Diverse innate stimuli activate basophils through pathways involving Syk and IκB kinases

Significance

Basophils contribute to allergic reactions when surface-bound IgE is cross-linked by allergens. How they respond to innate stimuli to produce cytokines during tissue inflammation and homeostasis is less known. Here, we comprehensively identify and compare which innate and homeostatic stimuli, among cytokines, growth factors, hormones, metabolites, and bacterial components, can activate primary basophils. We find that activated basophils display phenotypic heterogeneity at the single-cell level, suggesting that mutually exclusive functional basophil subsets exist in vivo. Furthermore, using selective kinase inhibitors we show that distinct receptor families, including IL-1R, GPCR, FcR, CD200R, and cytokine receptors, mediate basophil activation through Syk and IκB kinases. We propose these pathways have a rheostat function controlling basophil activation to diverse stimuli.

Abstract

Mature basophils play critical inflammatory roles during helminthic, autoimmune, and allergic diseases through their secretion of histamine and the type 2 cytokines interleukin 4 (IL-4) and IL-13. Basophils are activated typically by allergen-mediated IgE cross-linking but also by endogenous “innate” factors. The aim of this study was to identify the innate stimuli (cytokines, chemokines, growth factors, hormones, neuropeptides, metabolites, and bacterial products) and signaling pathways inducing primary basophil activation. Basophils from naïve mice or helminth-infected mice were cultured with up to 96 distinct stimuli and their influence on basophil survival, activation, degranulation, and IL-4 or IL-13 expression were investigated. Activated basophils show a heterogeneous phenotype and segregate into distinct subsets expressing IL-4, IL-13, activation, or degranulation markers. We find that several innate stimuli including epithelial derived inflammatory cytokines (IL-33, IL-18, TSLP, and GM-CSF), growth factors (IL-3, IL-7, TGFβ, and VEGF), eicosanoids, metabolites, TLR ligands, and type I IFN exert significant direct effects on basophils. Basophil activation mediated by distinct upstream signaling pathways is always sensitive to Syk and IκB kinases-specific inhibitors but not necessarily to NFAT, STAT5, adenylate cyclase, or c-fos/AP-1 inhibitors. Thus, basophils are activated by very diverse mediators, but their activation seem controlled by a core checkpoint involving Syk and IκB kinases.

Basophils are rare circulating granulocytes activated by immunoglobulin E (IgE)-mediated cross-linking of the high affinity receptor for IgE (FcεRI), which induces their degranulation and synthesis of the type 2 cytokines, interleukin 4 (IL-4) and IL-13 in autoimmune, allergic diseases and helminthiases (1). Basophils are also sensitive to innate signals, including the cytokine IL-3, the alarmins IL-18, IL-33, and Thymic stromal Lymphopoietin (TSLP), the prostaglandin D2 (PGD2), ATP, and various chemokines or growth factors (1–4). Some of these stimuli might regulate basophil immunoregulatory functions during homeostasis (5). Secretion of IL-4 and histamine release by human basophils is sensitive to FK506 (a calcineurin inhibitor), while secretion of IL-13 is not affected, indicating that distinct signaling pathways regulate the expression of these type 2 cytokines (6–8). By contrast, activation of basophils by IgE or IL-18/33 involves distinct receptors and upstream signaling pathways but results in expression of both IL-4 and IL-13 (5, 7, 9–11). Despite common evolutionary ancestry, it is currently thought that IL-4 and IL-13 show divergent functions in immunity and physiology (12, 13). Basophil IL-4 secretion has been implicated in T helper type 2 (Th2) differentiation and M2 skewing. Basophils are an important source of IL-13 in the lungs, where they control the phenotype of alveolar macrophages during development (1, 5).

The Il4/13 locus is subject to a differential regulation between distinct cell types but has been mostly studied in T cells. The locus contains elements regulated by cAMP, c-fos/AP-1, NFAT, NF-κB, GATA3, STAT5, and STAT6 that have been associated with the regulation of Il4/13 expression (6, 14–19). Basophils produce IL-4 and IL-13 after the cross-linking of their surface-bound IgE by antigen or by exposure to IL-3. IgE- or IL-3-mediated basophil activation are both controlled by the tyrosine kinase Syk (6, 20, 21). IgE-induced cytokine expression is also promoted by IκB kinase (IKK) through NF-κB-dependent or -independent signals (22, 23).

To date, studies of murine basophils have focused on cultured bone marrow-derived basophils (BMBa) or primary bone marrow (BM) basophils, with circulating mature basophils too few in number for useful studies (6). We used B8 × 4C13R IL-4/IL-13 triple reporter mice (24) to follow and compare the ex vivo responses of primary mature basophils to 96 common stimuli and found that basophils were highly sensitive to type 2 and epithelial-derived cytokines and growth factors. Single-cell analysis revealed that basophils displayed distinct phenotypes upon stimulation with IL-3, expressing IL-4, and/or IL-13, Ly6C, or the degranulation marker CD63 (25). During helminth infection, basophils were found to be hyper- and hyporeactive to distinct stimuli (“Hp-basophils”). Purified Hp-basophils were also sensitive to homeostatic growth factors and antiviral response elements. Basophil reactivity was always sensitive to Syk and IKKs specific inhibitors. In conclusion, we found basophils to be sensitive to a complex array of distinct innate stimuli that worked through core common signaling pathways.

Results

Activation Induces Distinct Basophil Phenotypes.

Basophils are present at very low levels in vivo. Previous studies of murine basophils centered on IL-3-cultured BMBa, but they present an activated phenotype (6, 8, 26). Indeed, using B8 × 4C13R reporter mice we identified that primary spleen or BM basophils of naïve mice do not express high levels of IL-4/IL-13 as BMBa (Fig. 1A). Given that basophils isolated from the spleen have similar gene transcription signatures to peripheral blood circulating mature basophils (27), we used splenocytes as a convenient source to study primary mature basophil activation ex vivo.

Fig. 1.

Basophils with distinct phenotype are induced by IL-3. (A) Dot plots depicting the expression of AmCyan (IL-4) and dsRed (IL-13) from B8 × 4C13R eYFP+, BM-derived, or primary basophils. (B) Overlay of contours plots displaying the phenotype of eYFP+ spleen basophils after 24 h stimulation ex vivo with 1 ng/mL IL-3. (C) tSNE analysis of concatenated 1,000 unstimulated and 1,500 IL-3 stimulated eYFP+ basophils (n = 4) as in B, represented as contour plots or heatmap dot plots. This is representative of at least three independent experiments. For D, as in B, contour plots depict the phenotype of basophils after stimulation with different compounds.

Activation of naïve basophils with IL-3 dose dependently increased survival, expression of the cytokines IL-4 and IL-13, the degranulation marker CD63, the activation markers CD200R, CD9, CD44, CD11b, and Ly6C, and decreased CD62L expression (SI Appendix, Fig. S1A). Activation also induced an heterogenous single cell phenotype. Most IL-13- or CD63-expressing basophils also expressed IL-4. IL-13+ basophils did not express CD63 and vice-versa (SI Appendix, Fig. S1B and Fig. 1B). Furthermore, IL-3 induced a time-dependent increase in IL-4/13 double positive cells, in contrast with CD63 expression, which was transient (SI Appendix, Fig. S1C). Basophil expression of Ly6C was independent on IL-4, IL-13, or CD63. Other activation markers (CD9, CD11b, and CD200R) were associated with the CD62L^lo phenotype and CD63 expression (Fig. 1B and SI Appendix, Fig. S2). Indeed, unbiased neighboring clustering analysis identified six coexisting activation patterns: CD63+ IL-4+, Ly6C+ IL-4+, Ly6C+ IL-4−, IL-4+ IL-13+ (Ly6C/CD63− to +), and IL-13+ IL-4− Ly6C− CD63− (Fig. 1C) and two unstimulated phenotypes (CD62L^hi or ^lo). Basophils showing evidence of exocytosis (CD63+) all expressed IL-4, activation markers such as CD200R and low levels of CD62L but not necessarily IL-13 or Ly6C. Most IL-13+ basophils expressed IL-4 but were not associated with CD63, Ly6C, CD62L, CD200R, CD9, or CD11b. Ly6C+ basophils were not associated with any other activation markers analyzed.

Interestingly, the basophil activating agents Leukotriene B4 (LTB4), PGD2, and TSLP induced only IL-4 expression; IL-33, Ba13 (αCD200R3), MARI (αFcεRIα) also induced IL-13 expression. IL-13 was not exclusively associated with IL-4, and IL-4+, IL-13+, and IL-4/IL-13+ distinct subpopulations were induced by these agents. Our results demonstrate that upon activation with various stimuli, basophils appear with distinct phenotypes related to exocytosis (CD63), cytokine expression (IL-4 and IL-13), or unknown processes (Ly6C). This supports the existence of a basophil functional heterogeneity, as identified recently in the human blood (28).

Diverse Stimuli Activate Naïve Primary Mature Basophils.

We followed the expression of type 2 cytokines by basophils, CD4+ T cells and ILC2s in the skin by flow cytometry using an established gating strategy in the B8 × 4C13R mice (24, 29). Four distinct skin inflammation models were compared, including irritant contact dermatitis (30), helminth-induced skin allergic inflammation (31, 32), atopic dermatitis (24, 33), and ear punch wound healing (34). In each model the kinetics of IL-4 and IL-13 expression by basophils, CD4+ T cells or ILC2s varied considerably, and basophils were found to preferentially express IL-4 over IL-13. For example, CD4+ T cells expressed IL-13 during irritant contact dermatitis, while basophils expressed IL-4, and the expression of type 2 cytokines by CD4+ T cells or ILC2s was not significant in a skin wound healing model, while basophils expressed IL-4 for 3 wk (SI Appendix, Fig. S3). Notwithstanding the caveat that these experiments only follow variations in the transcriptional regulation of Il4 and Il13, this confirms that basophils IL-4 and IL-13 expression is regulated by different stimuli and signaling pathways than CD4+ T cells or ILC2s in vivo, as previously described (35). We therefore sought to identify which stimuli are able to regulate basophils Il4 and Il13 transcriptional expression. Primary naïve B8 × 4C13R splenocytes were stimulated with 96 common inflammatory or homeostatic stimuli (SI Appendix, Table S1), and MCPT8+ basophils were analyzed by flow cytometry in terms of IL-4 and IL-13 expression or activation (CD200R) (24, 36). Individual results are represented in descending order in Fig. 2 A–C. We pooled all results excluding exogenous stimuli (papain, PMA, and ionomycin) after correcting for multiple comparisons using the Holm–Sidak method (Fig. 2 D–F). We also did analyze basophil survival and exocytosis in these conditions (SI Appendix, Fig. S4). The effects of each stimulus are summarized in SI Appendix, Table S2. Few stimuli induced the expression of both IL-4 and IL-13 (ionomycin > IL-3 > IL-13 > ⍺CD200R3 > IL-33 > ⍺FcεRI⍺ > papain > ⍺CD16/32 > ⍺CD49d > TGFβ1 > ⍺CD62L), but numerous distinct stimuli induced only IL-4 (TSLP> PGD2 > ATP > IL-7 > adenosine > LTB4 > ⍺CD11b > C5a > ⍺IgE > butyrate > PGE2 > oxytocin > Substance P > Relm⍺ > ⍺CD45 > ⍺IgG1 > glutamate > GABA > NGF). Strikingly, PMA > IL-10 > SCF induced a significant expression of IL-13, without IL-4 expression. The expression of the basophil activation marker CD200R was not correlated with IL-4 or IL13 expression and was induced by numerous stimuli (⍺CD200R3 > IL-3 > IL-33 > ⍺FcεRI⍺ > IL-13 > TSLP > IL-7 > ⍺IgD > heparin > ⍺CD44 > ⍺CD45 > ⍺ASGM1 > amphiregulin > ⍺IgE > PGD2) and inhibited by others (IL-12 > IFNɣ > IL-2 > TGFβ1 > fMLP).

Fig. 2.

Basophils are heterogeneously activated by diverse stimuli. Whole splenocytes from naïve B8 × 4C13R mice have been stimulated ex vivo with the depicted stimuli and eYFP+ basophils activation was analyzed by flow cytometry. Results from each stimulation have been normalized to the same unstimulated splenocytes. Each dot represents a different mouse, and results represent the pool of at least three independent experiments for AmCyan [IL-4 (A)], DsRed [IL-13 (B)], or CD200R (C). (C–E) Volcano plots show the induction of AmCyan (IL-4), dsRed (IL-13), or CD200R as in A–C. Statistics are one sample t test against zero (A and B) or one (C) or unpaired t test after correction for multiple comparison using the Holm–Sidak method (n = 6 to 59, SI Appendix, Tables S1 and S2). (A–C) Results with P > 0.01 were considered nonsignificant and are not represented. **P < 0.01; ***P < 0.001; ****P = 0.0001. (D and E) Dashed lines represent an adjusted P < 0.05 in logarithmic scale or fold changes of ±3% (D and E) or 10% (F).

In summary, stimuli-activating basophils were associated with type 2 or epithelial inflammation (IL-3/18/33, TSLP, and PGD2), while signals associated with type 1 inflammation were associated with decreasing basophil survival or activation. Homeostatic neuromodulators or hormones (oxytocin, Substance P, GABA, and glutamate), metabolites (ATP, adenosine, and butyrate), and growth factors (IL-7, NGF, VEGF, SCF, and TGFβ1) induced a low-level significant activation of basophils in naïve conditions.

Naïve Basophil Survival and Activation Are Sensitive to Syk, IKK, and Calcineurin Inhibitors.

We explored which signaling pathways controlled naïve basophils activation ex vivo by using selective inhibitors of the main myeloid cells activating signaling pathways: the calcineurin and NFAT pathway (FK506, “Tacrolimus”) (7), the Syk kinase pathway (R406) (37), and the IKKs controlling the NF-κB pathway (BMS345541, “BMS”) (23, 38). IKK, Syk, and calcineurin inhibitors did not induce a significant cell death in whole splenocytes (BMS induced a small decrease of CD4+ T cells). However, they both induced a strong decrease in naïve basophils viability (Fig. 3A), which could be due to an interference with basophils prosurvival pathways. To test this hypothesis, we stimulated whole splenocytes with prosurvival stimuli in the presence of these inhibitors. The inhibitory effects of FK506 were overcome by IL-3, TSLP, IL-33, or MARI. The inhibitory effects of the BMS were overcome by IL-3, MARI, or the ⍺CD200R3 antibody Ba13 (39). The inhibitory effects of the Syk inhibitor R406 could be overcome by IL-3. PGD2 did not show any prosurvival effect and did not overcome any inhibitor effect (Fig. 3B). These results suggest that the survival of naïve basophils is dependent on multiple pathways relying on IKK, the calcineurin, or Syk. Interestingly, IL-3–induced human basophil survival has been described as independent on Syk, dependent on PI3K and Pim1, and sensitive to Ly294002 (21, 40). However, we found that Ly294002 did not inhibit the survival of naïve mouse basophils in the presence of IL-3, at concentrations that inhibited IL-3–induced IL-4 and IL-13 expressions (SI Appendix, Fig. S5).

Fig. 3.

Naïve basophils survival and phenotype is altered by inhibitors of the IkB kinase, calcineurin, and Syk pathways. Whole splenocytes from B8 × 4C13R mice were stimulated ex vivo with indicated compounds in the presence of calcineurin (FK506), Syk (R406), or IkB kinases (BMS) selective inhibitors and their phenotype was analyzed by flow cytometry (n = 6). (A) represents the proportion of DAPI+ dead cells, eYFP+ basophils, or CD4+ cells among DAPI-living cells in the presence of the various inhibitors. (B) The effects of the various inhibitors on basophils survival were analyzed in the presence of various stimuli, as depicted. (C–F) The expression of CD62L (C), CD200R3 (D), IL-4 as AmCyan (E), and IL-13 as dsRed (F) were analyzed on eYFP+ basophils stimulated ex vivo with inhibitors as in A. Results are from two to three pooled independent experiments (n = 6 to 9). Statistics above histograms compare the effects of the inhibitors to the DMSO control and statistics between histograms compare the relevant groups by a one-way ANOVA with a Sidak’s correction. Not depicted: not significant; ^#P < 0.06; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

In addition to affecting basophils’ survival, these inhibitors also altered the phenotype of primary naïve basophils: BMS down-regulated CD62L (Fig. 3C) and all inhibitors down-regulated CD200R3 expression (Fig. 3D). Naïve mature basophils constitutively express low levels of IL-4 transcripts (41), and the IKK inhibitor BMS appeared to down-regulate this constitutive expression (Fig. 3 E and F). Overall, the phenotype of naïve basophils seemed constitutively dependent on IKK activity (CD62L and IL-4 expression), but their survival and expression of CD200R3 were also redundantly dependent on the Syk and NFAT pathways.

Naïve Basophil Activation Is Sensitive to IKK or Syk Inhibitors.

Naïve basophils activated with IL-3, TSLP, ⍺CD200R3, and PGD2 down-regulated their CD62L expression (Figs. 1 B and C and 4A). IL-3– and TSLP-induced CD62L down-regulation depended on Syk activity, as found in neutrophils (42), but was not inhibited by FK506 (Fig. 4B). As the IKK inhibitor down-regulated CD62L expression on unstimulated basophils (Fig. 3C), it could not be analyzed. IL-3, TSLP, MARI, and Ba13-induced IL-4 expression was sensitive to BMS and R406 but not to FK506. IL-33 and PGD2-induced IL-4 expression was only sensitive to BMS (Fig. 4 C and D).

Fig. 4.

Naïve basophils activation by diverse stimuli is sensitive to Syk and IKK inhibitors. Whole splenocytes from B8 × 4C13R mice were stimulated ex vivo with indicated compounds or in the presence of calcineurin (FK506), Syk (R406), or IkB kinases (BMS) selective inhibitors and their phenotype were analyzed by flow cytometry. The expression of (A and B) CD62L, (C and D) AmCyan (IL-4), (E and F) dsRed (IL-13), (G and H) and the positivity for CD63 were quantified by flow cytometry on eYFP+ basophils. (D, F, and H) Data has been normalized on stimulated uninhibited paired values. Results are from two to three pooled independent experiments. For A and B, n = 6. For C–H, n = 9. Statistics are (A, C, E, and G) unpaired t tests against unstimulated values reporting the P value or (B) a one-way or (D, F, and H) two-way ANOVA with a Dunnett’s correction reporting adjusted P values. Not depicted: not significant; ^#P < 0.1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

The expression of IL-13 was only induced by IL-3, IL-33, MARI, and Ba13. R406, but not FK506, inhibited the expression of IL-13. IL-33–induced IL-13 production was inhibited by IKK inhibitors (Fig. 4 E and F). The expression of CD63 was induced by IL-3, TSLP, IL-33, MARI, and Ba13. It was sensitive to both Syk and IKK inhibitors but not to FK506 (Fig. 4 G and H). Only modest effects were observed for Ba13-induced basophil CD63 expression. BMS inhibits IKK2 at 0.4 µM but also IKK1 at 4 µM (38); however, only 5 µM inhibited basophils (SI Appendix, Fig. S6), suggesting a major role of IKK1 over IKK2 in basophils. Overall, naïve basophil activation was predominantly sensitive to Syk and IKK inhibitors.

Helminth Infection Modifies Basophil Reactivity.

Basophils are systemically expanded and activated during helminth infections with descriptions of them being either hyper- or hyporeactive (43–48). We analyzed basophils from Heligmosigmoides polygyrus (Hp)-infected B8 × 4C13R mice (“Hp-basophils”) and found a hyperactivated phenotype in vivo with distinct subsets showing a high expression of IL-4, Ly6C, CD41, CD49b, or ST2 (the IL-33 receptor) (SI Appendix, Fig. S7). IL-13 or the activation markers CD200R or CD9 were not overexpressed. Furthermore, they presented with a hypoactivated phenotype, showing less expression of the degranulation marker CD63, the activating receptors FcεRIα and CD200R3, and the integrin CD11b. Importantly, Hp infection–induced CD11b down-regulation and basophilia were dependent on TSLPR (SI Appendix, Fig. S7). Thus, basophils from Hp-infected mice present both a hypo- and hyperactivated phenotype, slightly dependent on TSLP, as observed in Trichinella spiralis infection (49). As we previously demonstrated that Hp induced basophil activation was dependent on IL-3 and antibody production, it is reasonable to conclude that both IL-3, TSLP, and Fc receptor signaling are involved in Hp infection–induced basophil phenotype (46, 47).

When basophils from Hp-infected mice were compared to naïve basophils ex vivo they showed enhanced expression of IL-4, IL-13, and CD63 and presented as distinct IL-4+, IL-4+ IL-13+, and IL-4+ CD63+ IL-13− subpopulations in the absence of any specific stimulation (Fig. 5A). Also, in contrast to naïve basophils, the survival of Hp-basophils was not increased by IL-3, IL-7, TSLP, or MARI (Fig. 5B). Similarly, Hp-basophils did not express more IL-4 in response to IL-3, IL-33, PGD2, LTB4, or oxytocin. However, they were more sensitive than naïve basophils to CD200R3 or FcεRI⍺ cross-linking and to butyrate, with respect to their expression of IL-4. Importantly, both naïve or Hp-basophils were similarly responsive to ATP, C5a, IL-7, and TSLP (Fig. 5 C–E). Hp-basophils were more prone to express IL-13 and more sensitive to ATP, Ba13 (⍺CD200R3), butyrate, and MARI (⍺FcεRI⍺) to do so (Fig. 5F). Hp-basophils lost their ability to produce IL-4, but still expressed IL-13 and CD63 in an unaltered fashion, in response to IL-3 and IL-33 (Fig. 5G). Ba13 or MARI stimulated a more profound degranulation profile (CD63) in Hp-basophils compared to basophils from naïve mice (Fig. 5G).

Fig. 5.

Helminth infection shapes basophils reactivity to diverse stimuli. Whole splenocytes from B8 × 4C13R mice infected with Hp or not (Naïve) were stimulated with various compounds ex vivo and eYFP+ basophils analyzed by flow cytometry. (A) Contour plots and histogram depicting differences between unstimulated basophils from both conditions. B shows differences in the proportion of basophils among living cells. Statistics represents the significance of each stimulation against the unstimulated condition from the same group. (C) Contour plots depicting differences between MARI stimulated basophils from both conditions. (D–G) Each stimulated condition was normalized to its unstimulated control (n = 6). Statistics above bars represent a one-sample t test against zero. Statistics between bars compares an unpaired t test Naïve versus Hp stimulated in the same way. ns: nonsignificant; ^#P = 0.06; *P < 0.05; **P < 0.01; ***P < 0.001; ****P = 0.0001. Results are from two independent experiments.

Taken together, our analyses reveal that helminth infection biases basophils to express IL-13 and become hyperreactive to FcεRI⍺ and CD200R3 cross-linking. However, helminth infection causes basophils to become hyporeactive to other signals, including IL-3, IL-33, and eicosanoids, indicating that basophil functional responses are profoundly altered by helminth infection.

Diverse Innate Stimuli Activate Purified Basophils.

We wished to analyze the functional responses of purified mature primary basophils in a standardized assay. Because naïve mice do not have sufficient numbers of circulating mature basophils amenable for such studies, we used basophilic Hp-infected mice spleen as a source of basophils. As mature basophils quickly undergo apoptosis, we used the two main types of basophil “survival conditions” to keep them responsive ex vivo: IL-3 (10 pg/mL) or TSLP (1 ng/mL). We compared the direct effects of 75 innate stimuli (SI Appendix, Table S1) of these purified basophils in these two conditions (SI Appendix, Fig. S8 B–E). IL-3 or TSLP conditioning did induce a different activation of basophils only in the presence of IL-4, IL-18, IL-33, the TLR7 ligand imiquimod, or CXCL6. As previously shown, IL-3 conditioning was more associated with degranulation than TSLP conditioning (47). The remaining 70 stimuli induced a similar phenotype in both conditions (SI Appendix, Fig. S8 B–E).

The raw reactivity of purified conditioned basophils is depicted in Fig. 6 A–C and represented as volcano plots after correction for multiple comparisons using the Holm–Sidak method (Fig. 6 D–F). We observed that IL-3, IL-18, IL-33, IL-7, TGFβ1, VEGF, TSLP, GM-CSF, IL-13, histamine, ATP, IFN-I (IFNβ and IFNα4), and TLR ligands (Pam3Cys, imiquimod, CpG, lipopolysaccharides [LPS], and FSL1) can directly induce basophil activation, IL-4 or IL-13 expression (Fig. 6 A–C and D–F), irrespective of the conditions used. VEGF and TGFβ1 could directly induce basophils IL-4 expression (Fig. 6 A and D), while Pam3cys and GM-CSF induced IL-13 expression (Fig. 6 B and E). Surprisingly, prostaglandins (PGD2, PGE2, and BW245c, a DP1-specific agonist), IL-2 and CCL17 directly inhibited conditioned basophils IL-4 expression (Fig. 6 A and D). Basophil activation could be induced by various TLR ligands (Pam3cys, imiquimod, CpG, and LPS), IFN type I (IFNβ and IFNα4), and by histamine, IL-13, and ATP (Fig. 6 C and F). The full detail of the effects of each stimulus is represented in SI Appendix, Table S3.

Fig. 6.

The reactivity of purified basophils. A total of 2,000 eYFP+ FACS sorted basophils from the spleens of Hp-infected B8 × 4C13R mice were conditioned ex vivo with 10 pg/mL IL-3 or 1 ng/mL TSLP and stimulated with one of 75 innate stimuli for 24 h (n = 6). Basophils expression of IL-4—AmCyan, IL-13—dsRed, or CD200R is represented (A–F). Data has been normalized on paired unstimulated conditioned control by ratio (A, C, D, and F) or subtraction (B and E). Results are from two pooled independent experiments. Statistics are one sample t test against zero (A–C) or (D–F) unpaired t tests against the control value with a Holm–Sidak correction for multiple comparisons. Horizontal dotted lines represent the significance threshold (P < 0.05) and vertical dotted lines ±5% of variation. Adjusted P value is represented in logarithmic scale. Principal component analysis of IL-4, IL-13, CD200R, and CD63 variability in the phenotype of (G) naïve but not purified basophils or (H) conditioned and purified Hp basophils. (A–C) Results with P > 0.01 were considered nonsignificant and are not represented. **P < 0.01; ***P < 0.001; ****P = 0.0001. Blue and red represent TSLP and IL-3 conditions, respectively.

We used a dimensionality reduction approach by principal component analysis in order to visualize which of the 75 innate stimuli did induce a similar or distinct activation of basophils. In these analyses, IL-4, IL-13, CD200R, and CD63 contributed equally to the main variability of the datasets (PC1). In naïve conditions (dataset from Fig. 2), CD200R > CD63 and IL-13 > IL-4 showed a divergent contribution to PC2 (Fig. 6G), while in purified conditions CD63 was the main positive contributor to PC2, against IL-4 and IL-13 (Fig. 6H). In both datasets, IL-3, IL-1R ligands (IL-18 and IL-33) and IL-7R ligands (IL-7 and TSLP) clustered in distinct areas, supporting the notion that each receptor mediates a distinct activation of basophils. Purified Hp-basophil responses to antiviral response elements including IFN-I (IFNα4 and IFNβ) and intracellular TLR (imiquimod and CpG) clustered together, while TLR2-induced activation (Pam3Cys) was close to TGFβ1. The growth factors TSLP, IL-7, GM-CSF, and VEGF clustered with histamine. Finally, PGD2, PGE2, and BW245c clustered close to IFNγ. This suggests eicosanoids can inhibit Hp-basophils activation in a similar way as Th1 cytokines (Fig. 6H). The fact that similar stimuli clustered closely, and that distinct stimuli were segregated (i.e., IL-3, TSLP, and IL-33), validates the measurements of basophil responsiveness in naïve whole splenocytes and purified Hp-basophils datasets.

Overall, basophils can be activated directly by epithelial-derived signals (IL-18, IL-33, TSLP, and GM-CSF), inflammatory signals (IL-3, CCL17, IL-13, histamine, IFN-I, TLR ligands, and ATP), and homeostatic signals (IL-7, VEGF, TGFβ1, and serotonin). They show distinct patterns of activation in response to dissimilar stimuli.

Responses of Basophils to Innate Stimuli Are Sensitive to IKK and Syk Inhibitors.

We evaluated the responses of purified Hp-induced basophils to identify activating agents in the presence or absence of selective inhibitors of the adenylate cyclase and cAMP pathway (2′,5′dideoxyAdenosine, “ddAde”) (50), the calcineurin and NFAT pathway (FK506, “Tacrolimus”) (7), Syk kinase activity (R406) (37), IKKs activity (BMS) (23, 38), STAT5 phosphorylation (Stat5i) (51), or c-fos/AP-1 binding (T-5224) (52). A “physiological” level of 10 pg/mL of IL-3 was included as a survival factor in these experiments. Hp-basophils stimulated with IL-3, TSLP, IL-33, MARI, or PMA all responded with strong expression of IL-4, IL-13, and CD200R, with the exception of PMA, which reduced the expression of IL-4 (Fig. 7A) (7).

Fig. 7.

Purified basophils activation is controlled mainly by IKK and Syk inhibitors. A total of 2,000 eYFP+ basophils were FACS sorted from the spleens of Hp-infected B8 × 4C13R mice and stimulated ex vivo with IL-3 (A and B), TSLP (A and C), IL-33 (A and D), MARI (A and E), or PMA (A and F) in the presence of selective inhibitors of adenylate cyclase (ddAde), calcineurin (FK506), Syk (R406), IkB kinases (BMS), Stat5 (Stat5i), or the c-fos/AP-1 pathway (T-5224) always in the presence of 10 pg/mL of IL-3 as a survival factor. Each data has been normalized on its unstimulated control (A) or on its stimulated uninhibited control (B–F). Results are pooled from two to three independent experiments for each condition (n = 6 to 9). Statistics represent the significance of a one sample t test against zero. ns: not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

IL-3 induces the phosphorylation of STAT5 and Syk (6, 21, 53). Basophils stimulated by IL-3 were sensitive to inhibitors of adenylate cyclase, Syk, IKK, and STAT5 but not to FK506 or T-5224. In contrast, IL-13 and CD200R expression were more dependent on Syk and IKK compared to the other pathways (Fig. 7B).

TSLP is reported to induce the phosphorylation of STAT5 (6). Here, it induced the expression of IL-4, IL-13, and CD200R by purified Hp-basophils. IL-4 expression was sensitive to adenylate cyclase, calcineurin, Syk, IKK, and Stat5 inhibitors, while IL-13 and CD200R expression were only inhibited by Syk, IKK, and STAT5 inhibitors (Fig. 7C).

IL-33 is reported to induce the expression of IL-4/IL-13 via a MyD88 and NF-κB pathway (9, 11, 26). Consistent with this, we observed that IL-33–induced basophil activation was sensitive to the IKK inhibitor BMS. IL-4 and IL-13 expression were also sensitive to Syk and STAT5 inhibitors (Fig. 7D). We confirmed that cross-linking of FcεRI⍺ by MARI induced a Syk- and/or IKK-dependent activation (6, 9, 54), while STAT5 was needed for the expression of IL-4 and IL-13 (Fig. 7E). PMA stimulates basophils through PKC activation, and on purified basophils, both IL-4 and IL-13 expression was dependent on the IKK and Syk pathways. Adenylate cyclase, calcineurin, IKK, and STAT5 inhibitors further down-regulated the expression of IL-4 in the presence of PMA (Fig. 7F).

Any interpretation of our results must consider the caveat that the various inhibitors could have blocked the prosurvival effects of added or autocrine IL-3. Indeed, activated basophils are a physiologically relevant source of autocrine IL-3, especially after IgE cross-linking and in vivo, the contribution of which has not been controlled for in this study (55, 56). Taken together, basophil IL-4 and IL-13 expression was sensitive to IKK, Syk, and STAT5 inhibitors (with the exception of PMA). The activation of basophils as measured by CD200R expression was sensitive to IKK and Syk inhibitors (with the exception of IL-33–induced activation). FK506 did show either moderate or no effect on basophils activation (7). IKK and Syk appear to be critical for basophil activation and IL-4 and IL-13 expression in response to diverse stimuli.

Our studies reveal that murine basophil activation can be heterogeneous and involves the development of phenotypically and functionally distinct basophil subtypes, as identified very recently in human blood (28). We found basophils to be sensitive to epithelial and type 2 inflammatory stimuli but also to some homeostatic growth factors, metabolites, or neuromediators. The activation of both naïve or helminth-induced basophils by these diverse stimuli was sensitive to Syk and IKK inhibitors. This strongly suggests that Syk and IKK are core signaling elements controlling basophil type 2 cytokine expression, exocytosis, and survival.

Discussion

Basophils are poised to express and secrete the type 2 cytokines IL-4 and IL-13 upon activation. Transcription of Il4 and Il13 genes in basophils appears to be distinct from that observed in other immune cell lineages (17, 57). We confirmed this by comparing four preclinical models of type 2 skin inflammation and resolution for expression of Il4/Il13 by CD4+ T cells, ILC2s, and basophils (SI Appendix, Fig. S3). We further sought to identify the responses of primary mature basophils to pathophysiological and physiological stimuli, with a particular focus on their effects on Il4 and Il13. We first established that basophils express Il4, Il13, CD63, Ly6C, or CD62L in distinct patterns upon activation ex vivo. This supports the existence of basophils with different functional phenotypes as observed in humans (28). Particularly, a dichotomy between IL-4 and IL-13 was observed in primary basophils stimulated with IL-3, IL-33, IL-18, papain, MARI, anti-IgE, or Ba13, in naïve and/or in Hp infection conditions. Thus, diverse stimuli can lead to a heterogeneous functional phenotype of MCPT8+ basophils.

As some stimuli induce only an expression of IL-4 (i.e., PGD2) or IL-13 (i.e., PMA), distinct signaling pathways seem to regulate basophil IL-4 or IL-13 expression (Figs. 1 and 2). As previously reported (7), we confirmed that FK506 inhibited basophils expression of IL-4 but not of IL-13 (Figs. 4 and 7). Basophils were more prone to secrete IL-4 than IL-13 in vivo during skin inflammation (SI Appendix, Fig. S3), Hp infection (SI Appendix, Fig. S5), and ex vivo in naïve conditions (Fig. 2). IL-13 was induced at high levels on unstimulated and stimulated Hp-basophils ex vivo (Fig. 5) and by stimulation of naïve basophils with high amounts of IL-3 (SI Appendix, Fig. S1 A and B). Of note, IL-4 is known to be secreted faster than IL-13 by basophils (7), and upon stimulation with IL-3, IL-13 increased for at least 3 d (SI Appendix, Fig. S1C). This data would suggest that basophil IL-13 expression in vivo may require a specific and potent chronic stimulation. We quantified CD63 24 h after stimulation ex vivo, whereas it is transiently expressed within minutes during anaphylactic degranulation (58). Thus, our measurements of CD63 are more compatible with slower exocytosis processes (25). Interestingly, expression of IL-4, but not IL-13, was associated with CD63 expression upon stimulation of naïve or Hp-basophils (Figs. 1 and 5), which supports previous findings (7), and the existence of basophils with a heterogeneous functional phenotype (28).

We used purified basophils to distinguish which stimuli have direct or indirect effects. To overcome the difficulties of yield, we used flow cytometry to purify primary mature basophils from basophilic Hp-infected mice and used prosurvival factors to ensure basophil viability ex vivo. We observed distinct basophils activation states in PCAs in both naïve and purified conditions after stimulation with the most potent distinct stimuli (IL-3, IL-33, and TSLP), which validates our methodology to investigate basophils responsiveness (Fig. 6 G and H). Purified basophils viability was ensured by the presence of a minimal amount of IL-3 in inhibition studies. A caveat is that the effect of the inhibitors could have been due partly to a blockade of IL-3 conditioning (Fig. 7). However, basophil activation can be partly dependent on an autocrine secretion of IL-3, especially in vivo during allergic inflammation (56). Furthermore, signaling inhibitors strongly inhibited the activation of both unconditioned naïve basophils (Figs. 3 and 4) and conditioned purified basophils from Hp-infected mice, supporting the notion that the inhibitory effects we observed transcended exogenous or autocrine low-level IL-3 signaling.

STAT5, Syk, and IKK were the main signaling components controlling basophils activation observed in this study. STAT5 is recruited and phosphorylated to the IL-7R and IL-3R α chains upon stimulation, with IL-7/TSLP, IL-3, or IgE, to induce type 2 cytokine production (6, 8, 56, 59). It is also important to control ST2 expression and thus the response to IL-33 (60). Here, STAT5 inhibition controlled IL-4 expression but showed only modest effects on IL-13 or CD200R (Fig. 7). Syk and IKK inhibitors were more efficient to block basophils activation induced by all stimuli tested. Syk is a tyrosine kinase activated by the FcRγ chain [when induced by IgE or IL-3 (6, 20, 21)], DAP12 (upon CD200R3 cross-linking) (39, 61), cytokines (IL-33, M-CSF, and GM-CSF), and multiple other signals (20, 60, 62). Syk promotes activation and survival in various cell types, and its expression is heavily regulated in basophils (63). Importantly, Syk induces the phosphorylation of IKK (64). IKK are kinases which phosphorylate IκB molecules to allow NF-κB translocation. Numerous stimuli can lead to IKK and NF-κB activation, including IL-33, IgE, cytokines, hormones, and physiological mediators (23). NF-κB promotes cell survival and cytokine secretion (23, 65). Here, IKKs were central to control basophils survival and activation by multiple stimuli. As IKKs also induce NF-κB-independent effects, the effects we observed are not necessarily NF-κB dependent (66), but our study supports a central role for NF-κB in basophils survival and activation, as described for Th2 cells and most innate cells (23, 65).Thus, basophils survival, exocytosis, general activation, and IL-4 and IL-13 expression seem to be all controlled by core signaling elements. We propose that Syk and/or IKKs are central rheostats mediating basophil activation and survival induced by various homeostatic, innate, and/or adaptive stimuli (SI Appendix, Fig. S9).

Materials and Methods

Mice and Treatments.

All mice were bred and housed in specific pathogen-free conditions at the Malaghan Institute of Medical Research Biomedical Research Unit. All experimental protocols were approved by the Victoria University of Wellington Animal Ethics Committee (Permit 23910). A total of 200 Hp bakeri were gavaged and mice analyzed at the peak of basophilia 15 d after infection (48).

Flow Cytometry and Stimulations.

Whole splenocytes single cell suspensions were blocked for 15 min at 4 °C and stained in the FACS buffer for 20 min at 4 °C with an optimized concentration of fluorophore-conjugated antibodies. Flow cytometry was performed on an Aurora cytometer (three lasers: 405, 488, and 640 nm, Cytek), and sorting on a BD Influx (Becton Dickinson, analyses were conducted using FlowJo 10 [TreeStar]). The 2 × 10⁶ whole splenocytes or 5,000 FAC sorted basophils were stimulated overnight for 20 h (SI Appendix, Table S1).

Statistical Analysis.

Statistical analyses were performed using Prism 8 and 9.0 (GraphPad).

Data Availability

All study data are included in the article and/or SI Appendix.