Complexes between nuclear factor-κB p65 and signal transducer and activator of transcription 3 are key actors in inducing activation-induced cytidine deaminase expression and immunoglobulin A production in CD40L plus interleukin-10-treated human blood B cells

Abstract

The signal transducer and activator of transcription 3 (STAT3) transcription factor pathway plays an important role in many biological phenomena. STAT3 transcription is triggered by cytokine-associated signals. Here, we use isolated human B cells to analyse the role of STAT3 in interleukin (IL)-10 induced terminal B cell differentiation and in immunoglobulin (Ig)A production as a characteristic readout of IL-10 signalling. We identified optimal conditions for inducing in-vitro IgA production by purified blood naive B cells using IL-10 and soluble CD40L. We show that soluble CD40L consistently induces the phosphorylation of nuclear factor (NF)-κB p65 but not of STAT3, while IL-10 induces the phosphorylation of STAT3 but not of NF-κB p65. Interestingly, while soluble CD40L and IL-10 were synergistic in driving the terminal maturation of B cells into IgA-producing plasma cells, they did not co-operate earlier in the pathway with regard to the transcription factors NF-κB p65 or STAT3. Blocking either NF-κB p65 or STAT3 profoundly altered the production of IgA and mRNA for activation-induced cytidine deaminase (AID), an enzyme strictly necessary for Ig heavy chain recombination. Finally, the STAT3 pathway was directly activated by IL-10, while IL-6, the main cytokine otherwise known for activating the STAT3 pathway, did not appear to be involved in IL-10-induced-STAT3 activation. Our results suggest that STAT3 and NF-κB pathways co-operate in IgA production, with soluble CD40L rapidly activating the NF-κB pathway, probably rendering STAT3 probably more reactive to IL-10 signalling. This novel role for STAT3 in B cell development reveals a potential therapeutic or vaccine target for eliciting IgA humoral responses at mucosal interfaces.

Introduction

Naive mature B cells express both immunoglobulins (Ig) M and D. Antigen and T cell-dependent or -independent activation induces class switch recombination (CSR) of differentiated B cell genes, a molecular mechanism involving Ig heavy chain (C_H) gene rearrangements. After such activation, B cells produce IgG, IgA or IgE antibodies [1]. Whatever the mechanism, antibody production involves activation-induced cytidine deaminase (AID), an enzyme strictly necessary for Ig heavy chain recombination [2].

IgA constitutes the most abundant antibody class in the gut, where it contributes to immune protection against certain pathogens. Within the gut, low- and high-affinity IgA is produced in the lamina propria (LP) and Peyer's patches, respectively [3]. Low-affinity IgA produced by T-independent mechanisms essentially acts in blocking adhesion of commensal bacteria to epithelial cells [4–6]. Recently, it was shown that APRIL (a-proliferation-inducing ligand) triggers the differentiation of IgM⁺ B cells into low-affinity IgA plasma cells within the LP in response to Toll-like receptor (TLR) stimulation of epithelial cells [7]. B cell activating factor (BAFF) belonging to the tumour necrosis factor (TNF) family was also shown to sustain the differentiation of IgM⁺ CD27⁺ marginal zone B cells into IgA plasma cells, independently of CD40 [7], in the subepithelial regions of the mucosa. In contrast, the T-dependent production of high-affinity IgA occurs in the germinal centres (GC) of the Peyer's patches and requires CD40–CD40L interactions [8]. During a T-dependent response, CSR is promoted by CD40–CD40L interactions and modulated by various cytokines that target specific C_H genes prior to germline transcription [9]. A panel of cytokines, including TGF-β, interleukin (IL)-10 and others can skew CSR towards IgA.

CD40L, BAFF and APRIL trigger the activation of both nuclear factor (NF)-κB₁ and NF-κB₂[10]; however, only the NF-κB₁ pathway leads to NF-κB p65 activation. The NF-κB subunits (p50, p52, p65, c-Rel, RelA and RelB) function as dimers and have been shown to be both differentially activated [11,12] and also to possess distinct target DNA binding site specificities [13,14] that depend upon dimer composition. The CD40/CD40L interaction activates and phosphorylates the latent cytoplasmic NF-κB/IκB complex. This process is followed by IκB proteolysis and the translocation of NF-κBp50 or p65 into the nucleus, where these NF-κB subunits up-regulate gene expression by binding κB site-containing gene promoters [15]. NF-κB₁ may also affect other independent pathways upon activation of TNF receptor-associated factors, such as Janus kinases (JAK) and signal transducers and activators of transcription (STAT) [16].

Complex interactions exist between NF-κB subunits and STAT3 that can differently modulate B cell responses to pathogens. Phosphorylated p65 dimer can bind to non-phosphorylated STAT3 and this complex can then bind to κB sites, but not on γ-activated sites (GAS–STAT component) [17]. Alternatively, the phosphorylated form of STAT3 can interact with the phosphorylated NF-κB p50. This complex enhances the transcription of GAS-dependent genes [18]. Moreover, phosphorylated STAT3 can form a complex with a non-phosphorylated NF-κB dimer and bind to κB sites [19]. The recruitment and activation of STAT3 can also induce downstream expression of numerous cytokine receptors, including IL-10 receptor (IL-10R). IL-10 participates in many biological responses, including cell proliferation, survival, apoptosis and differentiation [20,21], and is an important factor in the regulation of Ig production. IL-10 is reported to be a possible switch factor for human IgG1, IgG3 and IgA, and is known to be required for sustaining the terminal differentiation of all Ig classes [12].

Here, we explore the translocation pathways required for soluble CD40L–IL-10 and TGF-β-induced IgA production in humans (irrespective of any antibody specificity) and address – in a cell culture model – the respective roles of the NF-κB and STAT3 pathways. Using a combination of blocking peptides to NF-κB subunits, we show that co-operation between NF-κB p65 and STAT3 activates downstream CD40 and IL-10-R, respectively, and is required for full IgA production. This occurs independently of IL-6 production by B cells. Our data help to define a novel role for IL-10-induced STAT3 in terminal B cell differentiation and in IgA production as a characteristic read-out of IL-10 signalling.

Materials and methods

B cell isolation

Buffy-coats were recovered from whole fresh blood from healthy volunteers who provided informed consent at the Auvergne-Loire Regional Blood Bank, as described previously [14]. Peripheral blood mononuclear cells (PBMC) were isolated by gradient density centrifugation using Histopaque-1077 (Sigma-Aldrich, Saint Quentin Fallavier, France). Total B cells were isolated with mixture of monoclonal antibodies towards CD2, CD3, CD7, CD14, CD16a, CD16b, CD36, CD43 and glycophorin A, using a B cell-negative isolation kit (Dynal; Invitrogen SARL, Cergy Pontoise, France) with a purity score ≥ 96% [14]. Allophycocyanin-conjugated CD19 monoclonal antibody (5 µg/10⁶ cells; clone HIB19; BD Biosciences, Le Pont de Claix, France) [22] and fluorescein isothiocyanate (FITC)-labelled anti-CD3 (clones SK7; BD Biosciences) were used to verify the purity before and after B cell isolation (Fig. 1a).

Fig. 1.

Determination of cell populations. (a) Example of B cell purity. Peripheral blood mononuclear cells (PBMC) were isolated from blood from healthy donors, as described in Materials and methods. Cells were stained with 7-aminoactinomycin D (7-AAD) (viability dye), anti-CD3-fluorescein isothiocyanate (FITC) (T cell marker) and anti-CD19-allophycocyanin (APC) (B cell marker) for 15 min at room temperature in the dark. The experience was repeated before and after B cell negative selection. (b) Example of naive B cell purity. PBMC were isolated from blood from healthy donors as described previously. Cells were stained with 7-AAD, anti-CD19-APC and anti-CD27-phycoerythrin (PE) (memory cell marker) for 15 min at room temperature in the dark. The staining was repeated before and after naive B cell negative selection. (c) Example of B cell population in blood. B cells were isolated from blood from healthy donors, as described in Materials and methods. Cells were stained with 7-AAD, anti-CD19-APC, anti-immunoglobulin (Ig)D or IgA or IgG or IgM-FITC and anti-CD27-PE for 15 min at room temperature in the dark. Data are representative of five independent experiments.

To characterize the enriched B cell populations, dead cells were excluded using 7-aminoactinomycin D (7-AAD) (BD Biosciences). Then, cells were labelled with anti-CD19-allophycocyanin (APC) (BD Biosciences) [22], anti-IgM-phycoerythrin (PE) or anti-IgD-FITC (clones G20-127 and IA6-2; BD Biosciences). To determine the percentage of memory IgA⁺, IgG⁺ or IgM⁺ B cells, CD19⁺ cells were stained with anti-CD27-PE plus anti-IgA, IgG or IgM-FITC (clones M-T271 and G20-359, G18-145 or G20-127; BD Biosciences). Labelling was analysed on a FACSCalibur (BD Biosciences) with FlowJo software (TreeStar Inc.). A total of 10⁴ events (CD19⁺ B cells) were recorded for each analysis.

For selected experiments, peripheral blood CD19⁺ B cells were magnetically sorted into enriched naive (CD27^-) or memory CD27⁺ B cells with CD27 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany) with a purity greater than 98% (Fig. 1b). The Raji B cell line (American Type Culture Collection, Manassas, VA, USA) was used for an experimental control.

B cell culture for IgA production

B cells were incubated at 37°C in a humidified atmosphere with 5% CO₂ for 12 days with human soluble trimeric CD40L (sCD40L, 0–200 ng/ml; Alexis-Coger, Paris, France), IL-10 (0–200 ng/ml) and/or TGF-β (0–2 ng/ml) [14,23,24]. To observe the role of IL-6, B cells were cultured with sCD40L (50 ng/ml) and IL-6 (5 ng/ml) in the presence or absence of IL-10 (100 ng/ml). All cytokines were from R&D Systems (Abingdon, UK). After stimulation with cytokines, B cells were washed with phosphate-buffered saline (PBS) containing 10% fetal bovine serum (FBS, endotoxin-free; Cambrex, Verviers, Belgium) and their phenotype was analysed by flow cytometry as described above. Cell-free supernatants were stored at −20°C until utilized. Using naive CD27^- B cells, we measured the level of Ig produced after CSR. In our experiments, the majority (90·5 ± 4·6%) of freshly isolated B cells were naive IgD⁺IgM⁺ B cells. In certain experiments, B cells were cultured for 120 min in supplemented Iscove's modified Dulbecco medium (IMDM). Blocking antibodies (5 µg/ml) against IL-6R, IL-10Rα and/or IL-10Rβ (clones 17506, 37607 and 90220, respectively; R&D Systems, Lille, France) were added with sCD40L and cytokines at the start of B cell culturing and monitored for 12 days. Binding of the IL-6R blocking antibodies on B cells was assessed by flow cytometry daily throughout the culture period (12 days, data not shown) [25].

Quantification of IL-6 and Ig in cell-free supernatants

IL-6 (48 h) and Ig total (12 days) levels in cell-free supernatants were quantified using a commercial specific enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems), according to the manufacturer's instructions [14,23,24].

ELISA plates (BD Biosciences) were coated with F(ab′)₂ of goat IgG anti-human IgA, IgG or IgM (33 ng/ml; MP Biomedical, Illkirch, France). After an overnight incubation at 4°C and four washes, plates were blocked for 60 min with PBS containing 1% bovine serum albumin (BSA). Supernatants at a 1:10 dilution were applied to the samples and incubated for 60 min at 37°C. After incubating for 45 min at 37°C, the plates were washed and bound Ig was detected with a horseradish-peroxidase (HRP)-labelled goat F(ab′)₂ IgG of anti-human IgA, IgG or IgM (Sigma-Aldrich). After four washes, O-phenylendiamine dihydrochloride (Sigma-Aldrich) was added and the plates were incubated at room temperature in the dark for 20 min. The reaction was stopped by addition of 1 M HCl (Sigma-Aldrich).

Transcriptional analysis of STAT3 and NF-κB

Purified B cells were incubated for 30 min, as described previously [26], with 50 ng/ml of sCD40L and 100 ng/ml of IL-10, with or without 5 ng/ml of IL-6. The cells were then washed with PBS–FBS (Cambrex) and treated with a nuclear extraction kit (Active Motif, Rixenart, Belgium), according to the manufacturer's instructions. Cytoplasmic and nuclear extracts were obtained for each condition and were stored at −80°C until used.

The levels of phosphorylated NF-κB p65 (pNF-κB p65, assay sensitivity = 0·5 µg/well) and phosphorylated STAT3 (pSTAT3, assay sensitivity = 0·6 µg/well) in the nuclear extracts of stimulated and non-stimulated B cells from each cell culture condition was determined using a transcription factor ELISA kit (active motif). Briefly, 2·5 µg of each nuclear extract was incubated in 96-well plates coated with a consensus sequence nucleotide binding site for pNF-κB p65 (5′-GGGACTTTCC-3′) or for pSTAT3 (5′-TTCCCGGAA-3′). Binding to the target oligonucleotide was detected by incubation with a primary antibody specific for the activated form of p65 or STAT3 and was visualized by incubation with an anti-IgG–HRP conjugate (diluted to optimal concentrations) and a developing solution, as instructed by the manufacturer. For each ELISA, the optical density was determined at 450 nm [optical density (OD)₄₅₀] using an ELISA reader (Multiskan EX; Labsystem, VWR International, Strasbourg, France), normalized with blanks and standards for each ELISA run.

As a control, the levels of pNF-κB or pSTAT3 were determined by Western blotting. Twenty-five µg of nuclear extract per well were separated by 10% acrylamide gel (Sigma-Aldrich) and transferred to a 0·45 µm nitrocellulose membrane (Amersham Pharmacia, Orsay, France) by electroblotting using transfer buffer supplemented with 20% methanol (Sigma-Aldrich). Membranes were blocked overnight at 4°C in PBS/0·1% Tween 20/1% BSA (I.D. Bio, Limoges, France) and incubated with a primary antibody to pNF-κB (0·4 µg/ml; Santa Cruz Biotechnology, Montrouge, France) or to pSTAT3 (0·4 µg/ml; Santa Cruz Biotechnology) for 90 min at room temperature. Thereafter, the membranes were washed three times for 10 min with blocking buffer then incubated for an additional 90 min with the secondary HRP-linked goat anti-rabbit antibody diluted to 1:5000 (Santa Cruz Biotechnology). Then, membranes were incubated with a chemiluminescent substrate according to the manufacturer's instructions (ECL; Amersham Pharmacia) and finally exposed to radiographic film (Sigma-Aldrich).

STAT3 and NF-κB studies

Purified B cells or PBMC were cultured at 1·0 × 10⁶ cells/ml and 2·0 × 10⁷ cells/ml, respectively, in IMDM (Sigma-Aldrich), supplemented as described previously [14]. The PBMC were tested to ascertain their viability and functionality after the addition of blocking peptides against pNF-κB p50 (Merck Chemicals Ltd, Nottingham, UK), pNF-κB p65 (one from Biosciences, San Diego, CA, USA and one from Santa Cruz Biotechnology, Montrouge, France) and/or pSTAT3 (one from eBiosciences, San Diego and one from Santa Cruz Biotechnology, Montrouge). The in vitro toxicity of these peptides was determined from the number of viable cells remaining after staining with the viability dye XTT (Sigma-Aldrich). To determine the optimal concentration and exposure time, for blocking peptides used against pNF-κB p50, pNF-κB p65 or pSTAT3, required to trigger B cell production of IgA, PBMC were stimulated in the presence or absence of these blocking peptides (0–10 µg/ml) at various time-points (from 0 to 240 min) prior to 12 days of cell culture.

B cell mRNA and AID investigation

Purified naive CD27^- B cells were stimulated with 50 ng/ml sCD40L and/or 100 ng/ml IL-10 for 4 days, washed with supplemented IMDM and the mRNA or DNA (positive control) was isolated using mRNA (Sigma-Aldrich) or DNA extraction kits following the manufacturer's instructions (Epicentre, Le Perray en Yvelines, France). Messenger RNA was extracted from 10⁶ B cells using the GenElute direct mRNA isolation kit (Sigma-Aldrich). One µg of the mRNA was reverse-transcribed into cDNA with a master mix of oligo-dT (20 µg/ml, Roche, Meylan, France), deoxyribonucleotide (dNTP) (16 µmol/ml; Invitrogen), RNase block (20 U/ml; Stratagene, Amsterdam, the Netherlands) and reverse transcriptase (50 U/ml; Invitrogen). The cDNA was then PCR-amplified with β-actin housekeeping gene-specific primers (R&D Systems) designed to amplify a portion of the coding sequences (7·5 pmol/µl), dNTP (8 µmol/ml) and Taq polymerase (1·25 U/ml; Sigma-Aldrich). Raji B cells were used as positive amplification controls and a master mix without added cDNA was used as a negative control. The cDNA expression was detected on a 1·5% agarose gel. The final product of the β-actin housekeeping gene was 298 base pairs (bp) in size.

To analyse AID gene expression, a nested reverse transcription–polymerase chain reaction (RT–PCR) assay was used. We selected the conserved active site of cytidine deaminase as the primary target. Primers were designed as follows: external 5′ GAAGAGGCGTGACAGTGCT 3′ (sense) and 5′ CGAAATGCGTCTCGT AAGT 3′ (anti-sense); internal 5′ CCTTTTCACTGGACTTTGG 3′ (sense) and 5′ TGATGGCTATTTGCACCCC 3′ (anti-sense). The final product of the AID gene was 656 bp in size [27]. Quantification of band intensity was carried out by Image J version 1·42q software (National Institutes of Health, Bethesda, MD, USA) and expressed as the mean of the optical density of five independent blots ± standard error of the mean (s.e.m.). Band intensity was normalized to the optical density of the actin-β housekeeping control loaded onto the same blot.

Statistical analysis

Interexperimental comparisons of the cell culture conditions were analysed by a Mann–Whitney unpaired test. Differences were considered statistically significant for P < 0·05.

Results

In-vitro IgA production by purified blood B cells

The peripheral blood of normal healthy donors (n = 15) showed large variation in the frequencies of the peripheral B cell subsets (Fig. 1c), with 68·3 ± 8·9% IgD⁺CD27^-, 11·5 ± 5·2% IgD⁺CD27⁺ and 22·9 ± 7·8% IgD^-CD27⁺ B cells. The IgD^-CD27⁺ B cells population could be subdivided further into 13·1 ± 3·2% IgD^-CD27⁺IgG⁺ or IgD^-CD27⁺IgA⁺ and 9·8 ± 3·6% IgD^-CD27⁺IgM⁺ B cells.

The optimal concentration of activators in this culture system required a balance between the best readout (IgA synthesis determined by ELISA) and B cell pathway activation (determined by Western blot).

In agreement with previously published culture conditions, we selected the concentrations of 50 ng/ml for sCD40L, 100 ng/ml for IL-10 and 0·2 ng/ml for TGF-β. Although sCD40L or IL-10 alone increased IgA production significantly by approximately 10-fold and approximately 30-fold, respectively, IgA production after the simultaneous addition of sCD40L and IL-10 was statistically similar to that observed with addition of IL-10 alone (Fig. 2a). An additive effect was observed for IgA production when sCD40L was used at 50 ng/ml and IL-10 from 80 to 120 ng/ml (Fig. 2b). Of note, the consistent IgA production seen with addition of IL-10 alone was attributable to memory B cells that are more sensitive to stimulation by IL-10 [28].

Fig. 2.

Immunoglobulin (Ig)A production by purified B cells stimulated with sCD40L, interleukin (IL)-10 and/or transforming growth factor (TGF)-β. (a) B cells were cultured with or without sCD40L (50 ng/ml), IL-10 (100 ng/ml) and/or TGF-β (0·2 ng/ml) for 12 days. (b) B cells were cultured with sCD40L (50 ng/ml) and IL-10 (20–120 ng/ml). IgA was then measured in cell-free supernatants by enzyme-linked immunosorbent assay (ELISA). P < 0·05 (*culture condition versus control: medium only) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

Although TGF-β can mediate B cell production of IgA in vitro in general, TGF-β alone under the present culture conditions did not alter B cell differentiation, nor did it augment the sCD40L- or IL-10-mediated IgA induction. Rather, IgA production induced by sCD40L and IL-10 was reduced significantly, albeit slightly, by addition of TGF-β (20·93 ± 6·09 µg/ml versus 34·71 ± 7·17 µg/ml, P < 0·05, Fig. 2a). Therefore, TGF-β was not used further in this study in addition to sCD40L and IL-10 as a differentiation/switch factor to induce B cell IgA production.

The respective roles of pNF-κB and pSTAT3 in IL-10 and sCD40L-stimulated B cells in driving IgA production

Next, we examined if our culture conditions engaged the intracellular phosphorylation of the classical NF-κB (Fig. 3a) and STAT3 (Fig. 3b) pathways. We used ELISA to detect pNF-κB p65 and pSTAT3 in nuclear extracts from B cells stimulated with sCD40L (50 ng/ml) and/or IL-10 (100 ng/ml) for 30 min. The sCD40L + IL-10 combination and, to a lesser extent, sCD40L alone, increased the pNF-κB p65 levels significantly in cultured B cells. IL-10 alone gave no signal over the control (Fig. 3a). In sharp contrast, sCD40L addition gave no signal over control signal for STAT3 phosphorylation, of which IL-10 was shown to be a powerful stimulator. No significant gain in pSTAT levels was observed when IL-10 was combined with sCD40L (Fig. 3b). Thus, in the in vitro conditions that initiate purified human blood B cell differentiation into IgA-secreting cells, sCD40L was able to induce the phosphorylation of NF-κB p65 but not of STAT3, while IL-10 induced the phosphorylation of STAT3 but not of NF-κB p65. Whereas sCD40L and IL-10 did not increase IgA production levels synergistically compared to sCD40L or IL-10 alone (Fig. 2a), IL-10 clearly increased CD40L-mediated activation of NF-κB p65 (Fig. 3a).

Fig. 3.

Detection of phosphorylated nuclear factor (pNF)-κB p65 (a) and phosphorylated signal transducer and activator of transcription 3 (pSTAT3) (b) in B cells stimulated with sCD40L and/or interleukin (IL)-10. Purified B cells were cultured with or without activator (50 ng/ml of sCD40L; 100 ng/ml of IL-10) for 30 min. pNF-κB p65 and pSTAT3 were detected by enzyme-linked immunosorbent assay (ELISA) in the cell nuclear extracts. P < 0·05 (*culture condition versus control: medium only) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

IL-10 exerts a direct IL-6-independent role in IgA synthesis

IL-6 has long been considered to be involved in Ig (particularly IgA) production [29]. Recently, IL-6 was also found to be one the main cytokines that is capable of inducing phosphorylation of STAT3 [30]. Moreover, IL-6 is released quickly by B cells after activation. We then asked whether IL-6 could behave as a mediator between IL-10 signalling and STAT3 phosphorylation. We hypothesize that IL-10 (through IL-10R) induces IL-6 release from B cells. This IL-6 could then be recaptured by B cells (through IL-6R) and activates STAT3. To test whether the IL-10-driven activation of the STAT3 pathway is direct or indirect, we measured both B cell production of IL-6 and IgA and also STAT3 phosphorylation in the presence or absence of IL-6R or IL-10R blocking antibodies.

B cells were incubated with IL-6R or IL-10R blocking antibodies for 120 min and were then stimulated by IL-6 or IL-10 for 30 min. The level of STAT3 phosphorylation was measured by ELISA (Fig. 4a). In the absence of inhibitors, both IL-6 and IL-10 significantly induced STAT3 phosphorylation. When IL-10R was specifically blocked, the STAT3 phosphorylation of IL-10-stimulated B cells decreased significantly. Conversely, blocking IL-6R did not alter the level of STAT3 phosphorylation in B cells incubated with IL-10, indicating that it did not rely on IL-6 production, as also indicated by measuring IgA level by ELISA (Fig. 4b). IL-6 increased IgA production by approximately twofold compared to untreated cells and IL-10 increased IgA production by more than 10-fold. Addition of the IL-10R blocking antibody to IL-10-treated B cells significantly decreased IgA production to nearly baseline levels, whereas the addition of the IL-6R blocking antibody did not affect IgA production. Moreover, when B cells were incubated for 120 min with blocking peptides against pNF-κB p65 and/or pSTAT3 and then stimulated with sCD40L and IL-10, the additional IgA production following stimulation was unaffected by blocking IL-6R (data not shown).

Fig. 4.

Role of interleukin (IL)-6 in the phosphorylated signal transducer and activator of transcription 3 (STAT3) pathway and in immunoglobulin (Ig)A production by purified B cells. Purified B cells were incubated with or without 5 µg/ml of blocking antibodies against the subunits α and β of IL-10R (anti-IL-10R) or against IL-6R (anti-IL-6R) for 120 min. Then B cells were stimulated with IL-6 (5 ng/ml) or IL-10 (100 ng/ml). (a) After 30 min, pSTAT3 was detected by enzyme-linked immunosorbent assay (ELISA) in cell nuclear extracts. (b) After incubation for 12 days, supernatant IgA levels were measured by ELISA. P < 0·05 (*culture condition versus control: medium only) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

B cells were also incubated with an IL-6R blocking antibody to rule out instantaneous binding (recapture) of released IL-6 to IL-6R. B cells were stimulated with sCD40L alone, IL-10 alone or sCD40L + IL-10 for 0–60 min and then IL-6 production by stimulated B cells was assayed by ELISA. IL-6 was not detected in any of the B cell cultures after 1–2 days (data not shown). We therefore conclude that IL-10 has a direct role in IgA production without an IL-6 shift and that IL-6 does not play an essential role in CD40L–IL-10-driven IgA production.

Inhibition of IgA production by blocking NF-κB p65 or STAT3 pathways

PBMC were stimulated in the presence or absence of blocking peptides against pNF-κB p65 and/or pSTAT3 at various concentrations (0–10 µg/ml; Fig. 5a) before initiation of the 12-day culture experiments. IgA ELISAs were performed to identify the optimal concentration for each peptide. IgA synthesis decreased in parallel with increased concentrations of blocking peptide against pNF-κB p65 and/or pSTAT3, with the lowest IgA level being observed at a concentration of 5 µg/ml.

Fig. 5.

Effects of phosphorylated nuclear factor (pNF)-κB p65 and phosphorylated signal transducer and activator of transcription 3 (pSTAT3) blocking peptides on immunoglobulin (Ig)A production by peripheral blood mononuclear cells (PBMC). (a) Blocking peptides against pNF-κB p65 and/or pSTAT3 (0–10 µg/ml) were added separately or together in PBMC cultures for 240 min. After incubation for 12 days, IgA were measured in cell-free supernatants by enzyme-linked immunosorbent assay (ELISA). (b) Blocking peptides against pNF-κB p65 and/or pSTAT3 (5 µg/ml) were added separately or together in PBMC cultures in a time-dependent manner. After incubation for 12 days, IgA were measured in cell-free supernatants by ELISA. P < 0·05 (*culture condition versus control: medium only) using the Mann–Whitney test. Data represent the mean (± standard error of the mean) of five experiments.

Next, PBMC were stimulated in the presence or absence of the same blocking peptides against pNF-κB p65 and/or pSTAT3 (5 µg/ml) at various time-points (0–240 min; Fig. 5b) before initiation of the 12-day culture experiments. IgA synthesis decreased in parallel with longer incubation times of blocking peptide against pNF-κB p65 and/or pSTAT3, with the lowest IgA level being observed at an exposure time of 120 min.

The pNF-κB p50 blocking peptide was tested under similar conditions and was not shown to be associated with a significant decrease in IgA synthesis at any of the blocking peptide concentrations tested (data not shown). Inhibition of IgA production was not due to in vitro toxicity of the blocking peptides against pNF-κB p50 or pNF-κB p65 or pSTAT3, as determined by counting the viable cells after 120 min of exposure to XTT during the 12 days of culture (Materials and methods, data not shown).

In this set of experiments, we used PBMC in order to determine the optimal concentration and incubation time for the inhibitory peptides. Because large numbers of cells were required for these assays, we chose to use PBMC for practical and economic reasons. However, to be sure that isolated B cells do not exhibit a different sensitivity to the blocking peptides, we ran the IgA and XTT assays for the optimal conditions only. The results were not different using PBMC or B cells.

Inhibition of AID expression by blocking NF-κB p65 or STAT3 pathways

Because AID is required for CSR, we examined the impact of either NF-κB p65 or the STAT3 pathways on the transcription of AID. Transcript levels for AID in naive B cells were measured by RT–PCR before or after culturing with sCD40L, IL-10 or sCD40L and IL-10. Messenger RNA encoding for AID was not observed in unstimulated naive B cells (Fig. 6a). AID transcript production was induced optimally by addition of sCD40L and IL-10 compared to the other cell culture conditions examined here in terms of signal-enhancing ability. Blocking the NF-κB or STAT3 pathways by incubating the cells for 120 min with blocking peptides (5 µg/ml) against pNF-κB p65 and/or pSTAT3 suppressed AID induction. Thus, blocking either the NF-κB p65 or the STAT3 pathway profoundly altered the production of mRNA for AID, an enzyme strictly necessary for CSR [31]. Transcript levels for AID were higher in the presence of sCD40L, IL-10 and sCD40L + IL-10 cell culture conditions (Fig. 6b).

Fig. 6.

Effects of nuclear factor (NF)-κB and signal transducer and activator of transcription 3 (STAT3) pathway inhibitors on activation-induced cytidine deaminase (AID) expression. (a) Purified, CD27^- naive B cells were incubated with or without blocking peptides against the phosphorylated (p)NF-κB p65 [α-phosphorylated (p)NF-κB] and/or pSTAT3 (α-pSTAT3) proteins (5 µg/ml) for 120 min, followed by stimulation with sCD40L (50 ng/ml) and IL-10 (100 ng/ml). After 4 days, mRNA and DNA transcripts were extracted, amplified by nested polymerase chain reaction (PCR) and detected on a 1·5% agarose gel. Raji B cells are used as positive control. This gel is representative of five independent experiments. (b) Fold-increase of AID expression in the sCD40L (50 ng/ml) and/or IL-10 (100 ng/ml)-treated samples relative to control samples. The intensity of the bands in (a) was quantified and the values from AID expression bands were normalized to those from the actin bands. P < 0·05 (*culture condition versus baseline) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

Both pNF-κB p65 and pSTAT3 are central in IgA production in in-vitro B cell culture conditions

Because the blocking peptides against pNF-κB p65 and pSTAT3 blocked AID transcription and IgA production in vitro, we next examined the impact of these peptides on IgG and IgM expression on B cells. First, we examined the B cell switch after 3, 4 and 5 days of incubation in the presence of the blocking peptides against pNF-κB p65 and pSTAT3 and activators (sCD40L + IL-10). The discrete B cell populations (IgD⁺, IgM⁺, IgA⁺, IgG⁺ or CD27⁺) were examined by flow cytometry for their individual sensitivity to the blocking peptides (Fig. 7a). Non-viable cells were excluded from the data shown by selective gating on 7-amino-actinomycin D (7AAD)-negative cells. IgM expression on B cells was not affected by the activators (sCD40L + IL-10); in contrast, IgA, IgG and CD27 expression increased by addition of the activators (Fig. 7b).

Fig. 7.

Role of nuclear factor (NF)-κB and signal transducer and activator of transcription 3 (STAT3) pathway inhibitors on purified B cell populations. B cells were incubated with or without blocking peptides against the phosphorylated (p)NF-κB p65, pNF-κB p50 and/or phosphorylated (p)STAT3 proteins (5 µg/ml) for 120 min, followed by stimulation with sCD40L (50 ng/ml) and interleukin (IL)-10 (100 ng/ml) for 30 min. B cells were stained with 7-aminoactinomycin D (7-AAD) (viability dye), anti-CD19-allophycocyanin (APC) and anti-immunoglobulin (Ig)A, IgG, IgM-fluorescein isothiocyanate (FITC) or anti-CD27-phycoerythrin (PE) for 15 min at room temperature in the dark. (a) Representative cytographs obtained in flow cytometry analysis. (b) CD27 and Ig expression on B cells with or without stimulation by sCD40L and IL-10. (c) IgA expression on B cells with or without blocking peptides followed by stimulation with sCD40L and IL-10. (d) IgG expression on B cells with or without blocking peptides followed by stimulation with sCD40L and IL-10. P < 0·05 (*culture condition versus baseline) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

Although the activators induced CSR towards IgA (and for control – towards IgG in short-term cultures), only the IgA⁺ population was affected by the blocking peptides against pNF-κB p65 and pSTAT3 (Fig. 7c); this population was decreased significantly in frequency (42·645 ± 0·295 % versus 14·04 ± 0·65 %; P < 0·05) by the inhibitors which caused a return to the baseline level. In addition, we observed that the blocking peptides against pNF-κB p50 decreased IgG expression, while anti-pSTAT3 did not seem to have an effect in this experimental model (Fig. 7d).

Incubation of purified blood B cells with blocking peptides against pNF-κB p65 or pSTAT3 (5 µg/ml, 120 min) induced a significant decrease in IgA production compared to the baseline level (Fig. 8a). IgA production, stimulated by sCD40L + IL-10, was reduced by 89% when the classical NF-κB pathway was blocked, by 98% when the STAT3 pathway was blocked and by approximately 100% when both pathways were blocked simultaneously. This last phenomenon was also observed when twofold, fourfold or eightfold lower concentrations of blocking peptides against pNF-κB p65 or pSTAT3 were used (data not shown).

Fig. 8.

Effect of the nuclear factor (NF)-κB and signal transducer and activator of transcription 3 (STAT3) pathways in immunoglobulin (Ig)A production in purified B cells. Cells were cultured with or without blocking peptides against the NF-κB p65 (α-NF-κB) or STAT3 (α-STAT3) pathways (5 µg/ml) for 120 min, followed by incubation with sCD40L (50 ng/ml) and IL-10 (100 ng/ml). (a) After 12 days, IgA levels in supernatants were measured by enzyme-linked immunosorbent assay (ELISA). (b) After 30 min, phosphorylated (p)NF-κB p65 was detected by ELISA in cell nuclear extracts. (c) After 30 min, phosphorylated (p)STAT3 was detected by ELISA in cell nuclear extracts. P < 0·05 (*culture activation condition versus inhibition) using the Mann–Whitney U-test. Data represent the mean (± standard error of the mean) of five experiments.

To assess the roles of NF-κB p65 and STAT3 in the later processes of cell differentiation (i.e. the final production of Ig), we sought to stimulate purified blood B cells with sCD40L + IL-10 while simultaneously blocking either one or both of the transcription pathways using specific blocking peptides against pNF-κB p65 or pSTAT3. The pNF-κB p65 blocking peptide led to a modest, but significant, 20% decrease in pNF-κB p65. The anti-pSTAT3 peptide alone had nearly the same effect, resulting in an 18% reduction in pNF-κB p65. Together, the blocking peptides against pNF-κB p65 and pSTAT3 reduced NF-κB p65 phosphorylation by 28% (Fig. 8b). Reciprocally, the anti-pSTAT3 peptide significantly reduced pSTAT3 by 45% (Fig. 8c), while the anti-pNF-κB p65 peptide reduced it by 30%. Combined, these blocking peptides reduced pSTAT3 by 73%. IgA production was completely inhibited; however, phosphorylation of NF-κB and STAT3 was not blocked completely. These observations were probably due to neo-phosphorylation induced by other stimuli or by the oscillations in NF-κB signalling, as could have been expected [32]. These data indicate that there is probably co-operation between the various transcription factor pathways, and in particular, an NF-κB influence on the STAT3 pathway. Furthermore, these results suggest that sCD40L acts first on purified B cells, promptly activating the classical NF-κB pathway and inducing IL-10R expression (experiments and data not shown), which then renders the STAT3 pathway reactive to IL-10 signalling.

Discussion

We aimed to elucidate some of the molecular pathways involved in providing purified B lymphocytes with the differentiation signals of non-cognate T cell surrogates, i.e. the classical sCD40L/CD40 + IL-10/IL-10R signals, leading to the skewed production of Ig towards IgA. We deliberately excluded from this investigation the addition of exogenous TGF-β, described classically as an IgA differentiation factor in a number of studies, on the basis of preliminary experiments (Fig. 2a and data not shown), having shown that TGF-β antagonized the differentiating role of sCD40L and IL-10 towards IgA class switch in this culture system. However, because these experiments were performed initially by culturing purified B lymphocytes in FCS-containing medium, the possibility that TGF-β eventually present in this serum may have biased our results was considered, as has been described, e.g. for the plasticity of T helper 17 (Th17) responses [33].

TGF-β1 induces IgA switching and secretion in stimulated B lymphocytes in mouse spleen. This has also been shown for IgG2b using mouse spleen B cells. For these reasons, TGF-P has been included with cytokines such as IL-4 and IFN-γ which promote selective Ig switching and therefore the ability to enhance particular antibody responses [34]. IL-4 is also a dominant cytokine which facilitates the IgA [35–37], but this point is still controversial. Although IL-4 definitely plays a role in mucosal immunity in Th2 responses, it was shown to be non-essential in mucosal IgA responses [38]. Secondly, in a mucosal context, one study reported than IL-4 is able to make IgA-positive cells switch to IgE-positive cells [39], which could have distorted our study. Thirdly, another study on PBMC stimulated with anti-CD40 monoclonal antibodies (mAb) showed that IL-4 and IL-10 co-operate, inducing a synergistic increase in IgA production only in IgA-deficient patients. Moreover, in a healthy subject group, the only cytokine able to significantly induce IgA production alone was IL-10 [37]. Moreover, while IL-4 and IL-21 increased the generation of IgG1⁽⁺⁾ cells synergistically from CD40L-stimulated B cells, IL-4 concomitantly abolished IL-21-induced switching to IgA [40].

Our primary interest was to determine the respective roles of STAT3, assumed to be activated directly by IL-10 and also of NF-κB, influenced by CD40L-ligation, with respect to the CSR of genes encoding IgA. A subsidiary interest was to eventually question the role of IL-6, a cytokine reported to affect STAT3 phosphorylation and reported to be instrumental in Ig production, that can be secreted via an endocrine pathway by activated/differentiated B cells [41].

To set up the conditions of the present study, we used blocking peptides against pNF-κB p65 and pSTAT3, which proved to efficiently block the NF-κB and STAT3 pathways for comparing IgA production in activated B cells. We found that these pathways were blocked more efficiently when anti-pNF-κB p65 and anti-pSTAT3 peptides (5 µg/ml) were incubated for 2 h with cells prior to long-term in vitro culture. Despite efficient inhibition of IgA production, we observed a difference between the inhibition of these two pathways and the inhibition of AID transcription, due probably to the low sensitivity of the AID assays. It remains that the sequence in which the CD40/CD40L stimuli are delivered to the B cell is still central to the outcome of terminal B cell differentiation into Ig-producing cells [14,42,43]. The cellular environment also appeared to play a substantial role in this process, as the presence of non-B cells (as with PBMC cultures) doubled the production of IgA compared to purified B cell cultures (unpublished data). This observation can be explained by the presence of our experimental model of monocyte-originating cytokines (e.g. IL-6 and IL-10) [44]; on one hand, it indicates the high level of complexity of cytokine intrications in B cell differentiation, and on the other hand a possible difference between effects mediated by purified cytokines and living-cell originating cytokines in ex vivo observations such as in this report. In all, our results confirmed suitable in vitro conditions in which to induce purified B cells to differentiate and produce IgA following exposure to a combination of sCD40L and IL-10 [45].

The present study next suggests that CD40 engagement, in the absence of other (known) stimuli, is sufficient to effectively induce IgA switching in human B cells, in a NF-κB-dependent manner [46]. IL-10 is the pleiotropic regulator of the immune system toward infection. It plays a central role in B cell proliferation, survival, isotype switching and differentiation [47]. Our results indeed confirm the involvement of IL-10 in IgA production; however, as IL-10 induced STAT3 and CD40L NF-κB, we next attempted to elucidate their respective influences on IgA production. The STAT3 protein is a STAT family member with diverse biological functions, including cell growth, cell survival, embryo development and cell motility [30,48,49]. STAT3 was shown to play a critical role in mouse B cell development, particularly in the thymodependent terminal differentiation of B cells into IgG plasma cells [50]. STAT3 was also identified recently as a major player in hyper-IgE syndrome [51]. Diehl et al. used human B cells to show that the inducible activation of STAT3 triggers blimp1 gene expression and promotes plasma cell differentiation and Ig production [52]. STAT3 and/or IL-10 mutations have been shown to be involved in inflammatory bowel disease, Crohn's disease or ulcerative colitis, impairing the signalling pathways [53]. STAT3 plays a major role in the IL-23/Th17 pathway, maintaining intestinal immune homeostasis [54]. However, it is becoming increasingly clear that IL-10 signalling appears to play a central role in inflammatory bowel disease pathogenesis, with germline variants associated with ulcerative colitis and Crohn's disease [55,56]. Here, we present evidence that the STAT3 pathway is also critical for either Ig (or more particularly IgA) production by human B cells or for export of IgA onto human B cells. Fan et al. showed that B cell stimulation by Ig triggering leads to STAT3 activation that depends on the combined effects of IL-6 and IL-10, whereas anti-Ig or pharmacological stimulation with phorbol myristate acetate (PMA)/ionomycin leads to STAT3 activation that depends primarily on IL-10 [57]. IL-10 also mediates the differentiation of germinal centre B cells into memory and plasma cells [58] and induces Janus kinase (JAK) proteins via the phosphorylation of STAT3 [59]. Here, we report that IL-10 by itself can lead to significant AID transcription and IgA production and that a combination of sCD40L and IL-10 induced comparable levels of IgA to those induced by IL-10 alone. Consequently, we propose that IgA synthesis by (in vitro) differentiated B cells is more dependent on the STAT3 pathway than on the NF-κB pathway. However, in the absence of IL-10 or when the STAT3 pathway is blocked, some IgA can still be produced by B cells, albeit in smaller quantities. IL-10 appears to substantially boost IgA production. The autocrine role of IL-10 in B cell differentiation was demonstrated further by the inhibitory effect of anti-IL-10 treatment on IgA secretion that was induced by the dual ligation of CD40 and antigen-receptor without alterations in cell growth [60].

Altogether, our experiments show that IL-10 directly activates the STAT3 pathway so that there is co-operation between the STAT3 pathway and the classical NF-κB pathway that is activated downstream of CD40 ligation (anti-pNF-κB p65 inhibited the STAT3 pathway and vice versa). Because blocking peptides to pNF-kB p50 did not interfere with IgA production, we suggest that p65 homodimers interact with pSTAT3 for enhancing/sustaining AID transcription and IgA production. As p50 does not possess a DNA binding motif, this complex would contain another Rel subunit to bind to κB motifs. It seems that complexes formed between p50 homodimers and STAT3 bind to GAS sites, whereas p65/STAT3 complexes bind to κB motifs, as was described previously in another model [18].

In this context, the NF-κB and STAT3 pathways affect each other via an unknown mechanism. It is plausible that after stimulation by IL-1 or IL-6 that STAT3 would form a complex with pNF-κB p65 to facilitate NF-κB binding to DNA [17]. However, we did not focus on IL-1 in this study because we found IL-1 to be unable to phosphorylate STAT3 (unpublished data and [26]). pSTAT3 is able to form a complex with unphosphorylated NF-κB dimers, which bind to κB sites [19]. Summarizing, we suggest that (i) CD40L stimulation induces pNF-κB dimers (interacting or not with unphosphorylated STAT3) to bind to κB sites, (ii) CD40L stimulation promotes IL-10R expression on the B cell surface, rendering STAT3 more reactive to IL-10 signalling and (iii) IL-10 stimulation induces pSTAT3 dimers to bind to GAS sites and pSTAT3 dimers interacting with unphosphorylated NF-κB to bind to κB sites. The fact that IL-10 induces the binding of dimers on both κB and GAS sites can account for the enhanced IgA production. Deciphering the machinery of IgA differentiation is valuable to mucosal immunology and vaccinology, as IgA represents the major protective barrier of mucosal surfaces. Immunological protection composed of a targeted, specific IgA response provided by either conventional or bioengineering vaccines, especially against invading microbes, may prove to be an achievable goal in the future.

Disclosure

The authors declared no conflict of interest.