PKCδ-Induced PU.1 Phosphorylation Promotes Hematopoietic Stem Cell Differentiation to Dendritic Cells§


  • Disclosure of potential conflicts of interest is found at the end of this article.

  • Author contributions: M.H.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; A.B.: data analysis and interpretation, manuscript writing, final approval of manuscript; S.S.: manuscript writing, final approval of manuscript; J.R.: final approval of manuscript; E.F.: conception and design, data analysis and interpretation, manuscript writing, final approval of manuscript. M.H. and A.B. contributed equally to this article.

  • §

    First published online in STEM CELLSEXPRESS November 23, 2010.


Human CD34+ hematopoietic stem cells (HSCs) exhibit the potential to differentiate into a variety of specialized blood cells. The distinct intracellular mechanisms that control cell fate and lineage commitment of these multipotent cells are not well defined. In this study, we investigate and modulate the signaling processes during HSC differentiation toward myeloid dendritic cells (mDCs). DC differentiation induced by the cytokines Granulocyte macrophage colony-stimulating factor (GM-CSF) and Interleukin-4 (IL-4) led to activation of the Extracellular-signal-regulated kinase (ERK), protein kinase C (PKC), and Janus kinase (JAK)/Signal Transducer and Activator of Transcription (STAT) but not the SAPK/c-Jun NH2-terminal kinase and p38 mitogen-activated protein kinase signaling pathways. From the activated signaling pathways the PKC isoform δ was found to phosphorylate the transcription factor PU.1, which is described as one of the key factors for myeloid HSC differentiation. On molecular level, PKCδ regulated PU.1 activity by affecting its transactivation activity, whereas its DNA binding activity remained unaffected. This was accompanied by PKCδ-induced phosphorylation of the PU.1 transactivation domain. Furthermore, treatment with PKC- and ERK1/2-specific signaling inhibitors impaired both HSC differentiation toward mDCs as well as phosphorylation-mediated transactivation activity of PU.1. Taken together, these results provide new insights into the molecular mechanisms promoting the differentiation process of HSCs toward mDCs and introduce the PKC isoform δ as critical mediator. STEM CELLS 2011;29:297–306


Hematopoiesis involves the maturation of multipotent hematopoietic stem/progenitor cells (CD34+ hematopoietic stem cells [HSCs]) toward the lymphoid or myeloid lineage, which is highly regulated and controlled by an interplay of growth factors and cytokines leading to the hierarchical activation of lineage-specific genes via distinct intracellular signaling pathways. Within the myeloid lineage, dendritic cells (DCs) are the most potent antigen-presenting cells of the immune system that differentiate from HSCs by stimulation with distinct growth factors [1]. It has been described that Granulocyte macrophage colony-stimulating factor (GM-CSF) and Interleukin-4 (IL-4) induce the differentiation to immature DCs, whereas tumor necrosis factor alpha (TNF-α) [2] and Transforming growth factor beta TGF-β [3] stimulation supports maturation in vitro. However, whether the distinct intracellular translation of GM-CSF and IL-4-induced stimulation and the associated signaling pathways are essential for differentiation of HSCs toward myeloid dendritic cells (mDCs) is still elusive.

During the last years, the impact of mitogen-activated protein kinase (MAPK) signaling pathways in the regulation of hematopoiesis has been underpinned. The three main MAPK signaling cascades represented by Extracellular-signal-regulated kinase (ERK), c-Jun NH2-terminal kinase (JNK), and p38 MAPK are considered to play a significant role in cytokine and stress-induced lineage commitment during HSC differentiation [4]. Several groups described that mitogen-activated protein kinase kinase 1 (MEK-ERK) activation is critical for myeloid and megakaryocyte differentiation [5, 6], whereas inactive ERK is a prerequisite for erythroid differentiation [7]. ERK activation results in expression of proliferation and differentiation-associated transcription factors as well as cytokines that regulate the balance between cell survival and expansion in an autocrine manner. Therefore, the ERK signaling pathway seems to play a key role in cell fate determination. Hierarchically located in parallel, the stress-induced p38 and the JNK pathway are indicated to be indispensable for erythroid development [8, 9].

A prominent role during hematopoiesis is also attributed to the protein kinase C (PKC) family of serine/threonine kinases that are classified according to their direct stimuli dependency. The class of novel PKC isoforms (δ, ε, θ, and η) comprises Ca2+-independent but phorbol ester-dependent kinases. Previous findings revealed that, in comparison with other members of the family, novel isoforms are predominantly expressed in hematopoietic cells [10], presumably fulfilling critical functions during lineage commitment. Cytokine-induced differentiation of myeloid progenitor cells toward monocytes and macrophages was described to involve activation of PKCα, PKCδ, PKCε, and PKCζ [11], whereas lymphoid progenitor differentiation was associated with PKCθ activity [12]. PKCδ was reported to be the predominant isoform present in myeloid cells [13] and its overexpression within COS7 cells and the murine myeloid cell line 32D allowed differentiation toward macrophages when treated with phorbol esters [14, 15]. However, the distinct PKC isoforms and key mechanisms that promote differentiation of HSCs toward myeloid DCs are still unknown.

Transcription factors are considered to be downstream targets of MAPK signaling pathways including PU.1, a prominent factor essential for differentiation processes such as the hematopoietic commitment toward the myeloid or lymphoid lineage and the erythroid or thrombocytic system from its progenitors [16–19]. PU.1 acts on stimuli for granulocyte-macrophage lineage commitment and is critical for the development of myeloid-derived but not lymphoid-related DCs [20]. Moreover, PU.1 phosphorylation was found to correlate with PKCδ and ζ activation in B cells [21], indicating their importance during cell maturation. However, to which extend ERK- or PKC-mediated PU.1 activation contributes to myeloid cell differentiation of HSCs has not yet been described.

In this study, we investigated the role of MAPK and PKC signaling pathways in human CD34+ hematopoietic progenitor cells during mDC development. We demonstrate that activation of ERK and the PKC isoform δ in HSCs are indispensable for GM-CSF/IL-4-induced DC differentiation. More intriguingly, our results indicate that PKCδ induces PU.1 activity by phosphorylation of its transactivation domain (TAD) to promote mDC-directed HSC differentiation.


Human CD34+ Culture, Differentiation, and Cytokine Array

CD34+-enriched cell suspensions were prepared from umbilical cord blood by positive selection (Lonza, Cologne, Germany, Increased cell numbers were obtained by culturing 0.5–1 × 106 CD34+ cells for 7 days in Stemspan SFEM (STEMCELL Technologies SARL, Grenoble, France, supplemented with Flt-3 ligand (100 ng/ml), SCF (100 ng/ml), IL-3 (20 ng/ml), and IL-6 (20 ng/ml). A total of 5 × 105 cells were seeded into a 12-well culture plate with 2 ml SFEM supplemented with GM-CSF (100 ng/ml) and IL-4 (20 ng/ml; GMP grade; Cellgenix, Freiburg, Germany, On day 7, cells were phenotypically and functionally analyzed for the immature DC state or cultivated for another 2 days in the presence of TNF-α (20 ng/ml) to induce maturation of the DCs. For the colony-forming assay, 5 × 102 CD34+ cells were embedded in MethoCult GF H4434 (STEMCELL Technologies SARL, Grenoble, France, according to the manufactures manual. The media supports the growth of erythroid progenitors (colony-forming unit erythroid (CFU-E) and burst forming unit erythroid (BFU-E)), granulocyte-macrophage progenitors (colony-forming unit granulocyte macrophage (CFU-GM)), colony-forming unit macrophage (CFU-M), colony-forming unit granulocyte (CFU-G)), and multipotential granulocyte, erythroid, macrophage, megakaryocyte progenitors (colony-forming unit granulocyte erythrocyte monocyte macrophage (CFU-GEMM). The embedded cells were plated on 35-mm dishes in triplicates and incubated at 37°C, 5% CO2 and 95% humidity for 10–14 days. The number of colonies was determined via microscopy. For the inhibition studies, the MethoCult media was supplemented with 5 μM U0126 (Promega, Mannheim, Germany, 1 μM Gö6986, or 250 nM Janus kinase (Jak) Inhibitor I (Calbiochem, Darmstadt, Germany, For activation of HSCs, we used 100 ng/ml GM-CSF and 20 ng/ml IL-4 or TPA (short stimulation ≤60 minutes, 100 ng/ml; long-term stimulation ≥24 hours, 20 ng/ml). For the cytokine array, DCs were stimulated with 10 μg LPS/ml medium and analyzed using a TNF-α instant enzyme-linked immunosorbent assay (ELISA) (Bender MedSystems, Vienna, Austria, and the “Human Cytokine Set 2 Array” (Biosource, Solingen, Germany, following the manufacturers' instructions.

Flow Cytometric Analysis

DCs were detached with trypsin/EDTA (0.05% trypsin) in the presence of 1% chicken sera, washed twice (PBS, 5% fetal calf serum FCS), and resuspended in FcR-blocking reagent (Milteny Biotec, Bergisch Gladbach, Germany, After 10 minutes, the cells were resuspended in staining buffer (PBS, 5% FCS, 0.1% sodium azide). HSCs were washed twice (PBS, 5% FCS) and resuspended in staining buffer. A total of 2 × 105 cells were stained with saturating concentrations of monoclonal antibodies against CD117 (104D2 (Biolegend, Fell, Germany,, CD34 (Beckman Coulter), and corresponding isotype controls. A total of 1 × 105 cells were analyzed using the LSR II flow cytometer (BD Biosciences, Heidelberg, Germany,

Cell lines, Transfection, and Inhibitors

HeLa, Alpha 1, NIH 3T3, and 293T cells were maintained in Dulbecco's modified Eagle's medium (supplemented with 10% fetal bovine serum, 0.29 mg/ml L-glutamine, and 100 units/ml penicillin/streptomycin (Invitrogen, Darmstadt, Germany, http:/ HOS cells were maintained in Eagle's minimum essential medium supplemented with 10% fetal bovine serum, 0.29 mg/ml L-glutamine, and 100 units/ml penicillin/streptomycin. Human A3.01 T cells (National Institute for Biological Standards and Control, U.K.), KG-1a, K-562, and U937 cells were grown in Roswell Park Memorial Institute medium (RPMI) 1640 medium supplemented with 10% fetal bovine serum, 0.29 mg/ml L-glutamine, and 100 units/ml penicillin/streptomycin. Cells were incubated at 37°C with 100% humidity in 5%–7% CO2. For transfection of NIH 3T3 and 293T cells, we used the reagent Lipofectamine Plus (Invitorgen), HeLa cells were transfected with FuGene (Roche, Mannheim, Germany,, and for suspension cells A3.1 and U-937 cells, we used DMRIE-C reagent (Invitrogen) following the manufacturer's instructions. Cell line activation with TPA was performed as described for HSC treatment. For the inhibition of the MAPK/ERK signaling pathway, we used the MEK inhibitor U0126 (5-10 μM, 1,4-diamino-2,3-dicyano-1,4-bis(2-aminophenylthio)butadiene; Promega), the “ERK inhibitor” (30 μM, 3-(2-aminoethyl)-5-((4-ethoxyphenyl)methylene)-2, 4-thiazolidinedione), the PKC inhibitor Gö6983 (1 μM, 2-[1- (3-dimethylaminopropyl)-5-methoxyindol-3-yl]-3-(1H-indol-3-yl)maleimide), and the JAK inhibitor I (250 nM, 2-(1,1-quinolin-7-one; all Calbiochem).

Plasmid Construction

The HA-PU.1 gene was generated by reverse transcription polymerase chain reaction (RT-PCR) (Superscript III, Invitrogen) on RNA isolated from human KG-1a and K-562 cells. To generate the HA-PU.1 insert, we used the forward primer HA-PU.1 sense (5′-CCACCATGTACCCCTACGACGTGCCCGAC TACGCCG GCGAAGG GTTTCCCCTCGTCCCCCCTCCATCAG AAGACCTGGTGCCC-3′) and reverse primer HA-PU.1 antisense (5′-TCAGTGGGGC GGGTGGCGCCGCTGGCCAGGC CCCCGC GGCCCAG-3′). All amplicons were generated using the Phusion High-Fidelity PCR Kit (Finnzymes, Espoo, Finland, according to the manufacturer's manual. The amplicon was cloned into the TA-cloning vector pTARGET (Promega) and the BamHI- and SalI-cut fragment was subcloned into pIRES2-EGFP (Clontech, Saint-Germain-en-Laye, France, The transactivation activity of PU.1 was visualized by a luciferase reporter system based on the TransLucent pTL-Luc plasmid (Panomics, Fremont, CA, The five GAL4 DNA binding motifs were replaced by five PU.1 binding sites AGGAAG. For the replacement of the motif, the pTL-Luc plasmid was digested with the restriction enzymes HindIII and XhoI and the backbone was ligated to the annealed primer pairs sense (5′-AGCTTGCATGCCGCAGGTGGAAGAGGAAGAGGAAGAGGAAGAGGAAGAAGCGAACTCTAGAGGGTATATAATG GA TCC-3′) and antisense (5′-CCGGGGATCCATTATATA CCTC TAGAGTCTCCGCTTCTTCCTCTTCCTCTTCCTCTTCCTCTTC CACCTGCAGGCATGCA-3′). To generate the plasmid pTL-PU.1 encoding for the PU.1 TAD-Gal4 DNA binding domain (DBD) fusion protein, the TAD domain encoding sequence of PU.1 was amplified by PCR using the primers sense (5′-GGGATC CATGGAGGGTTCCCCCGTCCC-3′) and antisense (5′-CCGAAT TCTCACTGCTCCAGCTCCATGTGGCGG-3′). The BamHI- and EcoRI-digested fragment was cloned into pTL-TAD (Panomics).

RT-PCR and siRNAs

RNA was prepared from human HSCs, positive selected (Miltenyi Biotec) monocytes (CD14+) cells from human blood peripheral blood mononuclear cells (PBMC) and myeloid DC generated from HSCs and monocytes using the RNeasy mini kit (Qiagen, Hilden, Germany, according to the manufacturer's protocol. Two micrograms of total RNA was used to generate cDNA using SuperScript III reverse transcriptase (Invitrogen; poly dT primer). PCR was performed using the following primer pairs: Cathepsin-C, forward (5′-TTTCTCAGCTCCCTGCAGCA-3′) and reverse (5′-CATGCACCCACCCAGTCATT-3′) DC-CK1, forward (5′-GCCAGGTGTC TCCTCCTAA-3′) and reverse (5′-GGCAC AATGTCTGCTGAGAA-3′) DC-SIGN, forward (5′-ACAGAGG AGAGCCCAACAACG-3′) and reverse (5′-GGTCGAAGGATGG GAGA GGA-3′) MCP-4, forward (5′-ATGACAGCAGCTTTCA ACCCC-3′) and reverse (5′-CTCCAAACCAGCAACAAGTCAA AT-3′) CD11b, forward (5′-CTCATGCTCAAAG AGCTGCTA-3′) and reverse (5′-CGAAAGAATGCCTTTAAGA′TCCTAG-3′) CCRD-BP-IMP1, forward (5′-GCTTTACATCGGCAACCTCAA CG-3′) and reverse (5′-TAGGAGACC TCAGGGCATGGTT-3′) 2-microglobulin, forward (5′-CTCGCTCCGTGGCCTTAGCTG GCTCGCGC-3′) and reverse (5′-TAACTTATGCACGCTTAAC TATC-3′). The cDNAs were amplified with 35 cycles of PCR using the Phusion polymerase (NEB, Frankfurt am Main, Germany, under the following conditions: denaturation for 10 seconds at 98°C, primer annealing for 30 seconds at the temperature of 50°C, and primer extension for 1 minute at 72°C. SiRNA-mediated knockdown of human ERK, PKCδ, and PU.1 in HeLa cells was performed by transfection of RNA duplexes: MAPK/ERK: Hs_MAPK3_5/Hs_MAPK1_10 siRNA (SI00300776, SI00300755; Qiagen) PKCδ: Hs_PRKCD_7 siRNA (SI00301329; Qiagen). For PU.1 silencing, we used a PU.1-specific siRNA with the target sequence 5′-AAGCTCACCT ACCAGTTC-3′ (Eurofins MWG Biotech, Ebersberg, Germany, Rmc0Q) as previously described [22]. For control, a nonspecific siRNA with the target sequence 5′-AAUUCUCCGAACGUGUCACGT-3′ was used (1022076; Qiagen). A total of 150–750 ng of siRNA were transfected into HeLa cells using the HiPerfect fast-forward protocol (Qiagen) for six-well plates according to the manufacturer's instructions. For cotransfection of siRNAs and the PU.1 reporter/expression plasmids, 5 × 105 HeLa cells were seeded in six-well plates. Next day, 1 μg of pPU.1-Luc reporter, 0.25 μg pIRES2-HA-PU.1-EGFP, and 750 ng of total siRNA were diluted in 50 μl of HEPES-buffered saline (HBS) buffer (50 mM HEPES, 150 mM NaCl, pH 7.4), and 12 μl of HiPerfect transfection reagent was diluted in 50 μl of HBS buffer. Both solutions were mixed and incubated at room temperature. After 15 minutes of incubation, the solution was added drop wise to the cells.

Immunoprecipitation and In Vitro Kinase Assay

HA-PU.1-expressing NIH 3T3 cells (transfected with 1 μg pIRES2-HA-PU.1-EGFP) were lysed in radio-Immunoprecipitation Assay (RIPA) buffer. The cleared lysate was incubated with HA-beads from Roche (Mannheim) for 1–2 hours at 4°C under gentle agitation. Subsequently, the beads were washed in lysis buffer. The precipitated proteins were either used for immunoblot analysis or for kinase assays. For kinase assays, the HA-PU.1 fusion protein was eluted from the HA-beads using 30 μl HA peptide solution (1 mg/ml HA peptide [Roche] in PBS). Elution was accomplished at 30°C for 10 minutes under gentle agitation and collection of the supernatant after centrifugation. In vitro kinase assays were performed using 2.5 ng recombinant ERK2 or PKCδ (Upstate) together with 10 μl HA-PU.1 eluate from immunoprecipitation, 20 μl MAP kinase or PKC buffer and 1 μl γATP32 (10 μCi) for 20 minutes at 30°C under gentle agitation. For the analysis of PU.1-TAD phosphorylation, a GST- fusion protein (GST-TAD-PU.1; Aa 1-120 of PU.1 Gene Bank Acession NM_003120, Abnova, Heidelberg, Germany, was used. As control substrate, recombinant Elk1 (Upstate, Schwalbach/Ts, Germany, was used for ERK2, and histone H1 was used for PKCδ. The specificity of the kinase activity was verified using 30 μM ERK inhibitor or 1 μM PKC inhibitor Gö 6983 (Calbiochem, Darmstadt).

Immunoblot Analysis

A total of 5 × 105 to 1 × 106 HSCs were stimulated for the indicated periods with 100 ng/ml GM-CSF + 20 ng IL-4 (Cellgenix) or 20–100 ng/ml TPA with or without 0.5 μM ionomycin and 10 μg/ml anisomycin (Calbiochem). The stimulation was stopped by the addition of ice-cold phosphate buffered saline (PBS). The cells were washed twice and lysed in RIPA buffer (25 mM Tris, pH 8.0, 137 mM NaCl, 1% glycerol, 0.5% sodium deoxycholate, 1% Nonidet P-40, 2 mM EDTA,pH8, 0.1% SDS, and protease inhibitors [Roche]). Lysates were cleared by centrifugation, boiled in Laemmli buffer (Carl Roth), and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (gradient gel 4%–12%, Invitrogen) followed by transfer to a nitrocellulose membrane (GE München, Germany, The proteins were detected using the following primary antibodies: anti-phospho p44/42 (Thr202/Tyr204), anti-phospho SAPK/JNK (Thr183/Tyr185), anti-phospho p38 (Thr180/Tyr182), anti-phospho JNK2 (Tyr 1007/1008), anti-phospho STAT5 (Tyr694), anti-phospho STAT6 (Tyr641), anti-phospho PKCδ (Thr505), anti-p44/42, anti-p42, anti-SAPK/JNK, anti-p38, anti-STAT5, anti-STAT6, anti-PKCδ (all Cell Signaling Technologies, Frankfurt am Main, Germany,, anti-JNK2 (Chemikon), anti-tubulin [YT1/2] (Abcam), anti-PU.1 [H-135] (Santa Cruz), anti-HA [12CA5/HA.11] Roche/Covance), and anti-CKII (BETHYL, Hamburg, Germany, using the dilutions suggested by the manufacturers. Secondary horseradish peroxidase-coupled antibodies were purchased from GE (Anti-mouse and anti-rabbit) as well as Abcam (anti-rat). Immunblot signals were visualized by ECL and ECL Plus (GE).

Luciferase Assay

HeLa cells were seeded in six-well tissue culture dishes at 5–7 × 105 cells per well. The following day, each well was transfected with 1 μg of reporter plasmid pPU.1-Luc Reporter and 0.25 μg of the expression plasmid pIRES2-HA-PU.1-EGFP or the control plasmid pIRES2-EGFP using 3.75 μl of FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's protocol. For transient transfection of A3.01 cells, DMRIE-C (Life Technologies, Invitrogen) was used as described earlier [23]. Twenty-four hours post-transfection, the inhibitors were added, and 30 minutes later, the cells were stimulated using 20 ng/ml TPA overnight. On the next day, cells were analyzed by the luciferase reporter assay system following the manufacturer's protocol. Relative light units were normalized to determined protein amounts.

Electrophoretic Mobility Shift Assay

For preparation of protein extracts, 1 × 106 CD34 cells were washed in cold PBS. The cell pellet was lysed by three freeze-thaw cycles. After centrifugation, the supernatant was used for incubation with radioactive-labeled DNA probes. Electrophoretic mobility shift assay (EMSA) probes were generated by annealing the following complementary oligonucleotides: probe I PU.1 motif: 5′-CTTTTACGGGAAGTCC-3′ and 5′-GGACTTCCCGT AAAAG-3′; probe II mutated PU.1 motif: 5′-CTTTTAGGAT CCGTCC-3′ and 5′-GGACGGATCCTAAAAG-3′ according to previously described procedures [24]. The double-stranded oligonucleotides were 5′ end-labeled using T4 polynucleotide kinase (New England Biolabs) and [γ-32P]-ATP (3,000 μCi/mmol, GE) and purified by using Nick G50 columns (Amersham, GE). For EMSA, 5 μg cellular extracts of CD34+ cells were preincubated on ice with 2 μg of poly(dI-dC) (Roche) as an unspecific competitor and 1 μg of bovine serum albumin in band shift buffer (50 mM Tris, 150 mM KCl, 5 mM EDTA, 2.5 mM dithiothreitol, 20% Ficoll) for 15 minutes. 32P-labeled oligonucleotides (50,000 c.p.m.) were added in a total volume of 20 μl, incubated on ice for 20 minutes and loaded onto 5% native polyacrylamide gels in Tris-borate-EDTA buffer. On fractionation, gels were dried and exposed for autoradiography. For competition experiments, 10-fold molar excess of the unlabeled NFκB oligonucleotide was added to the preincubation mixture. For supershift experiments, 2 μg PU.1 antibody (T-21; sc-352 x, Santa Cruz Biotechnology, Heidelberg, Germany, was added to the preincubation mixture for further 30 minutes.


DC Differentiation from Human CD34+ HSCs

To identify intracellular signaling pathways involved in the regulation of human cord blood-derived CD34+ HSC differentiation, we used a human ex vivo expansion [25, 26] and differentiation [1, 27] system allowing differentiation of HSCs into myeloid DCs. HSCs were expanded for 5 days and thereby maintained a typical round and blast stem cell morphology. Up to 95% of the expanded cells were positive for the stem cell markers CD34 and CD117 (Fig. 1A). Analyses of further cell-specific markers conform to the observed cell phenotype (Supporting Information Fig. 1). After expansion, the progenitor cells were incubated with GM-CSF and IL-4 to induce DC differentiation. Until day 7, the cells developed DC-characteristic veils (Fig. 1A). On day 10, generation of DCs has been confirmed by flow cytometric analysis (FACS) measuring downregulation of CD34 and CD117 (Fig. 1A). Co-occurring, the DC-specific surface markers CD1a, CD11c, major histocompatibility complex (HLA-DR), and immunoglobulin-like transcript 3 (ILT-3) were upregulated (Supporting Information Fig. 1). The mRNA levels of cathepsin-C, dendritic cell-derived chemokine-1 (DC-CK1), monocyte chemotactic protein-4 (MCP-4), Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin (DC-SIGN), CD11b, and Coding region determinant-binding protein/insulin-like growth factor II mRNA-binding protein (CRD-BP/IMP1) have been analyzed by reverse transcription polymerase chain reaction (RT-PCR) for HSC and HSC-derived DCs. These genes are considered to be specifically upregulated and downregulated during DC differentiation [28–30], and as expected, their mRNA expression profile matched to the phenotypes of HSCs and HSC-derived mDCs (Fig. 1B). Moreover, the obtained mRNA profile of mDCs was comparable with that of monocytes-derived DCs (Fig. 1B). The functional analysis of DCs in terms of cytokine secretion on Lipopolysaccharides (LPS) exposure revealed secretion of IL-6, monocyte chemoattractant protein-1 (MCP-1; CCL2)MCP-1) (CCL2), and RANTES (CCL5) within 6 or 24 hours, respectively (Fig. 1C) as well as TNF-α within 6 hours (data not shown). These results are in line with previous findings [31]. Next, the antigen uptake capacity of the generated cells were evaluated and revealed DC-typical behavior in this respect (Supporting Information Fig. 2). Taken together, these data demonstrate that HSC expansion and GM-CSF/IL-4 stimulation led to development of functional myeloid DCs.

Figure 1.

GM-CSF and IL-4 stimulation of expanded HSCs induce generation of functional myeloid dendritic cells. (A): HSC-derived DCs were generated by HSC incubation with 100 ng/ml GM-CSF and 20 ng/ml IL-4. On day 5 poststimulation, the cells were analyzed for their morphology and the expression of stem cell surface markers CD34 and CD117 via flow cytometric analysis (dotted line) compared with isotype control (solid line). (B): reverse transcription polymerase chain reaction (RT-PCR) of indicated marker gene mRNAs isolated from HSCs, monocytes, and dendritic cells derived from either HSCs or monocytes. (C): HSC-derived DCs were incubated with 10 μg/ml LPS or solvent control. At indicated time points, supernatants were subjected to cytokine array analysis. Boxes indicate secretion of MCP-1 (dotted), RANTES (black), and IL-6 (gray). The results shown are representative of three independent experiments. Abbreviations: CRD-BP/IMP1, Coding region determinant-binding protein/insulin-like growth factor II mRNA-binding protein; DC-CK1, Dendritic cell-derived chemokine-1; DC-SIGN, Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin; GM-CSF, Granulocyte macrophage colony-stimulating factor; HSC, hematopoietic stem cell; IL-6, Interleukin-6; LPS, Lipopolysaccharides; mDC, myeloid dendritic cell; MCP-4, Monocyte chemotactic protein-4.

ERK, JAK2/STAT5, and PKCδ Signaling Pathways Are Activated in GM-CSF/IL-4 Induced HSC Differentiation Toward DCs

To determine the role of MAPK signaling pathways involved in HSC differentiation, we focused on ERK, JNK, and p38 that play crucial roles in cell proliferation, stress response, and cell differentiation [32]. On induction of HSC differentiation by GM-CSF/IL-4, an immediate and time-dependent activation of ERK reaching a maximum within 5 minutes was observed by phospho-site-specific immunoblot analysis (Fig. 2A). Interestingly, JNK and p38 have not been activated by GM-CSF and IL-4 within this time period (Fig. 2A). We also analyzed PI-3K pathway activation in HSCs, because this pathway was implicated in myeloid cell differentiation before [33]. However, we found PI-3K not activated via GM-CSF and IL-4 stimulation (data not shown). GM-CSF-induced proliferation and differentiation of myeloid cells were associated with activation of the Jak2 signaling pathway [34]. Therefore, we examined whether GM-CSF/IL-4 might also induce Jak2 activation in HSCs. Jak2 was found activated until 5 and 7 minutes after treatment and dephosphorylation was observed between 15 and 30 minutes poststimulation (Fig. 2B). The activation of the Jak2 pathway was directly reflected in the downstream activation of Signal Transducer and Activator of Transcription 5 (STAT5), which was found 2–3 minutes delayed after Jak2 phosphorylation (Fig. 2B). Interestingly, the activity of the JAK2/STAT5 pathway was prolonged up to five in comparison with the ERK signaling pathway.

Figure 2.

GM-CSF and IL-4 induce activation of PKCδ, ERK1/2, and the JAK/STAT pathway. (A–C): hematopoietic stem cells were stimulated with 100 ng/ml GM-CSF and 20 ng/ml IL-4 for the designated periods. As positive control, cells were incubated with 100 ng/ml Phorbol-12-myristate-13-acetate (TPA), 0.5 μM ionomycin, and 10 μg/ml anisomycin (T/I/A) for 15 minutes. As negative control, unstimulated cells were used (C). Cell lysates were analyzed by immunoblot with antibodies against the indicated phosphorylated proteins and the whole proteins. The results shown are representative of three independent experiments. Abbreviations: ERK, Extracellular-signal-regulated kinase; GM-CSF, Granulocyte-macrophage colony-stimulating factor; IL-4 Interleukin-4; JAK, Janus kinase; JNK, c-Jun NH2-terminal kinase; PKC δ, protein kinase C δ; SAPK, Stress-activated protein kinase; STAT, Signal Transducer and Activator of Transcription.

As PKCδ was described to be the predominant isoform in myeloid cells [13], we focused on GM-CSF/IL-4-dependent activation of PKCδ in HSCs and observed an immediate and long-lasting phosphorylation of PKCδ, which was still apparent 75 minutes poststimulation (Fig. 2C). Overall, these observations indicate that DC-directed HSC differentiation by GM-CSF and IL-4 stimulation is associated with activation of ERK, JAK2/STAT5, and PKCδ.

Activation of the ERK and PKC Signaling Pathways by TPA Induces HSC Differentiation Toward DCs

The PKC activator Phorbol-12-myristate-13-acetate (TPA) was used previously for DC generation derived from the CD34+ cell line KG-1 [35] and primary CD34+ progenitor cells [36]. In comparison with cytokine stimulation, we aimed to analyze the role of the respective signaling pathways in TPA-induced HSC differentiation toward DCs to discriminate them from pathways that are activated by cytokine stimulation but might not be essential for differentiation. Five days after TPA stimulation, the cells showed a typical dendritic morphology (Fig. 3A). The surface marker profile corresponded to previously described characteristics of TPA-induced DCs (see [36] and Supporting Information Fig. 3). To determine which of the cytokine-activated signaling proteins, ERK, STAT5, or PKCδ, are also activated by TPA treatment, we performed immunoblot analysis. Interestingly, a specific activation of ERK and PKCδ was detected, whereas STAT5 phosphorylation was not induced by TPA (Fig. 3B). These data suggest that the differentiation of HSCs to DCs does not require JAK/STAT signaling per se. Nevertheless, the activation of the JAK/STAT pathway seems to be crucial for proliferation and survival of HSCs and progenitors derived thereof, as we observed a significant decrease of colony formation on JAK/STAT inhibition (Supporting Information Fig. 4). These findings show that both cytokine and TPA-induced HSC-differentiation toward myeloid DCs involves the activation of the ERK and PKCδ pathway, supporting their important roles in this process.

Figure 3.

TPA-induced HSC differentiation toward dendritic cells involves ERK1/2 and PKCδ signaling. (A): HSCs were stimulated with 100 ng/ml GM-CSF and 20 ng/ml IL-4 or 20 ng/ml TPA. The cell morphology was analyzed prior and 5 days poststimulation using phase-contrast microscopy. (B): HSCs were incubated with 20 ng/ml TPA for 25 minutes and analyzed for activation for ERK, STAT5, and PKCδ with immunoblot analysis using phospho-specific antibodies. The results shown are representative of five independent experiments. Abbreviations: DCs, dendritic cells; ERK, Extracellular-signal-regulated kinase; GM-CSF, Granulocyte-macrophage colony-stimulating factor; HSC, hematopoietic stem cell; IL-4, Interleukin-4; PKC δ, protein kinase C δ; STAT, Signal Transducer and Activator of Transcription; TPA, Phorbol-12-myristate-13-acetate.

PU.1 Is Phosphorylated by PKCδ

Lately, the role of PU.1 as critical regulator in DC differentiation became apparent (for review see [37]). However, the molecular mode of PU.1 regulation in this process is still unclear. Initially, we confirmed PU.1 expression in HSCs, the CD34+ myeloid leukemia cell line KG-1a, the lymphatic cell line A.301 but not in Molt 4/8 T, HOS, 293T, HeLa, and Alpha 1 cells (Fig. 4A). These cell-specific expression patterns coincide with the observed significance of PU.1 in differentiation processes of hematopoietic cells and differentiation-competent cell lines [38].

Figure 4.

PKCδ but not ERK2 phosphorylates PU.1. (A): Equal protein amounts of indicated cell line lysates were subjected to immunoblot analysis with a PU.1-specific antibody. (B): 3T3 cells were transfected with a plasmid encoding hemagglutinin (HA)-tagged PU.1. HA-PU.1 was isolated with anti-hemagglutinin (HA)-directed immunoprecipitation, eluted with HA peptide and incubated with γATP32 in the presence or absence of 2.5 ng recombinant ERK2 and 30 μM ERK inhibitor. The kinase reaction analyzed performing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) radiography and immunoblotting. (C): The phosphorylation assay was performed as described in (B) but with 20 ng recombinant PKCδ and 1 μM PKC inhibitor Gö6983. The results shown represent two independent experiments. Abbreviations: ERK, Extracellular-signal-regulated kinase; GST, Glutathione-S-transferase; HSC, hematopoietic stem cell; PKC δ, protein kinase C δ.

To determine whether ERK and PKCδ activation during HSC differentiation to DCs might induce phosphorylation of PU.1, we performed in vitro kinase assays. The radiography analysis showed no phospho-PU.1-specific protein band in the presence of ERK (Fig. 4B). The activity of ERK2 was confirmed by phosphorylation of the ERK substrate Elk-1, which was significantly circumvented in the presence of the ERK-inhibitor (Fig. 4B). The structural integrity of HA-PU.1 was confirmed by incubation with casein kinase II, which expectedly led to phosphorylation of HA-PU.1 (see Supporting Information Fig. 5).

Next, we analyzed the potential phosphorylation of PU.1 by PKCδ and used histone H1 as positive control substrate of PKCδ. Interestingly, we found a significant PKCδ-dependent phosphorylation of PU.1, which was suppressed by incubation with the PKC-specific inhibitor Gö6983 (Fig. 4C). These findings suggest that PKCδ-mediated direct phosphorylation of PU.1 promotes DC-directed HSC differentiation. Moreover, PKCδ is the only known isoform to translocate into the nucleus on TPA stimulation [39] and is therefore present in the same intracellular compartment as nuclear PU.1.

PU.1 Activity Is Regulated by an Interplay of the ERK1/2 and PKCδ Signaling Pathways

To get further insights into mutual effects of PKC and ERK signaling pathways in terms of PU.1 activation, we analyzed their potential in regulating PU.1-dependent transcription. A luciferase reporter construct harboring five PU.1 binding sites (Fig. 5A) was cotransfected with a HA-PU.1 expression plasmid into HeLa cells. To activate the ERK and PKC signaling pathways, the transfected cells were stimulated with TPA. The specificity of TPA-induced activation was confirmed by applying the inhibitor U0126 targeting the ERK upstream signaling kinase MEK or the PKC inhibitor Gö6983. Both pharmacological inhibitors were previously verified according to specificity and cell toxicity in a dose-dependent fashion (data not shown).

Figure 5.

PU.1 activity depends on interplay of the ERK1/2 and PKC signaling pathways. (A): Composition of the PU.1-dependent luciferase reporter pPU.1-Luc. (B): HeLa cells were transfected with pPU.1-Luc, a HA-PU.1-expressing plasmid or an empty plasmid. One day later and 16.5 hours prior to lysis, cells were incubated with 10 μM U0126, 1 μM Gö6983, or 30 μM ERK inhibitor. Thirty minutes later, cells were incubated with 20 ng/ml TPA or solvent. After cell lysis, equal protein amounts were analyzed for luciferase expression via luminometry. (C): HeLa cells were transfected with specific siRNAs against endogenous ERK1/2, PKCδ or cotransfected PU.1. Cells were analyzed for protein expression 48 hours post-transfection via Western blot with indicated antibodies. (D): HeLa cells were transfected with pPU.1-Luc, a HA-PU.1-expressing plasmid and 20 nM siRNAs against ERK1/2, PKCδ, or PU.1. Cells were incubated with TPA and analyzed as described in (B). The results shown here represent one of three independent experiments. Abbreviations: ERK, Extracellular-signal-regulated kinase; PKC δ, protein kinase C δ; RLU, relative light unit; TPA, Phorbol-12-myristate-13-acetate.

As expected, TPA induced significant PU.1-dependent transcription in comparison with nonstimulated cells (Fig. 5B). In MEK and ERK inhibitor-treated cells, PU-1 activity was fourfold reduced. Although PU.1 was not found to be a direct ERK phosphorylation substrate, these results indicate that PU.1 activity depends on MAPK signaling in TPA-treated cells. Inhibition of PKC by Gö6983 resulted even in a fivefold reduction of PU.1-dependent transcription (Fig. 5B), supporting the significance of PKC function in terms of TPA-mediated PU.1-activity. These observations suggest that a crosstalk of both the PKC and ERK signaling pathway is important for PU.1 transcriptional activity.

To determine whether PKCδ is the unique PKC isoform contributing to PU.1 activity, we performed siRNA experiments. From the endogenous PKC isoforms α, βII, γ, δ, ε, η, θ, and μ that we found expressed in HeLa cells (see Supporting Information Fig. 6), we selectively silenced PKCδ as well as endogenous ERK (Fig. 5C). HeLa cells transiently expressing PU.1 were cotransfected with the PU.1-dependent luciferase-reporter construct and with siRNAs specific for ERK, PKCδ, or PU.1 as control. Successful knockdown was confirmed by immunoblot analysis (Fig. 5C). Knockdown of ERK led to twofold impaired PU.1 activity in TPA-stimulated cells. Moreover, the knockdown of PKCδ resulted even in a 10-fold reduced PU.1 activity compared with ERK knockdown (Fig. 5D). Therefore, these results support that TPA-induced PU.1 activity strictly depends on the presence of the PKCδ isoform and to a minor degree on the presence of ERK1/2.

PKCδ Acts on PU.1 by Enhancing Its Transactivation Capacity Via Phosphorylation of Its Transactivator Domain

As TPA induces HSC differentiation and activation of PU.1, we analyzed its putative stimulation-dependent mode of regulation. To determine whether PU.1-DNA binding is regulated by stimulation, HSCs were cultivated in the presence or in the absence of TPA or GM-CSF/IL-4, for 2 and 6 hours, respectively. The cell lysates were incubated with radioactive-labeled DNA probes harboring the PU.1 consensus or mutant binding site. We observed a constitutive DNA binding of PU.1, which was not altered in the presence of TPA or GM-CSF/IL-4 stimulation (Fig. 6A). These data were controlled by super-shift experiments using a PU.1-specific antibody (Fig. 6A). Independently, potent TPA or GM-CSF/IL-4 stimulation of HSCs was confirmed by Nuclear factor kappa-B (NFκB)-specific band shift analysis (Supporting Information Fig. 7).

Figure 6.

ERK and PKCδ-promoted HSC differentiation is associated with enhanced transactivation activity of PU.1 via PKCδ-mediated phosphorylation of its TAD. (A): HSCs were incubated with 20 ng/ml TPA or 100 ng/ml GM-CSF/20 ng/ml IL-4 for the indicated period. Protein extracts of HSCs were incubated with a radioactive-labeled DNA probe comprising a wild-type or mutated PU.1-binding motif and analyzed via electrophoretic mobility shift assay. For PU.1 supershift detection, 2 μg anti-PU.1 antibody was added to the reaction prior gel separation. (B): The domain structure of the fusion protein harboring the GAL4-binding domain (top) and the PU.1 transactivation domain as well promoter composition of the luciferase reporter. HeLa cells were transiently transfected with plasmids encoding for GAL4-PU.1-TAD as well as the luciferase reporter (bottom). The cells were treated and analyzed as described in Figure 5B. (C): GST, GST-PU.1-TAD, or HA-PU.1 were incubated with γATP32 in the presence or absence of 20 ng recombinant PKCδ and 1 μM PKC inhibitor Gö6983. The kinase reaction was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), radiography, and immunoblotting. (D): Expanded HSCs were incubated with 20 ng/ml TPA in the presence or absence of 1 μM Gö6983, 5 μM U0126, or solvent. Five days later, fresh medium containing 20 ng/ml TPA was applied. At indicated days, the morphology was determined via phase-contrast microscopy. Each of the shown experiments was performed at least twice showing similar results. Abbreviations: DMSO, Dimethyl Sulfoxide; GM-CSF, Granulocyte-macrophage colony-stimulating factor; GST, Glutathione-S-transferase; HSC, hematopoietic stem cell; IL-4, Interleukin-4; PKCδ, protein kinase C δ; TAD, transactivation domain; TPA, Phorbol-12-myristate-13-acetate.

Besides DNA binding, PU.1 activity may depend on the activation of the TAD. To analyze whether ERK and PKC may regulate the TAD, a fusion protein consisting of the PU.1 TAD and a GAL4-DNA DBD was expressed to control the luciferase expression from a reporter plasmid harboring GAL4 binding sequences located upstream of a minimal thymidine kinase TK promoter (Fig. 6B). TPA stimulation of transfected U937 monocytes (data not shown) or A.301 T cells led to an approximate sixfold increase of PU.1 transactivation (Fig. 6B). Inhibiting the ERK pathway with the MEK inhibitor U0216 resulted in approximately 50% reduction of PU.1-dependent transcriptional activation. Most strikingly, the inhibition of PKC reduced transactivation up to 66% (Fig. 6B), indicating that PKC and ERK induce PU.1 activity by enhancing its transactivation capacity but not PU.1 DNA binding. As the TAD of PU.1 was sufficient to mediate PKC-induced activity, we analyzed whether the PU.1 TAD might be a direct phosphorylation target of PKCδ. In vitro kinase assays with GST-PU.1-TAD and full-length HA-PU.1 showed that both were phosphorylated by recombinant PKCδ, but not in presence of the PKC inhibitor Gö6983 (Fig. 6C), suggesting that PU.1 activation during HSC differentiation might be regulated by PKCδ-dependent phosphorylation of its TAD.

TPA-Induced Differentiation of HSCs Is Reversibly Arrested by Specific ERK1/2 or PKC Inhibitors

To test the concept of ERK- and PKCδ-mediated PU.1 regulation in TPA-induced HSC differentiation toward DCs, we aimed to knockdown PKCδ and ERK in HSCs but were not successful most probably due to very low transfection efficiencies in these primary cells. Alternatively, HSC differentiation was stimulated either in absence or presence of the ERK inhibitor U0126 or the PKC-specific inhibitor Gö6983. At day 5 poststimulation, HSC differentiation was found restrained in the presence of the ERK inhibitor U0126. We observed the formation of local proliferation clusters with small and blast cell morphology similar to undifferentiated, nonstimulated HSCs (Fig. 6D). In contrast, stimulated cells in the absence of inhibitors developed a typical dendritic morphology, as shown before. Intriguingly, application of the PKC-specific inhibitor Gö6983 resulted in even more apparent inhibition of DC formation (Fig. 6D). To control whether these cells still retain their differentiation potential, the inhibitors were removed and the cells were restimulated with TPA. On day 10, the cells exhibited DC morphology, indicating a restoration of HSC differentiation capacity after inhibitor withdrawal (Fig. 6D). These findings confirm that differentiation of HSC toward mDCs is mediated by activation of both the PKC and the ERK signaling pathway, furthermore, suggesting that the inhibition of the PKC pathway results in a differentiation block, which cannot be compensated by the ERK1/2 signaling pathway and vice versa.


In this study, we have investigated the role of PKC and MAPK signaling including downstream factors in CD34+ hematopoietic progenitor function, focusing on DC development. We used a human ex vivo cell culture system to expand undifferentiated HSCs [25, 26] derived from umbilical cord blood and verified the phenotypical and physiological properties of HSC-derived myeloid DCs [1, 27].

For the first time, we describe that initiation of HSC differentiation with GM-CSF/IL-4 induces rapid activation of the MAPK/ERK, JAK/STAT, and PKCδ signaling pathway, whereas the JNK and p38 MAPK signaling pathways were not activated by these cytokines. Surprisingly, all three pathways were inactivated very fast in comparison with activation kinetics of erythroid and myeloid cell lines [40].

Besides GM-CSF/IL-4, we confirm that the PKC activator TPA also drives HSCs toward DC differentiation [41], resulting in a more rapid DC formation but with a significantly lower number of effector cells. Interestingly, JAK/STAT signaling was not involved in TPA-mediated differentiation compared with GM-CSF/IL-4 triggered differentiation. We found JAK signaling to be indispensable for broad HSC differentiation capacity per se by influencing proliferative properties of the cells, which is in line with earlier findings that JAK/STAT signaling promotes proliferation and self-renewal of HSCs [42]. However, in terms of HSC-derived DC formation, our data indicate that the activation of the MAPK/ERK and PKCδ signaling pathway, but not JAK/STAT signaling pathway activation, is essential and sufficient. Furthermore, we show that the ERK and PKCδ differentiation-promoting capacity can be retained by reversible and nontoxic inhibition of these signaling targets. Unfortunately, there are no PKCδ-specific inhibitors available. Using Gö6983 in our experimental setups instead that, besides PKCδ, also inhibits PKC isoforms α, β, γ, and ζ might raise the concern that other PKC isoforms also contribute to TPA-induced PU.1 activation. However, our findings that knockdown of endogenous PKCδ in HeLa cells, that also express PKCα, βII, γ, ε, η, θ, and μ/PKD, abrogated PU.1 activation, clearly supports the unique role of PKCδ. In addition, from the PKC isoforms α, β, δ, γ, and ε, only PKCδ was reported to translocate into the nucleus on activation [39], corresponding to our data of PKCδ affecting the function of nuclear protein PU.1.

On molecular level, we show that the transcription factor PU.1, whose activation was reported to induce DC differentiation [43], is phosphorylated by PKCδ but not by ERK2. Analyzing PKCδ- and ERK-dependent PU.1 activity using specific siRNAs and inhibitors, we conclude that PU.1 activity depends on the presence of both kinases. As PU.1 is not phosphorylated by ERK, we suggest that PU.1 activity might be mediated by ERK-regulated proteins that are known to interact with PU.1 in a modulating manner such as GATA binding protein 1 (GATA-1), CREB binding protein (CBP), transducin-like enhancer of split 4 (TLE-4), and TATA box binding protein (TBP) [44–47]. PU.1 is a bifunctional protein as for instance CBP binding leads to PU.1 transcriptional activation [47], whereas TLE-4 was found to support PU.1-dependent repression [46]. However, the presence of a putative ERK-dependent PU.1-activating cofactor has to be determined yet. Besides ERK, our findings underline the prominent role of PKCδ in DC generation and strengthen the importance of PKC isoforms in hematopoietic signaling as reported by other groups [10]. Aiming to understand the distinct mode of PKCδ-mediated PU.1 regulation, we show that differentiation-inducing cytokine or TPA stimuli do not alter PU.1-DNA-binding activity in HSCs. Therefore, the stimuli-induced phosphorylation of PU.1 seems to have no impact on its DNA-binding capacity. Instead, our results demonstrate that the PU.1 transactivation capacity is positively regulated in a PKCδ- and ERK-dependent manner.

For instance, the molecular mechanism of the PU.1 repressor GATA-1 is controversially discussed. One group described GATA-1 to interfere in PU.1 transactivation and having no impact on PU.1 DNA binding [48], whereas another group concluded the contrary [49]. Notably, besides transcription of genes that is directly dependent on PU.1-mediated transactivation, PU.1 is also able to form complexes with cofactors as IRF-4 and IRF-8 to induce cofactor-dependent expression, for example, of Toll-like receptor 4 [50–52]. However, the molecular mechanism facilitating this complex-dependent transcription is not known. According to PU.1-dependent transcription in HSCs, we show that phosphorylation of the PU.1 TAD by PKCδ is associated with its enhanced transactivation activity thereby promoting DC generation.

Overall, these results provide new insights in the molecular mechanisms of HSC differentiation toward mDCs and underline the outstanding role of distinct PKC isoforms in hematopoietic processes. In principle, this mode of regulation may be conferrable to differentiation processes of other adult stem cells. Finally, the distinct intracellular mechanisms that drive the differentiation of adult stem cells such as HSCs are mostly unknown and their elucidation will be an important subject in respect of controlled cell/tissue development for stem cell-based therapies.


We acknowledge Renate Wenig and Erika Krebil for technical support as well as Heide Muckenfuss and Ralf Sanzenbacher for introduction in the EMSA technology and helpful discussions. We especially thank Carol Stocking, Heinrich-Pette-Institut, for excellent advice in hematopoietic stem cell cultivation as well as Denis Corbeil and Doreen Strauss, Biotechnology Center, TU Dresden. This work was supported by the German BMBF grant 0312113 (E.F.).


The authors indicate no potential conflicts of interest.