Although it has been proposed that the common myeloid progenitor gives rise to granulocyte/monocyte progenitors and megakaryocyte/erythroid progenitors (MEP), little is known about molecular switches that determine whether MEPs develop into either erythrocytes or megakaryocytes. We used the thrombopoietin receptor c-Mpl, as well as the megakaryocytic marker CD41, to optimize progenitor sorting procedures to further subfractionate the MEP (CD34+CD110+CD45RA−) into erythroid progenitors (CD34+CD110+CD45RA−CD41−) and megakaryocytic progenitors (CD34+CD110+CD45RA−CD41+) from peripheral blood. We have identified signal transducer and activator of transcription 5 (STAT5) as a critical denominator that determined lineage commitment between erythroid and megakaryocytic cell fates. Depletion of STAT5 from CD34+ cells by a lentiviral RNAi approach in the presence of thrombopoietin and stem cell factor resulted in an increase in megakaryocytic progenitors (CFU-Mk), whereas erythroid progenitors (BFU-E) were decreased. Furthermore, an increase in cells expressing megakaryocytic markers CD41 and CD42b was observed in STAT5 RNAi cells, as was an increase in the percentage of polyploid cells. Reversely, overexpression of activated STAT5A(1*6) mutants severely impaired megakaryocyte development and induced a robust erythroid differentiation. Microarray and quantitative reverse transcription-polymerase chain reaction analysis revealed changes in expression of a number of genes, including GATA1, which was downmodulated by STAT5 RNAi and upregulated by activated STAT5.
Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: S.G.O. and M.R.T.: performed research; S.F.: contributed to progenitor sorting procedures; A.L.D.: designed research, interpreted data; E.V. and J.J.S.: designed research, interpreted data, wrote the paper.
In the human system, megakaryocytes and erythrocytes arise from a common megakaryocyte/erythroid progenitor (MEP) that is derived from a common myeloid progenitor (CMP)  or possibly emerge via a more direct pathway from hematopoietic stem cells (HSCs) . The induction of MEP to the erythroid or megakaryocytic lineage is regulated by extrinsic factors, such as erythropoietin (EPO) and thrombopoietin (TPO), and by an intrinsic program of transcription factors [3, 4]. Two major transcription factors have been identified that are involved in CMP differentiation, namely, GATA-1, which drives differentiation of MEP, and PU.1, which regulates granulocyte-monocyte precursors (reviewed in ). The downregulation of PU.1 expression in the CMP is the first event associated with the restriction of differentiation to erythroid and megakaryocytic (MK) lineages . Instruction of the MEP toward the erythroid or MK lineage is dependent on the gene dosage of GATA-1 . Loss of GATA-1 leads to differentiation arrest and apoptosis of erythroid progenitors and accumulation of immature megakaryocytes [6, –8]. Similarly, low expression of GATA-1 in mice favors the megakaryocytic development , whereas forced GATA-1 expression reprogrammed the common lymphoid and myeloid progenitors to the MK/erythroid lineage . These instructive programs of GATA-1 are executed in association with a number of additional cofactors that reside in close proximity to GATA sequences in MK-specific promoters, such as the Ets family factors, or act independently, such as the proto-oncogene c-myb [3, 4].
The identification of JAK-2 mutation in patients with polycythemia vera (PV) and essential thrombocythemia (ET) [11, , , –15] has focused attention on the involvement of signal transducer and activator of transcription 5 (STAT5) in erythroid and megakaryocytic development . Because of the constitutive activation of JAK2, a number of downstream targets are persistently activated, including STAT5 and ERK . So far, STAT5 has been identified as a relevant transcription factor for the erythroid and stem cell compartment [17, –19]. Forced overexpression of STAT5 in CD34+ cord blood (CB) cells results in an expansion of the erythroid lineage, inhibits myeloid differentiation in part because of downregulation of C/EBP-α, and results in a two- to threefold expansion of the stem cell compartment [20, 21]. Knockdown studies using STAT5 RNAi constructs in CB CD34+ cells have shown that a reduced STAT5 expression impairs the number of stem and progenitor cells, as determined by long-term culture-initiating cell and colony-forming cell (CFC) assays, without affecting the differentiation program .
A number of recent studies have demonstrated that the commitment to the erythroid or MK lineage is dependent on the gene dose of relevant transcription factors. This not only applies to GATA-1 but also has been demonstrated for the Ets family of transcription factors and Runx1. Therefore, in the present study, we tested whether the gene dosage of STAT5 is relevant for MK or erythroid development by performing STAT5 RNAi and STAT5A(1*6) overexpression studies in CD34+ cells that had been cultured with TPO and stem cell factor (SCF). The results demonstrate that downregulation of STAT5 impairs erythroid differentiation and promotes MK development.
Materials and Methods
Cell Culture and Purification of CD34+ Cells
Mo7e cells were routinely propagated in RPMI 1640 (BioWhittaker, Lonza, Verviers, Belgium, http://www.lonza.com) supplemented with heat-inactivated fetal calf serum (FCS) (5%, vol/vol; Sigma-Aldrich, Zwijndrecht, The Netherlands, http://www.sigmaaldrich.com) and interleukin (IL)-3 (10 ng/ml). CD34-positive cells were obtained from healthy donors undergoing granulocyte colony-stimulating factor (G-CSF) treatment following institutional guidelines. CD34+ cells were isolated by EasySep immunomagnetic cell selection procedures (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com) as described by the manufacturer. To generate MK cells, CD34+ cells were grown in HPGM (Cambrex, Walkersville, MD, http://www.cambrex.com) supplemented with 40 ng/ml TPO (a kind gift from Kirin Brewery Co. [Tokyo, http://www.kirin.co.jp/english]) and 40 ng/ml SCF (Immunex, Seattle, http://immunex.com). To induce erythroid differentiation, CD34+ cells were grown in HPGM medium supplemented with 1 U/ml EPO and 40 ng/ml SCF. Cells were counted by trypan blue exclusion using a hemocytometer twice weekly, and fresh medium plus cytokines was added.
Flow Cytometry Analysis and Sorting Procedures
Sorting of the CD34+ cells into progenitor fractions was performed on the basis of the combinatorial expression of cell surface antigens, as previously reported [22, 23]: CMP as CD34+CD110−CD45RA−, granulocyte/monocyte progenitors (GMP) as CD34+CD110−CD45RA+, and MEP as CD34+CD110+CD45RA−. In addition, the MEP was separated in two fractions on the basis of the CD41+ expression. Erythroid progenitors consisted of CD34+CD110+CD45RA−CD41− cells, whereas megakaryocyte progenitors consisted of CD34+CD110+CD45RA−CD41+ cells. The fluorescence-activated cell sorting analyses were performed on a FACSCalibur (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com), and sorting of the cells was performed on a MoFlo (DakoCytomation, Carpinteria, CA, http://www.dakocytomation.com). Antibodies were obtained from Becton Dickinson, except phycoerythrin-labeled anti-CD41a and anti-CD42b, which were purchased from CLB (Amsterdam, The Netherlands, http://www.sanquin.nl). The CD61-positive cell fraction was purified from primary cultures after 7 days by MoFlo sorting as indicated. Data were analyzed using WinList 3D (Verity Software House, Topsham, ME, http://www.vsh.com) and FlowJo (Tree Star, Ashland, OR, http://www.treestar.com) software.
CFC Progenitor Assays
CFC assays were performed in 1.2% methylcellulose containing 30% FCS, 57.2 μM β-mercaptoethanol, and 2 mM glutamine, supplemented with 20 ng/ml IL-3, 20 ng/ml IL-6, 20 ng/ml G-CSF, 20 ng/ml c-Kit ligand, and 1 U/ml EPO (Eprex-Cilag, Brussels, Belgium, http://www.cilag.ch/indexE.htm). CFU-GM and BFU-E assays were performed in 1.2% methylcellulose containing 30% FCS, 57.2 μM β-mercaptoethanol, and 2 mM glutamine, supplemented with 10 ng/ml granulocyte-macrophage colony-stimulating factor and 10 ng/ml IL-3 or 2 U/ml EPO, respectively.
A commercially available kit for evaluation of colony-forming units-megakaryocytic progenitors (CFU-Mk) was used according to the manufacturer's instructions (MegaCult-C medium with cytokines; StemCell Technologies). In short, 2,000 purified CD34-positive cells were seeded per double chamber culture slide in serum-free medium containing thrombopoietin (50 ng/ml), IL-3 (10 ng/ml), IL-6 (10 ng/ml), and collagen (1.1 mg/ml). Cultures were incubated for 10–12 days, followed by dehydration and immunocytochemical staining of the slides. Megakaryocyte colonies were detected using the CD41 and alkaline phosphatase detection system and counterstained in Evan's Blue. Cultures were scored for the presence of pure MK colonies consisting of at least five nucleated cells expressing CD41, as previously described . For polyploidization analysis, cells were analyzed on cytospins, and multinucleated cells were counted.
293T human embryonic kidney cells (2.5 × 106 cells) were transfected with 3 μg of pCMV Δ8.91, 0.7 μg of VSV-G, and 3 μg of pTRIP Renilla RNAi or pTRIP STAT5 RNAi (kind gifts of Dr. H. Spits, Department of Cell Biology and Histology, Amsterdam Medical Center, Division of Immunology, Netherlands Cancer Institute, Amsterdam, The Netherlands). Target sequences have been described by Scheeren et al. . Twenty-four hours after transient transfection, medium was changed to HPGM, and after 12 hours, supernatant containing lentiviral particles was harvested and stored at −80°C. Mo7e cells were cultured in RPMI 1640 supplemented with 10% FCS and 10 ng/ml IL-3 for 4 hours at 37°C and 5% CO2, and CD34+ cells were cultured in HPGM supplemented with c-Kit ligand, Flt-3 ligand (both from Amgen, Thousand Oaks, CA, http://www.amgen.com), and TPO (100 ng/ml each) for 16 hours at 37°C and 5% CO2 prior to transductions. Transductions of CD34+ cells were performed in two consecutive rounds of 8–12 hours with lentiviral supernatant supplemented with c-Kit ligand/Flt-3 ligand/TPO (100 ng/ml each) and polybrene (4 μg/ml). Mo7e cells were transduced in one round of 12 hours with lentiviral supernatant supplemented with 10% FCS, 10 ng/ml IL-3, and 4 μg/ml polybrene. Transduction efficiency was measured by fluorescence-activated cell sorting (FACS) analysis, and knockdown was investigated by means of Western blot. Alternatively, knockdown was investigated by means of quantitative reverse transcription (RT)-polymerase chain reaction (PCR). Retroviral transductions with constitutively activated STAT5A(1*6) were performed as described previously .
Preparation of Cell Lysates and Western Blotting
Total cell lysates were performed by centrifugation of equal amounts of cells followed by lysing cell pellets in Laemmli sample buffer. SDS-polyacrylamide gel electrophoresis and immunoblotting were performed according to standard procedures. Antibodies used for Western blotting were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com; antibodies against STAT5, ERK, and β-Actin). Antibodies against tyrosine-phosphorylated STAT5 (pY694, C11C5) were obtained from Cell Signaling Technology (Leiden, The Netherlands, http://www.cellsignal.com). Detection was performed according to the manufacturer's guidelines (ECL; Amersham Biosciences, Little Chalfont, U.K., http://www.amersham.com).
Quantitative PCR and Illumina Microarray Analysis
Target gene expression was investigated by means of quantitative reverse transcription-polymerase chain reaction (Q-PCR). Total RNA was isolated from 1 × 105 to 1 × 106 cells using the RNeasy kit from Qiagen (Venlo, The Netherlands, http://www1.qiagen.com) according to the manufacturer's recommendations. RNA was reverse transcribed with M-MuLV reverse transcriptase (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com). For real-time RT-PCR, 2-μl aliquots of cDNA were real-time amplified using iQ SYBR Green Supermix (Bio-Rad, Veenendaal, The Netherlands, http://www.bio-rad.com) on a MyIQ thermocycler (Bio-Rad) and quantified using MyIQ software. As a negative control, RNA minus reverse transcriptase prepared cDNA was used in PCRs. To verify the correct sizes of PCR fragments, 10-μl aliquots were run on 1.5% agarose gels. β-Actin expression was used to normalize expression of all genes investigated. Genome-wide expression analysis was performed on Sentrix Human-6 BeadChip Arrays (Illumina, Inc., San Diego, http://www.illumina.com) (46,000 probe sets) on an outsource basis by ServiceXS (Leiden, The Netherlands, http://www.servicexs.com). mRNA (0.5–1 mg) was used in labeling reactions and hybridization with the arrays according to the manufacturer's instructions. Data were analyzed using the BeadStudio v3 Gene Expression Module (Illumina).
Data were expressed as mean ± SEM. Differences between samples were calculated using Student's t test. A two-sided p value of <.05 was considered statistically significant.
STAT5 Downregulation Promotes Megakaryocytic Expansion from CD34+ Cells
To determine the efficacy of STAT5 downmodulation by our lentiviral RNAi approach, the Mo7e cell line was transduced using Renilla RNAi control vectors or STAT5 RNAi vectors that target both STAT5A and STAT5B [19, 25]. Cell lysates were prepared for Western blot analysis 2 days after the transduction procedure, and as depicted in Figure 1A, 70%–80% downregulation of STAT5 protein could be obtained. As loading control, blots were stripped and reprobed against ERK1 and ERK2, and no differences in expression were observed (Fig. 1A). Similarly, normal CD34+ cells isolated from peripheral blood were transduced with the STAT5 or Renilla RNAi constructs, sorted, and cultured with TPO and SCF (Fig. 1B). Q-PCR analysis of STAT5 mRNA after 10 days of culture demonstrated a 60%–70% reduction in STAT5 expression (Fig. 1B). To study the cellular consequences of the reduced STAT5 expression, transduced CD34+ cells (n = 3) were cultured in CFU-MK, CFU-GM, and erythroid progenitors (BFU-E) assays. Downregulation of STAT5 resulted in a 1.9-fold increase in the number of megakaryocytic progenitors (CFU-MKs) compared with CD34+ cells transduced with the control vector (Fig. 1C; p = .002). No change was observed in the number of CFU-GM, whereas a 0.6-fold reduction in BFU-E colony formation was shown (Fig. 1C; p < .01). Furthermore, megakaryocytic differentiation was monitored in serum-free liquid culture assays driven by SCF and TPO. Under these conditions, a fivefold expansion was reached within 2 weeks (Fig. 1D), and 20%–50% of the cells expressed the megakaryocyte markers CD41 and CD42b (Fig. 1E). Although the expansion was unaffected by STAT5 downmodulation (Fig. 1D), an increase in cells expressing CD41 (42% ± 12% vs. 52% ± 15%; p = .002) and CD42b (19% ± 10% vs. 30% ± 15%; p = .002) was observed under these conditions (numbers indicate the average of the individual data sets shown in Fig. 1E). In addition, STAT5 downregulation resulted in an approximately threefold increase of polyploid cells (Fig. 1F, 1G). Together, these data indicate that downmodulation of STAT5 impairs the development of erythroid progenitors and promotes megakaryocytic differentiation.
Constitutive Active STAT5A(1*6) Induces Erythroid Differentiation at the Expense of Megakaryocyte Development
As the studies described above indicated that reduction of STAT5 is sufficient to promote megakaryopoiesis and impair erythropoiesis, we wished to study the reverse process by elevating STAT5 activity. CD34+ cells were transduced with constitutively active STAT5A(1*6), and MoFlo-sorted green fluorescent protein-positive cells were cultured in suspension with TPO and SCF. As depicted in Figure 2A, STAT5A(1*6) overexpression induced a strong proliferative advantage over MiGR1 control cells. Whereas megakaryocytic differentiation was induced in control cells, a strong reduction in CD41- and CD42b-positive cells was observed in the STAT5A(1*6)-transduced group (Fig. 2B). Instead, erythroid differentiation was induced by active STAT5, as demonstrated by the strong increase in CD36 and GPA-positive cells (Fig. 2B). Representative FACS profiles are shown in Figure 2C, and the induction of erythropoiesis was confirmed by morphological analysis of week 2 cells (Fig. 2D). Thus, these data indicate that constitutive active STAT5A(1*6) induces erythroid differentiation at the expense of megakaryocyte development.
Involvement of Endogenous STAT5 in TPO-Induced Megakaryopoiesis and EPO-Induced Erythropoiesis
Megakaryopoiesis was induced by stimulating CD34+ cells with TPO and SCF. Differentiation was monitored by FACS analysis, and as shown in Figure 3A, a strong increase in CD41+ cells was observed, whereas no increase in GPA+ cells was detected. STAT5 expression levels were determined throughout the differentiation period, and no changes in expression levels were observed (Fig. 3B). Some low levels of STAT5 tyrosine phosphorylation were observed in the initial phase of the experiment that could no longer be observed after 7 days of differentiation (Fig. 3B). In response to EPO, GPA+ cells were efficiently generated, and the percentage of CD41+ cells was reduced (Fig. 3C). Furthermore, a strong upregulation of STAT5 expression was observed in response to EPO, which coincided with a strong upregulation of STAT5 tyrosine phosphorylation (Fig. 3D). These data indicate that the endogenous STAT5 expression and activation patterns coincide with erythroid commitment in response to EPO.
Optimalization of Megakaryocyte Progenitor (CFU-MK) Sorting Protocols
To establish protocols to isolate megakaryocyte progenitors, CD34+ cells were first isolated from the peripheral blood of healthy donors, after which CMP, GMP, and MEP were sorted on the basis of CD45RA and CD110 expression, as shown in Figure 4A. Progenitor assays were performed with sorted cells, and as shown in Figure 4B and 4C, these sorting procedures yielded a high purity of myeloid and erythroid populations. The CD34+/CD110+/CD45RA− MEP population was further sorted into CD41high and CD41low populations (Fig. 4A). As shown in Figure 4B and 4C, the CD34+/CD110+/CD45RA−/CD41+ population contained the majority of CFU-MK progenitors (73%), whereas the majority of the BFU-E was retained in the CD34+/CD110+/CD45RA−/CD41− population (85%). Thus, we propose that the MEP compartment can be further subdivided into erythroid and megakaryocytic progenitors on the basis of CD41 expression (schematically depicted in Fig. 4D). The progenitor frequencies that we observed ranged from 5% to 10%, approximately, indicating that within the defined FACS gates, we also isolated cells that were most likely differentiated beyond the progenitor stage.
To further study the specificity of our sorting procedures, Mk progenitors were sorted and CFU-MK assays were initiated with TPO in the absence or presence of EPO. Megakaryocytic progenitors were generated by TPO, and the number of CFU-MKs was not affected by addition of EPO (Fig. 4E). Sorted Mk progenitors were also differentiated in liquid culture conditions, and upon stimulation with TPO and SCF, multinucleated megakaryocytes were readily generated (Fig. 4F). In contrast, no megakaryocytes could be generated from sorted Mk progenitors in response to EPO and SCF, and cells retained an immature megakaryocytic-erythroid progenitor morphology (Fig. 4F). Reversely, erythroid progenitors were sorted and CFC assays were performed in the absence or presence of EPO, and these experiments revealed that BFU-E were generated in an EPO-dependent manner (Fig. 4G). In liquid culture conditions, erythroid progenitors could be differentiated toward an erythroid fate in response to EPO, whereas an immature progenitor phenotype was retained in the presence of TPO (Fig. 4H). Thus, these data indicate that the CD34+/CD110+/CD45RA−/CD41+ Mk population responds to TPO and can be differentiated toward mature megakaryocytes, whereas the CD34+/CD110+/CD45RA−/CD41− erythroid progenitor (Ery) population predominantly responds to EPO.
STAT5 Downregulation in CMP, GMP, MEP, MK, and Erythroid Progenitors
Since the effects of STAT5 downregulation were observed in the balance between the megakaryocytic and erythroid lineage, we questioned whether we could further identify these modulating effects by isolation of megakaryocytic and erythroid progenitors. Therefore, CD34+ cells were transduced with STAT5 RNAi or Renilla RNAi control vectors and then further sorted in CMP (CD110−/CD45RA−), GMP (CD110−/CD45RA+), MEP (CD110+/CD45RA−), MK (CD110+/CD45RA−/CD41+), and erythroid (CD110+/CD45RA−/CD41−) progenitors. Subsequently, the effects of STAT5 downregulation were studied on the progenitor frequency and composition of the different subsets. As presented in Figure 5A and 5B in two representative transductions with CD34+ cells, a clear reduction in BFU-E and an increase in CFU-MKs was observed upon downmodulation of STAT5, whereas the number of myeloid progenitors was much less affected.
Microarray Analysis on Transduced PB CD34+ Cells
To correlate the effects of STAT5 depletion on megakaryoctye/erythroid lineage fate decisions with changes in gene expression, we performed microarray analysis on STAT5- and Renilla RNAi-transduced CD34+ cells that were expanded under serum-free conditions with SCF and TPO. Upon downmodulation of STAT5 expression, 53 genes were downregulated and 33 genes were upregulated (Table 1; fold change, >2; p < .05). The list of downmodulated genes contained a large number of genes involved in regulating erythropoiesis, including GATA1; hemoglobin ε1, γG, γA, and β1; CD36; and glycophorin B. Furthermore, known STAT5 target genes, such as Pim2, were found to be downregulated in STAT5 RNAi cells. Q-PCR analysis confirmed that Gata1 expression is reduced upon downmodulation of STAT5, and the same was true for the STAT5 target gene Bcl-XL and the cell cycle inhibitor p21 (Fig. 6A). Importantly, many of these genes were found to be significantly upregulated in cells expressing constitutively activated STAT5A(1*6), including GATA1 and p21 (Fig. 6B). Moreover, in line with the erythroid commitment induced by STAT5, we observed a strong upregulation of hemoglobins (Fig. 6B). The microarray analysis further revealed reduced expression of CXCR4 and the tetraspanin family members CDC151 and TSPAN4 upon downmodulation of STAT5, and inhibition of these membrane proteins has been associated with enhanced megakaryocyte development.
Table Table 1.. Downregulated and upregulated genes in STAT5 RNAi-transduced PB CD34+ cells
The present study demonstrates that the STAT5 expression level is a critical factor that determines erythroid versus megakaryocytic lineage fate. Reduction in STAT5 expression promoted lineage commitment to the megakaryocytic lineage at the expense of the erythroid lineage. Reversely, overexpression of activated STAT5 enhanced expansion of the erythroid compartment, whereas the generation of CD41+/CD42b+ megakaryocytic cells was impaired.
Our previous studies with STAT5 knockdown in human HSCs have recently shown that STAT5 activation is a critical factor at different levels in the hematopoietic system . Reduced STAT5 expression in CB CD34+ cells causes reduced numbers of stem cells and progenitors, which was partly attributed to changes in the cell cycle status . The present study demonstrates that STAT5 is also a critical factor for the lineage commitment to the erythroid and MK lineage. With CD34+ cells, as well as progenitor cells sorted on the basis of CD110, CD45RA, and CD41 expression, it was shown that downregulation of STAT5 changes the balance between the erythroid and megakaryocytic commitment in favor of the MK lineage. A higher number of CFU-MK was noticed in conjunction with an enhanced differentiation along the megakaryocytic lineage. Reversely, expression of activated STAT5 resulted in an induction of erythroid commitment, in line with our previous observations [20, 21]. Not only was the strong induction of erythropoiesis observed by overexpression of the STAT5A(1*6) mutant, but similar results have been obtained using the single mutant STAT5A (S711F) or by using fusion proteins of wild-type STAT5A and STAT5B with the ligand-binding domain of the estrogen receptor, whereby STAT5 activity is induced by administration of 4-hydroxytamoxifen (A.T.J. Wierenga et al., manuscript submitted for publication).
Recently, Buet et al. demonstrated that megakaryocytic progenitors can also be reprogrammed into erythroid progenitors by overexpression of p210Bcr−Abl . The number of CD41- and CD42-positive cells was strongly reduced upon expression of p210Bcr−Abl, whereas the number of cells expressing the erythroid marker GPA was strongly enhanced, in line with our observations using overexpression of activated STAT5. Thus, as STAT5 is one of the main signaling components downstream of p210Bcr−Abl, it is conceivable that the erythroid lineage conversion at the expense of megakaryocyte development involves p210Bcr−Abl-induced STAT5 activity as well. Fli1 was identified as one of the main targets that was downmodulated by p210Bcr−Abl . Whether Fli1 downmodulation is also a prerequisite for STAT5-induced erythropoiesis remains to be elucidated, but our preliminary data do indicate that Fli1 expression is reduced upon STAT5 activation in human cord blood CD34+ cells (unpublished observations).
A number of genes have been identified that have a crucial function in the switch of MEP to the erythroid or MK lineage. GATA1 is indispensable for the differentiation of MEPs toward terminally differentiated erythrocytes, as maturation is arrested at an early proerythroblast-like stage in GATA1-deficient embryos, and GATA1-deficient mice die between embryonic days 10.5 and 11.5 of gestation because of severe anemia . Various studies have shown that reduced expression of GATA-1 promotes the megakaryocytic development. In a megakaryocyte lineage-specific knockout model, it has been shown that depletion of GATA1 in primary megakaryocytes results in hyperproliferation and expansion of the megakaryocytic compartment, although maturation toward mature platelets was impaired, finally resulting in myelofibrosis [6, 27]. Our microarray analysis and Q-PCR data indicated that GATA1 expression is reduced upon depletion of STAT5, whereas introduction of activated STAT5 in human CD34+ cells resulted in an increase in GATA1 expression. These data suggest that the STAT5-induced change in lineage fate determination toward erythroid development might, at least in part, be mediated by changes in GATA1 expression. Megakaryopoiesis induced by downmodulation of STAT5 coincided with reduced levels of Bcl-Xl mRNA, which might be in line with previous data, indicating that megakaryopoiesis is impaired or delayed by overexpression of Bxl-Xl [28, 29]. STAT5-induced erythroid commitment was further demonstrated by upregulation of erythroid-specific genes, such as glycophorin A, CD36, hemoglobin α, and hemoglobin γ, in line with our previously published observations . Upon downmodulation of STAT5, many of these erythroid-specific genes were reversibly downregulated as well (including CD36; hemoglobins α, β, ε, and γ; and glycophorin), further strengthening the view that STAT5 is one of the key denominators that determine lineage commitment from the MEP, and it will be challenging to determine in future studies whether these STAT5-mediated changes in erythroid-specific gene expression are all mediated via GATA1.
Our modified progenitor sorting protocol allows further elucidation of molecular mechanisms involved in lineage fate decisions from the MEP into megakaryocytic or erythroid progeny. Efficient separation of myeloid and megakaryocyte/erythrocyte progenitors on the basis of CD45RA and TpoR expression was recently demonstrated by Edvardsson et al.  as a modification to the original protocol proposed by Manz et al. , in which CD45RA and CD123 were used. We now included CD41 in our sorting procedures as well and could efficiently isolate megakaryocytic (CD34+/CD110+/CD45RA−/CD41+) and erythroid (CD34+/CD110+/CD45RA−/CD41−) progenitors. Whether our sorting procedures using CD110 and CD41 are also valid for bone marrow cells remains to be elucidated, as one previous report suggested that expression of these markers does not indicate commitment to the megakaryocyte lineage in bone marrow cells , in contrast to our findings. Another study in mice indicated that CD9 might also be a useful marker for isolating megakaryocytic progenitors . These cells also coexpressed CD41, and it will be of interest to determine whether CD9 will aid in further purification of megakaryocyte progenitors from human populations as well. Our RNAi studies indicate that depletion of STAT5 from CMP, Ery, and MK populations results in an increase in megakaryocyte progenitors and a decrease in erythroid progenitors, whereas no changes were observed in the GM compartment.
So far, constitutive STAT5 activation has especially been linked to the expansion of the erythroid and MK compartment. V617F JAK2 mutations have recently been identified as an important genetic marker for PV and ET [11, , , –15]. In 80%–90% of PV patients, the JAK2 mutation can be demonstrated, compared with approximately 50% of patients with ET and idiopathic myelofibrosis (IM). A study by Teofili et al. demonstrated recently that the concomitant expression of phosphorylated STAT5 (pSTAT5) by the affected bone marrow cells differs significantly between the separate disorders . Uniformly increased pSTAT5 was demonstrated in PV, whereas reduced pSTAT5 was shown in ET and in patients with IM. This variability in pSTAT5 expression could not be related to the predominance of the myeloid or erythroid lineage. Apparently, the downstream targets of the JAK2 mutation are cell type-dependently affected and modulated. What is remarkable is the finding that in both disorders in which the megakaryocytic lineage is predominantly affected, a reduced expression of pSTAT5 was observed. These findings are in line with the results of the present study, which demonstrates a critical role for STAT5 expression levels in the commitment toward erythroid and megakaryocytic lineages.
Our data presented here indicate that we can utilize the thrombopoietin receptor c-Mpl in combination with the megakaryocytic marker CD41 to optimize progenitor sorting procedures to further subfractionate the MEP (CD34+CD110+CD45RA−) into erythroid (Ery: CD34+CD110+CD45RA−CD41−) and megakaryocytic (Mk: CD34+CD110+CD45RA−CD41+) progenitors from peripheral blood (PB). Furthermore, we have identified STAT5 as a critical denominator that determined lineage commitment between erythroid and megakaryocytic cell fates. Depletion of STAT5 from PB CD34+ cells by a lentiviral RNAi approach in the presence of TPO and SCF resulted in an increase in CFU-Mk while BFU-E were decreased. Activation of STAT5A resulted in the onset of erythropoiesis at the expense of megakaryopoiesis.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
We acknowledge Dr. H. Spits from the Department of Cell Biology and Histology, Amsterdam Medical Center, Division of Immunology, Netherlands Cancer Institute, Amsterdam, for the kind gifts of pTRIP Renilla RNAi and pTRIP STAT5 RNAi, and Amgen Inc. and Kirin for providing cytokines. This work was supported by an VENI grant from the Nederlandse organisatie voor Wetenschappeliijk Onderzoek (2004–2008) grant (to J.J.S.) and by a grant from the Sanquin Blood Supply Foundation (PPOC-03-013).