IL-3 Induces Basophil Expansion In Vivo by Directing Granulocyte-Monocyte Progenitors to Differentiate into Basophil Lineage-Restricted Progenitors in the Bone Marrow and by Increasing the Number of Basophil/Mast Cell Progenitors in the Spleen¹

Recent work has established important roles for basophils in initiating (1, 2, 3) and augmenting Th2-type immune responses to allergen challenges and parasitic infection (4, 5, 6), mediating anaphylaxis (7), and maintaining B cell memory responses (8). However, due to the difficulty of obtaining a sufficient numbers of basophils, basophil biology has not yet been studied extensively.

Basophils are a minor cell population, constituting <1% of peripheral blood and bone marrow cells. Basophils are normally generated from GMPs, which also give rise to neutrophils, monocytes, eosinophils, and mast cells. These cells share many common characteristics with mast cells, such as expression of a high affinity immunoglobin (IgE) receptor (FcεRI), and contain many of the granules that can be found in mast cells (9, 10). Basophils expand dramatically in number after infection with intestinal nematode, Nippostrongylus brasiliensis (11). They also increase in number after protease allergen challenge (2).

IL-3, produced primarily by CD4⁺ T cells (12, 13, 14), has been shown to be pivotal in expanding basophils. In vitro, IL-3 has been demonstrated to induce basophil differentiation and to enhance acute IL-4 production in mouse basophils (15, 16, 17). Galli and colleagues (18) reported that mice lacking IL-3 failed to show increased numbers of basophils and failed to expel nematode Stronglyoides. Nonetheless, mechanisms underlying basophil expansion remain obscure. Mature basophils have a short half-life and respond to IL-3 stimulation with a limited proliferation capacity (13, 19). Thus, it is likely that IL-3 might induce basophil expansion by enhancing basophil lineage commitment.

In this study, we report that administrating the IL-3 complex (IL-3 plus anti-IL-3 Ab) in vivo greatly facilitated the differentiation of GMPs into basophil lineage-restricted progenitors (BaPs)⁴ specifically in a STAT5 signaling-dependent manner and increased the number of basophil/mast cell progenitors (BMCPs) in the spleen. We showed that GMPs, but not common myeloid progenitors (CMPs), expressed low levels of IL-3 receptor. The IL-3 receptor expression was dramatically up-regulated in BaPs, but not eosinophil lineage-restricted progenitors (EoPs). We showed that ∼38% of BMCPs expressed the IL-3 receptor. The IL-3 receptor expression patterns might explain why IL-3 specifically expanded basophils in vivo.

C57BL/6J and B6.SJL-Ptprc^a Pepc^b/BoyJ (CD45.1) mice were purchased from The Jackson Laboratory. A pair of STAT5^+/− breeder mice (20) was provided by Dr. Lothar Hennighausen of the National Institutes of Health (Bethesda, Maryland). These mice have been bred and maintained under pathogen-free conditions at the animal facility of National Jewish Health. A pair of IL-3^−/− mice (18) was provided by Dr. Glenn Dranoff of Dana-Farber Cancer Institute (Boston, MA) via Dr. Shigeo Koyasu of Keio University School of Medicine (Tokyo, Japan). IL-3^−/−, CD45.1, and C57BL/6J mice were used at 6–14 wk of age. E14.5 STAT5^−/− fetuses were used for generating mice chimera. Animals were handled according to protocols approved by the National Jewish Health Animal Care and Use Committee.

IL-3 (10 μg) was mixed with either anti-IL-3 Ab (5 μg; MP2–8F8, BD Pharmingen) at room temperature for 1 min based on published methods (21). The cytokine and Ab mixture (in 0.2–0.3 ml of PBS) was injected into mice via the lateral tail vein. Three days after injection, mice were killed and cells were obtained for flow cytometric analysis.

For flow cytometric analysis of basophils, mast cells, and eosinophils, cells were incubated with anti-FcγRII/III (2.4G2) at 4°C for 10 min and stained with FITC-labeled anti-c-kit (2B8), biotin-labeled anti-FcεRIα (MAR-1), and PE-labeled Siglec-F (E50–2440) Abs, followed by streptavidin-allophycocyanin. Basophils were further characterized with PE-labeled anti-Thy1.2 (30-H12) or PE-labeled anti-CD49b (DX5) Abs. Basophils were defined as FcεRIα⁺c-kit⁻, while mast cells were defined as FcεRIα⁺c-kit⁺ cells. Eosinophils were marked as Siglec-F⁺ or Siglec-F⁺ and CCR3⁺ cells (22, 23). For detecting donor-derived basophils, Pacific Blue-labeled anti-CD45.2 Ab was used (clone 104, Biolegend).

CMPs, GMPs, BaPs, BMCPs, and EoPs were FACS sorted or analyzed according to the published protocols (24, 25, 26, 27) with some of the Abs labeled with different fluorochromes. CMPs and GMPs were identified as Lin⁻IL-7Rα⁻Sca-1⁻c- kit⁺CD34⁺FcγRII/III^low and FcγRII/III^high cells, respectively. For staining, bone marrow cells were incubated with biotin-labeled lineage markers (CD3 (CT-CD3), CD4 (RM4–5), CD8 (53–6.7), B220 (RA3–6B2), Gr-1 (RB6–8C5), and CD19 (1D3), biotin-labeled IL-7Rα (A7R34)), followed by streptavidin-PE-Cy5 (BD Biosciences). PE-labeled anti-FcγRII/III, allophycocyanin-Cy7-labeled anti-c-kit, Alexa Fluor 647-labeled anti-CD34 (RAM34), and Pacific Blue-labeled anti-Sca-1 (D7) mAbs were also added to the staining mix. EoPs were defined as Lin⁻Sca-1⁻CD34⁺c-kit^lowIL-5Rα⁺ cells. Staining was conducted by incubating bone marrow cells with 2.4G2 at 4°C for 10 min, before adding the biotin-labeled lineage markers (followed by streptavidin-PE-Cy5), followed by PE-labeled anti-IL5Rα (Τ21.2), allophycocyanin-Cy7-labeled anti-c-kit, Alexa Fluor 647-labeled anti-CD34 (RAM34), and Pacific Blue-labeled anti-Sca-1 (D7) mAbs were also added to the staining mix. BMCPs were identified as Lin⁻c-kit⁺β7^highFcγRII/III^high cells. For staining of BMCPs, spleen cells were incubated with biotin-labeled lineage markers, followed by streptavidin-PE-Cy5 (BD Biosciences). PE-labeled anti-FcγRII/III (BD Biosciences), allophycocyanin-labeled anti-c-kit (BD Biosciences), and FITC-labeled anti-β7 integrin (Biolegend, FIB504) mAbs were also added to the staining mix. The number of BMCPs was calculated by multiplying the total number of spleen cells by the percent of Lin⁻c-kit⁺β7^highFcγRII/III^high cells. BaPs were identified as Lin⁻CD34⁺FcεRIα⁺c-kit⁻ cells. Staining was conducted by incubating bone marrow cells with 2.4G2 at 4°C for 10 min, before adding the biotin-labeled lineage markers (followed by streptavidin-PE-Cy5), allophycocyanin-Cy7-labeled anti-c-kit, Alexa Fluor 647-labeled anti-CD34, and PE-labeled anti-FcεRIα Abs. Cells were acquired by CyAn (DakoCytomation) and analyzed using FlowJo software (Tree Star). The number of CMPs and GMPs were calculated by multiplying the total number of bone marrow cells by the percent of Lin⁻Sca-1⁻IL-7Rα⁻c-kit⁺ and the percent of CMP or GMP. The number of BaPs was calculated by multiplying the total number of bone marrow cells by the percent of Lin⁻CD34⁺ and the percent of BaPs.

For detecting donor-derived BaPs, the procedure was the same as that described above for analyzing BaPs, except that Pacific Blue-labeled anti-CD45.2 Ab was also used. For IL-3 receptor staining on CMPs, GMPs, and BaPs, bone marrow cells were stained using the same procedures as described above, except that FITC-labeled anti-IL-3 receptor Ab was also added.

Cell sorting was conducted using a MoFlo machine (DakoCytomation). CMPs, GMPs, BaPs, and basophils were sorted using the surface phenotype defined above, except that PE-Cy5-labeled anti-Sca-1 Ab instead of Pacific Blue-labeled anti-Sca-1 Ab was used for sorting CMPs and GMPs. All Abs used for flow cytometry analysis and sorting were purchased from BD Pharmingen, eBioscience, or Caltag Laboratories.

Single cell suspensions were prepared from E14.5 fetal livers of STAT5^+/+ and STAT5^−/− mice. Cells (2 × 10⁶) were injected into sublethally irradiated CD45.1 congenic recipient mice (600 rad) i.v. Mice were kept under pathogen-free conditions with water-containing antibiotics. Four to six weeks later, cells were isolated from the bone marrow of the reconstituted mice and were subjected to flow cytometric analysis.

All of the error bars in this report represent the SD. The difference between two samples was analyzed with Student’s t test.

Treatments with the IL-3 complex (IL-3 plus anti-IL-3 Ab) for 8 days induced the expansion of basophils, mast cells, and eosinophils (6). Because basophils might have expanded more rapidly than mast cells, we tested a short-term expansion protocol. Three days after a single injection of the IL-3 complex, the total number of bone marrow, spleen, and peritoneal cells was comparable between two groups with and without the IL-3 complex injection. Interestingly, we observed an increase in the number of FcεRIα⁺c-kit⁻ basophils by about 8-fold in the bone marrow (Fig. 1, A, C, and D), 8-fold in the peripheral blood, and 5-fold in the spleen (Fig. 1, C and D). The increase caused by a single injection lasted for 6 days (Fig. 1,B). By contrast, injection with IL-3 or with the sham control (anti-IL-3 Ab alone) did not increase the number of FcεRIα⁺c-kit⁻ basophils (Fig. 1,A). Binding IL-3 with anti-IL-3 Ab has been shown to increase the half-life of IL-3 significantly (28). No significant changes in the number of FcεRIα⁺c-kit⁺ mast cells were detected in the peritoneal cavity (Fig. 1, C and D) (p > 0.05). Injection of the IL-3 complex only caused a <2-fold increase in the number of Siglec-F⁺ eosinophils in the bone marrow, peripheral blood, and the spleen (Fig. 1, C and D). Further phenotypical and morphological analyses of the isolated FcεR1α⁺c-kit ⁻ cells revealed that expanded FcεR1α⁺c-kit⁻ cells resembled those basophils reported previously (Fig. 2 and Ref. 29). These results indicate that short-term IL-3 treatment specifically induces expansion of basophils in vivo.

It has been shown that IL-3 activates STAT5 (30) and that STAT5 signaling is required for the mast cell development, survival, and functions (31, 32). Although basophils and mast cells share many similarities, basophils do differ significantly from mast cells. Thus, it merits an investigation of the role of STAT5 in basophil development and IL-3-driven basophil expansion. In this study, we chose to evaluate the role of STAT5 in basophils in an experimental system where the STAT5 protein was completely absent. The complete absence of STAT5 results in perinatal death (20). We generated STAT5^−/− chimera mice by injecting embryonic day 14.5 fetal liver (E14.5 FL) cells into sublethally irradiated (600 rad) recipient mice. Mice reconstituted with STAT5^−/− FL cells contained 10% of donor-derived cells, while mice reconstituted with WT littermate control FL cells contained 80% of donor-derived cells. Total number of basophils was greatly reduced in STAT5^−/− chimeric mice (Fig. 3, a 93% reduction), demonstrating that basophil development was greatly impaired in the absence of STAT5 signaling. Although the percentage of STAT5^−/− FL cell-derived basophils increased in response to IL-3, the total number of STAT5^−/− FL cell-derived basophils failed to increase significantly (Fig. 3). These results demonstrate that STAT5 is required for basophil development and IL-3-driven expansion of basophils.

To analyze the mechanisms by which IL-3 induces the dramatic expansion of basophils in vivo, we first established the methods for analyzing CMPs, GMPs, and BaPs in the bone marrow according to published protocols (Fig. 4,A and Ref. 24, 25, 26). Using the defined sorting gates, we were able to achieve a 52-fold enrichment for CMPs, a 36-fold enrichment for GMPs, and a 5-fold enrichment for BaPs (Fig. 4,B). The BaPs-derived colonies were highly enriched with basophils, consisting of 87% basophils (Fig. 4 B). Mature basophils failed to generate any colonies (data not shown). Thus, the relative low colony-forming efficiency of the enriched basophil progenitors may reflect a lower proliferation capacity for such committed basophil progenitors.

Upon treatment with the IL-3 complex, we observed that the number of BaPs increased dramatically by ∼8-fold (Fig. 5, A and B). CMP numbers did not change significantly by the IL-3 treatment, while the number of GMPs increased moderately (Fig. 5, A and B). The IL-3 treatment did not significantly enhance the development of GMPs into EoPs (Fig. 5,C). Total number of BaPs was greatly reduced in STAT5^−/− chimeric mice (Fig. 5,D, a 93% reduction). Although the percentage of STAT5^−/− FL cell-derived BaPs increased in response to IL-3, the total number of STAT5^−/− FL cell-derived BaPs failed to increase significantly (Fig. 5 D). Together, these results demonstrate that IL-3 expands basophils by directing GMPs to differentiate into BaPs in the bone marrow.

An additional pathway for basophil differentiation has also been reported. Akashi and colleagues (24) have described a subset of BMCPs in the spleen, which expresses high levels of β7 integrin. BMCPs have been proposed as the immediate progeny of GMP in the bone marrow and are thought to give rise to mature basophils without going through the intermediate BaP developmental stage. We found that the IL-3 complex treatment resulted in ∼4-fold increase in the number of BMCPs in the spleen (Fig. 5, A and B). This result demonstrates that the IL-3 complex treatment also expands basophils in the spleen by directing BMCPs to differentiate into basophils.

To understand why a short-term IL-3 treatment resulted in the expansion of basophils in remarkable fashion, we examined the developmental stage at which IL-3 receptor begins to be expressed. The IL-3Rα-chain expression on CMPs, GMPs, BMCPs, BaPs, basophils, EoPs, and eosinophils was measured using six-color FACS analysis. We found that CMPs did not express a detectable IL-3Rα-chain while ∼7% of GMPs expressed the IL-3Rα-chain (Fig. 6,A), consistent with previous report (33). Approximately 38% of BMCPs expressed the IL-3Rα-chain. Nearly all BaPs and basophils expressed the IL-3Rα-chain at high levels. The up-regulation of IL-3 receptor expression did not occur on EoPs or eosinophils (Fig. 6 A).

Although IL-3 treatment increased the percentage of IL-3 receptor-expressing GMPs by 2-fold (presumably by inducing the division of the IL-3 receptor-expressing GMPs), IL-3 treatment did not increase the mean fluorescence intensity of IL-3 receptor expression on progenitors at any developmental stages examined (Fig. 6,A; p > 0.05). The IL-3 receptor expression on myeloid progenitors was not affected in the absence of IL-3 (Fig. 6,B) or in the absence of STAT5 signaling (Fig. 6 C). Together, our data indicate that selective up-regulation of the IL-3 receptor on BaPs and BMCPs might explain the remarkable expansion of baosphils as a result of the IL-3 treatment.

Expansion of basophils is imperative to the development of type-2 immunity. In this study, we sought to examine the mechanisms by which IL-3 dramatically increases the number of basophils. We used a short-term in vivo expansion system to examine the effects of IL-3 on the differentiation of myeloid progenitors into basophil lineage. In our short-term expansion experimental system, we found that IL-3 induced basophil-specific expansion. This result differs from that observed by Finkleman and colleagues (6). They used an 8-day IL-3 treatment protocol and observed that IL-3 induced dramatic expansion of mast cells and modest expansion of basophils and eosinophils in the spleen (6). The discrepancy might be explained by the different time points at which we chose to analyze basophil numbers. Three-day treatment with the IL-3 complex resulted in a 4-fold increase in the number of basophils in the spleen but did not increase the number of mast cells or eosinophils. Consistent with these observations, we found that the 3-day IL-3 treatment regimen caused a 4-fold increase in the number of BMCPs. Compared with the degree of increase in the number of BaPs, the fold of increase in the number of BMCPs was lower. We believe that this lower increase is due to a lower percentage of BMCPs that express the IL-3 receptor. Mast cells are known to live longer and undergo a slower maturation process (19, 34). Eight days of treatment with the IL-3 complex might preferentially induce mast cell expansion. Our time course experiment, showing that IL-3-induced basophil expansion that lasted up to seven days, was consistent with this notion. Taken together, our data demonstrate that a short-term IL-3 treatment leads to basophil-specific expansion in the bone marrow and in the spleen.

Based on in vitro data, IL-3 is believed to be a hematopoietic factor that has broad effects; it has been shown to promote macrophage, eosinophil, basophil, and mast cell differentiation (35). In vitro, we have consistently observed that bone marrow progenitors gave rise to 10–20% of basophils in the presence of IL-3. Yet, the in vivo effect of IL-3 stimulation is remarkably specific in expanding the number of basophils. Our analysis of IL-3 receptor expression patterns on progenitors may provide part of the explanation. Acquisition of IL-3 receptor expression by myeloid progenitors indicates the readiness of particular myeloid progenitors to respond to IL-3 robustly. We found that at the GMP developmental stage, 7% of GMPs began to express low levels of IL-3 receptor. At the BaP stage, all BaPs expressed high levels of the IL-3 receptor. Up-regulation of the IL-3 receptor did not occur on EoPs. Acquisition of IL-3 receptor expression appears to be controlled stochastically. Mice deficient in IL-3 have normal IL-3 receptor expression throughout basophil development. Thus, IL-3 does not appear to regulate its own receptor expression. Furthermore, expression of IL-3 receptor on basophils did not require STAT5 signaling.

The present study revealed that the development of STAT5^−/− FL-derived basophils was greatly impaired in the chimera model when sublethally irradiated recipient mice were used. We also observed that the mice reconstituted with STAT5^−/− FL cells contained only 10% of donor-derived cells, while those reconstituted with WT FL cells contained 80% of donor-derived cells. STAT5 is activated by multiple cytokines including IL-2, IL-3, IL-5, IL-7, SCF, Flt-3 ligand, GM-CSF, and erythropoietin (36) and was shown to play a critical role in hematopoietic stem cell functions (37, 38, 39); thus, we cannot rule out the possibility that STAT5 deficiency might have affected the earlier hematopoietic stem cell development. Indeed, STAT5 has been demonstrated to play a critical role in hematopoietic stem cell functions (40). One of the caveats of interpreting the STAT5 data obtained in this study is that we cannot identify the exact developmental stage at which STAT5 is required. STAT5 could be required at the GMP to BaP stage; alternatively, it could act at a much earlier stage. Ideally, if basophil-specific Cre mice were available, we would cross the basophil-specific Cre mice with STAT5 flox/flox mice. Because basophil-specific Cre mice were not available, a more definitive answer to which developmental stage STAT5 is required awaits the generation of basophil-specific Cre mice. Although our results do not determine the exact developmental stage at which STAT5 is necessary, our study demonstrates that STAT5 is required for basophil development and for IL-3 to induce basophil expansion in vivo.

We propose that up-regulation of the IL-3 receptor on GMPs can serve as a mechanism for innate type cells to coordinate with adaptive immunity to generate strong immune responses. Under various disease conditions, CD4⁺ T cells have been shown to be the major cellular source of IL-3 (12, 13, 14). It is conceivable that IL-3 receptor-expressing GMPs can respond robustly to IL-3 produced by CD4⁺ T cells. As the result of the interaction between CD4⁺ T cells and myeloid progenitors, basophils can expand to critical numbers within a short period of time and these expanded basophils might then produce IL-4 to augment Th2 immune responses (4, 6, 12, 41) or to maintain B cell memory responses (8).

We thank Dr. Lothar Hennighausen of the National Institute of Health (Bethesda, Maryland) for providing us with a pair of STAT5^+/− breeder mice and Dr. Glenn Dranoff of Dana-Farber Cancer Institute (Boston, MA) and Dr. Shigeo Koyasu of Keio University School of Medicine (Tokyo, Japan) for providing us with a pair of IL-3^−/− breeder mice. We are grateful to Josh Looms and the FACS core facility for excellent cell sorting and thank Tom Startz for technique assistance and Jen Hanson for electronic microscope analysis. We thank J. D. Williams for assistance in manuscript preparation.

The authors have no financial conflict of interest.