The stem cell leukemia (SCL or tal-1) gene was initially identified as a translocation partner in a leukemia that possessed both lymphoid and myeloid differentiation potential. Mice that lacked SCL expression showed a complete block in hematopoiesis; thus, SCL was associated with hematopoietic stem cell (HSC) function. More recent studies show a role for SCL in murine erythroid differentiation. However, the expression pattern and the role of SCL during early stages of human hematopoietic differentiation are less clear. In this study we chart the pattern of human SCL expression from HSCs, through developmentally sequential populations of lymphoid and myeloid progenitors to mature cells of the hematopoietic lineages. Using recently defined surface immunophenotypes, we fluorescence-activated cell–sorted (FACS) highly purified populations of primary human hematopoietic progenitors for reverse transcription–polymerase chain reaction (RT-PCR) analysis of SCL expression. Our data show that SCL mRNA is easily detectable in all hematopoietic populations with erythroid potential, including HSCs, multipotential progenitors, common myeloid progenitors, megakaryocyte/erythrocyte progenitors, and nucleated erythroid lineage cells. SCL mRNA expression was present but rapidly downregulated in the common lymphoid progenitor and granulocyte/monocyte progenitor populations that lack erythroid potential. SCL expression was undetectable in immature cells of nonerythroid lineages, including pro-B cells, early thymic progenitors, and myeloid precursors expressing the M-CSF receptor. SCL expression was also absent from all mature cells of the nonerythroid lineages. Although low levels of SCL were detected in lymphoid- and myeloid-restricted progenitors, our studies show that abundant SCL expression is normally tightly linked with erythroid differentiation potential.
The stem cell leukemia (SCL or tal-1) gene was initially identified as a translocation partner in a leukemia that possessed both lymphoid and myeloid differentiation potential . In subsequent studies, aberrant expression of SCL, arising from translocations and from site-specific recombinase-mediated deletions in the SCL gene, was associated with T-cell acute lymphocytic leukemia (T-ALL) [2–6].
The SCL gene encodes a basic helix-loop-helix transcription factor  that heterodimerizes with the E2A gene products [7, 8] and interacts with a subunit of the basal transcription factor TFIIH . In addition to the E2A gene products, SCL has been shown to interact with GATA-1, LM0 (LIM-only proteins), Lbd1, Sp-1, and others to form larger protein complexes [10, 11] that can either repress or enhance expression of target genes.
Murine models that examined hematopoiesis in the absence of SCL expression suggested that SCL was important in hematopoietic stem cell (HSC) function. Initial studies of SCL knockout mice showed a failure in primitive yolk sac hematopoiesis that resulted in embryonic lethality [12,13]. Similarly, studies of chimeric animals generated by the injection of SCL−/− embryonic stem (ES) cells into normal blastocysts showed that definitive adult hematopoiesis could only arise from cells that expressed SCL [14, 15]. The recent development of mouse models in which SCL expression can be modulated has aided in precisely identifying the murine hematopoietic activities dependent on the expression of SCL. Studies of SCL-conditional mutants showed that the production but not self-renewal or engraftment of long-term repopulating HSCs (LT-HSCs) is dependent on SCL . These mice also showed that SCL is required for the production of erythrocytes and megakaryocytes but not for lymphoid or myeloid differentiation [16–18]. However, the abrogation of SCL resulted in reduced short-term repopulating HSC (ST-HSC) function  and skewed multilineage repopulation . Studies of murine hematopoietic repopulation from HSCs in which levels of SCL were manipulated via expression of a dominant-negative form of SCL or by overexpression of wild-type SCL provided evidence that SCL may play a role in myeloid versus lymphoid differentiation .
Studies of human cell lines in which expression of either SCL or SCL antisense was enforced suggest that SCL may play several roles in human hematopoiesis. These roles include providing protection from apoptosis, enhancing cell-cycle progression and long-term culture-initiating cell (LTC-IC) activity, inhibiting myeloid differentiation, and promoting erythroid differentiation [20–25]. Evidence from the few studies of SCL function in primary human hematopoietic cells involved in vitro differentiation assays and suggests that at least some of these are legitimate roles for SCL in normal human hematopoiesis. Enforced SCL expression in primary human progenitors increases erythroid and megakaryocyte colony formation, selectively increases the size of erythroid colonies, and, in some cases, reduces granulocyte colony formation [26, 27]. Conversely, when SCL expression is decreased by treating primary human progenitors with SCL antisense, erythroid colony formation is significantly decreased, although no effect on the formation of granulocyte colonies is observed . Little is known of SCL expression during normal human hematopoietic differentiation.
Given that SCL was originally identified in a leukemia with both lymphoid and myeloid differentiation potential  and that murine studies suggest a role for SCL in lymphoid and myeloid lineage commitment [16, 18], we were interested in whether SCL might play some role in early stages of normal human lymphoid and myeloid differentiation. The classic model of hematopoiesis proposes that pluripotent HSCs in umbilical cord blood (CB) (CD34+ CD38−CD7−) and bone marrow (BM) (CD34+ CD38−) give rise to progenitors that have committed to either a lymphoid cell fate (common lymphoid progenitor [CLP]) or myeloid/erythroid cell fate (common myeloid progenitor [CMP]). CLPs are believed to ultimately give rise to single-lineage progenitors committed to becoming T, B, or natural killer (NK) cells. CMPs are postulated to generate two populations of bipotential progenitors, the granulocyte/monocyte progenitor (GMP) and the megakaryocyte/erythroid progenitor (MEP).
Recent studies have identified, based on surface immunophenotype, several of the lymphoid, myeloid, and erythroid differentiation intermediates predicted by the classic model of hematopoietic differentiation. Fluorescence-activated cell sorting (FACS) of Lin−CD34+CD38+ CB and BM progenitors based on coexpression of CD45RA and the interleukin-3 receptor alpha (IL-3Rα) allows the isolation of functional CMP (IL-3Rα+CD45RA−), GMP (IL-3Rα+CD45RA+), and MEP (IL-3Rα−CD45RA−) populations. The population with CLP function within human BM has been defined as Lin−CD34+CD10+. More recently, studies in our laboratory have shown that the human CB population that functions as a CLP is CD34+CD38−CD7+ .
Determining SCL expression at precise points in normal hematopoietic differentiation is key to clarifying the roles of SCL in this process and may provide clues to the mechanisms through which the SCL gene product acts. SCL expression has not been examined in precisely defined populations of lymphoid, myeloid, or erythroid progenitors isolated from primary human hematopoietic sources. Our studies follow SCL expression from the pluripotent HSCs to the more restricted CLP, CMP, GMP, and MEP and then through early lineage-committed progenitors to mature T, B, NK, myeloid, and erythroid lineage cells.
Materials and Methods
Cell Sources and Preparation
Human umbilical CB, BM, peripheral blood (PB), and postnatal thymus from pediatric patients undergoing cardiac surgery were obtained under protocols approved by the Childrens Hospital Los Angeles Committee on Clinical Investigation (Investigational Review Board). Thymic portions were cut into thin slices and washed with phosphate-buffered saline (PBS) to free cells from connective tissue. Mononuclear cells were isolated from CB, BM, PB, and thymus cell suspensions as previously described .
All hematopoietic populations were isolated by FACS using the FACSVantage (Becton, Dickinson Immunocytometry Systems, San Jose, CA, http://www.bdbiosciences.com/immunocytometry_systems/). In most cases (see Results), sorted populations were resorted to ensure purity. HSCs and progenitor populations from respective tissues were sorted by FACS from mononuclear cells that had been enriched for CD34+ cells using the MiniMacs CD34 Progenitor Isolation Kit (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com/) according to manufacturer's instructions. CD4+CD8+ (double-positive) and CD4+CD8− and CD4− CD8+ (single-positive) T-cell precursors were sorted by FACS from total thymus. Nucleated erythroid lineage cells, early monocytic cells (M-CSFR+ cells), and mature T, B, and NK cells were isolated from total mononuclear cells or from the CD34− cells rejected by the MiniMacs CD34 Progenitor Isolation Kit. Granulocytes were isolated (within 6 hours of delivery) from fresh umbilical CB cells that were located below the buffer layer after Ficoll-Paque Plus (Amersham Biosciences, Uppsala, Sweden, http://www.amersham.com) separation.
Antibody Staining for FACS
Surface immunophenotypes described in the Introduction and Results sections were used to isolate hematopoietic populations by FACS. Suspensions of mononuclear cells from the respective tissues were incubated with antibodies for 15 minutes on ice, washed with PBS, resuspended in PBS, and then FACS-sorted. Antibodies directed against the following surface markers and directly labeled with indicated fluorophore were used to stain before FACS: from Becton, Dickinson (BD) Biosciences/Pharmingen, San Diego (http://www.bdbiosciences.com/pharmingen/): CD1A (fluorescein isothiocyanate [FITC]), CD2 (FITC or phycoerythrin [PE]), CD3 (FITC or allophycocyanin [APC]), CD4 (FITC or APC), CD8 (FITC or PE), CD10 (FITC, PE, or APC), CD11B (PE or APC), CD11C (PE), CD14 (FITC, PE, or APC), CD15 (FITC), CD19 (FITC or APC), CD20 (PE), CD34 (FITC, APC, or PerCP-Cy5.5), CD38 (APC or PE-Cy7), CD56 (FITC, PE, or APC), CD57 (FITC), CD66B (FITC), IL-3Rα (PE clone 9F5), glycophorin A (APC); from Coulter-Immunotech (Westbrook, ME, http://www.beckman.com): CD7 (PE), CD34 energy-coupled dye (ECD), CD45RA (FITC), glycophorin A (PE or FITC), CD71 (FITC); from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, http://www.scbt.com): M-CSF receptor (PE). Antibodies to the following markers were used to sort indicated lineage-negative (Lin−) populations: for Lin−CD34+CD38− BM HSCs, antibodies against CD3, CD14, CD19, CD56, glycophorin A, and, in some cases, CD10, CD15, and CD57 were used; for sorting Lin−CD34+CD38−CD7+ CB CLPs and Lin−CD34+CD38−CD7− CB HSCS, antibodies against CD1a, CD3, CD8, CD11C, CD14, CD15, CD56, CD57, and glycophorin A were used; for sorting Lin−CD34+CD7+ thymic progenitors, antibodies against CD3, CD14, CD19, CD56, CD57, and glycophorin A were used; for sorting Lin− CMP, GMP, and MEP populations in CB, antibodies against CD2, CD3, CD4, CD7, CD8, CD10, CD11B, CD14, CD19, CD20, CD56, and glycophorin A were used.
Reverse Transcription–Polymerase Chain Reaction Analysis of RNA Expression
Cell populations isolated by FACS were deposited into 5% fetal calf serum and PBS, and RNA extraction was performed using RNA Stat-60 (Tel-test, Friendswood, TX, http://www.isotexdiagnostics.com). Complementary DNA (cDNA) was generated using an oligo dT primer and RNAguard (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) with the Omniscript RT kit (Qiagen, Valencia, CA, http://www.qiagen.com) per the manufacturer's instructions. cDNA from equivalent cell numbers was subjected to polymerase chain reaction (PCR) using HotStarTaq (Qiagen) per the manufacturer's directions in a DNA Thermal Cycler 480 (PerkinElmer Cetus, Norwalk, CT, http://www.perkinelmer.com). To ensure that PCR products did not arise from contaminating genomic DNA (gDNA), all primers were selected to span an intron.
To provide a positive control for PCR and to assess the quality of the cDNA preparation, in every case PCR analyses to detect transcripts for the housekeeping gene β2-microglobulin were performed alongside those to detect SCL. The following primers were used to generate a 330-base β2-microglobulin product (all from Invitrogen Life Technologies, Paisley, PA, http://www.invitrogen.com): 5′-CTC GCG CTA CTC TCT CTT TC-3′ and 5′-CAT GTC TCG ATC CCA CTT AAC-3′. β2-Microglobulin PCR conditions were as follows: 94°C for 1 minute, 58°C for 1 minute, 72°C for 2 minutes (32 cycles), and 72°C for 10 minutes (1 cycle).
Human SCL expression was assessed using the following primers that give rise to a 100-base product: 5′-TTC CCT ATG TTC ACC ACC AA-3′ and 5′-AAG ATA CGC CGC ACA ACT TT-3′. Touchdown PCR with higher annealing temperatures early on (giving high specificity and low yield) and lower annealing temperatures later (giving high yield once specific product has been generated) was used to assess SCL expression. Touchdown PCR conditions were as follows: 94°C for 1 minute, 62°C for 1 minute, 72°C for 2 minutes (1 cycle), 94°C for 1 minute, 61°C for 1 minute, 72°C for 2 minutes (1 cycle), 94°C for 1 minute, 60°C for 1 minute, 72°C for 2 minutes (1 cycle), 94°C for 1 minute, 58°C for 1 minute, 72°C for 2 minutes (1 cycle), 94°C for 1 minute, 56°C for 1 minute, 72°C for 2 minutes (1 cycle), 94°C for 1 minute, 64°C for 1 minute, 72°C for 2 minutes (30 cycles), and 72°C for 10 minutes (1 cycle).
PCR products were run on a 1% Nusieve GTG agarose plus 1% SeaKem LE agarose gel (BioWhittaker Molecular Applications, Rockland, ME, http://www.bmaproducts.com) containing ethidium bromide. Gels were photographed using Eagle Eye II (Stratagene, La Jolla, CA, http://www.stratagene.com).
SCL mRNA Is Expressed in CD34+CD38− HSCs from Human BM
Given that SCL is required for the production of murine HSCs, we first examined the expression of human SCL in HSCs and in progenitor populations isolated from normal BM. CD34+CD38− HSCs and CD34+CD38+ progenitors were FACS-sorted using sort gates shown in Figure 1A. SCL mRNA was detectable by reverse transcription (RT)–PCR analysis in CD34+CD38− HSCs isolated from human BM (Fig. 1B). SCL mRNA was also detected in the CD34+CD38+ BM progenitor population (Fig. 1B), which includes a mixture of lymphoid, myeloid, and erythroid progenitors. To determine if SCL expression in this mixed progenitor population was due entirely to erythroid-committed cells, we also examined SCL mRNA expression in FACS-sorted CD34+CD38+ progenitors that lacked expression of the erythroid lineage marker glycophorin A (CD34+CD38+GlyA−). SCL message was detectable in CD34+CD38−GlyA− cells, although at lower levels than in the CD34+C38+ population (Fig. 1B).
We then took several steps to ensure that the SCL mRNA detected in the BM HSC and progenitor populations was not due to lineage-committed cells or to low purity of the FACS sort in general. First, CD34+CD38− BM progenitors were isolated based on the absence of hematopoietic lineage markers (Lin−CD34+CD38−, gates shown in Figs. 1A and 1C: R1, R2, and R4) and then resorted a second time to ensure purity. RT-PCR analysis of 1,500 double-sorted cells at 35, 38, and 40 cycles (Fig. 1F) shows that SCL transcripts were present in highly purified Lin−CD34+CD38− HSCs isolated from human BM. Next, to preclude the possibility that the SCL expression we detected was due to contamination with very early erythroid-committed cells, we sorted HSCs and progenitors that were negative for both the erythroid lineage marker glycophorin A and the transferrin receptor CD71. (CD71 is thought to be expressed on erythroid-committed cells before glycophorin A .) We isolated CD34+CD38−GlyA−CD71− (gates shown in Figs. 1A and 1D: R1, R2, and R5) and CD34+ CD38+ GlyA−CD71− (gates shown in Figs. 1A and 1E: R1, R3, and R5) cells. SCL transcripts were detectable by RT-PCR in both populations (Fig. 1G), although depletion of cells expressing erythroid markers markedly reduced the expression level in CD34+CD38+ populations.
Expression of SCL During Lymphoid Differentiation
SCL was initially identified in a leukemia with both lymphoid and myeloid potential  and later associated with T-ALL [2–6]. Recent murine studies of SCL conditional mutants suggest that SCL may promote B versus T lymphopoiesis . Given these links between SCL and lymphopoiesis, we next determined whether SCL expression is extinguished during normal lymphoid differentiation. To address this question, we examined SCL expression in developmentally sequential populations of lymphoid differentiation intermediates isolated from human umbilical CB and thymus.
From human CB, HSCs (CD34+CD38−Lin−CD7−) and CLPs (CD34+CD38−Lin−CD7+) were sorted as shown in Figures 2A and 2B. RT-PCR showed that in CB, as in BM, HSCs express SCL message (Fig. 2D, left panel). SCL message was also detectable in CLPs but downregulated relative to HSCs (Fig. 2D, left panel). Using sort gates shown in Figure 2C, CD34+CD19+ proB cells and more mature CD34−CD19+ B lineage cells were also FACS-sorted from human CB. SCL transcripts were not detected in CD34+CD19+ pro-B cells or CD34−CD19+ B lineage cells isolated from CB (Fig. 2D, right panel). The complete absence of detectable SCL transcripts in CD19+ B lineage cells provides evidence that the SCL detected in the CLP population was not the result of contamination from more common B cell progenitors. Thus, SCL mRNA is expressed in pluripotent HSCs, down-regulated in the CLP population after lymphoid commitment, and completely lost by the time lymphoid progenitors acquire CD19 and commit to B lymphopoiesis.
To determine SCL expression during T-lymphocyte differentiation, we isolated T-cell precursor populations from human thymus. Very early thymic T-cell progenitors were obtained by FACS-sorting Lin−CD34+CD7+ from human thymus that had been enriched for CD34+ cells by magnetic separation (sort gates shown in Figs. 3A and 3B). We also isolated the classic CD34− T-cell precursor populations, double-positive (CD4+CD8+), CD4 single-positive (CD4+CD8−), and CD8 single-positive (CD4−CD8+), from total thymus using sort gates as shown in Figure 3C. All populations were isolated based on the absence of glycophorin A (gate not shown) and double-sorted to ensure purity. RT-PCR analysis showed that SCL expression is absent in all thymic T-cell precursor populations, including the very early Lin−CD34+CD7+ thymic progenitors (Fig. 3D).
Expression of SCL During Myeloid and Erythroid Differentiation
SCL has been postulated to play a role in lineage commitment [18, 19]. Therefore, we examined SCL expression at critical branch points in myeloerythroid differentiation. CMP, GMP, and MEP populations were isolated from human CD34+-enriched CB using recently defined surface immunophenotypes  (sort gates shown in Figs. 4A and 4B). SCL transcripts were easily detectable in the CMP and MEP populations, both of which possess erythroid differentiation potential (Fig. 4E, left panel). SCL mRNA was detectable but markedly downregulated in the GMP (Fig. 4E, left panel), a population that is restricted to nonerythroid lineages.
Next we examined myeloid and erythroid progenitors downstream of the erythroid-myeloid branch point. To determine if SCL is expressed after commitment to monocyte differentiation, we FACS-sorted myeloid precursors from CB based on expression of the M-CSFR as shown in Figure 4C. To isolate early erythroid-committed precursors, we FACS-sorted nucleated erythroid lineage cells that were glycophorin A+CD34+, along with more mature glycophorin A+ CD34−cells (Fig.4D). SCL transcripts were completely undetectable in M-CSF-R+ CB cells after 40 cycles of PCR (Fig. 4E, right panel). In contrast, SCL transcripts were readily detectable in glycophorin A+ CD34+ and glycophorin A+ CD34− cells (Fig. 4E, right panel).
SCL Expression in Mature Hematopoietic Cells
To determine SCL expression in mature cells of the various hematopoietic lineages, we isolated cells based on expression of lineage-specific surface markers or combinations of markers as shown in Figure 5. RT-PCR analysis showed that SCL transcripts continued to be expressed in nucleated glycophorin A+ CD34− erythroid lineage cells isolated from BM (Fig. 5), as well as CB (Fig. 4E). SCL was not detectable in mature B cells, NK cells, CD4+ T cells, CD66B+ granulocytes, or CD14+CD11B+ monocytes (Fig. 5).
Surprisingly, few reports have described SCL expression during normal human hematopoietic differentiation. In this study we provide the first comprehensive analysis of SCL mRNA expression in highly purified, FACS-sorted hematopoietic progenitors at critical points in lymphoid, myeloid, and erythroid lineage commitment and differentiation (Fig. 6).
The report of SCL expression in mature hematopoietic cells and in human BM progenitors is limited to one early study . SCL message was detected in erythroid lineage cells and mega-karyocytes but absent from mature lymphoid and myeloid cells in normal PB. The study also detected SCL in basophilic granulocytes obtained from a patient with chronic myelogenous leukemia. In human BM, SCL was reported to be absent from CD34+CD38− HSCs, although it was detected in the CD34+CD38+ population, which includes a mixture of lymphoid, myeloid, and erythroid progenitors . In contrast to the previously published report, we detected SCL mRNA in CD34+CD38− HSCs isolated from BM and from CB (Fig. 1). It is likely that the range of PCR cycles used in our assay provided a much more sensitive assessment of SCL than that previously reported. Our results are unlikely to be due to contaminating erythroid cells, because HSC populations were resorted to ensure purity and other populations that were similarly sorted were negative for SCL expression, even after 40 cycles of PCR (Fig. 2).
Although we did not examine basophilic granulocytes, we did not detect SCL mRNA in normal CB granulocytes (Fig. 5) that express CD66B, a neutrophil marker. Consistent with the previous report, our studies showed that CD34+CD38+ BM cells, a heterogeneous population of progenitors, express SCL (Fig. 1).
Three studies have examined changes in SCL expression in unilineage differentiation cultures initiated with primary human hematopoietic progenitors. In these in vitro models, SCL was upregulated in erythroid and megakaryocyte differentiation but downregulated in granulocyte and monocyte differentiation [28, 26, 34]. Consistent with these reports, we observed that SCL expression was abundant in the CMP and MEP (populations with both erythroid and megakaryocyte potential) but downregulated in the GMP population that gives rise to granulocytes and monocytes.
One recent study provided evidence that increased SCL expression might promote myeloid differentiation at the CLP/ CMP branch point . Although we did not directly compare relative levels of SCL expression between CLP and CMP populations, SCL expression in CMPs was readily detectable at a level similar to that observed for the erythroid-committed MEP population (Fig. 4E). In contrast, although SCL transcripts were present within the CLP population, their expression was at the threshold of detection (Fig. 2D).
Two studies have examined recovery of lymphoid and myeloid lineages after secondary transplant of progenitors from SCL conditional mutants. One of these studies showed that lymphoid reconstitution is skewed toward T lymphopoiesis at the expense of B lineage repopulation . The other showed a defect in repopulation of lymphoid and myeloid lineages . Thus, it has been suggested that in the mouse, SCL functions in lymphoid and myeloid lineage commitment and that SCL plays a role in self-renewal of progenitors that provide short-term reconstitution of multiple hematopoietic lineages . If SCL plays a role in T-lymphocyte versus B-lymphocyte differentiation, it is likely to be acting at the CLP stage. Consistent with this, we observed that SCL expression was detectable, although downregulated compared with HSCs, in the human CLP populations. Similarly, we observed detectable, but downregulated, SCL expression in human GMPs.
Enforced SCL expression in human CD34+ progenitors has recently been shown to increase in vitro LTC-IC activity . The heterogeneous population of CD34+CD38+ progenitors in which we and others  have detected SCL expression are likely to include cells responsible for at least a portion of LTC-IC activity and capable of short-term hematopoietic repopulation in humans.
Studies of human erythropoiesis show that SCL increases human erythroid colony formation [26–28] and participates in a multifactorial complex that functions in expression of glycophorin A, a marker of human erythroid lineage cells . These studies provide strong evidence of a role for SCL in human erythroid differentiation. Our studies add weight to this evidence, showing that SCL is abundantly expressed only in a series of developmentally sequentially hematopoietic populations with erythroid potential, from the CD34+CD38− HSCs to the CD34− glycophorin A+ nucleated erythroid lineage cells.
Conclusion and Summary
SCL mRNA is abundant in all hematopoietic populations with erythroid potential, including HSCs, multipotential progenitors, CMPs, MEPs, and nucleated erythroid lineage cells. SCL mRNA was rapidly downregulated in the CLP and GMP populations that lacked erythroid potential. SCL was not expressed in immature cells of nonerythroid lineages, including pro-B cells, early thymic progenitors, and myeloid precursors expressing the M-CSF receptor. SCL expression was also absent from the mature cells of the nonerythroid lineages. Although low levels of SCL message were detected in lymphoid- and myeloid-restricted progenitors, our studies show that abundant SCL expression is normally tightly linked with erythroid differentiation potential (Fig. 6).
Y.Z. and K.J.P contributed equally to this study. This work was supported by the following grants: RO1HL77912, PO1 CA59318, P50HL54850 (to G.M.C.), and K01 DK066163 (to K.J.P.). G.M.C. is a Scholar of the Leukemia and Lymphoma Society. K.J.P is the recipient of a Research Career Development Award from the Saban Research Institute at Childrens Hospital Los Angeles.