investigate what product of the sterol synthesis pathway (SSP) matters to the cells. We found that the growth of the cultured tumor cells was dependent on an increased supply of a particular class of molecules, the oxysterols. These molecules are chemically altered forms of cholesterol. There are many kinds, and we found that some are much more potent than others in supporting growth. If the SSP was inhibited, for example with the famous statins that are used to control sterol levels in millions of people, cultured medulloblastoma cells grow slowly. We discovered that certain oxysterols are powerful activators of the Hh signaling pathway, as potent as Shh itself, and we are now exploring how oxysterols influence components of the pathway.
Using SSP inhibitors as one way to control medulloblastoma growth is appealing because the drugs are already approved, but there is a serious problem in the form of the blood-brain barrier; most statins have been engineered to work well in the liver but not get into the brain. However the first pressing step is to find out whether the SSP is in fact important for tumor growth in an animal model. We can do this in two ways: using mouse mutants that have altered sterol metabolism, and using SSP-inhibiting drugs under conditions that allow penetration of the blood-brain barrier. We do not expect SSP drugs alone to be sufficient to eliminate tumors, but they might be powerful and gentle tools to use in combination with other treatments. The experiments described in this grant are designed to test these ideas.
I have a new friend, a medulloblastoma patient who is a cub scout. He and his pack sold popcorn to raise money for the Center for Children’s Brain Tumors that I am co-chairing at Stanford. Getting to know him, and appreciate what he and his family have gone through, is about as strong a motivator to do something important as I have ever experienced. We shall try.
