A Targeted shRNA Knockdown Screen of Phosphoinositide Modulator Genes Reveals INPP5B As Required for Cell Cycle Progression of MLL-AF9 Acute Myeloid Leukemia Cells

Abstract 2444

Phosphoinositides are pivotal regulatory and scaffolding lipids that coordinate key aspects of cellular physiology. For example, phosphatidylinositol (4,5)-bisphosphate (PIP2) regulates endocytosis, exocytosis, membrane ruffling and the function of actin, and phosphatidylinositol (3,4,5)-trisphosphate (PIP3), downstream of phosphoinositide-3-kinase (PI3K), is a critical mediator of signal transduction. In acute myeloid leukemia (AML) mutations of FLT3 or KIT induce constitutive activation of PI3K/AKT signaling; and PTEN and SHIP1, two PI3K antagonists, have been linked to hematopoietic stem cell maintenance, with experimental deletion of these genes leading to leukemia. Little is known about the role of other phosphoinositide modulators in normal or leukemic hematopoiesis. We hypothesized that a genetic knockdown screen, utilising lentivirally delivered short hairpin RNAs (345 pLKO.1 puro vectors) targeting 92 genes with known or supposed roles in the regulation of phosphoinositide metabolism, would reveal novel regulators of proliferation and differentiation in AML, representing candidate therapeutic targets.

In an initial screen, hairpins were tested for their ability to impact on proliferation or survival of human THP-1 AML cells using fold change in fluorescence of the cell viability indicator alamarBlue relative to control cells (infected with a non-targeting control hairpin) over four days as a surrogate. THP-1 cells harbor a t(9;11) chromosomal translocation which is the cytogenetic hallmark of MLL-AF9. Lentiviral supernatants, prepared individually in 96 well plates with one vector per well, were used to infect target cells, again in 96 well plate format with one lentiviral supernatant per well. 24 hours following infection puromycin was added. The alamarBlue dye assay was performed on replicate samples 48 hours and six days later. All non-infected cells were killed by puromycin exposure within 48 hours. 38 shRNAs, including two targeting PTEN, increased proliferation or enhanced survival, whereas 158, including multiple constructs targeting AKT1, AKT2 and PI3KR2, reduced proliferation or induced apoptosis (versus control cells fold change± 2SD). These results confirmed that the screening strategy was effective because it correctly identified known pro- and anti-proliferative regulators of phosphoinositide metabolism.

Next, we expanded the screen to include six other AML cell lines (MonoMac1, HL60, K562, NB4, Kasumi-1 and U937). 14 genes met our criteria for a “hit”: at least two knockdown constructs per gene needed to reduce proliferation on average by at least 55% compared to non-targeting shRNA control across all cell lines screened. We focused our further investigations on type II inositol-1,4,5-trisphosphate 5-phosphatase (INPP5B), because Inpp5b^−/- mice are viable into adulthood with no known hematopoietic phenotype, suggesting INPP5B might selectively regulate leukemia cells. INPP5B catalyses the conversion of inositol-1,4,5-trisphosphate (IP3) to inositol-1,4-bisphphate (IP2) and hence contributes to cellular calcium signalling mediated through phospholipase C. Knockdown of Inpp5b also reduced proliferation of murine MLL-AF9 AML cells in liquid culture using three separate shRNA constructs that reduced transcript levels by 70–80%, and largely abolished colony formation in methylcellulose; residual colonies were much smaller than control colonies. While wild-type KIT⁺ bone marrow stem and progenitor cells were robustly immortalized by infection with an MLL-AF9 retroviral vector, Inpp5b^−/− KIT⁺ cells either failed to immortalize or exhibited weak immortalisation, with fewer, smaller colonies, reminiscent of the knockdown phenotype. In cell cycle analyses of cells harvested from the third round of serial replating in methylcellulose, 90% of Inpp5b^−/− MLL-AF9 cells were in G1 versus 41% of WT cells. No increase in apoptosis was observed.

These data implicate INPP5B as a novel and selective regulator of leukemic versus normal hematopoiesis.