Leukemia stem cells (LSCs) are regarded as the origins and key therapeutic targets of leukemia, but limited knowledge is available on the key determinants of LSC ‘stemness’. Using single-cell RNA-seq analysis, we identify a master regulator, SPI1, the LSC-specific expression of which determines the molecular signature and activity of LSCs in the murine Pten-null T-ALL model. Although initiated by PTEN-controlled β-catenin activation, Spi1 expression and LSC ‘stemness’ are maintained by a β-catenin-SPI1-HAVCR2 regulatory circuit independent of the leukemogenic driver mutation. Perturbing any component of this circuit either genetically or pharmacologically can prevent LSC formation or eliminate existing LSCs. LSCs lose their ‘stemness’ when Spi1 expression is silenced by DNA methylation, but Spi1 expression can be reactivated by 5-AZ treatment. Importantly, similar regulatory mechanisms may be also present in human T-ALL.

PI3K/AKT signaling is known to regulate cancer metabolism, but whether metabolic feedback regulates the PI3K/AKT pathway is unclear. Here, we demonstrate the important reciprocal crosstalk between the PI3K/AKT signal and pentose phosphate pathway (PPP) branching metabolic pathways. PI3K/AKT activation stabilizes G6PD, the rate-limiting enzyme of the PPP, by inhibiting the newly identified E3 ligase TIRM21 and promotes the PPP. PPP metabolites, in turn, reinforce AKT activation and further promote cancer metabolic reprogramming by blocking the expression of the AKT inhibitor PHLDA3. Knockout of TRIM21 or PHLDA3 promotes crosstalk and cell proliferation. Importantly, PTEN null human cancer cells and in vivo murine models are sensitive to anti-PPP treatments, suggesting the importance of the PPP in maintaining AKT activation even in the presence of a constitutively activated PI3K pathway. Our study suggests that blockade of this reciprocal crosstalk mechanism may have a therapeutic benefit for cancers with PTEN loss or PI3K/AKT activation.

Combining immune checkpoint therapy (ICT) and targeted therapy holds great promises for broad and long-lasting anti-cancer therapies. However, combining ICT with anti-PI3K inhibitors have been challenging because the multifaceted effects of PI3K on both cancer cells and immune cells within the tumor microenvironment. Here we find that intermittent but not daily dosing of a PI3Kα/β/δ inhibitor, BAY1082439, on Pten-null prostate cancer models could overcome ICT resistance and unleash CD8+ T cell-dependent anti-tumor immunity in vivo. Mechanistically, BAY1082439 converts cancer cell-intrinsic immune-suppression to immune-stimulation by promoting IFNα/IFNγ pathway activation, β2-microglubin expression and CXCL10/CCL5 secretion. With its preferential regulatory T cell inhibition activity, BAY1082439 promotes clonal expansion of tumor-associated CD8+ T cells, most likely via tertiary lymphoid structures. Once primed, tumors remain T cell-inflamed, become responsive to anti-PD-1 therapy and have durable therapeutic effect. Our data suggest that intermittent PI3K inhibition can alleviate Pten-null cancer cell-intrinsic immunosuppressive activity and turn “cold” tumors into T cell-inflamed ones, paving the way for successful ICT.

Targeting the PI3K pathway is a promising strategy for treating prostate cancers with PTEN-loss. However, current anti-PI3K therapies fail to show long lasting effects. We find that not only the PI3Kα- and PI3kβ-isoforms, but also PI3Kδ, are associated with the epithelial-mesenchymal transition (EMT), a critical process distinguishing indolent from aggressive prostate cancer. This suggests that cotargeting PI3Kα/β/δ could preempt the rebound activation of the parallel pathways induced by α- or β-isoform-selective inhibitor and prevent EMT. Indeed, BAY1082439, a new selective PI3Kα/β/δ inhibitor, is highly effective in inhibiting null prostate cancer growth and preventing EMT in the mutant metastatic model. The anti-PI3Kδ property of BAY1082439 further blocks B-cell infiltration and lymphotoxin release, which are tumor microenvironment factors that promote castration-resistant growth. Together, our data suggest a new approach for the treatment of prostate cancer by targeting both tumor cells and tumor microenvironment with PI3Kα/β/δ inhibitor. .

Prostate cancers generally lack T cell infiltration and display resistance to immune checkpoint therapies (ICT). We found that intermittent but not daily dosing of PI3Kapha/beta/gamma inhibitor BAY1082439 on a Pten-null spontaneous prostate cancer model could overcome ICT resistance and unleash CD8+ T cell-dependent anti-tumor immunity in vivo. Mechanistically, BAY1082439 converts Pten-null cancer cell-intrinsic immune-suppression to immune-stimulation by promoting IFNalpha/gamma pathway activation, B2M expression and CXCL10/CCL5 secretion. Together with its preferential Treg inhibition activity, BAY1082439 promotes clonal expansion of tumor-associated CD8+ T cells. Once primed, tumors remain as T cell-inflamed and become responsive to anti-PD-1 therapy. Our data suggest that intermittent PI3K inhibition can alleviate Pten-null cancer cell-intrinsic immunosuppressive activity and turn "cold" tumors into T cell-inflamed ones, paving the way for successful ICT.

1 0 PPP metabolites promote AKT activation by negatively regulating

<p>Figure S1. Effects of BAY1082439 on cell apoptosis and cell cycle; Figure S2. BAY1082439 is well tolerated in vivo; Figure S3. BAY1082439 treatment impedes prostate tumor invasion; Figure S4. BAY1082439 treatment inhibit CRPC growth in castrated CP mice; Figure S5. B cell infiltration promote CRPC growth in CP model; Figure S6. Effects of BAY1082439 on splenic B cell growth and tumor infiltrating CD8 T cell</p>

<div>Abstract<p>Targeting the PI3K pathway is a promising strategy for treating prostate cancers with PTEN-loss. However, current anti-PI3K therapies fail to show long lasting <i>in vivo</i> effects. We find that not only the PI3Kα- and PI3kβ-isoforms, but also PI3Kδ, are associated with the epithelial–mesenchymal transition (EMT), a critical process distinguishing indolent from aggressive prostate cancer. This suggests that cotargeting PI3Kα/β/δ could preempt the rebound activation of the parallel pathways induced by α- or β-isoform–selective inhibitor and prevent EMT. Indeed, BAY1082439, a new selective PI3Kα/β/δ inhibitor, is highly effective <i>in vivo</i> in inhibiting <i>Pten-</i>null prostate cancer growth and preventing EMT in the mutant <i>Pten/Kras</i> metastatic model. The anti-PI3Kδ property of BAY1082439 further blocks B-cell infiltration and lymphotoxin release, which are tumor microenvironment factors that promote castration-resistant growth. Together, our data suggest a new approach for the treatment of prostate cancer by targeting both tumor cells and tumor microenvironment with PI3Kα/β/δ inhibitor. <i>Mol Cancer Ther; 17(10); 2091–9. ©2018 AACR</i>.</p></div>