The FLT3 Internal Tandem Duplication Mutation Prevents

Apoptosis in Interleukin-3-Deprived BaF3 Cells Due to

Protein Kinase A and Ribosomal S6 Kinase 1–Mediated

BAD Phosphorylation at Serine 112

Xinping Yang,1Liyun Liu,

2David Sternberg,

3Liren Tang,

4Ilene Galinsky,

1

Daniel DeAngelo,1and Richard Stone

1

1Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts; 2Department ofMolecular and Cellular Biology, Harvard University, Cambridge, Massachusetts; 3Division of Hermatology/Oncology,Mount Sinai School of Medicine, New York, New York; and 4Department of Medicine, University of British Columbia,Vancouver Hospital, Vancouver, British Columbia, Canada

Abstract

Internal tandem duplication (ITD) mutations in the FLT3tyrosine kinase have been detected in f20% of acute myeloidleukemia (AML) patients. Patients harboring FLT3/ITD muta-tions have a relatively poor prognosis. FLT3/ITD results inconstitutive autophosphorylation of the receptor and factor-independent survival. Previous studies have shown that FLT3/ITD activates the signal transducers and activators oftranscription 5 (STAT5), p42/p44 mitogen-activated proteinkinase [MAPK; extracellular signal-regulated kinase (ERK)1/2], and phosphatidylinositol 3-kinase/Akt pathways. Weherein provide biochemical and biological evidence thatribosomal S6 kinase 1 (RSK1) and protein kinase A (PKA)are the two principal kinases that mediate the antiapoptoticfunction of FLT3/ITD via phosphorylation of BAD at Ser112.Inhibiting both MAPK kinase (MEK)/ERK and PKA pathwaysby a combination of U0126 (10 Mmol/L) and H-89 (5 Mmol/L)reduced most of BAD phosphorylation at Ser112 and inducedapoptosis to a level comparable with that induced by FLT3inhibitor AG1296 (5 Mmol/L) in BaF3/FLT3/ITD cells. RNAinterference of RSK1 or PKA catalytic subunit reduced BADphosphorylation and induced apoptosis. The MEK inhibitorU0126 and/or the PKA inhibitor H-89 greatly enhanced theefficacy of the FLT3 inhibitor AG1296, suggesting that combin-ing FLT3/ITD downstream pathway inhibition with FLT3inhibitors may be a viable therapeutic strategy for AML causedby a FLT3/ITD mutation. (Cancer Res 2005; 65(16): 7338-47)

Introduction

Certain mutations in receptor tyrosine kinases (RTK) causeconstitutive activation of these enzymes, potentially resulting inmalignancy (1). Among these RTKs, FLT3 ( fms-like tyrosinekinase) has recently been found to be involved in the pathogenesisof acute myeloid leukemia (AML). Two types of FLT3 mutationshave recently been detected in patients with AML: (a) internaltandem duplications (ITD; inserted repeat sequences spanningfrom <7 to >30 amino acids) in the juxtamembrane domain inf20% (2–4) and (b) substitution mutations in the activation loop(usually D835Y) in f7% (5, 6). Patients harboring FLT3/ITD

mutations have a relatively poor prognosis (7–9), especially whenthe other allele is mutated and/or lost (10). Both FLT3/ITD andFLT3/D835 mutations result in autophosphorylation and activa-tion of the receptor kinase (11–14). Syngeneic mice injected with32D cells carrying FLT3/ITDs develop myeloproliferative disease(14, 15). Similarly, transplantation of FLT3/ITD-transduced bonemarrow cells causes myeloproliferative disease in another mousemodel (16).Normal hematopoietic cells depend on growth factors, such as

interleukin-3 (IL-3), for survival and proliferation, whereas leukemiccell lines and primary leukemic cells often become partially orcompletely factor independent. IL-3-dependent survival in BaF3cells is mediated by various signaling pathways, including theMAPK kinase (MEK)/extracellular signal-regulated kinase (ERK;refs. 17–19), phosphatidylinositol 3-kinase (PI3K; refs. 20–23),protein kinase A (PKA; ref. 24), and signal transducers andactivators of transcription 5 (STAT5; ref. 25) pathways.The Bcl-2 family members are downstream of antiapoptotic

signals, such IL-3. Such proteins include both proapoptotic(e.g., BAD) and antiapoptotic (e.g., Bcl-2 and Bcl-XL) species(26–28). The BAD function is inactivated by phosphorylation inresponse to survival factors, such as IL-3, nerve growth factor, andinsulin-like growth factor (18, 20, 24, 29–31). UnphosphorylatedBAD forms a heterodimer with the Bcl-2 homologue Bcl-XL,thereby inhibiting its antiapoptotic function (31–33). When BAD isphosphorylated at Ser112 and/or Ser136, it forms a complex with the14-3-3 protein in the cytosol preventing binding to Bcl-2 or Bcl-XLon the mitochondrial membrane (31, 34, 35). IL-3-mediatedactivation of the MEK (17, 18), PI3K (20), and PKA (24) pathwaysresults in BAD phosphorylation. IL-3-mediated activation of STAT5up-regulates the expression of Bcl-XL and promotes cell survival(25). Constitutive activation of STAT5 leads to expression of a setof its target genes, including Bcl-XL, and confers IL-3 independence(36). Ribosomal S6 kinase 1 (RSK1) mediates the MEK/ERKpathway cell survival signal via BAD phosphorylation at Ser112 (18).PKA also have been shown to mediate IL-3-induced BADphosphorylation at Ser112 (24). Akt, in the PI3K pathway, has beenimplicated in the phosphorylation of BAD at Ser136 (20, 29), but theimportance of the role of Akt in BAD phosphorylation has beenquestioned (24).In association with factor-independent survival and growth in

stably transduced 32D and BaF3 cells, FLT3/ITD has been shownpreviously to activate STAT5, p42/p44 mitogen-activated proteinkinase (MAPK), and PI3K/Akt pathways (13, 14). ConstitutiveSTAT5 phosphorylation has been observed widely in AML patients,

Requests for reprints: Richard Stone, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, 44 Binney Street, Boston, MA 02115.Phone: 617-632-2214; Fax: 617-632-2933; E-mail: rstone@partners.org.

I2005 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-04-2263

Cancer Res 2005; 65: (16). August 15, 2005 7338 www.aacrjournals.org

Research Article

including that caused by FLT3 phosphorylation (37). InhibitingFLT3/ITD reduced the phosphorylation of STAT5a and expressionof Bcl-XL in FLT3/ITD transduced cells (38). FLT3/ITD inactivatesthe proapoptotic BAD by phosphorylation. However, inhibitingboth MEK/ERK and PI3K pathways is not sufficient for inducingapoptosis in FLT3/ITD-transduced BaF3 cells (38).This study provides biochemical and biological evidence that

p90RSK (RSK1) in the MEK/ERK pathway and PKA are the twoprincipal kinases that mediate the antiapoptotic function of FLT3/ITD via phosphorylation of BAD at Ser112. Pharmaceuticalinhibition of FLT3/ITD, MEK, and/or PKA reduced BAD phosphor-ylation at Ser112 and induced apoptosis. Combined inhibition ofMEK and PKA essentially eliminated the phosphorylation of BAD atSer112 and synergistically induced apoptosis. Inhibition of RSK1 orPKA with small interfering RNAs (siRNA) supports the conclusionmade from pharmaceutical inhibition. These findings provide anaugmented understanding of FLT3/ITD-induced leukemogenesisand suggest potential therapeutic targets for AML patients.Moreover, inhibition of FLT3/ITD and its downstream MEK/ERKor/and PKA pathways greatly improves the efficacy of FLT3inhibitor drugs in inducing apoptosis of BaF3/FLT3/ITD cells.

Materials and Methods

Antibodies. Anti-human FLT3 rabbit polyclonal antibody, anti–PKAcatalytic subunit (PKAc) rabbit polyclonal antibody, anti-RSK1 rabbit

polyclonal antibody, anti-BAD rabbit polyclonal antibody, and anti-Bcl-2

and Bcl-XL monoclonal antibodies were purchased from Santa Cruz

Biotechnology (Santa Cruz, CA). Anti-BAD (total) monoclonal antibodywas obtained from Upstate, Inc. (Waltham, MA). The following antibodies

were from Cell Signaling Technologies, Inc. (Boston, MA): anti-phospho-

FLT3 (Tyr591) polyclonal antibody, anti-phospho-BAD (Ser112) monoclonal

antibody, phospho-BAD (Ser136) polyclonal antibody, anti-BAD polyclonalantibody (does not recognize phospho-Ser112-BAD), anti-phospho-MEK1/2

(Ser217/Ser221) rabbit polyclonal antibody, anti-MEK1/2 rabbit polyclonal

antibody, anti-phospho-p44/42 ERK1/2 (Thr202/Tyr204) rabbit polyclonal

antibody, anti-p42/MAPK 3A7 monoclonal antibody, and anti-phospho-RSK1 (Ser381) rabbit polyclonal antibody. All the phosphosites refer to those

on human proteins. Anti-RSK1 and anti-PKAc monoclonal antibodies and

phycoerythrin (PE)– or FITC-conjugated goat anti-mouse or goat anti-rabbit antibodies were from BD PharMingen (San Diego, CA).

Inhibitors. The FLT3 inhibitor AG1296, the MEK inhibitor UO126, and

the PKA inhibitor H-89 were obtained from Calbiochem-Novabiochem

Corp. (San Diego, CA) and Cell Signaling Technologies.Isolation of full-length FLT3 cDNA and construction of expression

plasmids. We screened for ITDs in FLT3 in genomic DNA from AML

patients’ bone marrow or peripheral blood (after informed consent on

DFCI protocol 01-017 was obtained) using primers flanking the ITDregion (3, 7). Total RNA was isolated from Ficoll-separated mononuclear

bone marrow cells from an AML patient whose blasts displayed an ITD

(26 amino acids) in the juxtamembrane domain of FLT3. Full-lengthFLT3 cDNA clones both with or without ITD were obtained by reverse

transcription-PCR. The cDNAs were sequenced from both ends to

confirm the sequences. The deduced amino acid sequence in the ITD

region is shown in Fig. 1A . The cDNA clone without an ITD mutationwas found to have 1 bp deleted. This cDNA clone was changed to

normal wild-type (WT) sequence using site-directed mutagenesis

(Invitrogen Corp., Carlsbad, CA) following the manufacturer’s protocol

and further confirmed by sequencing. The cDNAs of the FLT3/ITD andFLT3/WT were subcloned into a murine retroviral vector MSCVneo

(kindly provided by R. Hawley, University of Toronto, Toronto, Ontario,

Canada) under the control of a long terminal repeat.

Cell culture and generation of stable cell lines expressing humanFLT3/ITD. The murine IL-3-dependent hematopoietic cell line, BaF3, was

maintained in RPMI 1640 containing 10% fetal bovine serum (FBS) and

2 ng/mL murine IL-3 (R&D Systems, Inc., Minneapolis, MN). 293T cells weremaintained in DMEM with 10% FBS. The leukemic cell line MV4-11

harboring FLT3/ITD (10 amino acids) was cultured in RPMI 1640 with

20% FBS.

The FLT3/ITD and FLT3/WT cDNA constructs or the vector werecotransduced with a packaging construct, pIK69 (Genesys, Redwood City,

CA), into 293T cells using Superfect (Qiagen, Valencia, CA) according to the

manufacturer’s instructions. Supernatants were collected 48 hours after

transduction, filtered (0.45 Am), and used to transduce BaF3 cells asdescribed previously (16). Transduced BaF3 cells were selected in

methylcellulose medium (StemCell Technology, Vancouver, British Colum-

bia, Canada) with 1 mg/mL G418 and 2 ng/mL IL-3 to obtain G418-resistant

colonies. Expression of FLT3 was verified by Western blotting and bystaining with PE-conjugated monoclonal anti-human FLT3 antibody and

analysis with flow cytometry (data not shown).

Treatment of interleukin-3, FLT3 ligand, and inhibitors. BaF3/FLT3/WT or BaF3/FLT3/ITD cells were cultured at a starting density of 2 � 105

cells/mL in RPMI 1640 with 10% FBS with IL-3 (20 ng/mL) or FLT3 ligand

(100 ng/mL) or without cytokine for 24 hours before cells were harvested

for apoptosis analysis. To investigate the effects of the inhibitors of FLT3/ITD and its downstream pathways on apoptosis, the FLT3 inhibitor AG1296

(2.5 and 5 Amol/L), the MEK1/2 inhibitor U0126 (12.5 and 25 Amol/L), the

PKA inhibitor (2.5 and 5 Amol/L), or a combination of these inhibitors

were added to the culture medium. MV4-11 cells were cultured at astarting density of 2 � 105 cells/mL in RPMI 1640 with 10% FBS. AG1296

(2.5 Amol/L), U0126 (10 Amol/L), H-89 (10 Amol/L), or a combination of

these inhibitors were added to the medium for 48 hours before harvestingfor apoptosis analysis.

RNA interference. SMARTpool siRNAs were obtained from Dharmacon

(Chicago, IL) and GeneSilencer siRNA Transfection Reagent from Gene

Therapy Systems, Inc. (San Diego, CA). The siRNA transfection was doneaccording to the manufacturer’s instruction. siGuard RNase inhibitor (Gene

Therapy Systems) was added to the culture just before transfection to

protect siRNA from degradation during transfection. To increase the

efficiency of RNA interference, siRNA transfection was done for a secondtime 24 hours after first transfection. Cells were harvested 48 hours after the

second transfection.

Analysis of apoptosis and protein using flow cytometry. The cellswere stained with Annexin V-FITC and propidium iodide (PI) before flow

cytometry analysis using the apoptosis detection kit protocol (CN

Biosciences, Inc., Boston, MA).

The analysis of protein using flow cytometry included cell fixing,permeabilization, and immunostaining. The fix buffer, permeabilization

buffer, and washing/staining buffer were purchased from Santa Cruz

Biotechnology and the process was done according to the manufacturer’s

instructions.Protein kinase A kinase assay. BaF3/FLT3/WT and BaF3/FLT3/ITD

cells were placed in RPMI 1640 with 1% FBS for 8 hours, washed with PBS,

and harvested with cold extraction buffer containing 25 mmol/L Tris-HCl

(pH 7.4), 0.5 mmol/L EDTA, 0.5 mmol/L EGTA, 10 mmol/L h-mercaptoe-thanol, 1 Ag/mL leupeptin, 1 Ag/mL aprotinin, and 0.5 mmol/L phenyl-

methylsulfonyl fluoride (PMSF). PKA kinase activity was assessed using the

Nonradioactive Cyclic AMP (cAMP)–Dependent Protein Kinase Assay kit(Promega, Madison, WI). Protein concentration of the crude lysates were

quantitated, and equal mounts of protein were added to a reaction mixture

(kinase buffer) containing 40 mmol/L Tris-HCl (pH 7.4), 20 mmol/L MgCl2,

0.1 mg/mL bovine albumin, 100 Amol/L Pep-Tag A1 peptide, and 0.5 mmol/LATP with or without 1 mmol/L cAMP. Experiments were done in parallel

with addition of PKAc as positive control and no addition of PKA as

negative control. The phosphorylated and nonphosphorylated Pep-Tag

peptides were separated by electrophoresis on 0.8% agarose gel andvisualized under UV.

Immunoprecipitation and Western blotting. Cell were harvested and

rinsed with ice-cold PBS. Ice-cold lysis buffer [0.5 mL; 20 mmol/L Tris(pH 7.5), 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1% Triton

X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L h-glycerophosphate,1 mmol/L Na3VO4, 1 Ag/mL leupeptin, 1 mmol/L PMSF] was added to

FLT3/ITD Prevents Apoptosis via PKA and RSK1

www.aacrjournals.org 7339 Cancer Res 2005; 65: (16). August 15, 2005

1 � 107 cells and sonicated on ice four times for 5 seconds each followed by

microcentrifugation for 10 minutes at 4jC. Primary antibody (10 AL) wasadded to 200 AL of the cell lysate and incubated with gentle rocking at 4jCfor 2 hours. Protein A-agarose beads (Upstate; 20 AL of 50% beads slurry)were added and incubated with gentle rocking at 4jC for 2 hours followed

by microcentrifugation for 30 seconds at 4jC, washing thrice for Western

blotting analysis, or washing twice with lysis buffer and twice kinase buffer

for the kinase assay.

For Western blotting, cells were lysed in Laemmli sample buffer,

sonicated for 10 to 20 seconds, heated for 5 minutes, separated byelectrophoresis in SDS-polyacrylamide gel, and transferred to PROTRAN

membranes. The membranes were blotted with primary antibodies and

Figure 1. Antiapoptotic function of FLT3/ITD. A, a full-length FLT3 cDNA with an ITD (FLT3/ITD) from an AML patient was cloned and sequenced. The deducedamino acid sequence around the juxtamembrane domain is shown and the duplicated region is underlined. B, BaF3 cells expressing FLT3/ITD were starved for atleast 8 hours in medium containing 1% FBS and then treated with the indicated concentrations of AG1296 for 90 minutes. Western blotting was done to detectphospho-FLT3 (Tyr591). The membranes were stripped and reprobed with anti-human FLT3 antibodies to show equal loading. C, BaF3 cells transduced with FLT3/WTor FLT3/ITD were seeded in six-well plates at a density of 2� 105 cells/mL in RPMI 1640 with 10% FBS, with IL-3, with FLT3 ligand, or without cytokine for24 hours. Apoptosis was analyzed by Annexin V-FITC and PI followed by flow cytometric analysis. The experiments were repeated at least thrice. Percentage ofAnnexin V–positive apoptotic cells. Columns, mean of triplicate data points from one of two independent experiments; bars, SE. D, BaF3/FLT3/ITD cells were culturedseeded at 2 � 105 in six-well plates in RPMI 1640 with 10% FBS and indicated concentrations of AG1296 for 24 hours. Apoptosis was assayed as above.The experiments were repeated at least thrice and representative quadrants of flow cytometric analysis are shown. Percentage of Annexin V–positive cells.Columns, mean of triplicate data points from a representative experiment; bars, SE. E, BaF3/Flt3/WT cells were seeded at 2 � 105 in six-well plates in RPMI1640 with 10% FBS, 2 ng IL-3/mL, and AG1296 (5 Amol/L) for 24 hours. Apoptosis was assayed as described above. The experiments were repeated at leastthrice and representative quadrants of flow cytometric analysis are shown. Percentage of Annexin V–positive cells. Columns, mean of triplicate datapoints from representative experiment; bars, SE.

Cancer Research

Cancer Res 2005; 65: (16). August 15, 2005 7340 www.aacrjournals.org

then with horseradish peroxidase–conjugated secondary antibodies. Thesecondary antibodies were detected by enhanced chemiluminescence

(Amersham Biosciences, Piscataway, NJ). The same blots were stripped

and reprobed the blots with desired antibodies to confirm equal loading.

Immunoprecipitation kinase assay: BAD phosphorylation in vitro .Immunoprecipitation was done as described above. The agarose-protein A-

immunoprecipitate complex was washed twice with kinase buffer. Beads

were then suspended in 40 AL of 1� kinase buffer supplemented with 500

Amol/L ATP and 5 Ag soluble BAD (Upstate). After incubation at 30jC for30 minutes, the reaction was terminated with 20 AL SDS sample buffer.

Samples were heated to 95jC to 100jC for 5 minutes and analyzed by

Western blotting.

Results

FLT3/ITD prevents apoptosis induced by interleukin-3deprivation in BaF3 cells. To investigate the mechanism ofFLT3/ITD-mediated apoptosis inhibition, a full-length humanFLT3 cDNA with a 78-nucleotide ITD (encoding 26 amino acids;Fig. 1A) from an AML patient and WT human FLT3 cDNA werecloned into the retrovirus vector MSCV and transduced into BaF3cells. BaF3 cells depend on IL-3 for their continued growth andsurvival (39, 40). Cells transduced with vector or FLT3/WT couldnot survive in the absence of IL-3, whereas cells transduced withFLT3/ITD showed no significant growth rate difference with orwithout IL-3. To confirm constitutive activation of FLT3/ITD intransduced BaF3 cells, anti-phospho-FLT3 (Tyr591) polyclonal anti-body was used to detect the phosphorylation (activation) of FLT3/ITD. As shown in Fig. 1B (lanes 1 and 2), phospho-FLT3 (Tyr591)was detected in BaF3/FLT3/ITD cells but not in BaF3/FLT3/WTcells. As shown in Fig. 1C , BaF3/FLT3/WT cells underwentapoptosis after 24 hours of IL-3 deprivation. If FLT3 ligand wasadded to the culture medium, FLT3/WT cells remained viable,suggesting that FLT3/WT signaling prevents apoptosis. However,after deprivation of IL-3 or FLT3 ligand, BaF3/FLT3/ITD cellsremained viable.

AG1296, a FLT3 inhibitor (41), was used to inhibit the activationof FLT3/ITD. Exposure to 5 Amol/L AG1296 for 90 minutesessentially eliminated autophosphorylation of FLT3/ITD (Fig. 1B).BaF3/FLT3/ITD cells cultured in medium containing AG1296(2.5 and 5 Amol/L) for 24 hours underwent apoptosis in a dose-dependent manner, which could be reversed by IL-3 addition(Fig. 1D). AG1296 (5 Amol/L) did not result in apoptosis in BaF3/FLT3/WT cells in the presence of IL-3 (Fig. 1E), suggesting that thisdrug specifically inhibited FLT3/ITD signaling in the absence ofIL-3. These data suggest that FLT3/ITD substitutes for theantiapoptotic function of IL-3 in BaF3 cells.FLT3/ITD induced BAD phosphorylation at serine 112 and

Bcl-XL expression. Survival signals, such as IL-3, inhibit apoptosisvia BAD phosphorylation at Ser136 through the PI3K/Akt and Ser112

through the MEK/ERK and PKA pathways (17, 18, 20, 24). Wedetermined whether FLT3/ITD-mediated activation of one or moreof these pathways lead to BAD phosphorylation at Ser112 and/orSer136. To detect endogenous BAD phosphorylation in BaF3/FLT3/WT and BaF3/FLT3/ITD cell lines, we did immunoprecipitationand Western blot analysis using anti-phospho-Ser136-BAD–specificantibody, anti-phospho-Ser112-BAD–specific antibody and antibodyfor total BAD. As shown in Fig. 2A , when IL-3 or FLT3 ligand waspresent, endogenous BAD was phosphorylated at Ser112 in bothBaF3/FLT3/WT and BaF3/FLT3/ITD cells. However, when IL-3 orFLT3 ligand was removed, BAD remained phosphorylated at Ser112

in BaF3/FLT3/ITD cells but not in BaF3/FLT3/WT cells (Fig. 2A).No phosphorylation of endogenous BAD phosphorylation at Ser136

was detected by immunoprecipitation/Western blotting analysiswith anti-phospho-BAD (Ser136) antibody. In vitro phosphorylatedBAD was used as positive control.To confirm that the phosphorylation of BAD at Ser112 was due to

the functioning of FLT3/ITD, FLT3/ITD inhibitor AG1296 was usedto inhibit autoactivation of FLT3/ITD. The phosphorylation at Ser112

was eliminated by AG1296 (Fig. 2B). However, phosphorylation ofBAD at Ser136 was not detected in BaF3/FLT3/ITD cells (Fig. 2B)

Figure 2. FLT3/ITD induced BAD phosphorylation at Ser112 and Bcl-XL expression. A, BaF3/FLT3/WT and BaF3/FLT3/ITD cells were starved in RPMI 1640with 1% FBS for 8 hours and exposed to IL-3 or FLT3 ligand for 5 minutes. Immunoprecipitation/Western blotting were done to detect BAD phosphorylation withindicated antibodies. In vitro phosphorylated BAD was used as control in detecting phospho-BAD Ser136 (bottom, lane 1). B, BaF3/FLT3/ITD cells were starvedof serum as described above and then exposed to FLT3 inhibitor AG1296 (2.5, 5, and 10 Amol/L) for 90 minutes. Western blotting was carried out to detect thephosphorylation of BAD at Ser112. The same membranes were stripped and reprobed with anti-BAD antibody (for total BAD). Immunoprecipitation/Western blotting wasused to detect Ser136. The blots were stripped and reprobed anti-BAD. In vitro phosphorylated BAD was used as a control for detection phospho-BAD Ser136

(bottom, lane 1 ). C, BaF3/FLT3/WT cells were deprived of cytokines for 6, 12, and 24 hours. Western blotting was carried out to detect Bcl-XL with an anti-Bcl-XL

antibody. The same blot was stripped and reprobed with anti-Bcl-2 antibody to detect Bcl-2 and h-tubulin to verify equal loading. D, BaF3/FLT3/ITD cells weretreated with AG1296 (5 Amol/L) in standard medium for 0, 6, 12, 24, and 36 hours. Western blotting was used to detect the expression of Bcl-XL and Bcl-2 asdescribed above. The blots were reprobed with anti-h-tubulin to show equal loading. E, BAD was immunoprecipitated with anti–full-length BAD rabbit polyclonalantibody and protein G. Western blotting was used to detect the binding of Bcl-XL and Bcl-2. ITD lanes represent two BaF3/FLT3/ITD clones transfected with thesame construct.

FLT3/ITD Prevents Apoptosis via PKA and RSK1

www.aacrjournals.org 7341 Cancer Res 2005; 65: (16). August 15, 2005

and at a minimal level if the dephosphorylation was blocked bycalyculin A (data not shown). These data confirm our conclusionthat phosphorylation at Ser112 is the principal endogenous BADphosphorylation induced by FLT3/ITD and that BAD phosphoryla-tion at Ser112 (not Ser136) might be related to the antiapoptoticfunction of FLT3/ITD.BaF3/FLT3/WT cells depended on IL-3 for Bcl-XL expression.

When deprived of IL-3, Bcl-XL decreased in BaF3/FLT3/WT cells(Fig. 2C). BaF3/FLT3/ITD cells exhibited constitutive expression ofBcl-XL after IL-3 deprivation. Inhibition of FLT3/ITD by AG1296reduced the expression of Bcl-XL (Fig. 2D). Phosphorylation ofBAD leads to its binding to 14-3-3 protein and thus releases Bcl-XL and Bcl-2 from BAD/Bcl-2/Bcl-XL. Bcl-XL binds BAX andinhibits cytochrome c release from mitochondria, which is a keystep in apoptosis (31–35). In FLT3/ITD cells, BAD was shown tobe associated with Bcl-XL and, to a lesser degree, with Bcl-2(Fig. 2E), suggesting the involvement of the interaction of BADand other Bcl-2 family members during FLT3/ITD-mediated, IL-3-independent cell proliferation.FLT3/ITD activates RSK1 via MEK/ERK pathway, and RSK1

phosphorylates BAD at serine 112. The MEK/ERK pathwaymediates the antiapoptotic function of survival factors, such asIL-3 (17, 18). RSK1 has been shown to mediate BAD phosphor-ylation at Ser112 as a downstream effector of the MEK/ERKpathway after IL-3 stimulation (18). The MEK/ERK pathway hasbeen reported previously to be activated by FLT3/ITD in theabsence of IL-3 (13, 14). We have confirmed that BaF3/FLT3/WTcells depend on IL-3 or FLT3 ligand for the activation of theMEK/ERK pathway, although this pathway is constitutivelyactivated in BaF3/FLT3/ITD cells (Fig. 3A). Inhibition of FLT3/ITD by AG1296 abolished phosphorylation of MEK1/2 (Ser217/Ser221), ERK1/2 (Thr202/Tyr204), and RSK1 (Ser381; Fig. 3B). WhenBaF3/FLT3/ITD cells were treated with the MEK inhibitor U0126,phosphorylation on ERK1/2 (Thr202/Tyr204) and RSK1 (Ser381) was

eliminated (Fig. 3C), confirming that FLT3/ITD induced activationof RSK1 via the MEK/ERK pathway. At the concentration thateliminates phosphorylation of ERK1/2, U0126 reduced but did noteliminate the BAD phosphorylation at Ser112 (Fig. 3C), suggestingthe involvement of other pathway(s) in FLT3/ITD-induced BADphosphorylation at Ser112.To determine if RSK1 mediates BAD phosphorylation in BaF3/

FLT3/ITD cells, we conducted an immunoprecipitation/kinaseassay. RSK1 was immunoprecipitated from BaF3/FLT3/ITD cells

Figure 3. FLT3/ITD constitutively activates RSK1 via MEK/ERK and RSK1phosphorylates BAD at Ser112. A, BaF3/vector, BaF3/FLT3/WT, andBaF3/FLT3/ITD cells were starved in RPMI 1640 with 0.5% FBS for 8 hoursbefore stimulation with IL-3 (20 ng/mL) or FLT3 ligand for 5 minutes. Westernblotting was done to detect phospho-ERK1/2 (Thr202/Tyr204). The same blotswere stripped and reprobed with anti-ERK2 antibody (3A7 monoclonal antibody)to show equal loading. B, BaF3/FLT3/ITD were starved of serum as describedabove and treated with AG1296 (2.5, 5, and 10 Amol/L) for 90 minutes.Cells lysates were analyzed by Western blotting. The blots were probed withanti-phopho-MEK1/2 (Ser217/Ser221), phospho-ERK1/2 (Thr202/Tyr204), andphospho-p90RSK (Ser381). The same blots were stripped and reprobed withanti-MEK, anti-ERK2 (3A7 monoclonal antibody), and anti-RSK1 antibodies toshow equal loading. C, BaF3/FLT3/ITD cells were starved of serum as describedabove and treated with U0126 (12.5, 25, and 50 Amol/L) for 90 minutes. Cellslysates were analyzed by Western blotting. The blots were probed withanti-phospho-ERK1/2 (Thr202/Tyr204), anti-phospho-p90RSK (RSK1; Ser381),and phospho-BAD (Ser112). The blots were stripped and reprobed withanti-ERK2 (3A7 monoclonal antibody), anti-p90RSK (RSK1), and anti-BADantibodies to show equal loading. D, BaF3/FLT3/ITD cells lysates wereincubated with anti-RSK1 polyclonal antibody or IgG (control) for 2 hoursand then with protein A-agarose for 2 hours. The kinase activity ofimmunoprecipitates was assayed and Western blotting analysis was done todetect BAD phosphorylation at Ser112. Purified RSK1 (active) was used aspositive control in the reaction (lane 1 ). Rabbit IgG was used as negative control(lane 2). The exogenous BAD protein was excluded in one reaction (lane 3).E, BaF3/FLT3 cells were treated with U0126 (20 Amol/L) for 0, 12, 24, 36 hoursin RPMI 1640 with 10% FBS. Western blotting was used to detect the proteinlevel of Bcl-XL and Bcl-2. h-Tubulin was used to show equal loading.F, BaF3/FLT3/ITD cells were treated seeded in six-well plates at a density of2 � 105 cells/mL in RPMI 1640 with 10% FBS and with U0126 (0, 12.5, and25 Amol/L) for 24 hours and stained with Annexin V-FITC and PI. Apoptosis wasanalyzed by flow cytometry. The experiments were repeated at least thrice.Percentage of Annexin V–positive cells. Columns, mean of triplicate data pointsin one representative experiment; bars, SE.

Cancer Research

Cancer Res 2005; 65: (16). August 15, 2005 7342 www.aacrjournals.org

with anti-RSK1 polyclonal antibody and used to phosphorylateBAD in vitro . RSK1 from BaF3/FLT3/ITD cells had the ability tophosphorylate BAD at Ser112 (Fig. 3D), suggesting that RSK1 is aBAD kinase that is activated by FLT3/ITD via the MEK/ERKpathway as has been shown in Fig. 3B and C . When BAD proteinwas not included in the reaction mixture, no BAD phosphorylationwas noted, suggesting that BAD was phosphorylated by RSK1 inthe in vitro kinase reaction rather than having been immunopre-cipitated with RSK1.Our study excluded the role of MEK/ERK pathway in Bcl-XL

expression, which was constitutively induced by FLT3/ITD.Inhibition of MEK by U0126 did not reduce the protein level ofBcl-XL (Fig. 3E). We concluded that the MEK/ERK pathway mightcontribute to FLT3/ITD’s antiapoptotic signaling at least partiallythrough BAD phosphorylation by RSK1. Inhibition of the MEK/ERK/RSK1 pathway by U0126 induced apoptosis in BaF3/FLT3/ITDcells in a dose-dependent manner (Fig. 3F). Compared with theFLT3 inhibitor AG1296 (Fig. 1D), U0126 had much less effect oninducing apoptosis, indicating that other pathways are involved inFLT3/ITD’s antiapoptotic function.FLT3/ITD activates PKA and PKA phosphorylates BAD at

Ser112. Because the MEK/ERK pathway only partially mediatedFLT3/ITD-induced BAD phosphorylation at Ser112 (Fig. 3D), weattempted to find other pathways that might lead to BADphosphorylation at Ser112. PKA was activated in BaF3/FLT3/ITDcells but not in BaF3/FLT3/WT cells after IL-3 deprivation(Fig. 4A). The PKA inhibitor H-89 (2.5, 5, and 10 Amol/L) reducedthe phosphorylation level of BAD at Ser112 (Fig. 4B). To confirmthat PKA mediated BAD phosphorylation in BaF3/FLT3/ITD cells,we showed that immunoprecipitated PKA from BaF3/FLT3/ITDcells had the ability to phosphorylate BAD at Ser112 with a BADabsent control, indicating that the phosphorylated BAD was notfrom immunoprecipitation (Fig. 4C). We also excluded the role ofPKA in Bcl-XL expression because the PKA inhibitor H-89 (5 Amol/L)did not affect the protein level of Bcl-XL (Fig. 4D). As expected, thePKA inhibitor H-89 (2.5 and 5 Amol/L) induced apoptosis in BaF3/FLT3/ITD cells in a dose-dependent manner (Fig. 4E).RNA interference for RSK1 or PKAc reduced BAD phos-

phorylation. To confirm our conclusion that RSK1 and PKAcontribute to FLT3/ITD-induced BAD phosphorylation at Ser112,we used siRNA SMARTpool to partially shutdown the expression ofRSK1 and PKAc and checked the phospho-BAD (Ser112) level andapoptosis in MV4-11, a cell line harboring FLT3/ITD. As shown inFig. 5, RNA interference of RSK1 or PKAc reduced BADphosphorylation (Ser112) and induced apoptosis, thereby confirm-ing the results noted in BaF3/FLT3/ITD cells when inhibitors ofthese pathways were used (Figs. 3 and 4).U0126 and H-89 synergistically induced apoptosis in BaF3/

FLT3/ITD cells. As shown above, FLT3/ITD induced BADphosphorylation predominantly at Ser112 (Fig. 2). The MEK/ERKpathway mediated BAD phosphorylation at Ser112 via RSK1(Fig. 3) and PKA also mediated phosphorylation at the same site(Fig. 4). Inhibition of either MEK (Fig. 3) or PKA (Fig. 4) induced alesser degree of apoptosis compared with that induced by theFLT3 inhibitor AG1296 (Fig. 1D). To determine if MEK/ERK andPKA pathways are the principal pathways responsible forantiapoptotic function of FLT3/ITD, we blocked the two pathwaysat the same time. The MEK inhibitor U0126 and the PKAinhibitor H-89 synergistically induced apoptosis to a levelcomparable with that induced by AG1296 (5 Amol/L; Fig. 6Aand C), suggesting that MEK/ERK and PKA pathways are required

Figure 4. FLT3/ITD constitutively activates PKA and PKA phosphorylates BADat Ser112. A, BaF3/FLT3/WT and BaF3/FLT3/ITD cells were deprived of IL-3for 8 hours. The PKA kinase activity in cell extracts was assessed using thePep-Tag Assay for Nonradioactive Detection of cAMP-Dependent ProteinKinase kit. The phosphorylated and nonphosphorylated Pep-Tag peptides wereseparated by electrophoresis on 0.8% agarose gel and visualized with UV.B, BaF3/FLT3/ITD cells were placed in RPMI 1640 with 0.5% FBS for 8 hoursand then treated with H-89 (2.5, 5, and 10 Amol/L) for 90 minutes. Westernblotting was done to detect phospho-BAD (Ser112). The blots were stripped andreprobed with anti-BAD antibodies to show equal loading. C, BaF3/FLT3/ITDcell lysates were incubated with anti-PKAc polyclonal antibody or IgG (control)for 2 hours and then with protein A-agarose for 2 hours. The kinase activity ofimmunoprecipitates was assayed and Western blotting analysis was done todetect the phosphorylation of BAD at Ser112. Purified PKA (active) was used aspositive control (lane 1). Rabbit IgG was used as negative control (lane 2).The exogenous BAD protein was excluded in one reaction (lane 3). D, BaF3/FLT3/ITD cells were treated with H-89 (5 Amol/L) for 0, 12, 24, and 36 hoursin RPMI 1640 with 10% FBS and Western blotting was used to detected theprotein level of Bcl-XL and Bcl-2. h-Tubulin was used to show equal loading.E, BaF3/FLT3/ITD cells were treated with H-89 (0, 2.5, and 5 Amol/L)for 24 hours in RPMI 1640 with 10% FBS and stained with Annexin V-FITC andPI. Apoptosis was analyzed by flow cytometric analysis. The experimentswere repeated at least thrice. Percentage of Annexin V–positive cells.Columns, mean of triplicate data points from a representative experiment;bars, SE.

FLT3/ITD Prevents Apoptosis via PKA and RSK1

www.aacrjournals.org 7343 Cancer Res 2005; 65: (16). August 15, 2005

for the antiapoptotic function of FLT3/ITD. U0126 or H-89, aloneor in combination, did not reduce the expression of Bcl-XL (Figs.3, 4, and 6B); yet, the combination of these two drugs reducedmost of the BAD phosphorylation at Ser112 (Fig. 6B).U0126 and/or H-89 enhanced the efficacy of FLT3 inhibitor

AG1296 on inducing apoptosis in BaF3/FLT3/ITD cells. Wetested the inhibition of FLT3/ITD in combination with inhibition ofits downstream MEK/ERK or/and PKA pathways in BaF3/FLT3/ITD cells. The FLT3 inhibitor AG1296 (2.5 and 5 Amol/L) wascombined with U0126 (10 and 20 Amol/L) or H-89 (5 Amol/L) in thetreatment of BaF3/FLT3/ITD cells for 24 hours. Apoptosis, BADphosphorylation, and Bcl-XL expression were each comparedamong different combinations. As shown in Fig. 6A , combinationof the FLT3 inhibitor AG1296 (2.5 Amol/L) with the MEK inhibitorU0126 (10 Amol/L) or/and the PKA inhibitor H-89 (5 Amol/L)greatly enhanced the efficacy of these drugs on inducing apoptosisin BaF3/FLT3/ITD cells, especially when all three drugs werecombined. As we have described above, the combination of U0126

with H-89 reduced most of the BAD phosphorylation at Ser112

(Fig. 6B) and synergistically induced apoptosis (Fig. 6A and C).

U0126 (10 Amol/L) or/and H-89 (5 Amol/L) combined with AG1296

(2.5 Amol/L) further reduced BAD phosphorylation but to a lesser

degree (Fig. 6B). However, AG1296 (2.5 Amol/L) significantly

reduced the protein level of Bcl-XL (Fig. 6B), suggesting that the

synergistic effects of U0126 or/and H-89 with AG1296 might be due

to the inhibition of BAD phosphorylation and/or Bcl-XL expression.

To show that coinhibition of the PKA and MEK pathways was

at least additive in other systems, we did a similar experiment in

a FLT3/ITD-bearing human leukemia cell line, MV4-11 (Fig. 6D),

and found similar results to those obtained in the Baf3/FLT3/ITD

cells (Fig. 6A).

Discussion

Activating mutations in FLT3 due to an ITD in the juxtamem-brane domain or a point mutation at D835 in the activation loopresult in constitutive activation of this RTK (2–6). Constitutivelyactivated FLT3 leads to IL-3-independent growth and activatesMAPK, PI3K, and STAT5 in 32D and BaF3 cells (12–14). Theinhibition of apoptosis by IL-3 is mediated by the MEK/ERK, PI3K,and PKA pathways (17–24). BAD phosphorylation plays animportant role in the antiapoptotic signaling mediated by IL-3(17, 18, 20, 21, 24). However, the mechanisms involved in theantiapoptotic functions of FLT3/ITD are still largely unknown. Inthis study, we introduced a human FLT3 gene with an ITDmutation into BaF3 cells and investigated the effects of the ITDmutation on cell survival and the mechanism thereof. Our resultsshow that FLT3/ITD leads to constitutive activation of MEK/ERKand PKA pathways and that these pathways are required for IL-3-independent survival. FLT3/ITD induces BAD phosphorylation atSer112 and constitutive expression of Bcl-XL. RSK1 and PKA areresponsible for the endogenous BAD phosphorylation at Ser112,which is the predominant phosphorylation site induced by FLT3/ITD. Therefore, inhibition of both MEK/ERK/RSK1 and PKApathways synergistically induces apoptosis (Fig. 6). The proposedmechanism of FLT3/ITD-mediated activation of the MEK/ERK/RSK1 and PKA pathways to phosphorylate BAD and inhibitapoptosis is summarized in Fig. 7.A recent study (38) shows that FLT3/ITD induces phosphor-

ylation of BAD at Ser136 via the PI3K pathway when BAD isoverexpressed. In this study, we have detected endogenousphosphorylation of BAD predominantly at Ser112 and muchlower level of endogenous phosphorylation of BAD at Ser136 evenwhen the dephosphorylation of BAD was inhibited. Experiments

Figure 5. RNA interference of RSK1 and PKA. MV4-11 cells were transfected with siRNAs for RSK1 and PKAc twice at 0 and 24 hours. Cells were harvested at72 hours for analysis. Three independent experiments were done. Representative results. A, the expression level of RSK1 and PKAc was assayed by immunostainingand flow cytometry analysis. Cells were fixed, permeabilized, and stained with anti-RKS1 or anti-PKAc monoclonal primary antibody and PE anti-mouse secondaryantibody. Filled histograms, RNA interference–negative controls. B, the phospho-BAD (Ser112) level was assayed by immunostaining and flow cytometry analysis.The cells were treated as above. Anti-pBAD (Ser112) rabbit polyclonal primary antibody and FITC anti-rabbit secondary antibody were used in the staining.Filled histograms, RNA interference–negative controls. C, cells were harvested and stained with Annexin V-FITC and PI. Apoptosis was analyzed by flow cytometryanalysis. The experiments were repeated at least thrice. Percentage of Annexin V–positive cells. Columns, mean of triplicate data points from one representativeexperiment; bars, SE.

Cancer Research

Cancer Res 2005; 65: (16). August 15, 2005 7344 www.aacrjournals.org

Figure 6. Combination of AG1296, U0126, and H-89synergistically induced apoptosis in BaF3/FLT3/ITDcells. A, BaF3/FLT3/ITD cells were treated with drugs(U0126 10 Amol/L, H-89 5 Amol/L, and AG12962.5 Amol/L) in RPMI 1640 with 10% FBS in differentcombinations as indicated for 24 hours. Apoptosis wasassayed as described. The experiments were repeatedat least thrice. Percentage of Annexin V–positive cells.Columns, mean of triplicate data points from onerepresentative experiment; bars, SE. B, BaF3/FLT3/ITD cells were treated as described above and Westernblotting was done to detect the phosphorylation of BADat Ser112 and the protein levels of Bcl-XL and Bcl-2.Intensity was standardized with h-actin. C, BaF3/FLT3/ITD cells were treated with above combinations but withhigher concentrations of U0126 (20 Amol/L) andAG1296 (5 Amol/L) for 24 hours. Apoptosis wasanalyzed as described above. Percentage of AnnexinV–positive cells. Columns, mean of triplicate data pointsfrom one representative experiment. D, MV4-11 cellswere treated with U0126 (10 Amol/L), H-89 (10 Amol/L),and AG1296 (2.5 Amol/L) in RPMI 1640 with 10% FBSin different combinations as indicated for 48 hours.Apoptosis was assayed as described. The experimentwas conducted in triplicate on two occasions.Percentage of Annexin V–positive cells. Columns,mean of triplicate data points in one of the twoexperiments.

FLT3/ITD Prevents Apoptosis via PKA and RSK1

www.aacrjournals.org 7345 Cancer Res 2005; 65: (16). August 15, 2005

that rely on overexpressed BAD may not reflect the situation inthe cell, because overexpression may artificially increase thechance of BAD to interact with kinases, such as Akt.Nevertheless, the PI3K inhibitor LY294002 potentiates the abilityof the MEK inhibitor U0126 to induce apoptosis in BaF3/FLT3/ITD cells (38), suggesting that the PI3K/Akt pathway has somerole in the antiapoptotic function of FLT3/ITD but notnecessarily via BAD phosphorylation. Our studies detected alow level of BAD phosphorylation at Ser136 only after BADphosphatase was inhibited, indicating PI3K pathway was not themain pathway in BAD phosphorylation.Our results suggest that the MEK/ERK and PKA pathways are the

two principal pathways for BAD phosphorylation and playimportant roles in antiapoptotic function of FLT3/ITD but do notexclude the involvement of other pathways. For example, when theMEK inhibitor U0126 (10 Amol/L) and PKA inhibitor H-89 (5 Amol/L)were combined, only 60% of FLT3/ITD cells underwent apoptosis(Fig. 6A). As shown in Fig. 2, AG1296 inhibition of FLT3/ITD reducedthe expression of Bcl-XL. However, U0126 or H-89 did not reduce theexpression of Bcl-XL (Figs. 3 and 4). The interaction of thesepathways is diagramed in Fig. 7.Our data strongly support that FLT3/ITD might play a crucial

role in leukemogenesis by constitutive signaling through both

MEK/ERK and PKA pathways that in turn suppress apoptosis andpromote cell survival. Specific FLT3 inhibitors are currently underdevelopment as therapeutic agents for AML, which couldpotentially benefit those patients whose blasts show activatingmutations in FLT3. Understanding the downstream signalingpathways of FLT3/ITD could help to dissect the pathogenesis ofAML as well as to develop new antileukemic agents. Inhibitorscurrently used as clinical drugs, such as PKC412 (42), CT53518 (43),and SU5614 (44), have been shown to inhibit autophosphorylationof FLT3/ITD. Blocking downstream signaling could be an effectiveway to enhance the efficacy of these tyrosine kinase inhibitors, evenafter drug resistance begins to develop. Our results show that thecombination of the FLT3 inhibitor AG1296 with the MEK inhibitorU0126 and/or the PKA inhibitor H-89 greatly enhanced the efficacyof AG1296 on inducing apoptosis. This strategy may represent anovel approach to treat AML patients with FLT3/ITD mutations.

Acknowledgments

Received 6/30/2004; revised 4/26/2005; accepted 5/19/2005.Grant support: 2PO1CA66996-06.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Figure 7. Diagram of the mechanism of the antiapoptotic functionof FLT3/ITD. FLT3/ITD activates MEK/ERK/RSK1 and PKApathways and leads to BAD phosphorylation at Ser112.Phospho-BAD is sequestered by protein 14-3-3 and thus releasesBcl-XL from BAD-Bcl-XL complex. Bcl-XL binds BAX andinhibits the release of cytochrome c from mitochondria, which isa key step in apoptosis.

References1. Reilly JT. Class III receptor tyrosine kinases: role inleukaemogenesis. Br J Haematol 2002;116:744–57.

2. Nakao M, Yokota S, Iwai T, et al. Internal tandemduplication of the flt3 gene found in acute myeloidleukemia. Leukemia 1996;10:1911–8.

3. Yokota S, Kiyoi H, Nakao M, et al. Internal tandemduplication of the FLT3 gene is preferentially seen inacute myeloid leukemia and myelodysplastic syndromeamong various hematological malignancies. A study ona large series of patients and cell lines. Leukemia 1997;11:1605–9.

4. Xu F, Taki T, Yang HW, et al. Tandem duplication of theFLT3 gene is found in acute lymphoblastic leukaemia aswell as acute myeloid leukaemia but not in myelodys-

plastic syndrome or juvenile chronic myelogenousleukaemia in children. Br J Haematol 1999;105:155–62.

5. Abu-Duhier FM, Goodeve AC, Wilson GA, Care RS,Peake IR, Reilly JT. Identification of novel FLT-3 Asp835mutations in adult acute myeloid leukaemia. Br JHaematol 2001;113:983–8.

6. Yamamoto Y, Kiyoi H, Nakano Y, et al. Activatingmutation of D835 within the activation loop of FLT3 inhuman hematologic malignancies. Blood 2001;97:2434–9.

7. Kiyoi H, Naoe T, Nakano Y, et al. Prognosticimplication of FLT3 and N-RAS gene mutations in acutemyeloid leukemia. Blood 1999 May 1;93:3074–80.

8. Abu-Duhier FM, Goodeve AC, Wilson GA, et al. FLT3internal tandem duplication mutations in adult acutemyeloid leukaemia define a high-risk group. Br JHaematol 2000;111:190–5.

9. Meshinchi S, Woods WG, Stirewalt DL, et al.Prevalence and prognostic significance of Flt3 internaltandem duplication in pediatric acute myeloid leuke-mia. Blood 2001;97:89–94.

10. Whitman SP, Archer KJ, Feng L, et al. Absence ofthe wild-type allele predicts poor prognosis in adultde novo acute myeloid leukemia with normal cytoge-netics and internal tandem duplication of FLT3:a Cancer and Leukemia Group B Study. Cancer Res2001;61:7233–9.

11. Kiyoi H, Towatari M, Yokota S, et al. Internal tandemduplication of the FLT3 gene is a novel modality ofelongation mutation which causes constitutive activa-tion of the product. Leukemia 1998;12:1333–7.

12. Fenski R, Flesch K, Serve S, et al. Constitutiveactivation of FLT3 in acute myeloid leukaemia and its

Cancer Research

Cancer Res 2005; 65: (16). August 15, 2005 7346 www.aacrjournals.org

consequences for growth of 32D cells. Br J Haematol2000;108:322–30.

13. Hayakawa F, Towatari M, Kiyoi H, et al. Tandem-duplicated Flt3 constitutively activates STAT5 and MAPkinase and introduces autonomous cell growth in IL-3-dependent cell lines. Oncogene 2000;19:624–31.

14. Mizuki M, Fenski R, Halfter H, et al. Flt3 mutationsfrom patients with acute myeloid leukemia inducetransformation of 32D cells mediated by the Ras andSTAT5 pathways. Blood 2000;96:3907–14.

15. Zhao M, Kiyoi H, Yamamoto Y, et al. In vivo treatmentof mutant FLT3-transformed murine leukemia with atyrosine kinase inhibitor. Leukemia 2000;14:374–8.

16. Kelly LM, Liu Q, Kutok JL, Williams IR, Boulton CL,Gilliland DG. FLT3 internal tandem duplication muta-tions associated with human acute myeloid leukemiasinduce myeloproliferative disease in a murine bonemarrow transplant model. Blood 2002;99:310–8.

17. Leverrier Y, Thomas J, Perkins GR, Mangeney M,Collins MK, Marvel J. In bone marrow derived Baf-3 cells, inhibition of apoptosis by IL-3 is mediatedby two independent pathways. Oncogene 1997;14:425–30.

18. Shimamura A, Ballif BA, Richards SA, Blenis J. Rsk1mediates a MEK-MAP kinase cell survival signal. CurrBiol 2000;10:127–35.

19. Deng X, Ruvolo P, Carr B, May WS Jr. Survivalfunction of ERK1/2 as IL-3-activated, staurosporine-resistant Bcl2 kinases. Proc Natl Acad Sci U S A 2000;97:1578–83.

20. del Peso L, Gonzalez-Garcia M, Page C, Herrera R,Nunez G. Interleukin-3-induced phosphorylation ofBAD through the protein kinase Akt. Science 1997;278:687–9.

21. Mathieu AL, Gonin S, Leverrier Y, et al. Activation ofthe phosphatidylinositol 3-kinase/Akt pathway protectsagainst interleukin-3 starvation but not DNA damage-induced apoptosis. J Biol Chem 2001;276:10935–42.

22. Leverrier Y, Thomas J, Mathieu AL, Low W, BlanquierB, Marvel J. Role of PI3-kinase in Bcl-X induction and

apoptosis inhibition mediated by IL-3 or IGF-1 in Baf-3cells. Cell Death Differ 1999;6:290–6.

23. Yamaguchi H, Wang HG. The protein kinase PKB/Aktregulates cell survival and apoptosis by inhibiting Baxconformational change. Oncogene 2001;20:7779–86.

24. Harada H, Becknell B, Wilm M, et al. Phosphorylationand inactivation of BAD by mitochondria-anchoredprotein kinase A. Mol Cell 1999;3:413–22.

25. Dumon S, Santos SC, Debierre-Grockiego F, et al. IL-3dependent regulation of Bcl-XL gene expression bySTAT5 in bone marrow derived cell line. Oncogene 1999;18:4191–9.

26. Cory S. Regulation of lymphocyte survival by the bcl-2 gene family. Annu Rev Immunol 1995;13:513–43.

27. Adams JM, Cory S. The Bcl-2 protein family: arbitersof cell survival. Science 1998;281:1322–6.

28. Chao DT, Korsmeyer SJ. BCL-2 family: regulators ofcell death. Annu Rev Immunol 1998;16:395–419.

29. Datta SR, Dudek H, Tao X, et al. Akt phosphorylationof BAD couples survival signals to the cell-intrinsicdeath machinery. Cell 1997;91:231–41.

30. Pastorino JG, Tafani M, Farber JL. Tumor necrosisfactor induces phosphorylation and translocation ofBAD through a phosphatidylinositide-3-OH kinase-dependent pathway. J Biol Chem 1999;274:19411–6.

31. Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ.Serine phosphorylation of death agonist BAD inresponse to survival factor results in binding to 14-3-3not BCL-X(L). Cell 1996;87:619–28.

32. Kelekar A, Chang BS, Harlan JE, Fesik SW, ThompsonCB. Bad is a BH3 domain-containing protein that formsan inactivating dimer with Bcl-XL. Mol Cell Biol 1997;17:7040–6.

33. Ottilie S, Diaz JL, Horne W, et al. Dimerizationproperties of human BAD. Identification of a BH-3domain and analysis of its binding to mutant BCL-2 andBcl-XL proteins. J Biol Chem 1997;272:30866–72.

34. Yano S, Tokumitsu H, Soderling TR. Calciumpromotes cell survival through CaM-K kinase activationof the protein-kinase-B pathway. Nature 1998;396:584–7.

35. Wang HG, Pathan N, Ethell IM, et al. Ca2+-inducedapoptosis through calcineurin dephosphorylation ofBAD. Science 1999;284:339–43.

36. Nosaka T, Kawashima T, Misawa K, Ikuta K, Mui AL,Ktamura T. STAT5 as a molecular regulator ofproliferation and apoptosis in hematopoietic cells.EMBO J 1999;18:4754–65.

37. Birkenkamp KU, Geugien M, Lemmink HH, Krui jerW, Vellenga E. Regulation of constitutive STAT5phosphorylation in acute myeloid leukemia blasts.Leukemia 2001;15:1923–31.

38. Minami Y, Yamamoto K, Kiyoi H, Ueda R, Saito H,Naoe T. Different antiapoptotic pathways between wild-type and mutated FLT3: insights into therapeutic targetsin leukemia. Blood 2003;102:2969–75.

39. Rodriguez-Tarduchy G, Collins M, Lopez-Rivas A.Regulation of apoptosis in interleukin-3-dependenthemopoietic cells by interleukin-3 and calcium iono-phores. EMBO J 1990;9:2997–3002.

40. Collins MK, Marvel J, Malde P, Lopez-Rivas A.Interleukin 3 protects murine bone marrow cells fromapoptosis induced by DNA damaging agents. J Exp Med1992;176:1043–51.

41. Tse KF, Novelli E, Civin CI, Bohmer FD, Small D.Inhibition of FLT3-mediated transformation by useof a tyrosine kinase inhibitor. Leukemia 2001;15:1001–10.

42. Weisberg E, Boulton C, Kelly LM, et al. Inhibition ofmutant FLT3 receptors in leukemia cells by the smallmolecule tyrosine kinase inhibitor PKC412. Cancer Cell2002;1:433–43.

43. Kelly LM, Yu JC, Boulton CL, et al. CT53518, anovel selective FLT3 antagonist for the treatment ofacute myelogenous leukemia (AML). Cancer Cell 2002;1:421–32.

44. Spiekermann K, Dirschinger RJ, Schwab R, et al. Theprotein tyrosine kinase inhibitor SU5614 inhibits FLT3and induces growth arrest and apoptosis in AML-derived cell lines expressing a constitutively activatedFLT3. Blood 2003;101:1494–504.

FLT3/ITD Prevents Apoptosis via PKA and RSK1

www.aacrjournals.org 7347 Cancer Res 2005; 65: (16). August 15, 2005