Activation of protein phosphatase 2A in FLT3+ acute myeloid leukemia cells enhances the cytotoxicity of FLT3 tyrosine kinase inhibitors

Constitutive activation of the receptor tyrosine kinase Fms-like tyrosine kinase 3 (FLT3), via co-expression of its ligand or by genetic mutation, is common in acute myeloid leukemia (AML). In this study we show that FLT3 activation inhibits the activity of the tumor suppressor, protein phosphatase 2A (PP2A). Using BaF3 cells transduced with wildtype or mutant FLT3, we show that FLT3-induced PP2A inhibition sensitizes cells to the pharmacological PP2A activators, FTY720 and AAL(S). FTY720 and AAL(S) induced cell death and inhibited colony formation of FLT3 activated cells. Furthermore, PP2A activators reduced the phosphorylation of ERK and AKT, downstream targets shared by both FLT3 and PP2A, in FLT3/ITD+ BaF3 and MV4-11 cell lines. PP2A activity was lower in primary human bone marrow derived AML blasts compared to normal bone marrow, with blasts from FLT3-ITD patients displaying lower PP2A activity than WT-FLT3 blasts. Reduced PP2A activity was associated with hyperphosphorylation of the PP2A catalytic subunit, and reduced expression of PP2A structural and regulatory subunits. AML patient blasts were also sensitive to cell death induced by FTY720 and AAL(S), but these compounds had minimal effect on normal CD34+ bone marrow derived monocytes. Finally, PP2A activating compounds displayed synergistic effects when used in combination with tyrosine kinase inhibitors in FLT3-ITD+ cells. A combination of Sorafenib and FTY720 was also synergistic in the presence of a protective stromal microenvironment. Thus combining a PP2A activating compound and a FLT3 inhibitor may be a novel therapeutic approach for treating AML.

Human bone marrow samples were obtained from AML patients according to institutional guidelines as previously described [28]. Studies were approved by human ethics committees of the Royal Adelaide Hospital, the University of Newcastle, the South Eastern Sydney and Illawarra Area Health Service, and the Hunter New England Area Health. Mononuclear cells were isolated by Ficoll-Hypaque density-gradient centrifugation, washed, and resuspended at 5 × 10 5 /mL in IMDM, 0.5% FCS, 1% penicillin/streptomycin. In some experiments cells were cultured in the presence of 50 ng/ml FL as indicated. All samples had a minimum of 75% AML blasts. FLT3 status was determined using routine pathology, and was confirmed by genomic DNA sequencing utilizing juxtamembrane domain (Forward-5′-GCAATTTAGGTATGAAAGCCAGC-3′, Reverse-5′-CTTTCAGCATTTTGACGGCAACC-3′) and kinase domain (Forward-5′-TTCACAGAGACCTGGCCG-3′, Reverse-5′-GCCCAAGGACAGATGTGATG-3′) specific primers and sequenced by Sanger sequencing (AGRF; Brisbane, Australia). FACS purification of primary human CD34 + CD38 -CD123 + leukemic stem and progenitor cells (LSPCs) was performed as previously described [28].
For primary human CD34 + bone marrow cells, 500 cells were seeded in duplicate into 6-well culture plates with 1.1 ml methocult (StemCell Technologies, #H4034) in 0, 1 or 3 µM FTY720 or AAL(S). Total colonies were counted after 7 days and differential colonies counted after 14 days according to manufacturer's instructions.

Statistical analysis
Statistical significance of differences between samples was assessed using a Student's t-test. Values shown are the mean ± SEM. All analyses were performed using GraphPad Prism 5 software (GraphPad Software Inc, CA, USA). IC 50 values were calculated using cubic spline regression analysis using GraphPad Prism 5.0. For co-culture combination assays, synergy was determined by the Fractional Product method, calculated according to Webb 30 . Figure S1: Expression, proliferation and PP2A activity of FLT3 transduced BaF3 and FDC.P1 cells.

Supplementary
(A) FLT3 surface expression was monitored routinely using an anti-FLT3/FLT2 (SF1.340; Santa Cruz) antibody and an APC-conjugated secondary antibody and analyzed on a FACSCalibur flow cytometer (BD Biosciences). (B) BaF3 parental cells were stably transduced with empty vector (EV), WT, D835V, D835Y or ITD forms of human FLT3 and grown in the presence of a range of FLT3 ligand (rFL) concentrations for 48 hr. Cell proliferation was determined using a resazurin assay and expressed as fluorescence units. (C) BaF3 and (D) FDC.P1 cells transduced with human FLT3-WT were grown in the presence of IL3, GM-CSF or FL as indicated. Cell proliferation was assessed by seeding triplicate wells of a 96-well plate with 1 × 10 4 cells/well, and growth determined by resazurin reduction over time. (E) PP2A complexes were isolated from FDC.P1 mouse myeloid cells transduced with WT FLT3, and grown in the presence of granulocytemacrophage-colony stimulating factor (GM) or FLT3 ligand (FL), treated with or without 3 µM FTY720 or AAL(S) for 12 h, using immunoprecipitation with an anti-PP2A-C antibody. PP2A activity was determined by incubating the isolated PP2A-C complex with a PP2A-specific phosphopeptide and measuring free phosphate release using a colorimetric assay. Activity was calculated as a percentage of control WT-GM cells Columns; mean, bars; SEM, *p < 0.05, compared to WT-GM or untreated cells, as indicated; Students t test. Figure S2: (A) Immunoblot of BaF3 cells expressing an empty vector (EV) or FLT3, and FLT3-ITD + MV4-11 cells +/-3 µM FTY720, 24 hr. Numbers underneath blot show the densitometric ratio of the pY307-PP2Ac band divided by the total PP2Ac band. β-Actin is shown as a loading control. (B) Quantitation of immunoblots of BaF3 parents and FLT3-ITD cells showing expression of pY307PP2A-C normalised to total PP2A-C, PP2A-C, PP2A-A, PP2A-B55α, -B55δ -B56α, -B56γ, -B56ε -B"α 130 kDa, -B"α 72 kDa,-B"β. β-actin was used as a loading control. Quantitation of densitometric volume relative to BaF3/Parental cells reveals a significant difference in several PP2A subunits. Columns, mean densitometric volume relative to β-actin normalised to BaF3 parental cells, n = 4; bars, SEM. *p < 0.05, **p < 0.01; Student's t test compared to BaF3. (C). Quantitation of immunoblots of BaF3 cell lysates showing expression of pJAK2(Y221)/JAK2, p-p38MAPK(T180/Y182)/p38MAPK, pERK(1/2)(T202/Y204)/ERK1/2), pMEK(S221)/MEK and pAKT(T308)/AKT. β-actin was used as a loading control. Quantitation of OD relative to BaF3/Parental reveals a significant difference in several signalling proteins. Columns, mean OD (n = 4) relative to β-actin normalised to parental; bars, SEM. *p < 0.05; **p < 0.01; Student's t test compared to parental. (B) Gene expression data from U133+2.0 microarrays was normalised in R using the JustRMA() function in the affy library. For both datasets, the mean expression level of each PP2A subunit gene in FLT3 mutation positive samples was compared to non-mutated normal karyotype samples using a two-tailed t-test with p < 0.05. Box and whiskers plots show the median, first quartile and third quartile (box) and upper and lower limits of expression (whiskers) for each gene in FLT3 mutation negative and positive samples.

Supplementary
were treated with FTY720, AG1296 or CEP701 for 48 hr. Survival is the percentage of annexin V negative cells. (D-E) Normal bone marrow CD34+ cells (n = 4-6) were obtained from StemCell Technologies. Cells were seeded in duplicate full cytokine complement methylcellulose and treated with FTY720 or AAL(S) at the indicated concentrations. (D) After 7 days total colony number was counted. (E) After 14 days differential colonies were scored and categorized as colony forming unit erythroid (CFU-E), burst forming unit erythroid (BFU-E), or colony forming unit granulocyte, macrophage (CFU-GM). (F-G) Normal human bone marrow CD34+ cells were purchased from Lonza (n = 2) and seeded in duplicate in full cytokine complement methylcellulose and treated with FTY720 or AAL(S) at the indicated concentrations. (F) After 7 days total colony number was counted. (G) After 14 days differential colonies were scored.Columns; mean, bars; SEM. There were no significant differences with any drug treatments compared to control untreated cells.