GATA2 decommissioning of enhancers: A mechanism to constrain inflammatory signaling

SUMMARY

GATA2 establishes transcriptomes governing hematopoietic stem/progenitor cell development. In progenitors, GATA2 represses inflammatory genes (Il6st and Il6ra) encoding IL6ST/GP130 and IL6RA receptor subunits mediating IL-6 signaling. While IL6ST heterodimerizes with IL6RA, IL-11, IL-27, oncostatin M, and leukemia inhibitory factor receptors, IL6RA heterodimerizes exclusively with IL6ST to confer IL-6 signaling. As GATA2-dependent repression is not well understood, we devised a multi-omics strategy to elucidate mechanisms underlying repression and applied the approach to the cytokine/chemokine receptor gene family. Identifying accessible distal and intronic chromatin sites in GATA2-deficient (GATA2^low), but not wild-type, progenitors unveiled that GATA2 represses Il6st and Il6ra enhancers. GATA2 loss instigates a chromatin transition to increase accessibility, occupancy by the hematopoietic activator PU.1, transcription, and cell signaling capacity. Ablating Il6st enhancers in GATA2^low cells abrogates Il6st expression and IL-6 and IL-27 signaling. Il6ra enhancer loss selectively abrogates Il6ra expression and IL-6 signaling. Thus, GATA2 decommissions enhancers to constrain inflammatory signaling.

Graphical abstract

In brief

Jung et al. discovered that the critical regulator of blood cell development, GATA2, suppresses the activity of transcriptional enhancers that activate genes encoding key regulators of inflammatory signaling. This GATA2-dependent process to prevent ectopic inflammatory signaling illustrates an important intersection between GATA transcription factor and inflammatory mechanisms.

INTRODUCTION

Inflammation resulting from infection or diverse stresses alters the generation and function of hematopoietic stem and progenitor cells (HSPCs) through cell-autonomous and non-cell-autonomous mechanisms.^1,2 HSPC mutations disrupting the expression or activity of hematopoietic regulators can elevate inflammation and inflammatory signaling. Loss of GATA2, a master regulator of hematopoiesis,³ upregulates inflammatory (e.g., Irf8, Tlr1, Tlr2, and Tlr6) and cytokine receptor genes (e.g., Il6st and Il6ra) in fetal progenitors.^4–7 Upregulated interferon-gamma (IFNγ) and toll-like receptor (TLR) pathway components in GATA2-deficient (GATA2^low) fetal progenitors render the progenitor genome hypersensitive to inflammation.⁸ Tet2^+/− mutation increases IL-1 production and mutant clone expansion.⁹ Tet2 ablation elevates IL-6, and the mutant HSPCs are hypersensitive to inflammation.¹⁰ Runx1 mutations render granulocyte monocyte progenitors and neutrophils hypersensitive to inflammation.^11,12 A murine model of Runx1 mutant familial platelet disorder exhibits systemic inflammation¹³ and blocking inflammatory signaling in patient cells rescues hematopoiesis.¹⁴ Whether mechanisms mediating inflammatory signaling in these diverse contexts are shared is unknown.

Analogous to the pathogenic consequences of human RUNX1 germline genetic variation,¹⁵ GATA2 genetic variation causes GATA2 deficiency syndrome involving myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML).^16–19 Pathogenic variants commonly alter the zinc-finger domain and, less frequently, other GATA2 coding and enhancer sequences.^20–23 GATA2 transcription is activated by a conserved enhancer 77 kb upstream of the murine Gata2 promoter (−77) and intronic +9.5 kb enhancer, both occupied by GATA2 and additional factors.^24,25 These enhancers are essential for hematopoietic stem cell (HSC) emergence during embryogenesis,^22,26 suppressing the development of bone marrow failure and AML in mouse and man.^21,22

Ablating the −77 kb enhancer enabled the generation of GATA2^low primary and ER-HOXB8-immortalized progenitors (hi-77^−/−) with ~75% lower GATA2^4,5 and rescue assays in primary and immortalized cells to quantify wild type vs. GATA2 pathogenic variant activities; qualitatively similar results were obtained in both systems.^5,27–29 Despite similar numbers of GATA2-activated and -repressed genes in the hi-77^−/− cell rescue assay, GATA2-regulated chromatin sites revealed more WGATAR motifs at GATA2-activated vs. repressed loci. GATA2 occupancy was less frequent at GATA2-repressed vs. -activated loci.⁵

By contrast to repression mediated by the erythroid and megakaryocytic GATA factor GATA1,^30–33 how GATA2 represses transcription is unclear. GATA1 regulation of target genes often requires GATA1 binding to the Friend of GATA1 (FOG1) coregulator,^33–36 which binds the nucleosome remodeling and deacetylase (NuRD) repressive complex.^37–39 Certain GATA1 targets do not require FOG1.^33,36,40 FOG1 levels are low in GATA2-expressing erythroid precursors before GATA1 upregulation, and little to no FOG1 exists in GATA2-expressing myeloid progenitors. The relationships between GATA2-induced repression, FOG1-independent GATA1 repression, and repression by other GATA factors are unknown. GATA2-mediated repression can involve antagonism of the ETS transcription factor PU.1,⁵ which is implicated in GATA1-mediated repression.^41–44 GATA1 binds the PU.1 ETS domain through a basic tail within its carboxy-terminal zinc (C-) finger, displacing c-JUN, which is proposed to function as a coactivator in this context, thus reducing PU.1 transcriptional activity.^43,45 GATA1-PU.1 interaction recruits a repressive complex containing retinoblastoma protein (pRB), histone methyltransferase SUV39H, and heterochromatin protein 1 (HP1).⁴⁶ The interaction also prevents acetylation of lysine residues in the GATA1 C-finger basic tail by the acetyltransferase CREB-binding protein (CBP), which participates in GATA1-regulated erythroid differentiation.^47,48 The basic tail of the GATA1 C-finger is highly homologous to that of GATA2 and the PU.1-binding domain of PU.1-coactivator c-JUN.⁴³ Whether GATA2 inhibits PU.1 through a similar mechanism is unknown. GATA2 loss elevates PU.1 activity, which derepresses myeloid and B-lineage genes.⁵ Single-molecule imaging demonstrated that GATA1, GATA2, and PU.1 mRNAs are co-expressed in the same single cell of common myeloid progenitors (CMPs), granulocyte monocyte progenitors (GMPs), and megakaryocyte erythroid progenitors (MEPs),⁴⁹ providing a biological rationale for cross-regulation. It is unclear whether GATA2-PU.1 antagonism explains all GATA2-mediated repression, which occurs at chromatin sites of the repressed genes or involves mutually exclusive chromatin binding.

The GATA2-repressed genes Il6st and Il6ra encode IL-6 receptor heterodimeric subunits IL6ST and IL6RA.^50,51 Elevated IL-6 signaling in hi-77^−/− cells was reversed by the expression of GATA2 but not a pathogenic human clinical variant with a nine-amino acid insertion between the zinc fingers (9aa-ins).^5,27 While the 9aa-ins variant retained some capacity to bind chromatin at select GATA2 target genes, it was defective in occupying Il6st and Il6ra loci.⁵ As excessive IL-6 signaling is implicated in bone marrow failure and AML,^10,52,53 GATA2 repression of IL6 receptor genes has pathological implications. IL6ST/GP130 is a shared receptor subunit for IL-6 and IL-6 family members IL-11, IL-27, oncostatin M, and leukemia inhibitory factor.⁵¹ Dysregulated IL6ST in GATA2^low progenitors may therefore disrupt a signaling network involving, but not limited to, IL-6. Il6st ablation in mice reduces hematopoietic progenitors in the liver, causes myocardial defects, and is embryonic lethal.⁵⁴ Dominant-negative human IL6ST mutations truncate the cytoplasmic domain containing STAT3-binding tyrosine residues, impair STAT3 activation, and cause immunodeficiency.⁵⁵ A mouse homozygous knock-in Il6st Y757F variant that eliminates inhibitor (SOCS3 and SHP2) binding to phosphorylated tyrosines causes STAT3 hyperactivation with splenomegaly, lymphadenopathy, and thrombocytosis.⁵⁶ Stringent control of IL6ST-mediated signaling is vital for normal physiology.

Herein, we analyzed mechanisms underlying the inverse relationship between GATA2 and inflammatory signaling and provided evidence for a paradigm in which GATA2 renders enhancers dormant (GATA2-mediated enhancer decommissioning, GED), thereby constraining cytokine/chemokine receptor gene transcription and signaling. GATA2 reduced chromatin accessibility at PU.1-occupied distal and intronic conserved sites of GATA2-repressed genes that harbored enhancer attributes. When GATA2 was limiting, PU.1 occupied and activated Il6ra and Il6st enhancers. Normal GATA2 levels restricted PU.1 chromatin occupancy and decommissioned enhancers. GED occurs in a broader cadre of inflammatory genes.

RESULTS

GATA2-induced enhancer decommissioning restricts expression of Il6st encoding the shared cytokine receptor subunit IL6ST/GP130

Our prior multi-omics analysis⁵ demonstrated Il6st upregulation in GATA2^low hematopoietic progenitors, GATA2 rescues Il6st repression, and GATA2 loss increases chromatin accessibility at sites 24, 30, and 35 kb upstream (−24 kb, −30 kb, and −35 kb) of the transcriptional start site (Figure 1A). Upregulated Il6st expression elevated IL-6-induced STAT3 activation in GATA2^low progenitors (Figure 1A). As GATA2 antagonizes PU.1 to repress select myeloid/B-cell lineage genes,⁵ we asked if PU.1 derepresses Il6st when GATA2 is limiting. GATA2 loss does not significantly alter PU.1 levels.⁵ We used a rescue system with GATA2-and PU.1-deficient progenitors in which Spi1 (encoding PU.1) −14 kb upstream regulatory element (URE) was ablated in hi-77^−/− progenitors (hi-77^−/− URE^−/−), which decreased Spi1 expression by 50%.⁵ Previously, we generated RNA sequencing (RNA-seq) and CUT&Tag (GATA2 and PU.1) datasets with hi-77^+/+ (WT), hi-77^−/− (GATA2^low), and hi-77^−/− URE^−/− (GATA2^lowPU.1^low) cells (Figure 1B) and analyzed how progenitor genomes sense inflammation.⁵⁷ Herein, we used these datasets to elucidate how GATA2 represses myeloid/B-lineage genes in the steady state. Relative to GATA2^low cells, Il6st expression was ~85% lower in GATA2^lowPU.1^low cells (p < 0.0001) (Figure 1C). We asked if GATA2 and PU.1 occupy accessible Il6st distal sites. Il6st distal sites were largely inaccessible in WT cells, and 75% GATA2 reduction markedly increased chromatin accessibility (Figure 1D). The residual GATA2 occupied −24 kb and −35 kb sites in GATA2^low cells (Figures 1D and 1E). GATA2 occupancy at Il6st −24 kb and −35 kb sites was considerably lower in WT and GATA2^lowPU.1^low cells. PU.1 occupied both sites only in GATA2^low cells. H3K4me3 was enriched at the Il6st promoter only in GATA2^low cells, consistent with transcriptional activity when GATA2 is limiting (Figure 1D). We compared these results to GATA2-mediated activation. GATA2-activated Cebpe +6 kb and Ms4a3 +27.6 kb enhancers²⁸ were accessible in WT and GATA2^low cells (Figure S1). GATA2 occupancy at the enhancers was significantly lower in GATA2^low cells compared to WT. In GATA2^lowPU.1^low cells, residual GATA2 retained the capacity to occupy Cebpe, but considerably less was retained at Ms4a3. PU.1 did not occupy the enhancers in any context (Figure S1).

Figure 1. Opposing GATA2 and PU.1 functions establish Il6st expression and IL-6 signaling in GATA2-deficient progenitors.

(A) GATA2 loss increases chromatin accessibility at Il6st and induces Il6st expression and IL-6/STAT3 signaling.⁵

(B) Assays to analyze genomic (RNA-seq [GEO: GSE279155] and CUT&Tag [GEO: GSE279156]) and biological (western blot and flow cytometry) functions of GATA2 and PU.1 with WT, GATA2^low (G^low), and GATA2^lowPU.1^low (GP^low) ER-HOXB8-immortalized myeloid progenitors.

(C) Il6st TPMs in WT, G^low, and GP^low cells mined from RNA-seq.

(D) GATA2 and PU.1 CUT&Tag in WT, G^low, and GP^low cells.

(E) CUT&Tag signals from all of the four replicates at Il6st −24, −30, and −35 kb.

(F) Representative western blot to detect IL-6-induced STAT3 phosphorylation (pSTAT3) in GP^low, WT, and G^low cells (n = 3).

(G) p-STAT3/STAT3 quantification (n = 3) of (F). Data were normalized to IL-6-treated G^low.

(H) Representative flow cytometry plots of undifferentiated (Diff−, +β-est) and differentiated (Diff+, −β-est) WT (n = 5), G^low (n = 5), and GP^low (n = 4) cells.

(I) Quantification of undifferentiated (CD11b⁻ CD115⁻), granulocytic (CD11b⁺CD115⁻, Gr), and monocytic (CD11b⁺CD115⁺, Mo) progeny.

Error bars represent mean ± SEM. Statistical comparisons utilized one-way ANOVA followed by Tukey’s multiple comparisons test (C) or followed by Dunnett’s multiple comparisons test (G) and (I). **p < 0.01, ***p < 0.001, and ****p < 0.0001; n.s., not significant. See also Figure S1.

To test if PU.1-activated Il6st expression has important functional consequences in GATA2^low progenitors, we compared IL-6 signaling and differentiation in GATA2^low and GATA2^lowPU.1^low progenitors (Figure 1B). Compared to elevated activity in GATA2^low progenitors, a 50% PU.1 reduction in GATA2^lowPU.1^low progenitors reduced IL-6-induced STAT3 phosphorylation by 97% (p < 0.0001) to a level resembling WT cells (Figures 1F and 1G). Although differentiating GATA2^low progenitors yielded monocytic and granulocytic progeny, GATA2^lowPU.1^low progenitors differentiated into exclusively granulocytic progeny (Figures 1H and 1I).

We hypothesized that accessible, distal chromatin sites represent enhancers mediating Il6st expression and IL-6 signaling when GATA2 is limiting. Enhancers regulating Il6st and GATA2 activity to suppress enhancers have not been described. We identified sites with enriched enhancer marks H3K27ac and H3K4me1 and sites with mouse-to-human conservation.^58–60 Mining chromatin immunoprecipitation sequencing (ChIP-seq) data revealed PU.1 occupancy and H3K27ac and H3K4me1 enrichments in murine bone marrow-derived macrophages⁵⁸ and monocytes⁶⁰ at −24 kb, −30 kb, and −35 kb. These enrichments were not detected with myeloid progenitors⁵⁹ (Figure S2A). The −24 kb and −30 kb, but not −35 kb, sites were conserved in human and corresponded to −43 kb and −56 kb, respectively (Figures S3A and S3B). The other IL-6 signaling receptor subunit Il6ra, harbored an accessible chromatin region 1.7 kb downstream (+1.7 kb) at intron 1 only when GATA2 was limiting.⁵ The +1.7 kb site corresponded to human IL6R +2.2 kb (Figure S3C), and H3K27ac and H3K4me1 were enriched in all cell types (Figure S2B). These studies identified GATA2- and PU.1-bound candidate enhancers to establish IL-6 signaling.

Mechanisms underlying GATA2-induced enhancer decommissioning

As the Il6st-encoded receptor subunit mediates signaling by multiple IL-6 family cytokines,⁵¹ mechanisms governing its expression would establish and/or maintain multiple signaling systems. To test whether the candidate enhancers are essential for, modulate, or do not regulate Il6st transcription, we used CRISPR-Cas9 to ablate the −24 kb and −30 kb sites in GATA2^low progenitors with elevated Il6st expression (Figure S4A). We generated six −24 kb^−/− clonal lines (Δ−24), five −30 kb^−/− lines (Δ−30), and four −24^−/−;−30 kb^−/− double-mutant lines (Δ–24Δ–30) (Figures S4B and S4C). Relative to the GATA2^low control, −24 kb deletion in GATA2^low progenitors decreased Il6st expression by 68% (p < 0.0017) (Figure 2A). −30 kb deletion did not significantly alter expression (Figure 2A). Deletion of −24 and −30 kb sites collectively down-regulated the expression by 87% (p < 0.0001). Individual and dual site deletions did not affect Il6ra and Irf8 expression (Figure 2A).

Figure 2. Il6st enhancers are active when GATA2 is limiting and inactive when GATA2 levels are normal.

(A) GATA2-repressed gene (Il6st, Il6ra, and Irf8) mRNA levels in GATA2^low (G^low; n = 7) vs. GATA2^low cells lacking −24 kb (Δ–24; n = 6), −30 kb (Δ–30; n = 5), or −24 and −30 kb (Δ–24Δ–30; n = 4) sites.

(B) GATA2-repressed gene (Il6st, Il6ra, and Irf8) mRNA levels in undifferentiated (Diff–, +β-est) and differentiated (Diff+, −β-est) G^low control (n = 7) vs. G^low Δ–24 (n = 6), Δ–30 (n = 5), or Δ–24Δ–30 (n = 4) cells. Each data point represents an individual clone.

(C) Models depicting GATA2-induced Il6st repression.

(D) mRNA levels of GATA2-repressed genes (Il6st and Il6ra) in undifferentiated (Diff–, +β-est) and differentiated (Diff+, −β-est) G^low (n = 5), WT(n = 6), and WT cells lacking −24 and −30 kb sites (WT Δ–24Δ–30; n = 6). Each data point represents an individual clone.

(E) Representative western blot of endogenous PU.1 and exogenous HA-PU.1 in WT cells infected with empty control virus (WT empty), GATA2^low with empty control (G^low empty), GATA2^lowPU.1^low with empty control (GP^low Empty), and GATA2^lowPU.1^low with HA-PU.1 virus (GP^low PU.1) (n = 6).

(F) PU.1 quantification (n = 6) of (E). Data are normalized to WT Empty.

(G) mRNA levels of Il6st, Il6ra, and PU.1-activated genes (Cd79b, Rag2, and Tifab) in WT empty vs. G^low empty vs. GP^low empty vs. GP^low PU.1 (n = 6).

Error bars (A), (B), (D), and (G) represent mean ± SEM. Box-and-whisker plots (F) are presented with bounds from the 25^th to the 75^th percentiles, the median line, and whiskers ranging from minimum to maximum values. Statistical calculations utilized one-way ANOVA followed by Tukey’s multiple comparison test (A), one-way ANOVA followed by Dunnett’s multiple comparison test (F), paired Student’s t-tests (B) and (D), or unpaired Student’s t-test (G). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; n.s., not significant. See also Figures S2, S3, S4, and S5.

To test if the enhancers activate transcription exclusively in GATA2^low undifferentiated progenitors or also in differentiated cell progeny, we quantified Il6st, Il6ra, and Irf8 expression in undifferentiated and differentiated GATA2^low cells with or without the enhancers. Differentiation increased Il6st, Il6ra, and Irf8 expression. Ablating −24 and −30 kb enhancers collectively reduced Il6st, but not Il6ra and Irf8, expression (Figure 2B). While single deletions of −24 kb or −30 kb reduced IL-6 signaling (Figures S5A and S5B), the double deletion abrogated IL-6 signaling in all mutant clones (Figures S5C and S5D). Flow cytometric analysis revealed the generation of granulocytic, but not monocytic, progeny with mutant clones (Figures S5E and S5F). Thus, when GATA2 levels are limiting in undifferentiated and differentiated cells, −24 kb and −30 kb enhancers elevate Il6st and IL-6 signaling.

We extended the multi-omics strategy to determine if GATA2 controls IL-6 signaling exclusively through regulation of genes encoding IL-6 receptor subunits or also by repressing loci encoding additional proteins operating in the IL-6 signaling pathway and/or regulating IL-6 generation/function. We identified candidate enhancers at Hck and Slc11a1 (Figure S6A). Hck encodes hematopoietic cell kinase (HCK) that physically and functionally interacts with IL6ST,⁶¹ and IL-6 activates HCK kinase activity.⁶² Slc11a1 encodes a metal transporter NRAMP1,⁶³ and its ablation decreases IL-6 generation from CD11c⁺CD103⁻ dendritic cells.⁶⁴ RNA-seq analysis revealed that both genes are elevated in GATA2^low progenitors, and GATA2 rescue decreased Hck and Slc11a1 expression by 3-fold (p < 0.0001) and 21-fold (p < 0.0001), respectively (Figure S6B). RT-qPCR analysis confirmed the RNA-seq data (Figures S6C and S6D). Thus, beyond repressing the expression of genes encoding IL-6 receptor subunits, GATA2 represses other components implicated in IL-6 function.

We considered models to explain how GATA2 represses Il6st: (1) GATA2 functions at −24 kb and −30 kb enhancers to repress Il6st; (2) GATA2 antagonizes enhancer occupancy by PU.1, an Il6st activator (Figure 2C). We ablated −24 kb and −30 kb sites in WT cells to ask if the sequences confer repression and if their deletion de-represses Il6st (Figure S4C). Eliminating both enhancers in WT did not de-repress Il6st, which remained at undetectable levels (Figure 2D). The enhancers are therefore not active in cells expressing normal GATA2 levels, or essential mediators of Il6st and Il6ra transcription are lacking. Although PU.1 occupancy at −24 kb infers PU.1 function through the site, we conducted genetic rescue with PU.1 in GATA2^lowPU.1^low cells to establish if PU.1 reactivates the reduced Il6st expression. PU.1 expression in GATA2^lowPU.1^low cells elevated PU.1 protein levels to resemble WT and GATA2^low progenitors (Figures 2E and 2F). Under these conditions, the expression of Il6st and Il6ra and PU.1 target genes Cd79b, Rag2, and Tifab⁵ was reactivated (Figure 2G). Thus, GATA2 decommissions enhancer function and represses Il6st by antagonizing PU.1.

GATA2-induced enhancer decommissioning impacts multiple inflammatory genes

We tested whether GATA2 antagonizes enhancers at other cytokine/chemokine receptor genes to more broadly constrain inflammatory signaling. Using our RNA-seq and ATAC-seq data,⁵ we devised a strategy to identify candidate GATA2-repressed enhancers genome-wide and applied the strategy to all members of the cytokine/chemokine receptor gene family expressed in progenitors (95) (Figure 3A). We eliminated genes with low expression and prioritized GATA2-regulated genes (transcript per million, TPM > 0.5, |log₂(fold change)| ≥1) (Figure 3B). The majority (7 of 8) of GATA2-repressed cytokine/chemokine receptor genes (Il6ra, Csf1r, Flt3, Ccr2, Tnrsf1a, and Cxcr2) harbored accessible chromatin sites when GATA2 was limiting or functionally defective (9aa-Ins pathogenic variant) (Figure 3C). Our analysis revealed the Csf1r +2.6 kb enhancer that had been validated functionally but not demonstrated to be GATA2-repressed.⁶⁵

Figure 3. GATA2-induced enhancer decommissioning in the cytokine/chemokine receptor gene family.

(A) Cytokine/chemokine receptor genes and TPM values from hi-77^−/− (G^low) empty vs. G^low GATA2 RNA-seq.⁵

(B) Log₂(fold change) of cytokine receptor genes (cutoff: TPM > 0.5 in either G^low empty or G^low GATA2) that are GATA2-activated or GATA2-repressed.

(C) Representative ATAC-seq profiles of GATA2-repressed cytokine/chemokine receptor genes activated upon loss of GATA2 (G^low empty) or expression of a pathogenic 9aa-Ins variant (G^low 9aa-ins).

(D) mRNA levels of GATA2-repressed cytokine/chemokine genes in WT empty, G^low empty, and G^low GATA2 cells (n = 8).

(E) Flt3 and Tnfrsf1a expression in G^low control (n = 8) vs. Flt3 +7.4 kb enhancer mutants (Flt3 Δ+7.4; n = 8).

(F) Tnfrsf1a and Flt3 expression in G^low control (n = 8) and Tnfrsf1a +11 kb enhancer mutants (Tnfrsf1a Δ+11; n = 10). Data points in (E) and (F) represent single clones.

(G) Il6ra and Il6st expression in G^low control (n = 10) vs. Il6ra enhancer mutants (n = 4). As deletion sites are proximal to the Il6ra promoter, Il6ra primary transcripts with reverse transcriptase control were analyzed. Il6st mRNA levels were detected as a negative control. Data points represent two biological replicates of each of the two mutant clones.

Error bars represent mean ± SEM. Statistical calculations utilized unpaired one-way ANOVA followed by Tukey’s multiple comparisons test (D) or paired student t-test (E–G). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; n.s., not significant. See also Figures S6, S7, and S8.

GATA2 rescue repressed Csf1r, Il6ra, Flt3, and Tnfrsf1a (Figure 3D). Candidate intronic enhancers at Il6ra +1.7, Flt3 +7.4, and Tnfrsf1a +11 kb were conserved between mouse and human (Figures S3C and S7). We used CRISPR-Cas9 to ablate Il6ra, Flt3, and Tnfrsf1a intronic sites accessible in GATA2^low progenitors (Figure S8). Deleting the candidate enhancers reduced Flt3 expression by 37% (p = 0.003) and Tnfrsf1a by 33% (p = 0.0029) (Figures 3E and 3F). Deleting the Il6ra candidate enhancer reduced the expression by 85% (p < 0.0001) (Figure 3G). CUT&Tag data with WT, GATA2^low, and GATA2^lowPU.1^low cells revealed that PU.1 occupied Flt3, Tnfrsf1a, and Il6ra intronic enhancers. Occupancy was higher with limiting GATA2 (Figures S9A and S9B). Although Flt3 and Tnfrsf1a expression was insensitive to the 50% PU.1 reduction, PU.1 rescue in GATA2^lowPU.1^low cells increased Flt3 expression 1.8-fold (p = 0.0012) and Tnfrsf1a 2-fold (p = 0.0304) (Figure S9C). GATA2-PU.1 antagonism contributes to Flt3 and Tnfrsf1a repression, albeit to a lesser extent than the Il6st and Il6ra enhancers. Identifying distal and intronic sites with GATA2-repressed chromatin accessibility revealed GATA2-repressed enhancers at cytokine/chemokine receptor loci.

To evaluate the relationship between GATA2 and the human enhancers, we analyzed a human GATA2 CUT&Tag dataset that we generated and a published ChIP-seq and ATAC-seq dataset.⁶⁶ Our CUT&Tag data from human erythroid progenitor cells (HUDEP2) detected endogenous GATA2 occupancy at IL6ST −43 kb (mouse −24 kb) and FLT3 +6 kb (mouse +7.4 kb) (Figure 4A). No GATA2 occupancy was detected at MPO encoding myeloperoxidase, an abundant neutrophil marker (Figure 4A). ChIP-seq and ATAC-seq data from human HSPCs⁶⁶ revealed GATA2 occupied IL6ST −43 kb and −56 kb (mouse −24 kb and −30 kb) (Figure 4B). Chromatin was accessible in CMP and MEP populations at −43 kb and HSCs at −56 kb (Figure 4B). GATA2 did not occupy IL6R +2.2 kb (mouse +1.7 kb) in HSPCs (Figure 4C). Chromatin was accessible in the GMP population and, to a lesser extent, HSC and CMP (Figure 4C). GATA2 occupied FLT3 +6 kb (mouse +7.4 kb) in all populations, and chromatin was accessible in HSC, CMP, and MEP (Figure 4D).

Figure 4. GATA2 occupies and regulates human IL-6 receptor enhancers.

(A) Endogenous GATA2 occupancy at IL6ST −43 kb and FLT3 +6 kb in HUDEP2 cells determined by CUT&Tag.

(B–D) GATA2 occupancy and accessible chromatin in HSPCs mined from published ChIP-seq and ATAC-seq⁶⁶ at (B) IL6ST −43 kb and −56 kb (mouse: −24 kb and −30 kb); (C) IL6R +2.2 kb (mouse: +1.7 kb); and (D) FLT3 +6 kb (mouse: +7.4 kb).

(E) Left: representative western blot of endogenous GATA2 in human NOMO-1, THP-1, and MOLM-13 cells. Right: GATA2 quantification (n = 4). Error bars represent mean ± SEM.

(F) GATA2, IL6ST, and IL6R expression in NOMO-1, THP-1, and MOLM-13 cells.

(G) Representative western blot of NOMO-1 control vs. NOMO-1 expressing HA-GATA2.

(H) IL6ST, IL6R, and IL1RL1 expression in NOMO-1 control and NOMO-1 expressing HA-GATA2 (n = 4–5).

Error bars represent mean ± SEM. Statistical analysis utilized one-way ANOVA followed by Dunnett’s multiple comparisons test (E) and (F) or unpaired Student’s t-test (H). *p < 0.05, **p < 0.01, and ***p < 0.001; n.s., not significant.

We tested whether GATA2 represses genes encoding human IL-6 receptor subunits (IL6ST and IL6R) by quantifying GATA2 levels in human AML cell lines NOMO-1, THP-1, and MOLM-13. Relative to murine hi-77^+/+ and hi-77^−/− cells, human NOMO-1, THP-1, and MOLM-13 cells had low, intermediate, and high GATA2 levels, respectively, corresponding to protein levels (Figure 4E). Conversely, IL6ST and IL6R expression was the lowest in MOLM-13, with the highest GATA2 (Figure 4F). Expressing HA-GATA2 in GATA2-low NOMO-1 cells (Figure 4G) decreased IL6ST (p = 0.0124) and IL6R (p = 0.0053) expression 2-fold, while IL1RL1 expression remained unchanged (Figure 4H). Thus, GATA2 constrains IL-6 receptor subunit expression in mouse and human systems.

GATA2-induced enhancer decommissioning constrains cytokine signaling and STAT3 phosphorylation

We asked if Il6st and Il6ra enhancers increase signaling only when GATA2 levels are limiting in progenitors. Because IL6ST is a shared subunit for IL-6 and IL-6 family cytokines, including IL-27⁵¹ (Figure 5A), and the progenitors express Il27ra encoding the IL-27 receptor A-subunit, we tested whether Il6st enhancers control IL-6- and IL-27-mediated STAT3 phosphorylation. Il27ra was expressed similarly in WT empty, GATA2^low empty, and GATA2^low GATA2 conditions and not GATA2-regulated in the rescue assay (Figure 5B). Il27ra expression was comparable in GATA2^low vs. GATA2^low Δ–24Δ–30 (Il6st enhancer deletion) progenitors, yet compared to GATA2^low cells, Il27ra expression was 2.2-fold higher in GATA2^low Δ+1.7 (Il6ra enhancer deletion) progenitors (p = 0.003) (Figure 5B).

Figure 5. Enhancer-dependent expression of the shared cytokine receptor subunit IL6ST and multi-cytokine signaling only when GATA2 is limiting.

(A) IL6ST is a shared subunit of the IL-6 and IL-27 receptors.⁵¹ Fold change of hi-77^−/− Empty (G^low empty) vs. hi-77^−/− GATA2 (G^low GATA2) mined from RNA-seq ⁵ are shown in parentheses.

(B) Left: Il27ra mRNA quantification from GATA2 rescue assay in WT empty, G^low empty, and G^low GATA2 cells (n = 8). Right: Il27ra mRNA analysis of WT, G^low (n = 17), G^low Il6st Δ–24Δ–30 (n = 10), and G^low Il6ra Δ+1.7 cells (n = 5).

(C) Representative western blot to detect IL-6 or IL-27-induced p-STAT3 in WT, G^low, G^low Il6st Δ–24Δ–30, and G^low Il6ra Δ+1.7 cells.

(D and E) p-STAT3/STAT3 quantification (n = 7) of (C). Data were normalized to IL-6-treated G^low.

(F) Representative western blot to detect IFNγ-induced p-STAT3 in WT, G^low, G^low Il6st Δ𠄴24Δ–30, and G^low Il6ra Δ+1.7 cells.

(G) p-STAT3/STAT3 quantification (n = 6) of (F). Data were normalized to IFNγ-treated G^low.

Error bars (B) represent mean ± SEM. Box-and-whisker plots (D), (E), and (G) are presented with bounds from the 25^th to the 75^th percentiles, the median line, and whiskers ranging from minimum to maximum values. Statistical comparisons utilized unpaired Student’s t-test (B) or paired one-way ANOVA followed by Dunnett’s multiple comparisons test (D), (E), and (G). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; n.s., not significant.

IL6ST-dependent and IL-6- and IL-27-induced STAT3 phosphorylation was reduced by Il6st enhancer deletion (Figures 5C and 5D). Deleting the Il6ra enhancer reduced IL-6-, but not IL-27-, induced STAT3 phosphorylation (Figures 5C and 5E). IFNγ signaling, which does not require IL6ST, was constant in WT, GATA2^low, GATA2^low Δ–24Δ–30, and GATA2^low Δ+1.7 progenitors with or without IFNγ (Figures 5F and 5G). Limiting GATA2 levels commissions a PU.1-activated enhancer, elevating the expression of the shared cytokine receptor subunit IL6ST, and increasing IL-6 and IL-27 signaling. Under these conditions, an Il6ra enhancer is also commissioned, elevating the expression of IL6RA, the receptor subunit dedicated to IL-6 signaling, thereby increasing IL-6-, but not IL-27-, induced STAT3 phosphorylation.

DISCUSSION

As pro-inflammatory cytokines, chemokines, and lipid mediators potently orchestrate acute and chronic inflammation,⁶⁷ stringent mechanisms control their synthesis, secretion, and membrane receptor activation. Deleterious consequences of excessive inflammation include promoting MDS to AML progression^10,52,53 and hyperactivating immune responses, resulting in potentially fatal cytokine storms associated with cancers and immunotherapies. Canonical mechanisms integral to inflammation involve cytokine/chemokine-dependent cell signaling and transcription factor activation to alter genome function. Inflammation-activated transcription factors, e.g., NF-κB and AP1, form enhancer and/or promoter complexes that induce genes mediating inflammation.⁶⁸ Whereas the activating mechanisms are well studied, considerably less is known about transcriptional mechanisms constraining and resolving inflammation. Such mechanisms involve those inactivating signal-dependent transcription factors, suppressor of cytokine/chemokine signaling (SOCS) proteins that restrict inflammatory signaling,⁶⁹ and bioactive lipids (resolvins) that promote inflammation resolution.⁶⁷

Our multi-omics strategy unveiled GED at cytokine/chemokine receptor genes that remain dormant until derepressed. Distal and intronic enhancers at genes encoding inflammatory receptors are inactive until GATA2 declines, which enables binding of the activator PU.1, elevates chromatin accessibility, and activates cytokine/chemokine receptor genes. In cells with normal GATA2 levels, deleting the sequences had little to no impact on low-level Il6st expression. GED governs function of IL-6 and IL-27 signaling systems that control diverse hematopoietic and pathogenic processes. IL-6 and IL-27 receptors share the IL6ST subunit, activate MAPK and JAK/STAT pathways, exert pro-inflammatory functions, yet can also function distinctly.⁷⁰ IL-6 promotes granulopoiesis in the mouse⁷¹ and regulates macrophage colony-stimulating factor⁷² to promote macrophage differentiation,⁷³ consistent with the attenuated monocytic differentiation upon Il6st enhancer deletion (Figures S5E and S5F). As abrogating IL-6 signaling by ablating the Il6ra enhancer in GATA2^low myeloid progenitors upregulated Il27ra expression (Figure 5B), in certain contexts, the pathways may have opposing activities.

Elevated IL-6 and IL-27 can disrupt HSPC functions and are implicated in hematologic malignancies. Chronic inflammation involving IL-6 determines leukemic progenitor fate and induces chronic myelogenous leukemia development.⁷⁴ In mouse xenograft models, excessive IL-6 signaling promotes the MDS to AML transition⁵² and AML development.⁵³ Elevated IL-27 signaling during aging impairs HSC function.⁷⁵ As GATA2-PU.1 antagonism repressed Il6st enhancers to restrict IL-6 and IL-27 signaling, GED may have broader implications for immunopathology.

While GED antagonizes PU.1 function in fetal murine wild type and GATA2^low fetal progenitors, multiple lines of evidence suggest that the mechanisms can be extrapolated to adult contexts. In an adult AML mouse model harboring a compound Nras^G12D/+; p53^R172H/+ mutation, Gata2 expression decreased in adult bone marrow HSPCs, and Il6st and Il6ra expression increased,⁷⁶ consistent with GED. GATA2 occupied the Il6st enhancers in human bone marrow HSPC populations and adult erythroid progenitor (HUDEP2) cells (Figures 4A–4D), and the enhancers are evolutionarily conserved (Figure S3A). Furthermore, elevating GATA2 expression in human AML NOMO-1 cells with low GATA2 repressed IL6ST and IL6R (Figures 4G and 4H).

In summary, we provided evidence for GED in which GATA2 maintains enhancers in a dormant state, constraining expression of IL-6 receptor IL6ST and IL6RA subunits, and decreasing progenitor responsiveness to IL-6. GATA2 loss derepressed the enhancers, and PU.1 occupancy activated expression, increasing signaling. Additional genes encoding members of the cytokine/chemokine receptor family harbored GATA2-repressed enhancers. GATA2-repressed Il6st −24 and −30 enhancers confer IL-6 and IL-27 functions, whereas the Il6ra +1.7 kb intronic enhancer exclusively elevates IL-6 signaling (Figure 6). A plethora of inflammation-dependent mechanisms activate gene transcription to mediate innate immune responses. Constraining enhancer-dependent cytokine/chemokine receptor expression, e.g., via GED, may constitute a vital regulatory layer to counteract excessive inflammatory mediator production which would be deleterious to HSPCs and non-hematopoietic cells in the microenvironment. Regarding cells competent to express high levels of cytokines/chemokines, in principle, defective GED may cause deleterious immune cell hyperactivation and cytokine storm. It will be instructive to ask how GED is interlinked to and/or orchestrated with other mechanisms that constrain cytokine/chemokine signaling, as well as those mediating the resolution of inflammation.

Figure 6. Model of GATA2-induced enhancer decommissioning to constrain cytokine signaling.

GATA2 limits enhancer accessibility and precludes PU.1 chromatin occupancy to decommission conserved Il6st and Il6ra enhancers, which constrains IL-6 signaling. The Il6st enhancers regulate the expression of the shared cytokine receptor subunit IL6ST to mediate IL-6 and IL-27 signaling.

Limitations of the study

While our data support a model in which GATA2 decommissions inflammatory gene enhancers by opposing PU.1 occupancy, the precise molecular basis for the opposition remains to be elucidated. GED is characterized by relatively low GATA2 occupancy (Figures 1D and 1E), suggesting that GATA2 occupies these enhancers to inhibit PU.1; yet mechanisms underlying the low GATA2 signal detected by CUT&Tag are not established. PU.1 can function as a pioneer factor, binding nucleosomal DNA and recruiting the chromatin remodeling BRG1/BRM-associated factor (cBAF) complex.^77,78 PU.1 can also be chemically redirected from its native A-rich binding sites to GC-rich sites, increasing accessibility at ectopic loci.⁷⁹ In GATA2^low cells, reduced GATA2 was associated with increased PU.1 occupancy and enhancer accessibility at Il6st, Il6ra, Flt3, and Tnfrsf1a (Figures 1D, S9A, and S9B), supporting a GED mechanism by which GATA2 constrains PU.1-dependent chromatin remodeling to restrict enhancer activation.

Analogous to GATA1, GATA2-mediated repression might involve recruiting repressive complexes or interfering with PU.1 and its coactivators. While GATA1/FOG1 confers repression via the NuRD complex,³⁸ GATA2 corepressors remain unknown. GATA1 can inhibit PU.1-mediated activation by disrupting the PU.1-c-JUN interaction,^43,45 a function GATA2 might share given its C-finger sequence similarity. In principle, GATA2 might act through transient binding, consistent with a “hit-and-run” mechanism,⁸⁰ that recruits corepressors and precludes PU.1 access. In this type of mechanism, GATA2 might antagonize PU.1 transiently on chromatin, disrupting its pioneer activity and/or coactivator recruitment at PU.1-dependent loci. In addition, we cannot unequivocally exclude the possibility that GATA2 regulates a repressor of Il6st and other inflammatory genes. Though our findings introduce an unexpected mechanistic concept of GED, further studies in GATA2^lowPU.1^low cells are required to define the detailed molecular basis of GATA2-PU.1 antagonism at endogenous target genes.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Emery H. Bresnick (ehbresni@wisc.edu)

Materials availability

All unique/stable reagents generated in this study are available from the lead contact without restriction.

Data and code availability

The RNA-seq, ATAC-seq, and CUT&Tag datasets utilized during this study were previously published and are available at GEO: GSE199464 (RNA-seq), GSE201968 (ATAC-seq); and GEO: GSE279155 (RNA-seq), GSE279156 (CUT&Tag).⁵⁷ HUDEP2 GATA2 CUT&Tag data are available at GEO: GSE288701.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell culture

The generation of ER-HOXB8-immortalized (hi) hi-77^+/+ (WT), hi-77^−/− (GATA2^low), and hi-77^−/− URE^−/− (GATA2^lowPU.1^low) progenitor lines was described.^4,5 hi-progenitors were generated from mouse fetal liver Lin⁻ cells immortalized by retroviral expression of Estrogen-regulated HOXB8. Cells were cultured in Opti-MEM (31985-070; Gibco) with 10% Fetal Bovine Serum (FBS: 100–106; GeminiBio), 1% Penicillin-Streptomycin (SV30079.01; Cytiva), 1% Stem Cell Factor (SCF)-conditioned medium, 30 mM β-Mercaptoethanol (M3148; Sigma-Aldrich), 1 μM β-Estradiol, and 500 μg/mL G418 (400-111P-001; Gemini Bio). The AML cell lines NOMO-1, THP-1, and MOLM-13 (gift from Dr. Rebecca Richards, UW Madison) were maintained in RMPI 1640 (11875135; Gibco) supplemented with 10% FBS, 1% Penicillin-Streptomycin, and 1% L-Glutamine. Cells were cultured in a humidified 5% CO₂ incubator at 37°C.

To induce differentiation of hi-cells, cells were pelleted and washed with 1x PBS and resuspended in differentiation medium consisting of OPTI-MEM supplemented with 10% FBS, 1% Penicillin-Streptomycin, 1% SCF-conditioned medium, 1% IL3-conditioned medium, and 30 mM β-Mercaptoethanol. Cells were cultured for 3 d at 37°C and analyzed for monocytic and granulocytic populations. For flow cytometry, cells were pelleted and washed with 1x PBS, then stained at 4°C for 30 min using 1:200 diluted combinations of APC-CD11b (Invitrogen) or PE-Cy7-CD11b (eBioscience), PE-CD115 (BioLegend) in PBS with 10% FBS. After staining, cells were washed once with PBS containing 2% FBS and analyzed on an Attune NxT Flow Cytometer (Thermo Fisher Scientific). Differentiated cell populations were analyzed using FlowJo v10.8.0 software (BD Biosciences).

METHOD DETAILS

Mouse vs. human conservation analysis

Sequences under ATAC-seq peaks accessible in GATA2^low cells were assessed for evolutionary conservation⁸¹ and converted from mm10 mouse genome to hg38 human genome through the UCSC genome browser. The obtained mouse and human sequence were aligned using NCBI blastn and assessed for sequence conservation. Sequences with alignment scores over 70 were deemed as highly conserved. Alignment scores were determined by NCBI BLAST nucleotide scoring system with a positive score for matches, negative score for mismatches, and penalties for open and extended gaps (https://www.nlm.nih.gov/ncbi/workshops/2023-08_BLAST_evol/blast_score.html). Sequence alignments with extra 50 bp upstream and downstream are presented as box-shade plots (Figures S3 and S7).

Gene editing with CRISPR-Cas9

Guide sequences targeting distal and intronic enhancers of murine Il6st, Il6ra, Flt3, and Tnfrsf1a were designed using online tools (http://www.idtdna.com/). Chemically modified crRNAs and tracrRNA were purchased from IDT, and Cas9 protein was from Aldevron. 100 pmol each of crRNA was mixed with 100 pmol of tracrRNA and incubated at 95°C for 5 min. Cas9 protein (8 mg) was added and incubated at 37°C for 15 min to form RNP complexes. GATA2^low cells were resuspended in 20 μL P3 buffer with 22% supplement (Lonza) and added to the RNP complex. Electroporation was conducted using EO-100 program on Nucleofector 4D (Lonza). 72 h after nucleofection, genomic DNA was extracted from cell pools, and genomic DNA flanking the target site was amplified by PCR with GoTaq G2 DNA Polymerase (Promega) to estimate editing efficiency. Serial dilutions were conducted in 0.5 cell (150 μL)/well concentration in 96 wells for clone isolation. Clonal lines were screened by PCR using primers flanking the targeted region. Amplicons from clones harboring homozygous deletions were purified using QIAGEN PCR purification kit (QIAGEN) and mixed with 2 pmol primer for Sanger sequencing (Functional Biosciences). Clones were sequenced separately with forward and reverse primers.

Sequencing mutant clones

Homozygous clones and the estimated size of deletion were determined by PCR. Amplified PCR products from primers flanking the deletion were purified using QIAGEN PCR purification kit (QIAGEN) and mixed with 2 pmol primer for Sanger sequencing (Functional Biosciences). Clones were sequenced separately with forward and reverse primers. Endpoints common to both alleles were predicted to have the same deletion. Mixed sequencing results with no common endpoint were predicted to have different deletions. For different deletions, the endpoints of mixed sequences were determined to predict the maximum deletion size. All clones for Il6st Δ-24, Δ-30, Δ-24Δ-30 and Il6ra Δ+1.7 in GATA2^low cells were sequenced. Four to six representative clones were sequenced for Il6st Δ-24Δ-30 in WT cells and Flt3 Δ+7.4 and Tnfrsf1a Δ+11 in GATA2^low cells.

PU.1 rescue assay

PU.1 was expressed in hi-77^−/− URE^−/− (GATA2^lowPU.1^low) progenitors with retrovirus harboring murine Spi1 cDNA in the murine stem cell virus plasmid with GFP (MSCV-PIG).²⁹ Ecotropic virus was packaged in 293T cells, and retrovirus-containing supernatants were collected 48 h post-transfection. Cells were transferred to IMDM containing 2% FBS and incubated with supernatant by spinoculation for 90 min at 2,800 rpm at 30°C. Cells were cultured for 3 d in media described above. GFP⁺ cells were sorted by FACS with FACSAria cell sorter (BD Biosciences) and subjected to protein and RNA analysis. RNA was isolated with TRIzol (Invitrogen).

Gene expression analysis

Total RNA was purified from 1-5 × 10⁵ cells with TRIzol (Invitrogen) and 2 μg RNA was treated with DNase I (Thermo Fisher Scientific) for 15 min at room temperature. After heat inactivation of DNase I with EDTA for 10 min at 65°C, RNA was incubated with 250 ng of a 4:1 mixture of Oligo dT primers and Random Hexamer at 68°C for 10 min. RNA/primers were incubated with Moloney Murine Leukemia Virus Reverse Transcriptase (M-MLV RT) (Thermo Fisher Scientific), 5× First Strand Buffer (Thermo Fisher Scientific), 10 mM Dithiothreitol (Thermo Fisher Scientific), RNAsin (Promega), and 0.5 mM Deoxynucleoside Triphosphates (New England Biolabs) at 42°C for 1 h and then heat inactivated at 95°C for 5 min. Quantitative gene expression analysis was performed by real-time RT-qPCR using Power SYBR Green Master Mix (Applied Biosystems) and analyzed on a ViiA 7 Real-Time PCR System (Applied Biosystems). Control reactions without M-MLV RT yielded little to no signal. Relative expression of mRNA was determined from a standard curve of serial dilutions of cDNA samples, and values were normalized to 18S rRNA expression.

CUT&Tag

CUT&Tag was conducted as described⁵ with HUDEP-2 cells and anti-GATA2 rabbit polyclonal antibody.⁸² Concanavalin A magnetic beads (BP531; Bangs Laboratories) were activated by washing twice in Binding Buffer (20 mM HEPES pH 7.9, 10 mM KCl, 1 mM CaCl₂, 1 mM MnCl₂, and 0.1% BSA). Cells were gently crosslinked with 0.1% Formaldehyde for 2 min and quenched with 75 mM Glycine. Cells were washed twice in 1.5 mL of Wash Buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 0.83 mM Spermidine, 0.1% BSA, and one Roche Complete Protease Inhibitor EDTA-free tablet) and resuspended in 1 mL Wash Buffer. For one antibody, 150,000 cells in 1 mL final volume of Wash Buffer were mixed with 40 μL activated bead suspension and rotated for 30 min at room temperature. On a magnetic rack, wash buffer was replaced with 150 μL of Antibody Buffer (Wash Buffer plus 0.1 μg/mL Digitonin and 2 mM EDTA). 1.2 μg affinity-purified rabbit polyclonal anti-GATA2 antibody was added. Tubes were rotated overnight at 4°C. After a quick spin, tubes were placed on a magnetic stand, and Antibody Buffer was replaced with 150 μL Dig-Wash Buffer (Wash Buffer plus 0.1 μg/mL Digitonin) containing secondary antibody (1:100 dilution) for rabbit antibody. After an hour on a nutator at room temperature, cells were washed three times with Dig-Wash Buffer. Resuspended cells in 150 μL Dig-300 Buffer (20 mM HEPES pH 7.5, 300 mM NaCl, 0.83 mM Spermidine, 0.1 μg/mL Digitonin, 0.1% BSA, and one Roche Complete Protease Inhibitor EDTA-free tablet) containing pA-Tn5 pre-loaded with MEDS adapters (1:250 dilution). After nutating at room temperature for 1 h, cells were washed three times with Dig-300 Buffer. Added 300 μL Tagmentation Buffer (Dig-300 Wash plus 10 mM MgCl₂) and nutated tubes at 37°C for 1 h. Cells were isolated by centrifugation, and 10 μL 0.5 M EDTA was added to stop tagmentation, followed by addition of 3 μL 10% SDS and 3 μL Proteinase K (20 mg/mL, P8107S; NEB). Cells were incubated at 37°C overnight on a thermomixer at 600 rpm. DNA was purified by Phenol-Chloroform extraction and Ethanol precipitation. DNA (1.31 μL) was analyzed by RT-qPCR in a 25 μL reaction containing 10 μM i5 and i7 primers and PowerUp SYBR Green Master Mix. Libraries were prepared with the NEB Next HiFi 2x PCR Master Mix (M0514S) following the manufacturer’s protocol. The number of PCR cycles was determined based on RT-qPCR. Sample concentration was determined using Qubit 1× dsDNA HS Assay (Q33231; ThermoFisher). The quality of the library was assessed using Bioanalyzer High Sensitivity DNA Analysis (Agilent). Samples were pooled and sequenced on NovaSeq 6000 using 50 bp paired-end sequencing at about 10 M reads per sample. CUT&Tag FASTQ files were processed by removing the adapter ‘CTGTCTCTTATACACATCT’ using the Cutadapt software (version 4.5) and aligned to the mouse genome (version mm10) by Bowtie2 (version 2.5.1). Duplicated fragments were marked and removed by Picard (version 3.1.1). The CUT&Tag raw data were deposited to GEO with the accession number GSE288701.

Protein analysis for cytokine signaling

To induce IL-6/STAT3, IL-27/STAT3, or IFNγ/STAT3 signaling, WT, hi-77^−/−, and hi-77^−/− Il6st or Il6ra enhancer mutant progenitors were serum-starved for 2 h and treated with 20 ng/mL mouse IL-6 (PeproTech), 100 ng/mL mouse IL-27 (PeproTech), or 1 ng/mL mouse IFNγ for 30 min. Samples were lysed and boiled for 10 min in SDS Lysis Buffer. Proteins were resolved on 8% SDS-Polyacrylamide gels. Proteins were analyzed by semi-quantitative Western blotting with ECL Plus (Thermo Fisher Scientific) and West Femto (Thermo Fisher Scientific). The antibodies used were rabbit monoclonal anti-pSTAT3 (9145), rabbit monoclonal STAT3 (12640), and mouse monoclonal anti-β-Actin (3700) from Cell Signaling Technology. Blots were developed by LI-COR Odyssey Imaging System (LI-COR Biosciences) and quantified by Image Studio Lite (version 5.2.5) (LI-COR Biosciences).

Exogenous GATA2 expression in NOMO-1

Human GATA2 was expressed in NOMO-1 cells by infection with GATA2-expressing lentivirus. The lentivirus-containing supernatant was harvested after 48 and 72 h post-transfection, snap-frozen, and stored at −80°C. The GATA2 lentivirus was spinoculated with NOMO-1 cells for 90 min at 2,800 rpm at 30°C. Cells were expanded in RPMI 1640 medium for 72 h. GFP^high cells were sorted using BD FACSAria Cell Sorter with 130 μm nozzle and a pressure of 14 psi. Cells were harvested for RNA isolation and RT-qPCR analysis.

QUANTIFICATION AND STATISTICAL ANALYSIS

Unless otherwise stated, RT-qPCR analysis results were presented as the mean ± SEM, and Western Blot analysis results were presented as box-and-whisker plots with bounds from the 25^th to the 75^th percentiles, the median line, and whiskers ranging from minimum to maximum values. Statistical comparisons and the number of biological replicates for each experiment were presented in the figure legends. Outliers were identified using the ROUT test. All statistics had a significance cutoff of P less than 0.05 and were calculated using GraphPad Prism 10 (GraphPad Software).

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.116344.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-pSTAT3 rabbit monoclonal antibody (Y705)	Cell Sig Tech	9145; RRID: AB_2491009
Anti-STAT3 rabbit monoclonal antibody	Cell Sig Tech	12640; RRID: AB_2629499
Anti-PU.1	Cell Sig Tech	2258; RRID: AB_2186909
Anti-β-Actin mouse monoclonal antibody	Cell Sig Tech	3700; RRID: AB_2242334
Anti-α-Tubulin mouse monoclonal antibody	Cell Sig Tech	2144; RRID: AB_2210548
Anti-GATA2 rabbit polyclonal antibody	Jung et al.5	N/A
Anti-HA rabbit monoclonal antibody	Cell Sig Tech	3724; RRID: AB_1549585
PE-CD115	BioLegend	135506; RRID: AB_1937253
APC-CD11b	BioLegend	101212; RRID: AB_312795
PE-Cy7-CD11b	eBioscience	25-0112-82; RRID: AB_469588
Chemicals, peptides, and recombinant proteins
Recombinant mouse IL-6	PeproTech	216–16
Recombinant mouse IL-27	R&D Systems	2799
Recombinant mouse IFNγ	PeproTech	315–05
Critical commercial assays
P3 Primary Cell 4D Nucleofector X Kit S	Lonza	V4XP-3032
SYBR Green for RT-qPCR	Applied Bioscience	AB1163A
Deposited data
hi-77−/− RNA-seq	Jung et al.5	GEO: GSE199464
hi-77+/+ (empty), hi-77−/− (empty), and hi-77−/− (GATA2) ATAC-seq	Jung et al.5	GEO: GSE201968
hi-77−/− and hi-77−/− URE−/− RNA-seq	Jung et al.5	GEO: GSE279155
hi-77−/− and hi-77−/− URE−/− CUT&Tag (GATA2, PU.1)	Tran et al.57	GEO: GSE279156
HUDEP2 GATA2+/− CUT&Tag (GATA2, HA-tag)	This paper	GEO:GSE288701
Experimental models: Cell lines
hi-77+/+ (WT)	Johnson et al.4	N/A
hi-77−/− (GATA2low or Glow)	Johnson et al.4	N/A
hi-77−/− URE−/− (GATA2lowPU.1low or GPlow)	Jung et al.5	N/A
NOMO-1	N/A	N/A
MOLM-13	N/A	N/A
THP-1	N/A	N/A
HUDEP2	N/A	N/A
Oligonucleotides
Il6st −24 crRNA: GGCAAAGAGACTTACTAAGTGGG GGTTTTCTAAGTGGATAATCTGG	This paper	N/A
Il6st −24 PCR: TGGCCGAACCCAAACTTCTA ACTGTGTACTGTGTAACTTTCTTTG	This paper	N/A
Il6st −30 crRNA: ACATCAACCTAAGACTGGTGTGG TTAACGTCTTGTGAGTCCAGAGG	This paper	N/A
Il6st −30 PCR: TCACCCTCAACAGGATGGG TCTGAGGCCTTATGCCTGG	This paper	N/A
Il6ra +1.7 crRNA: ACCGAAAGAGGTTGTCCCAGAGG ACACGTTTTTAATCCTCTACAGG	This paper	N/A
Il6ra +1.7 PCR: TTTAATGAGTGGCCAGGGGT TGCACTCTGCCTTGTAGGTC	This paper	N/A
Flt3 +7.4 crRNA: GAATCCGCTATAGTCTACACTGG TCTCACCAACAAGATCGGCCTGG	This paper	N/A
Flt3 +7.4 PCR: GAAGACAGAATTTAGTGTGTACCTTA ACGTACTGCATAGCTTAGTCTAC	This paper	N/A
Tnfrsf1a +11 crRNA: CCGTAATTCTCTTGTCCCTGGG TCCCTTTCAACGCTGACCTTGGG	This paper	N/A
Tnfrsf1a +11 PCR: GTTATCTTGCTAGGTCTTTGCCTTC TCCCTTCTCACCTCTTTGACAG	This paper	N/A
Il6st RT-qPCR: GACCAAAGATGCCTCAACTTG CCCTTCCCACTGTCCTTAATG	This paper	N/A
Il6ra RT-qPCR: CAAGAATCCTCGTCCATGTCC TCGTCTTGCTTTCCTTCTCAG	This paper	N/A
Il6ra (primary) RT-qPCR: GGGACTGAAGAACTGTGAGTG AGGTCCTGTTGCTGTCTTAGA	This paper	N/A
Irf8 RT-qPCR: GACCAAGAGGAGCCCATCC CCATCTGGGAGAAAGCTGAATG	This paper	N/A
Cd79b RT-qPCR: CCCATCTTCCTGCTACTTGAC TCCTACCGACCACTTTACCTC	This paper	N/A
Rag2 RT-qPCR: GAGTATTTCACATCCACAAGCAG TCATAAGTGAGAAGCCTGGTTG	This paper	N/A
Tifab RT-qPCR: GTCCTGTCTGCTCATCTGTTC GGGCCTCTCCATAGATCCAG	This paper	N/A
Flt3 RT-qPCR: ATCATAAGAACAAGCCAGGAGAG CATCAGAGGAACAGGACGC	This paper	N/A
Tnfrsf1a RT-qPCR: GCAGGGAGTGTGAAAAGGG GGTTCTCCTTACAGCCACAC	This paper	N/A
Csf1r RT-qPCR: TGGCCACAGTTTGGCATG TCCACACTGCCATTGCTCA	This paper	N/A
Ccr2 RT-qPCR: GCTCTACATTCACTCCTTCCAC ACCACTGTCTTTGAGGCTTG	This paper	N/A
Cxcr2 RT-qPCR: CCCTGCCCATCTTAATTCTACG AGGAAGCCAAAAGTCTGAGG	This paper	N/A
Il27ra RT-qPCR: GGATCAGGTGGAAAGCTCTG ACTTGTGTCGCCAGTGTAAG	This paper	N/A
Human IL6ST RT-qPCR: GCAACATTCTTACATTCGGACAG TCCCACTCACACCTCATTTTC	This paper	N/A
Human IL6R RT-qPCR: TCAGGGTTGTGGAATCTTGC GAGCTCAAACCGTAGTCTGTAG	This paper	N/A
Human IL1RL1 RT-qPCR: CCTACCATTGACCTCTACAACTG GTAATCACCTGCGTCCTCAG	This paper	N/A
Hck RT-qPCR: GTCTGGTCCTTTGGTATCCTG GTTATCTGGTCGAGGCATACG	This paper	N/A
Slc11a1 RT-qPCR: GGCCTCAACGATCTACTCAATG GCTGGTGAAAGTCAAAATGGG	This paper	N/A
Recombinant DNA
Mouse GATA2-HA retrovirus	Katsumura et al.29	N/A
Mouse PU.1-HA retrovirus	This paper	N/A
Human GATA2-HA lentivirus	This paper	N/A
Human T354M-HA lentivirus	This paper	N/A
Software and algorithms
Graphpad Prism 10	GraphPad	https://www.graphpad.com/
Image Studio Lite 5.2.5	LiCor Biosciences	https://www.licor.com/bio/image-studio-lite/
FlowJo 10	BD Biosciences	https://www.bdbiosciences.com/en-us/products/software/flowjo-v10-software
QuantStudio Software 1.6.1	Applied Biosystems	https://www.thermofisher.com/us/en/home/technical-resources/software-downloads/applied-biosystems-viia-7-real-time-pcr-system.html

Highlights.

• GATA2 decommissions conserved enhancers to constrain cytokine signaling

• GATA2 represses enhancer accessibility and precludes PU.1 chromatin occupancy

• Enhancer decommissioning and PU.1 antagonism control multiple inflammatory signaling systems