Activating mutations of PIK3CA, the gene encoding for the p110a catalytic subunit of PI3K, is the most frequent genomic alteration in breast cancer and is a promising target for cancer therapy. Selective PI3K p110a (PI3Kα) inhibitors have shown preferential antitumor activity against PIK3CA-mutant cancer cells and we have observed high clinical activity in early clinical trials in patients with PIK3CA-mutant breast cancer. However, not all patients benefit and even those who initially respond inevitably develop resistance over time. Our project is aimed at identifying mechanisms of resistance to PI3Kα inhibitors in breast cancer and at developing hypothesis-driven therapeutic strategies to prevent or delay its occurrence. In preliminary work with samples from patients initially responding to thePI3Kα inhibitor BYL719 we have observed a series of PTEN genomic alterations in the biopsies collected at the time of disease progression, leading to a convergent PTEN- null phenotype. Our hypothesis is that PTEN loss and the resulting PI3Kß activation may render PI3KCA mutant tumors resistant to PI3Kα inhibitors. We plan to validate PTEN loss as a mechanism of acquired resistance to PI3Kα inhibitors in several models. Both constitutive and inducible loss of PTEN expression will be modeled in cells sensitive to PI3Kα inhibitors. Moreover, we will study the effects of inducible PTEN loss in transgenic mice by crossing transgenic mice with inducible PIK3CA knock-in and PTEN knock-out. We will also investigate whether the genomic alterations in PTEN observed in the metastasis were present in rare cell populations in the primary tumor or they were acquired during treatment. We will design droplet digital PCR (ddPCR) assays against these patient-specific somatic mutations to test their presence in cfDNA isolated from samples down to 0.01% frequency. In tumors with intrinsic resistance to anti-PI3Kα agents, we have preliminary findings that suggest that expression/activation of PDK1 acts as an AKT-independent mechanism mTORC1 activation that would result in resistance. The role of PDK1 in limiting the therapeutic effects of PI3Kα inhibition will be studied by overexpressing PDK1 in cells sensitive to PI3Ka inhibitors or suppressing PDK1 in cells resistant to these agents by either mRNA knock-down or pharmacological inhibition. PI3Ka blockade in combination with PDK1 inhibition will be tested both in vitro and in vivo. Moreover, we plan to dissect the molecular mechanisms by which PDK1 can activate mTOR by-passing AKT. These findings will be validated in clinical samples from patients treated with PI3Ka inhibitors. We will analyze samples from patient treated with either BYL719 or GDC-0032 to determine the frequency of PTEN loss at time of disease progression. PDK1 expression/activity will also be measured in samples collected during acute treatment (4-8 weeks). This proposal will allow selecting patients that are most likely to benefit from these therapies and to design rational combinatorial approaches to counteract the emergence of resistance.