Interferon Regulatory Factor-8-Dependent Innate Immune Alarm Senses GATA2 Deficiency to Alter Hematopoietic Differentiation and Function

Structured Abstract

Purpose of review-

Recent discoveries have provided evidence for mechanistic links between the master regulator of hematopoiesis GATA2 and the key component of interferon and innate immunity signaling pathways, Interferon-Regulatory Factor-8 (IRF8). These links have important implications for the control of myeloid differentiation in physiological and pathological states.

Recent Findings-

GATA2 deficiency resulting from loss of the Gata2 −77 enhancer in progenitors triggers an alarm that instigates the transcriptional induction of innate immune signaling and distorts a myeloid differentiation program. This pathological alteration renders progenitors hyperresponsive to IFNγ, Toll-like Receptor (TLR) and Interleukin-6 (IL-6) signaling and impaired in Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF) signaling. IRF8 upregulation in −77^−/− progenitors promotes monocyte and dendritic cell differentiation while suppressing granulocytic differentiation. As PU.1 promotes transcription of Irf8 and other myeloid and B-lineage genes, GATA2-mediated repression of these genes opposes the PU.1-dependent activating mechanism.

Summary-

As GATA2 deficiency syndrome is an immunodeficiency disorder often involving myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML), elucidating how GATA2 commissions and decommissions genome activity and developmental regulatory programs will unveil mechanisms that go awry when GATA2 levels and/or activities are disrupted.

Introduction

The transcription factor GATA2 exerts essential functions at multiple stages of hematopoiesis, including promoting hematopoietic stem cell (HSC) generation, conferring progenitor cell function during embryogenesis, and sustaining steady-state and stress hematopoiesis in the adult [1]. A complete loss of Gata2 leads to failure to produce the earliest blood cells in the embryo, yielding rapid lethality [2,3]. Though a single GATA2 allele suffices to sustain life, loss of one allele is not without consequences. In humans, germline GATA2 haploinsufficiency causes a disorder termed GATA2 deficiency syndrome, which is characterized by immunodeficiency with loss of monocytes, dendritic cells, B cells and natural killer (NK) cells and a predisposition to developing MDS/AML [4–7]. Numerous mutations have been identified in patients diagnosed with GATA2 deficiency syndrome [8], with most mapping within the GATA2 coding sequence, but others within a conserved intronic enhancer (+9.5 in mice) [9,10]. Disease onset is often associated with accumulation of secondary mutations and might involve stresses, such as infection [10–16]. The consequence of Gata2 haploinsufficiency in mice is age-dependent with enhanced HSC quiescence and apoptosis in young mice and increased proliferation and myeloid-biased differentiation of HSCs in aged mice [17,18*]. Although there has been remarkable progress in defining this human germline genetic disorder, given the highly context-dependent actions of GATA2, there are many unanswered questions as to how a single GATA2 allele disruption corrupts molecular processes that confer hematopoietic stem and progenitor cell generation and function.

Text of Review

Physiological and Pathological GATA2 Functions

GATA2 autoregulates its own expression and the related factor GATA1 represses Gata2 transcription during erythroid differentiation in a process termed GATA switching, in which GATA1 replaces GATA2 at Gata2 enhancers [1,19,20]. The +9.5 enhancer was identified as one of five conserved GATA switch sites within the Gata2 locus [21]. An ensemble of transcription factors (LYL1, TAL1, LMO2, FLI1, ERG and RUNX1) are detected by chromatin immunoprecipitation at GATA2 occupancy sites [22–24]. Multi-factor occupancy occurs at hundreds of genomic sites and has been characterized as a heptad complex [23,25,26]. Besides a commonly observed E-box-spacer-WGATAR composite element [25,27–29], diverse cis-elements often reside near GATA factor-occupied WGATAR motifs [30–33], and the rules governing GATA2 function through chromatin are incompletely defined.

Deletion of the conserved E-box-GATA composite element from the +9.5 of mice resulted in failure to generate hematopoietic stem cells in the aorta-gonad mesonephros (AGM) and embryonic lethality by embryonic day (E) 14.5 [9,34]. GATA2-regulated endothelial to hematopoietic transition [34,35] involves the regulation of many genes in hemogenic endothelial cells and hematopoietic stem and progenitor cell (HSPC) progeny, include the transcription factor GFI1b [34], which promotes the endothelial to hematopoietic transition downstream of GATA2 [36*]. Although loss of a +9.5 ETS motif, or a single-nucleotide alteration in this motif that can occur in GATA2 deficiency syndrome patients can be tolerated during development and homeostasis, these alterations create a vulnerability in stress hematopoiesis [10,37,38]. Modeling the pathogenic human single-nucleotide variant in mice modestly reduced HSCs in the embryo, and this alteration was not sustained into adulthood [37]. Whereas steady-state HSC levels are normal in +9.5 Ets^−/− adult mice, 5-fluorouracil (5-FU)–induced myeloablation revealed impaired HSPC expansion and abrogation of HSC long-term repopulating activity in a transplantation assay. +9.5 Ets^−/− mice were more likely to succumb to chronic stress induced by repeated doses of polyinosinic:polycytidylic acid (pI:C). Hematopoietic regeneration capacity was more severely impaired in compound heterozygous mice with a deletion of E-box and GATA motifs from one allele and the ETS single-nucleotide variant on the other allele [38]. The ETS motif variant constitutes a “conditionally-pathogenic allele” that is pathogenic only when triggered by secondary genetic or epigenetic aberrations [38].

Unlike the +9.5, the Gata2 −77 enhancer is dispensable for HSC generation in the AGM, and embryonic lethality is delayed until after E15.5 in −77^−/− mice [39]. −77 chromatin is largely inaccessible in the lineage⁻ (Lin⁻)Sca1⁺c-Kit⁺ multipotent population enriched in HSCs and multipotential progenitors (MPPs), whereas the chromatin is accessible in common myeloid progenitors (CMPs) and granulocyte-monocyte progenitors (GMPs). Deletion of the −77 substantially reduced Gata2 mRNA in these populations. As the −77 enhancer is not essential for HSC generation, this model provided a unique opportunity to analyze GATA2 mechanisms in later stages of hematopoiesis. −77^−/− fetal livers are hypocellular due to a depletion of Ter119⁺ erythroid cells and megakaryocyte-erythroid progenitors (MEP) [39]. Despite the ~80% decrease in Gata2 transcription, fetal liver GMP numbers were unaltered, while CMPs increased [40*]. In vitro differentiation to assess progenitor functionality revealed that −77^−/− fetal liver progenitors produced colonies containing predominantly macrophages, and the generation of monocytic progeny increased with a commensurate decrease in lymphoid progeny in vivo [39]. GATA2-null hematopoietic progenitors derived from human embryonic stem cells retained macrophage, but not granulocytic or erythroid, colony forming potential [41]. Consistent with loss of granulocytic colony forming potential in −77^−/− embryos, granulocyte progenitor (GP) and monocyte progenitor (MP) populations within the GMP pool, which normally exist in similar proportions, become shifted disproportionately to favor MPs [40*,42]. Increased numbers of monocyte-dendritic cell progenitors (MDP) and common dendritic progenitors (CDP) contributed to the increased cellularity of the −77^−/− CMP pool.

Proteomic and transcriptomic (with bulk and single CMP/GMP cells) analyses on −77^−/− fetal liver progenitors revealed innate immune gene induction resembling a response to interferon signaling [40*]. The upregulated genes included the transcription factor IRF8, a regulator of interferon signaling and monocyte and dendritic cell differentiation (see below). In single cells within the CMP/GMP population, Gata2 mRNA inversely correlated with Irf8 mRNA [42]. Genetic ablation of Irf8 in −77^−/− embryos restored CDP and MP to near normal levels in the fetal liver while expanding the GP population [40*]. Irf8 ablation did not alter the high MDP numbers. In vitro differentiation with −77^−/− and Irf8^−/− single- and double-mutant progenitors demonstrated that failure of −77^−/− Ly6C⁻ GMPs to generate granulocytic colonies was partially rescued by IRF8 loss. The enhanced differentiation of −77^−/− CMPs to CD11C⁺CD86⁺ dendritic cells was restored to wildtype levels in the double mutant. These results support a model in which GATA2 loss in fetal progenitors leads to IRF8 upregulation, which contributes to defective hematopoiesis during embryogenesis in −77^−/− mice. In the text below, we describe links between GATA2 and IRF8 mechanisms and pathological implications.

Physiological and Pathological IRF8 Functions

IRF8 (originally termed interferon consensus sequence-binding protein, ICSBP) was discovered as an inducer of IFNγ response genes [43] and demonstrated to be essential for myeloid differentiation [44,45]. IRF8-deficient mice develop a chronic myelogenous leukemia-like disorder with neutrophil and granulocyte progenitor accumulation [45]. Human IRF8 mutations cause susceptibility to infection, monocyte and DC deficiency and lymphoid defects [46–48]. The roles of IRF8 in hematopoietic development, immunity and inflammation have been studied extensively [49–52]. IRF8 partners with a cohort of transcriptional regulators to modulate its own expression to direct monocytic and dendritic cell differentiation and to mediate the capacity of these cells to respond to IFNγ signaling during infection. Though IRF8 possesses a helix-turn-helix DNA binding motif, interactions with other DNA binding partners, mediated by its IRF association domain (IAD), confer added specificity to chromatin occupancy [43,53]. IRF8 and its partners can bind composite motifs: IRF1(or IRF2)-IRF8 complexes bind IFN-stimulated response elements (ISREs); PU.1-IRF8 complexes bind ETS-IRF composite elements (EICEs); activator protein 1 (AP-1) family members complexed with IRF8 bind AP1-IRF composite elements (AICEs) [54,55**]; and the lymphoid-restricted IRF factor Pip forms a complex with PU.1 [56].

Irf8 expression is low in myeloid-biased MPP subset 3 (MPP3), lymphoid-primed multipotent progenitors (LMPP) and GMP, whereas its expression increases in MDP and CMP to promote differentiation into monocytes [57**]. Further increases in Irf8 expression are required in CDP to generate cDC type 1 (cDC1) and plasmacytoid DC (pDC).

Modulation of Irf8 expression during myeloid differentiation is achieved through the actions of cell-type specific enhancer complexes. An Irf8 enhancer 56 kb downstream from the transcriptional start site (+56) is active throughout the myeloid lineage and upregulates Irf8 as early as the MPP3 and LMPP stages [57**]. This enhancer includes ETS and RUNX motifs and confers the high Irf8 expression to generate dendritic cells, while being dispensable for monocyte differentiation.

BAC transgenic analysis identified an Irf8 −50 enhancer that requires PU.1 occupancy to induce a higher-order chromatin transition that brings the −50 site near the Irf8 promoter [58]. ChIP-seq data from myeloid progenitors and macrophages illustrates PU.1 occupancy at the Irf8 locus including −50 and +56 kb. Loss-of-function studies, involving small hairpin RNA (shRNA)-mediated knockdown of PU.1 in RAW264.7 cells, together with rescue assays in the PU.1^−/− myeloid progenitor line PU.1ER, provide evidence that PU.1 activates Irf8 transcription [58]. Mice lacking PU.1 motifs within the −50 enhancer produced monocytes and macrophages with reduced IRF8 levels and an impaired capacity to respond to Salmonella infection [59]. The −50 deletion did not alter IRF8 levels of MDP, CDP and pre-cDC1 or impair the generation of these cells.

For IRF8-dependent dendritic cell differentiation, enhancers at +41 and +32 kb are required for specification and development cDC1, respectively [59–61]. The +32 enhancer maintains high level Irf8 expression in cDC1 via AICE motifs engaged by BATF3 [60].

GATA2/PU.1/IRF8 Axis

GATA2 and IRF8 are expressed reciprocally during hematopoiesis, with GATA2 primarily in HSCs, multipotent progenitors, and erythroid progenitors and IRF8 being highest in progenitors with restricted monocyte and dendritic cell potential. Reciprocal expression also characterizes individual myeloid progenitor cells that express Gata2 and Irf8 [42]. The mechanism by which GATA2 represses Irf8 expression was unclear. ATAC-seq analysis in HoxB8-immortalized −77^−/− progenitors (with high Irf8 expression) revealed that Irf8 −50 and +56 enhancers, but not +32 and +41 DC enhancers, are accessible [62**]. Genetic rescue by expressing physiological levels of GATA2 downregulates Irf8 and decreases accessibility. GATA2-mediated regulation of accessibility at these enhancers is not associated with GATA2 occupancy at these sites. Thus, GATA2 may oppose other mechanisms that establish or maintain Irf8 expression.

PU.1 regulates the expression of Irf8 and numerous other myeloid genes and B-lineage genes. Though GATA2 occupies the Spi1 locus encoding PU.1 and regulates its expression in the G1ME erythroid progenitor cell line [63], Spi1 transcripts and PU.1 protein levels are similar in wildtype and −77^−/− primary myeloid progenitors [39,42]. GATA2 interacts with and antagonizes PU.1 [64,65] and reduced GATA2 levels in −77^−/− progenitors may therefore derepress PU.1 activity, leading to Irf8 transcriptional activation. The importance of PU.1 for activating Irf8 and other myeloid and B-lineage genes in −77^−/− progenitors was demonstrated by deletion of a −14 kb enhancer from Spi1, which reduced Spi1 mRNA 2-fold and lowered Irf8 expression similarly [62**]. Suppression of PU.1 activity, but not expression, appears to be a broadly utilized mechanism by which GATA2 represses transcription.

In −77^−/− progenitors, elevated PU.1 activity and upregulated Irf8, may profoundly affect the transcriptome, either through the individual activities of these factors or via PU.1-IRF8 complexes at composite motifs. Loci of many upregulated genes in −77^−/− progenitors are occupied by PU.1 and/or IRF8 at sites of GATA2-regulated chromatin accessibility [42,62**]. GATA2 represses multiple TLR family members (Tlr1, Tlr2 and Tlr6) in fetal liver progenitors and HoxB8-immortalized cell lines [66*]. PU.1 is implicated in activating Tlr1 and Tlr2 expression [67]. PU.1 serves a critical role in the innate defense against Aspergillus fumigatus via dendritic cell-associated C-type lectin receptor-1 and Toll-like receptors-2 and 4 in macrophages [68]. Whether ectopically elevated expression of genes encoding TLRs and other innate immune regulators further impairs the aberrant phenotype of GATA2-deficient progenitors is unresolved.

Inflammation and Loss of GATA2 in HSPCs

−77 enhancer-mutant fetal liver myeloid progenitors exhibit upregulated expression of Irf8 and other innate immune genes, including TLRs, increased expression of IL-6 receptors Il6st and Il6ra, and reduced expression of Csf2rb encoding a subunit of the GM-CSF receptor that is shared with other cytokine receptors [62**,66*] (Figure 1). These changes render the cells hyperresponsive to signaling by IFNγ, IL-6 and TLR1/2 and TLR2/6 agonists and less responsive to GM-CSF [62**,66*]. In mast cell lines, GATA2 functions through an IL6 enhancer to increase basal and induced IL-6 expression [69]. Compared to −77^−/− progenitors that were rescued by GATA2 expression, HoxB8-immortalized −77^−/− progenitors displayed a more robust induction of Irf8 expression in response to IFNγ and of TLR target genes Tnf and Ccl3 in response to TLR1/2 and TLR2/6 agonists [42,66*]. IFNγ-TLR signaling crosstalk further elevated expression of a gene cohort including those encoding cytokines and chemokines. The immortalized −77^−/− cells were hyperresponsive to IL-6-dependent induction of STAT3 phosphorylation (pSTAT3) and exhibited greater monocytic differentiation [62**]. However, GM-CSF had reduced capacity to induce pSTAT5 in the mutant cells that exhibited reduced granulocytic differentiation.

Figure 1.

Model of how GATA2 deficiency in fetal myeloid progenitors disrupts genome regulation and differentiation. GATA2 establishes and maintains a genetic network that supports the generation of diverse myeloid cell progeny (top panel). Loss of GATA2 level/activity triggers an alarm that instigates upregulation of a host of innate immune genes through a mechanism involving derepression of PU.1 transcriptional activity (bottom panel). While genetic analysis demonstrated the vital importance of this mechanism [62**,66*], one cannot rule out the involvement of additional transcriptional regulators, which have not been identified, in GATA2-dependent repression of innate immune genes. Upregulation of IRF8 preferentially promotes monocytic/dendritic cell differentiation of the GATA2-deficient cells, while increased expression of TLR and IL-6 receptor subunits renders the progenitors and their progeny hypersensitive to extrinsic stimuli that activate these receptor systems.

While progress has been made on elucidating mechanisms underlying how decreased GATA2 in fetal progenitors triggers an alarm that instigates an innate immune response, there are many unanswered questions regarding the resulting functional consequences. Innate immunity protects against pathogens e.g., bacteria, viruses, fungi, and parasites, using pattern recognition receptors, including TLRs, to mount a response [70,71]. Besides promoting an immediate response to infection, inflammatory signaling can act on HSPCs to alter proliferation, differentiation and cytokine production, though the response may differ depending on whether the cells are subjected to acute or chronic inflammation [72]. HSPC pools can be biased toward production of myeloid populations in response to inflammation due, in part, to the upregulation of PU.1 as occurs in HSCs upon exposure to TNFα or IL-1, which are upregulated by TLR signaling [73]. Loss of GATA2 and the resulting derepression of PU.1 may further sensitize HSPCs to inflammatory signals or induce an intrinsic immune response in the absence of external stimulation. Secondarily, PU.1 activation of Irf8 would reinforce a program favoring production of monocytic and dendritic cells and upregulation of interferon-responsive genes. In −77^−/− embryos, cell intrinsic upregulation of innate immune genes in myeloid progenitors, due to loss of GATA2 suppression of the PU.1/IRF8 regulatory network, may initiate an inflammatory response that impacts function of cells in the microenvironment, though this has not been demonstrated.

Conclusion

GATA2-mediated activation of transcription through direct engagement of target genes constitutes a canonical mechanism of controlling hematopoietic stem and progenitor cell generation and function. By contrast, mechanisms of GATA2-mediated repression have been elusive [62**], contrasting with the well-studied mechanism in which GATA1 utilizes FOG1 to repress transcription [1,20,74]. This review describes a model in which GATA2 restricts an IRF8- and PU.1-dependent monocytic gene expression program and prevents PU.1 from activating Irf8 and other myeloid and B-lineage genes. The innate immune response instigated by GATA2 loss in fetal progenitor cells serves as an alarm to trigger molecular/cellular transitions that impact the capacity of progenitors and their progeny to elaborate cytokines/chemokines, which impact HSPCs and diverse cell types in the microenvironment. This amalgamated hematopoietic cell-intrinsic and -extrinsic mechanism may customize hematopoietic cell generation requirements to contend with the unique pathological state of GATA2 deficiency during fetal development. It will be instructive to compare this mechanism with the corresponding adult mechanism to define common threads and/or the unique cast of molecular/cellular characters.

Key Points.

• GATA2 suppresses innate immune gene expression and signaling in fetal myeloid progenitors.

• GATA2 deficiency in fetal myeloid progenitors triggers an alarm that instigates innate immune signaling and distorts differentiation.

• GATA2 represses innate immune genes, encoding IRF8 and other myeloid and B-lineage proteins, by antagonizing PU.1.

• GATA2 suppression of IRF8 expression impedes monocytic and dendritic cell differentiation.