Approximately 35% of follicular thyroid carcinomas harbor a chromosomal translocation that fuses paired box gene 8 (PAX8) with the peroxisome proliferator-activated receptor gamma gene (PPARG), resulting in production of a PAX8-PPARG fusion protein denoted PPFP. PPFP contains the full sequence of the nuclear receptor PPARG1, and hence PPFP binds to PPARGresponsive genes and to PPARG ligands. We have created the first transgenic mouse model of this cancer. The cancer is locally invasive and forms lung metastases. Treatment with the PPARG agonist pioglitazone (Pio) shrinks the thyroid almost to control size and eliminates metastatic disease. Most remarkably, this therapeutic response is characterized by a trans differentiation-type process whereby the remaining thyroid cells develop large lipid droplets and express a wide array of PPARG- inducible adipocyte genes. Since PPARG is the master regulator of adipogenesis, these results indicate that, in the presence of Pio, PPFP is strongly PPARG-like. We postulate that the anti-tumor action of Pio in PPFP cancers is tied to this adipocyte trans differentiation-like effect;i.e., the more the thyroi cancer cells acquire a mature adipocyte phenotype, the less they retain of their malignant phenotype.
In Aim 1, we will evaluate the oncogenic action of PPFP in the mouse model. There is evidence that, in the absence of Pio, PPFP can inhibit PPARG induction of some target genes, and that inhibition of endogenous PPARG may underlie the oncogenic nature of PPFP. Therefore, we will test whether the genetic deletion of PPARG mimics the expression of PPFP in terms of the development of thyroid cancer. We also will assess whether both the PAX8 and PPARG DNA binding domains within PPFP are important by studying mice in which PPFP has appropriate mutations. Analyses of histology, gene expression and DNA binding will provide insight into the genes regulated by PPFP that contribute to the development and progression of thyroid cancer. We will use a non-adipogenic PPARg ligand to test whether the adipogenic nature of the Pio response is critical to its therapeutic effect. We also will test whether arsenic trioxide is therapeutic, especially in combination with Pio. This hypothesis derives from the observation that PPFP + Pio strongly induces AQP7, a channel protein through which arsenic enters cells.
In Aim 2, we will perform a phase II clinical trial to determine whether Pio is therapeutic in patients with metastatic PPFP thyroid cancer not treatable by standard therapies. The primary endpoint will be a decrease in the size of metastases. A secondary endpoint will be measurement of lipid content of metastases that do not completely resolve, based upon the observation that treatment of our mouse model of this cancer with Pio results in an adipogenic response in the surviving thyroid cells. Other secondary endpoints include changes in serum thyroglobulin and testing the ability of Pio to induce radioiodine uptake in the cancer, followed b radioiodine therapy if indicated. Overall, these studies will help elucidate mechanisms through which PPFP contributes to thyroid cancer and will help identify novel therapies both in the transgenic mouse model and in patients.
Follicular thyroid cancers contain a protein called PPFP that is not found in normal cells and that underlies development of this cancer. These studies will investigate the mechanisms through which PPFP contributes to thyroid cancer, and will test novel therapies both in mice and patients with this disease.