Transcriptional Control of Megakaryocyte Development

AN Goldfarb

Department of Pathology, University of Virginia School of Medicine, VA, USA

Megakaryocytes are highly specialized cells that arise and erythroblasts represents probably the most striking from a bipotent megakaryocytic-erythroid progenitor example of sibling divergence in developmental biology. (MEP). This developmental leap requires coordinated acti- Megakaryocytes pursue a ‘hypertrophic’ pathway, en- vation of megakaryocyte-specific genes, radical changes in compassing vast cellular enlargement, acquisition of cell cycle properties, and active prevention of erythroid polyploidy and the elaboration of several unique differentiation. These programs result from upregulation cytoplasmic membranous structures. Erythroblasts opt of megakaryocyte-selective transcription factors, down- for an ‘atrophic’ pathway characterized by cellular regulation of erythroid-selective transcription factors and shrinkage, cytoplasmic simplification with jettisoning ongoing mediation of common erythro-megakaryocytic of several organelles, and a highly focused, almost transcription factors. Unlike most developmental pro- single-minded, gene expression program. Underlying grams, no single lineage-unique family of master regula- this phenotypic divergence lies an array of transcription tors exerts executive control over the megakaryocytic factors, some of which are restricted to either lineage, plan. Rather, an assemblage of non-unique factors and but many of which are shared by both lineages signals converge to determine lineage and differentiation. (Figure 1). Further complicating this picture, most of In human megakaryopoiesis, hereditary disorders of these transcription factors subsume multiple tasks, platelet production have confirmed contributions from repressing or activating a repertoire of target genes that three distinct transcription factor families. Murine models shifts as a function of lineage and stage. have extended this repertoire to include multiple addi- Insights into individual players in this process have tional factors. At a mechanistic level, the means by which been gained from comparisons of megakaryocytic and these non-unique factors collaborate in the establishment erythroid promoter/enhancers, from identification of of a perfectly unique cell type remains a central question. genetic defects in human hereditary disorders of Oncogene (2007) 26, 6795–6802; doi:10.1038/sj.onc.1210762 megakaryopoiesis and from mouse genetics, both targeted and random. The major challenge consists of Keywords: megakaryocyte; platelet; development; tran- developing and validating an integrated model for the scription factor cooperative function of these various factors as they collaborate in programming megakaryopoiesis.

Discordant offspring from a mutually endearing parent GATA and Friend in sickness and in health: X-linked (MEP): erythro-megakaryocytic divergence thrombocytopenia

Megakaryocytic differentiation begins with the bipotent A canonical feature of megakaryocytic promoters is the megakaryocytic-erythroid progenitor (MEP), an entity presence of binding sites for the GATA family of zinc- whose existence was inferred from in vitro clonal finger transcription factors: WGATAR. Clinical rele- hematopoietic differentiation assays (Debili et al., vance for one of these sites is illustrated in a patient with 1996) and confirmed through prospective cellular isola- Bernard–Soulier syndrome (macrothrombocytopenia, tion (Akashi et al., 2000; Manz et al., 2002). The MEP decreased platelet expression of the gpI-IX-V complex may arise from either a committed common myeloid and decreased platelet aggregation in response to risto- progenitor (CMP) (Akashi et al., 2000) or directly from cetin) in whoma GATA binding site within the glycopro- a more primitive, uncommitted short-term hematopoie- tein Ibb promoter underwent mutation (Ludlow et al., tic stemcell (ST-HSC) (Adolfsson et al., 2005), by an 1996). GATA-1 and GATA-2 represent the major unknown mechanism that may involve programming by GATA proteins expressed during erythro-megakaryo- the transcription factor GATA-1 (Stachura et al., 2006). cytic differentiation, with GATA-1 levels increasing and The further development of MEP into megakaryocytes GATA-2 levels decreasing during the differentiation of both lineages (Cantor and Orkin, 2002). Coexpressed with these GATA factors is FOG1, a large multifinger Correspondence: Dr AN Goldfarb, University of Virginia Health Sciences Center, Box 800904, Charlottesville, Virginia, VA 22908, protein serving as a dedicated GATA coregulator. USA. Mouse knockouts have implicated GATA-1 and E-mail: [email protected] FOG1 specifically in erythroid and megakaryocytic Transcription in megakaryopoiesis AN Goldfarb 6796 EKLF c-Myb p300 Mediator (?) Ery

Early Late GATA-2/FOG1 GATA-1/FOG1 NF-E2 SCL LMO2 GFI-1b

Mk GABP Fli-1 RUNX1

Figure 1 Diagram of transcription factors involved in lineage divergence from a common megakaryocyte-erythroid progenitor. The factors are organized on the y-axis according to degree of lineage selectivity. For example, RUNX1 and EKLF represent highly selective megakaryocytic and erythroid factors, respectively. In the middle are factors equally involved in both lineages, such as GATA-1 and FOG1. Orientation on the x-axis reﬂects the stage of differentiation in which the factors are involved. For example c-Myb most likely participates in early erythroid commitment of the MEP, whereas NF-E2 contributes to later development of both lineages. This diagram does not reﬂect the dual roles of some factors in early and late stages of differentiation, as may occur with RUNX1.

development; GATA-2 by contrast contributes to the associated with moderate to severe thrombocytopenia in proliferation of multipotent progenitors (Cantor and the absence of anemia, but the more structurally Orkin, 2002). disruptive substitution D218Y resulted in severe throm- Experiments of nature have provided glimpses of the bocytopenia and anemia, again suggesting that mega- role of GATA-1 in human megakaryopoiesis. As many karyopoiesis is more sensitive than erythropoiesis to as four distinct biochemical classes of GATA-1 muta- diminished GATA-1 function (Freson et al., 2001, tion have been found in human X-linked disorders of 2002). The single kindred identified with a mutation megakaryocyte development (Liew et al., 2005). The causing GATA-1s surprisingly displayed macrocytic first class of mutations comprises N finger amino acid anemia with normal platelet counts; however, platelet substitutions which impair FOG1 binding, two exam- morphology and function were abnormal (Hollanda ples of which have been reported: V205M and G208S et al., 2006). A fascinating feature shared by various (Nichols et al., 2000; Mehaffey et al., 2001). The second kindreds with GATA-1 mutations is a role reversal in the class involves N-finger substitutions which preserve erythro-megakaryocytic divergence, the erythroblasts FOG1 binding but impair binding to a DNA GATC showing enlargement with multinucleation and the motif found within paired GATA target sites: examples megakaryocytes developing as small cells with hypolo- are R216Q and R216W (Yu et al., 2002; Phillips et al., bulated nuclei, ‘micromegakaryocytes’ (Nichols et al., 2007). The third class also affects the N finger but 2000; Freson et al., 2001, 2002; Hollanda et al., 2006; disrupts neither FOG1 nor DNA binding: D218G Phillips et al., 2007). (Freson et al., 2001; Liew et al., 2005). The biochemical Multiple themes emerge from a survey of the human defect in this third class thus remains mysterious. An and murine GATA-1 mutant phenotypes: (1) GATA-1 interaction potentially affected by this, or other, N-finger plays crucial roles in both megakaryopoiesis and erythro- substitutions could be the recently described binding of poiesis, roles that cannot be fulfilled by GATA-2 despite the mediator complex to the GATA-1 N finger (Stumpf structural and biochemical similarities; (2) distinct et al., 2006). The fourth class of hereditary GATA-1 structural elements within GATA-1 contribute to mutation causes congenital expression of GATA-1s, an distinct aspects of the differentiation program, a finding amino terminally truncated form lacking the first 83 confirmed in ex vivo megakaryocytic rescue assays (Kuhl amino acids (Hollanda et al., 2006). et al., 2005; Muntean and Crispino, 2005); (3) mega- The disease phenotypes reveal influences fromboth karyopoiesis depends more on GATA-1 dosage than the site of substitution and the severity of loss of does erythropoiesis, a finding also supported by ex vivo biochemical function. As an example of the latter rescue assays (Stachura et al., 2006); (4) proper influence, megakaryopoiesis displays greater sensitivity segregation of the ‘hypertrophic’ and ‘atrophic’ pheno- than erythropoiesis to decrements in GATA-1 binding types during erythromegakaryocytic divergence relies on to FOG1. In particular, the V205M mutant shows the activity of GATA-1. greater loss of FOG1 binding than G208S, and patients with V205M have severe anemia and thrombocytopenia, whereas those with G208S have only severe thrombocytopenia with no anemia. As an example of the site At the core of megakaryopoiesis: RUNX1 and familial influence, substitutions of R216 affecting DNA binding platelet disorder (FPD/AML) produce moderate thrombocytopenia associated with a thalassemic type of anemia (Raskind et al., 2000; Yu Several megakaryocytic promoters possess core binding et al., 2002; Phillips et al., 2007). The D218G mutation, factor (CBF) or RUNX sites consisting of TGT/cGGT affecting neither FOG1 nor DNA binding, has been (Elagib et al., 2003; Heller et al., 2005; Xu et al., 2006).

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6797 The CBF family of transcription factors comprises three to RUNX1 levels; GATA-1 haploinsufficiency, by homologous DNA binding proteins, RUNX1-3, and a comparison, causes no platelet abnormalities. The non-DNA binding component CBFb. The complex of basis for this sensitivity most likely resides in the RUNX1 with CBFb participates in programming dependency of target genes, such as PKCy and hematopoietic ontogeny and represents the most com- c-MPL, on critical thresholds of RUNX transcriptional mon mutational target in human acute leukemias activity. (Tracey and Speck, 2000). RUNX1 and CBFb retain high expression during megakaryocytic differentiation of MEP but undergo extinction during early phases of erythroid differentiation (Kundu et al., 2002; Elagib Role of Ets factors: Fli-1 and the Paris–Trousseau/ et al., 2003; Lorsbach et al., 2004). Mature megakar- Jacobsen syndrome (PTS) yocytes express strikingly high levels of RUNX1 and low levels of RUNX3 (Levanon et al., 2001). Inducible Ets factor binding sites, containing a core GGAA motif, inactivation in mice of either RUNX1 or CBFb causes occur in most megakaryocytic promoters, often adjacent rapid and profound disruption of megakaryocytic to GATA sites (Deveaux et al., 1996; Bastian et al., differentiation, with minimal impact on erythropoiesis 1999; Pang et al., 2006). Three Ets factors directly (Ichikawa et al., 2004; Growney et al., 2005; Kuo et al., contribute to murine megakaryocytic development: 2006; Putz et al., 2006). In addition, mice bearing Fli-1, GABPa, and TEL1. Deletion of TEL1 within hypomorphic CBFb alleles show defective megakaryo- erythroid and megakaryocytic cells in the GATA-1- poiesis, confirming the lineage to be sensitive to the CRE::TEL1f/f mouse strain causes a B50% drop in dosage of core binding factor (Talebian et al., 2007). platelet counts with no effect on hemoglobin levels. Despite the variety of RUNX1 mutations associated Marrows fromthis strain also show a 5-fold increase in with the autosomal dominant FPD/AML, most kin- megakaryocytic colony forming cells (CFU-Mk) (Hock dreds show similar clinical manifestations, in contrast to et al., 2004). Similarly, inducible deletion of TEL1 in the phenotypic–genotypic correlations seen with GATA-1 adult MX1-CRE::TEL1f/f mice triggers rapid and mutations. Thus FPD/AML appears to result from a dramatic drops in platelet counts, with no effects on hemo- simple haploinsufficient dosage effect, although a domi- globin levels (Hock et al., 2004). Diminished GABPa nant negative function has been suggested for some of the expression in the enhancer trapped GABPaTP/TP strain mutant RUNX1 alleles (Michaud et al., 2002; Matheny results in a B50% decline in fetal liver megakaryocytes, et al., 2007). A central clinical feature of this disorder as well as diminished megakaryocytic expression of consists of a bleeding tendency greatly out of proportion gpIIb and c-Mpl (Pang et al., 2006). FLI1 À/À mice to the mild thrombocytopenia (Dowton et al., 1985). manifest embryonic lethality with hemorrhage and Platelet abnormalities include diminished dense core diminished hematopoiesis. One group has reported granules and decreased cellular adenosine diphosphate megakaryocyte-selective hematopoietic abnormalities (ADP) levels (Buijs et al., 2001; Michaud et al., 2002). in fetal liver assays of colony formation and maturation Functionally, the platelets show marked aggregation (Hart et al., 2000), whereas another group has found defects in response to collagen, epinephrine and arachi- quantitative and qualitative defects in both erythroid donic acid but retain responsiveness to ristocetin (Ho and megakaryocytic lineages within fetal liver (Spyr- et al., 1996; Walker et al., 2002; Heller et al., 2005). opoulos et al., 2000). Thus while Fli-1 clearly con- Marrow megakaryocytic morphology has varied from tributes to normal megakaryocytic differentiation, its reportedly normal (Song et al., 1999) to abnormal with status as a lineage-selective transcription factor in immature-appearing megakaryocytes that are increased in erythro-megakaryocytic divergence remains uncertain. number (Walker et al., 2002; Heller et al., 2005). The Functional comparisons between Fli-1 and GABPa greatly increased incidence of acute myeloid leukemia have suggested a role for the former in activating late within affected family members may relate to abnormal- stage megakaryocytic markers such as gpIX and a role ities in hematopoietic stem cell (HSC) homeostasis for the latter in programming early markers such as associated with diminished RUNX1 function (Sun and gpIIb (Pang et al., 2006). Downing, 2004). Relevance of Fli-1 to human megakaryopoiesis has Causes for the platelet dysfunction in FPD/AML are been suggested fromstudies of a rare congenital most likely multiple. The diminished dense core granules disorder, the PTS. Two cases described in 1995 involved and ADP may explain the ‘aspirin like’ aggregation a mother and son with mild thrombocytopenia (Breton- defects consisting of an absent second wave in response Gorius et al., 1995). Most of their circulating platelets to epinephrine stimulation (Ho et al., 1996). In addition, had normal morphology, but B15% of the platelets Sun et al. (2004) have identified in a patient defective showed enlargement with a solitary giant red granule, activation of the gpIIbIIIa integrin complex and identified on electron microscopy (EM) as an abnormal diminished pleckstrin phosphorylation, both of which alpha granule. The mother’s marrow contained a 3-fold correlate with abnormally low platelet levels of the increase in megakaryocytes, many of which had micro- kinase PKCy. In another kindred, diminished platelet megakaryocytic morphology. EM on the marrow expression and signaling of c-Mpl was identified paradoxically revealed abnormally small alpha granules (Heller et al., 2005). FPD/AML thus provides an insight within megakaryocytes, as well as a subset of mega- into the unique sensitivity of human megakaryopoiesis karyocytes undergoing ‘lysis’, cytoplasmic zonation into

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6798 regions devoid of organelles and regions with clusters of Antagonists of megakaryopoiesis: c-Myb and p300 dilated vacuoles and caniliculi. Another subset of megakaryocytes displayed ultrastructural features associated Three independent studies have applied whole mouse with activated platelets. Fluorescence in situ hybridiza- randommutagenesis to identify the transcription factor tion analysis of chromosomes from both subjects c-Myb as a potent negative modulator of megakaryo- identified a deletion of the distal portion of 11q, with poiesis. Two of these studies employed large-scale a breakpoint at 11q23.3. Subsequent analysis of 14 chemical mutagenesis followed by screening of platelet patients with the related Jacobsen syndrome confirmed levels (Carpinelli et al., 2004; Sandberg et al., 2005). In terminal 11q deletions, which in all cases led to both studies, hypomorphic alleles encoding transcrip- hemizygous deletion of FLI1 (Hart et al., 2000). tionally compromised c-Myb mutants, when bred to Further support for involvement of Fli-1 came from homozygosity, caused marked elevation of circulating rescue studies in which FLI1 transduction of PTS CD34 þ platelet levels and tissue megakaryocytes, confirming an cells restored several aspects of megakaryocytic differ- earlier association in C-MYBÀ/loxP mice between dimin- entiation, including levels of CD41, CD42 and vWF, ished c-Myb levels and enhanced megakaryopoiesis as well as polyploidization and proplatelet formation (Emambokus et al., 2003). In a third study, random (Raslova et al., 2004). Characterization of megakaryo- transgene insertion upstreamof C-MYB eliminated poiesis fromPTS CD34 þ cells also established a basis expression specifically within the MEP compartment for hemizygous deletion of FLI1 affecting development (Mukai et al., 2006), permitting assessment of the role of of a subset of megakaryocytes and platelets. At early c-Myb in erythromegakaryocytic lineage divergence. stages of human megakaryocytic differentiation, FLI1 Mice homozygous for the transgene insertion (Tg) normally undergoes a phase of monoallelic expression, displayed an increase in marrow CFU-Mk counter- and during this phase hemizygous deletion of PTS yields balanced by a decrease on erythroid colony forming subsets of megakaryocytes with either normal or absent units, suggesting an influence of c-Myb levels on MEP Fli-1 expression (Raslova et al., 2004). This mono- fate. Further evidence derived fromanalysis of highly allelic model thus explains both the pathogenecity of purified primary MEP cultured on OP9 stromal cells. haploinsufficient FLI1 and the discrete populations of Wild-type MEP gave rise to predominantly erythroid normal and abnormal megakaryocytes/platelets in PTS cells, while Tg homozygous MEP yielded mostly patients. megakaryocytes, an effect reversible by retroviral The ETS1 gene also undergoes deletion in PTS and transduction of C-MYB. potentially could contribute to megakaryopoiesis. In Transcriptional activation of target genes by c-Myb cultures of human CD34 þ cells, Ets-1 undergoes depends on the recruitment of the p300 coactivator via upregulation in megakaryocytic differentiation and the latter’s KIX domain. Notably, mice engineered with downregulation in erythroid differentiation, albeit in a KIX-defective p300 phenocopy the c-Myb deficiency, delayed manner (Lulli et al., 2006); enforced expression showing striking increases in circulating platelets and of Ets-1 in these cells enhances megakaryocytic and marrow megakaryocytes (Kasper et al., 2002). Genetic inhibits erythroid differentiation. Arguing against a direct interaction of p300 and c-Myb in this pathway was role for Ets-1 in PTS are the lack of megakaryocytic confirmed in P300/C-MYB compound heterozygous abnormalities in ETS1 À/À mice and a low expression mutant mice which showed platelet elevations analogous level in megakaryocytes, relative to Fli-1 (Raslova et al., to these seen in both homozygous mutants. Thus, c-Myb 2004). Nevertheless, the possibility remains that the most likely functions at the level of the MEP as a hemizygous deletion of ETS1 could underlie some of transcriptional activator that either drives erythroid the megakaryocytic abnormalities in PTS, particularly lineage commitment at the expense of megakaryopoiesis since FLI1À/À mice do not faithfully recapitulate all of or primarily blocks megakaryopoiesis and secondarily the abnormalities. shunts cells into the erythroid pathway.

Table 1 Human platelet disorders with transcription factor mutations Gene Transmission Disease Mutation Platelets Megakaryocytes

RUNX1 Autosomal FPD/AML Loss of function, runt Slightly decreased, nl. size, Increased, small, dominant domain or TAD defective aggregation hypolobulated, decreased granules GATA-1 X-linked XLT Loss of function, N-ﬁnger Mod.-markedly decreased, Increased, small, (three biochemical classes) enlarged, defective hypolobulated, decreased aggregation granules GATA-1 X-linked XL anemia Loss of function, GATA-1s Nl number, defective Increased, small, leukopenia expression aggregation hypolobulated, decreased platelet granules FLI1 Autosomal PTS/Jacobsen Hemizygous deletion Slightly decreased, subset Increased, small, dominant enlarged, subset with giant hypolobulated, small alpha granule and defective apha granules, subset activation showing ‘lysis’

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6799 Table 2 Mouse strains with GATA-1, RUNX1 or FLI1 mutations Strain Overall phenotype Platelets Megakaryocytes

GATA-1À/À Embryonic lethal E10.5–11.5 Not Not evaluable evaluable RUNX1À/À Embryonic lethal E11.5–12.5 Not Not evaluable evaluable FLI1À/À Embryonic lethal E11.5–12.5 Not Small, hypolobulated, increased colony formation, poorly evaluable formed demarcation membranes, decreased alpha granules GATA-1Lo (DneoDHS) Viable adult, shortened lifespan Sixfold Small, hypolobulated, decreased ploidy, hyperproliferative, decrease poorly formed demarcation membranes RUNX1f/fHMx1-CRE Viable adult, shortened lifespan Fivefold Small, hypolobulated, decreased ploidy, hyperproliferative, (pIpC induced deletion) decrease poorly formed demarcation membranes GATA-1S Normal adult Transient expansion Normal Hyperproliferative fetal megakaryocytes of embryonic megakaryocyte progenitors

Programmatic integration: discerning a megakaryocytic transcriptional network Fli-1 As emphasized in the Abstract, the decision to pursue a FOG1 megakaryocytic fate does not rest in the hands of any GATA-1 DBD single transcription factor or family but rather results Megakaryocytic from a conversation among multiple families. Tables 1 promoter and 2 illustrate points of phenotypic overlap among the FOG1 DBD Megakaryocyte Fli-1 several human diseases and mouse models discussed in GATA-1 this review, indicating a convergence of at least three Erythroid different transcription factor families on key pathways promoter within megakaryocytic growth and development. At a mechanistic level, this overlap highlights a number of mutually regulated target genes that contribute to the FOG1 EKLF megakaryocytic program. Presented below are three GATA-1 DBD examples of transcriptional cross talk likely to be Erythroid important in megakaryopoiesis. promoter GATA-1- and RUNX1-deﬁcient organisms share FOG1

Erythroblast DBD several megakaryocytic abnormalities in common. Both EKLF show thrombocytopenia with increased marrow mega- GATA-1 karyocytes. In both, the megakaryocytes display im- Megakaryocytic promoter mature morphology, diminished ploidy, increased colony forming capabilities, increased proliferation Figure 2 A model for programming of erythro-megakaryocytic and defective formation of demarcation membranes divergence, depicting contributions of lineage-shared factors (for example, GATA-1 and FOG1), lineage-selective factors (for (Vyas et al., 1999; Nichols et al., 2000; Walker et al., example, Fli-1 and EKLF) and promoter binding sites. This model 2002; Ichikawa et al., 2004; Growney et al., 2005; Heller accounts for megakaryocyte GATA-1/FOG1 complexes activating et al., 2005; Hollanda et al., 2006). Shared target genes megakaryocytic promoters and repressing erythroid promoters, include C-MPL, GPIba, JAK2 and GPIIb (Vyas et al., with erythroid GATA-1/FOG1 complexes activating erythroid 1999; Heller et al., 2005; Kuhl et al., 2005; Muntean and promoters and repressing megakaryocytic promoters. In this model, the transcriptional output is determined by the mode of recruitment Crispino, 2005; Liu et al., 2006; Xu et al., 2006). of the lineage-selective factors. If the lineage-selective factor is Biochemical studies have demonstrated physical inter- recruited into the complexes by engagement of its DNA binding action and functional cooperation of these two factors domain (DBD) with cognate promoter sites, then cross talk between in transcriptional activation of megakaryocytic promo- transcriptional activation domains promotes activation of target genes. If the lineage-selective factor is recruited by protein–protein ters (Elagib et al., 2003; Xu et al., 2006). Their crosstalk interactions, with no binding sites present within the promoter, then reflects a highly conserved pathway in evolution, with the exposed DBD may exert a repressive effect. parallels found in the Drosophila GATA and RUNX counterparts, Serpent and Lozenge, which cooperate in the programming of hematopoietic crystal cell develop- In vivo deficiency of Fli-1 also results in expansion of ment (Fossett et al., 2003; Waltzer et al., 2003). The immature, hyperproliferative megakaryocytes with Drosophila ortholog of FOG1, U-shaped, negatively granule and demarcation membrane defects (Breton- regulates this program(Fossett et al., 2003; Waltzer Gorius et al., 1995; Hart et al., 2000). Target genes of et al., 2003), raising interesting possibilities for the Fli-1 and GABPa that overlap with those of GATA-1 influence of FOG1 on the megakaryocytic cooperation and RUNX1 include GPIba, c-MPL and GPIIb (Pang of GATA-1 and RUNX1. et al., 2006). Biochemical analyses have confirmed

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6800 physical and functional interaction of GATA-1 with megakaryopoiesis, GATA-1 participates in complexes Fli-1 (Starck et al., 2003). Interestingly, Fli-1 and GABPa, that promote megakaryopoiesis and also repress ery- but not the Ets factor PU.1, exert a positive effect on thropoiesis. Conversely, in erythropoiesis, GATA-1 FOG1 complexed with GATA-1, promoting activation complexes promote erythropoiesis and repress mega- of the GPIIb promoter (Wang et al., 2002; Pang et al., karyopoiesis. It is likely that the complexes promoting 2006). Thus, whether FOG1 functions as a corepressor megakaryopoiesis are similar to, or the same as, those or coactivator for GATA-1 may depend on the that repress erythropoiesis. For example Fli-1 binding to transcription factors and promoter milieu with which a megakaryocytic promoter may engage neighboring GATA-1 associates. GATA-1/FOG1 to potentiate activation, while Fli-1 The ﬁnal example of crosstalk illustrates how three recruitment to an erythroid promoter through tethering distinct transcription factors may come together to coordi- to EKLF or GATA-1 may exert a repressive function nate divergence of the megakaryocytic and erythroid through its exposed DNA binding domain (see Figure 2). lineages. In particular, Starck et al. (2003) identify Cracking the code for erythro-megakaryocytic lineage GATA-1 interactions with Fli-1 in megakaryocytic divergence will require detailed understanding of the complexes and with EKLF in erythroid complexes. As web of functional, physical and signaling connectivity an additional dimension, Fli-1 and EKLF directly among the known players, as well as regulators yet to be interact and cross-inhibit one another. Interaction with named. EKLF causes Fli-1 to expose a surface normally engaged in DNA binding, resulting in putative core- Acknowledgements pressor recruitment. This type of binary model integrates several key We thank Kamal Elagib and Ivo Mihaylov for ongoing features of megakaryocytic transcriptional regulation. discussions on megakaryocytic transcriptional regulation. The Importantly, GATA-1 functions as a central broker study has been supported by NIH Grants CA100057 and in both megakaryopoiesis and erythropoiesis. In CA93735.

References

Adolfsson J, Mansson R, Buza-Vidas N, Hultquist A, Liuba K, Emambokus N, Vegiopoulos A, Harman B, Jenkinson E, Anderson G, Jensen CT et al. (2005). Identiﬁcation of FLt3+ lympho-myeloid Frampton J. (2003). Progression through key stages of haemopoiesis stemcells lacking erythro-megakaryocytic potential: a revised road is dependent on distinct threshold levels of c-Myb. EMBO J 22: map for adult blood lineage commitment. Cell 121: 295–306. 4478–4488. Akashi K, Traver D, Miyamoto T, Weissman IL. (2000). A clonogenic Fossett N, Hyman K, Gajewski K, Orkin SH, Schulz RA. (2003). common myeloid progenitor that gives rise to all myeloid lineages. Combinatorial interactions of Serpent, Lozenge, and U-shaped Nature 404: 193–197. regulate crystal cell lineage commitment during Drosophila hema- Bastian LS, Kwiatkowski BA, Breininger J, Danner S, Roth G. (1999). topoiesis. Proc Natl Acad Sci USA 100: 11451–11456. Regulation of the megakaryocytic glycoprotein IX promoter by the Freson K, Devriendt K, Matthijs G, Van Hoof A, De Vos R, Thys C oncogenic ets transcription factor Fli-1. Blood 93: 2637–2644. et al. (2001). Platelet characteristics in patients with X-linked Breton-Gorius J, Favier R, Guichard J, Cherif D, Berger R, Debili N macrothrombocytopenia because of a novel GATA1 mutation. et al. (1995). A new congenital dysmegakaryopoietic thrombocyto- Blood 98: 85–92. penia (Paris–Trousseau) associated with giant platelet a-granules Freson K, Matthijs G, Thys C, Marien P, Hoylaerts MF, Vermylen J and chromosome 11 deletion at 11q23. Blood 85: 1805–1814. et al. (2002). Different substitutions at residue D218 of the Buijs A, Poddighe P, van Wijk R, van Solinge W, Borst E, Verdonck L X-linked transcription factor GATA-1 lead to altered clinical et al. (2001). A novel CBFA2 single-nucleotide mutation in familial severity of macrothrombocytopenia and anemia and are associ- platelet disorder with propensity to develop myeloid malignancies. ated with variable skewed X inactivation. Hum Mol Genet 11: Blood 98: 2856–2858. 147–152. Cantor AB, Orkin SH. (2002). Transcriptional regulation of erythro- Growney JD, Shigematsu H, Li Z, Lee BH, Adelsperger J, Rowan R poiesis: an affair involving multiple partners. Oncogene 21: et al. (2005). Loss of Runx1 perturbs adult hematopoiesis and is 3368–3376. associated with a myeloproliferative phenotype. Blood 106: 494–504. Carpinelli MR, Hilton DJ, Metcalf D, Antonchuk JL, Hyland CD, Hart A, Melet F, Grossfeld P, Chien K, Jones C, Tunnacliffe A et al. Mifsud SL et al. (2004). Suppressor screen in MplÀ/À mice: c-Myb (2000). Fli-1 is required for murine vascular and megakaryocytic mutation causes supraphysiologic production of platelets in the development and is hemizygously deleted in patients with thrombo- absence of thrombopoietin signaling. Proc Natl Acad Sci USA 101: cytopenia. Immunity 13: 167–177. 6553–6558. Heller PG, Glembotsky AC, Gandhi MJ, Cummings CL, Pirola CJ, Debili N, Coulombel L, Croisille L, Katz A, Guichard J, Marta RF et al. (2005). Low Mpl receptor expression in a pedigree Breton-Gorius J et al. (1996). Characterization of a bipotent with familial platelet disorder with predisposition to acute erythro-megakaryocytic progenitor in human bone marrow. Blood myelogenous leukemia and a novel AML1 mutation. Blood 105: 88: 1284–1296. 4664–4670. Deveaux S, Filipe A, Lemarchandel V, Ghysdael J, Romeo P-H, Ho CY, Otterud B, Legare RD, Varvil T, Saxena R, DeHart DB et al. Mignotte V. (1996). Analysis of the thrombopoietin receptor (mpl) (1996). Linkage of a familial platelet disorder with a propensity to promoter implicates GATA and Ets proteins in the coregulation of develop myeloid malignancies to human chromosome 21q22.1-22.2. megakaryocyte-speciﬁc genes. Blood 87: 4678–4685. Blood 87: 5218–5224. Dowton SB, Beardsley D, Jamison D, Blattner S, Li FP. (1985). Hock H, Meade E, Medeiros S, Schindler JW, Valk PJM, Fujiwara Y Studies of a familial platelet disorder. Blood 65: 557–563. et al. (2004). Tel/Etv6 is an essential and selective regulator of adult Elagib KE, Racke FK, Mogass M, Khetawat R, Delehanty LL, hematopoietic stem cell survival. Genes Dev 18: 2336–2341. Goldfarb AN. (2003). RUNX-1 and GATA-1 coexpression and co- Hollanda LM, Lima CSP, Cunha AF, Albuquerque DM, Vassallo J, operation in megakaryocytic differentiation. Blood 101: 4333–4341. Ozelo MC et al. (2006). An inherited mutation leading to

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6801 production of only the short isoformof GATA-1 is associated with Nichols KE, Crispino JD, Poncz M, White JG, Orkin SH, Maris JM impaired erythropoiesis. Nat Genet 38: 807–812. et al. (2000). Familial dyserythropoietic anaemia and thrombocy- Ichikawa M, Asai T, Saito T, Yamamoto G, Seo S, Yamazaki I et al. topenia due to an inherited mutation in GATA-1. Nature Genet 24: (2004). AML1 is required for megakaryocytic maturation and 266–270. lymphocytic differentiation, but not for maintenance of Pang L, Xue H-H, Szalai G, Wang X, Wang Y, Watson DK et al. hematopoietic stem cells in adult hematopoiesis. Nature Med 10: (2006). Maturation stage-specific regulation of megakaryopoiesis by 299–304. pointed-domain Ets proteins. Blood 108: 2198–2206. Kasper LH, Boussouar F, Ney PA, Jackson CW, Rehg J, van Deursen Phillips JD, Steensma DP, Pulsipher MA, Spangrude GJ, Kushner JP. JM et al. (2002). A transcription-factor-binding surface of (2007). Congenital erythropoietic porphyria due to a mutation in coactivator p300 is required for haematopoiesis. Nature 419: GATA1: the first trans-acting mutation causative for a human 738–743. porphyria. Blood 109: 2618–2621. Kuhl C, Atzberger A, Iborra F, Nieswandt B, Porcher C, Vyas P. Putz G, Rosner A, Nuesslein I, Schmitz N, Buchholz F. (2006). AML1 (2005). GATA1-mediated megakaryocyte differentiation and deletion in adult mice causes splenomegaly and lymphomas. growth control can be uncoupled and mapped to different domains Oncogene 25: 929–939. in GATA1. Mol Cell Biol 25: 8592–8606. Raskind WH, Niakan KK, Wolff J, Matsushita M, Vaughan T, Kundu M, Chen A, Anderson S, Kirby M, Xu L, Castilla LH et al. Stamatoyannopoulos G et al. (2000). Mapping of a syndrome of X- (2002). Role of Cbfb in hematopoiesis and perturbations resulting linked thrombocytopenia with thalassemia to band Xp11–12: fromexpression of the leukemogenic fusion gene Cbfb-MYH11. further evidence of genetic heterogeneity of X-linked thrombocyto- Blood 100: 2449–2456. penia. Blood 95: 2262–2268. Kuo Y-H, Landrette SF, Heilman SA, Perrat PN, Garrett L, Liu PP Raslova H, Komura E, Le Couedic JP, Larbret F, Debili N, Feunteun et al. (2006). Cbf beta-SMMHC induces distinct abnormal myeloid J et al. (2004). FLI1 monoallelic expression combined with its progenitors able to develop acute myeloid leukemia. Cancer Cell hemizygous loss underlies Paris–Trousseau/Jacobsen thrombocyto- 9: 57–68. penia. J Clin Invest 114: 77–84. Levanon D, Brenner O, Negreanu V, Bettoun D, Woolf E, EilamR Sandberg ML, Sutton SE, Pletcher MT, Wiltshire T, Tarantino LM, et al. (2001). Spatial and temporal expression pattern of Runx3 Hogenesch JB et al. (2005). c-Myb and p300 regulate hematopoietic (Aml2) and Runx1 (Aml1) indicates non-redundant functions stemcell proliferation and differentiation. Dev Cell 8: 153–166. during mouse embryogenesis. Mech Dev 109: 413–417. Song W-J, Sullivan MG, Legare RD, Hutchings S, Tan X, Kufrin D Liew CK, Simpson RJY, Kwan AHY, Crofts LA, Loughlin FE, et al. (1999). Haploinsufficiency of CBFA2 causes familial thrombo- Matthews JM et al. (2005). Zinc fingers as protein recognition cytopenia with propensity to develop acute myelogenous leukemia. motifs: structural basis for the GATA-1/Friend of GATA interac- Nat Genet 23: 166–175. tion. Proc Natl Acad Sci USA 102: 583–588. Spyropoulos DD, Pharr PN, Lavenburg KR, Jackers P, Papas TS, Liu Y, Cheney MD, Gaudet JJ, Chruszcz M, Lukasik SM, Sugiyama Ogawa M et al. (2000). Hemorrhage, impaired hematopoiesis, and D et al. (2006). The tetramer structure of the Nervy homology two lethality in mouse embryos carrying a targeted disruption of the Fli1 domain, NHR2, is critical for AML1/ETO’s activity. Cancer Cell 9: transcription factor. Mol Cell Biol 20: 5643–5652. 249–260. Stachura DL, Chou ST, Weiss MJ. (2006). Early block to erythrome- Lorsbach RB, Moore J, Ang SO, Sun W, Lenny N, Downing JR. gakaryocytic development conferred by loss of transcription factor (2004). Role of RUNX1 in adult hematopoiesis: analysis of GATA-1. Blood 107: 87–97. RUNX1-IRES-GFP knock-in mice reveals differential lineage Starck J, Cohet N, Gonnet C, Sarrazin S, Doubeikovskaia Z, expression. Blood 103: 2522–2529. Doubeikovski A et al. (2003). Functional cross-antagonism Ludlow LB, Schick BP, Budarf ML, Driscoll DA, Zackai EH, Cohen A between transcription factors FLI-1 and EKLF. Mol Cell Biol 23: et al. (1996). Identification of a mutation in a GATA binding site of 1390–1402. the platelet glycoprotein Ibb promoter resulting in the Bernard– Stumpf M, Waskow C, Krotschel M, van Essen D, Rodriguez P, Soulier syndrome. J Biol Chem 271: 22076–22080. Zhang X et al. (2006). The mediator complex functions as a Lulli V, Romania P, Morsilli O, Gabbianelli M, Pagliuca A, Mazzeo S coactivator for GATA-1 in erythropoiesis via subunit Med1/ et al. (2006). Overexpression of Ets-1 in human hematopoietic TRAP220. Proc Natl Acad Sci USA 103: 18504–18509. progenitor cells blocks erythroid and promotes megakaryocytic Sun L, Mao G, Rao AK. (2004). Association of CBFA2 mutation with differentiation. Cell Death Diff 13: 1064–1074. decreased platelet PKC-theta and impaired receptor-mediated Manz MG, Miyamoto T, Akashi K, Weissman IL. (2002). Prospective activation of GPIIb-IIIa and pleckstrin phosphorylation: proteins isolation of human clonogenic common myeloid progenitors. Proc regulated by CBFA2 play a role in GPIIb-IIIa activation. Blood 103: Natl Acad Sci USA 99: 11872–11877. 948–954. Matheny CJ, Speck ME, Cushing PR, Zhou Y, Corpora T, Regan M Sun W, Downing JR. (2004). Haploinsufficiency of AML1 results et al. (2007). Disease mutations in RUNX1 and RUNX2 create non- in a decrease in the number of LTR-HSC while simultaneously functional, dominant-negative, or hypomorphic alleles. EMBO J inducing an increase in more mature progenitors. Blood 104: 26: 1163–1175. 3565–3572. Mehaffey MG, Newton AL, Gandhi MJ, Crossley M, Drachman JG. Talebian L, Li Z, Guo Y, Gaudet JJ, Speck ME, Sugiyama D et al. (2001). X-linked thrombocytopenia caused by a novel mutation of (2007). T-lymphoid, megakaryocyte, and granulocyte development GATA-1. Blood 98: 2681–2688. are sensitive to decreases in CBFb dosage. Blood 109: 11–21. Michaud J, Wu F, Osato M, Cottles GM, Yanagida M, Asou N et al. Tracey WD, Speck NA. (2000). Potential roles for RUNX1 and its (2002). In vitro analyses of known and novel RUNX1/AML1 orthologs in determining hematopoietic cell fate. Sem Cell Dev Biol mutations in dominant familial platelet disorder with predisposition 11: 337–342. to acute myelogenous leukemia: implications for mechanisms of Vyas P, Ault K, Jackson CW, Orkin SH, Shivdasani RA. (1999). pathogenesis. Blood 99: 1364–1372. Consequences of GATA-1 deficiency in megakaryocytes and Mukai HY, Motohashi H, Ohneda O, Suzuki N, Nagano M, platelets. Blood 93: 2867–2875. Yamamoto M. (2006). Transgene insertion in proximity to the Walker LC, Stevens J, Campbell H, Corbett R, Spearing R, Heaton D c-myb gene disrupts erythroid-megakaryocytic lineage bifurcation. et al. (2002). A novel inherited mutation of the transcription factor Mol Cell Biol 26: 7953–7965. RUNX1 causes thrombocytopenia and may predispose to acute Muntean AG, Crispino JD. (2005). Differential requirements for the myeloid leukaemia. Br J Haematol 117: 878–881. activation domain and FOG-interaction surface of GATA-1 Waltzer L, Ferjoux G, Bataille L, Haenlin M. (2003). Cooperation in megakaryocyte gene expression and development. Blood 106: between the GATA and RUNX factors Serpent and Lozenge during 1223–1231. Drosophila hematopoiesis. EMBO J 22: 6516–6525.

Oncogene Transcription in megakaryopoiesis AN Goldfarb 6802 Wang X, Crispino JD, Letting DL, Nakazawa M, Poncz M, Blobel with RUNX1 in acute megakaryoblastic leukemia accompanying GA. (2002). Control of megakaryocyte-specific gene expression by Down syndrome. Leukemia 20: 1002–1008. GATA-1 and FOG-1: role of Ets transcription factors. EMBO J 21: Yu C, Niakan KK, Matsushita M, Stamatoyannopoulos G, Orkin SH, 5225–5234. Raskind WH. (2002). X-linked thrombocytopenia with thalassemia Xu G, Kanezaki R, Toki T, Watanabe S, Takahashi Y, Terui K et al. from a mutation in the amino finger of GATA-1 affecting DNA (2006). Physical association of the patient-specific GATA1 mutants binding rather than FOG-1 interaction. Blood 100: 2040–2045.

Oncogene