Methods of treating blood cancers

Abstract

A method of treating a blood cancer comprising the administration of therapeutically effective amounts of a phosphatidylinositol 3-kinase gamma (PI3Kg) inhibitor and an antimetabolite, optionally also a therapeutically effective amount of a BCL-2 inhibitor, to a patient in need thereof. A method of identifying a blood cancer patient susceptible to treatment with a PI3Kg inhibitor. A method of treating a blood cancer comprising administration of a therapeutically effective amount of a p-21-activated kinase 1 (PAK-1) inhibitor, optionally also a therapeutically effective amount of a BCL-2 inhibitor, to a patient in need thereof.

Description

[0001] METHODS OF TREATING BLOOD CANCERS

[0002] FEDERAL FUNDING STATEMENT

[0003] This invention was made with government support under W81XWH-20-1-0684 awarded by the Defense Health Agency, Medical Research and Development Branch, and CA225191 awarded by the National Institutes of Health. The government has certain rights in the invention. FIELD

[0004] There are provided methods for the treatment of blood cancers. BACKGROUND

[0005] Despite recent progress with development of targeted therapies for some genetic subsets of acute leukemias, many disease subtypes lack mechanistically targeted treatment and most patients have poor long-term outcomes. Acute myeloid leukemia (AML) and acute lymphocytic leukemia (ALL) are the most common types of acute leukemia in adults and children, respectively. Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare and aggressive hematologic malignancy that originates from the plasmacytoid dendritic cell (pDC) lineage.

[0006] The mammalian phosphoinositide 3-kinase (PI3K, also called phosphatidylinositol 3- kinase) family contains 8 isoforms that can be divided into several classes. Class I PI3Ks, divided into class IA and IB, generate 3-phosphoinositide lipids to activate signal transduction pathways, while class II and class III PI3Ks are regulators of membrane traffic along the endocytic route. Class IA PI3Ks include three catalytic subunits a, , and 5 (encoded by PIK3CA, PIK3CB, and PIK3CD) with regulatory subunits p85a, p55a, p50a, p85p, and p55y (encoded by PIK3R1, PIK3R2, and PIK3R3) Class IB PI3K has only one catalytic subunit pl lOy (encoded by PIK3CG) and 2 regulatory subunits p!01(encoded by PIK3R5) and p84 / p87 (encoded by PIK3R6) and is activated by G protein-coupled receptors (GPCRs) via heterotrimeric G proteins. Several cancers harbor activated class IA PI3Ks, and numerous pathway inhibitors are approved or in development. In contrast, Class IB PI3K components (i.e., the enzymatic pl lOy and regulatory subunit(s)) have received less attention, and therapeutic focus has been limited to reprogramming macrophages with PI3Ky inhibitors for solid tumor immunotherapy. As a result, the role for PI3Ky as a cell-intrinsic cancer driver remains unclear.

[0007] WO2021 / 242859 proposes methods for reducing the viability of cells that express PI3Ky, treating myeloid malignancies (e.g. AML), and sensitizing cells, particularly cancer cells that express PI3Ky, to chemotherapeutic agents. The methods comprise contacting a cell or administering to the subject a PI3K inhibitor that inhibits PI3Ky in an isoform-specific manner. Examples of said PI3Ky inhibitors are IPI-549 (eganelisib), AS252424, and AS605240. However, WO2021 / 242859 discloses that therapies using PI3Ky inhibitors in combination with antimetabolites are not effective. In particular, chemosensitization mediated by PI3Ky inhibition with IPI-549 was not observed with cytarabine, azacytidine, decitabine, methotrexate, glasdegib, docetaxel, and oxaliplatin.

[0008] Similarly, the present inventors have previously discovered a dependency on PI3Ky signalling by BPDCNs (see Q. Luo LLS Award Abstract https: / / www.lls.org / award / defining- pik3r5-related-pi3k-gamma-dependency -novel -therapeutic-target-blood-cancers). They found that BPDCN is uniquely dependent on PIK3R5 and its partner PIK3CG and confirmed that a PI3Ky blocking drug can suppress BPDCN cells more effectively than AML cells.

[0009] Blood cancer therapies that operate via inhibition of PI3Ky are under-explored, and therefore present the possibility for discovering new therapeutic strategies for blood cancers, particularly treating aggressive leukemias, such as BDPCN.

[0010] SUMMARY

[0011] Further investigations have led to the development of such therapeutic strategies including the targeting of the PI3Ky pathway as a means for treating blood cancers.

[0012] In a first aspect, there is provided a method of treating a blood cancer comprising the administration of therapeutically effective amounts of a PI3Ky inhibitor and an antimetabolite to a patient in need thereof. Further embodiments of this first aspect include: a PI3Ky inhibitor and an antimetabolite for use in the treatment of a blood cancer; and the use of a PI3Ky inhibitor and an antimetabolite in the manufacture of a medicament or medicaments for the treatment of a blood cancer. In a further embodiment of the first aspect, there is provided a method of treating a blood cancer comprising the administration of therapeutically effective amounts of a PI3Ky inhibitor, an antimetabolite, and a BCL-2 antagonist to a patient in need thereof. In another embodiment, this includes: a PI3Ky inhibitor, an antimetabolite, and a BCL-2 antagonist for use in the treatment of a blood cancer; and the use of a PI3Ky inhibitor, an antimetabolite and a BCL-2 antagonist in the manufacture of a medicament or medicaments for the treatment of a blood cancer. In a second aspect, there is provided a method of identifying a blood cancer patient susceptible to treatment with a PI3Ky inhibitor comprising: (i) comparing the levels of a factor which is indicative of the expression of the gene PIK3R5 and the levels of a factor which is indicative of the expression of at least one gene selected from the group consisting of IRF7, LCP2, CCL5, CXCL9, PLSCR1, LYN, JAK2, IFNGR1, IRF8, CD86, TLR2, FAS, STAT1, TAPBP, B2M, CD74, TAPI, CASP1, LAP3, IL18R1, TNFRSF1B, SELL, IL10RA, TNFSF10, IFITM3, STAT2, IRF9, EIF2AK2, LY6E, BST2, and RTP4 in cancerous blood cells obtained from a blood cancer patient to the levels of those same factors in non-cancerous blood cells obtained from said patient, or to the levels of those same factors known to be present in non-cancerous blood cells of healthy individuals, or to the levels of those same factors known to be present in cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor; and (ii) selecting said patient for treatment if the levels of said factors in their cancerous blood cells are higher than the levels of said factors in their non-cancerous blood cells, or higher than the levels of said factors in the non-cancerous blood cells of healthy individuals, or higher than the levels of said factors in the cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3 Ky inhibitor. The 32 genes identified above are sometimes referred to herein as the “innate immune response signature” (IIRS).

[0013] In a third aspect, there is provided a method of treating a blood cancer comprising administration of a therapeutically effective amount of a p-21 -activated kinase 1 (PAK-1) inhibitor to a patient in need thereof. Further embodiments of this third aspect include: a PAK-1 inhibitor for use in the treatment of a blood cancer; and the use of a PAK-1 inhibitor in the manufacture of a medicament for the treatment of a blood cancer.

[0014] BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Aspects and / or embodiments will now be described in detail with reference to the accompanying drawings, in which:

[0016] Figure la shows the area under the entire dose response curve (AUC) of leukemia Patient Derived Xenograft PDX samples after ex vivo treatment (in triplicate) with increasing doses of eganelisib (0, 0.25, 0.5, 1, 2 and 4 pM) for 72 hours indicating vulnerability of AML, ALL, and BPDCN PDX models to PI3Ky inhibition. All leukemias with elevated PIK3R5 showed sensitivity to eganelisib, while those with low PIK3R5 expression did not. Data are means ± S.E.M. with Mann-Whitney test.

[0017] Figure lb shows top enriched gene sets from Gene Set Enrichment Analysis (GSEA) comparing high PIK3R5 (PIK3R5-high; FPKM>20) to low PIK3R5 (PIK3R5-low; FPKM<10) expression groups.

[0018] Figure 1c shows overall survival of IIRS± and PIK3R5± AML patients from The Cancer Genome Analysis (TCGA) Program log-rank test.

[0019] Figure 2a shows GSEA indicating OXPHOS to be the most suppressed gene set upon eganelisib treatment (1 pM for 48 hours) or PIK3R5 depletion (PIK3R5i) in CAL-1 cells.

[0020] Figure 2b shows Seahorse assays detecting mitochondrial respiration in the indicated leukemia cell lines upon treatment with eganelisib or LY294002 (each 1 pM for 48 hours). Data are means ± S.D. from 5 technical replicates.

[0021] Figure 2c shows Seahorse assays detecting mitochondrial respiration in control, PIK3R5 depleted, or PIK3CG depleted CAL-1 cells. Data are means ± S.D. from 5 technical replicates.

[0022] Figure 2d shows GSEA results indicating TNFa signalling via NFkB to be the most activated gene set upon eganelisib treatment or PIK3R5 depletion in CAL-1 cells.

[0023] Figure 2e shows significantly upregulated or downregulated genes in CAL-1 cells treated with eganelisib or depleted of PIK3R5. Five genes belonging to the TNFa signalling via NFkB gene set are labelled in red.

[0024] Figure 2f shows RT-qPCR detection of gene expression in CAL-1 cells treated with eganelisib (1 pM), with or without NFkB inhibitors IKK-16 (0.5 pM) or SC-514 (20 pM), for 2 days validating the involvement of NFKB in activation of these tumor suppressor genes because it was found that eganelisib no longer increased expression of those 5 genes. Data are means ± S.D. from 3 technical replicates.

[0025] Figure 2g shows Seahorse assays detecting mitochondrial respiration in CAL-1 cells treated with eganelisib (1 pM), with or without NFkB inhibitors IKK-16 (0.5 pM) or SC-514 (20 pM), for 2 days. Data are means ± S.D. from 5 technical replicates.

[0026] Figure 2h shows relative CAL-1 viability after treatment with eganelisib (1 pM), with or without NFkB inhibitors IKK-16 (0.5 pM) or SC-514 (20 pM), for 2, 4, and 6 days. Data are means ± S.E.M. from 3 biological replicates. Figure 3a shows a Volcano plot showing consistently upregulated (positive log2 fold change) or downregulated (negative log? fold change) phosphorylation sites in CAL-1 cells treated with eganelisib or depleted of PIK3R5 (PIK3R5i). Adjusted P values are shown.

[0027] Figure 3b shows western blotting for the indicated proteins in control, PIK3R5 depleted, or PIK3CG depleted CAL-1 cells.

[0028] Figure 3c shows western blotting for the indicated proteins in CAL-1 and DF-DCN-1 cells after treatment with eganelisib or LY294002 (each 1 pM for 48 hours).

[0029] Figure 3d shows western blotting for the indicated proteins in CAL-1 cells upon treatment with FRAX597, FRAX1036, or MK-2206 (each 1 pM for 48 hours).

[0030] Figure 3e shows relative numbers of CAL-1 cells upon treatment with FRAX597, FRAX1036, or MK-2206 (each 1 pM) for 2, 4, and 6 days. Data are means ± S.E.M. from 3 biological replicates.

[0031] Figure 3f shows relative cell numbers of CAL-1 cells expressing empty vector (EV) or PAK1 mutants after treatment with vehicle control (DMSO) or eganelisib (1 pM) for 2, 4, and 6 days. Data are means ± S.E.M. from 3 biological replicates.

[0032] Figure 3g shows Seahorse assays detecting mitochondrial respiration in CAL-1 cells expressing empty vector (EV) or PAK1 mutants after treatment with vehicle control (DMSO) or eganelisib (1 pM) for 2 days. Data are means ± S.D. from 5 technical replicates.

[0033] Figure 3h shows RT-qPCR detection of gene expressions in CAL-1 cells expressing empty vector (EV) or PAK1 mutants after treatment with vehicle control (DMSO) or eganelisib (1 pM) for 2 days. Data are means ± S.D. from 3 technical replicates.

[0034] Figure 3i shows AUC of bone marrow normal CD34+ cells and AML patient samples after ex vivo treatment with increasing doses of eganelisib (0-4 pM) for 72 hours. Normal, n=l; PIK3R5 / PIK3CG10Wp-PAKr , n=9; PIK3R5 / PIK3CGhlghp-PAKr , n=5; PIK3R5 / PIK3CGhlghp- PAK1+, n=9. Data are means ± S.E.M. with Mann-Whitney test.

[0035] Figure 4a shows tumor volume and representative images of BPDCN intradermal xenografts following treatment with control, cytarabine, eganelisib and cytarabine+eganelisib. Data are means ± S.E.M. from 5 biological replicates. Two-way ANOVA test.

[0036] Figure 4b shows quantitative analysis of p-PAKl and PAK1 expression levels in BPDCN intradermal xenografts following treatment with control, cytarabine, eganelisib and cytarabine+eganelisib. Data are means ± S.E.M. from 5 biological replicates. Figure 4c shows survival of mice xenografted with leukemia PDXs and treated with control, eganelisib, cytarabine, or both (combo). Log-rank test. PIK3R5-high: BP08, AML11, and BA92; PIK3R5-low: BP01, AML20, and BA106.

[0037] Figure 4d shows western blotting for the indicated proteins in untreated or cytarabine- persistent PDX cells (BP01 : 320 pM; AML20: 5 pM; BA106: 10 pM; each for 72 hours).

[0038] Figure 4e shows GSEA indicating G protein-coupled purinergic receptor signalling pathways are significantly elevated in cytarabine-persistent AML20 cells demonstrating that G protein-coupled purinergic receptor signalling, but not other GPCR or PI3K-AKT-mTOR signalling, was activated in the cytarabine-persistent population.

[0039] Figure 4f shows a proposed model illustrating leukemia dependency on noncanonical PI3Ky signalling.

[0040] Figure 5a shows western blotting for PIK3R5 and PIK3CG in various leukemia cell lines after depletion of PIK3R5 or PIK3CG.

[0041] Figure 5b shows relative cell numbers of various leukemia cell lines after 2, 4, and 6 days of treatment with eganelisib, duvelisib, or LY294002 (each at 1 pM) as compared to control cells. Strongly sensitive cell lines are CAL-1, DF-DCN-86, and DF-DCN-1; moderately sensitive cell lines are P31-FUJ, and THP-1; and insensitive cell lines are DF-DCN-2, U937, and MV-4-11. Data are means ± S.E.M. from 3 biological replicates.

[0042] Figure 6a shows RT-qPCR of PIK3R5 mRNA in leukemia PDXs. Data are means ± S.D. from 3 technical replicates.

[0043] Figure 6b shows inhibition curve of leukemia PDXs after ex vivo treatment with increasing doses of eganelisib (0-4 pM) for 72 hours. Data are means ± S.D. from 3 technical replicates.

[0044] Figure 6c shows relative caspase-3 / 7 activities in leukemia PDXs after ex vivo treatment with eganelisib (1 pM) for 72 hours as compared to vehicle control (DMSO). Data are means ± S.E.M. with MannWhitney test.

[0045] Figure 6d shows flow cytometry of CDl lb and CD14 in leukemia PDXs after ex vivo treatment with eganelisib (1 pM) or vehicle control (DMSO) for 72 hours. PI3Ky inhibition induced markers of terminal myeloid differentiation (CDl lb and CD14) in AMLs with high PIK3R5 (AML11 and AML45) but not low PIK3R5 (AML 42 and AML20). Figure 6e shows a heatmap showing expression scores of the 32 defined genes in the IIRS for PIK3R5-high AMLs and PIK3R5-low AMLs. The IIRS expression score was determined by first transforming the RNA expression values for each gene across all the leukemias to a distribution of mean of 0 and SD of 1, thus each gene is assigned a distribution value in each leukemia sample, then summing all the distribution values for the 32 genes in each leukemia to get the final IIRS expression score for each leukemia sample.

[0046] Figure 6f shows the primary data for the heatmap of Figure 6e.

[0047] Figure 6g shows correlation analysis of PIK3R5 and IIRS expression scores in leukemia PDX samples. Spearman correlation coefficients are shown.

[0048] Figure 6h shows overall survival of AML patients from TCGA. Log-rank test.

[0049] Figure 6i shows proportions of FAB subtypes of AML patients from TCGA.

[0050] Figure 7a shows Seahorse assays measuring glycolysis in the indicated leukemia cell lines upon treatment with eganelisib or LY294002 (each at 1 pM for 48 hours). Data are means ± S.D. from 5 technical replicates.

[0051] Figure 7b shows Seahorse assays measuring glycolysis in control, PIK3R5 depleted, or PIK3CG depleted CAL-1 cells. Data are means ± S.D. from 5 technical replicates.

[0052] Figure 7c shows RT-qPCR of gene expression in CAL-1 cells upon treatment with eganelisib or LY294002 (each at 1 pM for 48 hours). Data are means ± S.D. from 3 technical replicates.

[0053] Figure 7d shows RT-qPCR of gene expression in CAL-1 cells expressing nontargeting control or gene-targeting CRISPRi guide RNAs. Data are means ± S.D. from 3 technical replicates.

[0054] Figure 7e shows relative numbers of CAL-1 cells expressing nontargeting control or genetargeting CRISPRi guide RNAs. Data are means ± S.E.M. from 3 biological replicates.

[0055] Figure 7f shows western blotting for p65 phosphorylation in CAL-1 cells treated with eganelisib (1 pM), together with or without NFkB inhibitors IKK-16 (0.5 pM) or SC-514 (20 pM), for 2 days. G, Relative numbers of CAL-1 cells after treatment with eganelisib (1 pM), with or without NFkB inhibitors IKK-16 (0.5 pM) or SC-514 (20 pM), for 2, 4, and 6 days. Data are means ± S.E.M. from 3 biological replicates.

[0056] Figure 8a shows GSEA indicating that the PI3K-AKT-mTOR signalling pathway downstream genes were not significantly affected upon eganelisib treatment (1 pM for 48 hours) or PIK3R5 depletion (PIK3R5i) in CAL-1 cells. Figure 8b shows a Volcano plot showing the upregulated or downregulated proteins by mass spectrometry in CAL-1 cells treated with eganelisib or depleted of PIK3R5 (PIK3R5i). PIK3R5 (two dots represent different PIK3R5 Uniprot protein IDs Q8WYR1 and J3KSW1, respectively) occurred as the most significantly downregulated protein, confirming the knockdown efficiency of PIK3R5. PIK3CG protein level was also moderately decreased due to its reduced stability after PIK3R5 depletion. Adjusted ? values are shown.

[0057] Figure 8c shows western blotting for Flag and PAK1 in CAL-1 cells expressing empty vector (EV) or Flag -tagged PAK1 mutants.

[0058] Figure 8d Shows inhibition curve of bone marrow normal CD34+ cells and AML patient samples after ex vivo treatment with increasing doses of eganelisib (0-4 pM) for 72 hours. Data are means ± S.D.; n = 3. Representative results from three biological replicates are shown. Correspond to panel e: PIK3R5 / PIK3CGlowp-PAKl , samples #2, #6, #8, #10, #13, #15, #17, #20, and #23; PIK3R5 / PIK3CGhlghp-PAKl- , samples #4, #7, #11, #16, and #18; PIK3R5 / PIK3CGhlghp- PAK1+, samples #1, #3, #5, #9, #12, #14, #19, #21, and #22.

[0059] Figure 8e Shows western blotting for the indicated proteins in bone marrow normal CD34+ cells and AML patient samples.

[0060] Figure 9a shows heatmaps showing inhibition percentages and synergy scores of combined eganelisib and cytarabine treatment in CAL-1 and DF-DCN-86 cells.

[0061] Figure 9b shows the primary data for the heatmaps of Figure 9a.

[0062] Figure 9c shows western blotting for PAK1 phosphorylation in DF-DCN-86 intradermal xenografts. The standard sample was created by mixing an equal amount from each of the 5 control samples.

[0063] Figure 9d shows GSEA indicating that neither the PI3K-AKT-mTOR signalling pathway nor other non-purinergic G protein-coupled receptor signalling pathways were significantly elevated in cytarabine-persistent AML20 cells.

[0064] DETAILED DESCRIPTION

[0065] In a first aspect, there is provided a method of treating a blood cancer comprising the administration of therapeutically effective amounts of a PI3Ky inhibitor and an antimetabolite to a patient in need thereof. Further embodiments of this first aspect include: a PI3Ky inhibitor and an antimetabolite for use in the treatment of a blood cancer; and the use of a PI3Ky inhibitor and an antimetabolite in the manufacture of a medicament (or medicaments) for the treatment of a blood cancer. In a further embodiment of the first aspect, there is provided a method of treating a blood cancer comprising the administration of therapeutically effective amounts of a PI3Ky inhibitor, an antimetabolite, and a BCL-2 antagonist to a patient in need thereof. In another embodiment, this includes: a PI3Ky inhibitor, an antimetabolite, and a BCL-2 antagonist for use in the treatment of a blood cancer; and the use of a PI3Ky inhibitor, an antimetabolite, and a BCL- 2 antagonist in the manufacture of a medicament or medicaments for the treatment of a blood cancer.

[0066] The term “blood cancer” is used herein to describe any type of cancer (i.e. uncontrolled, abnormal growth of cells) that affects blood cells. This definition encapsulates the five subtypes of blood cancer: leukemia, lymphoma, myeloma, myelodysplastic syndrome (MDS), and myeloproliferative disorder (MPD).

[0067] “Leukemia” refers to a type of blood cancer that affects blood cells in bone marrow - usually white blood cells. Examples of leukemia include: acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), acute promyelocytic leukemia (APL), chronic lymphocytic leukemia (CLL), chronic myeloid leukemia (CML), childhood leukemia and blastic plasmacytoid dendritic cell neoplasm (BPDCN) (although the last of these has features different than traditional leukemias).

[0068] "Lymphoma” refers to a type of blood cancer that affects the immune system - specifically white blood cells called lymphocytes. Examples of lymphoma include: diffuse large B-cell lymphoma, follicular lymphoma, primary mediastinal B cell lymphoma, mantle cell lymphoma, post-transplant lymphoproliferative disorder, marginal zone lymphoma, Waldenstrom's macroglobulinemia, and Burkitt lymphoma.

[0069] “Myeloma” refers to a type of blood cancer that affects plasma cells. Examples of myeloma include: light chain myeloma, non-secretory myeloma, solitary plasmacytoma, extramedullary plasmacytoma, monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM), immunoglobulin D (IgD) myeloma, and immunoglobulin E (IgE) myeloma. “Myelodysplastic syndrome (MDS)” refers to a type of blood cancer characterized by the bone marrow producing faulty blood cells and insufficient healthy blood cells. Examples of MDS include: MDS with single lineage dysplasia (MDS-SLD), MDS with ring sideroblasts (MDS-RS), MDS with multilineage dysplasia (MDS-MLD), and MDS with excess blasts (MDS-EB).

[0070] “Myeloproliferative disorder (MPD)” refers to a type of blood cancer whereby the bone marrow produces too many of a particular type of blood cell. Examples of MPD include: polycythaemia vera (PV), essential thrombocythaemia (ET), and myelofibrosis (MF).

[0071] As used herein, “treating” or “treatment” describes the management and care of a subject for the purpose of combating a disease, condition, or disorder. Treating includes preventing the onset of symptoms or complications, alleviating or eliminating symptoms or complications, or eliminating the underlying disease, condition, or disorder. For example, treating blood cancer in a subject includes reducing, repressing, delaying or preventing the growth of cancerous blood cells as well as killing cancerous blood cells within the subject.

[0072] As used herein, the terms “administering” and “administration” refer to any method of providing the active substance(s) to a subject. Where the method involves administration of two or more active substances, it results in them exerting their desired pharmacodynamic effects at the same time within the patient's body. Thus, administration of the two or more active substances is not limited to simultaneous administration, nor administration via the same route. It encompasses separate, sequential and simultaneous administration via the same or different routes and in the same or different medicaments, provided that it results in them exerting their desired pharmacodynamic effects at the same time within the patient’s body. Methods of administration are well known to those skilled in the art and include, but are not limited to, oral administration, transdermal administration, administration by inhalation, nasal administration, topical administration, intraaural administration, rectal administration, sublingual administration, buccal administration, and parenteral administration, including injectable administration such as intravenous administration, intra-arterial administration, intramuscular administration, intradermal administration, intrathecal administration, and subcutaneous administration. Administration can be continuous or intermittent. While the above means of administration may provide systemic exposure to the active substance(s), local administration / exposure is also contemplated, e.g., to the bone marrow of cells. A “PI3Ky inhibitor” is a substance that interacts with the PI3Ky protein (for example by binding to the catalytic domain thereof) and thereby inhibits the activity of the PI3Ky protein (i.e. the phosphorylation of phosphatidylinositol (4,5)-bisphosphate (PI(4,5)P2) to phosphatidylinositol (3,4,5)-trisphosphate (PI(3,4,5)P3). Suitable PI3Ky inhibitors are known in the art or may be identified by their ability to inhibit the activity of the PI3Ky protein in a suitable inhibition assay. Suitable examples of known PI3Ky inhibitors are apitolisib, buparlisib, dactolisib, duvelisib (Copiktra), eganelisib, gedatolisib, GSK 1059615, omipalisib, paxalisib, samotolisib, tenalisib, taselisib, voxtalisib, wortmannin, AS252424, AS605240, AZD3458, ZX-lOla and ZX- 4081.

[0073] An “antimetabolite” is an active substance that inhibits DNA replication, and thereby reduces, represses, delays or prevents cell growth and cell division. The three main subtypes of antimetabolites are: (1) purine antagonists - which prevent cancer cells from making purines; (2) pyrimidine antagonists - which prevent cancer cells from making pyrimidines; and (3) folic acid antagonists (antifolates) - which prevent cancer cells from using folic acid to make DNA and RNA. Suitable examples of antimetabolites include: 5-FU [fluorouracil] (Carac®, Tolak®, Efudex®, Fluoroplex®), azacitidine (Onureg™, Vidaza®), capecitabine (Xeloda®), cladribine (Mavenclad®, Leustatin™), clofarabine (Clolar®), cytarabine (Cytosar-U®, DEPOCYT®), decitabine (Dacogen®), floxuridine (FUDR®), fludarabine phosphate (Fludara®, Oforta™), gemcitabine (Gemzar®, Infugem™), hydroxyurea (Droxia®, Hydrea®), methotrexate (Otrexup®, Rasuvo®, Trexall®), nelarabine (Arranon®), pemetrexed (Alimta®, Pemfexy™), pentostatin (Nipent™), pralatrexate (Folotyn®), thioguanine (Tabloid®), and trifluridine / tipiracil (Lonsurf®).

[0074] A “BCL-2 inhibitor” is a substance that interacts with the BCL-2 protein (for example by binding to it) and thereby inhibits the activity of the BCL-2 protein (i.e. regulation of apoptosis). Suitable BCL-2 inhibitors are known in the art or may be identified by their ability to inhibit the activity of the BCL-2 protein in a suitable inhibition assay. Suitable examples of known BCL-2 inhibitors include: venetoclax (ABT-199), S55746 (BCL201), lisaftoclax (APG-2575), oblimersen (G3139), AZD4320, AZD0466, pelcitoclax (APG-1252), BM-1197, S44563, ABT-737, navitoclax (ABT-263), obatoclax (GX15-070) and AT-101.

[0075] The term “patient” refers to a mammalian, preferably human, subject. In a preferred embodiment, the blood cancer is selected from the group consisting of leukemia, lymphoma, myeloma, MDS and MPN. More preferably, the blood cancer is leukemia. Even more preferably, the blood cancer is selected from the group consisting of ALL, AML, APL, CLL, CML, childhood leukemia, and BPDCN. Most preferably, the blood cancer is selected from the group consisting of ALL, AML, and BPDCN. In a further preferred embodiment, the blood cancer is resistant to cytarabine monotherapy.

[0076] In a further preferred embodiment, the PI3Ky inhibitor inhibits the kinase activity of the PI3Ky protein with an IC50 of 100 nM or less, more preferably 20 nM or less. A suitable method for determining the inhibition constant of a potential PI3Ky inhibitor is disclosed in ACS Med Chem Lett. 2016 Sep 8; 7(9): 862-867. In a further preferred embodiment, the PI3Ky inhibitor also has at least 10-fold selectivity for PI3Ky over PI3Koc; and / or at least 10-fold selectivity for PI3Ky over PI3K0; and / or at least 10-fold selectivity for PI3Ky over PI3K8. More preferably, the PI3Ky inhibitor also has at least 10-fold selectivity for PI3Ky over two of PI3Ka and PI3K0 and PI3K5. More preferably, the PI3Ky inhibitor also has at least 10-fold selectivity for PI3Ky over all three of PI3Ka and PI3K0 and PI3K8. A suitable method for determining the selectivity of a potential PI3Ky inhibitor is disclosed in ACS Med Chem Lett. 2016 Sep 8; 7(9): 862-867.

[0077] In a further preferred embodiment, the PI3Ky inhibitor is selected from the group consisting of apitolisib, buparlisib, dactolisib, duvelisib (Copiktra), eganelisib, edatolisib, GSK1059615, omipalisib, paxalisib, samotolisib, tenalisib, taselisib, voxtalisib, wortmannin, AS252424, AS605240, AZD3458, ZX-lOla and ZX-4081. More preferably, the PI3Ky inhibitor is selected from the group consisting of duvelisib and eganelisib. Most preferably, the PI3Ky inhibitor is eganelisib.

[0078] In a further preferred embodiment, the antimetabolite is selected from the group consisting of 5-FU [fluorouracil] (Carac®, Tolak®, Efudex®, Fluoroplex®), azacitidine (Onureg™, Vidaza®), capecitabine (Xeloda®), cladribine (Mavenclad®, Leustatin™), clofarabine (Clolar®), cytarabine (Cytosar-U®, DEPOCYT®), decitabine (Dacogen®), floxuridine (FUDR®), fludarabine phosphate (Fludara®, Oforta™), gemcitabine (Gemzar®, Infugem™), hydroxyurea (Droxia®, Hydrea®), methotrexate (Otrexup®, Rasuvo®, Trexall®), nelarabine (Arranon®), pemetrexed (Alimta®, Pemfexy™), pentostatin (Nipent™), pralatrexate (Folotyn®), thioguanine (Tabloid®), and trifluridine / tipiracil (Lonsurf®). More preferably, the antimetabolite is selected from the group consisting of fluorouracil, azacitidine, capecitabine gemcitabine, cytarabine, decitabine, and floxuridine. Most preferably, the antimetabolite is azacitidine or cytarabine, especially cytarabine.

[0079] In a particularly preferred embodiment of this first aspect, the blood cancer is selected from the group consisting of ALL, AML, and BPDCN, the PI3Ky inhibitor is eganelisib and the antimetabolite is azacitidine or cytarabine, especially cytarabine.

[0080] In a further preferred embodiment, the BCL-2 inhibitor (when present) inhibits the activity of the BCL-2 protein with an IC50 of 100 nM or less, more preferably 20 nM or less. A suitable method for determining the inhibition constant of a potential BCL-2 inhibitor is disclosed in Proc Natl Acad Sei USA. 2000 Jun 20; 97(13): 7124-7129 and suitable assay kits are commercially available (for example the BCL-2 TR-FRET Assay Kit from Bioscience. In a further preferred embodiment, the BCL-2 inhibitor also has at least 10-fold selectivity for BCL-2 over other antiapoptotic proteins in the BCL-2 family; in particular at least 10-fold selectivity for BCL-2 over BCL-XL; and / or at least 10-fold selectivity for BCL-2 over BCL-W; and / or at least 10-fold selectivity over MCL-1; and / or at least 10-fold selectivity over BFL-1. More preferably, the BCL- 2 inhibitor also has at least 10-fold selectivity for BCL-2 over two of BCL-XL and BCL-W and MCL-1 and BFL-1. More preferably, the BCL-2 inhibitor also has at least 10-fold selectivity for BCL-2 over three of BCL-XL and BCL-W and MCL-1 and BFL-1. More preferably, the BCL-2 inhibitor also has at least 10-fold selectivity for BCL-2 over all four of BCL-XL and BCL-W and MCL-1 and BFL-1.

[0081] In a preferred embodiment, the BCL-2 inhibitor is selected from the group consisting of venetoclax (ABT-199), S55746 (BCL201), lisaftoclax (APG-2575), oblimersen (G3139), AZD4320, AZD0466, pelcitoclax (APG-1252), BM-1197, S44563, ABT-737, navitoclax (ABT- 263), obatoclax (GX15-070) and AT-101. More preferably, the BCL-2 inhibitor is selected from the group consisting of venetoclax (ABT-199), S55746 (BCL201), lisaftoclax (APG-2575), and oblimersen (G3139). Most preferably, the BCL-2 inhibitor is venetoclax.

[0082] In a particularly preferred embodiment, the BCL-2 inhibitor is venetoclax, the PI3Ky inhibitor is eganelisib and the antimetabolite is azacitidine or cytarabine, especially azacitidine. In a second aspect, there is provided a method of identifying a blood cancer patient susceptible to treatment with a PI3Ky inhibitor comprising: (i) comparing the level s of a factor which is indicative of the expression of the gene PIK3R5 and the levels of a factor which is indicative of the expression of at least one gene selected from the group consisting of IRF7, LCP2, CCL5, CXCL9, PLSCR1, LYN, JAK2, IFNGR1, IRF8, CD86, TLR2, FAS, STAT1, TAPBP, B2M, CD74, TAPI, CASP1, LAPS, IL18R1, TNFRSF1B, SELL, IL10RA, TNFSF10, IFITM3, STAT2, IRF9, EIF2AK2, LY6E, BST2, and RTP4 in cancerous blood cells obtained from a blood cancer patient to the levels of those same factors in non-cancerous blood cells obtained from said patient, or to the levels of those same factors known to be present in non-cancerous blood cells of healthy individuals, or to the levels of those same factors known to be present in cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor; and (ii) selecting said patient for treatment if the levels of said factors in their cancerous blood cells are higher than the levels of said factors in their non-cancerous blood cells, or higher than the levels of said factors in the non-cancerous blood cells of healthy individuals, or higher than the levels of said factors in the cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor.

[0083] The terms “blood cancer”, “patient”, “treatment” and “PI3Ky inhibitor” have the same meaning in this second aspect as in the first aspect above.

[0084] Preferred identities for the “blood cancer”, “patient” and “PI3Ky inhibitor” are the same in this second aspect as in the first aspect above.

[0085] The phrase “factor which is indicative of the expression of [a gene]” as used herein includes an mRNA molecule which results from transcription of the gene’s DNA or a protein molecule which results from translation of the mRNA. Levels of mRNA can be quantified by methods known in the art, such as polymerase chain reaction (PCR) following reverse transcription, nucleic acid hybridization methods such as microarray, and RNA sequencing methods. Levels of protein can be quantified by methods known in the art, such as mass spectrometry or by antibody-based methods, such as immunohistochemistry, enzyme-linked immunosorbent assay (ELISA), Western blotting, flow cytometry, and immuno-electron microscopy.

[0086] In a further preferred embodiment the factor which is indicative of the expression of a gene is the mRNA molecule which results from transcription of the gene’s DNA. In further preferred embodiments, the method comprises: (i) comparing the levels of a factor which is indicative of the expression of the gene PIK3R5 and the levels of factors which are indicative of the expression of two or more, preferably four or more, more preferably eight or more, still more preferably sixteen or more, most preferably all of the genes selected from the group consisting of IRF 7, LCP2, CCL5, CXCL9, PLSCR1, LYN, JAK2, IFNGR1, IRF8, CD86, TLR2, FAS, STAT1, TAPBP, B2M, CD74, TAPI, CASP1, LAP 3, IL18R1, TNFRSF1B, SELL, IL10RA, TNFSF10, IFITM3, STAT2, IRF9, EIF2AK2, LY6E, BST2, and RTP4 in cancerous blood cells obtained from a blood cancer patient to the levels of those same factors in non-cancerous blood cells obtained from said patient, or to the levels of those same factors known to be present in non-cancerous blood cells of healthy individuals, or to the levels of those same factors known to be present in cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3I<7 inhibitor; and (ii) selecting said patient for treatment if the levels of said factors in their cancerous blood cells are all higher than the levels of said factors in their non-cancerous blood cells, or higher than the levels of said factors in the non-cancerous blood cells of healthy individuals, or higher than the levels of said factors in the cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor.

[0087] In a further preferred embodiment, the levels of said factors in the cancerous blood cells are considered to be higher than their levels in the non-cancerous blood cells when they are at least 110%, preferably at least 120%, more preferably at least 130%, even more preferably at least 150%, most preferably at least 200% of their levels in the non-cancerous blood cells.

[0088] In a third aspect, there is provided a method of treating a blood cancer comprising administration of a therapeutically effective amount of a PAK-1 inhibitor to a patient in need thereof. Further embodiments of this third aspect include: a PAK-1 inhibitor for use in the treatment of a blood cancer; and the use of a PAK-1 inhibitor in the manufacture of a medicament for the treatment of a blood cancer.

[0089] The terms “blood cancer”, “patient”, “administration” “treating” and “treatment” have the same meaning in this third aspect as in the first aspect above.

[0090] A “PAK-1 inhibitor” is a substance that interacts with the PAK-1 protein (for example by binding to the catalytic domain thereof) and thereby inhibits the activity of the PAK-1 protein (i.e. the phosphorylation of its downstream effector substrates). Suitable PAK-1 inhibitors are known in the art or may be identified by their ability to inhibit the activity of the PAK-1 protein in a suitable inhibition assay. Suitable examples of known PAK-1 inhibitors are FRAX597, FRAX1036, G-5555, FL172, PF-3758309, AZ13705339, IPA-3, and NVS-PAK1-1.

[0091] In a preferred embodiment, the blood cancer is selected from the group consisting of leukemia, lymphoma, myeloma, MDS, and MPN. More preferably, the blood cancer is leukemia. Even more preferably, the blood cancer is selected from the group consisting of ALL, AML, APL, CLL, CML, childhood leukemia, and BPDCN. Most preferably, the blood cancer is selected from the group consisting of ALL, AML, and BPDCN.

[0092] In a further embodiment, the PAK-1 inhibitor inhibits the activity of the PAK-1 protein with an IC50 of 100 nM or less, more preferably 50 nM or less. A suitable method for determining the inhibition constant of a potential PAK-1 inhibitor is disclosed in Chem Biol. 2008 Apr; 15(4): 322-331. More preferably, the PAK-1 inhibitor also has at least 2-fold selectivity for PAK-1 over PAK-2; and / or at least 2-fold selectivity for PAK-1 over PAK-3; and / or at least 2-fold selectivity for PAK-1 over PAK-4; and / or at least 2-fold selectivity for PAK-1 over PAK-6; and at least 2- fold selectivity for PAK-1 over PAK-5 / 7. More preferably, the PAK-1 inhibitor also has at least 2-fold selectivity for PAK-1 over two of PAK-2 and PAK3 and PAK-4 and PAK-6 and PAK-5 / 7. More preferably, the PAK-1 inhibitor also has at least 2-fold selectivity for PAK-1 over three of PAK-2 and PAK3 and PAK-4 and PAK-6 and PAK-5 / 7. More preferably, the PAK-1 inhibitor also has at least 2-fold selectivity for PAK- 1 over four of PAK-2 and PAK3 and PAK-4 and PAK- 6 and PAK-5 / 7. More preferably, the PAK-1 inhibitor also has at least 2-fold selectivity for PAK- 1 over all five of PAK-2 and PAK3 and PAK-4 and PAK-6 and PAK-5 / 7. A suitable method for determining the selectivity of a potential PAK-1 inhibitor is disclosed in Nat Biotechnol. 2005 Mar;23(3):329-3.

[0093] In a further embodiment, the PAK-1 inhibitor is selected from the group consisting of: FRAX597, FRAX1036, G-5555, FL172, PF-3758309, AZ13705339, IPA-3, and NVS-PAK1-1. Most preferably, the PAK-1 inhibitor is FRAX1036.

[0094] EXAMPLES

[0095] Materials and methods Cell cultures. 293T, U-937 and THP-1 cells were purchased from ATCC (Manassas, VA). P31-FUJ and MV-4-11 cells were from Dr. James Griffin (Dana-Farber Cancer Institute [DFCI], Boston, MA). CAL-1 cells were provided by Dr. Takahiro Maeda (Nagasaki University, Nagasaki, Japan). 293T cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) (F2442; Sigma-Aldrich, St. Louis, MO). U937, P31-FUJ and CAL-1 cells were cultured in RPMI 1640 supplemented with 10% FBS (RIO). THP-1 cells were cultured in RIO supplemented with 0.05 mM 2-mercaptoethanol (21985023; Thermo Fisher Scientific, Waltham, MA). MV-4-11 cells were cultured in IMDM supplemented with 10% FBS. Cell line identity was verified by short tandem repeat (STR) profiling in the Molecular Diagnostics Laboratory at DFCI, and cells were verified to be Mycoplasma-free by conventional methods.

[0096] TheBPDCN cell line DF-DCN-86 was established from a skin biopsy in aBPDCN patient. Briefly, the skin biopsy sample was cut into small pieces with a scalpel then digested with type II collagenase (17101015; Thermo Fisher Scientific) at 37 °C for 30 min. After filtering with a 70 pm cell strainer (352350; Corning, Glendale, AZ), the cells were washed once with 1X RBC lysis buffer (420301; BioLegend, San Diego, CA) then cultured with advanced DMEM / F-12 (12634028; Thermo Fisher Scientific) based leukemia adaptive medium (LAM) for 2 weeks: IX Penicillin- Streptomycin (15140122; Thermo Fisher Scientific), 100 pg / ml Primocin (ant-pm-1; InvivoGen, San Diego, CA), 0.5 pg / ml Caspofungin (SML0425; Sigma-Aldrich), IX HEPES (15630080; Thermo Fisher Scientific), IX Glutamax (35050061; Thermo Fisher Scientific), 10 pM Y27632 (130-106-538; Miltenyi Biotec, Bergisch Gladbach, North Rhine-Westphalia, Germany), 100 ng / ml R-spondin-1 (120-38; PeproTech, Cranbury, NJ), 100 ng / ml Noggin (120- 10C; PeproTech), IX B-27 (17504044; Thermo Fisher Scientific), 500 nM A 83-01 (SML0788; Sigma- Aldrich), 1 pM Prostaglandin E2 (2296; Tocris Bioscience, Bristol, UK), 1.25 mM N- Acetyl-L-cysteine (A8199; Sigma-Aldrich), 10 mM Nicotinamide (N0636; Sigma-Aldrich), 3 pM CHIR99021 (SML1046; Sigma-Aldrich), 10 pM SB 202190 (S7076; Sigma-Aldrich), 50 ng / ml EGF (AF-100-15; PeproTech), 10 ng / ml FGF-10 (100-26; PeproTech), 10 ng / ml FGF-basic (100- 18B; PeproTech), 50 ng / ml Flt3-Ligand (300-19; PeproTech), 10 ng / ml SCF (300-07; PeproTech), 10 ng / ml IL-3 (200-03; PeproTech), and 1 pM StemRegenin 1 (S2858; Selleckchem, Houston, TX). Then, the LAM was replaced with R10 by 10% increasing each week for 9 weeks until the culture contained 10% LAM and 90% R10, after which the percentage of LAM was further reduced to 5%, 2%, 1% and 0% each week. The DF-DCN-1 and DF-DCN-2 cell lines were established by a similar method from BPDCN primary bone marrow samples that were previously grown in vitro in coculture with MS-5 stromal cells. Those cells were then cultured with the mix of LAM and RIO (supplemented with Flt3-Ligand, SCF, IL3 and SR-1 as the same concentrations as in LAM) at a ratio of 1 : 1, and the percentage of LAM was further reduced to 40%, 30%, 20%, 10%, 5%, 2%, 1% and 0% each week. All the 3 cell lines were cryopreserved in FBS + 10% DMSO, and 10% LAM was supplemented in the culture medium during the first 2 passages after each freeze-thaw cycle or after lentivirus infection.

[0097] Reagents. Eganelisib (S8330), duvelisib (S7028), LY294002 (SI 105), IKK-16 (S2882), SC-514 (S4907), FRAX597 (S7271), FRAX1036 (S7989), MK-2206 (S1078), and cytarabine (SI 648) were purchased from Selleckchem for in vitro and ex vivo experiments. Eganelisib (HY- 100716) and cytarabine (HY-13605) were purchased from MedChemExpress (Monmouth Junction, NJ) for in vivo experiments.

[0098] Plasmids. Guide RNA sequences for CRISPRi were cloned into pXPR_050 vector (#96925; Addgene, Watertown, MA). shRNA sequences were cloned into pSIHl-Hl-puro vector (#26597; Addgene). cDNA sequences of PAK1 mutants with a C-terminus Flag tag were synthesized at BGI Genomics (Beijing, China) then cloned into pLVX-IRES-Neo vector (632184; Takara Bio, Kusatsu, Shiga, Japan).

[0099] Analyses of TCGA AML. Analyses of TCGA AML. All patients in the TCGA AML dataset with both RNA-seq and prognosis data available were included for overall survival analyses (n=l 61). The overall survival was calculated from the date of first treatment until death or last follow-up and comparisons were made by the Log-rank test. Sixteen patients with PML- RARA rearrangement were excluded from the dataset before regression analyses as those patients are clearly distinguished (belonging to French-American-British M3 pathologic subtype) from other AML patients and have a long-term survival rate up to 90%. Categorical variables were summarized as counts and percentages, and comparisons were made by Pearson’s chi- square test. Continuous variables were summarized as median and range, and comparisons were made by the Wilcoxon rank sum test. Cox proportional hazards regression model was fit in a stepwise backward selection to assess the effect of the HRS score adjusted for age and disease characteristics on OS. Log-rank test was performed by GraphPad Prism 9 (GraphPad Software, San Diego, CA), and other analyses were performed by Stata 18.0. (StataCorp, College Station, TX). Virus production and transduction. Lenti virus was produced by transfection of 293 T cells with the second-generation packaging system psPAX2 (#12260, Addgene) and pMD2.G (#12259, Addgene). 293T cells were seeded at a confluence of 95% and culture medium was replaced with fresh DMEM without serum or antibiotics before transfection. Transfections were performed using Lipofectamine 2000 Transfection Reagent (11668500, Thermo Fisher Scientific) according to the manufacture’s protocol. Six hours after transfection, the medium was replaced with regular DMEM culture medium that contains 10% FBS and IX Penicillin-Streptomycin. Viral supernatants were collected 48 and 72 hours after transfection and filtered through a 0.45 pm filter.

[0100] Lentivirus transduction was performed by spin infection. Briefly, 2-5 * 105cells were resuspended in 500 pl culture medium and seeded into a well of 12-well plate, then 1 ml lentivirus was added and Polybrene Infection / Transfection Reagent (TR-1003-G; Sigma-Aldrich) was supplemented at a final concentration of 8 pg / ml. After spinning at 930 g, 30 °C for 2 hours with zero brake, another 1 ml regular culture medium was slowly added to each well without disturbing the cells. One day after the spin infection, virus-containing medium was replaced with fresh culture medium and cells were allowed to recover for another day, then cells were selected with 2 pg / ml puromycin (631306; Takara Bio) or 500 pg / ml geneticin (10131027; Thermo Fisher Scientific) for 3-7 days before following experiments.

[0101] CRISPR interference screening and analysis. The CRISPR interference screen was performed in the BPDCN cell line CAL-1, with two biological replicates. KRAB-dCas9 (#96918; Addgene) was transduced by lentivirus and successfully transduced cells were selected by 6 pg / ml blasticidin (R21001; Thermo Fisher Scientific) for 1-2 weeks. The expression of KRAB-dCas9 was validated by western blotting and a similar growth rate was observed between KRAB-dCas9- expressing and parental CAL-1 cells. The virus for the genome-wide Dolcetto CRISPR interference library (Sanson, K. R. et al. Optimized libraries for CRISPR-Cas9 genetic screens with multiple modalities. Nat Commun 9, 5416 (2018)) was obtained from the Genetic Perturbation Platform of Broad Institute (Cambridge, MA), which includes Set A and Set B that contain 57,050 and 57,011 sgRNAs to target 18,901 and 18,899 genes, respectively. KRAB- dCas9-expressing CAL-1 cells were transduced with the Dolcetto library at a multiplicity of infection of 0.3 by spin infection. Two days after transduction, cells were selected with 2 pg / ml puromycin for another 6 days, then samples were collected as day 0 of the dependency screen. The remaining cells were passaged every 2-3 days for collection of day 14 and day 21 samples. The sgRNA coverage was >l,000x throughout the screen and sample collections. Genomic DNA was extracted using a NucleoSpin Blood XL kit (740950.50; Takara Bio) and further cleaned up by a OneStep PCR Inhibitor Removal Kit (D6030; Zymo Research, Irvine, CA) before PCR amplification. Sequencing and data analysis were performed by the Genetic Perturbation Platform of Broad Institute.

[0102] Short-term cell proliferation assay. Cells were washed once with serum-free medium then cultured with the corresponding base medium supplemented with 0.5-5% FBS. The starting cell density (day 0) was 1 x 105 / ml and the cell numbers were counted by a Countess II Automatic Cell Counter (AMQAX1000; Thermo Fisher Scientific) every 2 days till day 6. Relative cell number (%) was calculated by the ratio of experimental arm (gene silencing or inhibitor treated) to control arm (nontargeting control or DMSO treated) at each time point.

[0103] Ex vivo and in vitro drug sensitivity assays. For ex vivo drug sensitivity assays using PDXs, PDX cells were collected from the bone marrow and spleen of NOD.Cg- PrkdcscldIl2rmIW}lI zJ (NSG; 005557; The Jackson Laboratory, Bar Harbor, ME) mice, then purified with a Mouse Cell Depletion Kit (130-104-694; Miltenyi Biotec) and / or Dead Cell Removal Kit (130-090-101; Miltenyi Biotec) to ensure the human cell percentage and live cell percentage were both greater than 90%. PDX cells were seeded into white 96-well plates (3917; Corning) with 4 x 104cells per well in 100 pl LAM, then treated for 72 hours. Cell viability was determined by a CellTiter-Glo 2.0 Cell Viability Assay (G9242; Promega, Madison, WI) according to the manufacture’s protocol. In vitro drug sensitivity assays using leukemia cell lines were performed similarly except for adjusting the cell number per well to 0.5-4 x 104according to the growth rates of different cell lines. Synergy score was calculated using the SynergyFinder online tool ().

[0104] Caspase-3 / 7 activity measurement. Caspase-3 / 7 activity in leukemia cells were detected by Caspase-Gio 3 / 7 Assay System (G8091; Promega) according to the manufacturer’s protocol. Briefly, PDX cells were seeded into white 96-well plates (3917; Corning) with 4 x 104cells per well in 100 pl LAM, then treated for 72 hours. Then, 100 pl premixed Caspase-Gio 3 / 7 reagent was added into each detection well. After mixing the contents on a plate shaker for 2 minutes, the plate was incubated at room temperature for 1 hour before being measured by a SpectraMax M3 Multi-Mode Microplate Reader. Intradermal xenografting model. The BPDCN cell line DF-DCN-86 established from a patient skin tumor was used for an intradermal xenografting model to mimic the clinical characteristics of BPDCN patients. Briefly, 1.25-10>< 106DF-DCN-86 cells were suspended in 50 pl cold PBS, then mixed with 50 pl Cultrex UltiMatrix Reduced Growth Factor Basement Membrane Extract (BME001-05; R&D Systems, Minneapolis, MN) on ice. Hair of NSG mice around the xenografting area was shaved with an electric clipper, and the 100 pl cell mixture was injected intradermally using a 28-gauge insulin syringe. Tumor size was measured by a caliper and tumor volume was calculated using the following formula: 0.52 x length x width2. After the tumors had grown for the designated time, mice were euthanized and the tumors were harvested for protein extraction and western blotting.

[0105] Advanced stage PDX model. To evaluate the effects of eganelisb alone or in combination with cytarabine on overall survival, 1 x 106leukemia PDX cells were suspended in PBS containing 0.2% FBS then injected into a NSG mouse via tail vein. Treatment was started 3 weeks post- xenografting when the disease was in an advanced stage (with a peripheral blood leukemia burden approximately 1-5%). Eganelisib was given at 15 mg / kg daily for 14 days via oral gavage (dissolved in 10% DMSO, 40% PEG300, 5% Tween-80, and 45% saline sequentially), and cytarabine was given at 30 mg / kg daily for 5 days (day 3-7 of eganelisib treatment period) via intraperitoneal injection (dissolved in saline). The date when the first mouse in control group reached the endpoint was recorded as the start day for the assessment of prolonged survival. Mice were humanely sacrificed once moribund.

[0106] Residual disease leukemia model. To obtain a persistent leukemia population after cytarabine treatment, 1xio6acute leukemia PDX cells were suspended in PBS containing 0.2% FBS then injected into a NSG mouse via tail vein, and treatment was started 3 weeks post- xenografting. Cytarabine was given at 30 mg / kg daily for 5 days via intraperitoneal injection. One day after the last dose of treatment, mice were sacrificed and cells were collected from bone marrow. Human leukemia cells were labelled with a FITC anti-human CD45 antibody (368508; BioLegend) then sorted using a BD FACS Aria II Cell Sorter.

[0107] Genetic characterization of leukemia models. BPDCN cell lines and PDXs were characterized using OncoPanel, a custom capture-based next-generation sequencing based assay for detection of single-nucleotide variants, insertions / deletions, copy number alterations, and structural variants across 282 cancer genes (Garcia, E. P. et al. Validation of OncoPanel: A Targeted Next-Generation Sequencing Assay for the Detection of Somatic Variants in Cancer. Arch Pathol Lab Med 141, 751-758 (2017)) and / or a PCR amplification-based hematologic malignancy gene panel (Kluk, M. J. et al. Validation and Implementation of a Custom Next- Generation Sequencing Clinical Assay for Hematologic Malignancies. J Mol Diagn 18, 507-515 (2016)).

[0108] Reverse transcription quantitative real-time PCR (RT-qPCR). Total RNA was extracted using NucleoSpin RNA Plus XS (740990.250; Takara Bio) according to the manufacturer’s protocol. First strand cDNA synthesis was performed with LunaScript RT SuperMix Kit (E3010L; New England Biolabs, Ipswich, MA), and 100 ng-1 pg total RNA was used for one 20 pl RT reaction. qPCR analysis was conducted on a QuantStudio 6 Flex Real- Time PCR System (4485697; Thermo Fisher Scientific) with Luna Universal qPCR Master Mix (M3003X; New England Biolabs). The qPCR primer sequences were as follows (5’-3’); GAPDH (Forward-TCGGAGTCAACGGATTTG; Reverse-CAACAATATCCACTTTACCAGAG), PIK3R5 (Forward-TGACATGCTACTCTACTACTG; Reverse- GGAGTGGATGAAGATCTCTG), PIK3CG (Forward-TCAGGACATCTGTGTTAAGG; Reverse-GCATCCCGGATATATTCAATG), SPI1 (Forward-AGCCATAGCGACCATTAC; Reverse-CTCCGTGAAGTTGTTCTC), ASH2L (Forward-CTTTTGGATCAGGACCTTAG; Reverse-CAGAAAACAAAGGGTCACTC), RHOB (Forward- CAAGGAGAGGGAAAAGAAAC; Reverse-ACTGCCCTTTATCAAAACTG), EGR1 (Forward-CAAAATAAGGAAGAGGGCTG; Reverse-CTACAACATTCCAACTCCTG), CD69 (Forward-CTACTCTTGCTGTCATTGATTC; Reverse-GTTCCTTTTTCAGTCCAACC), SGK1 (Forward-GAGAAGCATATTATGTCGGAC; Reverse- TCTGGAGATGGTAGAACAAC), and EGR3 (Forward-CAGAGAATGTAATGGACATCG; Reverse-CATGAGGCTAATGATGTTGTC).

[0109] Western blotting. Cells were lysed with RIPA Lysis and Extraction Buffer (89901; Thermo Fisher Scientific) for 10 minutes on ice. After brief sonication, cell lysate was centrifuged for 10 minutes at 20,000 g, 4 °C, to remove insoluble debris. Protein concentrations were determined by Pierce BCA Protein Assay Kit (23227; Thermo Fisher Scientific) and samples were prepared by boiling at 99 °C for 10 minutes with Blue Loading Buffer (7722S; Cell Signalling Technology, Danvers, MA). Then, 5-20 pg of total protein was separated by NuPAGE 4-12% BisTris Mini Protein Gels (NP0336BOX; Thermo Fisher Scientific) and transferred to Immobilon-P PVDF Membrane (IPVH00010; MilliporeSigma, Burlington, MA). Following transfer, the membrane was blocked with 5% non-fat milk (NC9022655; Fisher Scientific, Waltham, MA) or bovine serum albumin (9998S; Cell Signalling Technology) for 45 minutes at room temperature. Primary antibodies were incubated at 4 °C overnight and secondary antibodies were incubated at room temperature for 90 minutes. Proteins were detected by Clarity Western ECL Substrate (1705061; Bio-Rad, Hercules, CA) and images were taken by a ImageQuant LAS 4000 or a ImageQuant 800 (Cytiva, Marlborough, MA). The following primary antibodies were used for Western Blotting: anti-PIK3R5 (1 :500; sc-390916; Santa Cruz Biotechnology), anti-PIK3CG (1 :500; sc-166365; Santa Cruz Biotechnology), anti-ubiquitin (1 : 1000; 43124S; Cell Signalling Technology), anti-SPIl (1 :1000; 2258S; Cell Signalling Technology), anti-ASH2L (1 :2000; abl76334; Abeam, Cambridge, UK), anti-p-PAKl-S144 (1: 1000; 2606S; Cell Signalling Technology), anti-p-PAKl-T423 (1 : 1000; ab2477; Abeam), anti-PAKl (1 : 1000; 2602S; Cell Signalling Technology), anti-p-AKT-S473 (1 :2000; 4060S; Cell Signalling Technology), anti- AKT (1: 1000; 4691S; Cell Signalling Technology), anti-p-PDKl-S241 (1: 1000; 3061S; Cell Signalling Technology), anti-PDKl (1 : 1000; 3062S; Cell Signalling Technology), and anti-P-actin (1 :5000; A5441; Sigma- Aldrich). Horse anti-mouse IgG (7076S; Cell Signalling Technology) and goat anti-rabbit IgG (7074S; Cell Signalling Technology) secondary antibodies were used at a dilution of 1 :2000.

[0110] Flow cytometry. Leukemia cells were resuspended in flow buffer (PBS containing 2% FBS) and cell surface proteins were stained with the following antibodies (5 pl per test in 100 pl flow buffer) at 4 °C for 20 minutes: FITC anti-human CD45 (368508; BioLegend), FITC antihuman CD14 (325604; BioLegend), and APC / Cyanine7 anti-mouse / human CDl lb (101226; BioLegend). After staining, cells were washed with 1 ml flow buffer once then resuspended in 200 pl flow buffer and proceed to detecting using a CytoFLEX cytometer (Beckman Coulter, Jersey City, NJ).

[0111] RNA sequencing (RNA-seq). RNA-seq was performed by the Molecular Biology Core Facilities at DFCI (http: / / mbcf.dfci.harvard.edu / ). The raw FASTQ data were analyzed by the VIPER pipeline to generate gene expression data, as previously described (Togami, K. et al. Sex- Biased ZRSR2 Mutations in Myeloid Malignancies Impair Plasmacytoid Dendritic Cell Activation and Apoptosis. Cancer Discov 12, 522-541 (2022)). Gene set enrichment analysis (GSEA) was performed with the GSEA software according to published guidelines (Mootha, V. K. et al. PGC- 1 alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 34, 267-273 (2003) and Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Set USA 102, 15545-15550 (2005)).

[0112] Quantitative Tandem Mass Tag (TMT) proteomics and phosphoproteomics. The streamlined (SL)-TMT method was used to quantify proteins and phosphopeptides, the streamlined (SL)-TMT method was followed (Navarrete-Perea, J., Yu, Q., Gygi, S. P. & Paulo, J. A. Streamlined Tandem Mass Tag (SL-TMT) Protocol: An Efficient Strategy for Quantitative (Phospho)proteome Profiling Using Tandem Mass Tag- Synchronous Precursor Selection-MS3. J Proteome Res 17, 2226-2236 (2018)). CAL-1 cells were lysed in an 8M urea buffer (8M Urea, 75 mM NaCl, 50 mM HEPES pH 7.5) supplemented with EDTA-free protease and phosphatase inhibitors (Roche). The lysates were then clarified by centrifugation at 17,000 x g for 15 min at 4°C and quantified. To reduce and alkylate cysteines, 200 pg of protein was sequentially incubated with 5 mM TCEP for 30 mins, 14 mM iodoacetamide for 30 mins (in the dark), and 10 mM DTT for 15 mins. All reactions were carried out at RT. Next, proteins were chloroform-methanol precipitated and the pellet resuspended in 200 mM EPPS pH 8.5. Then, the protease LysC (Wako) was added at 1 : 100 (LysC: protein) ratio and incubated overnight at RT. The following day, samples were further digested for 5 hours at 37°C with trypsin at 1 :75 (trypsin:protein) ratio. Both digestions were performed using an orbital shaker at 1,500 rpm. After digestion, samples were clarified by centrifugation at 17,000 x g for 10 min. The peptide concentration in the supernatant was quantified using a quantitative colorimetric peptide assay from Thermo Fisher Scientific (Cat. No. 23275). TMT labeling was carried out using the TMTpro-18plex kit from Thermo Fisher Scientific. Briefly, for each sample, 100 pg of peptides was brought to 1 pg / pl with 200 mM EPPS (pH 8.5), acetonitrile (ACN) was added to a final concentration of 30% followed by the addition of 200 pg of each TMT label. After 1 h of incubation at RT, the TMT labeling was quenched by adding 0.3% hydroxylamine (Sigma) for 15 min at RT. Labeled samples were combined, desalted using tC18 SepPak solid-phase extraction cartridges (100 mg, Waters) as described in the SL- TMT method above, and dried in a SpeedVac. Next, desalted peptides were subject to a phosphopeptide enrichment (“mini-phos”) using the High-Select Fe-NTA Phosphopeptide Enrichment Kit (Cat# A32992, Thermo Fisher Scientific) following manufacturer's instructions. The unbound fraction and first two washes were pooled, dried to near dryness using vacuum centrifugation, and subsequently fractionated though basic pH reversed phase chromatography using a HPLC equipped with a 3.5 pm Zorbax 300 Extended-C18 column (Agilent). The fractions were first collected into a 96-well plate and then combined into 24 samples. Twelve of them were desalted following the Cl 8 Stop and Go Extraction Tip (STAGE-Tip)71 and dried down in a SpeedVac. These samples correspond to the total proteome analysis. Eluted peptides from the “mini-phos” enrichment were desalted using STAGE-Tip and dried down in the SpeedVac before MS analysis. Each of the twelve fractions from the total proteome was analyzed once, while the sample containing the phosphopeptides from the “mini-phos” was analyzed twice. All samples were analyzed in an Orbitrap Fusion Lumos mass spectrometer equipped with a FAIMSpro module, operating in high-resolution MS2 (hrMS2) mode (Schweppe, D. K. et al. Characterization and Optimization of Multiplexed Quantitative Analyses Using High-Field Asymmetric-Waveform Ion Mobility Mass Spectrometry. Anal Chem 91, 4010- 4016 (2019), and Schweppe, D. K., Rusin, S. F., Gygi, S. P. & Paulo, J. A. Optimized Workflow for Multiplexed Phosphorylation Analysis of TMT-Labeled Peptides Using High-Field Asymmetric Waveform Ion Mobility Spectrometry. J Proteome Res 19, 554-560 (2020)). The mass spectrometer was coupled to a Proxeon NanoLC 1200 (Thermo Fisher Scientific) mounted with a 100 pm capillary column that was packed with 35 cm of Accucore 150 resin (2.6 pm, 150 A; Thermo Fisher Scientific). Peptides were separated at 525 nL / min flow rate using two buffers as mobile phases (Buffer A: 5% acetonitrile, 0.1% formic acid and Buffer B: 95% acetonitrile, 0.1% formic acid). For both total and phospho TMT analysis, different MS parameters were employed, including gradient length, FAIMS compensation voltage (CV), MSI orbitrap resolution, scan range, MSI maximum injection time, automatic gain control (AGC), MS2 isolation window, higher-energy collision dissociation (HCD), Orbitrap MS2 resolution, and MS2 maximum injection time and AGC.

[0113] A suite of in-house pipeline (GFY-Core Version 3.8, Harvard University) was used to obtain final protein quantifications from all RAW files collected from the Orbitrap Fusion Lumos. RAW data were converted to mzXML format using a modified version of ReAdW.exe. mzXML files were searched using the search engine Comet (Eng, J. K., Jahan, T. A. & Hoopmann, M. R. Comet: an open-source MS / MS sequence database search tool. Proteomics 13, 22-24 (2013)) against a human target-decoy protein database (downloaded from UniProt in April 2021) that included the most common contaminants (Elias, J. E. & Gygi, S. P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat Methods 4, 207-214 (2007) and Elias, J. E. & Gygi, S. P. Target-decoy search strategy for mass spectrometry-based proteomics. Methods Mol Biol 604, 55-71 (2010)). Precursor ion tolerance was set at 20 ppm and product ion tolerance at 0.02 Da. TMTpro tags on lysine residues and peptide N termini (+304.2071 Da) and carbamidomethylation of cysteine residues (+57.021 Da) were set as static modifications, while oxidation of methionine residues (+15.995 Da) was set as a variable modification. For phosphorylation analysis, phosphorylation (+79.966 Da) on serine, threonine, and tyrosine were set as variable modifications. Peptide-spectrum matches (PSMs) were adjusted to a 1% FDR using a linear discriminant analysis as described previously (Huttlin, E. L. et al. A tissue-specific atlas of mouse protein phosphorylation and expression. Cell 143, 1174- 1189 (2010)). For total proteome analysis, proteins were further collapsed to a final protein-level FDR of 1%. Phosphorylation site localization was determined using the AScore algorithm (Beausoleil, S. A., Villen, J., Gerber, S. A., Rush, J. & Gygi, S. P. A probability -based approach for high-throughput protein phosphorylation analysis and site localization. Nat Biotechnol 24, 1285- 1292 (2006).). AScore is a probability-based method designed for the high-throughput localization of protein phosphorylation sites. A threshold of 13 corresponds to a 95% confidence level in site localization. TMT quantitative values were obtained from MS2 scans. Only those PSMs with a summed signal -to-noise ratio (S / N) across all samples > 100 and an isolation specificity > 0.7 were used for quantification. These PSMs were summed for each protein to obtain their relative abundance. To account for equal protein loading in both TMT (total and phosphor), all TMT intensities were normalized so the sum of the TMT signal for all proteins quantified in each channel was equivalent.

[0114] Seahorse assay. OXPHOS and glycolysis of leukemia cells were assessed with a Seahorse XFe96 Analyzer (Agilent Technologies, Santa Clara, CA) using a Cell Mito Stress Test Kit (103015; Agilent Technologies) and a Glycolysis Stress Test Kit (103020; Agilent Technologies), respectively. XF RPMI Base Medium supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose (for the Cell Mito Stress Test Kit) or 1 mM glutamine (for the Glycolysis Stress Test Kit) was used as the assay medium (103681-100; Agilent Technologies). Briefly, Seahorse XF96 Cell Culture Microplates (101085; Agilent Technologies) were treated with Cell-Tak Cell and Tissue Adhesive (354240; Corning) for 20 minutes (22.4 pg / ml; 25 pl per well), after which each well was washed with 200 pl sterile water twice. Then, 2 x io4(CAL-1) or 5 x 104(DF- DCN-86 or DF-DCN-1) leukemia cells were resuspended in 50 pl of assay medium and plated into each well. The microplates were centrifuged at 200 g for 1 minute with zero braking, then transferred to a 37°C non-CCh incubator for 25-30 minutes. Subsequently, another 130 pl of assay medium was added into each well and the microplate was incubated in a non-CCh incubator for 15-25 minutes before being loaded into the XFe96 Analyzer. For the mitochondrial stress test, oligomycin, carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) and rotenone / antimycin A (Rot / AA) were sequentially injected to final concentrations of 1.5 pM, 0.5 (CAL-1) or 1.0 (DF-DCN-86 or DF-DCN-1) pM, and 0.5 pM, respectively. For the glycolysis test, glucose, oligomycin and 2-deoxy -D-glucose (2-DG) were sequentially injected to final concentrations of 10 mM, 1 pM, and 50 mM, respectively. Five replicates were included in each group, and data analyses were performed using Wave software (Agilent Technologies).

[0115] Statistical analysis. Two-tailed t test, Mann-Whitney test, two-way ANOVA test, or Kaplan-Meier test were used for statistical analyses. For all statistical analyses, differences for which P < 0.05 were considered statistically significant, and at least three biologically independent experiments with similar results were reported. The data are presented as means ± S.E.M. or means ± S.D. as indicated in the figure legends. The sample size (n) for each analysis was determined on the basis of pretests and previous similar experiments, and are indicated in the figure legends. Data analyses were performed using GraphPad Prism 9 (GraphPad Software, San Diego, CA).

[0116] Example 1 - PI3Ky inhibition synergizes with cytarabine in leukemia treatment

[0117] To explore the translational potential of targeting PI3Ky in leukemia, the combination of eganelisib with cytarabine was tested using two BPDCN cell lines with elevated PIK3R5 (CAL-1 and DF-DCN-86). Surprisingly, significant synergistic effects between eganelisib and cytarabine were observed in both cell lines (Figs. 9a and 9b). This is contrary to findings in WO2021 / 242859, which disclosed that therapies using PI3Ky inhibitors in combination with metabolic inhibitors show no synergy.

[0118] Next, the intradermal xenografting model of DF-DCN-86 was utilized to assess the dynamic response of leukemia cells to eganelisib and cytarabine in vivo, alone or in combination. It was found that single agent eganelisib was more effective than cytarabine alone, and the combination of these two drugs generated a synergistic suppressive effect on tumor growth (Fig. 4a). Moreover, western blotting of tumor lysates harvested while on treatment showed reduction of PAK1 S144, but not T423, phosphorylation, in animals receiving eganelisib or eganelisib plus cytarabine (Fig. 4b and Fig. 9c).

[0119] Last, the effect of PI3Ky inhibition on survival in six disseminated PDX models was measured, three each PIK3R5-high or PIK3R5-low. As predicted, single agent eganelisib significantly prolonged survival in AML, ALL, and BPDCN with high PIK3R5, while having no benefit in cases with low PIK3R5 (Fig. 4c). However, surprisingly, the combination of eganelisib and cytarabine substantially increased survival in all cases compared to single agent cytarabine, regardless of baseline PIK3R5 (Fig. 4c).

[0120] Based on the in vivo efficacy, investigations were undertaken to find out why the combination of eganelisib and cytarabine were synergistic even in leukemias without elevated PIK3R5. To begin, ex vivo treatment of the three PDXs with lowPIK3R5 expression at cytarabine concentrations resulting in -20% persistent leukemia cells was performed, to enable analysis of residual disease. Strikingly, the cytarabine-persistent cells exhibited substantially elevated phosphorylation of PAK1 S144, but not AKT S473, providing a potential explanation for their induced eganelisib sensitivity (Fig. 4d). There was a very modest increase of PIK3R5 and PIK3CG in some cytarabine-persistent leukemias (Fig. 4d), but this did not seem sufficient to explain the striking increase of PAK1 phosphorylation and the observed in vivo synergy between cytarabine and eganelisib.

[0121] Thus, investigations were undertaken to find out if there were other potential mechanisms contributing to PAK1 activation. An in vivo cytarabine-persistent residual disease model was established using the AML20 PDX. GSEA of RNA-seq comparing cytarabine-persistent to untreated cells indicated that G protein-coupled purinergic receptor signalling, but not other GPCR or PI3K-AKT-mT0R signalling, was activated in the cytarabine-persistent population (Fig. 4e and 9d).

[0122] Thus, our data suggest that activated G protein-coupled purinergic receptor-PI3Ky-PAKl signalling in cytarabine-persistent residual leukemia may also be targeted by P13Ky inhibition, expanding the potential clinical application of PI3Ky inhibitors to leukemias without baseline PIK3R5 activation (see Fig. 4f which shows the PI3Ky complex can be activated via two mechanisms: 1. Intrinsically (baseline inflammatory signature) or extrinsically (via TLRs) activated innate inflammatory signalling promotes high PIK3R5 expression and subsequent stabilization of the PI3Ky complex; 2. Leukemias without baseline elevated PIK3R5 acquire increased PI3Ky activity through G protein-coupled purinergic receptor signalling after cytarabine treatment. In either circumstance, PI3Ky activation leads to increased PAK1 phosphorylation, which drives leukemia via suppression of an NFKB mediated transcriptional network of tumor suppressor genes. Inhibition of PI3Ky reactivates the NFKB-related tumor suppressor genes, leading to compromised OXPHOS and leukemia cell death).

[0123] Taken together, these experiments show that eganelisib and cytarabine exert synergy in models with elevated PIK3R5. Furthermore, the combination of eganelisib and cytarabine was also effective in leukemias with low baseline PIK3R5.

[0124] Example 2 - PIK3R5 expression is correlated with an innate immune response signature (IIRS) and can be activated by innate inflammatory signalling

[0125] As indicated above, the present inventors previously discovered a dependency on PI3 Ky signalling by BPDCNs. PI3Ky signalling was activated by the elevation of PIK3R5 and drives malignant phenotypes. To further define PI3Ky complex dependency, the vulnerability of AML, ALL, and BPDCN PDX models to PI3Ky inhibition was tested. All leukemias with elevated PIK3R5 showed sensitivity to eganelisib, while those with low PIK3R5 expression did not (Fig. la and Fig. 6a, b). Supporting induction of apoptosis as the mechanism of action, eganelisib treatment activated caspase-3 / 7 in sensitive PIK3R5-high leukemias but not in PIK3R5-low cases (Fig. 6c). In addition, PI3Ky inhibition induced markers of terminal myeloid differentiation (CDl lb and CD14) in AMLs with high PIK3R5 (AML11 and AML45) but not low PIK3R5 (AML42 and AML20) (Fig. 6d). Taken together, these results indicated a leukemia intrinsic dependency on the PI3Ky complex shared by various pathological subtypes that can be predicted by PIK3R5 mRNA expression level.

[0126] To further elucidate characteristics of the leukemia subset with elevated PIK3R5, gene set enrichment analysis (GSEA) was performed comparing PIK3R5-high to PIK3R5-low cases. Intriguingly, all the top enriched gene sets were related to innate immune response pathways (Fig. Id). The same conclusion was obtained after removing BPDCN cases from the analysis, excluding the potential bias that might be caused by the dendritic cell origin of BPDCN (Fig. Id). An innate immune response signature (IIRS) of 32 genes was then defined that were part of the GSEA core enrichment in at least three out of the seven top gene sets (Fig. Id and Figs. 6e and 6f). The IIRS expression score of each patient’s leukemia was determined by first transforming the relative expression of each of the 32 genes to values resulting in a mean of 0 and SD of 1 across all patients, then summing up individual gene expression scores to obtain the overall IIRS expression score for each patient’s leukemia. Analysis of the TCGA AML showed that the IIRS expression score highly correlated with PIK3R5 mRNA expression (Fig. 6g) and was more significantly associated with overall survival than the PIK3R5 single gene score (Fig. 1c).

[0127] To further determine the associations between IIRS score and prognosis, we performed subgroup analysis by dividing the TCGAAML cohort into favourable, intermediate, and adverse subgroups based on the 2022 European LeukemiaNet classification. We found that patients with high IIRS had worse prognosis across each of the subgroups, although the difference was most pronounced in the favourable and adverse cohorts (Fig. 6h). In further support of the IIRS defining an inflammatory or differentiation state-associated leukemia subset, French-American-British (FAB) M4 and M5 (myelo-monocytic and monocytic) pathologic subtypes of AML were enriched among TCGA cases with positive IIRS (Fig. 6i).

[0128] Together, these results support a model that activation of innate inflammatory signalling contributes to elevated PIK3R5 expression and to dependency of those leukemia cells on the PI3Ky complex.

[0129] Example 3 - PI3Ky inhibition suppresses OXPHOS and activates an NFicB-related transcriptional network

[0130] To study the downstream events underlying leukemia cell-intrinsic PI3Ky dependency, integrated RNA-seq, proteomic, and phosphoproteomic analyses were performed using CAL-1 cells treated with eganelisib or depleted of PIK3R5 using CRISPRi (PIK3R5i). With inhibition or depletion, gene set enrichment analysis (GSEA) from cell transcriptomes indicated that the most significantly downregulated gene set after targeting PI3Ky was related to OXPHOS (Fig. 2a). This provides additional rationale to selectively target mitochondria in leukemia cells by inhibiting PI3Ky, especially in AML, which is typically deficient in glycolysis and dependent on OXPHOS.

[0131] Seahorse assays were performed to validate the effects of PI3Ky inhibition on cellular energy metabolism. Treatment with eganelisib, but not the class IA PI3K inhibitor LY294002, suppressed the OXPHOS in three different BPDCN cell lines (Fig. 2b). The on-target effect of eganelisib was confirmed using CRISPRi-mediated PIK3R5 / PIK3CG knockdown, which also suppressed OXPHOS (Fig. 2c). In contrast, glycolysis was not affected by eganelisib treatment or PIK3R5 / PIK3CG knockdown, indicating the effects of PI3Ky on energy metabolism is mainly related to suppression of OXPHOS (Fig. 7a).

[0132] The most significantly upregulated gene set upon PI3Ky inhibition was related to TNFa signalling-mediated NFKB activity (Fig. 2d). Five of eleven consistently elevated genes (RHOB, EGR1, CD69, SGK1, and EGR3) after eganelisib treatment and PIK3R5i are within the core enrichment group of this gene set (Fig. 2e) and several have known tumor suppressive roles in various cancers.

[0133] Thus, it was hypothesized that one or more of these core enriched genes contributed to the vulnerability of the PIK3R5-related leukemia subset to PI3Ky inhibition. To test this possibility, the upregulation of these five genes upon eganelisib, but not LY294002 treatment was validated, in CAL-1 and DF-DCN-1 cells (Fig. 7b). Then, each gene was depleted individually in CAL-1 cells and treated with eganelisib. Knockdown of 4 out of the 5 genes reduced sensitivity to PI3Ky inhibition (Fig. 7c, d). The involvement of NFKB in activation of these tumor suppressor genes was validated by treatment of CAL-1 cells with selective NFKB inhibitors IKK-16 and SC-514, and found that eganelisib no longer increased expression of those 5 genes (Fig. 2f and Fig. 7e). Metabolic assays showed that NFKB inhibition blocked suppression of OXPHOS upon PI3Ky inhibitor treatment (Fig. 2g). Finally, CAL-1 cells treated with NFKB inhibitors were rendered resistant to eganelisib (Fig. 2h). A mild growth suppressive effect of NFKB inhibitors on extended leukemia cell growth is not surprising considering the complexity of the NFKB functions and is consistent with previous observations of susceptibility in BPDCN.

[0134] Together, these data support a conclusion that the PI3Ky dependency is at least partially mediated by an NFKB -related transcriptional network.

[0135] Example 4 - PI3Ky inhibition suppresses a noncanonical PI3K pathway that is not dependent on AKT

[0136] Focus then moved to elucidating the substrate(s) of PI3Ky in cells with activated PIK3R5 / PIK3CG. As the most well-defined substrate of PI3K, AKT mediates nearly all the described functions of PI3K complexes, including PI3Ky. Surprisingly, as measured by phosphoproteomics, AKT phosphorylation was not significantly changed in CAL-1 cells after PI3Ky inhibition (Fig. 3a). GSEA also showed that the canonical PI3K-AKT-mT0R signalling target genes were not altered either after eganelisib treatment or PIK3R5 depletion (Fig. 8a). These results indicated that noncanonical PI3K mediators may contribute to the dependency of the PIK3R5-related leukemia subset on PI3Ky.

[0137] Upon further analysis of proteomics and phosphoproteomics data, it was discovered that the phosphorylation of several proteins were affected by both PI3Ky inhibition and PIK3R5 depletion (Fig. 3a).

[0138] PAK1 was selected for further study, as PAK family members play important roles in other oncogenic signal transduction, PAK1 has not previously been connected with PI3Ky, and there were validated PAK1 inhibitors and phospho-specific antibodies available. Western blotting results confirmed that phosphorylation of PAK1 at serine 144 (S144), but not phosphorylation of other known proteins in canonical PI3K signalling, was significantly reduced upon PIK3R5 / PIK3CG depletion (Fig. 3b). Moreover, treatment with eganelisib, but not LY294002, reduced PAK1 S144 phosphorylation in two cell lines (Fig. 3c). Of note, depletion of PIK3R5 or PIK3CG led to a complementary increase of AKT phosphorylation (Fig. 3b), whereas eganelisib induced modest suppression of AKT phosphorylation (Fig. 3c). This is consistent with the known activity of eganelisib on the suppression of AKT phosphorylation possibly via its weak inhibition of class IA PI3Ks, which is why focus was placed on shared events after both eganelisib treatment and PIK3R5 knockdown to provide specificity.

[0139] Additionally, across 23 AML patient samples compared to normal CD34+ hematopoietic progenitor cells, we found that leukemias expressing low PIK3R5 / PIK3CG and negative PAK1 S144 phosphorylation were not sensitive to PI3Ky inhibition (Fig. 3i and Fig. 8d,e). In contrast, patient samples expressing high PIK3R5 / PIK3CG and detectable PAK1 S144 phosphorylation were sensitive to eganelisib (Fig. 3i and Fig. 8d,e). There was a smaller proportion of primary samples with high PIK3R5 / PIK3CG but negative PAK1 phosphorylation, and these had intermediate sensitivity to eganelisib. Together, these data supported our proposed model that high PIK3R5 / PIK3CG and PAK1 S144 phosphorylation are associated with PI3Ky inhibitor sensitivity in AML

[0140] To further validate the functional involvement of PAK1 S144 phosphorylation in PI3Ky dependency, CAL-1 cells were treated with selective inhibitors of PAK1 or AKT. PAK1 inhibitor treatment mimicked the strong suppression of CAL-1 growth that was observed with PI3Ky inhibition, while the AKT inhibitor MK-2206 only minimally impaired cell viability despite completely abrogating AKT phosphorylation (Fig. 3d,e). In order to evaluate the role of S144 as a possible phosphorylation site that mediates PAK1 function downstream of PI3Ky, exogenous constitutively active (S144D) or inactive (SI 44 A) PAK1 phospho-site mutants were introduced into CAL-1 cells and treated with eganelisib. The introduction of active PAK1S144D, but not inactive PAK1S144A, conferred resistance to PI3Ky inhibition (Fig. 3f and Fig. 8c). As controls, introduction of either activated or inactivated PAK1 mutants at threonine 423 (T423) showed no effect on sensitivity to PI3Ky inhibition (Fig. 3f and Fig. 8c).

[0141] Last, the effect of PAK1 S144 phosphorylation on PI3Ky downstream events was tested. Seahorse assays indicated that introduction of PAK1S144Dabolished the suppression of OXPHOS by PI3Ky inhibition (Fig. 3g). Eganelisib-induced elevation of the NFicB-related tumor suppressor genes was also diminished in CAL-1 cells expressing activated PAK1S144D(Fig. 3h).

[0142] Taken together, these results indicated that PAK1 is a critical substrate of PI3Ky in the leukemia-intrinsic dependency pathway and phosphorylation of PAK1 S144 plays the essential role in signalling. Therefore, inhibition of PAK1 presents as another therapeutic node to target the leukemia intrinsic PI3Ky pathway.

Claims

CLAIMS:

1. A method of treating a blood cancer, comprising the administration of therapeutically effective amounts of a phosphatidylinositol 3 -kinase gamma (PI3Ky) inhibitor and an antimetabolite to a patient in need thereof.

2. A PI3Ky inhibitor and an antimetabolite for use in the treatment of a blood cancer.

3. Use of a PI3Ky inhibitor and an antimetabolite in the manufacture of a medicament or medicaments for the treatment of a blood cancer.

4. A method according to claim 1; a PI3Ky inhibitor and an antimetabolite for use according to claim 2; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3; wherein the blood cancer is selected from the group consisting of leukemia, lymphoma, myeloma, myelodysplastic syndromes (MDS), and myeloproliferative neoplasms (MPN).

5. A method according to claim 1 or 4; a PI3Ky inhibitor and an antimetabolite for use according to claim 2 or 4; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3 or 4; wherein the blood cancer is resistant to cytarabine monotherapy.

6. A method according to claim 1, 4 or 5; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4 or 5: or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4 or 5; wherein the PI3Ky inhibitor has an IC50 of 100 or less.

7. A method according to claim 1, 4, 5 or 6; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5 or 6; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, or 6; wherein the PI3Ky inhibitor has atleast 10-fold selectivity for PI3Ky over PI3Ka; and / or at least 10-fold selectivity for PI3Ky over PI3K0; and / or at least 10-fold selectivity for PI3Ky over PI3K8.

8. A method according to claim 1, 4, 5, 6, or 7; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, or 7; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, or 7; wherein the PI3Ky inhibitor is selected from the group consisting of apitolisib, buparlisib, dactolisib, duvelisib, eganelisib, edatolisib, GSK1059615, omipalisib, paxalisib, samotolisib, tenalisib, taselisib, voxtalisib, wortmannin, AS252424, AS605240, AZD3458, ZX-lOla and ZX-4081.

9. A method according to claim 1, 4, 5, 6, 7 or 8; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7 or 8; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7 or 8; wherein the antimetabolite is selected from the group consisting of 5-FU [fluorouracil], azacitidine, capecitabine, cladribine, clofarabine, cytarabine, decitabine, floxuridine, fludarabine phosphate, gemcitabine, hydroxyurea, methotrexate, nelarabine, pemetrexed, pentostatin, pralatrexate, thioguanine, and trifluridine / tipiracil.

10. A method according to claim 1, 4, 5, 6, 7, 8 or 9; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7, 8 or 9; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8 or 9; wherein the PI3Ky inhibitor is eganelisib and the antimetabolite is cytarabine11. A method according to claim 1, 4, 5, 6, 7, 8, 9 or 10; a PI3Ky inhibitor and an anti metabolite for use according to claim 2, 4, 5, 6, 7, 8, 9 or 10; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8, 9 or 10; wherein the patient to be treated has been previously identified as having low baseline PIK3R5.

12. A method of identifying a blood cancer patient susceptible to treatment with a PI3Ky inhibitor comprising:(i) comparing the levels of a factor which is indicative of the expression of the gene PIK3R5 and the levels of a factor which is indicative of the expression of at least one gene selected from the group consisting of IRF7, LCP2, CCL5, CXCL9, PLSCR1, LYN, JAK2, IFNGR1, IRF8, CD86, TLR2, FAS, STAT1, TAPBP, B2M, CD74, TAPI, CASP1, LAP3, IL18R1, TNFRSF1B, SELL, IL10RA, TNFSF10, IFITM3, STAT2, IRF9, EIF2AK2, LY6E, BST2, and RTP4 in cancerous blood cells obtained from a blood cancer patient to the levels of those same factors in non-cancerous blood cells obtained from said patient, or to the levels of those same factors known to be present in non-cancerous blood cells of healthy individuals, or to the levels of those same factors known to be present in cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor; and(ii) selecting said patient for treatment if the levels of said factors in their cancerous blood cells are higher than the levels of said factors in their non- cancerous blood cells, or higher than the levels of said factors in the non- cancerous blood cells of healthy individuals, or higher than the levels of said factors in the cancerous blood cells of blood cancer patients who are not susceptible to treatment with a PI3Ky inhibitor.

13. A method according to claim 12, wherein the levels of said factors in the cancerous blood cells are considered to be higher than their levels in the non-cancerous blood cells when they are at least 110%, preferably at least 120%, more preferably at least 130%, even more preferably at least 150%, most preferably at least 200% of their levels in the non-cancerous blood cells.

14. A method of treating a blood cancer comprising administration of a therapeutically effective amount of a p-21 -activated kinase 1 (PAK-1) inhibitor to a patient in need thereof.

15. A PAK-1 inhibitor for use in the treatment of a blood cancer.

16. The use of a PAK-1 inhibitor in the manufacture of a medicament for the treatment of a blood cancer.

17. A method according to claim 14, a PAK-1 inhibitor for use according to claim 15, or the use of a PAK-1 inhibitor according to claim 16, wherein the blood cancer is selected from the group consisting of leukemia, lymphoma, myeloma, MDS, and MPN.

18. A method according to claim 14 or 17; a PAK-1 inhibitor for use according to claim 15 or 17; or a use of a PAK-1 inhibitor according to claim 16 or 17; wherein the PAK-1 inhibitor has an IC50 of 100 or less.

19. A method according to claim 14, 17 or 18; a PAK-1 inhibitor for use according to claim 15, 17 or 18; or a use of a PAK-1 inhibitor according to claim 16, 17 or 18; wherein the PAK-1 inhibitor has at least 2-fold selectivity for PAK-1 over PAK-2; and / or at least 2-fold selectivity for PAK-1 over PAK-3; and / or at least 2-fold selectivity for PAK-1 over PAK-4; and / or at least 2-fold selectivity for PAK-1 over PAK-6; and / or at least 2-fold selectivity for PAK-1 over PAK-5 / 7.

20. A method according to claim 14, 17, 18 or 19; a PAK-1 inhibitor for use according to claim 15, 17, 18 or 19; or a use of a PAK-1 inhibitor according to claim 16, 17, 18 or 19; wherein the PAK-1 inhibitor is selected from the group consisting of: FRAX597, FRAX1036, G-5555, FL172, PF-3758309, AZ13705339, IPA-3, and NVS-PAK1-1,21. A method according to claim 1, 4, 5, 6, 7, 8, 9, 10 or 11; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7, 8, 9, 10 or 11; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8, 9, 10 or 11; comprising administration of a therapeutically effective amount of BCL-2 inhibitor in addition to a PI3Ky inhibitor and an antimetabolite.

22. A method according to claim 1, 4, 5, 6, 7, 8, 9, 10, 11 or 21; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7, 8, 9, 10, 11 or 21; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8, 9, 10, 11 or 21; wherein the BCL-2 inhibitor is selected from the group consisting of venetoclax (ABT-199), S55746 (BCL201), lisaftoclax (APG-2575), oblimersen (G3139), AZD4320, AZD0466, pelcitoclax (APG-1252), BM-1197, S44563, ABT-737, navitoclax (ABT-263), obatoclax (GX15-070) and AT-101.

23. A method according to claim 1, 4, 5, 6, 7, 8, 9, 10, 11, 21 or 22; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7, 8, 9, 10, 11, 21 or 22; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8, 9, 10, 11, 21 or 22; wherein the BCL-2 inhibitor is venetoclax.

24. A method according to claim 1, 4, 5, 6, 7, 8, 9, 10, 11, 21, 22 or 23; a PI3Ky inhibitor and an antimetabolite for use according to claim 2, 4, 5, 6, 7, 8, 9, 10, 11, 21, 22 or 23; or the use of a PI3Ky inhibitor and an antimetabolite according to claim 3, 4, 5, 6, 7, 8, 9, 10, 11, 21, 22 or 23; wherein the wherein the PI3Ky inhibitor is eganelisib, the antimetabolite is azacitidine and the BCL-2 inhibitor is venetoclax.

Images