Drissen et al. conducted a scRNA-seq analysis of mouse pre-granulocyte–macrophage progenitors (pre-GMs) and identified the presence of distinct pre-GM populations, namely GATA1⁺ and GATA1⁻ pre-GMs (62). In vitro culture experiments identified that GATA1⁻ pre-GMs preferentially differentiate into neutrophils and monocytes, whereas GATA1⁺ pre-GMs can differentiate into both eosinophil/mast cell and erythrocyte/megakaryocyte lineages. Tusi et al. conducted a scRNA-seq analysis of c-Kit⁺ cell populations isolated from adult mouse bone marrow by using the inDrops technique (54). Population balance analysis predicted that the differentiation trajectory of basophils/mast cells was coupled with erythrocyte differentiation rather than neutrophil or monocyte differentiation. Supporting this notion, ∼25% of single-cell cultured c-Kit⁺CD55⁺CD105⁺CD71⁻CD150⁺ progenitor cells produced both TER-119⁺ erythroid cells and FcεRI⁺ basophils. Weinreb et al. traced hematopoietic lineage differentiation by combining DNA barcoding with scRNA-seq technologies (55). They conducted scRNA-seq analysis of DNA-barcoded mouse hematopoietic stem cells cultured ex vivo for 2, 4, and 6 days. Lineage tracing identified the presence of progenitor cells that differentiate into basophil/mast cell/eosinophil and erythrocyte/megakaryocyte lineages. Lineage hierarchy tree constructed by clonal couplings inferred that the differentiation trajectory of basophil/mast cell/eosinophil is coupled with that of erythrocyte/megakaryocyte in vitro. However, the conclusion was different when the authors used scRNA-seq datasets obtained from in vivo differentiated DNA-barcoded hematopoietic stem cells, indicating further experimental evidence to verify the coupling of differentiation pathways between basophil/mast cell and erythrocyte lineages.

The association of basophil and erythrocyte differentiation is also reported in humans. Görgens et al. identified that CD133^lo/−CD34⁺ hematopoietic progenitor cells produce both basophils/eosinophils and erythrocytes/megakaryocytes, although these progenitor populations have limited potential to differentiate into neutrophils (48). Drissen et al. conducted scRNA-seq analysis of Lin⁻CD34⁺CD38⁺CD123⁺CD45RA⁻ CMPs in the human bone marrow and identified the heterogeneity within the CMP population (64). Thus, they showed that CMP populations can be subdivided by surface expressions of two markers: CD114 (G-CSFR) and CD131 (IL-3Rβ). Single-cell culture experiments showed that CD114⁺ CMPs produce neutrophil and monocyte lineages, whereas CD131⁺ CMPs produce both eosinophil/mast cell/basophil and megakaryocyte/erythrocyte lineages.

A series of scRNA-seq analyses further support the association of basophil/eosinophil differentiation and erythrocyte/megakaryocyte differentiation. Velten et al. combined flow cytometry and scRNA-seq analyses of human Lin⁻CD34⁺ adult human bone marrow cells and identified that hematopoietic stem cells gradually acquire lineage-committed gene expression profiles during their differentiation (56). Based on the gene expression profiles, early progenitors committed to eosinophils/basophils/mast cells expressing CLC, HDC, and PRG2 are identified in the Lin⁻CD34⁺CD38⁺ population. Progenitors of eosinophils/basophils/mast cells display CD34⁺CD38⁺CD10⁻CD45RA⁻CD135^mid MEP-like surface expression phenotype, indicating the close relationship between eosinophil/basophil/mast cell and erythrocyte/megakaryocyte differentiation. Similarly, scRNA-seq analysis of CD34⁺ human cord blood-derived progenitor cells revealed the coupling of eosinophil/basophil/mast cell differentiation to erythrocyte differentiation by using a minimum spanning tree algorithm (57). Pellin et al. conducted scRNA-seq analysis of Lin⁻ adult human bone marrow cells and showed the potential association between the differentiation trajectory of basophils and erythrocytes/megakaryocytes by using population balance analysis (58). Ex vivo culture experiments identified that Lin⁻CD34⁺CD135 (Flt3)⁺ bone marrow precursor cells preferentially differentiate into monocytes, whereas Lin⁻CD34⁺CD135⁻ cells can produce both basophils and megakaryocytes. A recent study identified that FcεRIα⁺Integrin-β7⁺CD203c⁻ CMPs can produce basophils, mast cells, and erythrocytes, whereas FcεRIα⁺CD203c⁺ CMPs produce only basophils and mast cells but not erythrocytes (65). These results support the notion that the differentiation trajectory of basophils/eosinophils/mast cells is associated with that of erythrocytes/megakaryocytes but not that of neutrophils/monocytes.

Besides the association of the differentiation trajectory of basophils with that of the erythrocytes/megakaryocytes, scRNA-seq analysis revealed the coupled differentiation pathways of basophils and mast cells in mice. Dahlin et al. confirmed the association between basophil and mast cell differentiation pathway by using scRNA-seq analysis of mouse hematopoietic progenitor populations in the bone marrow (40). Force-directed graph visualization identified that entry points into basophil and mast cell differentiation are close, indicating the coupled differentiation trajectory of basophils and mast cells. By contrast, the association of the basophil differentiation trajectory with that of eosinophils is less clarified in mice. The scRNA-seq datasets by Dahlin et al. mapped the eosinophil differentiation trajectory close to that of neutrophils (40). By contrast, scRNA-seq datasets of c-Kit⁺ bone marrow progenitor cells mapped basophils and eosinophils in the same cluster (54, 58), indicating the closely coupled differentiation trajectory. Assessment of the transcriptome of eosinophils by using single-cell RNA-seq is technically challenging, possibly because of the abundant amounts of RNase in their granules, their sensitivity to the shear stress, and degranulation. Recently, Gurtner et al. reported the scRNA-seq data of eosinophils isolated from various tissues in mice by preventing shear stress and resulting mRNA degradation (66). By using such approaches, the relationship among the ontogeny of eosinophil, mast cell, and basophil lineages would be clarified.

Contrary to the case of mice, the association of the basophil-differentiation trajectory with that of mast cells or eosinophils is less clear in humans. Because studies using scRNA-seq have positioned basophils close to mast cells or eosinophils (56, 57), the differentiation of basophils, mast cells, and eosinophils may be closely coupled to each other in humans.

In addition to the characterization of the basophil differentiation trajectory, single-cell transcriptomic analyses were also used to identify basophil-committed progenitor cell populations. By combining highly sensitive scRNA-seq (67) and flow cytometric analyses, Miyake et al. identified CLEC12A^hic-Kit^lo pre-basophils within the mouse bone marrow and bone marrow-derived basophils (BMBAs) (Figure 2) (41). CLEC12A^hi pre-basophils and CLEC12A^lo mature basophils showed distinct cell morphology and cell surface markers. CLEC12A^hi pre-basophils showed kidney-shaped indented nuclei with large cell bodies, whereas CLEC12A^lo mature basophils showed ring-shaped nuclei with smaller cell bodies. CLEC12A^hi pre-basophils and CLEC12A^lo mature basophils in the bone marrow showed FcεRIα^hiCD49b^int and FcεRIα^loCD49b^hi surface expression profiles, respectively. RNA velocity and pseudotime trajectory analyses inferred the sequential differentiation trajectory from Fcer1a^hiCd34^hiKit^hi pre-BMP-like populations into Clec12a^hiCd9^intKit^lo pre-basophils that further differentiate into Clec12a^loCd9^hiKit^lo mature basophils. Consistently, ex vivo culture of CLEC12A^hiCD9^int pre-basophils for 24 h promoted the differentiation into CLEC12A^loCD9^hi mature basophils even without IL-3. Notably, CD34 ⁺FcεRIα BaPs also showed CLEC12A^hiCD9^int phenotype, and only 10% of CLEC12A^hi populations showed positive but low expression of CD34, indicating the possibility that CD34⁺ BaPs are included in the CLEC12A^hi pre-basophil populations. Supporting this notion, scRNA-seq analysis revealed that a small fraction (10%–20%) of Clec12a^hi cells in the pre-basophil cluster showed Cd34^hiKit^lo phenotype. Transcriptomic analysis revealed that pre-basophils and mature basophils highly expressed genes associated with cell proliferation and effector functions, respectively. Consistently, pre-basophils showed higher proliferative capacity than mature basophils. By contrast, mature basophils showed greater capacity for degranulation and de novo cytokine production in response to antigen/IgE stimulation. Notably, when basophils are stimulated with non-IgE stimulation, such as IL-3, IL-3+ IL-18, IL-3 + IL-33, and IL-3 + LPS, pre-basophils produced a greater amount of IL-4 than mature basophils, indicating that the responsiveness to stimulants is different between pre-basophils and mature basophils. Similarly, CD34⁺ basophil progenitors produced a greater amount of type 2 and proinflammatory cytokines than mature basophils in response to IL-33 stimulation (68). Thus, the bulk transcriptomic analysis identified that pre-basophils showed limited gene expression changes in response to antigen/IgE stimulation although FcεRIα expression on pre-basophils was higher than mature basophils. Bulk transcriptomic analysis further showed that IL-3-stimulated pre-basophils and mature basophils showed distinct gene expression profiles.

Reanalysis of previously published scRNA-seq analysis dataset obtained by Weinreb et al. identified Clec12a^hi pre-basophils among Mcpt8⁺ basophil populations. Likewise, pre-basophil-like precursor populations were also reported by other groups. The scRNA-seq analysis of basophil lineage populations by using Myb-68 GFP⁺ cells identified two heterogeneous basophil populations, namely Fcer1a^hi basophil1 and Fcer1a^int basophil2 populations (16), which correspond to pre-basophils and mature basophils, respectively. Park et al. identified CD34⁻CD200R3⁺FcεRIα^hi transitional basophils (tBasos) in the bone marrow with high proliferative capacity (42).

Under homeostatic conditions, pre-basophils reside in the bone marrow and are rarely detected outside the bone marrow. By contrast, 7 days after infection with intestinal helminth Nippostrongylus brasiliensis (Nb), CLEC12A^hi pre-basophils egress the bone marrow and can be detected in the spleen, peripheral blood, and lungs. Notably, CLEC12A^hi pre-basophils detected in the infected lungs retained their proliferative capacity and expressed IL-4 comparable to mature basophils. Moreover, the scRNA-seq analysis of basophils accumulating in the skin infection site after the second Nb infection demonstrated that pre-basophils detected in the skin infection site show transcriptional profiles similar to those of pre-basophils in the bone marrow of noninfected mice.

IL-3 promotes blood basophilia during Nb infections (69, 70). Notably, mice treated with IL-3 complexes also showed the emergence of CLEC12A^hi pre-basophils in the peripheral blood. Therefore, IL-3 upregulation caused by helminth infection may promote the egress of pre-basophils from the bone marrow. Further mechanistic study identified that IL-3 promotes the downregulation of chemokine receptor CXCR4 on basophils in the bone marrow, which prevents the retention of pre-basophils in the bone marrow and promotes the appearance of pre-basophils in the periphery.

Contrary to the case of mouse basophils, the committed progenitors for human basophils remain less identified. Providing that human basophil-like cells can be induced by culturing CD34⁺ progenitor cells isolated from the adult human bone marrow in the presence of IL-3, progenitors possessing basophil-differentiation potential may be found in the CD34⁺ cells in the bone marrow (71). Similarly, human mast cells can be generated ex vivo by culturing CD34⁺c-Kit⁺ cells isolated from the peripheral blood (72, 73), indicating the presence of mast cell-committed progenitors in the peripheral blood CD34⁺c-Kit⁺ cells. In line with this, recent studies have discovered Lin⁻CD34^hic-Kit^int/hiFcεRIα⁺ mast cell-committed precursor cells in the peripheral blood (74, 75).

STAT5 and GATA2 regulate basophil and mast cell differentiation. Bone marrow chimeric mice reconstituted with STAT5-deficient cells showed a markedly reduced number of basophils and tissue mast cells (76, 77). Induced GATA2-deficiency significantly reduced the surface expression of FcεRIα and E-cadherin on basophils and c-Kit on mast cells (39, 78), which almost abolished the generation of basophils and tissue mast cells (38, 78). Considering that STAT5 can directly bind to the promotor region of GATA2 and GATA2 overexpression restores the phenotype of induced STAT5-deficiency, the STAT5–GATA2 signaling may be critical for basophil and mast cell differentiation. Moreover, the STAT5–GATA2 signaling is critical for the survival of mature basophils and mast cells. Notably, haploinsufficiency of Gata2 gene affects the generation of mast cells but not basophils, indicating that mast cells are more heavily dependent on GATA2 TF than basophils, especially regarding their generation. A recent study further identified that GATA2 directly promoted chromatin remodeling at super-enhancer regions thus robustly regulating the expression of mast cell-associated genes and responsiveness to antigen/IgE stimuli (84). Thus, STAT5–GATA2 signaling is critical for both the differentiation and maintenance of basophils and mast cells in mice. However, the factor(s) inducing STAT5 activity for basophil/mast cell differentiation under homeostatic conditions remains unclear.

IL-3 is a growth factor promoting basophil differentiation and proliferation. The ex vivo culture of bone marrow cells in the presence of IL-3 induces the generation of BMBAs. Moreover, systemic administration of IL-3 complexes in mice increases the number of mature basophils and basophil precursor cells in the bone marrow and spleen (77, 85). Thymic stromal lymphopoietin (TSLP) is another growth factor that can promote the differentiation from basophil precursor cells into mature basophils. Systemic administration of TSLP increases the number of mature basophils in the spleen in an IL-3–IL-3R-independent manner (86). Notably, IL-3-elicited basophils and TSLP-elicited basophils show transcriptionally and functionally distinct properties. IL-3-elicited basophils are superior to TSLP-elicited basophils in the degranulation capacity in response to IgE-dependent stimuli. By contrast, TSLP-elicited basophils are more responsive to stimulation by IL-3, IL-18, and IL-33, and produce a greater amount of IL-4 compared with IL-3-elicited basophils. However, neither IL-3 nor TSLP is essential for basophil differentiation, which was evidenced by both IL-3 receptor-deficient mice and TSLP-receptor mice showing a normal number of basophils in the bone marrow (86). Moreover, basophil number was also intact even in the deficiency in the IL-3 and TSLP receptors (86), indicating that factor(s) other than IL-3 or TSLP played critical roles for the STAT5-mediated basophil differentiation.

IRF8 plays critical roles in the development of basophils and mast cells. Mechanistically, IRF8 regulates GATA2 expression in granulocyte progenitors; therefore, IRF8 deficiency impairs the generation of bipotent pre-BMPs in the bone marrow, leading to impaired development of basophils and mast cells in the bone marrow (79). IRF8-deficient mice showed significantly reduced pre-BMPs, BaPs, mature basophils, and MCPs in the bone marrow. However, IRF8-deficient mice showed a normal number of MCPs and mature mast cells in the tissues (e.g., intestine, skin, peritoneal cavity). Considering that the number of BMCPs in the spleen was unaffected by IRF8 deficiency, multiple pathways might have existed to retain tissue mast cell numbers.

The role of GATA1 is implicated in basophil and mast cell differentiation, although the contribution of GATA1 in mast cell differentiation remains controversial (87–89). Among hematopoietic cell lineages, GATA1 plays critical roles in the differentiation of erythrocytes, megakaryocytes, and eosinophils. GATA1-deficient mice result in embryonic lethality due to severe anemia, whereas ΔdblGATA mice lacking the double-GATA enhancer site in the Gata1 promotor region show selective deficiency in eosinophil lineages with only mild anemia (90). The number of mature basophils and basophil progenitor cells in the bone marrow is partially reduced in ΔdblGATA mice, possibly due to the decreased Gata1 expression in basophils (80). Moreover, basophils from ΔdblGATA mice show impaired degranulation and cytokine production in response to antigen/IgE stimulation. By contrast, GATA1-low mutant mice (91) that lack the upstream enhancer and promotor sequences of Gata1 gene displayed an increased frequency of basophils in the bone marrow. Interestingly, basophils from GATA1-low mutant mice show reduced CD49b expression, indicating the possibility that GATA1 directly regulates CD49b expression or GATA1 promotes the differentiation from CD49b^lo pre-basophils into CD49b^hi mature basophils. Importantly, GATA1-low mutant mice showed increased Gata2 expression in basophils, whereas ΔdblGATA mice showed normal Gata2 expression in basophils (39, 80). Therefore, the discrepancy in the basophil numbers between ΔdblGATA mice and GATA1-low mutant mice might stem from the extent of the GATA2 expression in basophils.

A recent report identified that the use of the Myb-68kb enhancer region is largely restricted to basophil and mast cell lineages (16). CRISPR-mediated depletion of Myb-68 enhancer results in the reduced Myb expression and impaired generation of basophils and mast cells in vitro, indicating the role of Myb-68 enhancer in basophil and mast cell differentiation.

C/EBPα and MITF are critical TFs regulating basophil and mast cell differentiation, respectively (38). For differentiation into basophils, the STAT5–GATA2 signaling induces C/EBPα activity to promote basophil-associated gene expression. By contrast, for differentiation into mast cells, the STAT5 signaling induces MITF activity to promote mast cell-associated gene expression. In line with this, MITF-deficient mice lack mature mast cells in various tissues (81, 82). C/EBPα and MITF repress the expression of other genes by directly binding to the promotor region of MITF and CEBP/α gene, respectively. Consistently, inducible knockdown of the C/EBPα gene promotes the differentiation from pre-BMPs to mast cells. Interestingly, inducible loss of the C/EBPα gene in mature basophils induces the generation of c-Kit⁺ mast cell-like cells. Conversely, approximately half of the bone marrow cells from MITF-mutant mice failed to generate c-Kit⁺CD49b⁺ mast cells but generated c-Kit⁻CD49b⁺ basophil-like cells when cultured in the presence of IL-3 for 28 days. A similar phenomenon was observed in Ikaros-deficient mice (37). When bone marrow cells from Ikaros-deficient mice were cultured with IL-3 and stem cell factor (SCF) for 6 weeks, some cells showed c-Kit⁻FcεRIα⁺ basophil-like surface expression phenotype. Mechanistically, Ikaros directly binds to the promotor region of Cebpa and represses the C/EBPα expression possibly through histone modification.

RUNX1, a critical regulator of hematopoietic stem cells, is involved in basophil differentiation. Distal (P1) and proximal (P2) promotors regulate RUNX1 expression (92). Deficiency in P1 promotor results in significantly reduced mature basophils and basophil progenitor cells in the bone marrow (44). However, P1-RUNX1 deficiency has a limited impact on the number of mast cells and eosinophils, indicating the critical roles of P1-RUNX1 in basophil lineage differentiation. The zinc-finger TF promyelocytic leukemia zinc finger (PLZF encoded by the Zbtb16 gene) is essential for the development of several innate lymphoid cells, including NKT cells, mucosal-associated invariant T (MAIT) cells, and group 2 innate lymphoid cells (ILC2s) (93). A recent study showed that Zbtb16 was highly expressed in basophil progenitor cells (83). PLZF deficiency partially reduced the number of basophil progenitors and mature basophils in the bone marrow. PLZF deficiency also influenced IL-4 production of basophils in response to IgE-dependent stimulation. Therefore, PLZF plays key roles in the development and function of basophils. Although PLZF is expressed in mature mast cells, the effect of PLZF deficiency on mast cells remains unclear.

A recent report has identified that the TF NFIL3 is upregulated during basophil maturation. However, basophil-specific Nifl3-deficient mice show a normal number of mature basophil and basophil precursor (tBaso) populations, indicating the possible roles of NFIL3 in the effector functions of mature basophils. Indeed, basophil-specific Nifl3-deficiency partly impairs IgE-mediated activation of mature basophils (42). Moreover, basophil-specific Nifl3-deficient mice show reduced ear thickening in the hapten oxazolone-induced atopic dermatitis model (42), possibly due to the reduced IL-4 expression in skin-infiltrating basophils.

A genetic association study using whole genome sequencing data from Estonian Biobank identified the strong association of basophil counts with the SNPs in CEBPA and GATA2 gene loci (94), indicating that GATA2 and CEBP/α possibly regulate basophil differentiation in humans. The mutation in the +39-kb enhancer region of CEBPA (rs787444187) influences the basophil count but not the number of other myeloid lineages. Interestingly, ATAC-seq analysis identified that the open chromatin region containing CEBPA +39-kb enhancer is only present in CMPs but not in GMPs or MEPs, indicating that basophils possibly differentiate from CMPs but not from GMPs or MEPs in humans, which is consistent with the previous findings. When the CEBPA +39-kb enhancer region is mutated in human hematopoietic progenitor cells, the differentiation into basophils is partially blocked, but the differentiation into mast cells is rather enhanced (94), possibly through reduced CEBPA expression.

GATA2 deficiency syndrome is caused by heterozygous and loss-of-function mutation in GATA2 and displays severe abnormalities in multiple myeloid and lymphoid lineages, including monocytopenia, neutropenia, and dendritic cell deficiency (95). Moreover, patients with GATA2 deficiency syndrome develop myeloid neoplasms, including myelodysplastic syndrome and acute myeloid leukemia, with a median age of onset at 17 years. Consistent with the findings in Gata2-deficient mice, mast cells generated from hematopoietic progenitor cells isolated from GATA2 deficiency syndrome showed reduced FcεRIα and c-Kit expressions, resulting in the defective degranulation capacity in response to antigen/IgE and SCF stimulation (96). Moreover, patients with GATA2 deficiency syndrome have decreased surface expression of FcεRIα on peripheral blood basophils. Interestingly, patients with GATA2-deficiency syndrome had a reduced prevalence of IgE-medicated hypersensitivity syndrome possibly due to the reduced FcεRI expression on basophils and mast cells. However, whether GATA2 deficiency in humans affects the generation of basophils and mast cells in the bone marrow remains unclear, although the frequency of basophils in the peripheral blood is unaffected by GATA2 deficiency in humans.