Funded Grants

Genetic lesions in the BMP pathway as a mechanism underlying brain tumors

Tumors that originate in the brain, primary brain tumors, can occur in people of any age with an overall incidence of new cases at roughly 1 in 7000 people each year. Most primary brain tumors are gliomas, composed of cells that show characteristics of immature glia. They arise primarily in the cerebral hemispheres, which can be especially devastating since this part of the brain is the seat of our highest intellectual functions. The most common glioma in adults, the glioblastoma multiforme, is also the most malignant. The most common type of malignant brain tumor in children is the medulloblastoma, which arises in the cerebellum. Both medulloblastomas and glioblastomas are resistant to radiation and chemotherapies, underscoring the importance of developing novel therapies targeted specifically at fighting these tumors.

In general, malignant and normal embryonic precursor cells are similar in that they are highly proliferative, have few differentiated features, express similar genes, and can often generate a mixture of cell types. In the case of medulloblastomas and gliomas, malignant cells share features with precursor cells in the embryonic cerebellum and cerebral hemispheres, respectively. Based on these similarities, it is hypothesized that medulloblastomas and gliomas are derived from precursor cells that persist in postnatal and adult brains. In particular, medulloblastomas are believed to originate from cerebellar granule cells, which normally give rise to granular neurons. Similarly, gliomas are thought to originate from glia, which can have stem cell-like properties. Evidence supporting the premise that gliomas are derived from glia include studies in which forced expression of oncogenes in glia leads to gliomas in mice. The suspected glial origin of gliomas and granule cell origin of medulloblastomas suggest that cells in these tumors share a common molecular makeup with their normal counterparts and that understanding the biology of one will contribute to a better understanding of the other.

Although in most cases the causes of gliomas and medulloblastomas are unclear, mutations in genes that regulate normal cell proliferation are likely to play a direct role, as with any other cancer. For example, an active EGF (epidermal growth factor) signaling pathway can maintain the proliferation of normal glial precursor cells and an overactive EGF pathway is observed in nearly half of malignant gliomas. Likewise, another extracellular signaling pathway, the sonic hedgehog (SHH) pathway, promotes proliferation of cerebellar granule cells and is suspected of underlying the abnormal proliferation of medulloblastoma cells. However, the genetic pathways that regulate normal and abnormal proliferation in the cerebral hemispheres and cerebellum are just beginning to be understood. A more in depth characterization of these pathways will provide additional targets for potential drug treatments.

Recent technical advances in mouse genetics should prove instrumental in understanding both normal brain development and the underlying causes of primary brain tumors. Namely, "conditional" genetic approaches have been developed that allow a gene of interest to be mutated specifically in cerebral or cerebellar cells, without affecting the function of this gene in other tissues of the body. The ability to mutate a gene in specific cell types or areas of the brain has important advantages. First, restricting the mutation to a subset of cells in the brain allows one to bypass any embryonic lethality that occurs as a result of mutating the gene in the whole animal. This is particularly useful since many genes suspected of regulating the proliferation of both embryonic brain cells and postnatal tumor cells also play essential roles in early developmental processes such as gastrulation. A second advantage of restricting the mutation to specific brain cells is that secondary effects on the brain due to other tissues that require the gene for their normal development or function are minimized. For example, a heart defect could have a significant influence on cell proliferation in the brain, without informing us on the direct role, if any, that the gene plays in the brain. And third, and most importantly, the ability to mutate a gene in a small number of cells in the postnatal or adult brain provides a more accurate model for the spontaneous genetic lesions that underlie malignant brain tumors in humans.

Using "conditional" mouse genetic approaches to mutate genes specifically in the cerebral hemispheres has already proven successful in assessing the function of several signaling pathways during development. For example, secreted signaling factors of the Bone Morphogenetic Protein (BMP) family have been demonstrated to be essential in inhibiting the proliferation of cerebral precursors that give rise to the choroid plexus. However, these BMP factors are also suspected of inhibiting the proliferation of a wider range of precursor cells in both the developing cerebral hemispheres and cerebellum. Based on these putative functions and because of the similarities between normal precursor cells and malignant tumor cells, loss of BMP signaling is also hypothesized to underlie the formation of medulloblastomas and gliomas.

This proposal has two aims. The first of which is to test directly in vivo whether BMP signaling is required to inhibit the proliferation of normal cerebral and cerebellar precursor cells. At least five genes encoding BMP ligands are expressed in the embryonic cerebellum and cerebral hemispheres, whereas only two BMP receptor genes are expressed. Therefore, to more effectively disrupt BMP signaling, both BMP receptor genes will be deleted in cerebral and cerebellar precursor cells using a "conditional" genetic approach. An increase in precursor cell proliferation accompanied by either a decrease in cell differentiation or apoptosis is expected in these mutants.

The second aim of this proposal is to test in mice whether mutations in the BMP pathway can lead to the formation of medulloblastomas and gliomas in the postnatal cerebellum and cerebral hemispheres, respectively. This second aim will take advantage of an innovation to the current approach of generating cerebral and cerebellar-specific gene mutations. In this approach, BMP receptor genes are mutated not only in specific cell types of the brain, but also at any time during the life of the animal. Moreover, the fraction of cells that lose the genes can be adjusted to either maximize the chances of obtaining tumors or minimized to mimic the rarer spontaneous mutations that occur in humans. Given the evidence that BMP signaling acts to inhibit proliferation of normal precursor cells during development and that downstream effectors of BMP signaling act as tumor suppressor genes in cell lines and other types of cancer, the formation of medulloblastomas and gliomas is expected in postnatal mice in which the BMP receptor genes have been mutated. The EGF and SHH pathways can also contribute to the formation of glioblastomas and medulloblastomas, respectively. Hence, it will be determined if overactivation of these pathways combined with loss of BMP signaling synergistically increases the frequency of tumorigenesis.

These studies will demonstrate whether loss of BMP signaling, on its own or in combination with other signals, can contribute to the formation of two of the most malignant types of primary brain tumors, medulloblastomas and glioblastomas. In turn the results obtained will either validate or weaken the case for designing drug therapies aimed at activating downstream components of the BMP pathway (on their own or in combination with other therapies) as a means to combat the proliferation of medulloblastomas and glioblastomas.