Andrew Koff: Oligodendroglioma

The RCAS/Tv-A model combined with knockout mice allows us to carry out structure-function analysis and link biochemical and cellular changes directly to the phenotype of interest. Combining the RCAS/Tv-A system with the concept of rescuing phenotype in null mice allows us to use structure-function analysis to probe how proteins participate in tumor development. Enlarge Image The RCAS/Tv-A model combined with knockout mice allows us to carry out structure-function analysis and link biochemical and cellular changes directly to the phenotype of interest. Combining the RCAS/Tv-A system with the concept of rescuing phenotype in null mice allows us to use structure-function analysis to probe how proteins participate in tumor development.

Subsets of oligodendroglioma exist, each with distinct clinical responses that are dictated by the underlying biochemical and molecular profile of the tumor. Currently, the effectiveness of treatment strategies for glioma is poor and new therapeutic targets might be obtained through a better understanding of the nature of this disease.

Eric Holland’s group developed a mouse model of PDGF-amplified oligodendroglioma by overexpressing PDGF in nestin positive progenitor cells. This model faithfully recapitulates many hallmarks of the human disease. We showed that the CDK inhibitor p21 is required for proliferation in th subtype of oligodendroglioma driven by PDGF. It executes this function by promoting the nuclear accumulation of cyclin D1-CDK4 complexes. Also, in contrast to the molecular and cellular analysis that has been carried out in cultured epithelial and mesenchymal cells, the structurally related CDK inhibitor, p27, did not act in a redundant fashion. In PDGF-induced glial tumors, p27 is a tumor suppressor that facilitates DNA repair events. These distinct mechanistic roles may underlie the relationship of these CDK inhibitors to patient prognosis, with high p21 levels correlated with poor prognosis and high p27 levels correlated to good prognosis.

A cell cycle view of PDGF-induced ODG in mice. Enlarge Image A cell cycle view of PDGF-induced ODG in mice.

Consequently, current efforts in our laboratory are focused on addressing the following:

  1. How is p21 expression regulated to promote tumor development in growth factor driven glioma cells? Although the elevation of p21 protein can contribute to arrest in many cell types and in response to DNA damage checkpoints, p21 accumulation is associated with proliferation in growth-factor induced glial cells, vascular smooth muscle cells, and stimulated T-cells. To understand how p21 protein accumulation is controlled in these cell types, we isolated p21-containing protein complexes from growth-factor transformed glial cells and used mass-spectroscopy to identify interacting proteins. We identified a novel E3 ubiquitin-ligase regulating p21 in these cell types and have shown that expression of this gene product is reduced in human glioma and reduction of this gene product in mice can enhance tumor development in the RCAS-TvA mouse model. Currently, we are focusing our efforts toward (a) reconstituting this ubiquitination process in vitro, (b) identifying gene products that can modulate this activity by high-throughput siRNA screens and biochemical approaches, and (c) looking for this circuit in other cell types in which p21 is a growth-activator.
  2. What prevents the accumulated p21 from inhibiting cyclin D1-CDK4? Recent work by Stacy Blain’s group demonstrated that tyrosyl-phosphorylation of p27kip1 can interfere with its ability to inhibit bound cyclin D-CDK4. Due to the structural similarity of p27 and p21 in the cyclin-CDK binding domain, we have focused our efforts on this type of modification in p21. Furthermore, we can directly assess the significance of such a modification by using the RCAS/Tv-A system.
  3. Our data suggests that interfering with cyclin D1-CDK4 nuclear accumulation might hold therapeutic promise in this disease. To directly address this we have created a series of mice to probe the requirements for cyclin D1, CDK4 and CDK2 in this disease, as we have done for p21 and p27. Furthermore, we are leveraging the analogue sensitive kinase allele approach pioneered by Kevan Shokat to develop a CDK4 mutant animal in which we can modulate the amount of CDK4 activity using an ATP analogue in vivo to assess the tumor response in situ to such inhibtion. These analyses promise to identify whether or not the cyclinD1-CDK4 cell-cycle axis is a viable target in glioma.
  4. How does p27 deficiency contribute to genetic instability and accelerated progression of growth-factor induced glioma? Much of our previous work has been focused on defining the ways in which p27 loss contributes to the pathogenesis of cancer. Such work has defined apoptotic functions as well as a role in the maintenance of genomic integrity. Recently, in the RCAS-TvA model, we have shown that p27 deficiency leads to increased CDK2 activity which appears to be affecting the efficiency (or rate) of DNA repair and this is associated with increased genomic instability. However, to directly establish the role of CDK2 activity in accelerating tumor progression, we are using the RCAS-TvA model to complement a CDK2 deficiency with an analogue sensitive CDK2 allele. This animal model, and cell lines derived from this model, should allow us to identify relevant substrates.
  5. Given the extensive biochemical similarity of p27 and p21, and the fact that these act in a redundant fashion to promote cyclin D1-CDK4 assembly/nuclear accumulation in mesenchymal cells, why do they have such different roles in mouse models of glioma? Having such a sensitive system as the RCAS-TvA models to study this disease we believe that we are in an excellent position to understand at fine detail the differences and regulatory events that control the switch from inhibitor to activator.