GATA1 in Normal and Pathologic Megakaryopoiesis and Platelet Development

Abstract

GATA1 is a highly conserved hematopoietic transcription factor (TF), essential for normal erythropoiesis and megakaryopoiesis, that encodes a full-length, predominant isoform and an amino (N) terminus-truncated isoform GATA1s. It is consistently expressed throughout megakaryocyte development and interacts with its target genes either independently or in association with binding partners such as FOG1 (friend of GATA1). While the N-terminus and zinc finger have classically been demonstrated to be necessary for the normal regulation of platelet-specific genes, murine models, cell-line studies, and human case reports indicate that the carboxy-terminal activation domain and zinc finger also play key roles in precisely controlling megakaryocyte growth, proliferation, and maturation. Murine models have shown that disruptions to GATA1 increase the proliferation of immature megakaryocytes with abnormal architecture and impaired terminal differentiation into platelets. In humans, germline GATA1 mutations result in variable cytopenias, including macrothrombocytopenia with abnormal platelet aggregation and excessive bleeding tendencies, while acquired GATA1s mutations in individuals with trisomy 21 (T21) result in transient abnormal myelopoiesis (TAM) and myeloid leukemia of Down syndrome (ML-DS) arising from a megakaryocyte-erythroid progenitor (MEP). Taken together, GATA1 plays a key role in regulating megakaryocyte differentiation, maturation, and proliferative capacity. As sequencing and proteomic technologies expand, additional GATA1 mutations and regulatory mechanisms contributing to human diseases of megakaryocytes and platelets are likely to be revealed.

12.1. Overview

GATA1 (GATA-binding protein 1) is located on the X chromosome (Xp21–11) and encodes a zinc finger (ZF) transcription factor (TF) essential for normal erythrocyte, megakaryocyte, mast cell, basophil, and eosinophil development, as well as for Sertoli cells in the testis. It is widely conserved between lower vertebrates and mammals, with homology between the human, chicken, mouse, rat, cow, sheep, turtle, and zebrafish forms [1, 2]. In humans, the GATA1 protein contains a transactivation domain (TAD) in each of the amino and carboxy termini (N-TAD and C-TAD, respectively) and two ZFs that are identified by their relative proximity to either terminus (Fig. 12.1). In reporter assays, both the N-TAD and C-TAD function to transcriptionally regulate target genes and have both unique and redundant functions [3, 4]. The N-terminus zinc finger (N-ZF) stabilizes GATA1-DNA-binding interactions, specifically binds to the palindromic (A/T)GATA(A/G):GAT motif [5, 6], and associates with the cofactor FOG1 (friend of GATA1), while the C-terminus zinc finger (C-ZF) facilitates GATA1 binding to its single (A/T)GATA(A/G), palindromic (A/T) GATA(A/G):GAT, and complex inverted or repeated double/tandem (A/T)GATA(A/G) target motifs [7](Fig. 12.2).

Fig. 12.1.

Schematic of human GATA1 gene and protein isoforms (wild-type GATA1 and GATA1s). N-TAD amino-terminus activation domain, C-TAD carboxy-terminus activation domain, N-ZF amino-terminus zinc finger, C-ZF carboxy-terminus zinc finger

Fig. 12.2.

GATA1-binding motifs

Human GATA1 contains a noncoding first exon and five coding exons. In contrast, the murine Gata1 ortholog contains two alternative noncoding first exons, the distal testis promoter/first exon (IT) for expression in Sertoli cells and the proximal erythroid promoter/first exon (IE) for expression in hematopoietic cells, followed by five coding exons (Fig. 12.3). Gata1 also contains 3 DNase I hypersensitivity regions—HS I, HS II, and HS III—that surround IE and are largely conserved in humans [6, 8–12]. HS I, also referred to as G1HE (Gata1 hematopoietic enhancer), is an enhancer region located 3.9–2.6 kb upstream of the IE promoter. It contains a composite GATA-E-box element and functions in the activation and maintenance of Gata1 expression in hematopoietic cells [8, 10, 13]. The entire HS I sequence is essential for normal megakaryopoiesis, while only the 5′ region is required for erythropoiesis [12, 14]. HS II encompasses IE and corresponds to the human GATA1 promoter; it contains a double GATA motif and at least two CACCC elements just proximal to IE [6, 9, 10, 15]. The double GATA motif and at least one of the CACCC elements interacts with the GATA-E-box element of HS I in the regulation of GATA1 expression [9, 10, 12]. HS III corresponds to intron 1.

Fig. 12.3.

Schematic of murine Gata1 gene. G1HE Gata1 hematopoietic enhancer, G1HRD Gata1 hematopoietic regulatory domain, IT testis promoter/first exon, IE erythroid promoter/first exon, HS hypersensitivity region

There are two isoforms of GATA1: the predominant full-length 42.7-kDa protein is encoded using the canonical start site in exon 2, while the shortened ~40-kDa protein GATA1s (GATA1 short) is missing the first 83 amino acids, including the N-TAD, due to alternative mRNA splicing and use of an alternative start codon in exon 3 (Fig. 12.1). Multiple studies have demonstrated that GATA1s bind similar targets as the full-length isoform but predictably lack transactivation activity [3, 5, 16, 17]. All humans naturally produce a small amount of GATA1s, but its role in normal hematopoiesis, if any, remains unclear. However, germline mutations resulting in the exclusive production of GATA1s causes thrombocytopenia with variable degrees of anemia and neutropenia [18–21], and somatic GATA1s mutations in individuals with Down syndrome (DS), or trisomy 21 (T21), are associated with a leukemia-like condition known as transient abnormal myelopoiesis (TAM) and myeloid leukemia of Down syndrome (ML-DS) [22, 23]. The mechanisms by which GATA1s exerts its phenotypes remain an active area of study.

12.2. GATA1 Regulation in Megakaryopoiesis

GATA1 is expressed in hematopoietic stem cells (HSCs), and its levels gradually increase as differentiation progresses through the common myeloid progenitor (CMP), megakaryocyte-erythroid progenitor (MEP), and then proerythroblasts and megakaryocytes; it is not expressed in the common lymphoid progenitor (CLP) or granulocyte-macrophage progenitor (GMP) [24]. Hematopoiesis therefore requires intricate coordination of transcriptional, posttranslational, and functional regulation of GATA1.

The transcriptional regulation of GATA1 is primarily driven by the three DNase hypersensitivity regions, or cis elements, that are located within an 8.5-kb region known as the GATA1 hematopoietic regulatory domain (G1HRD) [25, 26]. HS I, the double GATA motif, and a CACCC box are necessary and sufficient for GATA1 expression in hematopoietic cells, and the relative orientation of these three elements appears critical [13]. The intervening sequences between the three elements also plays a role in Gata1 expression by directing methylation of the locus by DNA methyltransferase 1 (DNMT1); this methylation is critical to the suppression of aberrant Gata1 expression in HSCs [27]. It has been suggested that methylation of the region encompassing the G1HRD prevents GATA2 (GATA-binding protein 2) from binding and thereby activating Gata1 expression in HSCs, and that methylation decreases with the initiation of hematopoietic differentiation [27]. A cell line model similarly found that GATA1 expression in hematopoietic progenitors is regulated by methylation of the cis elements, although this study suggested that GATA2 mediates GATA1 expression through demethylation of H3K4 [28].

In a murine megakaryocyte model, GATA1 binds to the GATA motif of the HS I GATA-E-box element, suggestive of a positive feedback loop that was first studied in an erythroid model [29] and also demonstrated in zebrafish [30, 31]. The loss of Gata1 mRNA levels in megakaryocytes derived from mice lacking HS I also suggests that this positive feedback loop regulates the platelet lineage [32]. GATA1 may facilitate hyperacetylation of histone H3 in this region and thus enhance transcription [33]. SCL/TAL1, a TF that also interacts with GATA1 alone and with other binding partners [34], appears to bind the E-box of the HS I element via other proteins such as LMO2, although it is not clear whether these other binding partners are the members of the pentameric complex vs. other proteins [12, 13] and whether SCL/TAL1 is required for Gata1 transcription [13, 33, 35].

GATA2 has been shown to play a role in GATA1 regulation in a number of contexts but appears to be nonessential. Cell line studies investigating the megakaryocyte, erythrocyte, and myeloid lineages have diverged on whether GATA1 expression can be induced by forced expression of GATA2 [36–40]. GATA1 and GATA2 share redundant functions in GATA1 regulation, but GATA2’s role is more prominent in early hematopoietic development than after lineage commitment [24]. This is consistent with the unique interaction between GATA2 and methylation of the G1HRD in regulating GATA1 expression in HSCs [27, 28]; although GATA2 regulation of GATA1 expression has been primarily studied in the context of erythropoiesis, it is likely that differential methylation also marks the transition from HSC to megakaryocyte.

ZBP89 (zinc finger DNA-binding protein 89), a Krüppel-type zinc finger TF that has been primarily studied as a cofactor for the GATA1-FOG1 complex, also regulates GATA1 expression [27, 33, 41, 42]. It is enriched in the G1HRD in murine cell lines [33, 41], suggesting that it directly controls transcription although its role specifically in megakaryocytes is less clear [33].

12.3. GATA1-Binding Partners and Targets

GATA1 plays a distinct role in each of the hematopoietic lineages for which it is critical. In megakaryocytes, GATA1 expression remains consistent throughout development, and both the N-terminus and the N-ZF are required for normal megakaryopoiesis [32]. GATA1 may interact with its megakaryocyte target genes either independently or as part of a protein complex (Fig. 12.4).

Fig. 12.4.

Schematic of different patterns of GATA1 occupancy of its target genes. Unlabeled proteins are representative of a range of binding partners

12.3.1. FOG1-GATA1 Complexes

The most well studied of GATA1-binding partners is FOG1 (friend of GATA1), a ZF protein that regulates erythropoiesis and megakaryopoiesis both in association with and independently of GATA1 [43–50]. Depending on the gene and lineage context, FOG1 acts to activate or repress GATA1-mediated gene expression [43, 51]. Megakaryocyte genes that are regulated via GATA1-FOG1 complexes carry tandem GATA- and ETS (erythroblastosis virus 26 transformation-specific)-binding sites in their regulatory regions; ETS domains are characterized by a core GGA(A/T) sequence [52]. The GATA1 N-ZF interacts directly with FOG1, and this complex in turn binds to their target sites via GATA1 [47, 50].

Several studies have investigated mutations affecting the N-ZF, demonstrating that mutations such as V205M, V205G, G208E, and others are sufficient to disrupt the GATA1-FOG1 interaction in megakaryocytes and thus binding to target genes [32, 45–48, 50]. Interestingly, a study in murine G1ME (Gata1⁻ megakaryocyte erythroid) cells showed that the N-ZF mutant GATA1^V205G binds to chromatin much less selectively than wild-type (WT) GATA1, suggesting that FOG1 regulates GATA1 chromatin occupancy even though FOG1 itself does not bind chromatin directly. Genes relevant to megakaryocyte development and function that have been specifically shown to be dysregulated in the setting of N-ZF mutations include p45 Nfe2 (nuclear factor, erythroid 2), Gp1ba (glycoprotein 1b subunit alpha), c-Mpl (myeloproliferative leukemia protein), Gpix (glycoprotein IX), and Jak2 (Janus kinase 2) [47, 48]. Murine studies have also shown that the GATA1-FOG1 interaction is dispensable for early megakaryocyte development but is required for later megakaryocyte maturation, likely due to partial compensation by GATA2 [46, 53]. It is important to note, however, that in murine models FOG1 itself is required for both megakaryocyte and erythrocyte lineage specification of hematopoietic progenitors, indicating GATA1-dependent and GATA1-independent roles for FOG1 in hematopoiesis [14, 44, 54, 55].

As neither GATA1 nor FOG1 recognizes the ETS sequence, the interaction of the GATA1-FOG1 complex with the ETS site is mediated by an ETS family protein. Multiple platelet-specific genes were in fact identified as containing GATA and ETS sequences before the GATA1-FOG1 interaction with ETS proteins was understood: based on murine and human cell line studies of GP1BA [56], GPIIB (glycoprotein IIb)/ITGA2B (integrin subunit alpha 2b) [57–59], PF4 (platelet factor 4) [60], GPIX [61, 62], c-MPL [63, 64], and GPV (glycoprotein V) [65], this combined motif was considered a characteristic hallmark of megakaryocytes. Interestingly, the 3′ end of GATA1 HS1 contains a likely ETS site [12]. At least three ETS proteins—GABPα (GA-binding protein transcription factor subunit alpha) [66], FLI1 (friend leukemia integration 1) [47, 64, 66–68], and ETS1 (ETS proto-oncogene 1) [64, 67]—have since been identified as partners of the GATA1-FOG1 complex in human and murine cell line models.

GABPα is a component of the TF complex GABP and mediates DNA binding. It is active in early megakaryopoiesis and induces the expression of ITGA2B and c-Mpl [66]. FLI1 is active throughout megakaryopoiesis, although it preferentially binds to genes active in the later stages of this process [66]. While FLI1 is also able to enhance GATA1-FOG1 activation of ITGA2B, GAPBα has a stronger affinity to this promoter [47, 66], and instead, FLI plays a more critical role in the expression of GP1BA [66, 68], GPIX [66–68], PF4 [66], c-MPL [64, 67], and likely GPIIB [67]. Despite this importance of FLI1 in GATA1-FOG1 activity, studies have disagreed on the nature and stability of the GATA1-FOG1-FLI1 interaction [47, 50, 69]; some have suggested direct binding between GATA1 and FLI1, while others have demonstrated poor in vitro affinity and instead proposed a more allosteric role for FLI1. Finally, a combination of murine and human cell line studies have shown that ETS1 occupies the promoter regions of GPIX [67], c-MPL [64, 67], GPIIB [67], and likely PF4 [59].

GATA1-FOG1 also interacts with the multiprotein complex NuRD (nucleosome remodeling and deacetylase) via the N-terminus of FOG1 to mediate both transcriptional repression and activation by GATA1 and maintain hematopoietic lineage fidelity [70–73]. Studies in murine cells have shown that histone deacetylase (HDAC) activity is required for erythrocyte and megakaryocyte lineage commitment. Mice with a FOG1 mutation that disrupts FOG1-NuRD interaction have partial embryonic lethality due to anemia, extra-medullary erythropoiesis, and thrombocytopenia due to a block in megakaryocyte maturation; they also show aberrant expression of mast cell and eosinophil genes in megakaryocyte and erythroid lineages and fail to undergo the “GATA switch” that downregulates GATA2 upon terminal differentiation of hematopoietic progenitors into mature erythrocytes [71, 72].

When associated with FOG1, NuRD is a nine-protein complex containing Mi-2β (CHD4, chromodomain helicase DNA-binding protein 4), MTA1 (metastasis-associated 1), MTA2 (metastasis-associated 2), p66 (GATAD2A, GATA zinc finger domain containing 2A), RBAP46 (retinoblastoma-binding protein P46) (RBBP7, retinoblastoma-binding protein 7), RBAP48 (retinoblastoma-binding protein P48) (RBBP4, retinoblastoma-binding protein 4), and MBD3 (methyl-CpG-binding domain protein 3) [70]; in megakaryocytes derived from primary mouse fetal liver, MTA2 and RBAP46 were enriched at GATA1-FOG1 targets including Itga2b, Mpl, Gpvi (glycoprotein VI), Gpix, Gp1ba, and Pf4 [70]. The loss of NuRD binding to FOG1 results in significantly decreased activation of Itga2b, although the extent of the redundancy between the NuRD- and ETS-mediated GATA1-FOG1 regulation of this gene is unclear as this was demonstrated by a reporter assay in a murine cell line [73]. Similarly, the same study showed that expression levels of other megakaryocyte genes were affected to different extents by loss of the NuRD-FOG1 association [73], indicating alternative mechanisms for gene regulation. Megakaryocytes derived from the fetal livers of mice with FOG1 mutations preventing NuRD-FOG1 interactions also showed variable changes in mRNA levels of platelet proteins, with significantly decreased levels of P-selectin transcript but not Vwf (von Willebrand factor), Pf4, or Pbp (platelet basic protein) [74]. Of note, GATA1-FOG1 also interacts with the CTBP (carboxy-terminus-binding protein) complex, but it has only been shown to drive erythropoiesis with no effect on the megakaryocyte lineage [75].

Finally, in addition to directly occupying the Gata1 promoter region, ZBP89 appears to associate with GATA1-FOG1 multiprotein complexes in early megakaryocyte development and is required for normal megakaryocyte differentiation of murine embryonic stem cells (ESCs) [41]. Key megakaryocyte genes such as ITGA2B, GPVI, and c-MPL contain GATA and GC-rich binding sites, which are binding motifs recognized by GATA1 and PU.1/SPI1 (hematopoietic transcription factor PU.1/spleen focus-forming virus (SFFV) proviral integration protein); ZBP89 dimerizes with PU.1. Thus, although PU.1 is a hematopoietic TF critical for lymphoid and myeloid but not megakaryocyte or erythroid differentiation [76], it has been hypothesized that GATA1, ZBP89, and PU.1 may together regulate the expression of these genes [41]. PU.1 has separately been shown to regulate the expression of the megakaryocyte gene PBP [77], supporting this hypothesis of GATA1, ZBP89, and PU.1 interplay.

12.3.2. Non-FOG1-Binding Partners

In addition to FOG1, GATA1 interacts with a range of protein partners to regulate megakaryopoiesis. These include RUNX1 (runt-related transcription factor 1), FLI1 outside of the context of the GATA1-FOG1 interaction, SCL/TAL1 (stem cell leukemia/T-cell acute lymphoblastic leukemia 1), GATA2, and GATA1 itself; a ChIP-Seq (chromatin immunoprecipitation followed by deep sequencing) study in primary human megakaryocytes suggests that these TFs interact in different combinations, including all five together, to regulate key megakaryocyte genes such as RUNX1, ITGA2B, and GP1BB (glycoprotein 1b platelet subunit beta) [78]. RUNX1 and its cofactor CBFβ, the β subunit of the evolutionarily conserved transcription complex CBF (core-binding factor), are required for definitive hematopoiesis and are upregulated specifically in megakaryocytes but not erythrocytes [79]. The runt domain of RUNX1 binds to the N- and C-TADs of GATA1, although the C-TAD may play a more critical role in this interaction and subsequent transcriptional activity [79, 80]; interaction between GATA1 and RUNX1 has also been shown in Drosophila melanogaster and mice. It has been suggested that GATA1 targets RUNX1 for phosphorylation, thereby converting RUNX1 from a repressor to an activator and triggering a megakaryocytic transcriptional program [81].

Several mouse and cell line studies have demonstrated that GATA1 and RUNX1 also interact as part of larger complexes. The GATA1-RUNX1-CBFβ complex synergistically upregulates the ITGA2B [79] and GPIBA [80] promoters; this activity is mediated in part by P-TEFb (positive transcription elongation factor b), a kinase complex that includes cyclin T1 and CDK9 (cyclin-dependent kinase 9) [82], that functions in various contexts to phosphorylate RNA polymerase II (RNAPII) and promote transcription; chemical inhibition of CDK9 significantly decreased the ability of GATA1-RUNX1-CBFβ to upregulate ITGA2B expression [83]. It has also been suggested that GATA1, RUNX1, and CBFβ act within a larger complex with FOG1 and FLI1 to activate ITGA2B as well as other key megakaryocyte genes such as c-MPL [84].

In addition to complexing with GATA1 alone or in association with GATA2, RUNX1, and/or FLI1 [78], SCL/TAL1 also interacts with GATA1 as part of a pentameric complex with E2A (transcription factor E2 alpha), LDB1 (LIN-11, Isl-1 and MEC-3 (LIM) domain-binding 1), and LMO2 (LIM domain only 2) [7, 34, 85, 86]. This complex binds to a combined E-box (CAGGTG) and GATA motif on its target genes [34]. In megakaryocytes and a murine cell line, GATA1-SCL/TAL1-E2A-LDB1-LMO2 interacts with ETO2 (CBFA2T3, CBFA2/RUNX1 partner transcriptional co-repressor 3) [7, 85, 86], a highly conserved transcriptional corepressor that has also been studied extensively in erythroid differentiation [86, 87]. ETO2 appears to restrain the expression of certain terminal megakaryocyte genes and thus prevent premature megakaryocyte differentiation by binding to the promoter of the key platelet gene Pf4 [7]. LMO4 (LIM domain only 4) and CDK9 were found to associate with LDB1 in murine cells; both proteins are involved in the regulation of the cell cycle and definitive hematopoiesis [86].

GATA1 binds to itself via a reciprocal interaction between the N-ZF and C-ZF [88, 89]. It has been suggested that dimerization is important for acting on promoters that contain multiple GATA binding sites [89], and a subsequent murine study found that GATA1 homodimerization is required for binding to repeated double/tandem GATA motifs [5]. Overall, investigation into the function of GATA1 homodimers has been limited.

Finally, GFI1B (growth factor-independent 1B transcriptional repressor), a TF originally reported as a GATA1-binding partner in red cells [49], has subsequently been shown in a mouse model to be required for megakaryocyte maturation; fetal liver megakaryocyte colony-forming assays showed smaller colonies and decreased expression of such genes as Vwf and c-Mpl [90]. Gfi1b-mutated mice were also found to have macrothrombocytopenia, megakaryocyte dysplasia, and impaired cytoskeleton organization in megakaryocytes [91]. Based on a human cell line study with constitutive RAS (rat sarcoma virus) activation, GATA1 appears to be required in RAS-mediated differentiation of hematopoietic progenitors to megakaryocytes [92]. However, the mechanism by which GATA1 and the RAS pathway interact remains unclear.

12.4. GATA1 Regulation of Megakaryocyte Development

Through the above interactions, GATA1 induces the expression of numerous megakaryocyte genes that contain GATA-binding sites in their promoters and have been identified in primarily murine-based studies [14, 47, 48, 54, 56–68, 70, 74, 76, 78–80, 83, 84, 90]. GATA1 is essential for normal platelet production, growth, and differentiation; varied GATA1-mutant megakaryocytes are highly proliferative but do not undergo terminal maturation [14, 17, 32, 48, 54, 93, 94]. This has been primarily studied in mice with either the Gata1neoΔHS or the Gata1ΔneoΔHS construct, both of which lack HS I and the distal promoter [14, 95], with mutations leading to exclusive expression of GATA1s [17, 94], or with the Gata-1.05 mutation that express 5% of WT levels [93] (Fig. 12.5 and Table 12.1).

Fig. 12.5.

Schematic of gene changes in murine Gata1 models. IT testis promoter/first exon, IE erythroid promoter/first exon, HS hypersensitivity region

Table 12.1.

Phenotypes of murine Gatal models

	Gata1neo∆HS Gata1∆neo∆HS	Gata1s	Gata-1.05	Gata1 Plt13
Mutation effect	Four- to five-fold decrease in Gata1 expression	Translation of GATA1s only	Gata1 expression 5% of WT	Undetectable Gata1 expression
Embryonic/fetal phenotype	↑Megakaryocyte progenitors in the liver	↑↑Megakaryocyte progenitors in the liver	Lethal anemia by E12.5 (males) Thrombocytopenia (females)	Lethal anemia (males)
Adult phenotype	↑↑Megakaryocytes in the spleen and bone marrow Macrothrombocytopenia (males)	↑Immature megakaryocytes in the bone marrow Normal to decreased platelet count	Thrombocytopenia, anemia Splenomegaly with megakaryocyte trapping ↑Megakaryocytes in the liver	↑↑Megakaryocytes in the spleen and bone marrow ↑Megakaryocytes in the liver and lung Thrombocytopenia Splenomegaly with megakaryocyte trapping
Platelet phenotype	Small, dysplastic ↑Proliferation ↓Granules and DMS ↓Activation Impaired upregulation of platelet-specific genes	Small ↑Proliferation Moderate upregulation of platelet-specific genes Immature (but more mature than (∆)neo∆HS)	Morphologically normal Impaired terminal differentiation	Dysplastic ↑Ploidy ↓DMS Immature
References	[8, 14, 32, 48, 54]	[17, 48, 97, 98]	[10, 97]	[101]

DMS demarcation membrane system. WT wild-type

Transgenic mice harboring Gata1neoΔHS or Gata1ΔneoΔHS exhibit increased megakaryocyte progenitors in the fetal liver and markedly increased megakaryocyte number in the adult spleen and bone marrow (BM), consistent with enhanced proliferation [14, 32] (Table 12.1); in vitro experiments with Gata1neoΔHS or Gata1ΔneoΔHS megakaryocytes reveal a similar proliferative phenotype. The in vivo megakaryocytes are smaller than controls and are dysplastic, with scant granules, disorganization of the demarcation membrane system (DMS), and minimal proplatelet formation [14, 48, 54]; despite the smaller cell size, the megakaryocyte colonies are larger due to proliferation of the immature cells [32], as evidenced by decreased DNA content and ploidy [54]. Transgenic male mice are macrothrombocytopenic, reflecting the impaired proplatelet formation, whereas heterozygous females have a normal platelet count [14, 54]. Microarray [48, 54] and flow cytometry [32] analyses of the in vivo and in vitro megakaryocytes show dysregulation of canonical platelet genes such as Gp1ba, Gp1bb, Gpvi, Pf4, c-Mpl, and p45 Nfe2, as well as those involved in cytokine signaling, calcium signaling, and maintenance of the cytoskeleton. Consistent with this microarray profile and the paucity of intracellular structures, Gata1neoΔHS or Gata1ΔneoΔHS platelets have decreased activation in response to thrombin or combined ADP (adenosine diphosphatase) and epinephrine [54].

Similarly, Gata1-lox|Pf4-Cre mice with megakaryocyte-specific GATA1 loss exhibit macrothrombocytopenia despite increased megakaryocyte proliferation, aberrant megakaryocyte maturation, and decreased platelet functionality [96]. The GATA1-deficient megakaryocytes appear to have temporally dysregulated maturation, with earlier expression of platelet-specific surface markers such as CD41 or CD61 but lower ploidy when compared to WT. These platelets also have a significantly diminished response to stimulation with aggretin, collagen, convulxin, and botrocetin attributed to defects in receptors such as integrin β1 or GPVI. Taken together, these findings in distinct mouse models underscore the uncoupling of megakaryocyte proliferation and terminal maturation with GATA1 loss.

The effect of loss of the N-terminus of GATA1 (i.e., GATA1s) appears in mice to be developmental stage-dependent and are distinct from the consequences of Gata1neoΔHS and Gata1ΔneoΔHS [17]. GATA1s expression in adult mice results in a mild expansion of immature megakaryocytes in the BM [97], but when expressed at the embryonic and fetal stages, there is a significant increase in megakaryocyte progenitors in the yolk sac and fetal liver that are hyperproliferative, immature, and morphologically abnormal, similar to the megakaryocytes observed in Gata1neoΔHS and Gata1ΔneoΔHS mice (Table 12.1). However, Gata1s megakaryocytes appear to be more mature than their neoΔHS and ΔneoΔHS counterparts, with moderate upregulation of megakaryocyte-specific genes but impaired downregulation of non-megakaryocytic genes including five TFs (Myc (myelocytomatosis viral oncogene homolog), Myb (myeloblastosis viral oncogene homolog), Gata2, Pu.1, and Znfn1a1 (zinc finger protein, subfamily 1A, 1)); inadequate repression of these genes that control hematopoietic progenitor proliferation and lineage specification may contribute to the hyperproliferation of Gata1s megakaryocytes [17]. Megakaryocytes with exclusive expression of GATA1s form proplatelets at a higher frequency ex vivo compared to Gata1ΔneoΔHS but at a lower frequency compared to WT [17, 48], leading to a decreased platelet count overall. A human ex vivo study of hematopoietic stem/progenitor cells (HSPCs) isolated from fetal liver and geneedited to exclusively express GATA1s similarly demonstrated that GATA1s leads to increased cell proliferation with a specific increase in CD41⁺ cells within all colonies containing megakaryocytes, although this study did not evaluate platelet maturation [98].

In in vitro studies of murine ESC-derived hematopoietic cells, both GATA1 loss [99, 100] and exclusive expression of Gata1s [94] lead to increased production of immature, developmentally arrested megakaryocytes. Gata1s megakaryocytes generated from ESCs are smaller, more proliferative, and express cell surface CD41 but lack GPIB that is typically expressed on mature megakaryocytes [100] or retain the progenitor cell marker KIT (KIT proto-oncogene, receptor tyrosine kinase) [94]. Not surprisingly, these megakaryocytes express a more immature transcriptional profile when compared to their WT counterparts [94]. In vivo studies of murine yolk sacs [94] and fetal liver [100] yielded similar findings. Meanwhile, GATA1 rescue studies in G1ME cells, a GATA1-deficient murine bipotential MEP line originally derived from differentiation of ESCs, reinforced the role of GATA1 in inhibiting cell division to promote maturation [100].

A more global loss of GATA1 as induced by the Gata-1.05 mutation induces thrombocytopenia in female heterozygous embryos and adult mice with varying degrees of random X inactivation; male mice die at E12.5 from severe anemia [10, 93] (Table 12.1). The female mice exhibit severe anemia and develop variable degrees of marked splenomegaly in addition to the thrombocytopenia. The megakaryocytes trapped in the spleens appear morphologically normal and express CD41 but fail to undergo terminal differentiation to shed mature platelets [93]. Meanwhile, a mutation in the Gata1 start codon designated Plt13, which is predicted to result in exclusive production of GATA1s but instead results in undetectable levels of either protein isoform, results in moderate thrombocytopenia without anemia in female heterozygous mice, and is embryonic lethal in hemizygous males [101] (Table 12.1). The Gata1^Plt13 megakaryocytes also accumulate in the spleens of the heterozygous females; they are characterized by hyperchromatic nuclei, disorganized or absent DMS, and increased ploidy when compared to WT. In contrast to the findings in Gata1neoΔHS or Gata1ΔneoΔHS mice [32, 54], Gata1^Plt13 megakaryocytes have increased DNA content and form colonies of normal size; of note, increased megakaryocyte colony formation was restricted to Gata1^Plt13 fetal liver, similar to Gata1s mice [17].

Mutations that prevent interactions between GATA1 and its protein partners have also been studied in murine models of hematopoiesis. A near-complete to complete loss of hematopoietic progenitors committed to the megakaryocyte and erythroid lineages has been observed in the absence of FOG1, suggesting that gene regulation through the GATA1-FOG1 and other FOG1 complexes is critical for both erythropoiesis and megakaryopoiesis [44, 48, 55]. The V205G mutation, which eliminates the GATA1-FOG1 interaction, results in large megakaryocytes that are unable to form proplatelets; these megakaryocytes also lack expression of genes involved in terminal maturation [32, 48]. Mice with FOG1 mutations preventing NuRD-FOG1 interaction and thus formation of the GATA1-FOG1-NuRD complex exhibit severe macrothrombocytopenia with intracellular dysplasia and impaired activation, although proliferation and ploidy appear intact [73, 74]. Interestingly, although the α-granules of these platelets appear morphologically abnormal, only P-selectin is significantly decreased among the α-granule proteins [74]. These platelets also lack response to thrombin and convulxin, with absent PF4 release.

Studies in human-derived models are extremely limited and have primarily occurred in the context of DS [102–106]. Although euploid human induced pluripotent stem cells (iPSCs) expressing WT GATA1 vs. exclusively GATA1s generate similar numbers of multipotent hematopoietic progenitors, when cultured in megakaryocyte-specific media, the GATA1s progenitors are more proliferative compared to WT controls [102]. This is accompanied by increased expression of genes associated with megakaryocytes and myeloid cells and decreased expression of those associated with erythrocytes. In human T21 iPSCs, GATA1s mutations and GATA1s overexpression both suppress megakaryocyte commitment in hematopoietic progenitors but ultimately result in a population of proliferative, immature CD34⁺/CD41⁺ megakaryocytes assayed by cell surface expression markers, cell morphology, and response to thrombin stimulation [103–107].

12.5. GATA1 Mutations in Human Diseases of Platelet Development

12.5.1. Germline GATA1 Mutations

Mutations affecting GATA1 have been recurrently implicated in human hematopoietic disease [108–111]. Germline mutations affecting GATA1 have been reported in association with X-linked congenital cytopenias, a heterogeneous spectrum of hematologic conditions unified by platelet abnormalities. Pathogenic variants can be divided into those that result in loss of the N-TAD, or affecting the C-ZF, N-ZF, or C-TAD (Fig. 12.6 and Table 12.2). While some cases have historically been classified as a variant of Diamond-Blackfan anemia (DBA) [19, 21, 112–116], thalassemia [117–121], or congenital erythropoietic porphyria (CEP) [121–123], GATA1-related cytopenias likely represent a distinct clinical entity. Patients diagnosed with “GATA1-related DBA” were found to have clinical manifestations of DBA, including macrocytic anemia and variable neutropenia, but lack a mutation in one of the ribosomal proteins and instead carry GATA1 mutations in the N-terminus, resulting in exclusive expression of GATA1s [19, 21, 114, 115, 124], or disrupting the C-TAD [125]. These individuals often have additional hematologic abnormalities, for example, thrombocytopenia, thrombocytosis, or dysmegakaryopoiesis, while lacking extra-hematopoietic manifestations of DBA such as skeletal anomalies. It has been hypothesized that the clinical overlap between germline GATA1 mutations and DBA is due to the impaired translation of GATA1 that occurs as a result of the ribosomal insufficiency in DBA [21, 114], but the distinct phenotype of “GATA1-related DBA” suggests that it is more precisely termed GATA-1 related cytopenias.

Fig. 12.6.

Broad associations between affected region of GATA1 and resulting phenotypes. N-TAD amino-terminus activation domain, C-TAD carboxy-terminus activation domain, N-ZF amino-terminus zinc finger, C-ZF carboxy-terminus zinc finger, Lu Lutheran blood group antigen, MDS myelodysplastic syndrome

Table 12.2.

Summary of reported human GATA1 mutations and associated phenotypes

DNA nucleotide change	Predicted protein change	Megakaryocyte and platelet phenotype	Red cell phenotype	Other hematologic features	Non-hematologic features	Refs.
Pathogenic variants resulting in GATA1s
c.−21A>G	–	Normal to mild thrombocytosis Dysplastic megakaryocytes	Dyserythropoiesis Mild to severe anemia Macrocytic ↑Hgb F	Occasional neutropenia ML-DS with acquired T21 MDS, AML		[116, 132]
c.2T>C	p.Met1a	Normal to mild thrombocytosis Dysplastic megakaryocytes	Dyserythropoiesis Severe anemia Macrocytic “Clinical DBA”	Progression to MDS Megaloblastic changes ML-DS with acquired T21		[21. 114, 115. 132]
c.3G>A	p.Met1a	Fluctuating thrombocytopenia	Moderate to severe anemia normocytic “Clinical DBA”		Hypertelorism and flat nasal bridge Hepatomegaly	[124]
c.94delG	p.Val32PhefsTer105	Moderate to severe thrombocytopenia Dysplastic megakaryocytes	Mild to severe anemia Normocytic to macrocytic	TAM with acquired T21 Hypercellular, fibrotic marrow	Reported in heterozygous females only	[131]
c.220+2T>C	–	Mild thrombocytosis Dysplastic megakaryocytes	Dyserythropoiesis Moderate anemia Macrocytic “Clinical DBA”			[21]
c.220G>C	p.Val74Leu	Normal or mild thrombocytopenia Dysplastic megakaryocytes Decreased aggregation	Mild to severe anemia Macrocytic	Neutropenia		[19, 20, 113]
Pathogenic variants affecting N-ZF
c.613G>A	p.Val205Met	Mild to severe thrombocytopenia Macrocytosis	Moderate to severe anemia Dyserythropoiesis Fetal hydrops		Cryptorchidism	[53, 150]
c.617A>T	p.Asn206Ile	Severe thrombocytopenia Decreased granules	Mild dyserythropoiesis			[146]
c.622G>A	p.Gly208Arg	Mild to severe thrombocytopenia Macrocytosis	Moderate to severe anemia Dyserythropoiesis		Cryptorchidism Splenomegaly Hepatomegaly Phenotype may improve with age	[144, 147]
c.622_623delGGinsTC	p.Gly208Ser	Severe thrombocytopenia  Hypogranular platelets Macrocytosis Decreased aggregation				[142, 143]
c.647G>A	p.Arg216Gln	Normal to moderate thrombocytopenia Macrocytosis Decreased aggregation, prolonged bleeding time	Mild anemia “β-thalassemia”		Splenomegaly	[117–120, 123,126–128, 141]
c.646C>T	p.Arg216Trp	Moderate thrombocytopenia	Mild anemia “β-thalassemia”		Splenomegaly “X-linked CEP”	[121–123]
c.653A>G	p.Asp218Gly	Moderate to severe thrombocytopenia Macrocytosis Decreased aggregation	Dyserythropoiesis without anemia			[140, 148, 149]
c.652G>T	p.Asp218Tyr	Severe thrombocytopenia Macrocytosis	Severe anemia	Platelets in heterozygous female expressed only wild-type allele		[139]
c.652G>A	p.Asp218Asn	Severe thrombocytopenia Macrocytosis Decreased aggregation, prolonged bleeding time	Dyserythropoiesis without anemia		Splenomegaly	[137, 145]
Pathogenic variants affecting C-ZF
c.788C>T	p.Thr263Met	Moderate thrombocytopenia Dysplastic megakaryocytes	Mild dyserythropoiesis Mild anemia	Leukocytosis (neutrophilia) Hypercellular, brotic marrow	Reported in heterozygous females only	[138]
c.802C>A	p.Leu268Met	Worsening thrombocytopenia with age Decreased aggregation	Anisocytosis, poikilocytosis Progressive macrocytosis with age ↑Hgb F			[146]
c.865C>T	p.His289Tyr	Normal to mild thrombocytopenia Anisocytosis Decreased aggregation, prolonged bleeding time	Dyserythropoiesis Normal to mild anemia			[137]
Pathogenic variants affecting C-TAD
c.871–24C>T	–	Normal to moderate thrombocytopenia Macrocytosis Dysplastic megakaryocytes Decreased aggregation	Dyserythropoiesis Hydrops fetalis, severe infantile anemia Worsening anemia with illness ↑Hgb F	Occasional myeloid dysplasia Progression to aplastic anemia	Cryptorchidism Hypospadias	[112, 133]
c.886A>C	p.Thr296Pro	Moderate thrombocytopenia Macrocytosis Decreased aggregation	Mild dyserythropoiesis Mild anemia ↑Hgb F Lu(a-b-)			[135]
c.919C>T	p.Arg307Cys	Mild thrombocytopenia	Mild dyserythropoiesis Hemolysis “Clinical DBA”	ADA overproduction		[125]
c.920G>A	p.Arg307His	Mild thrombocytopenia Macrocytosis	Variable dyserythropoiesis Hyperchromia Macrocytosis Hemolysis ↑Hgb F Hydrops fetalis, severe infantile anemia Lu(a-b-) “Clinical DBA”	ADA overproduction	Low birth weight Hypospadias Splenomegaly	[125, 134]
c.1240T>C	p.Ter414Arg	Mild thrombocytopenia Macrocytosis	Lu(a-b-)			[136]
Unclassified pathogenic variants
c.515T>C	p.Phe172Ser	Severe thrombocytopenia	Severe anemia	Leukopenia Bone marrow hypoplasia	Mild splenomegaly	[130]

AML acute myeloid leukemia, ADA adenosine deaminase, CEP congenital erythropoietic porphyria, C-TAD carboxy-terminus transactivation domain, C-ZF carboxy-terminus zinc finger, DBA Diamond-Blackfan anemia, Hgb F fetal hemoglobin, Lu Lutheran blood group antigen, MDS myelodysplastic syndrome, ML-DS myeloid leukemia of Down syndrome, N-ZF amino-terminus zinc finger, T21 trisomy 21, TAM transient abnormal myelopoiesis

^aSingle nucleotide mutation that disrupts start codon; an alternative start codon results in GATA1s expression

Similarly, those diagnosed with “X-linked thalassemia” or “X-linked thrombocytopenia with thalassemia” (XLTT) were noted to have ineffective erythropoiesis but also to have incongruous features, such as macrothrombocytopenia, a disproportionately prolonged bleeding time or platelet dysfunction, and progressive BM fibrosis [117–121, 126]. Instead of a mutation in a globin gene, a GATA1 mutation affecting the N-ZF and therefore DNA binding, p.Arg216Gln (c.647G.A), has been recurrently identified in these individuals [117–120]. However, it is important to note that the same mutation [127] and other mutations affecting the same residue (e.g., p.Arg216Trp (c.646C > T)) [128, 123] have been described without thalassemia, suggesting that these mutations do not uniformly cause thalassemia and that the interpretation of the resulting phenotype may be context- or physician-dependent.

CEP, a group of disorders of heme biosynthesis, is typically caused by autosomal recessive mutations in UROS (uroporphyrinogen III synthase) [121, 128]. While the typical hematologic manifestations of CEP are chronic hemolytic anemia and transfusion dependence, those with “X-linked CEP” have been found to also have dyserythropoiesis (described as “thalassemia” in one individual), persistence of Hgb F (fetal hemoglobin), and thrombocytopenia in addition to the characteristic skin findings of porphyria [121, 128]. These individuals have been found to have the mutation p.Arg216Trp (c.646C > T) that affects the N-ZF of GATA1. Interestingly, although this mutation has been reported in two probands, the co-inheritance of a UROS mutation was reported in one family [128], while this GATA1 mutation was the only pathogenic variant identified in the other [121]; of note, GATA1 regulates UROS expression in developing erythrocytes [129], likely accounting for the clinical overlap. Again, the hematologic features inconsistent with the unifying diagnosis, as well as the overlap in symptoms and terminology with those carrying other diagnoses, implies that “X-linked CEP” is in fact a form of GATA-1-related cytopenias.

Indeed, the wide range of phenotypic manifestations reported in individuals with germline mutations affecting the N-TAD, N-ZF, C-ZF, and C-TAD of GATA1 suggests that GATA1-related cytopenias are primarily a platelet disorder with variable other hematopoietic and extra-medullary findings that may overlap with those of other conditions (Table 12.2). The heterogeneity is likely due to the differences in the DNA-binding sites and protein-protein interactions that are disrupted by each of the mutations. At the same time, the distinct roles ascribed to each of GATA1’s functional domains allow for broad genotype-phenotype associations. The management of GATA1-cytopenias is primarily supportive, although curative HSC transplants have also been reported [20, 21, 53, 112, 121, 128, 130].

Pathogenic variants resulting in exclusive GATA1s expression with loss of the N-TAD occur due to changes in the exon 2 start codon, premature stop codons, or affect the exon 2 donor splice site. They tend to have a more severe red cell phenotype and variable thrombocytopenia despite megakaryocyte dysplasia and may lead to marrow fibrosis or failure including MDS (myelodysplastic syndrome), ML-DS with GATA1s expressed in the blast cells only (see GATA1 mutations in Down syndrome (DS) below), or ML-DS-like AML (acute myeloid leukemia) with acquired T21 [19–21, 113–116, 124, 131, 132]. Of note, only one variant (p.Val74Leu) is reported to result in decreased platelet function [20], although platelet aggregation testing was not pursued in most cases likely as bleeding symptoms were minor or consistent with the degree of thrombocytopenia.

Clinical phenotypes of mutations affecting the C-TAD are generally similar to those affecting the N-TAD, but with overall milder thrombocytopenia and more severe dyserythropoiesis and/or anemia. While most of these mutations are single-amino acid changes, one is intronic and hypothesized to affect splicing [112, 133]. There is a tendency for the anemia to be most severe in infancy, with multiple reports of hydrops fetalis [112, 133] as well as severe neonatal anemia requiring multiple transfusions [112, 133, 134], but there are also reports of possible progression to aplastic anemia later in life [112]. These C-TAD mutations are uniquely associated with Lu(a-b-) erythrocytes [135, 136], adenosine deaminase (ADA) overproduction and hemolysis [125], and an elevated Hgb F fraction [112, 133–135]. In comparison to germline GATA1s mutations, C-TAD variants appear more likely to cause abnormal platelet function [112, 133] or a bleeding diathesis, with easy bruising and petechiae [135], epistaxis [133, 135], gastrointestinal bleeding [133], and/or excessive bleeding with trauma or surgery [133].

Meanwhile, variants affecting either the N-ZF or C-ZF are associated with more severe macrothrombocytopenia, decreased platelet function and prolonged bleeding time, and dyserythropoiesis with variable anemia. This group of variants includes p.Val205Met, which disrupts GATA1-FOG1 binding and has been extensively studied in primarily murine-derived models [46, 47, 50, 53, 70, 72, 123]. Interestingly, even carrier females may have mild cytopenias or bleeding tendencies [53, 128, 137, 138], and both female heterozygotes [127, 139, 140] and male hemizygotes [119, 141] have rarely been reported to have two platelet populations on peripheral smear. The reported bleeding in individuals with GATA1 mutations affecting either ZF range from none [138, 142] to severe, regardless of tested platelet function, and include easy bruising and petechiae [126, 127, 137, 143–147], recurrent epistaxis [126, 130, 137, 143, 145, 147], other mucocutaneous bleeding [119, 145–148], excessive bleeding with trauma or surgery [119, 126, 137, 143, 148, 149], gastrointestinal hemorrhage [143, 149], menorrhagia [137], fetal cerebral hemorrhage [150], and other non-specified bleeding requiring medical evaluation with or without platelet transfusions or other intervention [53, 118, 120, 127, 139, 144]. Of note, two related individuals with the p.Leu258Met mutation had a worsening phenotype with age [146], in contrast to the stable cytopenias seen with the majority of the variants. Another mutation, p.Thr263Met, has been reported only in two related heterozygous females [138].

12.5.2. GATA1 Mutations in Down Syndrome (DS)

GATA1s uniquely interacts with DS, or T21, to result in TAM and ML-DS [16, 22, 151, 152]. Formerly known as transient myeloproliferative disorder (TMD), TAM is a preleukemic or leukemia-like condition that occurs exclusively in ~10% of infants with DS in the first 3 months of life [153–155]. In the majority of cases, the clonal proliferation of blast cells remits spontaneously, and no chemotherapy is required unless the blast burden causes end-organ dysfunction, most commonly massive hepatomegaly that may lead to cardiopulmonary compromise. Despite this, however, the 5-year overall survival (OS) for TAM is 80–90%, and ~25% of those who survive subsequently develop ML-DS, a true leukemia that requires standard chemotherapy, within 4 years [153–155]. The blasts of both TAM and ML-DS carry a GATA1s mutation while ML-DS blasts harbor at least one additional mutation, often in epigenetic regulators or members of the cohesin complex, suggesting that ML-DS arises from a residual TAM subclone that has acquired an additional oncogenic “hit” [23, 156–158] (Fig. 12.7). Both conditions are thought to arise from an aberrant MEP clone, and the blasts typically carry a megakaryocyte-like phenotype by flow cytometry.

Fig. 12.7.

Progression from germline trisomy 21 state to TAM with acquisition of GATA1s mutation, then to ML-DS with acquisition of “third-hit” mutation. TAM transient abnormal myelopoiesis, ML-DS myeloid leukemia of Down syndrome

T21 itself perturbs fetal liver hematopoiesis [159–161], and the acquisition of the GATA1s mutation, while somatic, is early and often during fetal life [162, 163]. Given that TAM can present on the day of birth or even cause in utero complications such as hydrops fetalis, it is conceptualized as a disorder of fetal hematopoiesis [154, 164], and its self-limiting nature is therefore likely due to the postnatal transition from fetal to adult hematopoiesis. Interestingly, the frequency of GATA1s mutations in the DS population may be as high as 30% when assessed by targeted next-generation sequencing instead of standard Sanger sequencing, indicating that ~20% of infants with DS have “silent TAM” [22]. Whether this is because some infants with T21 and a GATA1s mutation truly do not develop clinical TAM or because the disease process peaked and spontaneously remitted in fetal life remains unclear.

ML-DS is both genetically and clinically distinct from myeloid leukemias not associated with DS [165]. In addition to the hallmark GATA1s mutation, there are recurrent oncogenic mutations that are more common in ML-DS or do not carry the same prognostic significance when compared to non-DS leukemias. ML-DS blasts are exquisitely chemosensitive, especially when compared to other myeloid leukemias, leading to a >90% 5-year OS although the survival rate for relapsed cases remains <20% [166, 167]. Individuals without T21 but with germline GATA1 mutations do not have a leukemic predisposition. Yet there also exist case reports of patients without T21 but with germline GATA1s mutations who developed TAM [131] or ML-DS-like leukemia [132] and were found to have acquired T21 in the blast cells only. Furthermore, there are case reports of individuals without either germline T21 or GATA1 mutations who developed TAM and/or ML-DS-like leukemia and were found to have acquired T21 and GATA1s in the blast cells only [168]. The characteristics of these patients mirrored what would be expected for those with DS with regards to age of onset, response to chemotherapy, and associated genetic changes in the blasts. Given that in vivo and in vitro murine studies have demonstrated that GATA1s increase megakaryocyte proliferation at the expense of terminal maturation [14, 32, 48, 54, 93, 94, 96, 99, 100], it is not clear why a leukemic risk is increased only in the setting of T21. These cases therefore emphasize the unique interaction between chromosome 21 gene dosage imbalance and GATA1s, regardless of the order in which the genetic events occur.

Studies in human fetal liver and neonatal peripheral blood samples have demonstrated differential methylation patterns induced by both T21 and GATA1s [169, 170]. While the addition of the third copy of chromosome 21 appears to trigger hypomethylation of genes globally implicated in the broad phenotype of DS, the addition of a GATA1s mutation is associated with aberrant hypermethylation specifically of genes involved in hematopoiesis and cell cycle regulation, including the GATA1 target and key platelet gene NFE2 [169]. Interestingly, there appear to be limited differences between the epigenetic signature and transcriptome of TAM and ML-DS blasts.

As demonstrated in whole-exome and whole-genome sequencing studies, the additional mutation(s) found in ML-DS blasts when compared to those of TAM are typically those that affect chromatin organization and gene expression, with CTCF (CCCTC-binding factor), EZH2 (enhancer of zeste homolog 2), NRAS (neuroblastoma RAS viral oncogene homolog), RAD21 (double-strand-break repair protein Rad21 homolog), STAG2 (cohesin subunit SA-2), and TP53 (tumor antigen P53) among the recurrently mutated genes [23, 156]. A study of human fetal liver xenografts in mice demonstrated that the addition of a STAG2 mutation to a GATA1s background significantly increases engraftment and leukemic progression regardless of chromosome 21 status; in the T21 context, GATA1s alone is sufficient to drive blast production, but the addition of a STAG2 mutation amplifies this effect [98]. Another study of human iPSCs with T21, a GATA1s mutation, and haploinsufficiency of the cohesin protein SMC3 (structural maintenance of chromosomes 3) found that the combination of all three genetic alterations enhanced megakaryocyte colony-forming potential, colony size, and proliferation [106]. SMC3 haploinsufficiency appears to exacerbate the megakaryocyte maturation arrest at the megakaryoblast stage, giving rise to a leukemic phenotype.

ML-DS blasts demonstrate increased expression of GATA1 and of genes such as GATA2 and MYC that are typically downregulated by GATA1 [171]. The human iPSC model with T21, GATA1s, and SMC3 haploinsufficiency found that NFE2, which was found to be hypermethylated in TAM fetal liver samples [169], is also downregulated in this context and that forced NFE2 expression partially restores the maturation defect in megakaryocytes [106]. These examples of gene dysregulation imply that without the N-TAD, GATA1s is functionally unable to regulate these targets. Taken together, these findings suggest that altered chromatin structure and accessibility induced by “third-hit” mutations exacerbate the impaired regulation of hematopoietic genes by GATA1s to give rise to ML-DS.

In summary, GATA1 plays a key role in regulating megakaryocyte differentiation, maturation, and proliferative capacity. Varied murine and cellular models have elucidated the functions of the N- and C-TAD and ZFs. Human GATA1 mutations in these regions also provide mechanistic insight, and as sequencing technologies expand, additional mutations will likely be identified that contribute to GATA1-related cytopenias, including thrombocytopenia. With single-cell-based technology, future studies in human primary cells and/or iPSCs harboring distinct GATA1 mutations may improve our understanding of GATA1 regulatory mechanisms and how their disruptions lead to human diseases of megakaryocytes and platelets.

Abbreviations

ADA

Adenosine deaminase

ADP

Adenosine diphosphatase

AML

Acute myeloid leukemia

BM

Bone marrow

C-TAD

Carboxy (C)-terminus transactivation domain

C-ZF

Carboxy (C)-terminus zinc finger

CBF

Core-binding factor

CDK9

Cyclin-dependent kinase 9

CEP

Congenital erythropoietic porphyria

ChIP-Seq

Chromatin immunoprecipitation followed by deep sequencing

CLP

Common lymphoid progenitor

CMP

Common myeloid progenitor

CTBP

Carboxy (C)-terminus binding protein

CTCF

CCCTC-binding factor

DBA

Diamond-Blackfan anemia

DMS

Demarcation membrane system

DNMT1

DNA methyltransferase 1

DS

Down syndrome

E2A

Transcription factor E2 alpha

ESC

Embryonic stem cell

ETO2 (CBFA2T3)

Core binding factor 2 (CBFA2)/Runt-related transcription factor 1 (RUNX1) partner transcriptional co-repressor 3

ETS

Erythroblastosis virus 26 transformation-specific

ETS1

ETS proto-oncogene 1

EZH2

Enhancer of zeste homolog 2

FLI1

Friend leukemia integration 1

FOG1

Friend of GATA1

G1HE

Gata1 hematopoietic enhancer

G1HRD

GATA1 hematopoietic regulatory domain

G1ME

Gata1⁻ megakaryocyte erythroid

GAPBα

GA-binding protein alpha

GATA1

GATA-binding protein 1

GATA1s

GATA1 short (isoform)

GATA2

GATA-binding protein 2

GFI1B/Gfi1b

Growth factor-independent 1B transcriptional repressor

GMP

Granulocyte/macrophage progenitor

GP1BA/Gp1ba

Glycoprotein 1b subunit alpha

GP1BB/Gp1bb

Glycoprotein 1b subunit beta

GPIIB/Gpiib

Glycoprotein IIb

GPIX/Gpix

Glycoprotein IX

GPV/Gpv

Glycoprotein V

GPVI/Gpvi

Glycoprotein VI

HDAC

Histone deacetylase

Hgb F

Fetal hemoglobin

HS I, II, III

DNase I hypersensitivity region I, II, III

HSC

Hematopoietic stem cell

HSPC

Hematopoietic stem/progenitor cell

IE

Erythroid promoter/first exon

iPSC

Induced pluripotent stem cell

IT

Testis promoter/first exon

ITGA2B/Itga2b

Integrin subunit alpha IIb

Jak2

Janus kinase 2 KIT: KIT proto-oncogene, receptor tyrosine kinase

LDB1

LIN-11, Isl-1 and MEC-3 (LIM) domain-binding 1

LMO2

LIM domain only 2

LMO4

LIM domain only 4

Lu

Lutheran blood group antigen

MBD3

Methyl-CpG-binding domain protein 3

MDS

Myelodysplastic syndrome

MEP

Megakaryocyte-erythroid progenitor

Mi-2β (CHD4)

Chromodomain helicase DNA-binding protein 4

ML-DS

Myeloid leukemia of Down syndrome

c-MPL/c-Mpl

Myeloproliferative leukemia protein

MTA1

Metastasis-associated 1

MTA2

Metastasis-associated 2

Myb

Myeloblastosis viral oncogene homolog

MYC/Myc

Myelocytomatosis viral oncogene homolog

N-TAD

Amino (N)-terminus transactivation domain

N-ZF

Amino (N)-terminus zinc finger

NFE2/Nfe2

Nuclear factor, erythroid 2

NRAS

Neuroblastoma RAS viral oncogene homolog

NuRD

Nucleosome remodeling and deacetylase

P-TEFb

Positive transcription elongation factor b

P66 (GATAD2A)

GATA zinc finger domain containing 2A

PBP/Pbp

Platelet basic protein

PF4/Pf4

Platelet factor 4

PU.1/SPI1

Hematopoietic transcription factor PU.1/spleen focus-forming virus (SFFV) proviral integration protein

RAD21

Double-strand-break repair protein Rad21 homolog

RAS

Rat sarcoma virus

RBAP46 (RBBP7, retinoblastoma-binding protein 7)

Retinoblastoma-binding protein P46

RBAP48 (RBBP4, retinoblastoma-binding protein 4)

Retinoblastoma-binding protein P48

RNAPII

RNA polymerase II

RUNX1

Runt-related transcription factor 1

SCL/TAL1

Stem cell leukemia/T-cell acute leukemia 1

SMC3

Structural maintenance of chromosomes 3

STAG2

Cohesin subunit SA-2

T21

Trisomy 21

TAM

Transient abnormal myelopoiesis

TAD

Transactivation domain

TF

Transcription factor

TMD

Transient myeloproliferative disorder

TP53

Tumor antigen P53

VWF/Vwf

von Willebrand factor

WT

Wild-type

XLTT

X-linked thrombocytopenia with thalassemia

ZBP89

Zinc finger DNA-binding protein 89

ZF

Zinc finger

Znfn1a1

Zinc finger protein, subfamily 1A, 1