mNotch1 signaling and erythropoietin cooperate in erythroid differentiation of multipotent progenitor cells and upregulate β-globin

Experimental Hematology 35 (2007) 1321–1332

mNotch1 signaling and erythropoietin cooperate in erythroiddifferentiation of multipotent progenitor cells and upregulate b-globin

Konstanze Henninga,*, Timm Schroederb,c,*,Ralf Schwanbecka, Nikolaus Rieberb, Emery H. Bresnickd, and Ursula Justa,b

aDepartment of Biochemistry, Christian-Albrechts University Kiel, Kiel, Germany; bInstitute of Clinical Molecular

Biology and Tumour Genetics; cInstitute of Stem Cell Research, GSF-National Research Centre for Environment andHealth, Munich, Germany; dDepartment of Pharmacology, University of Wisconsin Medical School, Madison, Wis., USA

(Received 20 July 2005; revised 23 May 2007; accepted 24 May 2007)

Objective. In many developing tissues, signaling mediated by activation of the transmembranereceptor Notch influences cell-fate decisions, differentiation, proliferation, and cell survival.Notch receptors are expressed on hematopoietic cells and cognate ligands on bone marrowstromal cells. Here, we investigate the role of mNotch1 signaling in the control of erythroiddifferentiation of multipotent progenitor cells.

Materials and Methods. Multipotent FDCP-mix cell lines engineered to permit the conditionalinduction of the constitutively active intracellular domain of mNotch1 (mN1IC) by the4-hydroxytamoxifen (OHT)-inducible system were used to analyze the effects of activatedmNotch1 on erythroid differentiation and on expression of Gata1, Fog1, Eklf, NF-E2, andb-globin. Expression was analyzed by Northern blotting and real-time polymerase chain reac-tion. Enhancer activity of reporter constructs was determined with the dual luciferase systemin transient transfection assays.

Results. Induction of mN1IC by OHT resulted in increased and accelerated differentiation ofFDCP-mix cells along the erythroid lineage. Erythroid maturation was induced by activatedNotch1 also under conditions that normally promote self-renewal, but required the presenceof erythropoietin for differentiation to proceed. While induction of Notch signaling rapidlyupregulated Hes1 and Hey1 expression, the expression of Gata1, Fog1, Eklf, and NF-E2remained unchanged. Concomitantly with erythroid differentiation, activated mNotch1 upre-gulated b-globin RNA. Notch signaling transactivated a reporter construct harboring a con-served RBP-J (CBF1) binding site in the hypersensitive site 2 (HS2) of human b-globin.Transactivation by activated Notch was completely abolished when this RBP-J site wasmutated to prevent RBP-J binding.

Conclusions. Our results show that activation of mNotch1 induces erythroid differentiation incooperation with erythropoietin and upregulates b-globin expression. � 2007 ISEH - Societyfor Hematology and Stem Cells. Published by Elsevier Inc.

Notch proteins are a family of highly conserved transmem-brane receptors that transduce signals involved in control ofcell fate determination [1,2]. Notch receptors are activatedby specific transmembrane ligands of the Delta and Serrate/Jagged family. After ligand binding, the intracellulardomain of Notch (NotchIC) is proteolytically cleaved

*K. Henning and Dr. Schroeder contributed equally to this work.

Offprint requests to: Ursula Just, M.D., Department of Biochemistry,

Christian-Albrechts University Kiel, Olshausenstraße 40, 24098 Kiel, Ger-

many; E-mail: [email protected]

0301-472X/07 $–see front matter. Copyright � 2007 ISEH - Society for Hema

doi: 10.1016/j.exphem.2007.05.014

from the transmembrane region and translocated to thenucleus, where it associates with the transcriptional repres-sor RBP-J, also termed CBF1 [3]. After binding of NotchIC,RBP-J is converted to a transcriptional activator and in con-junction with chromatin remodeling enzymes, componentsof the transcriptional machinery and activity of other cofac-tors induces transcription of downstream target genes [2].In addition to this pathway, RBP-J–independent pathwayshave been reported [4–6]. Presence of several Notch recep-tors (Notch 1–4 in mammals), multiple ligands (Delta 1, 3and 4, Jagged1, and Jagged2 in mammals) and additional

tology and Stem Cells. Published by Elsevier Inc.

mailto:[email protected]

1322 K. Henning et al. / Experimental Hematology 35 (2007) 1321–1332

components that modulate Notch signaling, such as Deltex,Fringe, and Numb, add further diversity to the Notch sig-naling system [2].

Due to the high complexity of the Notch signaling system,the effects of Notch activation are context-dependent andvary depending on experimental models and conditions[1]. Thus, inhibition or delay of terminal differentiation byNotch activation has been described in some systems[7–11], while in other cases Notch signaling was shown topromote differentiation [12–18]. Similarly, Notch activationcan promote proliferation and cell-cycle progression[19–22], while under different conditions it blocks cell-cycleprogression leading to growth arrest [23–25]. Further, Notchactivation can promote survival by inhibiting apoptosis[26,27], but under certain conditions, Notch signaling canalso induce cell death [25,28]. Notch receptors and ligandsare expressed in the mammalian hematopoietic system andseveral studies have implicated Notch signaling in the con-trol of cell fate choices, differentiation, proliferation, andapoptosis during various stages of hematopoietic develop-ment [29,30]. While an important role of Notch signalingin differentiation of lymphocytes is well-established [31],the role of activated Notch in erythroid development ispoorly understood. Differentiation of erythroid cells is tak-ing place in the bone marrow and is tightly controlled bya complex network of different soluble cytokines and bydirect cell-to-cell contact between hematopoietic progenitorcells and bone marrow stroma cells. The essential solubleglycoprotein hormone erythropoietin positively regulateserythropoiesis by preventing apoptosis and stimulating mat-uration and proliferation of erythroid progenitor cellsand erythroblasts by, at least in part, upregulation of theerythroid-specific transcription factor GATA1 [32,33].Although Notch/RBP-J signaling is not essential for genera-tion of primitive and definitive erythroid cells early in devel-opment [10,30,34], several studies indicate an influence ofNotch on erythropoiesis. In cell culture but not in vivo, num-bers of primitive erythroid progenitors were expanded in theabsence of Notch signals [30]. Conversely, activated Notchinhibits development of erythroid cells [35], possibly by sup-pressing GATA1 activity [36], in the human erythroleukemiacell line K562. On the other hand, Notch signaling wasshown to promote erythroid differentiation from human um-bilical cord blood CD34þ cells [37] and to be necessary formaturation to proceed in mouse erythroleukemia cells [26].

In a search for regulators of b-globin transcription, theNotch signaling interactor protein RBP-J was found tobind to an enhancer region within the hypersensitive site2 (HS2) of the human b-globin locus [38,39]. Howevera functional role for this binding has not been established.In this study, we show that Notch1 signaling promoteserythroid differentiation of multipotent hematopoieticprogenitor cells and that Notch/RBP-J signaling upregu-lates b-globin expression via transactivation of HS2 of theb-globin locus.

Materials and methods

Cell cultureK562 were propagated in Iscove’s modified Dulbecco’s medium(IMDM) containing 10% fetal calf serum (FCS). FDCP-mix cellswere maintained (self-renewal conditions) in IMDM supple-mented with 20% pretested horse serum and mouse interleukin(IL)-3–conditioned medium (mIL-3 CM [40]) at a concentrationthat stimulated optimal cell growth, which corresponds to 100 Urecombinant IL-3 per mL. FDCP-mix cells were kept in a densitybetween 6 � 104 to 106 cells per mL. Cells were carefully con-trolled for normal growth rates, factor dependency, and differenti-ation [41]. In some experiments, 5 U per mL erythropoietin(Roche, Mannheim, Germany) were added to the self-renewal me-dium. For activation of the 4-hydroxytamoxifen (OHT)–induciblemN1IC (NERT) FDCP-mix cells, OHT (RBI, St. Louis, MO, USA)was added to the medium at the concentrations indicated. All cellswere regularly checked to be free of mycoplasma contaminationusing a mycoplasma polymerase chain reaction (PCR) enzyme-linked immunosorbent assay kit (Roche).

Erythroid differentiation (E-differentiation) of cells was in-duced by washing the cells once in IMDM and plating 1 � 105

cells per mL in IMDM containing 20% pretested FCS, 5 U permL erythropoietin (Roche), and 5 U per mL IL-3 (Roche). Gran-ulocyte/macrophage differentiation (G/M-differentiation) of cellswas induced by washing cells once in IMDM and plating1 � 105 cells per mL in IMDM containing 20% pretested FCS,150 U per mL granulocyte colony-stimulating factor (G-CSF;Neupogen, Amgen, USA), 2 U per mL IL-3 (Roche), and 1% mu-rine granulocyte macrophage-CSF (GM-CSF) CM, which corre-sponds to 100 U recombinant murine GM-CSF per mL [42].Aliquots were removed at time points indicated for analysis. Toensure optimal growth and differentiation conditions, cells were splitto constant density and fed with fresh differentiation medium every4th day. Viable cells were counted by trypan blue dye exclusion.Differentiation of FDCP-mix cells was monitored by morphologicalscoring of May-Grunwald-Giemsa and O-Dianiside–stained cyto-spin preparations. Considerable care was taken to validate accuratedifferential counts. All differential counts were done on 100 to 200cells in a blinded fashion by T.S. and U.J.

Northern blot analysisTotal RNA extracts were harvested using RNA Stat 60 (Tel-TestInc., Friendswood, TX, USA) according to manufacturer’s instruc-tions. Electrophoresis and Northern blotting was performed usingstandard procedures. Twenty micrograms total RNA was separatedper lane. For hybridization, 32P (Amersham Pharmacia Biotech,Freiburg, Germany) labeled (Random primed DNA labeling kit,Roche) cDNA probes specific for b-globin, GATA1, and GAPDHwere used. To inhibit de novo protein synthesis, 50 mg mL�1 cy-cloheximide (Sigma, St Louis, MO, USA) was added where indi-cated 30 minutes before addition of 1 mM OHT.

Quantitative PCRReal-time PCR was performed as described previously [43]. Briefly,1 mg RNA prepared with the RNeasy kit (Qiagen, Hilden, Germany)was reverse-transcribed into cDNA using the First Strand cDNASynthesis Kit in a 20 mL volume (Fermentas, St. Leon-Rot, Ger-many). Relative expression levels of the genes Hes1, Hey1, Gata1,Fog1 (Zfpm1), Eklf, NF-E2 (Nfe2), and Hbb-b2 (b-globin minor)

1323K. Henning et al./ Experimental Hematology 35 (2007) 1321–1332

were screened by real-time PCR on a 7900HT Fast Real-Time PCRSystem (Applied Biosystems, Foster City, CA, USA) in 384-wellPCR plates (ABgene) using the TaqMan Gene Expression Assays-on-Demand (Applied Biosystems; assay numbers Mm00468601_m1, Mm00468865_m1, Mm00484678_m1, Mm00494336_m1,Mm00516096_m1, Mm000801891_m1 and Mm00731743_mH,respectively).

Expression of the genes was normalized to Gapdh (AppliedBiosystems) as an endogenous control and quantified using a rela-tive standard curve. Induction by OHT treatment was calculatedusing normalized relative quantities.

Reporter constructs, transienttransfections, and luciferase assaysMutation of the RBP-J binding site within the HS2 site (HS2gmut,see Fig. 5A) was performed by site-directed PCR mutagenesis us-ing the High Fidelity PCR Enzyme Mix (Fermentas). pHS2gluc inthe pGL3 basic vector [38] was used as a template and HS2mut for(50-cactctaggcaccgtccatctgggcacacaccct-30) and HS2mut rev (50-gcccagatggacggtgcctagagtgatgactcct-30) as primers. TemplateDNA was digested with DpnI. Mutation of the RBP-J bindingsite was confirmed by sequencing.

FDCP-mix and K562 cells were transfected with the mN1IC-GFP or CMVeGFP plasmids by electroporation as described[18]. Luciferase reporter assays were done as described [17]. Inbrief, 5 � 106 cells were transfected by electroporation with 18mg RBP-J-luc (pGa981-6, firefly luciferase reading frame underthe control of a minimal b-globin promoter and 12 RBP-J bindingsites [44]), 0-luc (as RBP-J-luc without the RBP-J binding sites[44], pHS2gluc (firefly luciferase reading frame under the controlof a minimal g-globin promoter and the HS2 core region (NCBI:U01317 nt 8485–8872 [38]), pgluc (as pHS2gluc without HS2[38]), or pHS2gmut and 3 mg pTK-renilla plasmids (constitutiveexpression of Renilla-luciferase for transfection efficiency con-trol), respectively. Cells were treated with different concentrationsof OHT for 16 hours and measurement of luciferase activitieswere performed using the Dual Luciferase Kit (Promega) accord-ing to manufacturer’s instructions.

FACS analysisPhycoerythrin- and allophycocyanin-conjugated monoclonal anti-bodies directed against Ly76 (clone Ter119) and CD11b (Mac-1,clone M1/70), respectively, or their respective isotype controlswere used (all Pharmingen, Europe). Cells were harvested by cen-trifugation and resuspended in phosphate-buffered saline contain-ing 3% FCS. Fc-Block (#01241D, Pharmingen, Europe) wasadded at a dilution of 1:100 for 5 minutes at room temperature.Subsequently, antibodies were added at a dilution of 1:100. Afteran incubation for 20 minutes at room temperature in the dark, cellswere washed and resuspended in phosphate-buffered saline con-taining 1% FCS and 1 mg per mL propidium iodide or 7-amino-actinomycin D for dead cell exclusion. Fluorescent-activated cellsorting (FACS) analysis was done with a Becton Dickinson FACS-Calibur machine and Cell Quest software, or a Becton DickinsonFACSCanto machine and FACSDiva software using standardprocedures.

Determination of apoptotic statusThe percentage of cells undergoing apoptosis was determinedquantitatively using the ApoAlert FITC-Annexin V Apoptosis

Kit (Clontech, Mountain View, CA, USA) as described previously[23]. Propidium-iodide negative, Annexin V–positive cells werescored as apoptotic.

Statistical analysisStatistical differences were assessed using Student’s t-test forpaired data.

Results

Expression of the activatedintracellular domain of mNotch1 promoteserythrocyte differentiation of FDCP-mix cellsRecently, we established clones of FDCP-mix cells(rNERTneo [17]), which express a fusion protein of theactivated intracellular domain of mNotch1 and thehormone-binding domain of the human estrogen receptor(NERT [18]), thus allowing conditional induction ofmNotch1/RBP-J signaling by the OHT-inducible system.In the absence of OHT, the NERT protein is exclusivelylocated in the cytoplasm and the Notch/RBP-J signalingpathway is not activated. After addition of OHT, theNERT protein translocates to the nucleus, and the Notch/RBP-J signaling pathway is activated in a concentration-dependent manner [17]. For this study, five independentlyderived rNERTneo FDCP-mix clones (clone 22, 24, 25,26, and 32) were used. Control FDCP-mix clones carryingthe geneticin resistance conferring vector only (rneo 1, 3,and 5) do not transactivate the Notch/RBP-J signaling path-way in the presence or absence of OHT [17].

Self-renewal and differentiation of the multipotentialFDCP-mix progenitor cell line may be modulated by theculture conditions used (Fig. 1A, G/M-differentiation andE-differentiation): In the presence of high IL-3 and horseserum, FDCP-mix cells proliferate as undifferentiatedblasts. However, when the cells are cultured in the presenceof fetal calf serum instead of horse serum, the concentrationof IL-3 in the medium is reduced and lineage-specific cyto-kines are added to the medium, FDCP-mix cells differenti-ate into mature hematopoietic cells [45]. In the presence oferythropoietin (erythroid differentiation conditions [46]),differentiation of FDCP-mix cells into mature enucleatederythroid cells is paralleled by upregulation of GATA1and b-globin expression (Fig. 1B). To assess the effect ofactivated mNotch1 on erythroid differentiation of FDCP-mix cells, rNERTneo FDCP-mix cells were thus culturedin erythroid differentiation conditions and in the presenceor absence of OHT, and were monitored for changes inmorphology. Irrespective of the addition of OHT, rNERT-neo FDCP-mix cells and rneo control clones differentiatedboth predominantly, as the parental FDCP-mix cells, alongthe erythroid lineage (Fig. 2B). Thus, under these condi-tions, activation of mNotch1 signaling neither blockserythroid differentiation of FDCP-mix cells nor shifts


Figure 1. Expression levels of RNAs specific for erythroid FDCP-mix differentiation. To analyze regulation of specific RNAs during differentiation of

FDCP-mix cells, A4 FDCP-mix cells were differentiated under erythroid and G/M conditions for up to 11 days (in the absence of activated Notch). The

differentiation status of the cultures was followed by analyzing the morphology of May-Grunwald-Giemsa–stained cytospin preparations every 2 days.

(A) Morphologies of cells at the start and end of the differentiation are shown. (B) Total RNA was harvested at the time points indicated and expression

levels of specific RNAs were analyzed by Northern blot analysis.

differentiation to another lineage. However, in the rNERT-neo clones, time course experiments revealed an acceler-ated onset of erythroid differentiation concomitant withan increase of mature, benzidine-positive erythroid cellsin the presence of OHT, when compared with cells differen-tiated in the absence of OHT (Fig. 2A). These results werefurther confirmed by FACS analyses using an antibodydirected against the erythroid cell surface marker Ter119(Fig. 3). Although in line with our earlier results [17,23],the total cell numbers of the rNERTneo FDCP-mix cellswere reduced in the presence of OHT in erythroidconditions to about 60% (data not shown), differentiatingcultures of rNERTneo FDCP-mix cells contained a signifi-cantly higher absolute number of differentiated erythroidcells in the presence of OHT (p ! 0.01). Differentiationof rneo FDCP-mix was unaltered by the addition of OHT(Figs. 2 and 3), indicating that OHT itself has no effecton erythroid differentiation of FDCP-mix cells. Thesedata show that under conditions that allow erythroid differ-entiation of FDCP-mix cells, activated mNotch1 acceleratesand enhances the differentiation into erythroid cells.

Because we have recently shown that activated mNotch1induces myeloid differentiation of FDCP-mix cells underconditions that favor self-renewal [17], we asked whetheractivated mNotch1 would also promote erythroid differen-tiation of FDCP-mix cells in the presence of high

concentrations of IL-3 and horse serum (self-renewalconditions), when erythropoietin is added to allow forterminal erythroid differentiation. In the absence of OHT,all cells showed a predominantly blast cell morphology(data not shown). In contrast, in the presence of OHT, allrNERTneo FDCP-mix cell clones differentiated into matureerythroid cells, granulocytes, macrophages, and dendriticcells, whereas in control clones the cells remained mostlyundifferentiated (Table 1). Taken together, these data indi-cate that activated mNotch1 induces differentiation ofFDCP-mix cells along the myeloid and erythroid lineagesunder conditions that favor self-renewal of FDCP-mix cells.

We next analyzed if activated mNotch1 accelerates ery-throid differentiation via modulation of apoptosis. Apopto-sis was assessed using Annexin V in the presence orabsence of OHT under erythroid differentiation conditions,and under self-renewal conditions with erythropoietin pres-ent in the medium. Neither rNERTneo nor control rneoFDCP-mix showed a difference in the proportion of mye-loid and erythroid cells staining for Annexin V in the pres-ence or absence of OHT (Table 2). Thus, a substantialinfluence of activated mNotch1 on apoptosis was notdetectable in differentiating FDCP-mix cells in the presenceof erythropoietin. This rules out that the increase in ery-throid differentiation by activated mNotch1 is mediatedby selective induction or protection of apoptosis.


Figure 2. Activated mNotch1 accelerates cytokine induced erythrocyte differentiation. Three rneo and four rNERTneo clones were cultured in duplicates

under cytokine conditions that favor erythroid differentiation (reduction of interleukin-3, addition of erythropoietin) and in the presence or absence of 4-hy-

droxytamoxifen (OHT; 50 nM, 250 nM, or 500 nM, respectively). Differentiation of the cultures was followed by scoring the morphology of May-Grunwald-

Giemsa–stained cytospin preparations of the cells. Results for one representative rneo and rNERTneo clone treated with 50 nM OHT are shown. The exper-

iment was repeated six times with virtually identical results. Acceleration of cytokine-induced differentiation by activated Notch is statistically significant for

all rNERTneo clones and time points analyzed (p ! 0.001). (A) Morphology of differentiating rNERTneo and rneo FDCP-mix cells. Arrows depict some

early erythroid cells. The inserted picture shows mature, enucleated erythrocytes, as identified by histochemical staining of hemoglobin by Dianisidine (olive

green staining, filled arrowhead). After 6 days, these cells were frequently observed in OHT-treated rNERTneo cultures, but were not seen without OHT

treatment or in OHT-treated rneo control cultures. Cells were stained with May-Grunwald-Giemsa. Original magnification is 1000�. (B) Time course of

erythroid differentiation of rNERTneo and rneo FDCP-mix cells. Values shown represent the mean 6 SEM of the percentage of the different lineages of

differentiation. B 5 primitive blast cells; GM 5 granulocytic cells and macrophages; E 5 erythroid cells.


Figure 3. Activated mNotch1 accelerates the expression of cell surface markers for erythrocyte maturation. Three rneo and five rNERTneo clones were

cultured under cytokine conditions that favor erythroid differentiation and in the presence or absence of 50 nM 4-hydroxytamoxifen (OHT). Differentiation

in the cultures was followed by a two-color fluorescent-activated cell sorting analysis of living cells using antibodies against Mac-1 and Ter119. Consistent

with their normal counterparts, undifferentiated multipotent FDCP-mix cells express low levels of Mac-1. When FDCP-mix cells differentiate along the

macrophage lineage, Mac-1 expression increases to high levels, whereas Mac-1 expression is lost, and subsequently Ter119 expression is induced, when

the cells differentiate along the erythroid lineage. To follow erythroid differentiation, cells were gated for undifferentiated cells (Mac-1lowTer119�, lower

right quadrant), early erythoid cells that have lost expression of Mac-1 (lower left quadrant) and mature erythroid cell that express Ter119 (upper left quad-

rant). Note that, in line with our previous results [17], cells also differentiate in mature myeloid cells in some clones, as indicated by high Mac-1 expression

levels. Values represent the percentage of cells within a quadrant. Results for representative rneo and rNERTneo clones at day 7 are shown. Ter119 positivity

increased to 68% in rNERTneo clones after addition of OHT to induce Notch signaling, whereas in control cells 13% acquired Ter119 positivity (data not

shown). The experiment was repeated two times with virtually identical results. The acceleration of differentiation by mN1IC is statistically significant (p !0.001 for all rNERTneo clones and time points analyzed).

Activated mNotch1 does notinfluence expression of Gata1, Fog1, Eklf,and NF-E2 but upregulates expression of b-globinTo elucidate the mechanism(s) of differentiation inductionby activated mNotch1, we analyzed the expression levelsof the transcription factors Gata1 [47,48], Fog1 [49], Eklf[50], and NF-E2 [51,52] that are implicated in erythroiddifferentiation, in the presence or absence of OHT byreal-time PCR. During differentiation of FDCP-mix cellsinduced by FCS, low IL-3 and erythropoietin, the expres-sion of Gata1 is upregulated at 4 hours after differentia-

tion induction and is further increased as erythroiddifferentiation proceeds (Fig. 1B). Despite induction oferythroid differentiation in rNERTneo cells by OHT(compare Fig. 2 and Table 1), the expression levels ofGata1, Fog1, Eklf, and NF-E2 remained unchanged com-pared to uninduced cells after Notch activation, whereasthe known Notch target genes Hes1 and Hey1 wereclearly induced (Fig. 4A). This suggests that inductionof erythroid differentiation by activated mNotch1 is notmediated by an upregulation of Gata1, Fog1, Eklf, orNF-E2 expression.


Because the Notch signaling interactor protein RBP-Jwas shown to bind to an enhancer region within the HS2of the human b-globin locus [38,39] and erythroid differen-tiation of FDCP-mix cells was accelerated and maturation

Table 1. Activated Notch induces erythroid differentiation under

cytokine conditions that favor self-renewal

d6 d8

FDCP-mix

clone

Culture

conditions Myeloid Erythroid Myeloid Erythroid

rneo 3 �OHT 3 7 5 11

þ OHT 6 10 1 13

rNERTneo24 �OHT 24 2 30 4

þ OHT 67 9 65 16


þ OHT 30 34 45 25


þ OHT 25 25 46 26


þ OHT 6 41 6 41


þ OHT 3 72 18 51


þ OHT 12 17 17 23

Cells were cultured under cytokine conditions that favor self-renewal and

in the presence of erythropoietin with or without 50 nM of 4-hydroxyta-

moxifen (OHT). Differentiation of the cultures was followed by scoring

the morphology of May-Grunwald-Giemsa–stained cytospin preparations

of the cells. Data from a representative experiment are shown. The exper-

iment was repeated three times with virtually identical results. The in-

crease in erythroid differentiation by activated Notch is statistically

significant for all rNERTneo clones (clones 24–26 and clones 31–33)

and time points analyzed (p ! 0.001), but not for rneo cells (clones 1,

3, and 5; p O 0.5).

Table 2. Activated Notch has no influence on apoptosis of FDCP-mix

cells under erythroid differentiation conditions

Apoptotic cells (%)

FDCP-mix clone Culture conditions Day 4 Day 7

rneo1 �OHT 21 26

þ OHT 1 day 19 27

rneo5 �OHT 5 27

þ OHT 1 day 12 15

rNERTneo24 �OHT 6 14

þ OHT 1 day 6 13


þ OHT 1 day 17 45


þ OHT 1 day 9 7

Cells were cultured under erythroid differentiation conditions. At days 3 or

6, 50 nM 4-hydroxytamoxifen (OHT) was added to part of the cultures and

the apoptotic status of the cultures was measured the following day, re-

spectively. Data from a representative experiment are shown. The experi-

ment was repeated three times with virtually identical results. No

statistically significant induction or reduction of apoptosis by activated

Notch in rNERTneo clones (p O 0.8) or by OHT alone in rneo control

clones (p O 0.7) could be detected.

enhanced by activated Notch, we determined if activatedmNotch1 would directly regulate b-globin expression.Thus, b-globin RNA levels were analyzed in rNERTneocells and rneo control cells under self-renewal conditionsafter induction of mNotch1/RBP-J signaling by OHT andin the presence of cycloheximide to inhibit translation.When erythroid differentiation is induced by cytokines inFDCP-mix cells, b-globin RNA expression is upregulatedby day 2 and strongly increases with the emergence of cellswith erythoid morphology (Fig. 1). Even under self-renewalconditions and in the absence of erythropoietin, b-globinRNA was upregulated already 4 hours after addition ofOHT in rNERTneo, but not in rneo, FDCP-mix cells(Fig. 4B and C). Despite rapid upregulation, cycloheximidereduced induction of b-globin RNA by OHT in rNERTneocells, indicating that either protein synthesis or an unstableprotein in addition may be required to increase b-globinexpression (Fig. 4C).

To clarify this further, we tested whether activated Notchwould regulate HS2 activity in reporter assays. Luciferasereporter constructs containing either the functional RBP-Jbinding site of the HS2 [38] or a mutated sequence(Fig. 5A) that does not bind RBP-J [53] or showed reducedRBP-J binding [38] were transiently transfected intorNERTneo and rneo FDCP-mix cells and luciferase activitywas determined in the presence or absence of OHT. Induc-tion of Notch signaling by the addition of OHT resulted ina dose-dependent transactivation after induction of the con-struct competent to bind RBP-J (Fig. 5B). Addition of OHTto control rneo FDCP-mix cells did not activate transcrip-tion. As expected, the mutant HS2gmut compromised inRBP-J binding was completely impaired in transactivationby OHT (Fig. 5C). These results provide evidence thatNotch signaling upregulates b-globin expression by RBP-J–mediated transactivation via the RBP-J binding sitewithin HS2 of the b-globin locus control region.

In contrast to our results shown here, activated Notch in-hibits erythroid maturation in the human erythroleukemiacell line K562 [35]. To test whether this inhibition in K562cells correlates with a lack of b-globin upregulation by Notchsignaling, we measured the reporter activity of the HS2reporter constructs in K562 cells after transient transfectionof a vector encoding the constitutively active cytosolic do-main of Notch1 (NIC). While NIC strongly transactivateda RBP-J–dependent reporter construct, showing that theRBP-J–dependent pathway is functional in K562 cells,NIC did not alter the transcriptional activity of the HS2reporter construct in K562 cells (Fig. 6). This suggests thatother components necessary for upregulation of b-globinexpression by Notch signaling are missing in K562 cells.

DiscussionNotch receptors play a critical role in cellular differentia-tion in a highly cell-type– and context–dependent manner.


Figure 4. Activated mNotch1 upregulates expression levels of b-globin RNA, but not of Gata1, Fog1, Eklf, and NF-E2 in FDCP-mix cells. FDCP-mix cells

were cultured under erythroid conditions in the absence or presence of 50 or 500 nM 4-hydroxytamoxifen (OHT). Total RNA was harvested at different time

points and expression levels of RNAs were analyzed by real-time polymerase chain reaction. (A) Gata1, Fog1, Eklf, and NF-E2 RNA expression levels are

not altered by 4 hours of Notch activation. RNA level of the known Notch target genes Hes1 and Hey1 are statistically significant upregulated in the presence

of OHT (p ! 0.01). (B) b-globin RNA expression is increased by Notch activation after 4 and 8 hours (p ! 0.02). (C) FDCP-mix cells were cultured under

self-renewal conditions in the absence or presence of OHT and cycloheximide (CHX) for the times indicated. Total RNA was harvested and expression levels

of RNAs were analyzed by Northern blot analysis. The experiment was repeated twice with virtually identical results.

Although components of the Notch pathway were found tobe expressed in hematopoietic cells several years ago [29],the importance of Notch signaling in hematopoiesis is notwell understood. An exception is the regulation of T-celldevelopment, in which the role of Notch signaling hasbeen firmly established [31]. The slow progress indefining the involvement of Notch signaling in hematopoi-etic stem cells and in erythroid differentiation is most likelyrelated to several factors. First, the considerable complexityof the Notch signaling system with membrane-bound andsoluble ligands of two ligand families, that can have anactivating as well as an inhibitory activity in an unclarifieddependence of the cellular context, makes interpretationsdifficult. Second, manipulations of the Notch pathway indefined primary hematopoietic cells are equally difficultto achieve. Third, although cell lines represent a usefulmodel to study the outcomes of Notch signals, the cell linesused thus far to analyze the role of Notch signaling inerythropoiesis are derived from leukemic cells and do notrespond to normal physiological cues. Our inducible Notchactivation system described here has several importantadvantages that should allow molecular and biochemical

analyses of the mechanisms by which Notch controls hema-topoietic cell fate. FDCP-mix cell lines are nonleukemic,nonleukemogenic, and respond to the normal physiologicalstimuli, such as cytokines like erythropoietin and stromalcells in the same way as their normal progenitor cell coun-terparts [45]. Thus, our system combines the advantage ofa cell line in which Notch signaling can be regulated in a de-fined and dose-dependent manner with a normal phenotype.

In the present study we have shown that activatedNotch promotes erythroid differentiation of FDCP-mix cellsin cooperation with erythropoietin. Under conditions thatsupported erythroid differentiation, the activation of Notch1signaling resulted in increased and accelerated differentia-tion of FDCP-mix cells along the erythroid lineage. Further-more, even under conditions that normally promote self-renewal, erythroid differentiation was induced by activatedNotch1, although erythropoietin was still required for sur-vival and proliferation of the generated erythroid cells. Thesedata are in line with our previous work that Notch does notinhibit differentiation of hematopoietic progenitor cells, butpromotes differentiation along several myeloid lineages inthe presence of the respective cytokines GM-CSF, G-CSF,


Figure 5. The RBP-J binding site at HS2 of the b-globin locus is essential for enhancer activity by activated mNotch1. (A) The diagram shows the murine

b-globin locus. Binding sites for transcription factors within the HS2 of the used construct are shown in the middle: GC, a GC-rich region; NF-E2; GT,

repetitive GT residues; E box; RBP-J; GATA1; USF and YY1. Below are part of the DNA sequences of the pHS2gluc and pHS2gmut, in which the

RBP-J binding site is mutated. (B) pHS2gluc and the control plasmid pgluc were transiently transfected into FDCP-mix cells. Cells were cultured under

self-renewal conditions in the presence of different amounts of 4-hydroxytamoxifen (OHT) as indicated, and cell lysates were prepared after 16 hours.

Luciferase activity was measured and normalized to the activity of the renilla luciferase. The rNERTneo24 values of pHS2gluc for 100 and 250 nM

OHT were statistically significant (p ! 0.01). (C) pHS2gluc, pHS2gmut and the control plasmid pgluc were transiently transfected into FDCP-mix cells.

Cells were cultured under self-renewal conditions in presence of 250 nM OHT and cell lysates were prepared after 16 hours. Luciferase activity was measured

and normalized to the activity of the renilla luciferase. The rNERTneo24 value of pHS2gluc was statistically significant (p ! 0.01).

or M-CSF [17,18,43]. The recently shown cross-talk betweenNotch/Hes and JAK-STAT signaling pathways in the nervoussystem [54] raises the possibility that Notch signaling pro-motes erythroid and myeloid differentiation by increasingSTAT phosphorylation and activation induced by hematopoi-etic cytokines. It will be of interest to study the molecularmechanisms of potential signaling cross-talk between Notchand lineage-affiliated cytokine signaling on the hematopoi-etic system. Taken together, our data show that activatedmNotch1 induces differentiation of FDCP-mix cells alongthe granulocyte, macrophage, and erythrocyte lineage butdoes not affect lineage decisions between these three myeloidlineages in the presence of the respective lineage-affiliatedcytokines.

Our results differ from other studies suggesting thatNotch suppressed erythroid differentiation [35,36,55].These studies were based on experiments conducted with

the human embryonic leukemia cell line K562 and themouse adult erythroleukemia cell line F5-5. However,tumor cells differ in many aspects from their normal coun-terparts, e.g., they frequently lack functional p53, which hasbeen shown to regulate Notch signaling [56]. Thus, the out-come of Notch activation in tumor cells may be quite differ-ent or even opposing the effects of Notch signaling innormal cells. This view is further supported by recentdata that the Notch ligand Delta-4 promotes erythroid dif-ferentiation of primary primitive human hematopoietic cellsfrom cord blood [37].

Considering that Notch/RBP-J signaling influences cellfunction via changes in gene expression, induction of ery-throid differentiation may result from increased expressionof a transcription factor driving erythroid differentiation.However, the expression levels of the key erythroid tran-scription factors Gata1, Fog1, Eklf, and NF-E2 remained


unchanged by the activation of Notch1. Thus, the increasederythroid differentiation induced by activated Notch1 is un-likely to result from Gata1, Fog1, Eklf, or NF-E2 mRNAupregulation, although this does not exclude an indirector direct influence of Notch signaling on erythroid tran-scription factor activity.

Another mechanism how Notch increases erythroid dif-ferentiation could be the inhibition of apoptosis of differen-tiating erythroid cells as suggested by earlier studies [26].In our model presented here, apoptosis was not influencedby activated Notch, suggesting a direct influence ofNotch/RBP-J signaling on differentiation independent ofapoptosis. In line with our results, Jang et al. [57] showedthat the effects of Notch1 on cell death and differentiationin erythroleukemia cells are not linked.

Interestingly, Notch signaling upregulated expressionlevels of b-globin RNA. Both, erythropoietin and Notchpromoted erythroid maturation. However, mature erythroidcells appeared considerably more rapidly in response toadditional Notch/RBP-J signaling. Furthermore, upregula-tion of b-globin RNA was also induced by activated Notchin the absence of erythropoietin. Thus, it is unlikely thatactivated Notch promotes erythroid differentiation viaerythropoietin signaling. Rather, Notch/RBP-J signalingincreases further maturation by activating genes necessaryfor the specific function of erythrocytes. In this regard, itis of interest to note that Notch is required for erythroidmaturation of MEL cells to proceed [26]. In earlier studies,the Notch interactor protein RBP-J was identified as aDNA-binding protein that binds to a strong erythroid-specific enhancer within the b-globin locus control region[39]. We have now extended this finding, showing thatNotch signaling can transactivate a reporter gene via theRBP-J binding site located at the HS2 of the b-globin

Figure 6. Activated mNotch1 does not transactivate HS2 of the b-globin

locus in K562 cells. pHS2gluc and pgluc were transiently transfected into

K562 cells. As a control for functional Notch signaling, the vectors RBP-J-

luc and 0-luc were used. NIC-FL, a plasmid that contains the constitutively

active intracellular domain of Notch1, was cotransfected. After 40 hours,

cell lysates were prepared. Luciferase activity was measured and normal-

ized to the activity of the renilla luciferase. Only the RBP-J-luc construct

was statistically significant upregulated by activated Notch1 (p ! 0.05).

LCR in adult murine multipotent hematopoietic progenitorcells. While Notch signaling increased HS2 reporter activ-ity in a dose-dependent manner in the nonleukemic adultmultipotent progenitor cell line FDCP-mix, activated Notchwas completely unable to increase transcription by HS2 inthe human leukemic K562 cell line. K562 cells are ery-throid/megakaryocytic progenitor cells that only expressthe e- and g-globin genes and not the adult stage-specificb-globin genes. On the other hand, the more immature,multipotent FDCP-mix cells are derived from normal adultbone marrow long-term cultures and are considered as anadult model for murine hematopoietic stem cells. Globinswitching during ontogeny is regulated by a sequentialappearance of primitive and definitive lineages and at thetranscriptional level during maturation [58]. While tissue-specific basal transcription is controlled by the respectivepromotors, the enhancer activity of the b-globin LCR re-sides in HS2 and is also tissue-specific [59,60]. Thus, onepossibility could be that the responsiveness of the HS2RBP-J site to Notch signaling only in FDCP-mix cells butnot in K562 cells may be related to developmental specific-ity and/or the different precursor status of these cells.

Alternatively, while FDCP-mix cells are nonleukemic cellsresponding to and requiring the normal physiological stimuli,such as cytokines, erythropoietin, and stromal cells in thesame way as their normal progenitor cell counterparts, K562are factor-independent leukemic cells with impaired terminaldifferentiation. It is thus possible that, because of the leukemicintracellular environment, different complexes assemble atthe HS2 RBP-J site in K562 vs FDCP-mix cells, or a compo-nent required for transactivation may be missing, which canrestrict transactivation by activated Notch at this site.

Taken together, we have shown here that Notch signal-ing promotes erythroid differentiation and maturation incooperation with erythropoietin and upregulates b-globinexpression in multipotent progenitor cells. Our datastrongly suggest that Notch signaling directly targets theb-globin locus via the HS2 RBP-J binding site, but one can-not unequivocally rule out indirect actions. Additional stud-ies will be required to address this. To gain further insightinto the mechanisms of Notch-induced erythroid differenti-ation, it will be interesting to determine the proteinsbinding to the HS2 RBP-J binding site in responsive andnonresponsive cells and at different developmental stages,the epigenetic modifications at the globin LCR and pro-moters in response to Notch/RBP-J signaling and whetherNotch differentially regulates globin gene expressionduring ontogeny.

AcknowledgmentsWe thank P. Chambon for providing reagents, D. Gast, C. Kuklik-Roos, and S. Horn for expert technical assistance and G. Born-kamm and W. Ostertag for helpful discussions. This work wassupported by the Deutsche Forschungsgemeinschaft (SPP 1109


‘‘Embryonic and somatic stem cellsdRegenerative systems forcell and tissue repair’’ and SFB 415 ‘‘Signal Transduction’’ projectB8 to U.J.). E.H.B. is supported by NIH DK50107. This reportrepresents a part of the doctoral thesis by K.H. and N.R.

References1. Artavanis-Tsakonas S, Rand MD, Lake RJ. Notch signaling: cell fate

control and signal integration in development. Science. 1999;284:770–

776.

2. Lai EC. Notch signaling: control of cell communication and cell fate.

Development. 2004;131:965–973.

3. Honjo T. The shortest path from the surface to the nucleus: RBP-J kap-

pa/Su(H) transcription factor. Genes Cells. 1996;1:1–9.

4. Shawber CJ, Nofziger D, Hsieh JJ, et al. Notch signaling inhibits mus-

cle cell differentiation through a CBF1-independent pathway. Devel-

opment. 1996;122:3765–3773.

5. Yamamoto N, Yamamoto S, Inagaki F, et al. Role of Deltex-1 as a tran-

scriptional regulator downstream of the Notch receptor. J Biol Chem.

2001;276:45031–45040.

6. MacKenzie F, Duriez P, Larrivee B, et al. Notch4-induced inhibition of

endothelial sprouting requires the ankyrin repeats and involves signal-

ing through RBP-Jkappa. Blood. 2004;104:1760–1768.

7. Nye JS, Kopan R, Axel R. An activated Notch suppresses neurogene-

sis and myogenesis but not gliogenesis in mammalian cells. Develop-

ment. 1994;120:2421–2430.

8. Kuroda K, Tani S, Tamura K, Minoguchi S, Kurooka H, Honjo T.

Delta-induced Notch signaling mediated by RBP-J inhibits MyoD

expression and myogenesis. J Biol Chem. 1999;274:7238–7244.

9. Calvi LM, Adams GB, Weibrecht KW, et al. Osteoblastic cells regu-

late the haematopoietic stem cell niche. Nature. 2003;425:841–846.

10. Schroeder T, Fraser ST, Ogawa M, et al. Recombination signal

sequence-binding protein Jkappa alters mesodermal cell fate decisions

by suppressing cardiomyogenesis. Proc Natl Acad Sci U S A. 2003;

100:4018–4023.

11. Shen Q, Goderie S, Jin L, et al. Endothelial cells stimulate self-

renewal and expand neurogenesis of neural stem cells. Science.

2004;304:1338–1340.

12. Lowell S, Jones P, Le Roux I, Dunne J, Watt FM. Stimulation of

human epidermal differentiation by delta-notch signalling at the

boundaries of stem-cell clusters. Curr Biol. 2000;10:491–500.

13. Gaiano N, Nye JS, Fishell G. Radial glial identity is promoted by

Notch1 signaling in the murine forebrain. Neuron. 2000;26:395–404.

14. Morrison SJ, Perez SE, Qiao Z, et al. Transient Notch activation ini-

tiates an irreversible switch from neurogenesis to gliogenesis by neural

crest stem cells. Cell. 2000;101:499–510.

15. Uyttendaele H, Closson V, Wu G, Roux F, Weinmaster G, Kitajewski

J. Notch4 and Jagged-1 induce microvessel differentiation of rat brain

endothelial cells. Microvasc Res. 2000;60:91–103.

16. Tanigaki K, Nogaki F, Takahashi J, Tashiro K, Kurooka H, Honjo T.

Notch1 and Notch3 instructively restrict bFGF-responsive multipo-

tent neural progenitor cells to an astroglial fate. Neuron. 2001;29:

45–55.

17. Schroeder T, Kohlhof H, Rieber N, Just U. Notch signaling induces

multilineage myeloid differentiation and up-regulates Pu.1 expression.

J Immunol. 2003;170:5538–5548.

18. Schroeder T, Just U. Notch signalling via RBP-J promotes myeloid

differentiation. EMBO J. 2000;19:2558–2568.

19. Carlesso N, Aster JC, Sklar J, Scadden DT. Notch1-induced delay of

human hematopoietic progenitor cell differentiation is associated with

altered cell cycle kinetics. Blood. 1999;93:838–848.

20. Cereseto A, Tsai S. Jagged2 induces cell cycling in confluent fibro-

blasts susceptible to density-dependent inhibition of cell division.

J Cell Physiol. 2000;185:425–431.

21. Karanu F, Murdoch B, Gallacher L, et al. The Notch ligand Jagged-1

represents a novel growth factor of human hematopoietic stem cells.

J Exp Med. 2000;9:1365–1372.

22. Karanu F, Murdoch B, Miyabayashi T, et al. Human homologues of

Delta-1 and Delta-4 function as mitogenic regulators of primitive

human hematopoietic cells. Blood. 2001;97:1960–1967.

23. Schroeder T, Just U. mNotch1 signaling reduces proliferation of

myeloid progenitor cells by altering cell-cycle kinetics. Exp Hematol.

2000;28:1206–1213.

24. Rangarajan A, Talora C, Okuyama R, et al. Notch signaling is a direct

determinant of keratinocyte growth arrest and entry into differentia-

tion. EMBO J. 2001;20:3427–3436.

25. Morimura T, Goitsuka R, Zhang Y, Saito I, Reth M, Kitamura D. Cell

cycle arrest and apoptosis induced by Notch1 in B cells. J Biol Chem.

2000;275:36523–36531.

26. Shelly L, Fuchs C, Miele L. Notch-1 inhibits apoptosis in murine

erythroleukemia cells and is necessary for differentiation induced by

hybrid polar compounds. J Cell Biochem. 1999;73:164–175.

27. Jehn BM, Bielke W, Pear WS, Osborne BA. Cutting edge: protective

effects of notch-1 on TCR-induced apoptosis. J Immunol. 1999;162:

635–638.

28. Ohishi K, Varnum-Finney B, Flowers B, Anasetti C, Myerson D,

Bernstein ID. Monocytes express high amounts of Notch and undergo

cytokine specific apoptosis following interaction with the Notch

ligand, Delta-1. Blood. 2000;95:2847–2854.

29. Kojika S, Griffin JD. Notch receptors and hematopoiesis. Exp Hema-

tol. 2001;29:1041–1052.

30. Hadland BK, Huppert SS, Kanugo J, et al. A requirement for Notch1

distinguishes 2 phases of definitive hematopoiesis during develop-

ment. Blood. 2004;104:3097–3105.

31. Maillard I, Fang T, Pear WS. Regulation of lymphoid development,

differentiation, and function by the Notch pathway. Annu Rev Immu-

nol. 2005;23:945–974.

32. Pevny L, Simon MC, Robertson E, et al. Erythroid differentiation in

chimaeric mice blocked by a targeted mutation in the gene for tran-

scription factor GATA-1. Nature. 1991;349:257–260.

33. Gregory T, Yu C, Ma A, Orkin SH, Blobel GA, Weiss MJ. GATA-1

and erythropoietin cooperate to promote erythroid cell survival by reg-

ulating bcl-xL expression. Blood. 1999;94:87–96.

34. Mancini SJ, Mantei N, Dumortier A, Suter U, MacDonald HR, Radtke

F. Jagged1-dependent Notch signaling is dispensable for hematopoi-

etic stem cell self-renewal and differentiation. Blood. 2005;105:

2340–2342.

35. Lam L, Ronchini C, Norton J, Capobianco A, Bresnick EH. suppres-

sion of erythroid but not megakaryocytic differentiation of human

K562 erythroleukemic cells by Notch-1. J Biol Chem. 2000;275:

19676–19684.

36. Ishiko E, Matsumura I, Ezoe S, et al. Notch signals inhibit the devel-

opment of erythroid/megakaryocytic cells by suppressing GATA-1

activity through the induction of HES1. J Biol Chem. 2005;280:

4929–4939.

37. Sugimoto A, Yamamoto M, Suzuki M, et al. Delta-4 Notch ligand pro-

motes erythroid differentiation of human umbilical cord blood CD34þ

cells. Exp Hematol. 2006;34:424–432.

38. Lam L, Bresnick EH. A novel DNA-binding protein, HS2FN5, inter-

acts with a functionally important sequence of the human beta-globin

locus control region. J Biol Chem. 1996;271:32421–32429.

39. Lam LT, Bresnick EH. Identity of the beta-globin locus control region

binding protein HS2NF5 as the mammalian homolog of the notch-

regulated transcription factor suppressor of hairless. J Biol Chem.

1998;273:24223–24231.

40. Karasuyama H, Melchers F. Establishment of mouse cell lines which

constitutively secrete large quantities of interleukin 2, 3, 4 or 5,

using modified cDNA expression vectors. Eur J Immunol. 1988;18:

97–104.


41. Spooncer E, Heyworth CM, Dunn A, Dexter TM. Self-renewal and

differentiation of interleukin-3-dependent multipotent stem cells are

modulated by stromal cells and serum factors. Differentiation. 1986;

31:111–118.

42. Schroeder T, Lange C, Strehl J, Just U. Generation of functionally ma-

ture dendritic cells from the multipotential stem cell line FDCP-mix.

Br J Haematol. 2000;111:890–897.

43. Schroeder T, Meier-Stiegen F, Schwanbeck R, et al. Activated Notch1

alters differentiation of embryonic stem cells into mesodermal cell

lineages at multiple stages of development. Mech Dev. 2006;123:

570–579.

44. Strobl LJ, Hofelmayr H, Stein C, et al. Both Epstein-Barr viral nuclear

antigen 2 (EBNA2) and activated Notch1 transactivate genes by inter-

acting with the cellular protein RBP-Jk. Immunobiology. 1997;198:

299–306.

45. Just U, Boettiger D, Kan O, Dexter TM, Spooncer E. Insertional mu-

tagenesis as a route to identifying genes involved in self renewal of

haemopoietic stem cells. Curr Top Microbiol Immunol. 2000;251:

27–34.

46. Heyworth CM, Alauldin M, Cross MA, Fairbairn LJ, Dexter TM,

Whetton AD. Erythroid development of the FDCP-Mix A4 multipo-

tent cell line is governed by the relative concentrations of erythropoi-

etin and interleukin 3. Br J Haematol. 1995;91:15–22.

47. Evans T, Felsenfeld G. The erythroid-specific transcription factor

Eryf1: a new finger protein. Cell. 1989;58:877–885.

48. Tsai SF, Martin DI, Zon LI, D’Andrea AD, Wong GG, Orkin SH.

Cloning of cDNA for the major DNA-binding protein of the erythroid

lineage through expression in mammalian cells. Nature. 1989;339:

446–451.

49. Tsang AP, Visvader JE, Turner CA, et al. FOG, a multitype zinc finger

protein as a cofactor for transcription factor GATA-1 in erythroid and

megakaryocytic differentiation. Cell. 1997;90:109–119.

50. Miller IJ, Bieker JJ. A novel, erythroid cell-specific murine transcription

factor that binds to the CACCC element and is related to the Kruppel

family of nuclear proteins. Mol Cell Biol. 1993;13:2776–2789.

51. Ney PA, Andrews NC, Jane SM, et al. Purification of the human

NF-E2 complex: cDNA cloning of the hematopoietic cell-specific sub-

unit and evidence for an associated partner. Mol Cell Biol. 1993;13:

5604–5612.

52. Andrews NC, Erdjument-Bromage H, Davidson MB, Tempst P, Orkin

SH. Erythroid transcription factor NF-E2 is a hematopoietic-specific

basic-leucine zipper protein. Nature. 1993;362:722–728.

53. Tun T, Hamaguchi Y, Matsunami N, Furukawa T, Honjo T, Kawaichi

M. Recognition sequence of a highly conserved DNA binding protein

RBP-J. Nucleic Acids Res. 1994;22:965–971.

54. Kamakura S, Oishi K, Yoshimatsu T, Nakafuku M, Masuyama N, Go-

toh Y. Hes binding to STAT3 mediates crosstalk between Notch and

JAK-STAT signalling. Nat Cell Biol. 2004;6:547–554.

55. Kumano K, Chiba S, Shimizu K, et al. Notch1 inhibits differentiation

of hematopoietic cells by sustaining GATA-2 expression. Blood. 2001;

98:3283–3289.

56. Laws AM, Osborne BA. p53 regulates thymic Notch1 activation. Eur

J Immunol. 2004;34:726–734.

57. Jang MS, Miao H, Carlesso N, et al. Notch-1 regulates cell death in-

dependently of differentiation in murine erythroleukemia cells through

multiple apoptosis and cell cycle pathways. J Cell Physiol. 2004;199:

418–433.

58. Kingsley P, Malik J, Emerson R, et al. ‘‘Maturational’’ globin switch-

ing in primitive erythroid cells. Blood. 2006;104:1665–1672.

59. Li Q, Peterson K, Fang X, Stamatoyannopoulos G. Locus control

regions. Blood. 2002;100:3077–3086.

60. Stamatoyannopoulos G. Control of globin gene expression during

development and erythroid differentiation. Exp Hematol. 2005;33:

259–271.

Documents

mNotch1 signaling and erythropoietin cooperate in erythroid differentiation of multipotent progenitor cells and upregulate β-globin