Constitutive activation of Flt3 and STAT5A enhances self-renewal and alters differentiation of hematopoietic stem cells

Experimental Hematology 35 (2007) 105–116

Constitutive activation of Flt3 and STAT5A enhances self-renewal and alters differentiation of hematopoietic stem cells Malcolm A.S. Moore, David C. Dorn, Jan Jacob Schuringa*, Ki Young Chung, and Giovanni Morrone** Moore Laboratory, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Objective. To model human leukemogenesis by transduction of human hematopoietic stem cells (HSC) with genes associated with leukemia and expressed in leukemic stem cells. Methods. Constitutive activation of Flt3 (Flt3-ITD) has been reported in 25 to 30% of patients with acute myeloid leukemia (AML). Retroviral vectors expressing constitutively activated Flt3 and STAT5A were used to transduce human cord blood CD34+ cells and HSC cell selfrenewal and differentiation were evaluated. Results. We have demonstrated that retroviral transduction of Flt3 mutations into CD34+ cells enhanced HSC self-renewal as measured in vitro in competitive stromal coculture and limiting-dilution week-2 cobblestone (CAFC) assays. Enhanced erythropoiesis and decreased myelopoiesis were noted together with strong activation of STAT5A. Consequently, transduction studies were undertaken with a constitutively active mutant of STAT5A (STAT5A[1*6]) and here also a marked, selective expansion of transduced CD34+ cells was noted, with a massive increase in self-renewing CAFC detectable at both 2 and 5 weeks of stromal coculture. Differentiation was biased to erythropoiesis, including erythropoietin independence, with myeloid maturation inhibition. The observed phenotypic changes correlated with differential gene expression, with a number of genes differentially regulated by both the Flt3 and STAT5A mutants. These included upregulation of genes involved in erythropoiesis and downregulation of genes involved in myelopoiesis. The phenotype of week-2 self-renewing CAFC also characterized primary Flt3-ITD+ AML bone marrow samples. Isolation of leukemic stem cells (LSC) with a CD34+, CD38L, HLA-DRL phenotype was undertaken with Flt3-ITD+ AML samples resulting in co-purification of early CAFC. Gene expression of LSC relative to the bulk leukemic population revealed upregulation of homeobox genes (HOXA9, HOXA5) implicated in leukemogenesis, and hepatic leukemia factor (HLF) involved in stem cell proliferation. Conclusion. Myeloid leukemogenesis is a multi-stage process that can involve constitutively activated receptors and downstream pathways involving STAT5, HOX genes, and HLF. Ó 2007 International Society for Experimental Hematology. Published by Elsevier Inc.

The cytokine receptor Flt3 is a member of the type II receptor tyrosine kinase subfamily and is expressed on the majority of human CD34þ cells, including CD34þ, CD38, NOD/SCID repopulating hematopoietic stem cells (SRC) (reviewed [1,2]). Flt3 ligand (FL) is a type I transmembrane protein expressed by stromal cells in the bone marrow.

*Currently at the Department of Hematology, The University Medical Center, Groningen, The Netherlands. **Currently at the University of Catanzaro Magna Graecia, Catanzaro, Italy. Offprint requests to: Malcolm A.S. Moore, D. Phil, Moore Laboratory, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021; E-mail: [email protected]

In combination with c-kit ligand (KL), thrombopoietin (TPO), and interleukin-6 (IL-6), FL supported extensive, long-term (w3 months) expansion of CD34þ cells, progenitors, and SRC in suspension culture of cord blood CD34þ cells [3]. FL binding to the Flt3 receptor induces homodimerization and coupling of cytoplasmic domains leading to transphosphorylation of specific tyrosine residues in the JM domain and subsequent autophosphorylation in the tyrosine kinase domain [2]. A number of signal transduction pathways involved in proliferation and apoptosis are activated. Weak JAK-independent phosphorylation of STAT5 by wild-type (WT) Flt3 and subsequent DNA binding has been reported in one study [4] but this has not been confirmed [5,6]. Somatic mutations of Flt3 occur in w30%

0301-472X/07 $–see front matter. Copyright Ó 2007 International Society for Experimental Hematology. Published by Elsevier Inc. doi: 10.1016/j.exphem.2007.01.018

106

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

of AML patients, most frequently as an in-frame internal tandem duplication (ITD) at the juxtamembrane domain of the receptor [7,8]. The mutation is detected in all acute myeloid leukemia (AML) FAB subtypes with the highest frequency in M3 and the least in M2 [9]. Less frequently, activating point mutations occur that also lead to constitutive activation and these are found in w7% of AML [10]. The downstream signaling pathway of mutated Flt3 differs from that of the ligand-stimulated WT Flt3. In a gene array comparison, 767 of 6586 genes differed in expression in a comparison of cells expressing the WT versus mutated receptor [11]. Genes regulated by Flt3-ITD resembled more closely those of the IL-3 receptor signaling pathway. Pim2 was shown to be important for clonal growth of 32D cells. PU.1 and C/EBPa were induced by ligand activation of the WT receptor but inhibited by Flt3 mutants [1,2]. Thus, while partially mimicking the IL3R pathway, other parts of the Flt3 mutant signaling pathway may be novel and can antagonize the differentiation-inducing effects of the WT receptor. FL stimulation increased phosphorylation of the Src kinase Lyn, and this pathway may be involved in the proliferative effects of Flt3 mutations [12]. Flt3-ITD also prevents apoptosis in IL-3-deprived BaF3 cells due to protein kinase A and ribosomal S6 kinase-1–mediated Bad phosphorylation [13]. Flt3-ITD constitutively activates Akt, and Akt phosphorylation was found in myeloid blasts of 86% of patients, indicating an important role in AML [14]. Akt is necessary for increased survival, proliferation, and leukemic transformation by Flt3-ITD, possibly by inactivating the Foxo transcription factors. Flt3-ITD may cooperate with Wnt signaling in leukemic signal transduction. Flt3-ITD expression in 32D cells elevated expression of Frizzled-4, a Wnt ligand receptor, and also elevated bcatenin protein expression and induced the Wnt target cMyc [15]. The majority of AML samples with Flt3-ITD expressed high b-catenin levels, whereas WT cases did not. Introduction of Flt3-ITD into mouse hematopoietic stem/progenitor cells blocks maturation and induces a lethal myeloproliferative disease (MPD) characterized by a marked neutrophil leukocytosis, extramedullary hematopoiesis, and death at 40 to 60 days [16]. Flt3-ITD and Flt3-TKD mutants produce different diseases in mouse models, the former an oligoclonal MPD, and the latter an oligoclonal lymphoid disorder with longer latency and distinct hematological manifestation [6]. Flt3-TKD did not induce strong STAT5 pathway activation in contrast to the Flt3-ITD, suggesting differences in cell signaling pathways mediated by different Flt3 mutations [17]. Evidence for a two-hit model of leukemogenesis [1] was provided by transduction of Flt3 mutants plus AML1-ETO in a mouse BM transplant model resulting in development of acute leukemia [18]. We have shown that retroviral transduction of cord blood (CB) CD34þ stem/progenitor cells with Flt3-ITD mutants provides a proliferative advantage to these cells as revealed

in competitive proliferation assays in long-term stromal cocultures [19]. Differentiation was biased to erythroid relative to myeloid in both clonogenic assay and by phenotypic analysis of culture suspension cells. The appearance of a high frequency of cobblestone areas developing beneath the stroma by 7 to 14 days characterized Flt3-ITDtransduced cells since comparable cobblestone formation was absent in control cultures. Early cobblestone areas expressed stem cell properties due to their ability to undergo secondary recloning. The downstream signaling pathway of Flt3-ITD-transduced cells revealed a panel of genes that were both upregulated and downregulated, and strong activation of STAT5A. We developed a constitutively activated STAT5A(1*6) vector to transduce CB CD34þ cells and observed a phenotype that shared a number of features to that observed with Flt3-ITD; namely, excessive production of erythroid at the expense of myeloid cells, excess of BFUE relative to CFU-GM, and a high frequency of early CAFC that could undergo serial passage [20]. In addition (and in contrast to the Flt3-ITD studies), there was also a high frequency of late week-5 CAFC. In this report we further evaluate the proliferation and differentiation potential and the gene expression profile of CD34þ cells transduced with Flt3-ITD and STAT5(1*6). We compare these features to primary AML cells, with or without Flt3 mutations, and to CD34þ, CD38, and HLA-DR leukemic stem cells (LSC) isolated from AML patients with Flt3-ITD.

Materials and methods Primary cell isolation Neonatal cord blood was obtained from healthy full-term pregnancies from the Cord Blood Bank of the New York Blood Center. Human CD34þ cells (O95% pure) were derived from cord blood using MiniMACS (Miltenyi Biotech, Auburn, CA, USA). Ficoll-separated mononuclear cells were obtained from the peripheral blood and bone marrow of AML patients with and without Flt3 mutations under an Institutional Review Board–approved clinical protocol at Memorial Sloan-Kettering Cancer Center. AML cells were further fractionated by sterile FACS sorting into subsets differentially expressing CD34, CD38, or HLA-DR. Cell culture and cell lines H29 cells were cultured in Dulbecco’s modified Eagle’s medium (DME, Gibco-Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS, HyClone Laboratories, Logan, UT, USA), penicillin, streptomycin, 200 mM glutamine, 2 mg/mL tetracyclin, 300 mg/mL geneticin, and 2 mg/mL puromycin (all obtained from Gibco-Life Technologies). PG13 cells were grown in DME supplemented with 10% FBS, penicillin and streptomycin, and 200 mM glutamine (Sigma, St. Louis, MO,

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

USA) MS5, OP9 and aorta-gonad-mesonephros-S2 (AGMS2) murine stromal cells were propagated in a-Eagle minimum essential medium (aMEM, Gibco-Life Technologies) supplemented with 10% FBS, penicillin and streptomycin and 200 mM glutamine. For MS5 coculture experiments and long-term culture-initiating cell (LTC-IC) assays, cells were grown in aMEM supplemented with heat-inactivated 12.5% FBS, heat-inactivated 12.5% horse serum (HyClone, Logan, UT, USA), penicillin and streptomycin, 200 mM glutamine, 57.2 mM b-mercaptoethanol, and 1 mM hydrocortisone (LTC medium). For OP9 coculture experiments cells were grown in aMEM supplemented with 12.5% FBS, 12.5% horse serum, penicillin and streptomycin, 200 mM glutamine, 57.2 mM b-mercaptoethanol, and 50 mg/mL vitamin C. For AGM-S2 coculture experiments cells were grown in aMEM supplemented with 12.5% FBS, 12.5% horse serum, penicillin and streptomycin, 200 mM glutamine, and 57.2 mM b-mercaptoethanol. Retroviral vectors and constructs A murine stem cell virus (MSCV)-based retroviral expression vector (MIGR1) was used for all experiments. The MIGR1 retroviral backbone vector (kindly provided by Dr. Warren Pear, The University of Pennsylvania), contains a multi-cloning site upstream of an encephalo-myelocarditis virus (EMCV)-derived internal ribosomal entry site (IRES2) element and the coding sequence for the enhanced green fluorescent protein (EGFP) downstream of the latter. MIGR1 constructs carrying the human Flt3, W51 and W81 Flt3-ITD coding sequences were kindly provided by Dr. Gary Gilliland (Harvard Institute of Medicine). The constitutively active mutant murine STAT5A(1*6) was subcloned from pMXpuro-STAT5(1*6) into the EcoRI-SalI sites of the pIRES2-EGFP vector (Clontech, Palo Alto, CA, USA). Subsequently, the BglII-SalI fragment from pSTAT5A (1*6)-IRES2-EGFP was subcloned into the BglII-XhoI sites from the MSCV vector. All constructs were verified by sequencing. To generate stable, high-titer retroviral packaging cell lines, H29 cells were transiently transfected with 10 mg of vector DNA using the calcium-phosphate precipitation method. After 72 hours, the H29 supernatant, containing vesicular stomatitis virus–pseudotyped retrovirus, was used for cross-transduction of the PG13 packaging cell line, in the presence of polybrene (Sigma; 8 mg/mL). High-titer retrovirus-producing PG13 cell lines were selected by EGFPþ cell sorting and/or single cell cloning, as well as by high protein expression levels as determined by western blotting. Retroviral transduction protocol Cord blood–derived CD34þ cells were prestimulated for 48 hours in a cocktail containing QBSF-60 serum-free medium (Quality Biological, Inc., Gaithersburg, MD, USA) supplemented with TPO, FL, and KL (all at 100 ng/mL) (all

107

cytokines kindly provided by Kirin Corp, Tokyo, Japan). Retroviral supernatants were harvested from stable PG13 producers cultured in QBSF for 8 to 12 hours. Prior to transduction rounds, supernatants were collected; TPO, FL, and KL (all at 100 ng/mL) and 4 mg/mL polybrene were added, and supernatants were filtered through 0.45micron filters (Corning, Vineland, NJ, USA) and used immediately for transduction of CB CD34þ cells on RetroNectin-coated (Takara, Otsu, Japan) 6-well plates. Three consecutive transduction rounds of 8 to 12 hours were performed prior to starting various assays.

Flow cytometry analysis All antibodies were obtained from Pharmingen (San Diego, CA, USA). Cells were incubated with phycoerythrin-labeled antibodies at 4 C for 45 minutes. For blocking nonspecific binding to Fcg receptors, cells were treated with anti-Fcg antibodies for 15 minutes at 4 C. All fluorescence-activated cell sorter (FACS) analyses were performed on a FACScalibur (Becton-Dickinson, San Jose, CA, USA) and data was analyzed using FlowJo (Tree Star, Inc., San Carlos, CA, USA). Cells were sorted on a MoFlo (DakoCytomation, Carpinteria, CA, USA).

CFC, LTC-IC, and 2nd CAFC assays CFC assays and LTC-IC assays on MS5 stromal cells were performed as described previously [21]. Briefly, CFC assays were performed in 1.2% methylcellulose containing 30% FBS, 57.2 mM b-mercaptoethanol, 2 mM glutamine, 0.5 mM hemin (Sigma), 20 ng/mL interleukin-3 (IL-3), 20 ng/mL IL-6, 20 ng/mL G-CSF, 20 ng/mL KL, and 6 U/mL EPO. CFU-E assays were performed as CFC assays in the presence or absence of EPO only. Cobblestone area– forming cell (CAFC) and LTC-IC assays were performed as described [21] by plating transduced CB CD34þ cells (10– 60% EGFPþ) or FACS-sorted EGFPþCD34þ cells at 103 to 104 cells in T12.5 tissue-culture flasks (Becton-Dickinson, Franklin Lakes, NJ, USA) with confluent monolayers of MS5, OP9, or AGM-S2 stroma [22], in triplicate or at limiting dilutions in the range of 5 to 100 cells per well on MS5 stroma in 24- or 48-well plates in LTC medium. Half the medium and suspension cells were removed weekly and fresh medium added. Suspension cells were analyzed by FACS for EGFPþ and surface phenotype and plated for secondary CFC. Cobblestone areas #10 phasedark closely associated cells beneath the stroma were scored at 7 to 18 days and at 5 to 6 weeks. For 2nd CAFC assays, day-10 cobblestone areas were harvested by trypsinization of adherent cell populations and replated on fresh MS5 stroma or used for analysis or GFPþ sorting. Additional CAFC stromal cocultures were performed using sorted EGFPþ CD34þ/CD38 and CD34þ/CD38þ subsets of CB CD34þ cells and AML patient blast cells.

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

108

5’LTR

IRES2

EGFP

3’LTR

5’LTR

Flt3ITD

IRES2

EGFP

3’LTR

5’LTR

STAT5A(1*6)

IRES2

EGFP

3’LTR

Figure 1. Map of M iG1 retroviral vectors used in this study expressing enhanced green fluorescence protein (EGFP) alone (control vector), Flt3-ITD (W 51 or W78 mutants) with EGFP, or STAT5A(1*6) with EGFP.

Gene-array analysis Ten-mL aliquots were run on 1.5% agarose gels. For microarray analyses, total RNA was isolated from transduced GFPþ-sorted cells using the RNeasy kit from Qiagen and 4 mg of RNA was used for labeling reactions according to the manufacturer’s instructions and was hybridized to Affymetrix Human Genome U133A. MiGR1 and STAT5A(1*6) transcripts were hybridized independently and gene expression profiles were compared. Differences in gene expression were considered significant when the fold change was greater than 1.87 with a detection p value less than 0.05 and a signal value greater than 200.

Results Retroviral transduction of Flt3-ITD and STAT5A(1*6) into human CB CD34þ cells Cytokine-primed human CB CD34þ cells were transduced with the control bicistronic MiGR1-EGFPþ retroviral vector or with vectors expressing either the W51 or W78 Flt3-ITD mutants (Fig. 1). Gene transfer efficiencies ranged from 15 to 45% as determined by FACS for EGFP fluorescence 24 to 36 hours after final transduction. The mutated Flt3-ITD protein was detected in transduced cells by western blot. Constitutive activation of STAT5A has been reported in hematopoietic cells lines transfected with Flt3ITD [5,11]. We also observed strong expression and nuclear localization of STAT5A in CB CD34þ cells transduced with mutated Flt3-ITD but not in cells transduced with WT Flt3 and ligand (data not shown). In order to determine if the phenotype and downstream signaling events that we observed in Flt3-ITD-transduced cells could be explained predominantly by activation of the STAT5A pathway, we developed a MiGR1-based vector expressing constitutively activated STAT5(1*6) (Fig. 1) and used this to transduce CD34þ cells with an efficiency of 20 to 35% for the STAT5 and 35 to 60% for the MiGR1 control. As a further control we suppressed STAT5 activity by coexpression of a dominant-negative STAT5 (Y694F) in Flt3-ITDþ cells and showed complete suppression of the Flt3-ITD phenotype (CFC, early CAFC, and erythroid differentiation) (data not shown). Comparison of Flt3-ITD and STAT5A(1*6) proliferation in stromal coculture The expression of two Flt3-ITD mutants (W51, W78) provided a selective advantage to CD34þ cells in MS5 stromal

coculture. Increased cell proliferation relative to nontransduced cells or EGFP-transduced CD34þ control populations was detected at 2 weeks of culture and mutant cells progressively expanding relative to control cells through to week 5 (Fig. 2). The STAT5(1*6)-transduced cells showed comparable selective expansion relative to nontransduced or EGFP control cells in three different studies using three different stromal layers (MS5, OP9, AGM-S2) (Fig. 2) The selective growth advantage with STAT5(1*6) was seen a week earlier than in Flt3-ITD cultures and resulted in a greater fold expansion (4.6-fold vs 3.2-fold to 3.4-fold) by week 5. Morphologically, the cells generated in both Flt3-ITD and STAT5A(1*6) cultures showed a strong bias to erythroid differentiation with a spectrum of proerythroblasts, basophilic erythroblasts, polychromatic and orthochromatic erythroblasts, and terminally differentiated erythroid cells at week 2dall developing in the absence of exogenous erythropoietin. Flow cytometric analysis showed high numbers of cells, particularly in the STAT5A(1*6) cultures, expressing the erythroid markers CD36, CD71bright, and glycophorin A (Fig. 3A) with reduced CD45 expression. Myeloid cells, particularly those expressing the monocyte-macrophage marker CD14, were reduced relative to control cultures. Analysis of progenitors in the CFC assay at the time of transduction (day 0) showed a modest increase in CFC frequency with both Flt3-ITD- and STAT5A(1*6)-transduced populations with increased numbers of erythroid colonies (BFU-E) relative to myeloid (CFU-GM). The progenitor content of stromal cocultures at one and two weeks also revealed more CFC in the Flt3-ITD and STAT5A(1*6) cultures relative to controls, again with a strong bias to erythroid relative to myeloid colonies (Fig. 3A). EPO-independent erythroid colony formation (CFU-E) was seen with STAT5A(1*6), and to a lesser extent Flt3-ITDþ cells at 1 to 2 weeks of culture, and not in control cultures. These data suggest that Flt3-ITD signaling via STAT5A provides a proliferative advantage at the CD34þ stem/progenitor cell level; drives erythropoiesis, even in the absence of exogenous EPO; and blocks myeloid differentiation. Flt3-ITD- and STAT5A(1*6)-transduced CD34þ cells generate early cobblestone areas in stromal coculture and the CAFC possess stem cell capability The formation of cobblestone areas (CAs) of phase dark cells beneath the stroma has been used as an assay for detection of multilineage progenitor cells (week-2 ‘‘early’’

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116 2.5

A

2

MiGR1 W51 W78

4

1.5 1 0.5

1

3

2

30

35 0

MiGR1 STAT5A(1*6)

3.5

B

15

CAFC x 102

4

10

0

MiGR1 STAT5A(1*6)

25

MiGR1 Flt3ITD

20

STAT5A(1*6)

15

MiGR1

10

Flt3ITD

5

MiGR1

0

Days of Co-culture

3 2.5 2 1.5

5

1 0

20

25

30

35

STAT5A (1*6)

15

MiGR1

10

Flt3ITD

5

MiGR1

0

Flt3ITD

0

STAT5A (1*6)

0

0

MiGR1

0.5

MiGR1

Fold Change in % GFP+ Cells

BFU-E CFU-mix CFU-GM

1

0 -0.5

B

5

GPA+ CD14+

Total CFC x 103

2

Total cells x 105

Fold change in % GFP+ Cells

A

109

-0.5

Days of Co-culture Figure 2. (A) CB CD34þ cells were transduced with the MiGR1 control vector or vectors expressing Flt3-ITD mutants W51 or W78. Data expressed as fold increase in percentage of EGFPþ cells harvested weekly in the suspension phase. EGFPþ cells at input ranged from 15 to 45%. (B) CB CD34þ cells were transduced with the MiGR1 control vector or a vector expressing STAT5A(1*6) and cocultured on MS5, OP9, or AGM-S2 stroma (data shown as mean and SEM of three experiments each using a different stroma). Data expressed as fold increase in percentage of EGFPþ cells harvested weekly in the suspension phase. EGFPþ cells at input ranged from 20 to 60%.

CAFC) or true self-renewing hematopoietic stem cells (week-5þ ‘‘late’’ CAFC) [23]. We observed that the cytokine priming of CD34þ cells and subsequent culture to allow EGFP expression was associated with loss of early CAFC in control cultures and with a w50% reduction in week-5 CAFC. In contrast, Flt3-ITD- and STAT5A(1*6)transduced cells formed large numbers of early cobblestone areas by day 7 to 14 (1.8–4.0% in Flt3-ITD cultures and 10–12% in STAT5A(1*6) cultures) by limiting dilution (Fig. 3B). Flt3-ITD-transduced cells formed few late CAFC (!0.1%) while STAT5A(1*6) still showed large numbers of CAFC (w10%) by week 5 (Fig. 3B). The early cobblestone areas contained cells that were predominantly erythroblasts at various stages of differentiation, and

Day 10

Day 35

Figure 3. (A) Left panel: FACS analysis of suspension cells recovered after 2 weeks of MS5 coculture of CB CD34þ cells transduced with MiGR1, Flt3-ITD W51, or STAT5A(1*6) vectors. Total suspension cells that expressed the erythroid marker glycophorin A (GPA) or the monocytemacrophage marker CD14 are shown. Note that both Flt3-ITD and STAT5(1*6) significantly (p 5 0.001) increased GPAþ erythroid differentiation and decreased (p 5 0.03), monocyte-macrophage differentiation. Right panel: Progenitor cell assay of suspension cells recovered after 2 weeks of MS5 coculture of CB CD34þ cells transduced with MiGR1, Flt3-ITD W51, or STAT5A(1*6) vectors. Note the significant increase in total progenitors generated in Flt3-ITD- and STAT5A(1*6)-transduced populations due, predominantly, to increased numbers of BFU-E. (B) Early (day-10) and late (day-35) CAFC assay of CB CD34þ cells transduced with MiGR1, Flt3-ITD W51, or STAT5A(1*6) vectors on MS5 stroma at days 10 and 35. Cobblestone area formation determined by limiting dilution. All early cobblestone areas were EGFPþ. Note the absence of early CAFC in MiGR1 control-transduced CD34þ cells and a frequency of 1.8 to 4% early CAFC with Flt3-ITD expression and 10 to 12% with STAT5A(1*6) expression. The STAT5A(1*6) effect persisted to week 5 (10% CAFC) whereas the Flt3-ITD-transduced cells showed less than 0.1% late CAFC.

myelo-monocytic blasts. Normal early CAFC do not undergo secondary passage to generate further CAFC; however, Flt3-ITD-transuded CD34þ cells generated secondary and tertiary early CAs with a replating efficiency

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

CD38-PE

110

2nd CAFC CAFC CFC RT-PCR

2nd CFC

EGFP CAFC freq (day 10)

MiGR1 STAT5(1*6)

CFC (per 1000 cells plated) CD38+ CD38-

CD38-

CD38+

0

0

28

14

8

8

11

7

13.2%

3.7%

7

27

18

6

22

18

CFU-GM BFU-E mixed CFU-GM BFU-E mixed

2nd CAFC 2nd CFC

Figure 4. STAT5A(1*6) or MiGR1 control transduced CB CD34þ cells were FACS sorted for EGFPþ and high or low expression of CD38. Sorted cells were assayed for day-10 CAFC in limiting-dilution MS5 cocultures, as well as for CFC. In addition the secondary passage potential of both CAFC and CFC of the CD38lo/ and CD38þ fractions was determined. Note that the majority of the STAT5A(1*6) CAFC were in the CD34þ, CD38 fraction while CFC were equally distributed between CD38 and CD38þ fractions with both populations showing excess BFU-E and CFU-Mix relative to comparable fractions of the MiGR1 control population.

of greater than 1. Similarly, STAT5A(1*6) CAs continuously generated new early CAs over 18 weeks of 7 serial passages. STAT5A(1*6)-transduced CD34þ cells were sorted into EGFPþ, CD34þ, and CD38 or CD38þ subpopulations prior to MS5 stromal coculture (Fig. 4). The CD38 cells were markedly enriched for early CAFC relative to the CD38þ fraction but the frequency and morphology of CFC were comparable in the two fractions. Of particular note was the observation that only the CAFC and CFC from the CD34þ, CD38 fraction were capable of generating secondary CAFC and CFC. Molecular analysis of CD34þ stem/progenitor cells overexpressing constitutively active Flt3-ITD and STAT5A(1*6) To obtain further insight into the molecular basis for transformation induced by constitutively activated Flt3 and STAT5, microarray analysis was performed using the Affymetrix human genome U133A gene array containing w21,000 probe sets (Tables 1, 2). The bias to erythroid differentiation seen with both Flt3-ITD- and STAT5A(1*6)transduced cells is reflected in the gene arrays. Eryrthroid-associated genes upregulated by STAT5 include hemoglobins z, e1, a1, a2, gA, and glycophorins A and B, Ankyrin 1 erythrocytic, rhesus blood group–associated glycoprotein, Kell blood group precursor, and CD36, while erythroidassociated genes upregulated by Flt3-ITD include hemoglobins a1, a2 , e1, and Ankyrin 1. The impaired myeloid differentiation seen particularly with STAT5A(1*6) was associated with downregulation of myeloid-associated genes

including cathepsin G, myeloperoxidase, CD31, neutrophil elastase 2, C/EBPa, C/EBPd, CD44, and CD48. FosB is strongly upregulated by both Flt3-ITD and Stat5(1*6): FosB and other members of the Fos family hetero-dimerize with Jun proteins to form the AP-1 transcription factor complex. AP-1-regulated genes include important regulators of invasion and metastasis, proliferation, differentiation, and survival, which are important regulators of cell proliferation and differentiation [24]. The small GTP-binding proteins Rac1 and Rac2 are critically important in regulating multiple signal transduction pathways in eukaryotic cells, and have been implicated in hematopoietic stem cell proliferation, survival, and migration [25,26]. Haataja et al. [27] report isolation of a third member, Rac3, which differs from the others at the carboxy-terminal end, a domain associated with subcellular localization. Rac3 was upregulated in both Flt3-ITD and STAT5A(1*6) CD34þ gene arrays. Taken together, results suggest a role for both the Rac1 and Rac3 GTPases in human breast cancer progression and metastasis and RNA interference reveals a critical role for these GTPases in the invasive behavior of glioma and breast carcinoma cells [28]. Our data suggests a similar role for Rac3 in hematopoietic stem and progenitor cells, possibly facilitating their invasion of the stromal microenvironment and formation of cobblestone areas beneath the stroma. Primary AML blast cells generate early CAs Bone marrow and peripheral blood samples were obtained from 12 AML patient (FAB M4, M2) at diagnosis or in

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

111

Table 1. Differential gene expression in CB CD34þ cells expressing the W51 Flt3-ITD* Gene FBJ murine osteosarcoma viral oncogene homolog B Phosphatidylinositol binding clathrin assembly protein Interleukin-8 Cyclin-dependent kinase inhibitor 2D (p19, inhibits CDK4) Tumor necrosis factor receptor superfamily, member 9 Eukaryotic translation initiation factor 5A Leukemia inhibitory factor (cholinergic differentiation factor) Dual-specificity phosphatase 1 Hemoglobin, alpha 1 Thioredoxin interacting protein v-maf musculoaponeurotic fibrosarcoma oncogene homolog F TGFb inducible early growth response Ankyrin 1, erythrocytic Hemoglobin, e 1 Ras-related C3 botulinum toxin substrate 3(r family, Rac3) Cyclin D1 Ubiquitin-conjugating enzyme E2M (UBC12 homologue) Small inducible cytokine A3 Hemoglobin, a 1 Small inducible cytokine A3 Protein tyrosine phosphatase type IVA, member 3 Hemoglobin, a 2 Activated leukocyte cell adhesion molecule Selectin L (lymphocyte adhesion molecule 1) Chemokine (C-C motif) receptor 2 Zinc finger protein 304 B-cell CLL/lymphoma 6 (zinc finger protein 51) Tumor necrosis factor receptor superfamily member 1B Erythrocyte membrane protein band 4, 1-like 3 Homeobox A9 Src-like adaptor Homeobox A5 Transcription factor 8 (represses IL-2 expression) Metallothionein 1H C-type lectin, superfamily member 6

Unigene accession

Gene symbol

Fold change

Hs.75678 Hs.7885 Hs.624 Hs.29656 Hs.73895 Hs.119140 Hs.2250 Hs.171695 Hs.272572 Hs.179526 Hs.51305 Hs.82173 Hs.183805 Hs.117848 Hs.45002 Hs.82932 Hs.200478 Hs.73817 Hs.272572 Hs.73817 Hs.43666 Hs.347939 Hs.10247 Hs.82848 Hs.395 Hs.287374 Hs.155024 Hs.256278 Hs.103839 Hs.127428 Hs.75367 Hs.37034 Hs.232068 Hs.2667 Hs.115515

FOSB PICALM I18 CDKN2D TNFRSF9 EIF5A LIF DUSP1 HBA1 TXNIP MAFF TIEG ANK1 HBE1 RAC3 CCND1 UBE2M SCYA3 HBA1 SCYA3 PTP4A3 HBA2 ALCAM SELL CCR2 ZNF304 BCL6 TNFRSF1B EPB41L3 HOXA9 SLA HOXA5 TCF8 MT1H CLESCF6

8.00 8.00 3.73 3.25 2.83 2.64 2.64 2.46 2.30 2.14 2.14 2.00 2.00 2.00 2.00 2.00 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 2.00 2.00 2.00 2.00 2.14 2.14 3.48 7.46 9.85

*CB CD34þ cells were prestimulated for 48 hrs in QBSF with KL, FL, and TPO (100 ng/mL each) followed by three transduction rounds in the next 48 hrs. GFPþ cells were sorted and total RNA was isolated and used to hybridize gene arrays. Data shown is the comparison of GFPþ STAT5A(1*6) versus GFPþ MiGR1 cells. A change in gene expression was only considered significant when the fold change was greater than 1.87 with a statistical p value less than 0.05 and a signal value greater than 200.

relapse. Seven patients had Flt3 mutations: 6 with Flt3-ITD and one with an Asp 835 mutation. Ficoll-separated mononuclear cells were plated onto MS5 stroma (5  105 cells per T12.5 flask) and scored for early (day-10 to day-14) and late (day-35) CAFC. Early CAFC were detected only in cultures established from patients with Flt mutations (Table 3), indicating that this early cobblestone phenotype is a marker for subsets of leukemia with activating Flt3 mutations. Late (week-5) CAFC were present in both groups but were greater than 10-fold more frequent in the Flt3-ITD group, a feature shared with the STAT5A(1*6)-transduced CD34 population. Isolation and characterization of leukemic stem cells isolated from patients with Flt3-ITDþ AML Two of the Flt3-ITD patients shown in Table 3 had very high white blood cell counts and were subjected to thera-

peutic leukapheresis. We were able to obtain large numbers of cells that were sorted for both CD34, CD38 and HLADR expression (Fig. 5). In all, 0.1% of total leukemic cells were CD34þ, CD38, HLA-DR, and were enriched 150fold to 600-fold for early CAFC in MS5 coculture. This phenotype has been shown to characterize LSC with NOD/SCID engraftment potential [29] and co-association of this fraction with early CAFC further supports the status of this subpopulation as LSC. Affymetrix human genome U133A gene array analysis was undertaken in a comparison of genes upregulated or downregulated in both LSC fractions relative to the bulk AML population (Table 4). GATA3 and MYCN are the most overexpressed genes. The GATA3 nuclear transcription factor is reported to be expressed exclusively in T-ALL, not B-ALL or AML (which express GATA1 and 2) [30]. In birds, GATA3 is implicated in hematopoiesis, but in mammals it is mainly

112

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

Table 2. Differential gene expression in CB CD34þ cells expressing STAT5A(1*6) Gene Hemoglobin, z Glycophorin A (includes MN blood group) FBJ murine osteosarcoma viral oncogene homolog B Aquaporin 3 Leukemia inhibitory factor (cholinergic differentiation factor) Hemoglobin, a 2 Hemoglobin, a 1 Glycophorin B (includes Ss blood group) Monoamine oxidase A pim-1 oncogene Ankyrin 1, erythrocytic TNF receptor–associated factor 4 Cytokine-inducible SH2-containing protein Hemoglobin, g A Hemoglobin, e 1 Leptin receptor Rhesus blood group–associated glycoprotein Ankyrin 1, erythrocytic ras-related C3 botulinum toxin substrate 3 (r family, Rac3) Kell blood group precursor (McLeod phenotype) CD36 antigen (collagen type I receptor) PDGFA-associated protein 1 Son of sevenless homolog 1 (Drosophila) Vascular endothelial growth factor FOS-like antigen 2 Absent in melanoma 1 T cell receptor b locus CD44 antigen (homing function) RAB27A, member RAS oncogene family Platelet/endothelial cell adhesion molecule (CD31 antigen) TYRO protein tyrosine kinase binding protein Myeloperoxidase v-myc myelocytomatosis viral related oncogene (avian) Cathepsin G Elastase 2, neutrophil CCAAT/enhancer binding protein (C/EBP), d Immunoglobulin heavy constant m Caspase 1, apoptosis-related cysteine protease CD48 antigen (B-cell membrane protein) Chemokine (C-X-C motif), receptor 4 (fusin) B-cell CLL/lymphoma 6 (zinc finger protein 51) CCAAT/enhancer binding protein (C/EBP), a

Unigene accession

Gene symbol

Fold change

Hs.272003 Hs.108694 Hs.75678 Hs.234642 Hs.2250 Hs.347939 Hs.272572 Hs.343871 Hs.183109 Hs.81170 Hs.183805 Hs.8375 Hs.8257 Hs.266959 Hs.117848 Hs.226627 Hs.169536 Hs.183805 Hs.45002 Hs.78919 Hs.75613 Hs.278426 Hs.326392 Hs.73793 Hs.301612 Hs.161002 Hs.303157 Hs.169610 Hs.50477 Hs.78146 Hs.9963 Hs.1817 Hs.25960 Hs.100764 Hs.99863 Hs.76722 Hs.153261 Hs.2490 Hs.901 Hs.89414 Hs.155024 Hs.76171

HBZ GYPA FOSB AQP3 LIF HBA2 HBA1 GYPB MAOA PIM1 ANK1 TRAF4 CISH HBG1 HBE1 LEPR RHAG ANK1 RAC3 XK CD36 PDAP1 SOS1 VEGF FOSL2 AIM1 TRB CD44 RAB27A PECAM1 TYROBP MPO MYCN CTSG ELA2 CEBPD IGHM CASP1 CD48 CXCR4 BCL6 CEBPA

42.22 12.13 11.31 4.92 4.92 3.73 3.73 3.73 3.25 3.25 3.25 3.25 3.03 3.03 2.83 2.64 2.64 2.46 2.46 2.46 2.30 2.14 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 1.87 2.00 2.00 2.00 2.00 2.14 2.14 2.46 2.64 2.64 3.25

CB CD34þ cells were prestimulated for 48 hrs in QBSF with KL, FL, and TPO (100 ng/mL each) followed by three transduction rounds in the next 48 hrs. GFPþ cells were sorted and total RNA was isolated and used to hybridize gene arrays. Data shown is the comparison of GFPþ STAT5A(1*6) versus GFPþ MiGR1 cells. A change in gene expression was only considered significant when the fold change was greater than 1.87 with a statistical p value less than 0.05 and a signal value greater than 200.

detected in T lymphocytes and the embryonic nervous system [31]. GATA3-deficient mice die at embryonic day 11 to 12 with a failure of embryonic hematopoiesis and nervous system defects [32]. MYCN amplification strongly predicts adverse outcome of neuroblastoma [33] and is implicated in rhabdomyosarcoma [34]. The DNA-binding transcription factor hepatic leukemia factor (HLF) is also strongly expressed in Flt3-ITD LSC (Table 4). HLF expression in CD34þCD38 hematopoietic stem cells is greater in neonatal and fetal than adult cells [35]. HLF is closely related to zipper-containing transcrip-

tion factors that have a role in developmental stage–specific gene expression [36] and it has recently been shown to activate the LMO2 promoter [37] necessary for initiation of mammalian embryonic hematopoiesis. In addition, HLF was found to be involved in acute human t(17;19)-positive proB-leukemia when fused to the E2A basic-helix-loop-helix transcription factor [36,38] and was recently identified in two other stem cell molecular profiling studies [39,40]. Shojaei et al. [35] showed that transduction of HLF into cord blood CD34þ cells enhanced their in vivo repopulating potential, conferring antiapoptotic effects and preventing

Flt3 mutation Flt3 WT

7 5

138621 060

229619 2163

premature cell death that was mediated in part by bcl-2 regulation. Upregulation of the homeobox genes HOXA9 and HOXA5 in the LSC gene array has implications for leukemic development. HOXA9 is one of the top 20 genes distinguishing AML from ALL and correlates with poor prognosis [41]. Kumar et al. [42] have proposed the ‘‘HOX code,’’ minimally defined by the HOXA5-9 cluster as central to MLL leukemogenesis. HOXA9 transformed primary bone marrow cells through specific collaboration with meis 1 (reviewed [1,2]). Constitutive expression of HOXA5 in cord blood CD34þ cells inhibited human erythropoiesis and promoted myelopoiesis [43]. The reciprocal inhibition of erythropoiesis and promotion of myelopoiesis in the absence of any demonstrable effect on proliferation suggests that HOXA5 diverts differentiation at a multipotent progenitor stage away from the erythroid toward the myeloid pathway.

Discussion Enforced expression of Flt3-ITD in human cord blood CD34þ cells confers a selective, proliferative advantage to cells at the stem and progenitor level in competitive stroma coculture systems over 5 weeks, with increased erythroid differentiaton at the expense of myeloid. A major downstream pathway for the mutated receptor involves phosphorylation and activation of STAT5A and its nuclear localizations. This pathway is normally not activated by the wild-type receptor. Comparison of the Flt3-ITD- and STAT5(1*6)-transduced CD34þ cells in stromal coculture or cytokine-stimulated suspension culture revealed selective growth advantage only in the stromal coculture. This was associated with appearance of large numbers of cobblestone areas by 7 to 14 days with persistent cobblestone formation though week 5 with STAT5(1*6) only. These CAFC could be distinguished from normal early CAFC that are derived from progenitors and are rapidly lost in control cultures. In addition, the Flt3-ITD and STAT5A(1*6) week-2 CAFC could be serially passaged, in the latter case for more than 18 weeks, generating secondary and tertiary CAFC and progenitors (CFC-GM, BFU-E, CFU-Mix). Phenotypic analysis showed that the majority of early CAFC were CD34þ, CD38, a phenotype characteristic of normal and leukemic hematopoietic stem cells [29]. We have also shown in vivo engraftment of STAT5A(1*6) early CAFCs

FL2-unstained

No. Day-10–14 CAFC/5  105 Day-35 CAFC/5  105

0.43%

HLA-DR-APC

FL3-unstained

CD38-PE

AML patient

113

1.7%

FL2-CD38

Table 3. Early (day-10 to day-14) and late (day-35) CAFC incidence in MS5 stromal coculture of primary AML patient blast cells (mean 6 SEM): Comparison of patients with WT Flt3 to those with constitutively active Flt3 mutations (six with Flt3-ITD and one with an Asp 835 mutation)

FL2-unstained

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

0.4%

FL3-unstained

HLA-DR-APC

Flt3 ITD Blasts

Day 14 CAFC frequency (Limiting Dilution)

Fold CA enrichment

Unsorted

0.0001%

/

CD38-/ HLA-DR-

0.002%

20

CD34+/ CD38-/ HLA-DR-

0.20-0.06%

150-300

Figure 5. FACS sorting of leukemic cells from a patient with Flt3-ITDþ, M4 AML. Cells were sorted for CD34, CD38, and HLA-DR expression. The frequency of early (day-14) CAFC determined by limiting dilution on MS5 stroma was 89 6 8 per 5  105 AML cells. CAFC were enriched in the CD34þ, CD38 and in the CD34þ, CD38, HLA-DR fractions.

in NOD/SCID mice [20]. Furthermore, STAT5A(1*6) expressed in mouse embryonic stem cells facilitated enhanced hematopoietic differentiation in stromal coculture with generation of stem cells with at least short-term primary and secondary engraftment in adult irradiated mice [44]. Earlier studies employed STAT5A/B-deficient mice where targeted gene disruption of the first protein-encoding exon resulted in expression of an N-terminally truncated form of STAT5A/B. These mice displayed profound defects in competitively repopulating HSC [45]. Hoelbl et al. [46] have developed mice with complete deletion of the STAT5A/B gene locus. Most of these mice die at birth, but a few survive and have a severe combined immunodeficiency with lack of CD8 T cells and B cell maturation block at the pre-pro-B cell stage [46,47]. Constitutive STAT5 activation is reported in 22 to 80% of primary AML, only a portion of which could be explained by mutations in Flt3 [48]. Lewis et al. [49] generated activating mutants of zebrafish STAT5.1 and showed that the mutant protein increased tyrosine phosphorylation and transactivation activity compared to wild-type protein and expression led to a range of hematopoietic perturbations; hyperproliferation of the hematopoietic compartment increased numbers of early

114

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

Table 4. Differential gene expression in CD34þ/CD38/HLA-DR primary Flt3-ITDþ AML cells Gene GATA binding protein 3 Hepatic leukemia factor v-myc myelocytomatosis viral related oncogene, neuroblastoma derived (avian) Homeobox A9 Hepatic leukemia factor Histone 1, H2bg Serologically defined colon cancer antigen 8 Amyloid b (A4) precursor protein (protease nexin-II, Alzheimer disease) Histone 1, H2bc Homeobox A5 Potassium inwardly-rectifying channel, subfamily J, member 2 Quaking homolog, KH domain RNA binding (mouse) Ganglioside-induced differentiation-associated protein 1-like 1 Signal sequence receptor, delta (translocon-associated protein delta) E2IG2 protein Cytochrome c oxidase subunit VIIb Cytochrome c oxidase subunit Vb Ubiquinol-cytochrome c reductase core protein I SRY (sex determining region Y)-box 4 Cell death–regulatory protein GRIM19 Annexin A2 Cystatin A (stefin A) Ribonuclease, RNase A family, 2 (liver, eosinophil-derived neurotoxin) RA-regulated nuclear matrix–associated protein Insulin-like growth factor binding protein 7 Biliverdin reductase A Cell division cycle 2, G1 to S and G2 to M Lectin, galactoside-binding, soluble, 1 (galectin 1) Chromosome 10 open reading frame 3

Unigene accession

Gene symbol

Fold change

Hs.524134 Hs.196952 Hs.25960 Hs.110637 Hs.196952 Hs.347939 Hs.272572 Hs.343871 Hs.458395 Hs.595822 Hs.1547 Hs.510324 Hs.517059 Hs.409223 Hs.475387 Hs.522699 Hs.1342 Hs.119251 Hs.592947 Hs.534453 Hs.511605 Hs.518198 Hs.728 Hs.26490 Hs.479808 Hs.488143 Hs.334562 Hs.445351 Hs.14559

GATA3 HLF MYCN HOXA9 HLF HIST1H2BG SDCCAG8 APP HIST1H2BC HOXA5 KCNJ2 QKI GDAP1L1 SSR4 E2IG2 COX7B COX5B UQCRC1 SOX4 GRIM19 ANXA2 CSTA RNASE2 RAMP IGFBP7 BLVRA CDC2 LGALS1 C10orf3

20.0 8.0 8.0 6.0 5.0 3.0 3.0 3.0 2.5 2.5 2.0 2.0 2.0 3.5 3.5 4.0 4.0 4.5 4.5 4.5 5.0 6.0 6.5 6.5 6.5 7.0 7.0 7.0 9.5

Primary Flt3-ITDþ AML cells were sorted for CD34þ/CD38/HLA-DR and total RNA was isolated and used to hybridize gene arrays. Data shown is the comparison of CD34þ/CD38/HLA-DR sorted cells versus unsorted cells. A change in gene expression was only considered significant when the fold change was greater than 2 with a statistical p value less than 0.05 and a signal value greater than 200.

and late myeloid cells, erythrocytes, and B cells. STAT5A(1*6) expression in murine HSC renewal ex vivo and in vivo produced a fatal MPD [50]. Transduced committed pluripotent progenitors (CD34þ, KLS) were not induced to self-renew but expansion and differentiation were enhanced [50]. Mouse BM transduced with activated STAT5 into irradiated wild-type or Rag2/ mice revealed increase in erythroid cell numbers, hyperproliferation of myeloid cells, accumulation of myeloblasts, and leukemia [51]. STAT5 regulates expression of genes involved in cell survival and cell-cycle progression like bcl-xL, and D-type cyclins as well as genes encoding cytokines or growth factors (osm, igf-1) and the proto-oncogene pim-1. Persistent transactivation of these genes by STAT5A(1*6) may play a role in progression to leukemia. Activation of the PI3-K/Akt signaling cascade also plays an important role in STAT5-induced cell growth and survival via the scaffolding adapter Gab2. Akt regulates the activity of a number of substrates involved in cell apoptosis or proliferation like Bad, Forkhead, NFkB, and GSk3b [52]. The transcription factor CCAAT enhancer binding protein-a (C/EBPa), a crucial regulator of granulopoiesis, was downmodulated by STAT5A(1*6) and Flt3-ITD in

our CD34þ gene arrays (Tables 1 and 2 and Schuringa, Morronne, Ki, and Moore, unpublished observation), as well as in primary Flt3-ITD AML [53]. In CD34þ cells transduced with C/EBPa in-frame with the estrogen receptor ligand-binding domain, the addition of b-estradiol leads to granulocyte differentiation and inhibits erythroid differentiation [54]. Disruption of C/EBPa including dominantnegative mutations of C/EBPa are found in AML [55]. The t(8;21) AML1-ETO-positive AML blasts have eightfold lower level of C/EBPa mRNA and undetectable protein levels [56]. C/EBPa deficiency in mice results in progenitor hyperproliferation and macrophage and neutrophil differentiation defects marked by absence of receptors for M-CSF and G-CSF [57]. Conditional C/EBPa KO in mice blocked the differentiation of the common myeloid progenitor with myeloblast accumulation in marrow, absence of neutrophils, and enhancement of HSC competitive repopulating capacity and self-renewal [58]. We have recently presented data suggesting that C/EBPa might be a critical downstream gene whose downregulation following Flt3-ITD or STAT5A constitutive activation accounts for the proliferative and differentiative events observed [59]. We coexpressed an estrogen-inducible C/EBPa-ER

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

protein together with the STAT5A(1*6) mutant in CB CD34þ cells and showed that re-expression of C/EBPa restored a normal pattern of myelopoiesis, suppressing erythropoiesis and the generation of progenitors (CFC) and both short- and long-term culture-initiating cells.

Acknowledgments This work was supported in part by funds from the Italian Ministry for University and Research (MIUR, CLUSTER C-04 and Interlink grant) and from the Italian Association for Cancer Research (AIRC) to GM, by a grant from the EMBO (ALTF-412-2001) to JJS, and by the Leukemia & Lymphoma Society Specialized Center of Research Program and the Gar Reichman Fund of the Cancer Research Institute to MASM.

References 1. Moore MAS. Converging pathways in leukemogenesis and stem cell self-renewal. Exp Hematol. 2005;33:719–737. 2. Parcells BW, Ikeda AK, Simms-Waldrip T, Moore TB, Sakamoto KM. FMS-like tyrosine kinase 3 in normal hematopoiesis and acute myeloid leukemia. Stem Cells. 2006;24:1174–1184. 3. Gammaitoni L, Weisel KC, Gunetti M, et al. Elevated telomerase activity and minimal telomere loss in cord blood long-term cultures with extensive stem cell replication. Blood. 2004;103:4440–4448. 4. Zhang S, Fukuda S, Lee Y, et al. Essential role of signal transducer and activator of transcription (Stat)5a but not Stat5b for Flt3-dependent signaling. J Exp Med. 2000;192:719–728. 5. Hayakawa F, Towatari M, Kiyoi H, et al. Tandem-duplicated Flt3 constitutively activates STAT5 and MAP kinase and introduces autonomous cell growth in IL-3-dependent cell lines. Oncogene. 2000;19:624–631. 6. Grundler R, Miething C, Thiede C, Peschel C, Duyster J. FLT3-ITD and tyrosine kinase domain mutants induce 2 distinct phenotypes in a murine bone marrow transplantation model. Blood. 2005;105: 4792–4799. 7. Kiyoi H, Towatari M, Yokota S, et al. Internal tandem duplication of the FLT3 gene is a novel modality of elongation mutation which causes constitutive activation of the product. Leukemia. 1998;12: 1333–1337. 8. Kondo M, Horibe K, Takahashi Y, et al. Prognostic value of internal tandem duplication of the FLT3 gene in childhood acute myelogenous leukemia. Med Pediatr Oncol. 1999;33:525–529. 9. Gilliland DG, Griffin JD. The roles of FLT3 in hematopoiesis and leukemia. Blood. 2002;100:1532–1542. 10. Yamamoto Y, Kiyoi H, Nakano Y, et al. Activating mutation of D835 within the activation loop of FLT3 in human hematologic malignancies. Blood. 2001;97:2434–2439. 11. Mizuki M, Schwable J, Steur C, et al. Suppression of myeloid transcription factors and induction of STAT response genes by AML-specific Flt3 mutations. Blood. 2003;101:3164–3173. 12. Robinson LJ, Xue J, Corey SJ. Src family tyrosine kinases are activated by Flt3 and are involved in the proliferative effects of leukemia-associated Flt3 mutations. Exp Hematol. 2005;33:469–479. 13. Yang X, Liu L, Sternberg D, et al. 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. Cancer Res. 2005;65:7338–7347. 14. Brandts CH, Sargin B, Rode M, et al. Constitutive activation of Akt by Flt3 internal tandem duplications is necessary for increased survival, proliferation, and myeloid transformation. Cancer Res. 2005;65: 9643–9650.

115

15. Tickenbrock L, Schwable J, Wiedehage M, et al. Flt3 tandem duplication mutations cooperate with Wnt signaling in leukemic signal transduction. Blood. 2005;105:3699–3706. 16. Kelly LM, Liu Q, Kutok JL, Williams IR, Boulton CL, Gilliland DG. FLT3 internal tandem duplication mutations associated with human acute myeloid leukemias induce myeloproliferative disease in a murine bone marrow transplant model. Blood. 2002;99:310–318. 17. Choudhary C, Schwable J, Brandts C, et al. AML-associated Flt3 kinase domain mutations show signal transduction differences compared with Flt3 ITD mutations. Blood. 2005;106:265–273. 18. Schessl C, Rawat VP, Cusan M, et al. The AML1-ETO fusion gene and the FLT3 length mutation collaborate in inducing acute leukemia in mice. J Clin Invest. 2005;115:2159–2168. 19. Chung KY, Morrone G, Schuringa JJ, Wong B, Dorn DC, Moore MA. Enforced expression of an Flt3 internal tandem duplication in human CD34þ cells confers properties of self-renewal and enhanced erythropoiesis. Blood. 2005;105:77–84. 20. Schuringa JJ, Chung KY, Morrone G, Moore MA. Constitutive activation of STAT5A promotes human hematopoietic stem cell self-renewal and erythroid differentiation. J Exp Med. 2004;200:623–635. 21. Jo DY, Rafii S, Hamada T, Moore MA. Chemotaxis of primitive hematopoietic cells in response to stromal cell–derived factor-1. J Clin Invest. 2000;105:101–111. 22. Weisel KC, Moore MA. Genetic and functional characterization of isolated stromal cell lines from the aorta-gonado-mesonephros region. Ann N Y Acad Sci. 2005;1044:51–59. 23. Breems DA, Blokland EA, Neben S, Ploemacher RE. Frequency analysis of human primitive haematopoietic stem cell subsets using a cobblestone area forming cell assay. Leukemia. 1994;8:1095–1104. 24. Milde-Langosch K. The Fos family of transcription factors and their role in tumourigenesis. Review. Eur J Cancer. 2005;41: 2449–2461. 25. Arai A, Kanda E, Miura O. Rac is activated by erythropoietin or interleukin-3 and is involved in activation of the Erk signaling pathway. Oncogene. 2002;21:2641–2651. 26. Gu Y, Filippi M-D, Cancelas JA, et al. Hematopoietic cell regulation by Rac1 and Rac2 guanosine triphosphatase. Science. 2003;302:445–449. 27. Haataja L, Groffen J, Heistrekamp N. Characterization of RAC3, a novel member of the Rho family. J Biol Chem. 1997;272:20384– 20388. 28. Baugher PJ, Krishnamoorthy L, Price JE, Dharmawardhane SF. Rac1 and Rac3 isoform activation is involved in the invasive and metastatic phenotype of human breast cancer cells. Breast Cancer Res. 2005;7: R965–R974. 29. Hope KJ, Jin L, Dick JE. Acute myeloid leukemia originates from a hierarchy of leukemic stem cell classes that differ in self-renewal capacity. Nat Immunol. 2004;5:738–743. 30. Minegishi N, Morita S, Minegishi M, et al. Expression of GATA transcription factors in myelogenous and lymphoblastic leukemia cells. Int J Hematol. 1997;65:239–249. 31. Elefanty AG, Antoniou M, Custodio N, Carmo-Fonseca M, Grosveld FG. GATA transcription factors associate with a novel class of nuclear bodies in erythroblasts and megakaryocytes. EMBO J. 1996;15:319–333. 32. Pandolfi PP, Roth ME, Karis A, et al. Targeted disruption of the GATA3 gene causes severe abnormalities in the nervous system and in fetal liver haematopoiesis. Nat Genet. 1995;11:40–144. 33. Tang XX, Zhao H, Kung B, et al. The MYCN enigma: significance of MYCN expression in neuroblastoma. Cancer Res. 2006;66:2826– 2833. 34. Morgenstern DA, Anderson J. MYCN deregulation as a potential target for novel therapies in rhabdomyosarcoma. Expert Rev Anticancer Ther. 2006;6:217–224. 35. Shojaei F, Trowbridge J, Gallacher L, et al. Hierarchical and ontogenic positions serve to define the molecular basis of human hematopoietic stem cell behavior. Dev Cell. 2005;8:651–663.

116

M.A.S. Moore et al./ Experimental Hematology 35 (2007) 105–116

36. Inaba T, Roberts WM, Shapiro LH, et al. Fusion of the leucine zipper gene HLF to the E2A gene in human acute B-lineage leukemia. Science. 1992;257:531–534. 37. Crable SC, Anderson KP. A PAR domain transcription factor is involved in the expression from a hematopoietic-specific promoter for the human LMO2 gene. Blood. 2003;101:4757–4764. 38. Honda H, Inaba T, Suzuki T, et al. Expression of E2A-HLF chimeric protein induced T-cell apoptosis, B-cell maturation arrest, and development of acute lymphoblastic leukemia. Blood. 1999;93: 2780–2790. 39. Georgantas RW 3rd, Tanadve V, Malehorn M, et al. Microarray and serial analysis of gene expression analyses identify known and novel transcripts overexpressed in hematopoietic stem cells. Cancer Res. 2004;64:4434–4441. 40. Ivanova NB, Dimos JT, Schaniel C, Hackney JA, Moore KA, Lemischka IR. A stem cell molecular signature. Science. 2002;298: 601–604. 41. Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science. 1999;286:531–537. 42. Kumar AR, Hudson WA, Chen W, Nishiuchi R, Yao Q, Kersey JH. Hoxa9 influences the phenotype but not the incidence of Mll-AF9 fusion gene leukemia. Blood. 2004;103:1823–1828. 43. Crooks GM, Fuller J, Petersen D, et al. Constitutive HOXA5 expression inhibits erythropoiesis and increases myelopoiesis from human hematopoietic progenitors. Blood. 1999;94:519–528. 44. Schuringa JJ, Wu K, Morrone G, Moore MA. Enforced activation of STAT5A facilitates the generation of embryonic stem–derived hematopoietic stem cells that contribute to hematopoiesis in vivo. Stem Cells. 2004;22:1191–1204. 45. Bunting KD, Bradley HL, Hawley TS, Moriggl R, Sorrentino BP, Ihle JN. Reduced lymphomyeloid repopulating activity from adult bone marrow and fetal liver of mice lacking expression of STAT5. Blood. 2002;99:479–487. 46. Hoelbl A, Kovacic B, Kerenyi MA, et al. Clarifying the role of Stat5 in lymphoid development and Abelson-induced transformation. Blood. 2006;107:4898–4906.

47. Yao Z, Cui Y, Watford WT, et al. Stat5a/b are essential for normal lymphoid development and differentiation. Proc Natl Acad Sci U S A. 2006;103:1000–1005. 48. Ilaria Jr. RL. AML: Many new tools for discovery, not enough answers (yet). Leuk Res. 2006;30:1073–1075. 49. Lewis RS, Stephenson SE, Ward AC. Constitutive activation of zebrafish Stat5 expands hematopoietic cell populations in vivo. Exp Hematol. 2006;34:179–187. 50. Kato Y, Iwama A, Tadokoro Y, et al. Selective activation of STAT5 unveils its role in stem cell self-renewal in normal and leukemic hematopoiesis. J Exp Med. 2005;202:169–179. 51. Moriggl R, Sexl V, Kenner L, et al. Stat5 tetramer formation is associated with leukemogenesis. Cancer Cell. 2005;7:87–99. 52. Nyga R, Pecquet C, Harir N, et al. Activated STAT5 proteins induce activation of the PI 3-kinase/Akt and Ras/MAPK pathways via the Gab2 scaffolding adapter. Biochem J. 2005;390:359–366. 53. Zheng R, Friedman AD, Levis M, Li L, Weir EG, Small D. Internal tandem duplication mutation of FLT3 blocks myeloid differentiation through suppression of C/EBPa expression. Blood. 2004;103:1883–1890. 54. Cammenga J, Mulloy JC, Berguido FJ, et al. Induction of C/EBPa activity alters gene expression and differentiation of human CD34þ cells. Blood. 2003;101:2206–2214. 55. Frohling S, Dohner H. Function in acute leukemia. N Engl J Med. 2004;351:2370–2372. 56. Pabst T, Mueller BU, Harakawa N, et al. AML1-ETO downregulates the granulocytic differentiation factor C/EBPa in t(8;21) myeloid leukemia. Nat Med. 2001;7:444–451. 57. Heath V, Suh HC, Holman M, et al. C/EBPa deficiency results in hyperproliferation of hematopoietic progenitor cells and disrupts macrophage development in vitro and in vivo. Blood. 104:1639–1647. 58. Zhang P, Iwasaki-Arai J, Iwasaki H, et al. Enhancement of hematopoietic stem cell repopulating capacity and self-renewal in the absence of the transcription factor C/EBPa. Immunity. 2004;21:853–863. 59. Wierenga AT, Schepers H, Moore MA, Vellenga E, Schuringa JJ. STAT5-induced self-renewal and impaired myelopoiesis of human hematopoietic stem/progenitor cells involves down-modulation of C/EBPa. Blood. 2006;107:4326–4333.