Upregulated Expression of the IL-9 Receptor on TRAF3-Deficient B Lymphocytes Confers Ig Isotype Switching Responsiveness to IL-9 in the Presence of Antigen Receptor Engagement and IL-4

Interleukin-9 is a γ-chain family cytokine that has pleiotropic functions in the immune system (1–4). IL-9 binds to a heterodimeric receptor consisting of the unique subunit IL-9R and the common subunit γ-chain shared by multiple cytokines of the γ-chain family, including IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 (1–3). Under physiological and pathological conditions, IL-9 is produced by a variety of activated immune cell subsets (1, 2). Examples are IL-9–producing CD4 Th (Th9) cells (which originally received the numerical designation from IL-9), T_FH cells, type 2 innate lymphoid cells, mast cells, invariant NKT cells, mucosal-associated invariant T cells, NK cells, macrophages, basophils, eosinophils, IL-9–producing CD8 cytotoxic T (Tc9) cells, and Th2 and Th17 cells (1, 2, 4, 5). When produced by these immune cell subsets, IL-9 plays diverse roles in the host, ranging from protective immunity to immunopathology depending on the specific setting and microenvironment (2, 4–8).

It has been demonstrated that IL-9 has important functions in protective immunity against helminthic and mycobacterial infections (4, 9–12). However, IL-9–mediated inflammation exacerbates the damaging effects of some other infectious pathogens such as the respiratory syncytial virus–induced pathology, Staphylococcus aureus–induced pneumonia, Helicobacter pylori–induced gastritis, and barrier function loss caused by fungal infections (13–18). Furthermore, increasing evidence obtained from both human patients and mouse models indicates that IL-9 plays multiple pathogenic roles in autoimmune disorders (such as systemic lupus erythematosus, rheumatoid arthritis, multiple sclerosis, systemic sclerosis, and psoriasis), allergic diseases (such as asthma, dust mite allergy, and food allergy), inflammatory diseases (such as ulcerative colitis, atherosclerosis, vasculitis, hepatic and pulmonary fibrosis), and cancers (such as lymphomas and leukemias) (4, 5, 8, 19–29). In contrast, IL-9 and Th9/Tc9 exhibit potent antitumor properties in solid tumors (such as melanoma, lung cancer, breast cancer, and colon cancer), which has recently attracted extensive interest in this cytokine (2, 5, 30–35).

The diverse roles of IL-9 in protective immunity, immunopathology, and antitumor immunity are realized through the activities of its various target cells expressing IL-9R constitutively or inducibly in specific stimulatory settings and microenvironments (4, 5, 20, 26). A wide variety of immune and nonimmune cell types can respond to IL-9 under different physiological and pathological conditions. These include B lymphocytes, T lymphocytes, innate lymphoid cells, mast cells, polymorphonuclear cells, dendritic cells, platelets, hematopoietic progenitor cells, epithelial cells, vascular smooth muscle cells, keratinocytes, hepatic stellate cells, astrocytes, oligodendrocytes, and microglia (8, 19, 20, 22, 23, 26, 36, 37). By inducing the JAK-STAT pathway, IL-9 can regulate the survival, proliferation, differentiation, migration, activation, or effector function of different target cells under specific conditions (8, 19, 20, 22, 26, 36, 37).

Among the target cells of IL-9, B lymphocytes are the central player of humoral immunity and the only cell type in mammals that can produce Abs against various pathogens (38–40). IL-9 was initially reported to potentiate IL-4–induced production of IgE and IgG, but not IgM, in crude B lymphocytes prepared from human blood, human tonsils, and mouse spleens (41–44). Subsequent evidence revealed that IL-9R is not expressed on naive follicular (FO) B cells, although it is detected on B1 B cells, marginal zone (MZ) B cells, and some germinal center (GC) B cells (44–47). More recently, it has been shown that IL-9R is selectively induced on GC-derived memory precursor cells and memory B cells (6, 45). IL-9 accelerates the exit of memory precursor cells from the GC and induces the development, proliferation, and differentiation of memory B cells (6, 45). Interestingly, Il9r^−/− mice exhibit normal T-dependent primary Ab responses but impaired recall Ab responses (45). Thus, the effects of IL-9 on potentiating IL-4–induced production of IgE and IgG observed in the initial studies are most likely attributable to memory B cells, B1 B cells, or MZ B cells included in the crude preparation of human and mouse B lymphocytes (41–43), as naive FO B cells do not express IL-9R and are not responsive to IL-9.

In the current study, we found that the expression of IL-9R was strikingly upregulated in mouse naive FO B cells genetically deficient in TNFR-associated factor 3 (TRAF3), a critical regulator of B cell survival and function (48–56). The highly upregulated IL-9R on Traf3^−/− FO B cells conferred responsiveness to IL-9, including IgM production and STAT3 phosphorylation. Interestingly, IL-9 significantly enhanced class switch recombination (CSR) to IgG1 induced by BCR crosslinking plus IL-4 in Traf3^−/− FO B cells, which was not observed in littermate control (LMC) B cells. This evidence reinforces the notion that in the absence of TRAF3, B cell activation and CSR no longer require cognate T cell help (53). Furthermore, we demonstrated that blocking the JAK-STAT3 signaling pathway abrogated the enhancing effect of IL-9 on CSR to IgG1 induced by BCR crosslinking plus IL-4. Our study, to our knowledge, thus revealed a novel pathway that TRAF3 suppresses B cell activation and Ig isotype switching by inhibiting the IL-9R-JAK-STAT3 signaling axis.

Traf3^flox/floxCD19^+/Cre (B cell–specific TRAF3-deficient [B-Traf3^−/−]) and Traf3^flox/flox (LMC) mice were generated as previously described (49). All experimental mice for this study were produced by breeding of Traf3^flox/flox mice with Traf3^flox/floxCD19^+/Cre mice. All mice were kept in specific pathogen-free conditions in the Animal Facility at Rutgers University and were used in accordance with National Institutes of Health guidelines and under an animal protocol (PROTO999900371) approved by the Animal Care and Use Committee of Rutgers University. Equal numbers of male and female mice were used in this study.

Tissue culture supplements including stock solutions of sodium pyruvate, l-glutamine, nonessential amino acids, and HEPES (pH 7.55) were from Invitrogen (Carlsbad, CA). IL-9 and IL-4 were purchased from PeproTech (Rocky Hill, NJ). Goat anti-mouse IgM F(ab′)₂ was obtained from Jackson ImmunoResearch Laboratories (West Grove, PA). Agonistic anti-CD40 (HM40-3) was purchased from eBioscience (San Diego, CA). Peficitinib, baricitinib, decernotinib, stattic, and AS1517499 were purchased from Sigma-Aldrich (St. Louis, MO). EDTA-free protease inhibitor cocktail tablets were obtained from Roche Diagnostics (Indianapolis, IN). Phosphatase inhibitor mini tablets were purchased from Pierce (Rockford, IL). DNA oligonucleotide primers were obtained from Integrated DNA Technologies (Coralville, IA). Polyclonal or monoclonal rabbit Abs against total or phosphorylated STAT3, STAT6, STAT1, RelA, RelB, c-Rel, NF-κB1, noncanonical NF-κB (NF-κB2), YY1, activation-induced cytidine deaminase (AID), Bcl6, and Blimp1 were from Cell Signaling Technology (Beverly, MA). Polyclonal rabbit Abs to TRAF3 (H122) were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-actin Ab was from Chemicon (Temecula, CA). HRP-labeled secondary Abs were from Jackson ImmunoResearch Laboratories (West Grove, PA).

Mouse splenic B cells were purified using anti-mouse CD43-coated magnetic beads and a MACS separator (Miltenyi Biotec) following the manufacturer’s protocols as previously described (49). The purity of the isolated B cell population was monitored by FACS analysis, and cell preparations of >98% B220⁺CD3⁻ purity were used for culture experiments as well as RNA or protein preparation. Purified splenic B cells were cultured ex vivo with mouse B cell medium (15, 25) in the absence or presence of IL-9, IL-4, anti-IgM (surface BCR), anti-CD40, LPS, CpG, or different inhibitors, alone or in combination, for 1–5 d before preparation of RNA or proteins and analysis of Ig isotype switching.

Single-cell suspensions were made from mouse spleens. Immunofluorescence staining and FACS analyses were performed as previously described (49, 57). Erythrocytes from spleens were depleted with ACK (ammonium-chloride-potassium) lysis buffer. Cells (1 × 10⁶) were blocked with rat serum and FcR blocking Ab (2.4G2), and then incubated with various Abs conjugated to different fluorochromes for multiple color fluorescence surface staining. Intracellular proteins were stained after fixation and permeabilization of cells using an intracellular or phospho-flow staining kit (BD Biosciences, San Jose, CA). Intracellular staining of Ig isotypes, including IgG1, IgG2a, IgG2b, IgG3, and IgE, was performed as previously described (58). Analyses of cell surface markers included Abs to CD45R (B220), CD3, CD21, CD23, IgM, CD138, IL-4R, IL-21R, and IL-6R (BioLegend, San Diego, CA). FACS Ab specific for mouse IL-9R was previously described (45). FACS Abs specific for p-STAT3 (pY705), p-STAT1 (pY701), p-STAT5 (pY694), p-STAT6 (pY641), and p-ERK1/2 (pT202/pY204) were purchased from Becton Dickinson (Mountain View, CA). FACS data were acquired on a Northern Lights spectral flow cytometer (Cytek, Fremont, CA). The results were analyzed using FlowJo software (Tree Star, San Carlos, CA). Forward light scatter (area and height) and side light scatter (area and height) gatings were used to gate single live cells.

Total cellular RNA was extracted from splenic B cells using an RNeasy mini kit (Qiagen, Germantown, MD) according to the manufacturer’s protocol. RNA concentration and quality were assessed using a NanoDrop spectrophotometer. cDNA was prepared from RNA using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA). A TaqMan assay of mouse Il9r was performed using specific TaqMan primers and probe (FAM labeled) (Applied Biosystems) as previously described (59, 60). Each reaction also included the probe (VIC labeled) and primers for mouse Actb, which served as an endogenous control. Quantitative real-time PCR was performed using the Power SYBR Green PCR master mix (Applied Biosystems) and primers specific for mouse Cγ1 germline transcript (GLT, forward, 5′-GGC CCT TCC AGA TCT TTG AG-3′; reverse, 5′-CAG GGT CAC CAT GGA GTT AGT T-3′), Aicda (forward, 5′-CGC GCC TCT ACT TCT GTG AA-3′; reverse, 5′-GTC TGG TTA GCC GGA CAG AA-3′), and Actb (forward, 5′-TGG AAT CCT GTG GCA TCC ATG AAA C-3′; reverse, 5′-TAA AAC GCA GCT CAG TAA CAG TCC G-3′) as described (58). Reactions were performed on a QuantStudio 3 real-time PCR system (96-well, 0.1-ml block, Applied Biosystems). Reaction specificity of each gene was monitored by postamplification melting curve analyses. Real-time PCR data were analyzed using the QuantStudio design and analysis software v1.5.2 (Applied Biosystems). Specific mRNA levels of each gene were normalized to that of the housekeeping gene Actb. Relative mRNA expression levels of each gene were calculated using the comparative threshold cycle (Ct) method (ΔΔCt) as previously described (59, 60).

The γ1 switch circle transcripts (SCTs) were amplified from cDNA samples using the Iγ1F primer (5′-GGC CCT TCC AGA TCT TTG AG-3′) paired with the CµR primer (5′-AAT GGT GCT GGG CAG GAA GT-3′) and PCR cycle conditions as reported by Kinoshita et al. (61). Gapdh transcripts were also amplified as a control using specific primers (forward, 5′-ACC ACA GTC CAT GCC ATC AC-3′; reverse, 5′-TCC ACC ACC CTG TTG CTG TA-3′). The PCR products were separated by PAGE gel electrophoresis followed by ethidium bromide staining for size assessment and imaging analysis.

Total protein lysates were prepared by lysing cell pellets in 2× SDS sample buffer (62.5 mM Tris [pH 6.8], 1% SDS, 15% glycerol, 2% 2-ME, and 0.005% bromophenol blue), sonicated for 30 pulses, and then boiled for 10 min (52, 59). Cytoplasmic and nuclear extracts were prepared as previously described (49, 62). Proteins were separated by SDS-PAGE and immunoblotted with Abs to phosphorylated or total STAT3, STAT6, STAT1, RelA, RelB, c-Rel, NF-κB1, p100/p52 NF-κB2, AID, Blimp1, Bcl6, TRAF3, YY1, or actin, followed by HRP-conjugated secondary Abs (goat anti-rabbit or goat anti-mouse IgG) (52, 60). A chemiluminescent substrate (Pierce) was used to detect HRP-labeled Abs on the immunoblots. Chemiluminescence images of the immunoblots were acquired using a low-light imaging system (LAS-4000 mini, FUJIFILM Medical Systems USA, Stamford, CT) (52, 57, 62).

Statistical analyses were performed using the Prism software (GraphPad, La Jolla, CA). For direct comparison of the levels of IL-9R expression between LMC and Traf3^−/− B cells, statistical significance was determined with the unpaired t test for two-tailed data. For comparison of three or more groups of data, a one-way ANOVA was used to determine the statistical significance. A p value <0.05 was considered significant (indicated as *p < 0.05, **p < 0.01, and ***p < 0.001).

To elucidate the mechanisms of TRAF3-mediated regulation of B cell survival and function, we recently performed a transcriptome profiling to identify genes differentially expressed between LMC and Traf3^−/− splenic B cells purified from sex-matched young adult mice (National Center for Biotechnology Information GEO accession no. GSE113920) (https://www.ncbi.nlm.nih.gov/geo/) (51). Interestingly, we found that the expression of Il9r was strikingly upregulated in Traf3^−/− splenic B cells (GSE113920). Given the recent recognition of IL-9R in the memory B cell response (6, 45), we verified its striking upregulation (>30-fold) in Traf3^−/− B cells at the mRNA level using TaqMan quantitative real-time PCR (qRT-PCR) (Fig. 1A). To investigate the potential importance of Il9r upregulation in Traf3 deficiency–initiated B cell malignant transformation, we examined the transcript levels of Il9r in splenic B lymphomas spontaneously developed in seven different individual B-Traf3^−/− mice using the TaqMan assay. Our results showed that Il9r expression was moderately upregulated (2- to 5-fold) in six of the seven examined splenic B cell lymphomas (Fig. 1B). Thus, it appeared that the highly upregulated Il9r expression observed in nonmalignant Traf3^−/− B cells is not stringently maintained during or after B cell malignant transformation. Based on these results, we selected to focus on the Il9r upregulation in nonmalignant Traf3^−/− B lymphocytes in the current study.

We next verified the drastic upregulation of IL-9R on nonmalignant Traf3^−/− B cells at the protein level using flow cytometric analysis. We found that IL-9R was strikingly upregulated in both the splenic FO and MZ B cell subsets of young adult B-Traf3^−/− mice (Fig. 1C). In contrast, the expression of IL-9R was not altered in splenic T cells, and the expression of another cytokine receptor, IL-4R, was not changed in FO or MZ B cell subsets of B-Traf3^−/− mice (Fig. 1C). Our results, to our knowledge, thus revealed a novel link between TRAF3 and IL-9R in mature B lymphocytes, including both the FO and MZ B cell subsets.

To investigate the potential involvement of IL-9R in the activation of B lymphocytes, we analyzed the expression levels of IL-9R on cultured LMCs and Traf3^−/− splenic B cells after stimulation with various B cell stimuli, including agonistic anti-CD40 Abs, LPS (an agonist of TLR4), CpG2084 (an agonist of TLR9), BCR crosslinking Abs, IL-4, and IL-9. Our FACS data demonstrated that among the B cell stimuli examined, CD40 engagement most robustly induced IL-9R upregulation on the surface of LMC FO and MZ B cells (Fig. 1D, 1E, Supplemental Fig. 1A), confirming the previous reports that IL-9R expression is induced by CD40 stimulation in B cells (44, 45). We noticed that LPS or CpG also clearly induced IL-9R upregulation on LMC FO and MZ B cells (Fig. 1D, 1E, Supplemental Fig. 1A). Each of these stimuli further increased IL-9R expression on the surface of Traf3^−/− FO and MZ B cells (Fig. 1D, 1E, Supplemental Fig. 1A). In contrast, anti-BCR plus IL-4, in the absence or presence of IL-9, did not induce IL-9R expression on LMC or Traf3^−/− FO and MZ B cells (Fig. 1D, 1E, Supplemental Fig. 1A). Addition of IL-4 could not potentiate IL-9R upregulation induced by CD40 engagement in both genotypes of FO and MZ B cells (Fig. 1D, 1E, Supplemental Fig. 1A). These data suggest that IL-9R upregulation may have certain functional significance in B cell activation.

To understand how Il9r expression is upregulated, we analyzed the promoter region of the mouse Il9r gene and identified potential binding sites for the transcription factors AP-1, Sp1, and NF-κB. This is consistent with the initial report that the 5′ flanking region of the human IL9R gene also contains potential binding motifs for AP-1, AP-2, AP-3, Sp1, and NF-κB (63). We previously reported constitutively elevated levels of nuclear p52 NF-κB2 and RelB but normal levels of phosphorylated ERK, JNK, and p38 in Traf3^−/− B cells (49). NF-κB1 activation induced by CD40, LPS, or BCR signaling is also enhanced in Traf3^−/− B cells (53–55). In this context, we reasoned that NF-κB1 or NF-κB2 may be responsible for the upregulation of Il9r expression in B cells. We thus searched and analyzed available RNA sequencing datasets for the expression of Il9r in splenic B cells genetically deficient in different NF-κB subunits in comparison with wild-type (WT) B cells, in the absence or presence of stimulation with B cell activating factor (BAFF), CD40, or CD40 plus IgM (64–66). These include Nfkb2^−/− splenic B cells (GSE75761 and GSE75762: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE75761 and https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE75762), Rela^−/− splenic B cells (GSE58972: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58972), and cRel^−/− and Relb^−/−cRel^−/− FO B cells (GSE62559: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE62559) (64–66). The search results of these RNA sequencing datasets revealed that BAFF or CD40-induced Il9r expression was not compromised in Nfkb2^−/− splenic B cells (GSE75761 and GSE75762), whereas CD40+IgM–induced Il9r expression was moderately decreased in Rela^−/− splenic B cells (GSE58972) (Supplemental Fig. 1B). Interestingly, BAFF-induced Il9r expression was completely abolished in FO B cells doubly deficient in both c-Rel and Relb (Relb^−/−cRel^−/−), although cRel^−/− FO B cells exhibited increased Il9r expression as compared with WT cells (GSE62559) (Supplemental Fig. 1B). Taken together, these data suggest that the NF-κB subunits RelB and RelA contribute to the upregulation of Il9r expression in Traf3^−/− B cells or following B cell activation.

To explore the functional impacts of IL-9R upregulation on Traf3^−/− B cells, we first compared the effects of IL-9 stimulation on the survival, proliferation, and Ig production between LMC and Traf3^−/− B cells purified from sex-matched, young adult mice. We found that IL-9 stimulation did not promote the survival or proliferation in both genotypes of B lymphocytes (Supplemental Fig. 2A). However, IL-9 induced IgM production in Traf3^−/−, but not LMC, B lymphocytes as analyzed by both intracellular FACS staining of the cultured FO B cells and ELISA of the culture supernatants (55, 58, 67) (Fig. 2A, 2B). We did not detect any effects of IL-9 on the production of other Ig isotypes in cultured Traf3^−/− or LMC B cells, including IgG1, IgG2a, IgG2b, IgG3, IgA, and IgE (data not shown), which often requires T-dependent signals. Thus, IL-9 specifically induced IgM production, a T-independent effect, in Traf3^−/− B lymphocytes.

In light of the early evidence that IL-9 potentiates IL-4–induced production of IgG and IgE by B lymphocytes (41–43), we determined the effects of IL-9 and IL-4, alone or in combination, on LMC and Traf3^−/− B cells. We noticed that IL-4 markedly promoted the survival of LMC B cells and slightly helped the survival of Traf3^−/− B cells, which was not potentiated by the addition of IL-9 (Supplemental Fig. 2B). IL-9 and IL-4 did not induce IgG1 or IgE production in LMC B cells (Supplemental Fig. 2C, 2D). In contrast, addition of IL-9 augmented the frequency of IL-4–induced IgG1-producing, but not IgE-producing, Traf3^−/− B cells (Supplemental Fig. 2C, 2D). However, the combined treatment with IL-9 plus IL-4 only induced a very small population of IgG1-producing Traf3^−/− B cells in culture (<1.5%), suggesting that these two cytokines are not sufficient to induce a robust Ig isotype switching response in IL-9R^hi B cells.

Chen et al. (53) recently reported that BCR crosslinking plus IL-4 can significantly induce IgG1 CSR in Traf3^−/− B lymphocytes. We thus investigated the CSR in LMC and Traf3^−/− B cells after treatment with BCR crosslinking in combination with IL-9 or IL-4. Our results showed that addition of IL-9 to BCR crosslinking could not induce CSR in both genotypes of B cells. Interestingly, addition of IL-9 to BCR crosslinking plus IL-4 significantly potentiated Ig isotype switching to IgG1 in Traf3^−/− but not LMC B cells (Fig. 2C, 2D), although it could not promote the survival or proliferation in these B cells (Supplemental Fig. 3A). We verified that IL-9 potentiated BCR plus IL-4–induced CSR to IgG1 in a dose-dependent manner (Supplemental Fig. 3B, 3C). We also analyzed the effects of IL-9 on CSR induced by CD40 engagement plus IL-4, a combinatory stimulation commonly used in CSR studies. We found that addition of IL-9 to CD40 engagement plus IL-4 significantly potentiated Ig isotype switching to IgG1 in both genotypes of B cells and that this potentiation effect was more robust in Traf3^−/− than in LMC B cells (Fig. 2E, 2F). These results are consistent with the observation that IL-9R upregulation was induced by CD40 engagement plus IL-4 but not by BCR crosslinking plus IL-4 in both genotypes of B cells (Fig. 1D, 1E). In contrast, IL-9 did not have obvious effects on CSR to IgE, IgG2a, IgG2b, or IgG3 in the presence of BCR crosslinking plus IL-4 or CD40 engagement plus IL-4 in both genotypes of B cells (Fig. 2C, 2E, and data not shown). Taken together, our results indicate that upregulation of IL-9R is functionally important and confers the responsiveness to IL-9 in Traf3^−/− B lymphocytes, including IgM production induced by this cytokine alone as well as its potentiation of CSR to IgG1 induced by IL-4 in combination with BCR crosslinking or CD40 engagement. Thus, IL-9 alone is sufficient to induce a T‐independent effect, that is, IgM production in Traf3^−/− B cells, whereas IL-9 together with BCR crosslinking plus IL-4 are able to stimulate a T-dependent response, that is, IgG1 CSR in Traf3^−/− B cells.

CSR in B lymphocytes is preceded by germline transcription of specific switch regions in the IgH locus and is accompanied by excision of specific switch circles (61, 68). To elucidate how IL-9 could potentiate CSR to IgG1 induced by BCR crosslinking plus IL-4, we first examined the expression levels of the Cγ1 GLT in LMC and Traf3^−/− B cells. Interestingly, we found that even in the absence of IL-4, BCR crosslinking plus IL-9 induced significantly higher levels of the Cγ1 GLT in Traf3^−/− B cells than in LMC B cells (Fig. 3A). The addition of IL-4 drastically further increased the levels of the Cγ1 GLT in both genotypes of B cells (Fig. 3A). We also tested the possibility that IL-9 alone could induce the expression of the Cγ1 GLT in Traf3^−/− B cells. Our results demonstrated that this indeed was the case (Fig. 3B). We next analyzed the induction of the γ1 SCT, a hallmark of active de novo CSR in B cells, using reverse transcription PCR (61). Our results showed that BCR crosslinking plus IL-4 clearly induced the expression of the γ1 SCT and that addition of IL-9 robustly further increased the induction of the γ1 SCT expression in Traf3^−/− B cells (Fig. 3C), indicating the presence of active de novo CSR to IgG1 in these stimulated Traf3^−/− B cells.

CSR and production of switch circles require the critical enzyme AID, which is encoded by the Aicda gene (61, 69). We next determined the expression of AID at the mRNA and protein levels using qRT-PCR and Western blot analyses, respectively. We found that addition of IL-9 remarkably potentiated BCR crosslinking plus IL-4–induced expression of AID at both the mRNA and protein levels in Traf3^−/− but not LMC B cells (Fig. 3D, 3E). Fawaz et al. (44) previously suggested that IL-9 induces the upregulation of Bcl-6 in GC B cells, and Takatsuka et al. (45) recently reported that IL-9 regulates the expression of Blimp-1 in in vitro–induced GC-like B cells (70). Given that Bcl-6 and Blimp-1 are two transcriptional repressors especially crucial for B cell differentiation (71), we investigated their expression in our experimental model using Western blot analysis. We observed that addition of IL-9 did not further enhance the expression of Bcl-6 and Blimp-1 induced by BCR crosslinking plus IL-4 in both LMC and Traf3^−/− B cells (Fig. 3E), arguing against a role for Bcl-6 or Blimp-1 in the IL-9–induced effect on CSR. We noticed that the levels of Aicda and AID expression induced by IL-9 together with BCR crosslinking plus IL-4 (Fig. 3D, 3E) appear to be discordant with the extent of Cγ1 GLT induction (Fig. 3A), suggesting that Cγ1 GLT induction does not strictly require high protein levels of AID. Indeed, an early study by Berton and Vitetta (72) demonstrated that IL-4 alone can induce Cγ1 GLT expression in resting WT B cells. In contrast, the effects of IL-9 together with BCR crosslinking plus IL-4 on the Aicda transcript and AID protein expression (Fig. 3D, 3E) are consistent with their effect on the γ1 SCT expression (Fig. 3C), reinforcing the notion that AID is critically required for CSR and production of switch circles (61, 69). Taken together, our results revealed that IL-9, in the presence of BCR crosslinking plus IL-4, induces the expression of AID and de novo CSR to IgG1 in Traf3^−/− B cells.

It has been reported that IL-9 induces rapid activation of STAT3, STAT1, STAT5, and ERK1/2 in various target cells (2, 6, 28, 44, 45, 47). We analyzed and compared these proximal signaling events of IL-9R between LMC and Traf3^−/− B cells. Our results of phospho-flow cytometry demonstrated that IL-9 alone drastically induced the phosphorylation of STAT3 and also markedly induced the phosphorylation of STAT1 in Traf3^−/− FO B cells, but it barely induced any phosphorylation of STAT3 or STAT1 in LMC FO B cells (Fig. 4A, 4B, Supplemental Fig. 4A). Consistent with the previous observations that WT MZ B cells express IL-9R and respond to IL-9 treatment (1, 45, 47), we detected a robust induction of STAT3 and STAT1 phosphorylation by IL-9 alone in the MZ B cell subset of LMC mice (Fig. 4C, 4D, Supplemental Fig. 4A). In addition, IL-9–induced STAT1 phosphorylation, but not STAT3 phosphorylation, was significantly higher in Traf3^−/− than in LMC MZ B cells (Fig. 4C, 4D), which correlates with the elevated expression level of IL-9R detected on the surface of Traf3^−/− MZ B cells (Fig. 1C). In contrast, IL-9 alone did not induce the phosphorylation of STAT5 or ERK1/2 in FO and MZ B cell subsets of both genotypes (Fig. 4, Supplemental Fig. 4B, 4C). Our results thus identify STAT3 and STAT1 as two major signaling components of IL-9R in Traf3^−/− B lymphocytes.

We wondered whether the enhanced phosphorylation of STAT3 and STAT1 induced by IL-9 can be attributed entirely to the IL-9R upregulation or whether there is an additional effect on signaling conferred by Traf3 deficiency. To address this, we examined the phosphorylation of STAT3 and STAT1 induced by another cytokine of the γ-chain family, IL-21, using phospho-flow cytometry. Our results demonstrated that IL-21–induced phosphorylation of STAT3 and STAT1 was modestly enhanced in Traf3^−/− FO B cells (Supplemental Fig. 5A, 5B). In addition, Lin et al. (56) previously reported that IL-6–induced STAT3 phosphorylation and plasma cell differentiation are significantly increased in Traf3^−/− B cells. However, we did not detect any significant difference in the expression levels of IL-21R and IL-6R between Traf3^−/− and LMC FO B cells (Supplemental Fig. 5C), suggesting an effect of Traf3 deficiency on IL-21R and IL-6R signaling in B cells. In this regard, Lin et al. (56) revealed that TRAF3 inhibits IL-6R signaling by facilitating the association of the phosphatase PTPN22 with the kinase Jak1 to restrain STAT3 phosphorylation. Based on the above evidence, we speculate that the actions of IL-9 observed in Traf3^−/− FO B cells can be attributed mainly to the striking upregulation of IL-9R, although there is likely an additional effect of Traf3 deficiency on IL-9R signaling in B cells.

We further investigated the combined effects of IL-9 and BCR crosslinking plus IL-4 on these proximal signaling events. We found that BCR crosslinking plus IL-4 could not further increase IL-9–induced phosphorylation of STAT3 and STAT1 in Traf3^−/− FO and MZ B cell subsets (Fig. 4). Interestingly, BCR crosslinking plus IL-4 remarkably induced the phosphorylation of STAT6, a key player of the IL-4 signaling pathway (73), in FO and MZ B cell subsets of both genotypes (Fig. 4). However, addition of IL-9 could not further increase the phosphorylation of STAT6 induced by BCR crosslinking plus IL-4 (Fig. 4). These data suggest the potential importance of STAT3, STAT1, and STAT6 in CSR induced by IL-9 in the presence of BCR crosslinking plus IL-4 in Traf3^−/− B cells.

We next employed Western blot analysis of cytoplasmic and nuclear extracts to verify our findings obtained from phospho-flow cytometry. Our results demonstrated that addition of IL-9 markedly potentiated the phosphorylation and nuclear translocation of STAT3 induced by BCR crosslinking plus IL-4 in Traf3^−/− B cells (Fig. 5A). The results of Western blot analysis also verified that IL-9 could not further increase the phosphorylation and nuclear translocation of STAT6 induced by BCR crosslinking plus IL-4 in both genotypes of B cells (Fig. 5A). In contrast, addition of IL-9 did not clearly increase the phosphorylation (pY701) and nuclear translocation of STAT1 induced by BCR crosslinking plus IL-4 in Traf3^−/− B cells as determined by Western blot analysis (Supplemental Fig. 6), which did not reflect the data of phospho-flow cytometry (Fig. 4) for an unknown reason. We previously reported the constitutive NF-κB2 activation in Traf3^−/− B cells (49). In this study we showed that IL-9 together with BCR crosslinking plus IL-4 did not further increase the levels of nuclear NF-κB2 subunits, including p52 NF-κB2 and RelB, in Traf3^−/− B cells (Supplemental Fig. 6). It is known that BCR crosslinking induces NF-κB1 activation and that NF-κB1 plays a critical role in CSR (55, 58, 74). We found that addition of IL-9 did not potentiate the increase of cytoplasmic or nuclear NF-κB1 subunits, including p50 NF-κB1, RelA, and c-Rel, induced by BCR crosslinking plus IL-4 in both genotypes of B cells (Supplemental Fig. 6). Collectively, our results suggest that IL-9 and BCR crosslinking plus IL-4 activate distinct early signaling players such as STAT3, NF-κB1, and STAT6, which may cooperate at downstream signaling components to synergistically induce CSR in Traf3^−/− B cells.

It has been shown that IL-9R binds to JAK1 constitutively and that the common γ-chain receptor binds to JAK3 constitutively. Upon IL-9 engagement, IL-9R and γ-chain form a heterodimeric complex, and thus bring JAK1 and JAK3 to close proximity to induce their cross-activation, which in turn induces the phosphorylation of STAT3 (2, 28, 47). Similarly, IL-4 induces the phosphorylation of STAT6 via activating the IL-4R–associated JAK1 and γ-chain–associated JAK3 in lymphocytes (73). To elucidate whether these JAK-STAT signaling pathways are required for the upregulation of AID expression induced by IL-9 in combination with BCR crosslinking plus IL-4, we examined the effects of several pharmacological inhibitors specific for JAK or STAT proteins. These include peficitinib (a pan-JAK inhibitor), baricitinib (a JAK1/JAK2 inhibitor), decernotinib (a JAK3 inhibitor), stattic (a STAT3 inhibitor), and AS1517499 (a STAT6 inhibitor) (75, 76). As demonstrated by both the qRT-PCR data of the transcript level and Western blot results of the protein level, the induction of AID expression stimulated by IL-9 in combination with BCR crosslinking plus IL-4 was almost completely abolished by each of the three examined JAK inhibitors (peficitinib, baricitinib, and decernotinib) as well as the STAT6 inhibitor AS1517499, and it was also substantially blocked by the STAT3 inhibitor stattic in Traf3^−/− B cells (Fig. 5B, 5C). Hence, these results indicate that the JAK1/3-STAT3/6 signaling pathways are required for BCR+IL-4+IL-9–induced expression of AID, which is vital for mediating de novo CSR in activated B lymphocytes.

Given the blocking effects of the JAK and STAT inhibitors on AID expression observed in Traf3^−/− B cells stimulated with IL-9 in combination with BCR crosslinking plus IL-4, we next sought to investigate their subsequent functional impacts on Ig isotype switching. We found that the CSR to IgG1 induced by IL-9 in combination with BCR crosslinking plus IL-4 was almost completely abolished by each of the three examined JAK inhibitors (peficitinib, baricitinib, and decernotinib), was severely impaired by the STAT6 inhibitor AS1517499, and was also substantially suppressed by the STAT3 inhibitor stattic in Traf3^−/− B cells (Fig. 6). To further dissect out the IL-9–specific signaling components from those of IL-4, we determined the effects of these pharmacological inhibitors on CSR induced by CD40 engagement plus IL-4 in LMC and Traf3^−/− B cells. Interestingly, the CSR to IgG1 induced by CD40 engagement plus IL-4 was almost abolished by each of the three examined JAK inhibitors (peficitinib, baricitinib, and decernotinib) as well as the STAT6 inhibitor AS1517499, but it was not inhibited at all by the STAT3 inhibitor stattic in both genotypes of B cells (Fig. 6). We also verified that the JAK or STAT3 inhibitors did not suppress the expression of IL-9R on Traf3^−/− or LMC B cells, in the absence or presence of stimulation by CD40 engagement plus IL-4 (Supplemental Fig. 7). Taken together, our findings suggest that the JAK-STAT6 signaling pathway is induced by IL-4 and is therefore required for CSR induced by both BCR+IL-4+IL-9 and CD40+IL-4. In contrast, JAK-STAT3 signaling is specifically induced by IL-9 and is thus required for CSR induced by BCR+IL-4+IL-9, but not CD40+IL-4, in Traf3^−/− B lymphocytes (Fig. 7).

Regulation of B cell function is central to humoral immunity (38–40, 77). Inadequate B cell activation often leads to compromised Ab responses, infectious diseases, and chronic infections, whereas aberrant activation of B cells by innocuous or self-antigens causes allergic and autoimmune diseases (38–40, 77). A better understanding of how precisely B cell function is regulated in different immune contexts is required for improving the management of these disease conditions resulting from abnormal B cell activation. In the current study, to our knowledge, we provide novel insights into the molecular mechanisms of how an important cytokine IL-9 regulates B cell activation and Ig isotype switching in the context of genetic deficiency in TRAF3, a vital regulator of B cell survival and function (48–56). TRAF3 critically restrains the signaling events of multiple receptors essential for B cell physiology, including BAFF-R, CD40, BCR, TLRs, and IL-6R (48–50, 53–56, 78). In an effort to identify TRAF3 downstream target genes in B lymphocytes, we found that the expression of IL-9R was strikingly upregulated in Traf3^−/− FO B cells. The highly upregulated IL-9R expression conferred responsiveness to IL-9 in naive Traf3^−/− FO B cells. Taken together, our evidence presented in this study, to our knowledge, revealed a novel pathway that TRAF3 suppresses B cell activation and Ig CSR by inhibiting the IL-9R-JAK-STAT3 signaling axis.

Interestingly, IL-9–induced IL-9R-JAK-STAT3 signaling (but not STAT5 or STAT6 phosphorylation) remarkably and specifically potentiated CSR to IgG1 induced by BCR crosslinking plus IL-4 or CD40 engagement plus IL-4 in Traf3^−/− FO B cells. In line with our finding, STAT3 is known as a transcription factor that promotes IgG-specific plasma cell differentiation and inhibits CSR from IgG1 to IgE in GC B cells, which has been demonstrated by both mouse and human evidence (79–82). B cell–specific Stat3^−/− mice display profound defects in T-dependent IgG1 responses, but elevated serum IgE levels (79, 80). In humans, dominant-negative mutations in the STAT3 gene underlie sporadic and dominant forms of the hyper-IgE syndrome, a disease characterized by elevated serum IgE levels and recurrent infections (81, 82). Similar to IL-9, the FO helper CD4 T cell–derived cytokine IL-21 also induces the JAK-STAT3 signaling pathway to potentiate IgG1 CSR, while inhibiting IgE CSR in GC B cells (83). The importance of STAT3 in this process is further highlighted by the evidence that naive B cells from hyper-IgE syndrome patients with STAT3 mutations are unable to differentiate into IgG-producing cells after stimulation with IL-21 in the presence of CD40L (84, 85). However, IL-21R, but not IL-9R, is expressed on naive WT FO B cells, and therefore normal naive FO B cells respond to IL-21 but are irresponsive to IL-9 (45, 84, 86).

For normal naive FO B cells, full activation and Ig CSR require not only the engagement of BCR by specific Ags but also the signals provided by T_FH cells that recognize peptides presented in complex with MHC class II on the surface of B cells, including the membrane-bound CD40L expressed on the surface of T_FH cells and the cytokines such as IL-21 and IL-4 secreted by T_FH cells (84, 87–90). Such linked recognition of Ags by B cells and cognate T_FH cells serves to distinguish foreign Ags from self-antigens (84, 87–90). Upon Ag binding, BCR signaling induces the activation of NF-κB1, Akt-FOXO1, and ERK1/2-AP-1 (87, 88, 91). Engagement of CD40 by CD40L stimulates the activation of NF-κB1, NF-κB2, Akt, and ERK1/2 (87, 88, 91). IL-4 induces the principal IL-4R-JAK-STAT6 signaling pathway, whereas IL-21 primarily activates the IL-21R-JAK-STAT3 signaling pathway (84, 88–90). The concerted action of NF-κB1, NF-κB2, STAT6, and STAT3 potently induces the expression of AID, the enzyme that catalyzes CSR, thus mediating Ig isotype switching in activated B cells (84, 87–90). In the absence of CD40L and IL-21 provided by T_FH cells, BCR crosslinking by various Ags plus IL-4 can stimulate the BCR and IL-4R signaling pathways to induce the expression of certain transcription factors required for B cell differentiation such as Bcl-6 and Blimp1, but it cannot induce the expression of AID due to lack of NF-κB2 activation and STAT3 signaling (Figs. 3D, 3E, 5, Supplemental Fig. 6). Consequently, normal naive FO B cells cannot undergo CSR in response to BCR crosslinking triggered by self-antigens even if inflammatory cytokines IL-4 and IL-9 are present in the microenvironment (Fig. 7A).

However, for TRAF3-deficient naive FO B cells, full activation and Ig CSR do not stringently require the signals provided by T_FH cells (Fig. 7B). Previous evidence reveals that deletion of TRAF3 leads to constitutive activation of NF-κB2 and also results in elevated BCR signaling in B cells (49, 50, 53, 54). In response to BCR crosslinking plus IL-4, the induced activation of NF-κB1 and STAT6 together with the constitutive activation of NF-κB2 moderately promotes AID expression and CSR in Traf3^−/− B cells (Figs. 2C, 2D, 3C, 3E, 4A, 4B) (53). In the current study, we further demonstrated that loss of TRAF3 also causes strikingly upregulated expression of IL-9R on FO B cells and that addition of IL-9 robustly potentiates AID expression and CSR induced by BCR crosslinking plus IL-4 via JAK-STAT3 signaling in Traf3^−/− B cells (Figs. 2C, 2D, 3C–E, 4A, 4B, 5A, Supplemental Fig. 6). Thus, in the absence of intracellular TRAF3, constitutive activation of NF-κB2 provides surrogate CD40 signaling to remove the requirement of CD40L expressed on the surface of T_FH cells (53), whereas upregulated IL-9R confers IL-9–induced JAK-STAT3 signaling to replace the requirement of IL-21 produced by T_FH cells. As a consequence, full activation and Ig CSR of TRAF3-deficient naive FO B cells could be induced by self-antigens, as long as they can engage specific BCRs and the necessary cytokines such as IL-4 and IL-9 are present in the microenvironment (Fig. 7B). Consistent with this scenario, elevated serum levels of IL-4 and IL-9 have been detected in human patients with systemic lupus erythematosus and other autoimmune disorders (8, 21, 92–94). Therefore, although it is recognized that T‐dependent functions cannot be simply reduced to NF-κB2 activation and cytokine production and that B cell–T cell cognate interactions are necessary for affinity maturation and high-affinity Ab production, which is often found in autoimmune conditions (95), the above findings suggest that loss of TRAF3 in B cells would lead to increased risks of autoimmunity.

Indeed, B-Traf3^−/− mice exhibit autoimmune manifestations, including the presence of high levels of anti-dsDNA autoantibodies in the serum, spontaneous formation of GCs in the spleen, and deposition of immune complexes at glomeruli in the kidney (49). Relevant to the observed autoimmunity, Traf3^−/− B cells display enhanced Ab production and Ig CSR responses to TLR agonists (55). In line with the notion that autoimmune conditions are strong predisposing factors for B lymphoma development (96), we observed that >80% of B-Traf3^−/− mice spontaneously develop B cell lymphomas at the age of 9–18 mo (57). Deletions and inactivating mutations of the TRAF3 gene are frequently detected in a variety of human B cell malignancies (97–100). Upregulation of IL-9R has also been detected in malignant B cells of patients with diffuse large B cell lymphoma and Hodgkin lymphoma, while elevated serum levels of IL-9 have been observed in patients with chronic lymphocytic leukemia, diffuse large B cell lymphoma, and Hodgkin lymphoma (28, 101–105). Overexpression of IL-9 in transgenic mice results in a complex phenotype characterized by the development of autoantibodies and T cell lymphoma (19). Interestingly, IL-9 significantly increases in vitro production of anti–systemic scleroderma 70 (Scl70) autoantibodies by B cells isolated from the peripheral blood of patients with systemic sclerosis, suggesting a pathogenic role of IL-9/IL-9R on autoreactive B cells (22). Furthermore, the latent membrane protein LMP1 encoded by the EBV, a risk factor of B cell oncogenesis and autoimmunity (106–108), sequesters TRAF3 in B lymphocytes and can render EBV-infected B cells functionally TRAF3 deficient without genetic alterations of the TRAF3 gene (109, 110). In this context, our study lays the ground for further investigation on the TRAF3-IL-9R-JAK-STAT3 axis of B lymphocytes in the pathogenesis of autoimmunity, especially in human patients of autoimmune disorders associated with EBV or B cell malignancies.

Elevated levels of IL-9 and/or IL-4 are also detected in the specimens of patients with various allergic diseases, including asthma, atopic dermatitis, and food allergy (23, 111–113). Moreover, IL-9 promotes IL-4–induced IgE production in allergic B cells isolated from children with asthma (112), indicating that allergic B cells express IL-9R. Interestingly, the rs71421264 and rs56101042 single-nucleotide polymorphisms of TRAF3 are associated with increased susceptibility to asthma and atopic dermatitis, respectively, in human patients (114, 115). In light of our finding that upregulated IL-9R on Traf3^−/− B cells allows IL-9 to potentiate BCR+IL-4–induced CSR to IgG1, it is possible that the TRAF3-IL-9R-JAK-STAT3 axis plays a role in the pathogenesis of IgG1-dependent allergic diseases (116–118). Notably, IgG1⁺ memory B cells can undergo further CSR to generate IgE-producing plasma cells upon subsequent exposure to specific allergens (118–123). Based on the above evidence, we speculate that the TRAF3-IL-9R axis may be broadly implicated in the pathogenesis of both IgG1- and IgE-dependent allergic diseases in patients having repeated exposures to allergens, which requires further investigation in future studies.

In summary, our findings provide, to our knowledge, the novel insights into the TRAF3-IL-9R axis in B cell function and bear significant implications for the understanding and treatment of a variety of human diseases involving aberrant B cell activation, including autoimmune disorders and allergic diseases. Our study identifies the TRAF3-IL-9R-JAK-STAT3 pathway components as potential therapeutic targets to intervene in autoimmunity and allergy. Given the recent exciting developments of IL-9 and Th9/Tc9 in the immunotherapy of solid tumors (30, 31, 124, 125), our findings suggest that prudent considerations should be given on the potential risks of such therapy at inducing autoimmunity or allergy. This is particularly important for solid tumor patients with latent EBV infection, autoimmune or allergic history, or specific TRAF3 inactivation or single-nucleotide polymorphisms. Considering that autoreactive, allergic, and TRAF3-deficient B cells are responsive to IL-9 (5, 22, 26, 112), this knowledge will provide new perspectives to minimize adverse effects of IL-9 and Th9/Tc9 when treating solid tumors within different genetic and immune contexts.

The authors have no financial conflicts of interest.

We thank Dr. Sining Zhu, Judith Shakarchi, Shilpreet Brar, and Andrew Xia for technical assistance of this study.