A genome-wide relay of signalling-responsive enhancers drives hematopoietic specification.

Developmental control of gene expression critically depends on distal cis-regulatory elements including enhancers which interact with promoters to activate gene expression. To date no global experiments have been conducted that identify their cell type and cell stage-specific activity within one developmental pathway and in a chromatin context. Here, we describe a high-throughput method that identifies thousands of differentially active cis-elements able to stimulate a minimal promoter at five stages of hematopoietic progenitor development from embryonic stem (ES) cells, which can be adapted to any ES cell derived cell type. We show that blood cell-specific gene expression is controlled by the concerted action of thousands of differentiation stage-specific sets of cis-elements which respond to cytokine signals terminating at signalling responsive transcription factors. Our work provides an important resource for studies of hematopoietic specification and highlights the mechanisms of how and where extrinsic signals program a cell type-specific chromatin landscape driving hematopoietic differentiation.

Depletion of RUNX1/ETO in t(8;21) AML cells leads to genome-wide changes in chromatin structure and transcription factor binding.

The t(8;21) translocation fuses the DNA-binding domain of the hematopoietic master regulator RUNX1 to the ETO protein. The resultant RUNX1/ETO fusion protein is a leukemia-initiating transcription factor that interferes with RUNX1 function. The result of this interference is a block in differentiation and, finally, the development of acute myeloid leukemia (AML). To obtain insights into RUNX1/ETO-dependant alterations of the epigenetic landscape, we measured genome-wide RUNX1- and RUNX1/ETO-bound regions in t(8;21) cells and assessed to what extent the effects of RUNX1/ETO on the epigenome depend on its continued expression in established leukemic cells. To this end, we determined dynamic alterations of histone acetylation, RNA Polymerase II binding and RUNX1 occupancy in the presence or absence of RUNX1/ETO using a knockdown approach. Combined global assessments of chromatin accessibility and kinetic gene expression data show that RUNX1/ETO controls the expression of important regulators of hematopoietic differentiation and self-renewal. We show that selective removal of RUNX1/ETO leads to a widespread reversal of epigenetic reprogramming and a genome-wide redistribution of RUNX1 binding, resulting in the inhibition of leukemic proliferation and self-renewal, and the induction of differentiation. This demonstrates that RUNX1/ETO represents a pivotal therapeutic target in AML.

Aberrant expression of CD19 in AML with t(8;21) involves a poised chromatin structure and PAX5.

Correct hematopoietic differentiation requires the tightly regulated execution of lineage-specific and stage-restricted gene expression programs. This process is disturbed in hematological malignancies that typically show incomplete differentiation but often also display a mixed lineage phenotype. Co-expression of lymphoid and myeloid molecules is a well-known feature of acute myeloblastic leukemia (AML) with t(8;21). These cells consistently express the B-cell-specific transcription factor PAX5, and the B-cell-specific cell surface protein CD19. However, the functional consequences of PAX5 expression are unknown. To address this question, we studied the chromatin features of CD19, which is a direct target of PAX5 in cells with and without the t(8;21) chromosomal translocation. We show that CD19 chromatin exists in a poised configuration in myeloid progenitors and that this poised chromatin structure facilitates PAX5-dependent CD19 activation. Our results also show a positive correlation between PAX5 and CD19 expression in t(8;21)-positive AML cells and demonstrate that PAX5 binds to the promoter and enhancer of CD19 gene and remodels chromatin structure at the promoter. This study shows that expression of PAX5 in leukemic cells has functional consequences and points to an important role of a progenitor-specific chromatin configuration in myeloid leukemia.

A NF-kappa B/Sp1 region is essential for chromatin remodeling and correct transcription of a human granulocyte-macrophage colony-stimulating factor transgene.

The GM-CSF gene is expressed following activation of T cells. The proximal promoter and an upstream enhancer have previously been characterized using transfection and reporter assays in T cell lines in culture. A 10.5-kb transgene containing the entire human GM-CSF gene has also been shown to display inducible, position-independent, copy number-dependent transcription in mouse splenocytes. To determine the role of individual promoter elements in transgene function, mutations were introduced into the proximal promoter and activity assessed following the generation of transgenic mice. Of four mutations introduced into the transgene promoter, only one, in an NF-kappaB/Sp1 region, led to decreased induction of the transgene in splenocytes or bone marrow-derived macrophages. This mutation also affected the activity of reporter gene constructs stably transfected into T cell lines in culture, but not when transiently transfected into the same cell lines. The mutation alters the NF-kappaB family members that bind to the NF-kappaB site as well as reducing the binding of Sp1 to an adjacent element. A DNase I hypersensitive site that is normally generated at the promoter following T cell activation on the wild-type transgene does not appear in the mutant transgene. These results suggest that the NF-kappaB/Sp1 region plays a critical role in chromatin remodeling and transcription on the GM-CSF promoter in primary T cells.

Reconstitution of T cell-specific transcription directed by composite NFAT/Oct elements.

The complex nature of most promoters and enhancers makes it difficult to identify key determinants of tissue-specific gene expression. Furthermore, most tissue-specific genes are regulated by transcription factors that have expression profiles more widespread than the genes they control. NFAT is an example of a widely expressed transcription factor that contributes to several distinct patterns of cytokine gene expression within the immune system and where its role in directing specificity remains undefined. To investigate distinct combinatorial mechanisms employed by NFAT to regulate tissue-specific transcription, we examined a composite NFAT/AP-1 element from the widely active GM-CSF enhancer and a composite NFAT/Oct element from the T cell-specific IL-3 enhancer. The NFAT/AP-1 element was active in the numerous cell types that express NFAT, but NFAT/Oct enhancer activity was T cell specific even though Oct-1 is ubiquitous. Conversion of the single Oct site in the IL-3 enhancer to an AP-1 enabled activation outside of the T cell lineage. By reconstituting the activities of both the IL-3 enhancer and its NFAT/Oct element in a variety of cell types, we demonstrated that their T cell-specific activation required the lymphoid cofactors NIP45 and OCA-B in addition to NFAT and Oct family proteins. Furthermore, the Oct family protein Brn-2, which cannot recruit OCA-B, repressed NFAT/Oct enhancer activity. Significantly, the two patterns of combinatorial regulation identified in this study mirror the cell-type specificities of the cytokine genes that they govern. We have thus established that simple composite transcription factor binding sites can indeed establish highly specific patterns of gene expression.

Generation of an improved luciferase reporter gene plasmid that employs a novel mechanism for high-copy replication.

We have engineered the reporter gene plasmid pXPG by incorporating a novel high-copy origin of replication and a modified luciferase gene into a pXP1-derived vector that efficiently blocks read-through transcription in eukaryotic cells. pXPG contains the Luc+ luciferase gene derived from pGL3 that lacks a peroxisomal targeting sequence, thereby allowing accumulation of luciferase protein in the cytoplasm rather than subcellular organelles of transfected eukaryotic cells. pXPG has distinct advantages over pGL3, because it contains SV40 polyadenylation signals that appear to be more efficient at blocking read-through transcription than the synthetic polyadenylation signal present in pGL3. pXPG contains a novel mutation near the origin of replication that increases plasmid copy number in Escherichia coli. This mutation alters the -10 sequence in the RNA II promoter of the ColE1 origin of replication from TAATCT to TAATAT. As this sequence is a closer match to the consensus -10 element, we suggest that the mutation increases copy number by increasing the rate of transcription of the RNA II replication primer. This novel mechanism for increasing copy number may have more widespread applications than the commonly used pUC high-copy origin of replication mutation. Unlike pUC, which reverts to low copy number at 30 degrees C, the pXPG mutation supports a higher copy number at both 37 and 30 degrees C.

Nuclear factor of activated T cells contributes to the function of the CD28 response region of the granulocyte macrophage-colony stimulating factor promoter.

The granulocyte macrophage colony stimulating factor (GM-CSF) promoter contains a 10 bp element known as CK-1 or CD28RE that specifically responds to the co-stimulatory signal delivered to T cells via the CD28 surface receptor. This element is a variant NFkappaB site that does not function alone but requires an adjacent promoter region that includes a classical NFkappaB element, an Sp-1 site and a putative activator protein-1 (AP-1)-like binding site. The entire region is referred to as the CD28 response region (CD28RR). The GM-CSF CK-1 element has been shown to bind NFkappaB proteins, in particular c-Rel, whose binding and function is dependent on the architectural transcription factor HMGI(Y). It has been previously suggested that the nuclear factor of activated T cells (NFAT) family of proteins also plays a role in the activity of this region. We show here that recombinant NFATp but not AP-1 can bind to the GM-CSF CD28RR. NFATp present in activated Jurkat T cell extracts can also interact with the CD28RR. The binding of NFATp and Rel proteins requires the same core CK-1 sequences, and appears to be mutually exclusive. We investigated the functional significance of NFATp binding to CK-1 by overexpressing the protein in Jurkat T cells and found that NFATp cannot activate the CD28RR alone but can cooperate with signals generated by phorbol 12-myristate 13-acetate/calcium ionophore. The CD28RR is therefore a complex region that can bind and respond to a combination of transcription factors and signals.

The human granulocyte-macrophage colony-stimulating factor gene is autonomously regulated in vivo by an inducible tissue-specific enhancer.

The granulocyte-macrophage colony-stimulating factor (GM-CSF) gene is part of a cytokine gene cluster and is directly linked to a conserved upstream inducible enhancer. Here we examined the in vitro and in vivo functions of the human GM-CSF enhancer and found that it was required for the correctly regulated expression of the GM-CSF gene. An inducible DNase I-hypersensitive site appeared within the enhancer in cell types such as T cells, myeloid cells, and endothelial cells that express GM-CSF, but not in nonexpressing cells. In a panel of transfected cells the human GM-CSF enhancer was activated in a tissue-specific manner in parallel with the endogenous gene. The in vivo function of the enhancer was examined in a transgenic mouse model that also addressed the issue of whether the GM-CSF locus was correctly regulated in isolation from other segments of the cytokine gene cluster. After correction for copy number the mean level of human GM-CSF expression in splenocytes from 11 lines of transgenic mice containing a 10.5-kb human GM-CSF transgene was indistinguishable from mouse GM-CSF expression (99% +/- 56% SD). In contrast, a 9.8-kb transgene lacking just the enhancer had a significantly reduced (P = 0.004) and more variable level of activity (29% +/- 89% SD). From these studies we conclude that the GM-CSF enhancer is required for the correct copy number-dependent expression of the human GM-CSF gene and that the GM-CSF gene is regulated independently from DNA elements associated with the closely linked IL-3 gene or other members of the cytokine gene cluster.

A T cell-specific enhancer in the interleukin-3 locus is activated cooperatively by Oct and NFAT elements within a DNase I-hypersensitive site.

Interleukin-3 (IL-3) is a cytokine that is expressed primarily in activated T cells. Here we identified an inducible T cell-specific enhancer 14 kb upstream of the IL-3 gene that responded to activation of T cell receptor signaling pathways. The IL-3 enhancer spanned an inducible cyclosporin A-sensitive DNase I-hypersensitive site found only in T cells. Four NFAT-like elements exist within the enhancer. The two most active NFAT-like elements were located at the center of the DNase I-hypersensitive site. One of these NFAT-like elements encompassed overlapping Oct- and NFATp/c-binding sites, which functioned in a highly synergistic manner. We suggest that the T cell-specific expression of the IL-3 gene is partly controlled through the enhancer by cooperation between Oct and NFAT family proteins.

Signals for activation of the GM-CSF promoter and enhancer in T cells.

Granulocyte-macrophage colony-stimulating factor (GM-CSF) is one of the many cytokines produced following T-cell activation. It is also produced in a variety of other cell types, in particular following activation by inflammatory mediators. Changes in the rate of transcription are important in the control of GM-CSF expression in T cells and in fibroblasts and endothelial cells. The GM-CSF gene contains two distinct transcriptional control regions. These are the proximal promoter consisting of the first 120 bp from the transcription start site and an enhancer located approximately 3 kb upstream from the proximal promoter. Distinct regions of the proximal promoter respond to a wide array of signals such as phorbol myristate acetate (PMA) and Ca2+ ionophore or phytohemaglutinin (PHA), CD28 activation, human T leukemia virus (HTLV)-1 tax, TNF, and interleukin 1 (IL-1). The transcription factors that mediate these responses have mainly been defined, with the major inducible proteins being the NF-kappa B/rel and AP-I families of transcription factors. In contrast to the promoter, the enhancer responds only to PMA and Ca2+ ionophore signals and binds NFAT/AP-1 complexes that appear to mediate its function.

Regulation of GM-CSF gene transcription by core-binding factor.

GM-CSF gene activation in T cells is known to involve the transcription factors nuclear factor-kappa B, AP-1, NFAT, and Sp1. Here we demonstrate that the human GM-CSF promoter and enhancer also encompass binding sites for core-binding factor (CBF). Significantly, the CBF sites are in each case contained within the minimum essential core regions required for inducible activation of transcription. Furthermore, these core regions of the enhancer and promoter each encompass closely linked binding sites for CBF, AP-1, and NFATp. The GM-CSF promoter CBF site TGTGGTCA is located 51 bp upstream of the transcription start site and also overlaps a YY-1 binding site. A 2-bp mutation within the CBF site resulted in a 2-3-fold decrease in the activities of both a 69-bp proximal promoter fragment and a 627-bp full-length promoter fragment. Stepwise deletions into the proximal promoter also revealed that the CBF site, but not the YY-1 site, was required for efficient induction of transcriptional activation. The AML1 and CBF beta genes that encode CBF each have the ability to influence cell growth and differentiation and have been implicated as proto-oncogenes in acute myeloid leukemia. This study adds GM-CSF to a growing list of cytokines and receptors that are regulated by CBF and which control the growth, differentiation, and activation of hemopoietic cells. The GM-CSF locus may represent one of several target genes that are dysregulated in acute myeloid leukemia.

Regulation of granulocyte-macrophage colony-stimulating factor and E-selectin expression in endothelial cells by cyclosporin A and the T-cell transcription factor NFAT.

Nuclear factor of activated T cells (NFAT) was originally described as a T-cell-specific transcription factor athat supported the activation of cytokine gene expression and mediated the immunoregulatory effects of cyclosporin A (CsA). As we observed that activated endothelial cells also expressed NFAT, we tested the antiinflammatory properties of CsA in endothelial cells. Significantly, CsA completely suppressed the induction of NFAT in endothelial cells and inhibited the activity of granulocyte-macrophage colony-stimulating factor (GM-CSF) gene regulatory elements that use NFAT by 60%. CsA similarly mediated a reduction of up to 65% in GM-CSF mRNA and protein expression in activated endothelial cells. CsA also suppressed E-selectin, but not vascular cell adhesion molecule-1 (VCAM-1) expression in endothelial cells, even though the E-selectin promoter is activated by NF-kappa B rather than NFAT. Hence, induction of cell surface expression of this leukocyte adhesion molecule by tumor necrosis factor (TNF)-alpha was reduced by 40% in the presence of CsA, and this was reflected by a 29% decrease in neutrophil adhesion. The effects of CsA on endothelial cells were also detected at the chromatin structure level, as DNasel hypersensitive sites within both the GM-CSF enhancer and the E-selectin promoter were suppressed by CsA. This represents the first report of NFAT in endothelial cells and suggests mechanisms by which CsA could function as an antiinflammatory agent.

Multiple signals are required for function of the human granulocyte-macrophage colony-stimulating factor gene promoter in T cells.

The human granulocyte-macrophage CSF (GM-CSF) gene is expressed in T cells in response to TCR activation that can be mimicked by treatment of the cells with PMA and Ca2+ ionophore. The gene contains a proximal functional promoter region (-620 to +34), as well as a powerful enhancer located 3 kb upstream, both of which are involved in the response of the gene to TCR activation. The proximal promoter contains a region termed CLEO (-54 to -31) that consists of a purine-rich element abutting an activator protein-1 (AP-1)-like site, as well as an upstream nuclear factor-kappa B (NF-kappa B) site (-85 to -76) and a CK-1 element (-101 to -92). We show in this work that mutations in either the purine-rich region of the CLEO element or the NF-kappa B site result in reduced PMA/Ca2+ activation of a 620-bp human GM-CSF promoter-luciferase reporter construct in Jurkat T cells by 65% and 50%, respectively. The major inducible protein complex that binds to the human CLEO (hCLEO) element is an AP-1-like complex that is inducible by PMA alone, but shows increased binding in response to PMA together with Ca2+ ionophore. Although the binding of this complex is not cyclosporin-sensitive, promoter induction is inhibited by cyclosporin treatment. A second weak inducible complex resembling nuclear factor of activated T cells (NF-AT) was also observed binding to the hCLEO region. By using recombinant proteins, we confirmed that AP-1, NF-ATp, and a higher order NF-ATp/AP-1 complex could all form with the hCLEO element, and we have also defined the sequence requirements for binding of each of these complexes. We found that expression of a constitutively active form of calcineurin could substitute for Ca2+ ionophore and synergize with PMA to activate the GM-CSF promoter, and conversely that mutant-activated Ras could substitute for PMA and cooperate with Ca2+ ionophore. Co-expression of Ras and calcineurin, however, did not activate the GM-CSF promoter, but required the additional expression of NF-kappa B p65. These results imply that at least three signals are required to activate the GM-CSF proximal promoter, and that the signals impinge on distinct transcription factors that bind to the hCLEO and NF-kappa B regions of the promoter.

Transcriptional regulation of mouse granulocyte-macrophage colony-stimulating factor/IL-3 locus.

Granulocyte-macrophage (GM)-CSF and IL-3 are hemopoietic growth factors whose genes are closely linked in both humans and mice. In humans, the GM-CSF and IL-3 genes are regulated by a cyclosporin A-inhibitable enhancer located 3 kb upstream of the GM-CSF gene that is inducible by signals that mimic TCR activation. To search for a murine homologue of this enhancer we probed mouse genomic DNA and located a 400-bp element 2 kb upstream of the mouse GM-CSF gene that was 76% homologous with the human GM-CSF enhancer. Like the human GM-CSF enhancer, this element formed a cyclosporin A-inhibitable DNase I-hypersensitive site in the murine T cell line EL4 upon activation with phorbol ester and calcium ionophore. Transient transfection assays showed that this homologue of the human enhancer acted as an inducible enhancer of the thymidine kinase promoter, the mouse IL-3 promoter, and the human GM-CSF promoter. We observed, however, that the mouse GM-CSF promoter was significantly more active than the human GM-CSF promoter and found that it supported a level of activity equivalent to the combination of the human GM-CSF promoter and the human GM-CSF enhancer. Consequently, the activity of mouse GM-CSF promoter was not significantly elevated in the presence of the mouse GM-CSF enhancer. Because the mouse GM-CSF enhancer is considerably less active than its human homologue we suggest that the mouse GM-CSF gene has evolved with less dependence upon the upstream enhancer for its activation.

Human granulocyte-macrophage colony-stimulating factor enhancer function is associated with cooperative interactions between AP-1 and NFATp/c.

The promoter of the human granulocyte-macrophage colony-stimulating factor gene is regulated by an inducible upstream enhancer. The enhancer encompasses three previously defined binding sites for the transcription factor NFAT (GM170, GM330, and GM550) and a novel NFAT site defined here as the GM420 element. While there was considerable redundancy within the enhancer, the GM330, GM420, and GM550 motifs each functioned efficiently in isolation as enhancer elements and bound NFATp and AP-1 in a highly cooperative fashion. These three NFAT sites closely resembled the distal interleukin-2 NFAT site, and methylation interference assays further defined GGA(N)9TCA as a minimum consensus sequence for this family of NFAT sites. By contrast, the GM170 site, which also had conserved GGA and TCA motifs but in which these motifs were separated by 15 bases, supported strong independent but no cooperative binding of AP-1 and NFATp, and this site functioned poorly as an enhancer element. While both the GM330 and GM420 elements were closely associated with the inducible DNase I-hypersensitive site within the enhancer, the GM420 element was the only NFAT site located within a 160-bp HincII-BalI fragment defined by deletion analysis as the essential core of the enhancer. The GM420 element was unusual, however, in containing a high-affinity NFATp/c-binding sequence (TGGAAAGA) immediately upstream of the sequence TGACATCA which more closely resembled a cyclic AMP response-like element than an AP-1 site. We suggest that the cooperative binding of NFATp/c and AP-1 requires a particular spacing of sites and that their cooperativity and induction via independent pathways ensure very tight regulation of the granulocyte-macrophage colony-stimulating factor enhancer.

Transcriptional regulation and chromatin structure of the human CD34 gene promoter region.

The human CD34 surface antigen is selectively expressed on stem/progenitor cells within the hematopoietic system. Because CD34 expression is tightly linked to the immature status of hematopoietic cells, with expression being rapidly lost as hematopoietic cells mature and differentiate, an examination of its regulation may provide important insights into the molecular control of blood cell development. A comparison of the CD34 nuclear transcription rate in CD34+ and CD34- cells indicated that the CD34 gene is transcriptionally regulated in hematopoietic cell lines. In a previous report, we had identified two major clusters of CD34 transcription initiation sites by 5' RACE (rapid amplification of cDNA ends) analysis. In transient transfection experiments, we now demonstrate the ability of sequences encompassing each of these clusters to function as promoters of transcription in CD34+ cells. These promoters functioned at equivalent levels in CD34+ and CD34- cells, and the addition of 5' flanking sequences, extending as far as 3.7 kb upstream, to the core promoters did not differentially modify the level of expression in CD34+ versus the CD34- cell lines. An examination of DNase I hypersensitivity sites within an 18-kb segment of DNA, extending 9 kb either side of the proximal promoter, showed six sites that were primarily associated with CD34- expressing cells. Taken together, these data indicate that the CD34 promoter sequences alone do not confer tissue-or stage-specific expression. Appropriate transcriptional regulation of the CD34 gene must be controlled by chromatin structure, as identified by DNase I hypersensitivity, and/or by other tissue- and stage-specific elements present outside of the promoter region.

The granulocyte-macrophage colony-stimulating factor/interleukin 3 locus is regulated by an inducible cyclosporin A-sensitive enhancer.

Granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin 3 (IL-3) are pleiotropic hemopoietic growth factors whose genes are closely linked and induced in T lymphocytes in a cyclosporin A (CsA)-sensitive fashion. Since we found that the human GM-CSF and IL-3 proximal promoters were not sufficient to account for the observed regulation of these genes, we mapped DNase I hypersensitive sites across the GM-CSF/IL-3 locus in the Jurkat human T-cell line to identify additional regulatory elements. We located an inducible DNase I hypersensitive site, 3 kb upstream of the GM-CSF gene, that functioned as a strong CsA-sensitive enhancer of both the GM-CSF and IL-3 promoters. Binding studies employing Jurkat cell nuclear extracts indicated that four sites within the enhancer associate with the inducible transcription factor AP1. Three of these AP1 elements lie within sequences that also associate with factors resembling the CsA-sensitive, T cell-specific transcription factor NFAT. We provide additional evidence suggesting that an AP1-like factor represents one of the components of NFAT. We propose that the intergenic enhancer described here is required for the correctly regulated activation of both GM-CSF and IL-3 gene expression in T cells and that it mediates the CsA sensitivity of the GM-CSF/IL-3 locus.