AML1/RUNX1 is the most frequently mutated gene in leukaemia and is central to the normal biology of hematopoietic stem and progenitor cells. However, the role of different AML1 isoforms within these primitive compartments is unclear. Here we investigate whether altering relative expression of AML1 isoforms impacts the balance between cell self-renewal and differentiation in vitro and in vivo.
Methods and Findings
The human AML1a isoform encodes a truncated molecule with DNA-binding but no transactivation capacity. We used a retrovirus-based approach to transduce AML1a into primitive haematopoietic cells isolated from the mouse. We observed that enforced AML1a expression increased the competitive engraftment potential of murine long-term reconstituting stem cells with the proportion of AML1a-expressing cells increasing over time in both primary and secondary recipients. Furthermore, AML1a expression dramatically increased primitive and committed progenitor activity in engrafted animals as assessed by long-term culture, cobblestone formation, and colony assays. In contrast, expression of the full-length isoform AML1b abrogated engraftment potential. In vitro, AML1b promoted differentiation while AML1a promoted proliferation of progenitors capable of short-term lymphomyeloid engraftment. Consistent with these findings, the relative abundance of AML1a was highest in the primitive stem/progenitor compartment of human cord blood, and forced expression of AML1a in these cells enhanced maintenance of primitive potential both in vitro and in vivo.
These data demonstrate that the “a” isoform of AML1 has the capacity to potentiate stem and progenitor cell engraftment, both of which are required for successful clinical transplantation. This activity is consistent with its expression pattern in both normal and leukaemic cells. Manipulating the balance of AML1 isoform expression may offer novel therapeutic strategies, exploitable in the contexts of leukaemia and also in cord blood transplantation in adults, in whom stem and progenitor cell numbers are often limiting.
Citation: Tsuzuki S, Hong D, Gupta R, Matsuo K, Seto M, Enver T (2007) Isoform-Specific Potentiation of Stem and Progenitor Cell Engraftment by AML1/RUNX1 . PLoS Med 4(5): e172. https://doi.org/10.1371/journal.pmed.0040172
Academic Editor: Elaine Dzierzak, Erasmus University, The Netherlands
Received: June 26, 2006; Accepted: March 19, 2007; Published: May 15, 2007
Copyright: © 2007 Tsuzuki et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors received no specific funding for this study.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: BM, bone marrow; BrdU, 5-bromo-2′-deoxyuridine; CAFC;, cobblestone area-forming cell; CFC, colony-forming cell; CFU, colony-forming unit; CFU-G, granulocyte-restricted CFU; CFU-GM, granulocyte/macrophage-restricted CFU; CFU-meg, CFU megakaryocyte; CFU-mix, CFU mixed lineage; CSF, colony-stimulating factor; Epo, erythropoietin; FACS, fluorescence-activated cell sorter; 5-FU, 5-fluorouracil; GFP, green fluorescent protein; HSC, hematopoietic stem cell; IL, interleukin; IRES, internal ribosome entry site; KSL, c-Kit+ Sca-1+Lin−; LTC-IC, long-term culture initiating cell; meg, megakaryocyte; SCF, stem cell factor
Blood contains red blood cells (which carry oxygen round the body), platelets (which help the blood to clot), and white blood cells (which fight off infections). All these cells, which are regularly replaced, are derived from hematopoietic stem cells, blood-forming cells present in the bone marrow. Like all stem cells, hematopoietic stem cells self-renew (reproduce themselves) and produce committed progenitor cells, which develop into mature blood cells in a process called hematopoiesis. Many proteins control hematopoiesis, some of which are called transcription factors; these factors bind to DNA through their DNA-binding domain and then control the expression of genes (that is, how DNA is turned into proteins) through particular parts of the protein (their transcription regulatory domains). An important hematopoietic transcription factor is AML1—a protein first identified because of its involvement in acute myelogenous leukemia (AML, a form of blood cancer). Mutations (changes) in the AML1 gene are now known to be present in other types of leukemia, which are often characterized by overproliferation of immature blood cells.
Why Was This Study Done?
Because of AML1′s crucial role in hematopoiesis, knowing more about which genes it regulates and how its activity is regulated could provide clues to treating leukemia and to improving hematopoietic cell transplantation. Many cancer treatments destroy hematopoietic stem cells, leaving patients vulnerable to infection. Transplants of bone marrow or cord blood (the cord that links mother and baby during pregnancy contains peripheral blood stem cells) can replace the missing cells, but cord blood in particular often contains insufficient stem cells for successful transplantation. It would be useful, therefore, to expand the stem cell content of these tissues before transplantation. In this study, the researchers investigated the effect of AML1 on self-renewal and differentiation of hematopoietic stem and progenitor cells in the laboratory (in vitro) and in animals (in vivo). In particular, they have asked how two isoforms (closely related versions) of AML1 affect the ability of these cells to grow and differentiate (engraft) in mice after transplantation.
What Did the Researchers Do and Find?
The researchers artificially expressed AML1a and AML1b (both isoforms contain a DNA binding domain, but only AML1b has transcription regulatory domains) in mouse hematopoietic stem and progenitor cells and then tested the cells' ability to engraft in mice. AML1a-expressing cells engrafted better than unaltered cells and outgrew unaltered cells when transplanted as a mixture. AML1b-expressing cells, however, did not engraft. In vitro, AML1a-expressing cells grew more than AML1b-expressing cells, whereas differentiation was promoted in AML1b-expressing cells. To investigate whether the isoforms have the same effects in human cells, the researchers measured the amount of AML1a and AML1b mRNA (the template for protein production) made by progenitor cells in human cord blood. Although AML1b (together with AML1c, an isoform with similar characteristics) mRNA predominated in all the progenitor cell types, the relative abundance of AML1a was greatest in the stem and progenitor cells. Furthermore, forced expression of AML1a in these cells improved their ability to divide in vitro and to engraft in mice.
What Do These Findings Mean?
These findings indicate that AML1a expression increases the self-renewal capacity of hematopoietic stem and progenitor cells and consequently improves their ability to engraft in mice, whereas AML1b expression encourages the differentiation of these cell types. These activities are consistent with the expression patterns of the two isoforms in normal hematopoietic cells and in leukemic cells—the mutated AML made by many leukemic cells resembles AML1a. Because the AML1 isoforms were expressed at higher than normal levels in these experiments, the physiological relevance of these findings needs to be confirmed by showing that normal levels of AML1a and AML1b produce similar results. Nevertheless, these results suggest that manipulating the balance of AML1 isoforms made by hematopoietic cells might be useful clinically. In leukemia, a shift toward AML1b expression might slow the proliferation of leukemic cells and encourage their differentiation. Conversely, in cord blood transplantation, a shift toward AML1a expression might improve patient outcomes by expanding the stem and progenitor cell populations.
Please access these Web sites via the online version of this summary at http://dx.doi.org/10.1371/journal.pmed.0040172.
- Wikipedia has pages on hematopoiesis and hematopoietic stem cells (note: Wikipedia is a free online encyclopedia that anyone can edit; available in several languages)
- The US National Cancer Institute has a fact sheet on bone marrow and peripheral blood stem cell transplantation (in English and Spanish) and information for patients and professionals on leukemia (in English)
- The American Society of Hematology provides patient information about blood diseases, including information on bone marrow and stem cell transplantation
AML1 (also termed RUNX1 or CBFA2), which encodes a transcription factor, was originally identified as a component of the fusion gene generated as a consequence of the t(8; 21) chromosomal translocation characteristic of a subset of acute myelogenous leukaemia . Subsequently, mutations in the AML1 gene were shown to be associated with a number of malignant and premalignant conditions including acute myelogenous leukaemia , childhood acute lymphocytic leukaemia, familial platelet disorder, and myelodysplastic syndromes. AML1 is thus the most frequently mutated gene in human leukaemia [2,3]. Aml1 expression marks long-term repopulating hematopoietic stem cells (HSCs) in the midgestation mouse embryo , and gene targeting experiments have unequivocally shown Aml1 to be essential for the development of the hematopoietic system; Aml1-null mice die in utero of haemorrhage and lack adult-type hematopoiesis because of a failure in the generation of definitive HSCs [5,6]. In contrast, conditional gene-targeting experiments have revealed that Aml1 is dispensable for maintenance of adult HSCs. Aml1 deficiency in adult stages is accompanied by increases in myeloid progenitors and phenotypically defined HSCs, as well as impaired development of lymphoid and megakaryocytic cells [7–9].
While AML1 is known to regulate the expression of many hematopoietic genes, including those encoding cytokines and their receptors as well as other transcription factors, the full spectrum of its target genes in different cell types in normal hematopoiesis, or AML1-associated leukaemia, remains unknown [10,11]. Furthermore, the manner in which AML1 is itself regulated at the level of transcription and post-transcriptional processing is largely not understood . Addressing these issues will be essential to understanding how AML1 plays such a crucial role in both normal and leukaemic hematopoiesis. Potentially confounding factors in these endeavours are the large size of the AML1 locus, the complexity of its putative regulatory elements, and the prevalence within the AML1 gene of extensive alternative splicing [13–15].
Evidence is accumulating that isoforms of key hematopoietic transcription factors can play both distinct and key regulatory roles in hematopoiesis. A case in point is GATA1, which regulates the differentiation of erythroid and megakaryocytic cells . Full-length GATA1 is composed of an amino-terminal transactivation domain as well as two zinc fingers that mediate its binding to DNA. However, an isoform lacking the transactivation domain is also expressed. This short form is impaired in its ability to promote differentiation and is implicated in abnormal hematopoiesis associated with Down syndrome [17,18]. Similarly, truncated forms of SCL/TAL1 favour erythroid lineage development, whereas full-length forms enhance the development of the megakaryocytic lineage . An alternatively spliced isoform of TEL/ETV6 has recently been shown to inhibit, in a dominant-negative manner, the activity of full-length TEL and thereby negatively regulate the TEL-mediated erythroid differentiation of mouse erythroleukaemia cells . Finally, enforced expression of an isoform of IKAROS/IKZF (IKAROS 6), which lacks the DNA-binding domain, in human CD34+ cells limits B cell development . IKAROS isoforms are also implicated in the development of leukaemia [22–24].
The human AML1 gene generates three alternatively spliced variants, AML1a, AML1b, and AML1c . AML1b and AML1c both possess the DNA-binding region (runt domain) at the amino-terminal end as well as the carboxyl-terminal transcriptional regulatory domains; AML1b and AML1c are therefore considered to be broadly similar in function. In contrast, AML1a retains the DNA-binding domain, but lacks the transcriptional regulatory domains. AML1a is thus considered a potential functional antagonist of AML1b and AML1c . The opposing functions of AML1a and AML1b have been shown in myeloid-committed cell lines with respect to their granulocyte colony-stimulating factor (G-CSF)–induced differentiation programmes; ectopically expressed AML1a blocks granulocytic differentiation, and this blockade is released by enforced expression of AML1b. Differentiation arrest in this cell-line model is accompanied by uncompromised growth under culture conditions that normally induce differentiation in the parental cell line . AML1a transcripts have also been noted in selected leukaemia samples, as have mutant truncated forms of the AML1 protein that structurally resemble AML1a [26–29]. However, little is known about the potentially differential roles of AML1 isoforms in regulating normal or leukaemic cell fates in the stem and progenitor compartment of hematopoiesis. To address these questions, we expressed AML1a and AML1b in primary murine bone marrow (BM)-derived hematopoietic stem/progenitor cells and analysed the resulting effects on hematopoiesis both in vitro and in vivo. We then performed similar experiments in the corresponding human compartments. Such studies may lead to new strategies for the manipulation of haematopoietic progenitors and inform future treatments for haematological malignancies
Generation of Retroviruses
cDNAs for AML1a and AML1b  (generous gifts of Misao Ohki, National Cancer Center, http://www.ncc.go.jp) tagged with the FLAG-epitope at the amino terminal were cloned upstream of the internal ribosome entry site (IRES) element into the EcoRI/XhoI restriction site of the murine stem cell virus–internal ribosome entry site–green fluorescent protein (GFP) vector (MSCV-IRES-GFP) . The MSCV-IRES-human CD4 vector that expresses an extracellular portion of the human CD4 molecule was generously provided by Jason G. Cyster (University of California San Francisco, http://www.ucsf.edu) . Retrovirus vector plasmids were transiently transfected into Phoenix-Eco packaging cells (obtained from American Type Culture Collection, http://www.atcc.org) by a standard calcium phosphate precipitation method. Virus supernatants were then concentrated by centrifugation before use in infecting cells .
Retroviral Infection of Mouse BM Cells and Transplantation
BM cells were harvested from tibiae and femora of BALB/c mice 4 d after intravenous administration of 5-fluorouracil (5-FU) (150 mg/kg) and treated with ACK (0.15 M NH4Cl, 1.0 mM KHCO3, 0.1 mM EDTA) to lyse red blood cells. BM cells were then prestimulated with stem cell factor (SCF) (100 ng/ml) + interleukin (IL) 3 (IL3) (10 ng/ml) + IL6 (6 ng/ml) (Peprotech, http://www.peprotech.com) for 24 h and spin-infected with concentrated virus in the presence of cytokines and polybrene (4 μg/ml) (Sigma, http://www.sigmaaldrich.com) . Cells (1 × 106) were transplanted intravenously into lethally irradiated (7.5 Gy) recipient mice, which were maintained on acidified water throughout the experimental period . In some experiments, lin−c-kit+ hematopoietic progenitor cells sorted from C57BL6 mice were used, in place of 5-FU-treated BM cells. Results obtained using the two different sources of cells were essentially identical. For competitive repopulation, BM cells from the 5-FU-treated mice were infected with retroviruses for either GFP only, AML1a, or human CD4. Cells were then admixed to obtain mixtures of GFP-transduced and CD4-transduced cells, or AML1a-transduced and CD4-transduced cells for transplantation.
Flow Cytometry of Mouse Cells
After red blood cell lysis, cells were incubated with the indicated antibodies, conjugated with either one of phycoerythrin, biotin, or allophyocyanin (APC); biotinylated antibodies were then stained with streptavidin-APC or streptavidin-peridinin-chlorophyll-protein Cy5.5. Nonspecific binding was blocked by preincubation with an anti-Fc receptor antibody (2.4G2). Antibodies used were as follows: anti-CD3 (145–2C11), anti-CD4 (RM4–5), anti-CD8 (53–6.7), anti-B220 (RA3-6B2), anti-CD19 (1D3 and MB19–1), anti-TER-119 (TER-119), anti-Gr-1 (RB6-8C5), anti-Mac-1 (M1/70), anti-c-kit (2B8), anti-CD34 (RAM34), anti-Sca-1 (D7), and anti-human CD4 (RPA-4). Isotype-matched Ig was used as a control. Antibodies were obtained from BD Biosciences (http://www.bdbiosciences.com) and eBioscience (http://www.ebioscience.com). Cells were washed with PBS containing 1% fetal calf serum and analysed on a FACSCalibur (BD Biosciences) with CELLQUEST software. Cell sorting was conducted with a FACSVantage (BD Biosciences).
Western Blot Analysis
Total cell lysates were fractionated by SDS-PAGE and blotted onto a nylon membrane. AML1a and AML1b were detected with an anti-AML1 antibody (Ab-2, Oncogene Research Products, http://www.merckbiosciences.co.uk) or an anti-FLAG antibody (M2, Sigma). Anti-tubulin antibody (Sigma) was used to allow comparison of loaded protein amounts. Primary staining was visualized with a goat anti-rabbit Ig- horseradish peroxidase (HRP) conjugate, a goat anti-mouse Ig–HRP conjugate, or a secondary antibody by using enhanced chemiluminescence (Amersham Bioscience, http://www.amersham.com). Western blot analysis of primary human cord blood was conducted using a 12% acrylamide gel, and the AML1 proteins were detected using an anti-AML1 antibody (Ab-2) in combination with the highly sensitive West Femt reagent (Pierce, http://www.piercenet.com).
In Vitro Colony-Forming Assays of Mouse Cells
Cells were plated in triplicate in methylcellulose under multimyeloid differentiation conditions (SCF [50 ng/ml], IL3 [10 ng/ml], IL6 [10 ng/ml], and erythropoietin [Epo, 3 units/ml]) (Methocult M33342, Stem Cell Technology, http://www.stemcell.com), or myelomonocytic differentiating conditions (SCF [50 ng/ml], IL3 [10 ng/ml], IL6 [10 ng/ml], and granulocyte/macrophage-CSF [GM-CSF; 10 ng/ml]). Colonies were typed and counted at day 7. For colony-forming unit (CFU) megakaryocyte (meg) (CFU-meg), α-MEM supplemented with 0.8% methylcellulose, 1% BSA, 30% FBS, 0.1 mM 2-mercaptoethanol, and 10 units of mouse thrombopoietin (Peprotech) was used .
In Vitro Culture of Mouse Cells
Transduced cells were maintained in Iscove's modified Dulbecco's medium (Gibco, http://www.invitrogen.com) supplemented with 10% fetal calf serum in the presence of SCF (50 ng/ml) and IL3 (10 ng/ml). Cytokines were purchased from Peprotech.
Cells in culture were incubated with 10 μM of 5-bromo-2′-deoxyuridine (BrdU) (WAKO, http://www.wako-chem.co.jp/english) for 30 min and then fixed in 70% cold ethanol and treated with 4 N HCl. Incorporated BrdU was visualized by anti-BrdU antibody conjugated to fluorescein (BU.1, Chemicon International, http://www.chemicon.com) and analysed on a FACSCalibur along with DNA content as stained by propidium iodide.
Long-Term Culture–Initiating Cell and Cobblestone Area-Forming Cells Assays for Mouse Cells
For long-term culture-initiating cell (LTC-IC) assays, GFP+ and GFP− cells from the lin− compartment were sorted from mouse BM by flow cytometry and inoculated onto pre-established irradiated stromal cells by limiting dilution in Myelocult medium (Stem Cell Technology). Cells were cultured for 4 wk and then subjected to colony formation assay in Methocult M33342, following the manufacturer's instructions. For cobblestone area-forming cell (CAFC) assays, the stromal cell line FBMD-1 (a kind gift from Dr. Gary Van Zant, University of Kentucky, http://www.uky.edu) was used. Total BM cells sorted on the basis of GFP expression were inoculated onto the pre-established stromal cell layer by limiting dilution and cultured for 5 wk.
Cord blood cells were collected, with written informed consent, for use in research. Total CD34+ cells and total CD15+ cells were purified using MACS magnetic beads (Miltenyi Biotec, http://www.miltenyibiotec.com) according to the manufacturer's instructions. CD34+ subsets were isolated by cell sorting using antibodies against human CD14, CD19, CD33, and CD133. For the isolation of CD36+ cells, purified CD34+ cells were cultivated in vitro with SCF (50 ng/ml), IL3 (20 ng/ml), and IL6 (20 ng/ml) for 7 d and isolated using a CD36-PE antibody in combination with anti-PE magnetic beads . The resulting cells gave 98% erythroid output when plated in methylcellulose containing SCF, IL3, IL6, GM-CSF, G-CSF, and Epo.
Quantification of AML1 Isoforms in Human Cord Blood by Reverse Transcription-PCR
Primers 5′-GAGGGAAAAGCTTCACTCTGA-3′ (primer C), GTGTACCGGGATCCATGCTA-3′ (primer 3A), and 5′-GTTGAGAGTCGACTGGAAAG-3′ (primer 2B) were used to detect AML1a (primers C and 3A) and AML1b/AML1c (primers C and 2B) as described . PCR products thus obtained were then used as templates for additional PCR with the following primer sets to generate competitor DNAs: 5′-TTTGAATTCGAGGGAAAAGCTTCACTCTGAGAACCTCGAAGACATCGGCA-3′ (primer C′), 5′-TTTGAATTCGTGTACCGGGATCCATGCTA-3′ (primer 3A′), and 5′-TTTGAATTCGTTGAGAGTCGACTGGAAAG-3′ (primer 2B′). These primer sets were designed to generate products 157 bp shorter than the authentic PCR products and to facilitate subcloning. PCR products with primers C′ and 3A′, or with primers C′ and 2B′, were digested with EcoRI, and inserted into the EcoRI site of the pBluescript vector (Stratagene, http://www.stratagene.com) to obtain pBS/competitor A and pBS/competitor B. NotI/XhoI fragment of pBS competitor A was then inserted into NotI/XhoI site of the pMXs vector, then the EcoRV/XhoI fragment was inserted into the EcoRV/XhoI site of the pBS/competitor B. The plasmid thus obtained (pBS/competitor A + B) contained competitors for AML1a and AML1b in back-to-back orientation, with 1,250 bp of intervening spacer. This pBS/competitor A + B was used for the competitive PCR for both AML1a and AML1b/AML1c to quantitate AML1 isoforms in a given cell population . Total RNA was isolated from the fractionated cord blood cells using Trizol reagent (Invitrogen, http://www.invitrogen.com) and then reverse-transcribed using random primers and Superscript II enzyme according to the manufacturers' instructions. cDNA thus obtained was then mixed with known amounts of the competitor plasmid and PCR was conducted either with primer sets C and 3A, or with primer sets C and 2B. PCR products were then fractionated on a 2.5% agarose gel electrophoresis, and the PCR products from endogenous AML1 and from the competitor plasmid were quantitated densitometrically. The amounts of the endogenous AML1 isoforms were then estimated as described  and normalized on the basis of G6PD expression estimated by a quantitative real-time PCR .
Preparation of Lentivirus and Cord Blood Cell Transduction
An AML1a cDNA was cloned into the pHR-SINCSGW lentivirus , which carries an emerald-GFP reporter. Lentiviruses were pseudotyped with the vesicular stomatitis virus G (VSVG) protein by transient transfection of 293T cells. High-titre viral stocks were prepared by ultracentrifugation . CD34-enriched cord blood cells were infected for 24 h by lentivirus (either AML1a or control virus) with MOI of 100 in serum free medium Stem/Span (Stem Cell Technologies) supplemented with SCF (100 ng/ml), FLT-3 ligand (100 ng/ml), thrombopoietin (20 ng/ml), and IL6 (20 ng/ml) (all growth factors were from Peprotech) [38,39].
In Vitro Analysis of Human Cells
Colony-forming cell (CFC) assays with human cells were under serum-free conditions with recombinant human SCF, GM-CSF, IL3, IL6, G-CSF, and Epo (MethoCult H4536, StemCell Technologies). LTC-IC assays with human cells were performed in 96-well format in Myelocult H5100 medium (StemCell Technologies) on a feeder layer comprising a 1:1 mixture of irradiated M210B4 and S1/S1 mouse fibroblasts . For limiting dilution CAFC assays MS-5 stroma cells  were maintained in α-MEM medium complemented with 2 mM L-glutamine, 2 mM sodium pyruvate, 10% FBS (Stem Cell Technologies). After the cells were inoculated, weekly half medium changes were performed for the duration of the culture. CFAC were scored at 6 wk [42,43].
Nonobese Diabetic–Severe Combined Immunodeficient Mice and Transplantations
Nonobese diabetic–severe combined immunodeficient (NOD-SCID) mice were bred and maintained at the Weatherall Institute of Molecular Medicine (WIMM) animal facility in accordance with Home Office regulations (United Kingdom). Animals were handled under sterile conditions. Transplantations were performed by tail-vein or intrafemoral injection into 6- to 10-wk-old mice. Recipients received 350 cGy of total body irradiation, and BM engraftment was assessed at 8 wk post-transplantation. Human cell engraftment was assessed by staining of BM cells with anti-CD19-APC (DAKO), anti-CD34-PE (BD Biosciences Pharmingen), anti-CD45-APC (DAKO, http://www.dako.com), and CD38-FITC (DAKO).
Effects of AML1a and AML1b on Long-Term Engrafting Cells and Hematopoiesis In Vivo
BM cells were isolated from 5-FU-treated mice and transduced with murine stem cell virus -based retroviruses containing amino-terminal FLAG-tagged AML1a or AML1b inserted upstream of an IRES-GFP cassette (Figure 1A). GFP+ and GFP− cells (Figure 1B) were separated by flow cytometry, and expression of AML1a and AML1b were confirmed by Western blotting using an antibody directed against a peptide within the AML1 runt domain (Figure 1C), which also detects endogenous murine Aml1; the identity of the AML1a and AML1b were additionally confirmed by Western blotting with an antibody directed against the FLAG tag (Figure 1C). These results also showed that the expression level of exogenous AML1a and AML1b was less than 2-fold that of endogenous Aml1 in the cells used subsequently for transplantation.
(A) Recombinant viruses.
(B) FACS analysis of GFP expression in infected hematopoietic cells immediately after infection with control, AML1a-containing, or AML1b-containing viruses.
(C) Western blot analysis of lysates from cells sorted immediately after infection by using anti-AML1, anti-Flag, and anti-tubulin antibodies. The anti-AMLl antibody was raised against the Runt domain and therefore detected both human and endogenous mouse AML proteins.
(D) Western blot analysis of GFP+ and GFP− BM cells from animals 4 mo after the transplantation (antibodies as in [C]).
(E) Typical FACS analyses of BM cells from competitive repopulation assays of AML1a-transduced cells. BM cells of mice transplanted with mixtures of human CD4− and GFP-only-control-transduced cells (a) or human CD4− and AML1a-transduced cells (b) were analysed for expression of human CD4 and GFP.
Animals were transplanted in cohorts to facilitate kinetic studies of engraftment (Table 1). The contribution of control vector-transduced cells was relatively stable over time in all cohorts followed. AML1b-transduced cells, in contrast, never contributed to BM at 1 mo or thereafter. Strikingly, the contribution of AML1a-transduced cells increased with time; in a typical case (cohort 4), percentage GFP was 10.5% at 3.5 mo, 11.4% at 4 mo, 42.2% at 11.5 mo, and 53.4% at 12 mo. Western blotting showed that exogenous AML1a expression was 2- to 3-fold higher than endogenous Aml1 and confined exclusively to GFP+ cells within the graft (Figure 1D). These results suggest that while AML1b expression abrogates engraftment potential, AML1a expression confers a growth advantage on BM cells in vivo.
Kinetic Analysis of Engraftment of AML1a-, AML1b-, and Control Vector-Transduced BM Cells
We therefore examined this growth advantage using competitive repopulation assays. Competitor cells marked by expression of the extracellular domain of a human CD4 were mixed with either AML1a- or control vector-transduced cells just prior to transplantation into cohorts of lethally irradiated mice. In one cohort, a mixture of 6.2% of GFP+ control vector-transduced cells and 4.0% of human CD4+-transduced competitor cells were used for the transplantation. In the second cohort, a mixture of 4.5% GFP+ AML1a-transduced cells and 5.7% of human CD4+-transduced competitor cells were used. Subsequent analysis of engrafted cells—a typical example is shown in Figure 1E—revealed the engraftment potential of AML1a-expressing cells to be substantially higher than their control-vector transduced counterparts. Kinetic studies conducted over a 9-mo period (Table 2) showed the ratio of GFP and CD4 to be relatively stable in the control-vector/CD4 cohort. In contrast, the GFP/CD4 ratio increased with time in the AML1a/CD4 mice cohort (Table 2). This property was also evident in the secondary transplantation experiments: when BM cells from a primary recipient exhibiting GFP/CD4 ratio = 1.6 at 3 mo post-transplant were secondarily transplanted, the secondary recipients displayed ratios of 3.0 ± 1.41 (mean ± standard deviation) at 3 mo and 9.0 ± 3.29 at 8 mo post-transplant. This increase in the GFP/CD4 ratio was not seen in mice secondarily transplanted with BM cells taken from control (control-vector/CD4) cohorts. Collectively, these results suggest that expression of AML1a confers a high level of competitive engraftment potential.
Competitive Repopulation Assays of AML1a-Transduced Cells in First and Second Transplantations
Results from GFP+ and GFP− cells are shown in solid and open symbols, respectively.
(A) Percentages of KSL cells in GFP+ and GFP− fraction of individual animals are presented pairwise. All animals were analysed 4 mo after transplantation. Statistical significance was evaluated using Wilcoxon and Mann-Whitney tests.
(B) LTC-IC analysis of GFP+ and GFP− cells from the lin− BM compartment of AML1a and control animals. LTC-IC frequency within the lin− compartment and its 95% confidence interval (in parenthesis) are presented. Two independent experiments gave a similar result.
(C) CAFC analysis of GFP+ and GFP− cells from total BM of AML1a and control animals. Wells containing clusters of six or more cobblestone cells after 35 d of culture were scored as positive, and the frequency of CAFC calculated by Poisson statistics; CAFC frequency and its 95% confidence interval (in parenthesis) are presented. Two independent experiments gave a similar result.
(D) Colony-forming activity BM cells sorted from an AML1a animal and cultured under multimyeloid (CFU-GM, erythropoietic burst formation [BFU-E], and CFU-mix) and megakaryocytic (CFU-meg) conditions are shown. Representative data from three independent experiments conducted in triplicate are presented. Asterisks indicate statistical significance of the difference estimated by Mann-Whitney test.
(E) Photomicrographs show relative sizes of colonies generated in multimyeloid condition.
We explored this competitive potential further by examining the status of the primitive progenitor compartment in primary AML1a recipients. We determined the number of c-Kit+Sca-1+Lin− (KSL) cells within GFP+ and GFP− fractions of BM in each of six AML1a and six control animals 6 mo post-transplantation, and the results are shown pairwise for individual mice in Figure 2A. No difference in KSL frequency was observed between GFP+ and GFP− fractions from control animals. In contrast, the GFP+ fractions of AML1a mice contained significantly more KSL than the corresponding GFP− fractions of the same animals. Interestingly, however, percentage KSL was not significantly different between the GFP+ fractions of AML1a and control animals. This result may reflect the relative competitiveness of GFP+ and GFP− cells within the BM of a single animal, or alternatively percentage KSL may not entirely reflect the frequency of functional stem and multipotent progenitors. We therefore quantitated the number of functional primitive hematopoietic cells using LTC-IC and CAFC assays, and typical data from one such experiment are shown in Figure 2B and 2C. In the case of LTC-IC assays, GFP+lin− and GFP−lin− cells were isolated from transplanted animals by flow cytometry and cocultured using the limiting dilution method on pre-established BALB/c mouse BM stromal layers. After 4 wk of culture, the cells were harvested and assayed for CFC activity to enumerate the frequency of primitive hematopoietic cells. In the case of CAFC assays, total GFP+ or GFP− cells were purified from transplanted mice and cultured using the limiting dilution method on FBMD-1 cell line stromal layers; cobblestone areas, comprising six or more cells, were enumerated 5 wk later. These in vitro assays revealed a significant increase in the frequency of primitive hematopoietic cells in AML1a-positive BM-fraction; comparison of the GFP+ versus GFP− fractions showed a 4-fold increase (1/388.1 versus 1/1,666.2) in LTC-IC (Figure 2B) within lin− cells and a 14-fold increase (1/7,808.2 versus 1/10,9187.2) in CAFC (Figure 2C) within unfractionated cells. These increases were not seen in the corresponding fractions of cells from control animals (Figure 2B and 2C).
We next examined the committed progenitor compartment using CFC assays. GFP+ cells from AML1a mice displayed increased total CFC activity (Figure 2D) primarily in granulocyte- and/or macrophage-restricted (CFU-GM) and primitive mixed lineage (CFU-mix) progenitors. Furthermore, AML1a-expressing colonies (GFP+) were larger than those produced from control cells (GFP−) from the same animals (Figure 2E). AML1a expression did not alter the frequency of erythroid colonies (erythropoietic burst formation) generated but was associated with an increase in CFU-meg formation (Figure 2D), but cells in these CFU-meg colonies were smaller than controls, displaying hypolobulated nuclei indicative of impaired differentiation (unpublished data). Consistent with these CFC data, the lin−c-kit+ progenitor compartment was increased 4.0-fold in the GFP+ versus GFP− fraction in animals transplanted with AML1a-transduced BM, and these animals also displayed increases in immature blast-like cells (Figure S1). Morphological examination of sorted BM cells from AML1a and control mice also revealed an over-representation of myeloid versus lymphoid cells in the GFP+ fraction of AML1a-mice (Figure S1A), which was confirmed by immunophenotyping (unpublished data). There were no morphological or immunophenotypic indications of arrested differentiation in the AML1a-expressing myeloid compartment (Figure S1B and S1C).
AML1a and AML1b Display Opposing Effects on Myeloid Differentiation In Vitro
The abrogation of engraftment by AMLb precluded further analysis of its effects on haematopoietic cells in vivo. We therefore compared effects of AMLb and AML1a on the clonogenic activity of appropriately transduced 5-FU cells (see Figure 1A and 1B) in vitro using CFU in culture assays. The total number of colonies was similar in AML1a- and control vector-transduced cells, but significantly reduced in AML1b-expressing cells (Figure 3A). Upon serial replating—an assay of self-renewal—AML1b-expressing cells gave rise to significantly fewer colonies than control vector-transduced cells at all passages examined, whereas expression of AML1a produced more colonies. Furthermore, compared with colonies produced by control vector-transduced cells, colonies from AML1a-expressing cells were larger, whereas colonies from AML1b-expressing cells were smaller (Figure S2A). Analysis of cell-cycle status, as assessed by using BrdU incorporation and DNA content, revealed that AML1a-expressing cells were more proliferative than controls or AML1b-expressing cells (Figure 3B and 3C). Results of colony typing and morphological assessment of cells within primary colonies are shown in Figure 3D. AML1b-expressing cells produced no erythroid output and gave rise to significantly more granulocyte-restricted CFU (CFU-G) than control cells. The overall distribution of colony type produced by AML1a-expressing cells was similar to that seen in controls, but the granulocytic component within CFU-GM colonies was markedly decreased (unpublished data). AML1b-expressing cells did not produce CFU-meg colonies (Figure 3E). In contrast, AML1a cells produced more CFU-meg than controls (Figure 3E). When compared with vector-transduced control cells, the Gr-1-high mature granulocytic fraction was increased by AML1b expression and decreased by AML1a expression in both multimyeloid and myelomonocytic conditions. However, upon replating, AML1a-expressing cells started to express high levels of Gr-1 (Figure 3F), which was accompanied by the appearance of morphologically mature granulocytes (Figure 3G). Similar results were obtained when cells were plated under myelomonocytic- (SCF + IL3 + IL6 + GM-CSF) promoting culture conditions (unpublished data).
(A) Frequency of CFCs in serial replating experiments; we sorted 1 × 104 GFP+ cells immediately after infection with GFP vector-only, AML1a-containing, or AML1b-containing viruses, and they were plated. For replating, cultures were harvested on day 7 and 1 × 104 cells replated. Representative data from no fewer than three independent experiments conducted in triplicate are presented. Kruskal-Wallis test showed statistical significance of difference (p < 0.05) of colony numbers generated by the three populations (GFP, AML1a, and AML1b) in rounds 2, 3, and 4, but not in round 1. Difference of any given two populations was therefore analysed in rounds 2, 3, and 4, by Mann-Whitney test, with asterisks indicating statistical significance of the difference.
(B) Representative cell-cycle analyses of control and AML1a- or AML1b-transduced cells from the second round of replating are presented. Cells harvested from CFC assays were pulsed with BrdU and analysed for BrdU incorporation along with DNA content.
(C) Graphical summary of cell-cycle data from four independent experiments. Kruskal-Wallis test showed statistical significance of difference (p < 0.01) of the three populations (GFP, AMl1a, and AML1b) and Mann-Whitney tests showed statistical significance of the difference between any given two populations (p < 0.05).
(D) Colony types formed in multimyeloid conditions (upper bar graph) are shown. After typing on day 7, cultures were harvested for cytospin preparations and differential counting (lower bar graph).
(E) CFU-meg formation was analysed. BM cells from 5-FU-treated BALB/c mice were infected with viruses for vector-alone, AML1a, and AML1b, and then sorted into GFP+ and GFP− cells for colony assays (triplicate). Two experiments yielded a similar result. Asterisks indicate statistical significance of the difference estimated by Mann-Whitney test.
(F) FACS analysis for myeloid-differentiation markers (Gr-1 and Mac-1) on cells harvested from colonies at the first and second replating.
(G) Photomicrograph of AML1a-transduced cells harvested from second replating.
Collectively, these findings suggest that AML1b expression promotes granulocytic differentiation at the expense of alternative lineages, and AML1a expression retards but does not block granulocytic maturation of BM cells under the conditions used while promoting self-renewal and proliferation of progenitor cells. Furthermore, AML1a-transduced cells recovered from the third round of replating propagated in suspension culture, displaying a cumulative cell expansion of about 1014-fold in 9 wk (Figure 4A). The cultures comprised an adherent layer of macrophages and a mixture of floating monocytic, granulocytic, and blast-like cells (Figure S2B). These cultures were reproducibly established in both BALB/c and C57BL6-derived cells and continued to express exogenous AML1a as evidenced by Western blotting with anti-RHD and anti-FLAG antibodies (Figure 4B). Flow cytometric analysis of the nonadherent cells revealed the following subpopulations; (1) Mac-1−/c-kit+, (2) Mac-1low/c-kitlow, (3) Mac-1high/c-kit−/Gr-1−, and (4) Mac-1high/c-kit−/Gr-1+ (Figure 4C). Expression of lymphoid (CD3, CD4, CD8, CD5, CD25, B220, CD19, IL7-receptor alpha, AA4.1, BP-1, and NK1.1) and erythroid (TER-119) markers was not detected; a proportion of Mac-1−/c-kit+ cells expressed Sca1 (unpublished data). Morphological analysis (Figure 4D) and colony-forming assays (Figure 4E) showed Mac-1− cells to be blast-like with a colony-forming efficiency of 4%–7%, Mac-1low cells to be immature myeloid cells with a lower clonogenic capacity (0.5%–0.7%), and Mac-1high cells to be a mixture of monocytic and granulocytic cells with no colony-forming potential. The Mac-1high/Gr-1+ fraction contained exclusively granulocytes as judged by morphology and enzymatic assays (unpublished data). Finally, fluorescence-activated cell sorter (FACS)-purified Mac-1−/c-kit+ cells were able to generate all of the other cell types observed in the cultures (unpublished data). The use of Epo or ST-2 stromal cells to obtain in vitro erythroid or lymphoid outputs, respectively, was not successful, however (unpublished data). These results suggest that AML1a-expressing Mac-1−/c-kit+ cells are capable of self-renewal and myelomonocytic differentiation in vitro.
Cells harvested from the third replating of AML1a-transduced cells were cultured in medium containing SCF + IL3. Cells were counted weekly and replated at 1 × 105 cells per millilitre in fresh medium and cytokines. Successful long-term cultures were established in four out of four experiments using either cells derived from BALB/c or C57BL6 mice.
(A) Fold increase of cell number are presented on a logarithmic scale versus time.
(B) Western blot analysis of cultured AML1a-transduced cells using anti-AML1 (anti-RHD) and anti-FLAG antibodies.
(C) Flow cytometric analyses of the cultured AML1a-transduced cells using anti-c-kit and anti-Mac-1 (left) and anti-Gr-1 and anti- Mac-1 (right) antibodies.
(D) May-Grunwaldt-Giemsa staining of the sorted Mac-1−, Mac-1low, and Mac-1high cells in culture.
(E) CFC activity of sorted Mac-1−, Mac-1low, and Mac-1high cells in SCF and IL3. Typical data from three independent experiments conducted in triplicate are presented. Kruskal-Wallis test showed statistical significance (p < 0.05) of the three populations (Mac-1−, Mac-1low, and Mac-1high), and Mann-Whitney tests showed statistical significance of the difference (p < 0.05) of any given two populations.
(F and G) Engraftment and lineage potential were analysed. Mice were lethally irradiated and transplanted with 107 cultured AML1a-transduced cells and 5 × 105 fresh “radioprotective” BM cells to mitigate against immediate effects of lethal radiation. BM and spleen cells were gated for GFP and analysed for lineage markers ten days after transplantation. Gr-1 (granulocytic) and Mac-1 (myeloid) expression in BM and spleen were analysed (F). Given the predominance of Mac-1+ cells in the graft, Mac-1− cells were analysed for TER-119 (erythroid), B220 (B-lymphoid), and CD19 (B-lymphoid) (G). Grey shading shows marker expression, and the black dotted lines show profiles of isotype-matched control antibodies. Representative data from three independent experiments are shown.
We next transplanted cells from these proliferating in vitro cultures into lethally irradiated syngeneic mice together with a radioprotective dose of fresh BM cells. We found 30%–40% of harvested spleen and BM cells expressed GFP, demonstrating successful engraftment of cultured AML1a-transduced cells 10 to 20 d after the transplantation (unpublished data). Engraftment was predominantly myelomonocytic (Figure 4F), but GFP+ cells contributed to the erythroid (TER-119) and B-lymphoid (B220, CD19) lineages; given the predominance of myelomonocytic Mac+ cells within the GFP+ fraction, this contribution to erythroid and B-lymphoid lineages is most clearly apparent in the GFP+/Mac-1− fraction (Figure 4G). GFP+ cells did not contribute, however, to the thymus, nor were CD3+ cells seen in BM or spleen. These findings demonstrate that the cultured AML1a-transduced cells have erythroid and B-lymphoid as well as myeloid differentiation potential. The engraftment of these cultured AML1a-transduced cells was not persistent, however, with virtually no GFP+ cells being detectable in the BM or spleen of the transplanted animals at 5 wk post-transplant and thereafter. Collectively, these data suggest that AML1a expression maintains short-term reconstituting multipotent progenitors prolonged in vitro culture.
Analysis of AML1 Isoforms in Human Cord Blood
Taken together, the data obtained in mouse cells raised the possibility that AML1 may have some utility in the setting of expansion of human cord blood-derived stem and progenitor cells. We therefore next assessed the expression of AML1 isoforms in different cellular compartments of human cord blood, which is a rich source of stem and progenitor cells. For these experiments a quantitative competitive reverse transcription-PCR strategy was used (see Methods). Total AML1 (AML1a + AML1b/c) expression was similar in progenitor (CD34+) and myelomonocytic (CD15+) subsets, and by comparison substantially was reduced in the erythroid (CD36+) compartment (Figure 5A). In contrast, AML1a expression was only observed in progenitor (CD34+) but not myelomonocytic (CD15+) or erythroid (CD36+) compartments (Figure 5B). CD34+ cells were therefore subfractionated into CD133+ (primitive), CD33+ (immature), CD19+ (B-lymphoid), and CD14+ (macrophage/monocyte) progenitor subsets. Total AML1 expression was highest in CD34+CD133+ cells, followed by CD34+CD33+, then CD34+CD14+, and finally CD34+CD19+ cells (Figure 5A). The AML1a to AML1b/AML1c ratio was about 1:6 in CD34+CD33+, 1:30 in CD34+CD133+, and 1:100 in both CD34+CD19+ and CD34+CD14+ cells (Figure 5B). Collectively, these findings suggest that AML1 is relatively abundant in stem/early myeloid progenitors and under-represented in erythroid progenitors. Thus, while expression of AML1b/AML1c predominated in all fractions analysed, AML1a expression was restricted to CD34+ progenitor compartments and within these were relatively most enriched in the CD34+CD33+ subset. We next examined expression of AML1 isoforms in cord blood cells at the level of protein. Cell extracts of CD34+ and CD34− CB cells from two different individuals were analysed by Western blotting. Although technically difficult given the low cells numbers available, the results suggest that AML1a expression is restricted to the CD34+ compartment where it constitutes a significant fraction of the total AML expressed (Figure 5C). These expression data, particularly the protein data, raise the possibility that expression of AML1a may be relevant to the function of primitive human haematopoietic cell compartments. We explored this possibility functionally by constructing a lentiviral expression vector for AML1a (Figure 6A), using it to transduce CB cells, and subsequently assessed their behaviour in vitro and in vivo. AML1a lentivirus-transduced CD34+CD38− human CB cells were initially assessed in LTC-IC assays on M2-10B4 and SI/SI stromal layers. AML1a-expressing (GFP+) colonies appeared larger than their GFP-control counterparts (Figure 6B), and LTC-IC frequency was significantly enriched (approximately 3-fold) in the GFP+ fraction (Figure 6C). Similar results were obtained in CAFC assays conducted on MS-5 stromal cells (unpublished data), and in this case cells could be retrieved from the culture and examined immunophenotypically. The results show an increase (approximately 9-fold) in the proportion of primitive cells (CD34+CD38−) within the GFP+ fraction (Figure 6D). In light of these observations, we assessed the functional potential of AML1a-transduced cells in a xenotransplant setting. CD34+CD38− CB cells were transduced with AML1a lentivirus and injected into irradiated NOD-SCID recipients, and human cell engraftment assessed at eight weeks. AML1a-expressing cells successfully engrafted, and more importantly, the proportion of CD34+CD38− cells was approximately 3-fold higher within the GFP+ fraction of the graft (Figure 6E). Consistent with this, functional assessment of primitive cell activity within the GFP+ and GFP− fractions of the graft using LTC-IC assays revealed a 3-fold increase in the GFP+ (i.e., AML1a-expressing) population (Figure 6F). Taken together, these results indicate that AML1a has the capacity to enhance primitive hematopoietic cell function in human cord blood both in vitro and in vivo.
Human blood cells from three pooled individual umbilical cord samples were sorted on the basis of the cell surface molecules indicated, and the levels of transcripts corresponding to different AML1 isoforms were quantitated in triplicate and normalized relative to G6PD.
(A) Total AML1 transcript levels are presented as relative levels of that observed in CD36+ cells.
(B) The fraction of AML1a relative to total AML1 transcripts was determined. In (A) and (B) Kruskal-Wallis test showed statistical significance of difference among the fractions (p < 0.01). Mann-Whitney tests were then used to estimate the difference of any given two fractions, with asterisks indicating the statistical significance (p < 0.05). N.S., not significant.
(C) Western blot analysis of AML-isoform expression using the anti-RHD antibody in CD34+ and CD34− fractions of human cord blood (4 × 105 cells each) from two individual healthy donors. Analysis of actin expression in the same samples provide a control for loading.
(A) Lentiviral construct used for transduction.
(B) Photomicrograph of typical LTC-IC colonies derived from AML1a-expressing (GFP+) and control (GFP−) human cord blood cells.
(C) Frequency of LTC-IC in AML1a-expressing (GFP+) and control (GFP−) human cord blood cells.
(D) Representative FACS analysis of AML1a-transduced (GFP+) and control (GFP−) cells recovered after 5 wk of culture on MS-5 stromal layers. Right plot shows graphical summary of all experiments performed.
(E) Representative FACS analysis of AML1a-expressing (GFP+) and control (GFP−) human cells engrafted in the BM of NOD-SCID mice at 8 wk (left plot). Right plot shows graphical summary of all experiments performed.
(F) Frequency of LTC-IC in human cells derived from primary NOD-SCID mice engrafted with AML1a-transduced cells.
In (C), (D), (E), and (F), Mann-Whitney test showed statistical significance of the difference (p < 0.05).
Here, we show that AML1a increases the competitive engraftment potential and in vitro proliferation capacity of primitive BM compartments, while AML1b exhibits an essentially opposing activity, promoting differentiation and abrogating engraftment. Expression of AML1a in human cord blood-derived hematopoietic cells gave results broadly in line with those obtained in mouse cells.
Although AML1/RUNX1 has been shown to be critically involved in the generation and maintenance of hematopoietic stem and progenitor cells as well as the differentiation of their progeny, the precise roles of its different isoforms in these processes is not as yet fully understood [44,45]. Most studies have focused on the role of Aml1 in the specification of HSC either in knockout mice or using the hematopoietic differentiation of embryonic stem cells as a model [46–48]. These studies indicate that expression of Aml1b or an Aml1b molecule that lack its carboxyl-terminal (VPRWY) repression domain is sufficient to rescue definitive hematopoiesis in Aml1-deficient cells or animals. In contrast AML1 moieties that lack the transactivation domain, and hence may be considered AML1a-like, are unable to rescue hematopoiesis. Elegant as these studies are they do not directly address the physiological role of AML1a in homeostatic adult hematopoiesis, its potential as a nonphysiological but therapeutic regulator of the adult stem and progenitor compartment, nor its pathological activity in leukaemia. Here we have approached these issues and provided evidence that the AML1a and AML1b isoforms have opposing effects on hematopoietic stem and progenitor cells. Most notably, in murine cells we have shown that expression of AML1b abrogates hematopoietic reconstitution in an experimental transplant setting, whereas expression of AML1a significantly enhances competitive engraftment potential in both primary and secondary recipients. In mice transplanted with AML1a-transduced cells, the frequency of primitive progenitors was significantly enlarged. CAFC activity was increased 14-fold relative to controls, and the frequency of LTC-IC in the lin− compartment increased 4-fold; since the lin− compartment was itself more than 3-fold increased, the overall increase in total LTC-IC was at least 12-fold in AML1a-transduced versus untransduced cells. Committed progenitor activity (as judged by CFC assays) in the AML1a-expressing component of the grafts was also increased, but to a lesser degree, and the colonies produced were larger, suggesting increased proliferation, but they exhibited no block in myeloid differentiation. Since AML1b abrogated engraftment, further comparison with AML1a was necessarily conducted in vitro. Here AML1b promoted granulocytic differentiation whereas AML1a enhanced self-renewal as judged by colony replating and gave rise to larger colonies, which were somewhat retarded, although not arrested, in granulocyte differentiation.
Cell-cycle analysis suggested that this growth advantage was the result of increased proliferation as opposed to decreased apoptosis. These in vitro effects of AML1a are broadly in agreement with results obtained by forced expression of AML1a in myeloid cell lines in which granulocytic differentiation was blocked, and this blockade was accompanied by uncompromised cell growth . Most remarkably, AML1a-transduced BM cells were able to expand in a simple culture with SCF and IL3, with immature (c-kit+lin−) cells proliferating vigorously yet retaining the ability to produce terminally differentiated myelomonocytic cells in vitro and giving rise to short-term multilineage engraftment in vivo. These extended in vitro cultures established by AML1a-transduced cells thus recapitulate many aspects of hematopoiesis, retaining multipotency in a manner reminiscent of human CD34+ cells programmed to express the AML1-ETO oncoprotein .
Both AML1a and AML1-ETO are thought to interfere with AML1b/c, and possibly also AML2 and AML3, in a dominant-negative fashion, either by directly competing for AML-binding sites on DNA or by sequestering subunit b of core binding factor (CBFb), an obligate subunit required for binding of AML proteins to DNA . Thus comparison of the results obtained with our AML1a mice to those reported for AML1-ETO and Aml1-targetted animals should be informative with respect to understanding physiological and pathological AML1 activities. Like AML1a animals, “haploinsufficient” Aml1+/− mice also exhibit increased CFU-GM and CFU-mix, but unlike AML1a mice, the size of colonies generated is unaffected . BM cells of the Aml1+/− mice exhibit increased competitiveness accompanied by a nearly doubled frequency of immature hematopoietic stem and progenitors as assessed in CAFC assays . The KSL fraction is increased in the BM cells of conditional Aml1−/− mice, and hematopoietic reconstitution ability in transplanted mice maintained; yet Aml1−/− hematopoietic cells are defective in competitive repopulation capacity . Mice reconstituted with AML1a-transduced cells did not show appreciably inhibited myeloid differentiation, in contrast to mice reconstituted with hematopoietic cells retrovirally transduced with AML1-ETO [51,52]. These findings imply that while some aspects associated with forced AML1a expression mimic reduced dosage, although not complete loss, of Aml1, AML1a confers unique effects on stem and progenitor cell properties that cannot be achieved by reducing the dosage of the Aml1 gene or by expression of the AML1-ETO oncoprotein. Of note here also are our preliminary results obtained with a mutant of AML1 (AML1R139G), which can sequester CBFb but cannot bind DNA . Like AML1a, AML1R139G-transduced cells exhibited delayed granulocytic differentiation in vitro. Unlike AML1a, AML1R139G-transduced cells did not produce long-term cultures in vitro and did not contribute to long-term hematopoiesis in vivo (ST, unpublished data). Thus, the ability of AML1a to bind DNA appears crucial for cell expansion, thereby further emphasising that AML1a is not a simple antagonist of AML1b/c or AML2/AML3, but rather exerts distinct effects on hematopoietic cells by virtue of its DNA-binding activity. Although we have in this study concentrated on the role of AML1a, the observation that AML1b can abrogate engraftment, presumably through enhancing granulocytic differentiation, is of interest and perhaps at first glance surprising. However it is now well documented that transcription factor effects are highly dose sensitive. Aml1 is itself a case in point , and GATA-2 [55,56], PU1 , and OCT4  provide additional examples. Recent modelling approaches  (S. Huang and TE, unpublished data) are starting to illuminate how dose-sensitive bi-stable switches can be created as a result of the transcriptional motifs (coherent and incoherent feedforward motifs, crossantagonistic motifs, and the like) adopted by transcription factors. While such dynamic models and the genetic regulatory networks on which they are based  do not as yet include Aml1, the fact that Aml1is a key regulator of granulocyte/monocyte-affiliated growth factor receptors as well as transcription factors provides a plausible framework for understanding how increasing its expression could lead to enhanced differentiation at the expense of self-renewal.
In light of the AML1a activities highlighted by our experiments, it is intriguing that AML1a transcripts are evident in selected samples of leukaemia . AML1 mutants devoid of the carboxyl-terminal portion of the molecule, and thereby structurally resembling AML1a, are also found in BM cells of some patients with myelodysplastic syndrome and myeloid leukaemia [27–29]. Collectively, AML1a or disease-associated AML1a-like AML1 molecules could be implicated in increased proliferation or competitiveness at key stages of hematopoietic differentiation. Further analysis of the mechanisms involved will be of therapeutic interest in the context of hematopathology and may inform attempts to expand transplantable hematopoietic stem/progenitor cells for cell therapy. Of particular interest as a stem cell source is cord blood. It is widely available through cord-blood banks, is applicable across human leukocyte antigen mismatches, and carries low risks of infectious disease transmission . Despite these merits, slow engraftment and limited cell numbers contained in a given umbilical cord increases the likelihood of serious infectious complications and limits widespread application to adult patients. Modest expansion of stem and progenitor cells would be expected to circumvent these drawbacks , and thus AML1a-mediated mechanisms may therefore prove useful in the facilitation of transplantation therapeutics. Our experiments on human cord blood cells support this view. We have shown at the RNA level that AML1a is predominantly expressed within the primitive CD34+CD133+ and CD34+CD33+ compartments of human cord blood. Although human HSCs have not been fully characterized, these compartments are thought to contain the bulk of HSCs and multipotent progenitors [63–66]. Thus it is plausible that expansion of hematopoietic cells could be normally affected, in part, by AML1a expression by a combination of antagonism of AML1b but also through AML1a-specific pathways (see above). It should be stressed, however, that in all these compartments, at least at steady state, AML1b transcripts are significantly more abundant than those of AML1a. However, analysis of relative protein abundance in CD34+ versus CD34− cells suggests that AML1a may comprise a significant fraction of total AML1 protein in the CD34+ compartment. Whatever the physiological role of AML1a, our forced expression experiments indicate that it has the capacity to enhance primitive hematopoietic cell activity in cord blood; forced expression of AML1a in primitive CD34+CD38− cord blood-derived cells resulted in increased LTC-IC activity, increased proportions of immunophenoptypically primitive cells in long-term MS-5 stromal cultures, and most importantly, enhanced primitive cell activity in vivo assessed in the NOD-SCID mouse model. Identification and exploitation of processes that normally regulate the relative balance of AML1a and AML1b could therefore provide one strategy to potentiate the engraftment activity of human hematopoietic stem and progenitor cells. Alternatively direct introduction of AML1a protein as a trans-activating transduction-fusion, along the lines proposed for TAT-HOXB4 , could be considered. In any event, our results identify AML1a as one of the few molecules  with the capacity for increasing the engraftment potential of hematopoietic stem and progenitor cells. In summary, our findings are of interest with respect to both the physiological and pathological roles of AML1 isoforms and have clear implications for hematopoietic cell transplantation.
Figure S1. BM Cells from 5-FU-Treated Mice Were Transduced with AML1a-Containing or Control Retrovirus and Transplanted into Lethally Irradiated Syngeneic Hosts
The following analyses were conducted 4 mo after the transplantation.
(A) FACS analysis of the primitive c-kit+lin− (B220, CD3, CD4, CD8, CD5, Gr-1, Mac-1, and TER-119) progenitor compartment. Results are shown from the GFP+ and GFP− fractions of a typical AML1a animal. GFP+ cells from a typical mouse receiving vector-only-transduced cells provide a control.
(B) Proportion of cells representative of different myeloid maturation stages and erythroid and lymphoid lineages in GFP+ BM cells from a typical vector-only or AML1a animal.
(C) Cytospins of GFP+ BM cells from a typical vector-only animal (a), GFP− (b), and GFP+ (c) cells from a typical AML1a animal.
(5.5 MB PPT)
Figure S2. BM Cells from 5-FU-Treated Mice Were Transduced with AML1a-Containing, AML1b-Containing, or Control Retrovirus and Cultured under Multimyeloid Differentiation Conditions
(A) Photomicrograph of typical colonies derived from control, AML1a-, or AML1b-transduced BM cells at the second round of re-plating.
(B) Photomicrographs of cells from 8-wk-old suspension cultures derived from AML1a-expressing colonies obtained at the third round of replating (see text for details). May-Grunwaldt-Giemsa staining of the adherent (upper image) and floating cells (lower image).
(1.0 MB PPT)
The GenBank (http://www.ncbi.nlm.nih.gov/) accession numbers of the genes discussed in this paper are AML1a (D43967), AML1b (D43968), AML1c (D43969), Aml1 (NM_009821), GATA-1 (X17254), SCL/TAL (NM_003189), TEL/ETV6 (NM_001987), and IKAROS/IKZF (U40462).
Supported by programme grants from the Leukaemia Research Fund and Medical Research Council of Great Britain (TE); Wella Award (2005) from Japan Leukemia Research Fund (MS); grant-in-aid for the second-term comprehensive ten-year Strategy for Cancer Control from the Ministry of Health and Welfare (MS); grant-in aid for scientific research from The Ministry of Education, Culture, Sports, Science and Technology (MS and ST); grant-in-aid for scientific research from the Japan Society for the Promotion of Science (ST); and research grants from the Japanese Leukemia Research Fund (ST) and from Takeda Science Foundation (ST).
ST, DH, MS, and TE designed the study. ST, DH, KM, and TE analysed the data. ST, RG, and TE contributed to writing the paper. ST conducted experiments using mouse cells, and analysed relative amount of AML1 isoforms in human cells. DH designed and carried out the AML1a functional experiments in human cells under the supervision of TE. RG performed fractionation of cord blood compartments and preparation of RNA/cDNA and cloning of lentivirus. KM contributed to the article as a biostatistic expert and conducted all of the statistical analyses. MS made an initial contribution of the plasmid constructs used in the paper with ST This work was initiated as a collaboration between the laboratories of MS and TE.
- 1. Miyoshi H, Shimizu K, Kozu T, Maseki N, Kaneko Y, et al. (1991) t(8;21) breakpoints on chromosome 21 in acute myeloid leukemia are clustered within a limited region of a single gene, AML1. Proc Natl Acad Sci U S A 88: 10431–10434.
- 2. Ito Y (2004) Oncogenic potential of the RUNX gene family: “Overview”. Oncogene 23: 4198–4208.
- 3. Speck NA, Gilliland DG (2002) Core-binding factors in haematopoiesis and leukaemia. Nat Rev Cancer 2: 502–513.
- 4. North TE, de Bruijn MF, Stacy T, Talebian L, Lind E, et al. (2002) Runx1 expression marks long-term repopulating hematopoietic stem cells in the midgestation mouse embryo. Immunity 16: 661–672.
- 5. Okuda T, van Deursen J, Hiebert SW, Grosveld G, Downing JR (1996) AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 84: 321–330.
- 6. Wang Q, Stacy T, Binder M, Marin-Padilla M, Sharpe AH, et al. (1996) Disruption of the Cbfa2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis. Proc Natl Acad Sci U S A 93: 3444–3449.
- 7. Ichikawa M, Asai T, Saito T, Seo S, Yamazaki I, et al. (2004) AML-1 is required for megakaryocytic maturation and lymphocytic differentiation, but not for maintenance of hematopoietic stem cells in adult hematopoiesis. Nat Med 10: 299–304.
- 8. Growney JD, Shigematsu H, Li Z, Lee BH, Adelsperger J, et al. (2005) Loss of Runx1 perturbs adult hematopoiesis and is associated with a myeloproliferative phenotype. Blood 106: 494–504.
- 9. Putz G, Rosner A, Nuesslein I, Schmitz N, Buchholz F (2006) AML1 deletion in adult mice causes splenomegaly and lymphomas. Oncogene 25: 929–939.
- 10. Friedman AD (1999) Leukemogenesis by CBF oncoproteins. Leukemia 13: 1932–1942.
- 11. Blyth K, Cameron ER, Neil JC (2005) The RUNX genes: Gain or loss of function in cancer. Nat Rev Cancer 5: 376–387.
- 12. Levanon D, Groner Y (2004) Structure and regulated expression of mammalian RUNX genes. Oncogene 23: 4211–4219.
- 13. Bae SC, Ogawa E, Maruyama M, Oka H, Satake M, et al. (1994) PEBP2 alpha B/mouse AML1 consists of multiple isoforms that possess differential transactivation potentials. Mol Cell Biol 14: 3242–3252.
- 14. Zhang YW, Bae SC, Huang G, Fu YX, Lu J, et al. (1997) A novel transcript encoding an N-terminally truncated AML1/PEBP2 alphaB protein interferes with transactivation and blocks granulocytic differentiation of 32Dcl3 myeloid cells. Mol Cell Biol 17: 4133–4145.
- 15. Fujita Y, Nishimura M, Taniwaki M, Abe T, Okuda T (2001) Identification of an alternatively spliced form of the mouse AML1/RUNX1 gene transcript AML1c and its expression in early hematopoietic development. Biochem Biophys Res Commun 281: 1248–1255.
- 16. Cantor AB, Orkin SH (2002) Transcriptional regulation of erythropoiesis: An affair involving multiple partners. Oncogene 21: 3368–3376.
- 17. Wechsler J, Greene M, McDevitt MA, Anastasi J, Karp JE, et al. (2002) Acquired mutations in GATA1 in the megakaryoblastic leukemia of Down syndrome. Nat Genet 32: 148–152.
- 18. Xu G, Nagano M, Kanezaki R, Toki T, Hayashi Y, et al. (2003) Frequent mutations in the GATA-1 gene in the transient myeloproliferative disorder of Down syndrome. Blood 102: 2960–2968.
- 19. Calkhoven CF, Muller C, Martin R, Krosl G, Pietsch H, et al. (2003) Translational control of SCL-isoform expression in hematopoietic lineage choice. Genes Dev 17: 959–964.
- 20. Sasaki K, Nakamura Y, Maki K, Waga K, Nakamura F, et al. (2004) Functional analysis of a dominant-negative DeltaETS TEL/ETV6 isoform. Biochem Biophys Res Commun 317: 1128–1137.
- 21. Tonnelle C, Bardin F, Maroc C, Imbert AM, Campa F, et al. (2001) Forced expression of the Ikaros 6 isoform in human placental blood CD34+ cells impairs their ability to differentiate toward the B-lymphoid lineage. Blood 98: 2673–2680.
- 22. Sun L, Goodman PA, Wood CM, Crotty ML, Sensel M, et al. (1999) Expression of aberrantly spliced oncogenic ikaros isoforms in childhood acute lymphoblastic leukemia. J Clin Oncol 17: 3753–3766.
- 23. Nakase K, Ishimaru F, Avitahl N, Dansako H, Matsuo K, et al. (2000) Dominant negative isoform of the Ikaros gene in patients with adult B-cell acute lymphoblastic leukemia. Cancer Res 60: 4062–4065.
- 24. Yagi T, Hibi S, Takanashi M, Kano G, Tabata Y, et al. (2002) High frequency of Ikaros isoform 6 expression in acute myelomonocytic and monocytic leukemias: Implications for up-regulation of the antiapoptotic protein Bcl-XL in leukemogenesis. Blood 99: 1350–1355.
- 25. Miyoshi H, Ohira M, Shimizu K, Mitani K, Hirai H, et al. (1995) Alternative splicing and genomic structure of the AML1 gene involved in acute myeloid leukemia. Nucleic Acids Res 23: 2762–2769.
- 26. Tanaka T, Tanaka K, Ogawa S, Kurokawa M, Mitani K, et al. (1995) An acute myeloid leukemia gene, AML1, regulates hemopoietic myeloid cell differentiation and transcriptional activation antagonistically by two alternative spliced forms. Embo J 14: 341–350.
- 27. Christiansen DH, Andersen MK, Pedersen-Bjergaard J (2004) Mutations of AML1 are common in therapy-related myelodysplasia following therapy with alkylating agents and are significantly associated with deletion or loss of chromosome arm 7q and with subsequent leukemic transformation. Blood 104: 1474–1481.
- 28. Harada H, Harada Y, Niimi H, Kyo T, Kimura A, et al. (2004) High incidence of somatic mutations in the AML1/RUNX1 gene in myelodysplastic syndrome and low blast percentage myeloid leukemia with myelodysplasia. Blood 103: 2316–2324.
- 29. Osato M (2004) Point mutations in the RUNX1/AML1 gene: Another actor in RUNX leukemia. Oncogene 23: 4284–4296.
- 30. Persons DA, Allay JA, Allay ER, Ashmun RA, Orlic D, et al. (1999) Enforced expression of the GATA-2 transcription factor blocks normal hematopoiesis. Blood 93: 488–499.
- 31. Reif K, Ekland EH, Ohl L, Nakano H, Lipp M, et al. (2002) Balanced responsiveness to chemoattractants from adjacent zones determines B-cell position. Nature 416: 94–99.
- 32. Tsuzuki S, Seto M, Greaves M, Enver T (2004) Modeling first-hit functions of the t(12;21) TEL-AML1 translocation in mice. Proc Natl Acad Sci U S A 101: 8443–8448.
- 33. Freyssinier JM, Lecoq-Lafon C, Amsellem S, Picard F, Ducrocq R, et al. (1999) Purification, amplification and characterization of a population of human erythroid progenitors. Br J Haematol 106: 912–922.
- 34. O'Connell J, Houston A, Kelly R, O'Brien D, Ryan A, et al. (2002) Rapid development of a quantitative-competitive (qc) RT-PCR assay using a composite primer approach. Methods Mol Biol 193: 93–102.
- 35. Kasugai Y, Tagawa H, Kameoka Y, Morishima Y, Nakamura S, et al. (2005) Identification of CCND3 and BYSL as candidate targets for the 6p21 amplification in diffuse large B-cell lymphoma. Clin Cancer Res 11: 8265–8272.
- 36. Demaison C, Parsley K, Brouns G, Scherr M, Battmer K, et al. (2002) High-level transduction and gene expression in hematopoietic repopulating cells using a human immunodeficiency [correction of imunodeficiency] virus type 1-based lentiviral vector containing an internal spleen focus forming virus promoter. Hum Gene Ther 13: 803–813.
- 37. Guenechea G, Gan OI, Inamitsu T, Dorrell C, Pereira DS, et al. (2000) Transduction of human CD34+ CD38− bone marrow and cord blood-derived SCID-repopulating cells with third-generation lentiviral vectors. Mol Ther 1: 566–573.
- 38. Ailles L, Schmidt M, Santoni de Sio FR, Glimm H, Cavalieri S, et al. (2002) Molecular evidence of lentiviral vector-mediated gene transfer into human self-renewing, multi-potent, long-term NOD/SCID repopulating hematopoietic cells. Mol Ther 6: 615–626.
- 39. Piacibello W, Sanavio F, Severino A, Dane A, Gammaitoni L, et al. (1999) Engraftment in nonobese diabetic severe combined immunodeficient mice of human CD34(+) cord blood cells after ex vivo expansion: Evidence for the amplification and self-renewal of repopulating stem cells. Blood 93: 3736–3749.
- 40. Cashman JD, Lapidot T, Wang JC, Doedens M, Shultz LD, et al. (1997) Kinetic evidence of the regeneration of multilineage hematopoiesis from primitive cells in normal human bone marrow transplanted into immunodeficient mice. Blood 89: 4307–4316.
- 41. Itoh K, Tezuka H, Sakoda H, Konno M, Nagata K, et al. (1989) Reproducible establishment of hemopoietic supportive stromal cell lines from murine bone marrow. Exp Hematol 17: 145–153.
- 42. Nishihara M, Wada Y, Ogami K, Ebihara Y, Ishii T, et al. (1998) A combination of stem cell factor and granulocyte colony-stimulating factor enhances the growth of human progenitor B cells supported by murine stromal cell line MS-5. Eur J Immunol 28: 855–864.
- 43. Ohkawara JI, Ikebuchi K, Fujihara M, Sato N, Hirayama F, et al. (1998) Culture system for extensive production of CD19+IgM+ cells by human cord blood CD34+ progenitors. Leukemia 12: 764–771.
- 44. Cai Z, de Bruijn M, Ma X, Dortland B, Luteijn T, et al. (2000) Haploinsufficiency of AML1 affects the temporal and spatial generation of hematopoietic stem cells in the mouse embryo. Immunity 13: 423–431.
- 45. de Bruijn MF, Speck NA (2004) Core-binding factors in hematopoiesis and immune function. Oncogene 23: 4238–4248.
- 46. Nishimura M, Fukushima-Nakase Y, Fujita Y, Nakao M, Toda S, et al. (2004) VWRPY motif-dependent and -independent roles of AML1/Runx1 transcription factor in murine hematopoietic development. Blood 103: 562–570.
- 47. North T, Gu TL, Stacy T, Wang Q, Howard L, et al. (1999) Cbfa2 is required for the formation of intra-aortic hematopoietic clusters. Development 126: 2563–2575.
- 48. Okuda T, Takeda K, Fujita Y, Nishimura M, Yagyu S, et al. (2000) Biological characteristics of the leukemia-associated transcriptional factor AML1 disclosed by hematopoietic rescue of AML1-deficient embryonic stem cells by using a knock-in strategy. Mol Cell Biol 20: 319–328.
- 49. Mulloy JC, Cammenga J, Berguido FJ, Wu K, Zhou P, et al. (2003) Maintaining the self-renewal and differentiation potential of human CD34+ hematopoietic cells using a single genetic element. Blood 102: 4369–4376.
- 50. Sun W, Downing JR (2004) Haploinsufficiency of AML1 results in a decrease in the number of LTR-HSCs while simultaneously inducing an increase in more mature progenitors. Blood 104: 3565–3572.
- 51. de Guzman CG, Warren AJ, Zhang Z, Gartland L, Erickson P, et al. (2002) Hematopoietic stem cell expansion and distinct myeloid developmental abnormalities in a murine model of the AML1-ETO translocation. Mol Cell Biol 22: 5506–5517.
- 52. Schwieger M, Lohler J, Friel J, Scheller M, Horak I, et al. (2002) AML1-ETO inhibits maturation of multiple lymphohematopoietic lineages and induces myeloblast transformation in synergy with ICSBP deficiency. J Exp Med 196: 1227–1240.
- 53. Imai Y, Kurokawa M, Izutsu K, Hangaishi A, Takeuchi K, et al. (2000) Mutations of the AML1 gene in myelodysplastic syndrome and their functional implications in leukemogenesis. Blood 96: 3154–3160.
- 54. Song WJ, Sullivan MG, Legare RD, Hutchings S, Tan X, et al. (1999) Haploinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia. Nat Genet 23: 166–175.
- 55. Ling KW, Ottersbach K, van Hamburg JP, Oziemlak A, Tsai FY, et al. (2004) GATA-2 plays two functionally distinct roles during the ontogeny of hematopoietic stem cells. J Exp Med 200: 871–882.
- 56. Rodrigues NP, Janzen V, Forkert R, Dombkowski DM, Boyd AS, et al. (2005) Haploinsufficiency of GATA-2 perturbs adult hematopoietic stem-cell homeostasis. Blood 106: 477–484.
- 57. Stirewalt DL (2004) Fine-tuning PU.1. Nat Genet 36: 550–551.
- 58. Niwa H, Miyazaki J, Smith AG (2000) Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet 24: 372–376.
- 59. Chickarmane V, Troein C, Nuber HM, Sauro C, Peterson C (2006) Transcriptional dynamics of the embryonic stem cell switch. PLOS Computational Biology 2: 1080–1092.
- 60. Swiers G, Patient R, Loose M (2006) Genetic regulatory networks programming hematopoietic stem cells and erythroid lineage specification. Dev Biol 294: 525–540.
- 61. Brunstein CG, Wagner JE (2006) Umbilical cord blood transplantation and banking. Annu Rev Med 57: 403–417.
- 62. Sorrentino BP (2004) Clinical strategies for expansion of haematopoietic stem cells. Nat Rev Immunol 4: 878–888.
- 63. Yin AH, Miraglia S, Zanjani ED, Almeida-Porada G, Ogawa M, et al. (1997) AC133, a novel marker for human hematopoietic stem and progenitor cells. Blood 90: 5002–5012.
- 64. de Wynter EA, Buck D, Hart C, Heywood R, Coutinho LH, et al. (1998) CD34+AC133+ cells isolated from cord blood are highly enriched in long-term culture-initiating cells, NOD/SCID-repopulating cells and dendritic cell progenitors. Stem Cells 16: 387–396.
- 65. Jin CH, Takada H, Nomura A, Takahata Y, Nakayama H, et al. (2000) Immunophenotypic and functional characterization of CD33(+)CD34(+) cells in human cord blood of preterm neonates. Exp Hematol 28: 1174–1180.
- 66. Taussig DC, Pearce DJ, Simpson C, Rohatiner AZ, Lister TA, et al. (2005) Hematopoietic stem cells express multiple myeloid markers: Implications for the origin and targeted therapy of acute myeloid leukemia. Blood 106: 4086–4092.