Transcription Factor GATA1 Is Dispensable for Mast Cell Differentiation in Adult Mice

Abstract

Although previous studies have shown that GATA1 is required for mast cell differentiation, the effects of the complete ablation of GATA1 in mast cells have not been examined. Using conditional Gata1 knockout mice (Gata1^−/y), we demonstrate here that the complete ablation of GATA1 has a minimal effect on the number and distribution of peripheral tissue mast cells in adult mice. The Gata1^−/y bone marrow cells were capable of differentiating into mast cells ex vivo. Microarray analyses showed that the repression of GATA1 in bone marrow mast cells (BMMCs) has a small impact on the mast cell-specific gene expression in most cases. Interestingly, however, the expression levels of mast cell tryptases in the mouse chromosome 17A3.3 were uniformly reduced in the GATA1 knockdown cells, and GATA1 was found to bind to a 500-bp region at the 5′ end of this locus. Revealing a sharp contrast to that observed in the Gata1-null BMMCs, GATA2 deficiency resulted in a significant loss of the c-Kit⁺ FcεRIα⁺ mast cell fraction and a reduced expression of several mast cell-specific genes. Collectively, GATA2 plays a more important role than GATA1 in the regulation of most mast cell-specific genes, while GATA1 might play specific roles in mast cell functions.

INTRODUCTION

GATA1 is a key transcription factor involved in the differentiation of several hematopoietic cell lineages, including erythroid cells, megakaryocytes, eosinophils, and dendritic cells (1, 2). Gata1-deficient mice die in utero around embryonic day 11.5 (E11.5) due to defective primitive erythropoiesis in the yolk sac (3). Similarly, Gata1 knockdown mice, referred to as Gata1^G1.05/Y mice, generated by targeted inactivation of the hematopoietic-specific exon 1 (IE) promoter, die by E12.5 (4). Cre-loxP-mediated Gata1 gene deletion in adult mice results in erythroid aplasia and defective stress erythropoiesis, indicating that GATA1 is also necessary for postnatal erythropoiesis (5). The requirements of GATA1 for megakaryocyte differentiation and specification of the eosinophil lineage were demonstrated through the use of megakaryocyte- and eosinophil-specific GATA1 knockdown mice, respectively (6, 7).

In addition to these cell lineages, GATA1 is also expressed in mast cells. These cells have a central role in the innate immune system and allergic diseases (8). Several studies shown that GATA1 is not essential for the specification of the mast cell lineage but is critical for the later stage of mast cell development (9–14). In addition, several mast cell-specific genes, such as FcerIa and FcerIb, encoding the α and β subunits of the IgE receptor (FcεRIα, FcεRIβ), and Cpa3, encoding mast cell carboxypeptidase, have been shown to be regulated by GATA1 (13–15). Of note, these studies were conducted using Gata1 knockdown mice or cultured mast cell lines. The consequences of complete ablation of GATA1 on mast cell differentiation have never been examined. We previously noted that GATA2, another GATA family member, is abundantly expressed in mast cells, implying a functional redundancy between GATA1 and GATA2 (16).

GATA2 is essential for mast cell lineage specification in the in vitro differentiation of embryonic stem cells (17). We recently revealed that the GATA2 mRNA level was significantly increased, while GATA1 mRNA expression was maintained at low levels during the differentiation of mast cells derived from mouse bone marrow (BMMCs) (16). Furthermore, in a coculture system with Swiss 3T3 fibroblasts, Takano et al. reported that the expression level of GATA1 further declines to an undetectable level when BMMCs mature into connective tissue-type mast cells (18). Collectively, these data prompted us to reassess whether GATA1 plays an essential role in BMMCs.

In contrast to BMMC differentiation, the GATA2 expression in multilineage progenitors declines upon commitment to the erythroid lineage and is switched for GATA1 expression, which peaks at the late erythroid progenitor and proerythroblast stages. This dynamic transition of GATA factor expression is essential for correct erythroid differentiation and has been referred to as “GATA factor switching” (19, 20), which is mediated by two key cis-regulatory mechanisms through highly conserved GATA binding sites in the Gata1 and Gata2 loci. One is a direct repression of Gata2 gene expression by GATA1 through the conserved GATA boxes within the Gata2 locus (21, 22). The other is a positive autoregulation of Gata1 through several conserved GATA boxes, including the Gata1 gene hematopoietic enhancer (G1HE, also referred to as HS1 or mHS-3.5) located 3.9 kb upstream of IE (23, 24). Importantly, we showed that neither forced expression nor small interfering RNA (siRNA)-mediated knockdown of GATA1 affected the Gata2 gene expression in BMMCs, indicating that the GATA1-mediated Gata2 repression does not take place in mast cells (16). Furthermore, we found that the G1HE region is epigenetically inactivated and is dispensable for Gata1 gene expression in BMMCs and peritoneal mast cells by performing transgenic reporter mouse assays (16). Taking these findings into account, we surmised that, unlike erythroid differentiation, GATA2 might play a predominant role over GATA1 in mast cell differentiation.

In the present study, we wanted to define the specific roles of GATA1 in mast cell development. To this end, we examined the effects of complete ablation of GATA1 in mast cell differentiation using tamoxifen-inducible Gata1 knockout mice (Gata1^−/y). We demonstrate here that the role of GATA1 in mast cell differentiation is more limited than previously anticipated and that the null mutation of Gata1 is likely compensated for by GATA2.

MATERIALS AND METHODS

Mice.

Conditional Gata1 knockout mice (Gata1^flox/y) were generated as described previously (5). The knockin mice expressing a 4-hydroxy tamoxifen (4-OHT)-inducible Cre recombinase gene under the control of the Rosa26 promoter (Rosa26CreER^T2) were kindly provided by Anton Berns, Netherlands Cancer Institute. Since the Gata1 gene is X linked, the Gata1 knockout phenotype was examined in hemizygous male mice (Gata1^flox/y) expressing CreER^T2. These mice had been bred on a BDF1 background and were subsequently backcrossed to a C57BL/6 background. Cre-mediated recombination of the Gata1 gene was determined by genomic PCR, as described previously (16). Gata2^flox mice (25) were kindly provided by S. A. Camper, University of Michigan. C57BL/6-Kit^W-sh/W-sh mice were purchased from RIKEN BRC. Mice were maintained in the animal facility of Takasaki University of Health and Welfare in accordance with institutional guidelines.

Induction of the Cre transgenes in vivo.

To induce Cre recombinase, mice (8 to 10 weeks of age) were injected subcutaneously with tamoxifen (0.1 mg/g [body weight]; Sigma) dissolved in sunflower oil on experimental days 1 to 5 and 8 to 12. The body weight and hematocrit level were monitored weekly. The mice were euthanized and used for the analysis on experimental days 28 to 35.

Hematological analyses.

Blood samples were taken from the tail vein using heparin-coated microtubes. The hematocrit values were measured using a micro-hematocrit centrifuge (MC-150; Tomy Seiko).

qRT-PCR.

Total RNA was extracted from cells using NucleoSpin RNA II (TaKaRa). Reverse transcription (RT) reactions were performed using a ReverTra Ace qPCR RT kit (Toyobo) according to the manufacturer's instructions. Quantitative RT-PCR (qRT-PCR) was performed using the Go Taq qPCR master mix (Promega) and an Mx3000P real-time PCR system (Stratagene), as described previously (26). The data were normalized to the 18S rRNA or GAPDH (glyceraldehyde-3-phosphate dehydrogenase) levels and are shown as the averages ± the standard deviations (SD). The primer sequences used for PCR are shown in Table 1, unless previously described (16).

TABLE 1.

Primer sequences for qRT-PCR and qChIP analyses

Analysis and gene	Locus	Orientation	Sequence
qRT-PCR
Il1a		5′	TCGGGAGGAGACGACTCTAA
		3′	GTGCACCCGACTTTGTTCTT
Epor		5′	GACCCCAAGTTTGAGAGCAA
		3′	TGCAGGCTACATGACTTTCG
Ahr		5′	ATGGCTTTGTGCTGGGTTGTCACAG
		3′	ACTCCTTGTGCAGAGTCTGGGTTT
Il18		5′	ACGTGTTCCAGGACACAACA
		3′	ACAAACCCTCCCCACCTAAC
Tnfrsf11b		5′	ATGAACAAGTGGCTGTGCTG
		3′	CCTCACTGTGCAGTGCTGTT
Creg2		5′	GTGTCCACCCACGAAAAGAT
		3′	GTCTTCTGGGTCGACGATGT
Fceria		5′	AGAGCAAACCTGTGTACTTG
		3′	GTGGACTTTCCATTTCTTCC
Fcerib		5′	AGGCTACCCATTCTGGGGTG
		3′	GGCTGCCTCTCACCAGATAC
Fcerig		5′	CTCCTTTTGGTGGAAGAAGC
		3′	TGAGTCGACAGTAGAGCAGG
Mcpt1		5′	TTCCCTTGCCTGGTCCCT
		3′	GTTTTCCCCCAGCCAGCT
Mcpt4		5′	GAAGTGAAAAGCCTGACCTGC
		3′	CATGCTTTGTTGAACCCAAGG
Mcp9		5′	GGGTGGCCCATGGTATTGTA
		3′	CGGGTGAAGATTGCAGGG
Tpsg1		5′	GGTCACACTGTCTCCCCACT
		3′	GCATCCCAGGGTAGAAGTCA
Tpsb2		5′	CGACATTGATAATGACGAGCCTC
		3′	ACAGGCTGTTTTCCACAATGG
Tpsab1		5′	ATGACCACCTGATGACTGTGAGCCAG
		3′	AGGAACGGAGGTCATCCTGGATGTG
qChIP
Chr17A3.3a	Region A a	5′	AACCTTCGACGTGACCTTTG
		3′	GGCACAGGATTTGTGAGACC
	Region A b	5′	TGGCTACATGTTCCCTACCC
		3′	CAAAGGTCACGTCGAAGGTT
Kit	−114Ka	5′	AGCAATGGCCTCACGAGTTCTA
		3′	CCAGGAAAAGTTTGGCAGGAT
	−114Kb	5′	CGTGCACACAGGTTTGTTTC
		3′	TGCTGAGATGTGGCAATAGG
	Promoter	5′	CACCTCCACCATAAGCCGAAT
		3′	CTCCTAGACAATAAAGGACAACCA
Fcerib	Promoter a	5′	ACAGCAAGAGAAAGGAGTCACTGAT
		3′	CATGCGGAACCTACTTGTCAGA
	Promoter b	5′	ATGAAGCTGCCAAAGAGCAC
		3′	CTGCCAGAGATGTGGGATTT
	−2.8K	5′	GACAAGTAGGTTCCGCATGAA
		3′	TGGGCTATCCAGGAATGAAA

^aThat is, mouse chromosome 17A3.3.

Western blotting.

Cells were lysed by boiling in Laemmli buffer. The lysates were resolved by sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis, and the Western blot analyses were performed as described previously (26) using anti-GATA1 (N6; Santa Cruz), anti-GATA2 (H-116; Santa Cruz), anti-β-tubulin (Sigma), and anti-lamin B (M-20; Santa Cruz) antibodies.

Histological and cytological analyses.

Tissue specimens were fixed in 4% paraformaldehyde, followed by embedding in Tissue-Tek OCT compound (Sakura Finetechnical) and freezing in liquid nitrogen. Cryostat sections were prepared by using a Leica CM1515 cryostat (Leica). Cytospin preparations of the cells (5 × 10⁴ cells) were made using a Shandon Cytospin4 centrifuge (Thermo Electron Corp.). Tissue sections and cytospin preparations were stained with either Wright-Giemsa stain, toluidine blue, or Alcian blue-safranin O.

Cell culture.

Bone marrow mononuclear cells isolated from the femurs and tibiae were cultured for 2 weeks in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% fetal bovine serum and penicillin-streptomycin in the presence of 10 ng of recombinant murine interleukin-3 (IL-3; Peprotech)/ml and subsequently cultured for 2 weeks with IL-3 and 10 ng of recombinant murine stem cell factor (SCF; Peprotech)/ml. The cultures were replenished with fresh medium every 4 days and maintained at a cell density of 5 × 10⁵/ml.

FACS analysis.

For fluorescence-activated cell sorting (FACS) analysis, cells were stained with fluorescence-conjugated antibodies for 30 min and analyzed using a FACSCantoII flow cytometer (BD Biosciences). The following antibodies were used for the analyses: allophycocyanin-conjugated rat anti-mouse CD117 (c-Kit), phycoerythrin (PE)-conjugated mouse anti-mouse FcεRIα (eBioscience), fluorescein isothiocyanate-conjugated rat anti-mouse CD11b (MacI), and PE-conjugated mouse anti-mouse Ly6C/G (Gr1). The anti-mouse FcεRIα antibody was purchased from eBioscience, and the other antibodies were purchased from BD Pharmingen.

Degranulation assay.

The degranulation of mast cells was measured by the release of β-hexosaminidase, as described previously (26). Briefly, BMMCs were sensitized with 0.5 μg of monoclonal anti-dinitrophenyl (DNP) murine IgE antibody (clone SPE-7; Sigma)/ml overnight and stimulated with 200 ng of DNP-BSA (Calbiochem)/ml at 37°C for 30 min. After stimulation, supernatants were collected, and the cells were lysed with 0.5% Triton X-100 in Tyrode buffer. The supernatants and cell lysates were incubated with 2 mM N-acetyl-β-d-glucosamide (Sigma) in 0.1 M citrate buffer, and the absorbance at 405 nm was measured using a Sunrise Rainbow RC microplate reader (Tekan).

Reconstitution of mast cells in mast cell-deficient mice.

BMMCs (2.0 × 10⁶ cells) in 200 ml of phosphate-buffered saline (PBS) were injected intraperitoneally into 6-week-old male Kit^W-sh/W-sh mice, and the mice were used for the experiments 6 to 8 weeks after the injection.

Transfection of siRNA.

The siRNA duplexes for mouse Gata1 and Gata2 were purchased from Invitrogen. The control siRNA was purchased from Sigma. BMMCs (2.0 × 10⁶ cells) were transfected with 200 pmol of siRNA by electroporation using an Amaxa Nucleofector (Lonza). The cells were harvested 24 h after transfection and used for the analyses.

Methylcellulose colony formation assay.

Methylcellulose colony formation assays were performed with MethoCult M3434 (Stem Cell Corporations) according to the manufacturer's instructions. Bone marrow cells were suspended in culture medium containing methylcellulose and plated onto 35-mm culture dishes at a density of 2 × 10⁴ cells/dish. Colonies were counted after 8 days of incubation at 37°C.

Microarray analysis.

Wild-type (WT) BMMCs were transfected with control or Gata1 siRNA (200 pmol), and total RNA was isolated 24 h after transfection, as described above. The quality of the RNA samples was checked using an Agilent 2100 bioanalyzer platform (Agilent Technologies). For each siRNA treatment, two replicate samples from different animals were pooled and used for the analysis. Cy3 labeling of cRNA, hybridization, scanning, and data analysis were done by Miltenyi Biotec using Agilent whole-mouse-genome Oligo Microarrays 8x60K. The expression data were analyzed using the Gene Spring software program (Agilent Technologies). Heat maps were generated using the Cluster 3.0 (http://bonsai.hgc.jp/∼mdehoon/software/cluster/) and JAVA TreeView (http://jtreeview.sourceforge.net/) software programs. The classification of the selected genes according to their functions was done using the Ingenuity Pathway Analysis (IPA) software program (Ingenuity System). The P values, represented as the negative log ratio of the IPA results, were the probability based on the Fisher exact test.

ChIP.

A chromatin immunoprecipitation (ChIP) assay was performed using anti-GATA-1 (N6; Santa Cruz) and anti-GATA2 (H-116; Santa Cruz) antibodies as in a prior report (26). As described previously, the GATA1 ChIP assay was performed using an anti-rat IgG rabbit antibody (Jackson ImmunoResearch) as a secondary antibody to precipitate the immune complex (26). The GATA2 ChIP assay was performed without the use of a secondary antibody. The quantitative analyses of DNA purified from the ChIP samples were conducted as previously described (16). The primer sequences used for PCR are shown in Table 1, unless previously described (16).

Statistical analysis.

Comparisons between two groups were made using Student t test. The data are presented as the means ± the SD. For all of the analyses, statistical significance was defined as a P < 0.05.

Accession numbers for the microarray data.

The microarray data were deposited in the Gene Expression Omnibus under GEO accession number GSE52254.

RESULTS

The systemic administration of tamoxifen resulted in the complete ablation of GATA1 expression in the hematopoietic tissues of adult Gata1^−/y mice.

To examine the roles of GATA1 in mast cell differentiation in adult mice, we used conditional Gata1 knockout mice expressing tamoxifen-inducible Cre recombinase (Gata1^fl/y::ROSA26CreER^T2). We previously examined the roles of GATA1 in adult erythropoiesis by using a similar compound mutant mouse model (Gata1^fl/y::ROSA26CreER [5]). Consistent with our previous data (5), systemic administration of tamoxifen led to progressive anemia by 28 days after the first tamoxifen treatment (Fig. 1A). The PCR analyses using genomic DNA revealed that recombination of the Gata1 gene in Gata1^fl/y::ROSA26CreER^T2 mice occurred efficiently in the bone marrow (BM) and spleen by day 28 (Fig. 1B). Thus, we here refer to the Gata1^fl/y::ROSA26CreER^T2 mice treated with tamoxifen as Gata1^−/y mice. Mice expressing Cre recombinase with the normal Gata1 allele (Cre) also showed anemia to various extents, probably due to “Cre toxicity,” as previously reported (27, 28) (Fig. 1A). However, unlike the Gata1^−/y mice, all Cre mice showed a recovery from anemia by day 28 and had normal hematocrit levels thereafter. The Gata1 mRNA level was significantly decreased in the BM and spleen of Gata1^−/y mice, as shown by both the conventional and qRT-PCR analyses (Fig. 1C). Consistently, the GATA1 protein was undetectable in the BM and spleen of Gata1^−/y mice by a Western blot analysis (Fig. 1D). These results demonstrate that, upon tamoxifen treatment, GATA1 was almost completely ablated in the hematopoietic tissues of Gata1^−/y mice.

FIG 1.

The GATA1 expression in hematopoietic tissues was completely ablated by the administration of tamoxifen to Gata1^−/y mice. (A) Average change in the hematocrit levels of WT, Cre, and Gata1^−/y mice treated with tamoxifen (n = 4 for each group). (B) Results of the PCR analysis of recombination in the genomic DNA isolated from the BM and spleens of WT, Cre, and Gata1^−/y mice. Genomic DNA was isolated 28 days after the start of the tamoxifen treatment and was used for the PCR analysis. PCR amplicons of the G1HE region are shown as controls. (C) mRNA levels of GATA1 in the BM and spleens of WT, Cre, and Gata1^−/y mice measured by conventional (upper panels) and quantitative (lower panels) PCR analyses. PCR amplicons of GAPDH are shown as controls. For qRT-PCR, the value from WT mice was set to 1. The data were obtained from three independent experiments. (D) Protein levels of GATA1 and β-tubulin (control) in the BM and spleens of WT, Cre, and Gata1^−/y mice measured by a Western blot analysis.

The mast cell number and distribution in peripheral tissues were not affected by GATA1 ablation.

We next examined the number and distribution of mast cells in the peripheral tissues of Gata1^−/y mice (Fig. 2). Wild-type (WT) and Cre mice treated with tamoxifen were used as controls. The histological analyses of the ear pinna using both toluidine blue and Alcian blue-safranin O staining showed that the number and distribution of mast cells were comparable between all genotypes of mice (Fig. 2A and B). The FACS analysis demonstrated that c-Kit⁺ FcεRIα⁺ mast cells were detectable in the peritoneal fluid of Gata1^−/y mice. The average frequency of c-Kit⁺ FcεRIα⁺ cells was lower in Gata1^−/y mice (1.60% ± 0.70%) than in WT and Cre mice (2.39% ± 1.06% and 2.98% ± 2.76%, respectively), although the difference was not statistically significant (P = 0.113) (Fig. 2C, left panels, and D). Alcian blue-safranin O staining of the cytospin preparations of peritoneal cells showed differentiated mast cells containing safranin O-positive granules in all three genotypes of mice (Fig. 2C, right panels). The number and distribution of mast cells in the forestomach and glandular stomach were also similar in all three genotypes, as assessed by toluidine blue staining (Fig. 2E and F). Collectively, these results suggest that the ablation of GATA1 in adult mice does not significantly affect the number or distribution of peripheral tissue mast cells.

FIG 2.

The number and distribution of mast cells in peripheral tissues were not affected by Gata1 disruption. (A) Ear skin sections prepared from WT, Cre, and Gata1^−/y mice were stained with toluidine blue (left) and Alcian blue-safranin O (right). Mast cells are indicated by arrows. The scale bar represents 100 μm. (B) Mast cell numbers in ear skin sections. The values represent the average cell numbers ± the SD per field in three separate fields per section. The data were obtained from three mice per group. (C) In the left panels, representative flow cytometry plots show the expression of c-Kit and FcεRIα on peritoneal mast cells isolated from WT, Cre, and Gata1^−/y mice. In the right panels, cytospin preparations of peritoneal mast cells were stained with Alcian blue-safranin O. Original magnification, ×100. (D) Percentages of c-Kit/FcεRIα double-positive cells in peritoneal fluid isolated from WT, Cre, and Gata1^−/y mice. The data are average ± the SD obtained from seven mice per group. (E) Forestomach and glandular-stomach sections prepared from WT, Cre, and Gata1^−/y mice were stained with toluidine blue. Mast cells are indicated by arrows. (F) Mast cell numbers in stomach sections. The values represent the average cell numbers ± the SD per square millimeter of sections stained with toluidine blue. The data were obtained from four mice per group.

Gata1-null bone marrow cells are able to differentiate into mast cells in vitro.

To test the ability of GATA1-null BM cells to differentiate into mast cells in vitro, total BM cells were cultured in the presence of IL-3 and SCF, and then the frequency of c-Kit⁺ FcεRIα⁺ cells was examined by FACS on days 14 and 28 of culture (Fig. 3A). On day 14, the c-Kit⁺ FcεRIα⁺ cells appeared at a frequency of 20 to 30% in all three genotypes (Fig. 3A, upper panels). By day 28, the frequency of c-Kit⁺ FcεRIα⁺ cells was increased in all of the genotypes, but the average frequency of these cells was lower in the Gata1^−/y mice compared to the WT mice (86.7% ± 1.68% and 95.7% ± 1.85%, respectively; P = 0.009) (Fig. 3B). Wright-Giemsa staining of the cytospin preparations showed that the BMMCs were morphologically indistinguishable among these three genotypes (Fig. 3C).

FIG 3.

Gata1-null bone marrow cells differentiate into mast cells in vitro in the presence of IL-3 and SCF. (A) Representative flow cytometry plots showing the expression of c-Kit and FcεRIα on BM cells isolated from WT, Cre, and Gata1^−/y mice on culture day 14 and day 28. (B) Percentage of c-Kit/FcεRIα double-positive cells in BM cells prepared from WT, Cre, and Gata1^−/y mice on culture days 14 and 28. The values represent average cell numbers ± the SD. The data were obtained from four mice per group. **, P < 0.01 (compared to the data for WT day 28). (C) Cytospin preparations of BM cells stained with Wright-Giemsa stain. The scale bar represents 50 mm. (D) FcεRI-dependent β-hexosaminidase release from WT and Gata1^−/y BMMCs on culture days 14 and 28.

We next measured the release of β-hexosaminidase in WT and Gata1^−/y BMMCs to determine whether the antigen-induced mast cell degranulation was affected by the GATA1 ablation (Fig. 3D). The average percentage of β-hexosaminidase release was similar in WT and Gata1^−/y BMMCs (Fig. 3D). Thus, although GATA1 might play a facilitative role, it is not necessary for the formation and maturation of BMMCs.

GATA1 is dispensable for maintaining the mast cell characteristics in BMMCs.

To test whether GATA1 is required for maintaining the mast cell phenotype in BMMCs, BMMCs were generated from Gata1^flox/y mice without tamoxifen treatment, and then 4-hydroxy tamoxifen (4-OHT) was added to the culture medium on day 28 of culture. Twelve days after the first 4-OHT treatment, the Gata1 mRNA level was reduced to 26% of that observed in WT cultures (P = 0.004) (Fig. 4A). Consistent with this finding, the GATA1 protein was undetectable by a Western blot analysis (Fig. 4B). Despite the almost complete GATA1 ablation, most of the cells sustained both c-Kit and FcεRIα expression, as assessed by a FACS analysis (Fig. 4C). Moreover, toluidine blue staining of the cytospin preparations showed that the Gata1^−/y cells contained metachromatic granules in the cytoplasm similar to those observed in the WT and Cre BMMCs (Fig. 4D).

FIG 4.

GATA1 is dispensable for maintaining mast cell characteristics in BMMCs. (A) GATA1 mRNA levels in BMMCs isolated from WT, Cre, and Gata1^−/y mice measured by qPCR, with 18S rRNA used as an internal control. Total RNA was isolated on day 12 after the first 4-OHT treatment. *, P < 0.05 (compared to the data for WT). (B) The protein levels of GATA1 and lamin B (control) in nuclear extracts prepared from WT, Cre, and Gata1^−/y BMMCs were examined by Western blot analysis. The preparation of nuclear protein was done 12 days after the first 4-OHT treatment. (C) Representative flow cytometry plots show the expression of c-Kit and FcεRIα on BMMCs isolated from WT, Cre, and Gata1^−/y mice on days 0 and 12 of 4-OHT treatment. (D) Cytospin preparations of BMMCs stained with Wright-Giemsa stain on days 0 and 12 of 4-OHT treatment. (E) mRNA levels of Gata2, Tal1, Kit, and Cpa3 in Cre and Gata1^−/y BMMCs examined by qRT-PCR. The values from WT BMMCs are set to 1.0, and the relative values are shown. (F) BMMCs (2.0 × 10⁶ cells) isolated from WT or Gata1^−/y mice were injected intraperitoneally into Kit^W-sh/W-sh mice, and the peritoneal fluid was collected 6 to 8 weeks after the first injection. Upper panels show the results of the flow cytometric analysis of the peritoneal cells isolated from Kit^W-sh/W-sh mice injected with PBS, WT BMMCs, or Gata1^−/y BMMCs. Lower panels show cytospin preparations of peritoneal cells stained with Alcian blue-safranin O. Original magnification, ×100.

We then examined the expression levels of the GATA2 and Tal1/Scl transcription factors, both of which are critical for mast cell development (17, 29). As shown in Fig. 4E, the qRT-PCR analysis revealed that the expression of these genes was not affected by the GATA1 ablation. A previous study showed that GATA1 binds to the promoter region of the Kit and Cpa3 genes in BMMCs (30). Despite the fact that the GATA1 protein expression was decreased to an undetectable level, the expression levels of Kit and Cpa3 were unaffected in BMMCs (Fig. 4E). Collectively, these results indicated that GATA1 is not necessary for maintaining the general characteristics of mast cell in BMMCs.

We further examined whether Gata1^−/y BMMCs are able to terminally differentiate into mast cells in vivo. To this end, WT and Gata1^−/y BMMCs on day 12 of 4-OHT treatment were transferred into the peritoneal cavity of mast cell-deficient Kit^W-sh/W-sh mice. At 6 to 8 weeks after transplantation, c-Kit⁺ FcεRIα⁺ mast cells were detectable in the peritoneal fluid of the BMMC-transplanted Kit^W-sh/W-sh mice but not of the PBS-injected control mice. Furthermore, the peritoneal fluid of Kit^W-sh/W-sh mice transplanted with either WT or Gata1^−/y BMMC cells contained safranin O-positive differentiated mast cells (Fig. 4F). Therefore, Gata1^−/y BMMCs appear to be capable of further differentiating into peritoneal mast cells in vivo.

Mast cell colonies developed at a high frequency from Gata1^−/y BM cells in methylcellulose media.

A previous study showed that putative mast cell colonies were frequently formed in agar cultures of BM cells from heterozygous Gata1^Plt13/+ mice, which carry an N-ethyl-N-nitrosurea-induced mutation at the translation initiation codon of Gata1 (31). Because the mouse Gata1 gene is located on the X chromosome, cells having an activated Gata1^Plt13 allele do not produce detectable amounts of GATA1 protein. The mast cell colony-forming cells from the Gata1^Plt13/+ mice exclusively carried the active Gata1^Plt13 mutant allele in spite of the random X inactivation (31). Interestingly, despite the high frequency of mast cell colony-forming cells in the Gata1^Plt13/+ BM, the mast cell population in the peripheral tissues was normal in the Gata1^Plt13/+ mice. Based on this observation, we assumed that the number of mast cell colony-forming cells might be increased in the Gata1^−/y BM, even though the number of peripheral tissue mast cells was not significantly affected by the GATA1 ablation. Of note, the methylcellulose colony assays showed a significantly increased frequency of colony-forming cells with aberrant morphology in the Gata1^−/y BM, while the total colony numbers were comparable among all of the genotypes of BM cells (Fig. 5A and B).

FIG 5.

Mast cell colonies appeared at a high frequency from Gata1^−/y BM cells in methylcellulose culture. (A) Total numbers of colonies grown from WT, Cre, and Gata1^−/y BM cells in methylcellulose culture. Colony counts were performed on the eighth day of culture. (B) Frequencies of colony-forming cells in the WT, Cre, and Gata1^−/y BM. Colony counts were performed on the eighth day of methylcellulose culture. (C) Phase-contrast photomicrographs (left) and cytospin preparations (right) of CFU-Mast and CFU-GM colonies. For cytospin preparations, cells were stained with Wright-Giemsa stain. (D) Results of flow cytometric analysis of colony-forming cells stained for c-Kit/FcεRIα (left panels) or Mac1/Gr1 (right panels). Cells were picked up from CFU-Mast or CFU-GM colonies on the eighth day of methylcellulose culture.

The aberrant colonies developed from Gata1^−/y BM cells showed a large and loosely dispersed appearance, which is compatible with the characteristics of GATA1-deficient mast cell colonies described in a previous study (31) (Fig. 5C). Thus, we designated these colonies CFU-Mast. To further characterize the cellular morphology of the CFU-Mast, the colony-forming cells were picked off and stained with Wright-Giemsa stain on cytospin preparations. The cells from typical CFU-GM colonies were stained with Wright-Giemsa stain as controls (Fig. 5C, right panels). Most of the cells from CFU-GM showed characteristics of macrophages, with condensed nuclei in the peripheral region. Granulocyte-like cells with segmented nuclei were also observed, albeit at a lower frequency. In contrast, the cells from CFU-Mast were uniformly shaped, with rounded nuclei in the central region of the cells. Secretory granules were rarely developed in these cells.

Next, the cells picked up from the colonies were examined by a FACS analysis (Fig. 5D). The frequency of c-Kit⁺ FcεRIα⁺ cells was higher in the cells from CFU-Mast compared to those from CFU-GM (Fig. 5D, left panels). In contrast, the frequency of cells expressing myeloid cell markers, Mac1 and Gr1, was higher in the cells from CFU-GM compared to those from CFU-Mast (Fig. 5D, right panels). Taken together, these results demonstrated that GATA1 ablation led to an increase in CFU-Mast formation in methylcellulose cultures. Consistent with the known roles of GATA1 in erythropoiesis, no erythroid colony was formed from the Gata1^−/y BM cells (Fig. 5B).

GATA1 might play a specific role in the regulation of tryptase genes on Chr17A3.3.

Although the general characteristics of the mast cells were mostly maintained in the GATA1-deficient BMMCs, we assumed that GATA1 deficiency might affect more specific mast cell functions. To explore this issue, we comprehensively examined the gene expression profiles in GATA1-deficient BMMCs by means of microarray analyses using BMMCs transduced with either control or Gata1 siRNA. Because BMMCs are primary cultured cells, we were concerned about nonspecific, individual variability in the transcript abundance. Thus, instead of comparing BMMCs prepared from Gata1^−/y and WT mice, BMMCs prepared from the same individuals were transfected with either control or Gata1 siRNA and used for the microarray analysis. The transduction of Gata1 siRNA resulted in a 4.9-fold decrease in the Gata1 mRNA level compared to control cells. This analysis identified 1,670 differentially expressed genes (768 downregulated and 902 upregulated genes) in the Gata1 knockdown cells.

To gain further insight into the roles of GATA1 in mast cell gene expression, we functionally categorized these genes using the IPA software program. We found that the differentially expressed genes were strongly associated with hematological, inflammatory, and immunological diseases (Fig. 6A). However, only a few mast cell-related genes were listed among the differentially expressed genes in these categories (Fig. 6B). Consistent with the microarray data, the downregulation of Il1a, Epor, and Ahr and the upregulation of Il18, Tnfrsf11b, and Creg2 were confirmed by using qRT-PCR analyses (Fig. 6C).

FIG 6.

Large-scale gene expression profiles in BMMCs transfected with Gata1 siRNA. WT BMMCs were transfected with control or Gata1 siRNA (200 pmol), and total RNA was isolated 24 h after transfection. (A) Pathway analysis was performed using the IPA software program. A comparison of disease-related and biological functions among genes affected in the GATA1 knockdown BMMCs is shown. The red line indicates the threshold for a significant difference (P < 0.05) compared to the BMMCs transfected with control siRNA. (B) The expression levels of the genes listed in the categories of hematological and inflammatory diseases were compared in control and Gata1 siRNA-transfected BMMCs. Heat map comparisons of genes significantly affected (P < 0.05, 2-fold or larger change) in GATA1 siRNA-transfected BMMCs compared to controls are shown. Red indicates relatively high expression, green indicates low expression, and black indicates intermediate expression in GATA1 siRNA-transfected BMMCs compared to controls, as indicated by the color bar. (C) Results of qRT-PCR analysis of three downregulated (Il1a, Epor, and Ahr) and three upregulated (Il18, Tnfrsf11b, and Creg2) genes. The values of the wild-type cells were set to 1.0, and the relative values are shown. (D) Results of qRT-PCR analysis of mast cell-related genes. The values of the wild-type cells were set to 1.0, and the relative values are shown. (E) Configurations of group A tryptase gene loci on mouse chromosome 17A3.3. “Region A” located at the 5′ end of the gene cluster is indicated. (F) Nucleotide sequence of region A. Seven WGATAR sequence motifs are boxed and numbered. All GATA motifs except motif 1 were found to be evolutionally conserved between mouse and rat. The GATA motifs 2, 4, and 6 (in the red box) were identified in recent ChIP-seq studies. (G) The DNA-binding activity of GATA1 to region A, the double GATA region (dblGATA, positive control), and the sixth exon (Gata1 ex6, negative control) of the Gata1 gene was examined using qChIP analysis. The control experiments were performed using rat IgG in place of anti-GATA1 antibodies (N6). *, P < 0.05; **, P < 0.01 (compared to the data for IgG).

We then investigated the expression of 37 mast cell-related genes whose expression and/or function in mast cells have been previously reported (32–50) (Table 2). Again, we found that the fold change values were <2.0 for most of the mast cell-related genes on the list (Table 2). Nevertheless, we noted that the expression levels of mast cell proteases were uniformly decreased, except for that of carboxypeptidase E (Table 2). Moreover, the qRT-PCR analyses revealed that the expression levels of three mast cell tryptases—Tpsg1, Tpsb2 (Mcpt6), and Tpsab1 (Mcpt7)—in the Gata1 knockdown cells were reduced to less than half of those of the control siRNA-treated cells (0.46 ± 0.10, 0.47 ± 0.02, and 0.43 ± 0.10, respectively [n = 4]) (Fig. 6D). In contrast, the expression levels of three mast cell chymases Mcpt1, Mcpt4, and Cma1 were less affected by the Gata1 siRNA treatment (0.81 ± 0.19, 0.64 ± 0.09, and 1.14 ± 0.26, respectively [n = 4]) (Fig. 6D). Similarly, the expression levels of Fceria, Fcerig, and Cpa3 were not reduced in the Gata1 knockdown cells (Fig. 6D). It has been reported that the Tpsg1, Tpsb2, and Tpsab1 genes can be categorized as group 2 tryptase genes residing on mouse chromosome 17A3.3 (51, 52) (Fig. 6E). In order to examine whether GATA1 directly regulates the expression of these genes, we searched for consensus GATA-binding motifs that are evolutionally conserved among mammalian species on this region. We identified a 500-bp region that contains seven repetitive WGATAR sequences, six of which are conserved between mouse and rat, at the 5′ end of the group A tryptase gene cluster (here referred to as region A) (Fig. 6E and F). Furthermore, quantitative ChIP (qChIP) assays showed that the binding activity of GATA1 to this region was comparable to that of the double GATA site of the Gata1 gene, a previously characterized GATA1-binding region in BMMCs (16) (Fig. 6G). Interestingly, this region contains two “chromatin occupancy” consensus sequences for GATA1, (C/G)(A/T)GATAA(G/A/C)(G/A/C), and a composite half E-box-GATA motif (CTGN₈WGATAA), both of which are characterized as functional GATA-binding motifs according to genome-wide ChIP-seq studies (53–55). Therefore, while siRNA-mediated GATA1 knockdown has only a small impact on most mast cell-affiliated genes in BMMCs, GATA1 might play a role in more specific functions of mast cells by regulating group A tryptase genes.

TABLE 2.

Expression of mast cell-enriched genes in the control or GATA1 siRNA-treated BMMCs

Gene	Entrez full name	Gene ID	Fold change	P	Reference
Gata1	GATA binding protein 1	14460	–4.88	6.8E–15
Mast cell protease genes
Mcpt2	Mast cell protease 2	17225	–1.61	1.4E–03
Mcpt4	Mast cell protease 4	17227	–1.91	2.0E–05
Cma2	Chymase 2, mast cell	545055	–1.57	2.3E–03
Tpsab1	Tryptase alpha/beta 1 (Tpsab1)	100503895	–2.22	5.2E–07
Tpsb2	Tryptase beta 2	17229	–1.53	3.6E–03
Cpe	Carboxypeptidase E (Cpe)	12876	1.91	4.0E–05
Mast cell proteoglycan genes
Prg2	Proteoglycan 2	19074	–1.46	8.9E–03	32
Agrn	Agrin	11603	–2.19	3.0E–02	32
Gpc2	Glypican 2 (cerebroglycan)	71951	4.00	3.3E–02	32
Chad	Chondroadherin	12643	1.39	3.5E–02	32
Gpc1	Glypican 1	14733	1.35	3.8E–02	32
Other mast cell-related genes
CtsD	Cathepsin D	13033	1.40	2.1E–02	33
Ctsg	Cathepsin G	13035	–1.35	3.7E–02	34
Timp3	Tissue inhibitor of metalloproteinase 3	21859	1.29	1.9E–02	34
pld2	Phospholipase D family, member 2	18806	–1.45	1.0E–02	35
Dusp14	Dual specificity phosphatase 14	56405	1.71	4.5E–04	34
Maob	Monoamine oxidase B	109731	–1.98	1.0E–05	34
Trib2	Tribbles homolog 2 (Drosophila)	217410	–1.43	4.8E–02	34
Trib3	Tribbles homolog 3	228775	–1.41	1.8E–02	36
C3	Complement component 3 (C3)	12266	1.48	2.2E–16	37
C5ar1	Complement component 5a receptor 1	12273	1.47	1.2E–02	38
Grk5	G protein-coupled receptor kinase 5	14773	–1.64	8.6E–04	39
Tlr1	Toll-like receptor 1	21897	1.37	3.2E–02	40
Tlr2	Toll-like receptor 2	24088	1.24	2.1E–06	40
myd88	Myeloid differentiation primary response gene 88	17874	1.10	3.3E–02	40
Il13	Interleukin-13	16163	–1.42	1.6E–02	41
fgf2	Fibroblast growth factor 2	14173	–2.05	2.8E–45	42
Ccr2	Chemokine (C-C motif) receptor 2	12772	1.61	1.3E–03	43
Cx3cr1	Chemokine (C-X3-C) receptor 1	13051	–1.87	7.0E–05	44
Fyb	FYN binding protein	23880	–1.54	3.3E–03	45
Orai1	ORAI calcium release-activated calcium modulator 1	109305	1.44	1.3E–02	46
Mt1	Metallothionein 1	17748	1.03	1.0E–02	34
Fst	Follistatin	14313	–1.34	4.8E–02	34
Nlrp3	NLR family, pyrin domain containing 3	216799	–1.45	2.3E–02	47
Socs3	Suppressor of cytokine signaling 3	12702	1.38	2.6E–02	48
CD52	CD52 antigen	23833	1.75	1.9E–04	49
CD69	CD69 antigen	12515	1.44	1.2E–02	50

GATA2 plays a more important role than GATA1 in the regulation of most mast cell-specific genes.

We previously reported that the Gata2 mRNA levels are significantly increased during the maturation of BMMCs, whereas the Gata1 gene expression remains largely unaffected throughout BMMC differentiation (16). Given the robust GATA2 induction observed in the BMMCs, we hypothesized that GATA2 might compensate for GATA1 loss in the GATA1 knockdown BMMCs. To test this hypothesis, we examined the expression profiles of five mast cell-related genes, i.e., Kit, Ms4a2 (Fcer1b), Mcp9, Mitf, and Cpa3, in the BMMCs, in which either or both the levels of Gata1 and Gata2 were reduced using siRNAs. The downregulation of the Gata1 and Gata2 expression was confirmed by qRT-PCR (Fig. 7A), as well as Western blot analysis (Fig. 7B). The qRT-PCR analyses showed that expression levels of all mast cell-specific genes were most significantly decreased when both GATA factors were knocked down in comparison with that observed upon the knockdown of either GATA factor individually (Fig. 7C). Furthermore, the qChIP assays showed that both GATA1 and GATA2 bind to the Kit −114Kb enhancer and the Ms4a2 promoter regions, both of which harbor functional GATA-binding sites (30). These results clearly indicate the contribution of GATA1 to the regulation of these genes. Meanwhile, there remains a question as to whether GATA1 and GATA2 equally contribute to the expression of these mast cell-specific genes, or whether GATA2 plays a more dominant role than GATA1. In order to address this question, we used the Gata2^flox allele to conditionally disrupt the GATA2 activity (25). This mutant allele was designed to conditionally delete the Gata2 gene fifth exon encoding the DNA-binding C-finger domain. As described earlier, Gata2^flox mice were crossed with ROSA26CreER^T2 mice, and BMMCs were subsequently generated from the Gata2^flox/flox::ROSA26CreER^T2 mice. At 4 days after the 4-OHT treatment, we confirmed that the transcripts derived from the Gata2 fifth exon sequences were undetectable in the qRT-PCR analysis (Fig. 7F, GATA2ΔCF). Surprisingly, the FACS analysis performed on day 4 of the 4-OHT treatment showed that the frequency of c-Kit⁺ FcεRIα⁺ cells was significantly reduced in the GATA2ΔCF samples compared to that observed in the control BMMCs obtained from the ROSA26CreER^T2 mice (58.7% ± 6.9% and 92.7% ± 5.9%, respectively, P = 0.003 [n = 3]) (Fig. 7E). This observation reveals a sharp contrast with the subtle effects of Gata1-null deficiency noted in the BMMCs. Furthermore, the mRNA levels of Kit, Ms4a2, Mcp9, and Mitf were significantly reduced in the GATA2ΔCF cells, while the Gata1 expression was largely maintained at the wild-type level (Fig. 7F). Collectively, our data suggest that GATA2 plays a more important role than GATA1 in the regulation of most mast cell-specific genes.

FIG 7.

GATA2 plays a more important role than GATA1 in the regulation of most mast cell-specific genes. (A) mRNA levels of Gata1 and Gata2 in BMMCs transfected with the indicated siRNAs. The values from control siRNA-transfected cells were set to 1.0, and the relative values are shown. *, P < 0.05; **, P < 0.01 (versus control; n = 3 for each group). (B) Protein levels of GATA1, GATA2, and lamin B (control) in BMMCs transfected with the indicated siRNAs. (C) mRNA expression levels of Kit, Ms4a2 (Fcer1b), Mcp9, MITF, and Cpa3 genes in cells transfected with the indicated siRNAs. The values from control siRNA-transfected cells were set to 1.0, and the relative values are shown. *, P < 0.05 for the comparisons indicated (n = 3 for each group). (D) The DNA-binding activity of GATA1 and GATA2 to the Kit and Ms4a2 (Fcer1b) loci. Two primer sets were used to verify the Kit −114Kb region and the Ms4a2 promoter. The Kit promoter and the Ms4a2 −2.8Kb region were examined as negative regions. The gray bars indicate the results using rat (GATA1) or rabbit (GATA2) IgG in place of specific antibodies. *, P < 0.05; **, P < 0.01 (compared to the data for IgG). (E) Representative flow cytometry plots showing the expression of c-Kit and FcεRIα in the BMMCs isolated from Cre and GATA2ΔCF mice on days 0 and 4 of 4-OHT treatment. The percentage of c-Kit/FcεRIα double-positive cells among the Gata2ΔCF cells on days 0 and 4 is indicated (n = 3). (F) The mRNA levels of Gata1, the fifth exon of the Gata2 gene (Gata2 ex5), c-Kit, Ms4a2, Mcp9, and MITF in the Gata2ΔCF cells treated with 4-OHT. The ratios of the values obtained on day 4 versus day 0 are shown.

DISCUSSION

The present study demonstrates that the role of GATA1 in mast cell differentiation is more limited than previously anticipated based on in vivo (9–11) and in vitro (12, 56) studies. There are two plausible explanations for this discrepancy. First, while we examined mice lacking an entire Gata1 coding region, the preceding loss-of-function studies of GATA1 were performed using Gata1 knockdown mouse models with intact coding sequences. It was reported that there was defective mast cell differentiation in Gata1^low mutant mice (9), which were generated by a targeted deletion of Gata1 gene enhancer region designated as Gata1 hematopoietic enhancer (G1HE, also referred to as HS1) (6). We and another group (11, 12) previously showed that GATA1 plays an important role in mast cell differentiation using heterozygous Gata1 knockdown mice generated by the targeted inactivation of the hematopoietic specific IE promoter (4). In both mutant mouse models, the Gata1 mRNA expression was severely reduced because of the lack of a promoter or critical regulatory sequences.

The important point here is that the regulation of Gata1 gene expression is distinct between erythroid and mast cell lineages. We recently showed that the proximal double GATA region, but not G1HE, is essential for the Gata1 gene expression in BMMCs and peritoneal mast cells in vivo using transgenic mouse reporter assays (16). Interestingly, the targeted disruption of this region in mice (ΔdblGATA) resulted in an eosinophil-deficient phenotype in vivo, although the BM cells of these mutant mice are able to differentiate into eosinophils ex vivo (7, 57). The BM cells of ΔdblGATA mice are also capable of differentiating into mast cells ex vivo (7), while the formation of peripheral tissue mast cells in vivo has not been examined. Therefore, the precise roles of these regulatory regions in the Gata1 gene expression in mast cells can be determined by comparing mast cell formation between the two mutant mouse models, Gata1^low and ΔdblGATA, in vivo.

With regard to the IE promoter, a previous study demonstrated that the promoter usage of Gata1 gene in mast cells is distinct from that in the erythroid lineage cells, and a considerable amount of Gata1 mRNA is transcribed from an alternatively spliced first exon Ib that is regulated by PU.1 (58). Therefore, the mast cell defects reported in these previous studies using mice with a targeted Gata1 gene regulatory sequence or promoter might have resulted from the aberrant spatial and temporal regulation of Gata1 gene expression rather than reflecting the simple absence of GATA1 expression in differentiated mast cells.

Second, whereas we examined the effects of postnatal Gata1 inactivation in mast cells using inducible gene targeting in the present study, the Gata1 expression was attenuated from the embryonic stage onward in the previous mouse models. Thus, the differences in the results might have been due to the differential roles of GATA1 in mast cell differentiation during embryonic and postnatal hematopoiesis. Importantly, our data suggest a novel role of GATA1 in the regulation of group A tryptase genes located on chromosome 17A3.3. Recent gene targeting studies have shown that mMCP-6, a group A tryptase, plays a key role in the innate immune response against bacterial and parasitic infection (59, 60), as well as the formation of abdominal aortic aneurysms (61). Taken together, with our microarray data revealing that the expression of genes related to hematological, inflammatory, and immunological diseases was affected in the Gata1 knockdown BMMCs, GATA1 might play a central role in mast cell function under certain pathological conditions in adult mice. Because of their progressive anemia, we had limitations when trying to examine Gata1^−/y (Gata1^fl/y::ROSA26CreER^T2) mice under pathological conditions. Recently, several groups have succeeded in performing Cre-loxP-mediated gene targeting, specifically in mast cells, using the regulatory regions of the cpa3, mcpt5, and il4 genes (see references 62, 63, and 64, respectively). These strategies could be applied for generating mast cell-specific Gata1 knockout mice to examine their innate immune responses against bacterial and/or parasitic infection and a variety of mast cell functions observed under pathological conditions.

Our data from the methylcellulose colony assays using bone marrow cells from Gata1^−/y mice are consistent with the previous report of mast cell colony forming cells from chimeric Gata1^Plt13/+ bone marrow cells in agar cultures (31). Interestingly, the methylcellulose colony assays of bone marrow and spleen cells from Gata1^low mice resulted in an abnormal colony formation, with trilineage colonies containing erythroblasts, megakaryocytes, and mast cells (9, 19). Taken together with our present results, these findings suggest that residual low-level GATA1 expression level is required for the formation of this trilineage progenitor. An independent line of evidence also suggests that enhanced proliferation of progenitor cells in vitro might be a principle feature of GATA1 loss, because Gata1-null proerythroblasts cultured on OP9-stroma cells displayed high proliferative activity (65). Notably, these cells differentiate into mast cells in the presence of IL-3 without a restoration of GATA1 (65), supporting our contention that GATA1 is not essential for mast cell differentiation. It remains to be elucidated why Gata1^−/y mice showed normal mast cell numbers in vivo.

We have presented several lines of evidence suggesting that GATA1 contributes to the expression of mast cell-specific genes, although the vast majority of mast cell-specific genes evaluated here were not affected by GATA1 loss. Our data for the GATA2ΔCF BMMCs indicate that the null mutation of Gata1 is likely compensated for by GATA2, with the exception of the regulation of a limited group of mast cell-specific genes, such as mast cell tryptases. The sequence analysis of region A on chromosome 17A3.3 raised an interesting possibility regarding the DNA-binding specificity of GATA1 and GATA2 in the hematopoietic cell lineages. The preceding genome-wide studies of GATA1 ChIP-seq were conducted using either erythroid cell lines (53–55) or a multipotent progenitor cell line harboring the potential to differentiate into erythroid cells (66). Notably, region A contains two “chromatin occupancy” consensus sequences for GATA1, (C/G)(A/T)GATAA(G/A/C)(G/A/C), and a composite half E-box-GATA motif (CTGN₈WGATAA), both of which were characterized as functional GATA-binding motifs in these studies. The former motif has been identified to be the most frequent sequence among the WGATAR-containing peaks bound by GATA1 (53), while the latter has been found to be enriched in de novo GATA1 peaks without prior binding of GATA2 in differentiating erythroid progenitor cells (66). Interestingly, the latter sequence motif is also found to be frequently cooccupied by GATA1 and Scl/Tal1 (55). Therefore, it is conceivable that the properties of GATA factor binding specificity are conserved between erythroid and mast cell lineages and that GATA1, rather than GATA2, preferentially acts on this chromatin region.

In the final experiment, we showed that the ablation of the DNA-binding activity of GATA2 resulted in a significant reduction of the c-Kit⁺ FcεRIα⁺ mast cell fraction and a reduced expression of several mast cell-specific genes, suggesting that GATA2 plays a more important role than GATA1 in the regulation of most mast cell-specific genes. These data suggest that GATA2 is required for the maintenance of the mast cell identity in the differentiated state, in addition to cell lineage specification, as previously reported (17). Previous studies have demonstrated that the downregulation of a cofactor FOG-1 or the CEBP/α (CCAAT enhancer-binding protein α) is required for mast cell lineage specification (67, 68). Therefore, it is likely that GATA2 plays a role in maintaining the repressive state of these genes. Future analyses of GATA2ΔCF cells are expected to reveal mechanistic insight into the cell type-specific gene regulation of these critical innate immune cells.