Search

Precursor B-cell acute lymphoblastic leukemia (B-ALL) is associated with recurrent mutations that occur in cancer-initiating cells. There is a need to understand how driver mutations influence clonal evolution of leukemia. The E26-transformation-specific (ETS) transcription factors PU.1 and Spi-B (encoded by Spi1 and Spib) execute a critical role in B-cell development and serve as complementary tumor suppressors. Here, we used a mouse model to conditionally delete Spi1 and Spib genes in developing B cells. These mice developed B-ALL with a median time to euthanasia of 18 weeks. We performed RNA and whole-exome sequencing (WES) on leukemias isolated from Mb1-Cre?PB mice and identified single nucleotide variants (SNVs) in Jak1, Jak3, and Ikzf3 genes, resulting in amino acid sequence changes. Jak3 mutations resulted in amino acid substitutions located in the pseudo-kinase (R653H, V670A) and in the kinase (T844M) domains. Introduction of Jak3 T844M into Spi1/Spib-deficient precursor B cells was sufficient to promote proliferation in response to low IL-7 concentrations in culture, and to promote proliferation and leukemia-like disease in transplanted mice. We conclude that mutations in Janus kinases represent secondary drivers of leukemogenesis that cooperate with Spi1/Spib deletion. This mouse model represents a useful tool to study clonal evolution in B-ALL.

The E26 transformation-specific (Ets) transcription factor PU.1 is required to generate lymphoid progenitor cells from hematopoietic stem cells, but it is not required to generate B cells from committed B-cell lineage progenitors. We hypothesized that PU.1 function in B-cell differentiation is complemented by the related Ets transcription factor Spi-B. To test this hypothesis, mice were generated lacking both PU.1 and Spi-B in the B-cell lineage. Unlike mice lacking PU.1 or Spi-B, mice deficient in both PU.1 and Spi-B in the B-cell lineage had reduced frequencies of B cells as well as impaired B-cell differentiation. Strikingly, all PU.1 and Spi-B–deficient mice developed pre-B cell acute lymphoblastic leukemia before 30 weeks of age. Pre-B cells accumulated in the thymus resulting in massive thymic enlargement and dyspnea. These findings demonstrate that PU.1 and Spi-B are essential transcriptional regulators of B-cell differentiation as well as novel tumor suppressors in the B-cell lineage.

The E26 transformation-specific (Ets) transcription factor PU.1 is required to generate lymphoid progenitor cells from hematopoietic stem cells, but it is not required to generate B cells from committed B-cell lineage progenitors. We hypothesized that PU.1 function in B-cell differentiation is complemented by the related Ets transcription factor Spi-B. To test this hypothesis, mice were generated lacking both PU.1 and Spi-B in the B-cell lineage. Unlike mice lacking PU.1 or Spi-B, mice deficient in both PU.1 and Spi-B in the B-cell lineage had reduced frequencies of B cells as well as impaired B-cell differentiation. Strikingly, all PU.1 and Spi-B–deficient mice developed pre-B cell acute lymphoblastic leukemia before 30 weeks of age. Pre-B cells accumulated in the thymus resulting in massive thymic enlargement and dyspnea. These findings demonstrate that PU.1 and Spi-B are essential transcriptional regulators of B-cell differentiation as well as novel tumor suppressors in the B-cell lineage.

The author shares her insights about the role of interpreters in the story of women's rights in the U.S. She narrates the involvement of Elizabeth Cady Stanton leading to the First Women's Rights Convention in Seneca Falls, New York. She believes that despite the legacy of the women's rights movement, the equality of men and women remains a debated topic since 1851. She declares that the interpreter must strive to respect the visitors' personal connection to the convention's legacy.

PU.1 is a member of the Ets family of transcription factors required for the development of various lymphoid and myeloid cell lineages, but its role in natural killer (NK) cell development is not known. The study shows that PU.1 is expressed in NK cells and that, on cell transfer into alymphoid Rag2/γc−/−mice, hematopoietic progenitors of PU.1−/−fetal liver cells could generate functional NK cells but not B or T cells. Nevertheless, the numbers of bone marrow NK cell precursors and splenic mature NK cells were reduced compared to controls. Moreover,PU.1−/− NK cells displayed reduced expression of the receptors for stem cell factor and interleukin (IL)-7, suggesting a nonredundant role for PU.1 in regulating the expression of these cytokine receptor genes during NK cell development.PU.1−/− NK cells also showed defective expression of inhibitory and activating members of the Ly49 family and failed to proliferate in response to IL-2 and IL-12. Thus, despite the less stringent requirement for PU.1 in NK cell development compared to B and T cells, PU.1 regulates NK cell differentiation and homeostasis.

PU.1 is a member of the Ets family of transcription factors required for the development of various lymphoid and myeloid cell lineages, but its role in natural killer (NK) cell development is not known. The study shows that PU.1 is expressed in NK cells and that, on cell transfer into alymphoid Rag2/γc−/−mice, hematopoietic progenitors of PU.1−/−fetal liver cells could generate functional NK cells but not B or T cells. Nevertheless, the numbers of bone marrow NK cell precursors and splenic mature NK cells were reduced compared to controls. Moreover,PU.1−/−NK cells displayed reduced expression of the receptors for stem cell factor and interleukin (IL)-7, suggesting a nonredundant role for PU.1 in regulating the expression of these cytokine receptor genes during NK cell development.PU.1−/−NK cells also showed defective expression of inhibitory and activating members of the Ly49 family and failed to proliferate in response to IL-2 and IL-12. Thus, despite the less stringent requirement for PU.1 in NK cell development compared to B and T cells, PU.1 regulates NK cell differentiation and homeostasis.

Bone marrow is the major site of hematopoiesis in the adult mammal. Bone marrow contains a highly organized microenvironment for the support of hematopoietic stem and progenitor cells, including the production of growth factors. Bone marrow cells also produce negative regulatory factors which may regulate hematopoiesis and inflammatory responses. In this paper we describe Reptimed, a unique bone marrow-derived factor with inhibitory activity for myelopoiesis andin vitrogrowth of myeloid cell lines. Reptimed was partially purified from bone marrow supernatants using a combination of solid-phase extraction and size exclusion chromatography. Reptimed is <1000 Da MW and is water soluble. Reptimed inhibited growth of granulocyte–macrophage and macrophage colonies as well as proliferation of several myeloid leukemia cell lines. Reptimed may be part of a hemoregulatory circuit.

PU.1 is a transcription factor essential for myeloid development. High PU.1 expression leads to myeloid differentiation and cell cycle arrest while low expression leads to increased proliferation and self-renewal. Reduced PU.1 can lead to the development of acute myeloid leukemia. PU.1 induces microRNAs targeting genes involved in cell cycle and lipid metabolism. Included in these genes are E2f1and Aclythat are associated with lipogenesis. E2F1 is directly involved in cell cycle activation and lipid metabolism. Aclyencodes ATP citrate lyase (ACL) that is responsible for the production of acetyl-CoA from mitochondrial citrate. Acetyl-CoA is the first precursor molecule for lipid biosynthesis, and lipogenesis is essential for proliferation. Therefore, we hypothesized that PU.1 activates genes encoding microRNAs to downregulate E2F1 and ACL, leading to cell cycle arrest. To test this hypothesis, we utilized a PU.1 inducible cell line system that is derived from myeloid progenitor cells in the liver of fetal BN mice (Spi1BN/BN) that express PU.1 at 20% of normal levels (iBN cells). Following doxycycline administration, iBN cells can be induced to express high levels of PU.1, causing cell cycle arrest and differentiation of these myeloid progenitor cells into a macrophage like phenotype. MicroRNA (miR)-223 and miR-141 were identified as PU.1 inducible miRs that target E2f1and Acly, respectively. Supplementation with acetyl-CoA or acetate significantly rescued cell cycle inhibition caused by PU.1 induction. Inhibition of ACL activity by the chemical inhibitor BMS303141 (BMS) was sufficient to reduce cell cycle progression in cultured iBN cells. Supplementation with acetyl-CoA or acetate significantly rescued cell cycle progression in iBN cells inhibited with BMS. To determine if ACL levels play a role during normal macrophage differentiation, we investigated mRNA transcript levels of Aclyover the course of myeloid progenitor differentiation promoted by M-CSF or GM-CSF. We found that AclymRNA transcript levels inversely correlated with Adgre1(encoding F4/80) or Spi1(encoding PU.1) during macrophage differentiation. We conclude that during myeloid differentiation, PU.1 regulates the expression of genes involved in lipid metabolism through miRs, and influences cell cycle progression by limiting fatty acid synthesis required for proliferation. Elucidation of the mechanisms by which lipid metabolism may control cell cycle in myeloid progenitor cells is an important issue for further study.

B cell development is controlled by stage-specific expression of transcription factors. Aberrant expression of such factors can lead to B cell acute lymphoblastic leukemia (B-ALL). Deletion of genes encoding the E26 transformation-specific (ETS) transcription factors, PU.1 and Spi-B, in B cells (CD19+/CreSfpi1lox/loxSpib-/-mice, abbreviated to CD19-CreΔPB) leads to B-ALL at 100% incidence and with a median survival of 21 weeks. However, little is known about the target genes of PU.1 and Spi-B that explain leukemic transformation in these mice. In the current study, we investigated the developmental origins and mechanisms of leukemogenesis in CD19-CreΔPB mice. We found that B-ALL cells in CD19-CreΔPB mice had frequently rearranged both their heavy and light chain genes, but retained cell surface expression of interleukin-7 receptor (IL-7R), suggesting aberrant pre-B cell differentiation. Preleukemic CD19-CreΔPB mice had increased frequencies of pre-B cells compared to wild type mice. Pre-B cells, but not mature B cells, purified from the bone marrow of preleukemic CD19-CreΔPB mice could rapidly transfer disease to transplanted recipient mice. B-ALL cells from established tumors had uniform expression of markers indicating a pre-B cell phenotype and contained a high-frequency of leukemia-initiating cells as measured by transplantation assays. Genome-wide analysis of gene expression showed that B cell receptor signalling was the top impaired pathway in B-ALL cells from CD19-CreΔPB mice. Bone marrow cells from CD19-CreΔPB mice had increased responsiveness to IL-7R signalling and could be cultured as IL-7-dependent cell lines. Preleukemic or leukemic cells from CD19-CreΔPB mice expressed reduced levels of the gene encoding Bruton's tyrosine kinase (Btk), which we show is a target gene of PU.1 and/or Spi-B that in combination with reduced BLNK is sufficient to explain increased IL-7R responsiveness. We conclude that mutation of PU.1 and Spi-B predispose developing B cells to leukemogenesis by impairing expression of key genes, such as Btk, that are required for BCR signalling and are involved in attenuation of IL-7 receptor signaling.

No relevant conflicts of interest to declare.

The Ets transcription factor PU.1 (encoded by the gene Sfpi1) functions in a concentration-dependent manner as a hematopoietic cell fate determinant. PU.1 levels are uniform in early hematopoiesis, increase during myeloid differentiation, and decrease after erythrocyte and T cell/natural killer cell commitment. It is unknown how downstream target genes respond to changes in PU.1 concentration. To address this, we generated mice with two distinct hypomorphic alleles of Sfpi1 and analyzed interleukin-3 dependent cell lines from fetal liver cells homozygous for either allele. PU.1 was produced in these cells at ∼20% (Sfpi1BN/BN) or ∼2% (Sfpi1Blac/Blac) of wild type. These cells fail to terminally differentiate as a consequence of low PU.1 expression and can be maintained as cell lines. To determine what groups of genes are expressed in response to two distinct PU.1 concentrations, we performed whole-genome microarray analysis and compared gene expression in Sfpi1BN/BN and Sfpi1Blac/Blac cell lines to Sfpi1−/− cell lines. Groups of downstream target genes were activated or repressed in four modes in response to the two discrete concentrations of PU.1: at higher but not lower PU.1 concentration, at lower but not higher PU.1 concentration, at both lower and higher concentration, and in a gradient fashion. We decided to focus on genes regulated in a gradient manner, because dose-dependency suggests that these may be direct targets of PU.1. Genes activated in a gradient manner were mostly myeloid-specific and enriched for target genes of PU.1. Genes repressed in a gradient manner included erythroid-specific genes and, unexpectedly, T cell and natural killer cell-specific genes. T cell genes were also repressed by PU.1 in cultured progenitor-B cells. With this unique allelic system, we can study three discrete concentrations of PU.1 at 20%, 2%, and 0% to examine concentration-dependent effects of PU.1 on target genes and lineage decisions. Overall, our results suggest that PU.1 functions in a concentration-dependent manner to promote myeloid differentiation and repress T cell or natural killer cell development.

The Ets transcription factor PU.1 (encoded by the gene Sfpi1) functions in a concentration-dependent manner as a hematopoietic cell fate determinant. PU.1 levels are uniform in early hematopoiesis, increase during myeloid differentiation, and decrease after erythrocyte and T cell/natural killer cell commitment. It is unknown how downstream target genes respond to changes in PU.1 concentration. To address this, we generated mice with two distinct hypomorphic alleles of Sfpi1and analyzed interleukin-3 dependent cell lines from fetal liver cells homozygous for either allele. PU.1 was produced in these cells at ∼20% (Sfpi1BN/BN) or ∼2% (Sfpi1Blac/Blac) of wild type. These cells fail to terminally differentiate as a consequence of low PU.1 expression and can be maintained as cell lines. To determine what groups of genes are expressed in response to two distinct PU.1 concentrations, we performed whole-genome microarray analysis and compared gene expression in Sfpi1BN/BNand Sfpi1Blac/Blaccell lines to Sfpi1−/−cell lines. Groups of downstream target genes were activated or repressed in four modes in response to the two discrete concentrations of PU.1:

PU.1 is an Ets-family transcription factor that is critical for normal hematopoiesis. Knock-out studies of PU.1 demonstrate that it is required for the development of myeloid and lymphoid cells. PU.1 is expressed at graded levels in the immune system, with higher expression in myeloid lineage cells. Previous research has suggested that differences in PU.1 concentration are involved in hematopoietic cell fate decisions. However, the mechanism by which target genes in developing cells sense and respond to different PU.1 concentrations is unknown. To address this question, we inserted a beta-lactamase reporter gene and a neomycin resistance cassette into the first coding exon of PU.1. Mice homozygous for the targeted allele (PU.1BN), and an allele in which neomycin was removed (PU.1BLAC), had graded reductions in PU.1 expression and function. PU.1BN/BN mice display a substantial reduction in fetal and neonatal myelopoiesis and are macrophage deficient, resulting in a failure of vascular regression in the developing eye. A fraction of these mice survive until weaning and have an accumulation of c-Kit-expressing immature myeloid cells in the spleen and bone marrow. In addition, these mice acquire a myelomonocytic infiltration in the retina and hyaloid vessels of the eye. Mice homozygous for the second allele (PU.1BLAC/BLAC) survive less than one week and have a severe reduction in PU.1 activity and a block to myeloid development. Fetal livers from both types of mice lack any discernable mature myeloid cells and demonstrate a graded reduction in colony forming ability. To determine if these two alleles have an effect on gene expression, we established IL-7-dependent pro-B cell lines or IL-3-dependent cell lines from these mice as well as from wild type or PU.1 null controls. We found that PU.1-dependent target genes in these cell lines are expressed in a graded manner. In particular, expression of the gene encoding the low-affinity IgG Fc receptor (FcγRIIb) is directly responsive to PU.1 activity. In summary, these results demonstrate the critical importance of an appropriate threshold of PU.1 activity for normal hematopoiesis. The hypomorphic PU.1 alleles provide an excellent model for determining the mechanism of graded PU.1 levels in hematopoiesis.

PU.1 is an Ets-family transcription factor that is critical for normal hematopoiesis. Knock-out studies of PU.1demonstrate that it is required for the development of myeloid and lymphoid cells. PU.1 is expressed at graded levels in the immune system, with higher expression in myeloid lineage cells. Previous research has suggested that differences in PU.1 concentration are involved in hematopoietic cell fate decisions. However, the mechanism by which target genes in developing cells sense and respond to different PU.1 concentrations is unknown. To address this question, we inserted a beta-lactamase reporter gene and a neomycin resistance cassette into the first coding exon of PU.1. Mice homozygous for the targeted allele (PU.1BN), and an allele in which neomycin was removed (PU.1BLAC), had graded reductions in PU.1 expression and function. PU.1BN/BNmice display a substantial reduction in fetal and neonatal myelopoiesis and are macrophage deficient, resulting in a failure of vascular regression in the developing eye. A fraction of these mice survive until weaning and have an accumulation of c-Kit-expressing immature myeloid cells in the spleen and bone marrow. In addition, these mice acquire a myelomonocytic infiltration in the retina and hyaloid vessels of the eye. Mice homozygous for the second allele (PU.1BLAC/BLAC) survive less than one week and have a severe reduction in PU.1 activity and a block to myeloid development. Fetal livers from both types of mice lack any discernable mature myeloid cells and demonstrate a graded reduction in colony forming ability. To determine if these two alleles have an effect on gene expression, we established IL-7-dependent pro-B cell lines or IL-3-dependent cell lines from these mice as well as from wild type or PU.1 null controls. We found that PU.1-dependent target genes in these cell lines are expressed in a graded manner. In particular, expression of the gene encoding the low-affinity IgG Fc receptor (FcγRIIb) is directly responsive to PU.1 activity. In summary, these results demonstrate the critical importance of an appropriate threshold of PU.1 activity for normal hematopoiesis. The hypomorphic PU.1alleles provide an excellent model for determining the mechanism of graded PU.1 levels in hematopoiesis.

Leukemia arises as a result of a multistep progression of blood cells towards malignancy. Genetic blocks in cell development and uncontrolled proliferation are often key steps in leukemic transformation. PU.1 is a transcription factor that has several known roles in hematopoietic development. Knock-out studies of PU.1 demonstrate a requirement for PU.1 in the development of myeloid and lymphoid cells. In addition, PU.1 concentration can influence cell development in normal hematopoiesis. Low concentrations of PU.1 are required for lymphoid development, whereas high concentrations of PU.1 are required for myeloid development. Mice that have been genetically altered to express abnormally low concentrations of PU.1 develop acute myeloid leukemia (AML) due to a developmental block within the myeloid lineage. In contrast, abnormally high concentrations of PU.1 in mice are capable of inducing erythroleukemias. This can occur through Friend virus integration into the PU.1 gene, which leads to over-expression of PU.1, and the subsequent blocking of erythroid development due to PU.1’s ability to antagonize GATA-1 function. Here we report that ectopic expression of PU.1 is sufficient to immortalize hematopoietic progenitors in a GM-CSF-dependent manner. When fetal hematopoietic progenitors are infected with a PU.1 retrovirus and placed into culture with GM-CSF, a small fraction become immortalized. These PU.1-infected progenitors (PIP cells) are arrested in development and can grow for several months in culture. PU.1 Immortalized Progenitors (PIP) are GM-CSF restricted and dependent. In addition, PIP cells display blast-like morphology and express cell surface markers indicative of myeloid development. In contrast, gene expression analysis of clonal PIP cell lines demonstrates the expression of stem cell, myeloid, and erythroid specific genes. Finally, PIP cells display low levels of endogenous PU.1 production as shown by real-time PCR relative to wild-type cells. These observations support a model in which high levels of PU.1, in conjunction with signaling through the GM-CSF receptor, block cell differentiation at an early stage in hematopoiesis. This discovery provides new information into PU.1’s ability to control cell differentiation and its ability to induce leukemia.

The cytokine interleukin-7 (IL-7) is required for B and T lymphocyte development, and for the survival and maintenance of both naive and memory T lymphocytes. The receptor for IL-7 (IL-7R) is heterodimeric, consisting of a common gamma chain (γc) and an alpha subunit (IL-7Rα). The γc is expressed in most hematopoietic cell types, but the IL-7Rα gene is regulated in a cell type and developmental stage-specific manner. We have previously shown that the Ets-family transcription factor PU.1 is required to activate transcription of the IL-7Rα gene during fetal lymphocyte development. However, several questions remain unanswered. First, the IL-7Rα promoter is poorly characterized. Second, the IL-7Rα is expressed at high levels in the T cell lineage where PU.1 is not expressed. Third, the transcription factor early B cell factor (EBF) can activate IL-7Rα transcription in developing B cells in the absence of PU.1. To address these questions, we have characterized the IL-7Rα promoter region in detail. First, we determined that the major transcription start sites in the IL-7Rα gene are downstream of an Ets/PU.1 binding site. We found that the intact Ets site is required for IL-7Rα promoter activity, as well as to mediate enhancer action from a distance. IL-7Rα promoter activity depends on the proper orientation of the Ets site relative to functional initiator sequences. We found, using gel shift analysis, that both PU.1 and the Ets transcription factor GA binding protein (GABP) are expressed in developing B cells, and can interact with the Ets binding site in the IL-7Rα promoter. However, the function of PU.1 is distinct from GABP during B cell development. Retroviral transduction of PU.1 mutant progenitor cells with a PU.1 retrovirus robustly rescues IL-7Rα transcription and IL-7-dependent B cell development. In contrast, transduction with GABPα and GABPβ1 subunits fails to activate IL-7Rα transcription in PU.1 mutant progenitor cells. We conclude that activation of the IL-7Rα gene requires PU.1 during the earliest stages of lymphocyte development, but is alternatively utilized by PU.1 and GABP after commitment to the B cell lineage.

The cytokine interleukin-7 (IL-7) is required for B and T lymphocyte development, and for the survival and maintenance of both naive and memory T lymphocytes. The receptor for IL-7 (IL-7R) is heterodimeric, consisting of a common gamma chain (γc) and an alpha subunit (IL-7Rα). The γc is expressed in most hematopoietic cell types, but the IL-7Rα gene is regulated in a cell type and developmental stage-specific manner. We have previously shown that the Ets-family transcription factor PU.1 is required to activate transcription of the IL-7Rα gene during fetal lymphocyte development. However, several questions remain unanswered. First, the IL-7Rα promoter is poorly characterized. Second, the IL-7Rα is expressed at high levels in the T cell lineage where PU.1 is not expressed. Third, the transcription factor early B cell factor (EBF) can activate IL-7Rα transcription in developing B cells in the absence of PU.1. To address these questions, we have characterized the IL-7Rα promoter region in detail. First, we determined that the major transcription start sites in the IL-7Rα gene are downstream of an Ets/PU.1 binding site. We found that the intact Ets site is required for IL-7Rα promoter activity, as well as to mediate enhancer action from a distance. IL-7Rα promoter activity depends on the proper orientation of the Ets site relative to functional initiator sequences. We found, using gel shift analysis, that both PU.1 and the Ets transcription factor GA binding protein (GABP) are expressed in developing B cells, and can interact with the Ets binding site in the IL-7Rα promoter. However, the function of PU.1 is distinct from GABP during B cell development. Retroviral transduction of PU.1 mutant progenitor cells with a PU.1 retrovirus robustly rescues IL-7Rα transcription and IL-7-dependent B cell development. In contrast, transduction with GABPα and GABPβ1 subunits fails to activate IL-7Rα transcription in PU.1 mutant progenitor cells. We conclude that activation of the IL-7Rα gene requires PU.1 during the earliest stages of lymphocyte development, but is alternatively utilized by PU.1 and GABP after commitment to the B cell lineage.

Leukemia arises as a result of a multistep progression of blood cells towards malignancy. Genetic blocks in cell development and uncontrolled proliferation are often key steps in leukemic transformation. PU.1 is a transcription factor that has several known roles in hematopoietic development. Knock-out studies of PU.1 demonstrate a requirement for PU.1 in the development of myeloid and lymphoid cells. In addition, PU.1 concentration can influence cell development in normal hematopoiesis. Low concentrations of PU.1 are required for lymphoid development, whereas high concentrations of PU.1 are required for myeloid development. Mice that have been genetically altered to express abnormally low concentrations of PU.1 develop acute myeloid leukemia (AML) due to a developmental block within the myeloid lineage. In contrast, abnormally high concentrations of PU.1 in mice are capable of inducing erythroleukemias. This can occur through Friend virus integration into the PU.1 gene, which leads to over-expression of PU.1, and the subsequent blocking of erythroid development due to PU.1's ability to antagonize GATA-1 function. Here we report that ectopic expression of PU.1 is sufficient to immortalize hematopoietic progenitors in a GM-CSF-dependent manner. When fetal hematopoietic progenitors are infected with a PU.1 retrovirus and placed into culture with GM-CSF, a small fraction become immortalized. These PU.1-infected progenitors (PIP cells) are arrested in development and can grow for several months in culture. PU.1 Immortalized Progenitors (PIP) are GM-CSF restricted and dependent. In addition, PIP cells display blast-like morphology and express cell surface markers indicative of myeloid development. In contrast, gene expression analysis of clonal PIP cell lines demonstrates the expression of stem cell, myeloid, and erythroid specific genes. Finally, PIP cells display low levels of endogenous PU.1 production as shown by real-time PCR relative to wild-type cells. These observations support a model in which high levels of PU.1, in conjunction with signaling through the GM-CSF receptor, block cell differentiation at an early stage in hematopoiesis. This discovery provides new information into PU.1's ability to control cell differentiation and its ability to induce leukemia.

Current models of hematopoiesis suggest that in early, pluripotent progenitor cells, lineage-specific transcription factors are expressed at low levels. During differentiation, subsets of these transcription factors become dominantly expressed in a lineage-restricted fashion. Understanding how transcription factors are expressed in distinct cell-types is central to defining the regulatory events that occur during lineage selection. GATA-1 is an essential transcriptional regulator for the erythroid and megakaryocyte lineages, while it is absent in neutrophils and monocytes. PU.1, on the other hand, is a critical transcription factor for neutrophils and monocytes, but it is not abundantly expressed in erythroid cells. Although these two factors have been shown to be antagonistic in monocytic and erythroid cells, both GATA-1 and PU.1 are required for the normal development of the mast lineage (Migliaccio et al., 2003, Walsh et al., 2002). Here we show that mast cells express a unique mRNA isoform of GATA-1 that is distinct from the major erythroid/megakaryocyte isoform. It is related, but not identical to the Ib transcript that has been described as a minor expressed form in erythroid cells (Tsai et al., 1991) and as a major expressed form in RNA isolated from CFU-GM primary myeloid cultures (Seshasayee et al., 2000). This GATA-1 mast cell isoform (GATA-1mast) differs from the erythroid/megakaryocyte isoform by a unique, untranslated first exon that is alternatively spliced onto the downstream coding exons. In mast cells, GATA-1mast is expressed from a promoter separate from that utilized in megakaryocytic and erythroid cells. Comparative analysis of genomic sequence of the GATA-1 locus in this region reveals modules of extensive phylogenetic conservation in mammals, including stretches containing both highly conserved PU.1 and GATA binding sites. We have performed chromatin immunoprecipitation studies with GATA-1 antibodies and have defined multiple regions of in vivo binding within the GATA-1 locus in erythroid cells. Addtional studies are underway utilizing the Scanning ChIP procedure (Zeller et al., 2001) to determine in vivo GATA-1, GATA-2, and PU.1 binding sites of these factors to the GATA-1 locus in mast cells. In order to determine whether PU.1 positively regulates the expression of the mast cell GATA-1 isoform, we have examined GATA-1mast expression in PU.1 −/ − cells. PU.1 −/ − fetal liver cells cannot differentiate into mast cells in vitro; reintroduction of PU.1 expression restores mast cell differentiation. We show that PU.1 −/ − cells are deficient in expression of the GATA-1 mast cell mRNA isoform, and reintroduction of PU.1 into the PU.1 deficient cells markedly up-regulates the expression of GATA-1mast. Our findings demonstrate that PU.1 positively regulates a distinct GATA-1 isoform during mast cell differentiation. We propose a model in which GATA factors cooperate with PU.1 to direct cell-specific isoforms of transcriptional regulators during hematopoietic development.

Current models of hematopoiesis suggest that in early, pluripotent progenitor cells, lineage-specific transcription factors are expressed at low levels. During differentiation, subsets of these transcription factors become dominantly expressed in a lineage-restricted fashion. Understanding how transcription factors are expressed in distinct cell-types is central to defining the regulatory events that occur during lineage selection.