P3Ka: A Driver of Tumor Metastasis?

The link between PI3K and cancer is among the strongest in the genome. PI3K sits in a critical node in cancer signaling space where it integrates growth and survival signals from RTKs and Ras to the mTOR, MAPK, FOXO1 and GSK3β signaling pathways. The Class I PI3Ks are dimeric kinases comprised of one regulatory subunit (p85 α, β, or γ) and one catalytic subunit (p110 α, β, or γ) which phosphorylate phosphoinositides at the D-3 position of the inositol ring in response to environmental cues. Constitutive activation of the PI3Ks is transforming in experimental models and in human breast, colon and endometrial cancer, activating mutations in the gene for PIK3CA (p110α) are seen in 30% of the patient population. Inactivating mutations to the p85 regulatory subunit have been described in a subpopulation (~10%) of colorectal cancer patients. Moreover, the PTEN phosphoinositide phosphatase, nature’s brake to proliferative PI3K signaling, is a frequently mutated tumor suppressor gene [1, 2 and references therein]. Given the successes over the last decade in developing small molecule inhibitors targeting the BCR-ABL, EGFR, and BRAF oncogenic kinases, it is not surprising that numerous PI3K inhibitors have been developed and are currently being evaluated in human clinical trials. Despite the fact that among the p110 isoforms, p110α is preferentially mutated in cancer, the first wave of PI3K inhibitors to enter the clinic consisted of non-isoform selective (pan) molecules. While these inhibitors are broadly active in preclinical cancer models [3, 4] concern as to whether the lack of selectivity may adversely impact the therapeutic index has led to follow up efforts to develop isoform-selective PI3K inhibitors. Working from an imidazopyridine-based scaffold, Schmidt-Kittler et al optimized PI3Kα-selective inhibitors with nM potency against PI3Kα (J-series). Among class I PI3Ks (α, β, γ, δ) the J-series compounds averaged 20 to 30 fold selectivity in favor of the p110α isoform in vitro but did not discriminate between wt and mutant (E545K, H1047R) forms of the enzyme. Representative compounds from the series inhibited growth and phosphoAKT levels in HCT116 cells in the 100 nM range in vitro, independent of PI3K mutational status. An unexpected finding occurred during the in vivo evaluation of two such compounds from the series. Although phospho-AKT levels were robustly inhibited in HCT116 xenograft tumor tissue, this translated into only modest effects on tumor growth inhibition. In contrast, when the compounds were evaluated in a metastatic version of the HCT116 model, wherein tumor cells injected into the spleen spread to the liver and lungs, the emergence of distant metastases was significantly curtailed [5]. The conclusions from this work – that PI3Kα inhibitors may be more useful for limiting the spread of a primary tumor to distant sites than treating established tumors – would be strengthened by additional studies incorporating isoform specific shRNAs in one or more in vivo tumor models. Still, the findings are provocative and raise many interesting questions.