The role of PI3K p110alpha and p110beta isoforms in oncogenesis

Upon activation by many receptors, the ubiquitously expressed isoforms of Class IA phosphatidylinositol-3-kinases (PI3Ks), p110α and p110β, generate lipid second messengers, which initiate signal transduction cascades that begin with activations of serine/threonine protein kinases at plasma membrane and culminate in phosphorylations of transcription factors in cell nucleus. Deregulation of PI3K signaling has been widely implicated in many cancers, and the PIK3CA gene encoding p110α has been found to be frequently mutated in human tumors. Interestingly, mutations in the PIK3CB gene encoding p110β have not been reported yet. The objective of my dissertation has been to determine the oncogenic properties and importance in normal signal transduction of PI3K p110α and p110β isoforms. In my gain of function studies, two most common tumor-derived mutants of p110α, E545K and H1047R, potently activated PI3K signaling and robustly transformed human mammary epithelial cells. Intriguingly, membrane localization via myristoylation further enhanced the activity of these mutants, implying that the activation mechanisms of these tumorigenic alleles are not simply via translocation to the membrane. To begin to elucidate the oncogenic mechanism of these mutants, several p85-binding domain mutants were examined. The results suggest that one possible oncogenic mechanism of these tumor mutants is to relieve the inhibitory effects of p85, while leaving intact the stabilization afforded by p85-binding. Interestingly, membrane localization via myristoylation highly activated p110β and the resulting myristoylated-p110β was transforming in both soft agar and orthotopic tumor formation assay, indicating that p110β does possess oncogenic potential. Further analysis showed this potential could be evoked by multiple point mutations and the mechanism might be independent of p110β's kinase activity. In my loss of function approach, p110β conditional knockout mice and mouse embryonic fibroblasts were generated. Ablation of p110β had profoundly negative effects on cell proliferation, but surprisingly little effect on classical growth factor signaling. Notably, re-expression of kinase-inactive p110β rescued growth almost as well as the wild-type enzyme. These data suggest that p110β may have critical kinase-independent functions. Since deletion of p110β blocks transformation by various oncogenes, the results presented here may have important implications for development of PI3K inhibitors to treat cancer and other diseases.