Mice lacking p38γ, p38δ and p38γ/δ have been described (Sabio et al., 2005; Risco et al., 2012). All strains were backcrossed onto the C57BL/6 strain for at least nine generations. B cell-specific p38γ/δ deficient (CD19-Cre^KI/+p38γ/δ^f/f) mice were generated by breeding p38γ/δ^f/f to CD19-Cre^KI/+ mice (Rickert et al., 1997; Alsina-Beauchamp et al., 2018). Male and female 15–20 weeks old mice were used. Mice were housed in specific pathogen-free conditions in accordance with European Union regulations; work was approved by local CNB-CSIC ethical review and by the Bioethics Committee of the Community of Madrid PROEX316/15. Escherichia coli LPS was purchased from Sigma. Anti-p38α and TPL2 were from Santa Cruz. Antibodies to ERK1/2 and phospho-ERK1/2 (Thr202/Tyr204), Akt and phospho-Akt (Ser473, Thr408), phospho-NFκB1/p105 (Ser933; P-p105), phospho-p38MAPK (Thr180/Tyr182; this antibody recognises all four phosphorylated-p38 isoforms), IκBα and phospho-GSK3α/β (Ser21/9) were from Cell Signaling Technologies; anti-phospho-JNK1/2 (Thr183/Tyr185) was from Biosource. Anti-p38γ and -p38δ antibodies were raised and purified as described (Cuenda et al., 1997).

Thymus, spleen and lymph node cell suspensions were prepared; erythrocytes were lysed, and cells were counted. Cell samples were stained with combinations of fluorescently labelled antibodies to the cell surface markers CD19, CD3, CD4, CD8, CD43, CD25, Gr1, CD11b, B220, F4/80, IgD, CD21, CD23, CD69, CD86, CD95, GL7, PD-1 (all from BD Biosciences), CD138, CXCR5 (both from Biolegend) and IgM (Jackson Immunoresearch Lab.) for 30 min at 4°C. Cell analysis was performed in a FACScalibur, Beckman Coulter CYTOMIX FC500 MCL and LSR-II cytometer (BD Biosciences). Biotinylated goat anti-mouse IgG1, IgG2a, IgG2b, and IgG3 antibodies (Southern Biotech) were used to detect surface expression of IgG isotypes, followed by fluorescently labelled streptavidin (Molecular Probes). The profiles obtained were analysed with FlowJo (BD Biosciences) and Kaluza Analysis 2.11 (Beckman Coulter) software; B cells were gated as CD19⁺ cells when indicated.

Total splenocytes were obtained from freshly isolated spleens after tissue homogenisation and a Lympholyte step (Cedarlane Laboratories); residual erythrocytes were eliminated using erythrocyte lysis buffer (5 min, RT). For B cell isolation, total splenocytes were incubated with Dynabeads mouse pan-T (30 min, 4°C; Thy1.2, Dynal Biotech, Invitrogen) to eliminate the T cell fraction. The fraction enriched in B cells (>95% purity) was used for the in vitro experiments. Inguinal and popliteal lymph nodes were digested with collagenase-A plus DNAse-I (Roche; 15 min, 37°C), followed by homogenisation to isolate the lymphocyte compartment.

Purified B cells were stimulated for various times with anti-BCR (1 μg/ml) or LPS (2.5 μg/ml). Cells were lysed in lysis buffer [50 μM Tris-HCl pH 7.5, 1 μM EGTA, 1 mM EDTA, 0.15 M NaCl, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 50 mM sodium β-glycerophosphate, 5 mM pyrophosphate, 0.27 M sucrose, 0.1 mM phenylmethylsulphonyl fluoride, 1% (v/v) Triton X-100] plus 0.1% (v/v) 2-mercaptoethanol and complete proteinase inhibitor cocktail (Roche). Lysates were centrifuged (13,000 × g, 15 min, 4°C), supernatants removed, quick-frozen in liquid nitrogen, and stored at −80°C.

Protein samples were resolved in SDS-PAGE and transferred to nitrocellulose membranes, which were blocked (30 min) in 50 mM Tris/HCl (pH 7.5), 0.15 M NaCl, 0.05% (v/v) Tween (TBST buffer) containing 10% (w/v) non-fat dry milk, then incubated in TBST buffer with 10% (w/v) non-fat dry milk and 0.5–1 μg/ml antibody (2 h, room temperature or overnight, 4°C). Protein was detected using horseradish peroxidase-conjugated secondary antibodies and the enhanced chemiluminescence reagent (Amersham Pharmacia Biotech), using the Odyssey infrared imaging system.

Freshly isolated spleens were immersed in OCT and frozen with liquid nitrogen. Cryostat sections (10 μm) were fixed in 4% PFA [10 min, room temperature (RT)], blocked with PBS containing 1% BSA and 10% goat serum (30 min, RT), and stained with FITC-conjugated rat anti-mouse IgD (BD Bioscience), Cy5-goat anti-mouse IgM (Jackson Immunoresearch) and biotin-rat anti-mouse CD169/MOMA-1 antibody (Acris Antibodies) plus Cy3-streptavidin (Jackson Immunoresearch), at the appropriate dilution in PBS 1% BSA (45 min, RT); washes were done with PBS. Sections were then mounted in Fluoromount (Southern Biotech) and imaged on a Zeiss Axiovert LSM 510-META inverted microscope with 10x/air objective. Images were analysed using LSM 510 software (Zeiss).

Artificial planar lipid bilayers were prepared as previously described (Saez de Guinoa et al., 2011). In brief, unlabelled mouse GPI-linked ICAM-1-containing liposomes and/or liposomes containing biotinylated lipids were mixed with 1,2-dioleoyl-PC (DOPC) liposomes at various ratios to achieve specified molecular densities (ICAM-1 at 200 molecules/μm²; biotin-lipids, as indicated). Planar bilayers were assembled on sulphochromic solution-treated coverslips in FCS2 closed chambers (Bioptechs), then blocked with PBS 2% FCS (1 h, RT). For cell migration assays, ICAM-1-containing artificial bilayers were coated with recombinant murine CXCL13 (100 nM, 20 min, RT; Peprotech) immediately before imaging. For immune synapse formation assays, surrogate antigen (su-Ag) was anchored by incubating ICAM-1/biotin-lipids-containing membranes with AlexaFluor-647-streptavidin (Molecular Probes), followed by monobiotinylated rat anti-κ light chain mAb (BD Biosciences) (20 min, RT per incubation step). Monobiotinylation was obtained labelling the antibody with NHS-LC-LC-biotin at 1 μg/ml (30 min, RT, in PBS; Pierce), followed by dialysis and checked by flow cytometry using streptavidin-coated silica beads (5 μm-diameter; Bangs Laboratories). The number of GPI-ICAM-1 or anti-κ antibody molecules/μm² was estimated using an immunofluorimetric assay and anti-ICAM-1 or anti-rat IgG antibodies, respectively. The standard values were obtained from microbeads with various calibrated rat IgG-binding capacities (Bangs Laboratories).

Violet-tracer-labelled (see below) WT B cells mixed with unlabelled p38γ/δ-deficient B cells at 1:1 ratio (2 × 10⁶), or the other way around, were injected into the warmed chamber (37°C) and imaged. Confocal fluorescence (1 μm-optical section), differential interference contrast (DIC) and interference reflection microscopy (IRM) images were acquired; to monitor cell motility, images were taken every 30 s for 20 min. Planar bilayers assembly at FCS2 chambers and assays with cells were performed in chamber-buffer (PBS 0.5%FCS, 0.5 g/l D-glucose, 2 mM MgCl₂, 0.5 mM CaCl₂). Images were acquired on an Axiovert LSM 510-META inverted microscope with a 40X oil immersion objective (Zeiss), at 512 × 512 pixels quality. Imaris 7.0 software (Bitplane) was used for qualitative and quantitative analysis of cell dynamics parameters, fluorescence and IRM signals. The fluorescence signal of the planar bilayer in each case was set up as the background of fluorescence intensity. The frequency of adhesion (IRM⁺ cells) per imaged field was estimated as [n° of B cells showing IRM contact/total n° of B cells (estimated by DIC)] × 100; similarly, we calculated the frequency of immune synapse (i.e., B cells showing IRM contact and forming a detectable su-Ag central cluster) or the frequency of migration (i.e., B cells showing an IRM contact and moving over time).

Total splenocytes or purified B cells (3 × 10⁵) were cultured in flat-bottom p96 wells alone or with anti-BCR [purified F(ab’)₂ goat anti-mouse m heavy chain; Jackson Immunoresearch] at 1 or 10 μg/ml, hamster anti-mouse CD40 (clone HM40-3; BD Pharmingen) at 1 μg/ml or LPS (Sigma) at 2.5 μg/ml. After 24 h, cells were collected, stained for CD69, CD86, CD25, and CD19, and analysed by flow cytometry as described. For survival assays, purified B cells (3 × 10⁵) were cultured in flat-bottom p96 wells alone or with recombinant mouse BAFF (0.1 μg/ml; R&D Systems) and analysed for cell viability by flow cytometry after 24, 48, and 72 h; cell survival frequency was estimated by FSC/SSC morphological parameters. For proliferation assays, total splenocytes or purified B cells (5-10 × 10⁶ cells/ml) were labelled with 0.1 μM Violet Tracer (Molecular Probes) in PBS (10 min, 37°C); the labelling reaction was blocked with one volume of FCS (1 min, RT) and cells were washed with RPMI 10% FCS. Violet-labelled cells (3 × 10⁵) were cultured in flat-bottom p96 wells alone or with anti-BCR plus recombinant mouse IL-4 (20 ng/ml; Peprotech), anti-CD40 plus IL-4 or with LPS (Sigma) at the previously indicated doses. Cells were collected 96 h later, stained for CD19 for B cell identification, and analysed by flow cytometry. Assays were performed in RPMI 10% FCS. Proliferation index values were obtained from FlowJo analysis software (BD Biosciences); it is a ratio of the total number of divisions between the number of cells that went into division, both calculated at the time of analysis (96 h) in each experiment (see FlowJo website for more details).

For activation experiments with MZ and FO B cell subpopulations, spleens from WT and p38γδ^–/– mice were homogenised in PBS, treated with erythrocyte lysis buffer (5 min, RT), and then stained with FITC-conjugated rat anti-mouse CD21 and PE-conjugated rat anti-mouse CD23 (Biolegend; 30 min, 4°C). MZ (CD21^highCD23^low) and FO (CD21⁺CD23⁺) B cells were separated using a FACSAria Fusion flow cytometer and then cultured (5 × 10⁴) in round-bottom p96-wells plates in absence or presence of anti-BCR, anti-CD40 or LPS as described above. Cells were collected at 24 h, stained for CD69, CD86, and CD25, and analysed by flow cytometry.

Purified B cells (3 × 10⁵) from the spleen of WT or p38γ/δ-deficient mice were loaded on the top-reservoir of Boyden chambers (p24 transwell plates; 3 μm-diameter pore size; Costar) in 0.1 ml RPMI 10%FCS. In the bottom-reservoir, recombinant murine CXCL13 or CCL21 (10 μM; Peprotech) was added to 0.6 ml RPMI 10%FCS to get a final concentration of 400 and 100 nM, respectively. After 3 h in culture at 37°C, migratory B cells were collected from the bottom-well and counted by flow cytometry (1 min, at high flow rate). The migration frequency was estimated as [n° of migratory B cells at the bottom-well/n° of input B cells] × 100; the n° of input B cells was obtained by counting 3 × 10⁵ B cells in 0.6 ml of medium by flow cytometry. Assays were performed in duplicates.

Groups of four WT mice or mice deficient in p38γ, p38δ or both were immunised i.p. with NP-KLH (50 μg; Biosearch Technologies) in alum (100 μl; Thermo Scientific) or with TNP-Ficoll (50 μg; Biosearch Technologies), in a final volume of 200 μl. Similarly, CD19-Cre^KI/+ mice and B cell-specific p38γ/δ deficient (CD19-Cre^KI/+p38γ/δ^f/f) were immunised i.p. with NP-KLH in alum, as above. Peripheral blood samples were taken 1 day before immunisation, and at days 7, 14, 21, and 28 post-immunisation. Serum levels of NP- or TNP-specific antibodies of different isotypes (IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA) were evaluated by ELISA. Briefly, ELISA plates were coated with NP-BSA (ratio > 20) or TNP-BSA (both from Biosearch Technologies), blocked with PBS 2% BSA, and incubated with appropriate dilutions of serum samples (1:1000 for IgM and distinct IgG). Bound antigen-specific antibodies were developed with biotin-goat anti-mouse Ig isotype-specific antibodies (Southern Biotech) plus alkaline phosphatase (AP)-streptavidin (Sigma); finally, we added the colorimetric substrate p-nitrophenyl phosphate (Sigma) and measured absorbance at 405 nm wavelength. To measure the affinity maturation of the anti-NP antibodies generated in vivo, ELISA plates were coated with NP3-BSA (low NP ratio) or with NP30-BSA (high NP ratio) (both from Biosearch Technologies) and followed by the same steps as above. The increase in antibody affinity was estimated by the ratio between the NP3-derived absorbance value and NP30-derived value in each case and at each time. Higher NP3/NP30-ratio values over time indicated increased antibody affinity (affinity maturation).

For the analysis of germinal centre (GC) and plasma cells (PC) populations, groups of five WT mice or mice deficient in p38γ/δ were immunised i.p. with NP-KLH in alum as explained above; spleens and bone marrow of these animals were analysed at day 7 post-immunisation by flow cytometry. The following antibody mixes against surface markers were used: B220/IgD/CD95/GL7/CXCR4/ef780 for GC B cells; B220/PNA/CD19/IgD/CD38/ef780 for GC and PC in spleen; B220/IgG1/CD138/IgD/Sca-1/IgM/ef780 for PC in spleen and BM; CD3/CD4/CXCR5/PD1 for follicular T helper cells. The ef780 viability dye (eFluor-780, Invitrogen) was used for dead cell exclusion in the analysis, and anti-CD16/32 for blocking non-specific staining via FCγ receptors.

Purified B cells (1 × 10⁵) were cultured in round-bottom p96 wells alone or with anti-CD40 (1 mg/ml; clone HM40-3, BD Biosciences) plus IL-4 (20 ng/ml; Peprotech) or LPS (2.5 mg/ml; Sigma). After 96 h, supernatants were collected for antibody detection by ELISA; ELISA plates were coated with goat anti-κ light chain (1 μg/ml; Southern Biotech), blocked as before, incubated with appropriate dilutions of supernatants (1:100 for IgM; 1:10 for the distinct IgG), and developed as before.

Statistical data analysis between two groups was performed by applying the two-way ANOVA (immunisation analyses) and two-tailed unpaired Student’s t-test, using GraphPad Prism v. 7.0a. A p-value ≤ 0.05 was considered statistically significant.

We examined the implication of p38γ and p38δ in B cell development comparing different B cell subpopulations by cytometry in the bone marrow (BM) of wild type (WT) mice, p38γ and p38δ deficient (p38γ/δ^–/–) and B cell p38γ and p38δ deficient (CD19-Cre^KI/+-p38γ/δ^f/f). First, we checked, with specific antibodies, that WT B cells, but not B cells from CD19-Cre^KI/+-p38γ/δ^f/f or p38γ/δ^–/– expressed p38γ and p38δ (Supplementary Figure S1A). We also confirmed that in CD19-Cre^KI/+-p38γ/δ^f/f mice the expression of p38γ and p38δ was not affected in other tissues (Supplementary Figure S1B). The absolute cell numbers in BM were similar in p38γ/δ^–/– and WT mice, and in CD19-Cre^KI/+-p38γ/δ^f/f and p38γ/δ^f/f mice (Figure 1A). p38γ/δ^–/– lymphocytes differentiated into pro-B, pre-B, immature and mature B cells in the BM (Figures 1B–D). All p38γ/δ^–/–, CD19-Cre^KI/+-p38γ/δ^f/f, p38γ/δ^f/f and WT mice displayed similar frequency and absolute numbers of BM B cells (Figures 1C,D).

We next evaluated the impact of p38γ and p38δ deletion on peripheral B cell development. In the spleen, p38γ/δ^–/– and CD19-Cre^KI/+-p38γ/δ^f/f mice exhibited a decrease in total cell number compared to control mice (WT, p38γ/δ^f/f and CD19-Cre^KI/+) (Figure 2A and Supplementary Figure S2A), that in the case of p38γ/δ^–/– was of ∼40% and statistically significant (Figure 2A). Accordingly, spleens from p38γ/δ^–/– mice were perceptibly smaller than those from WT mice (Supplementary Figure S2B). The deletion of either p38γ or p38δ alone did not affect spleen total cell number indicating a redundant role of these two kinases (Supplementary Figure S2C). Flow cytometry analysis of the B cell population showed no significant differences in B cell frequency (Figure 2B); however, p38γ/δ^–/– mouse spleens showed lower absolute numbers of B cells than those from WT mice (Figure 2C). The alterations in frequency and absolute B cell numbers were not significant in p38γ- or p38δ-deficient mouse spleens (Supplementary Figures S2D,E). The diminished B cell number in the spleen of p38γ/δ^–/– mice was not due to impaired B cell-activating factor (BAFF) signalling, since the response to BAFF survival signal of p38γ/δ deficient B cells and WT B cells was similar (Figure 2D). In addition, absolute numbers, but not frequency, of other splenic immune cell populations from p38γ/δ^–/– mice were perceptibly smaller than those from WT mice (Table 1 and Supplementary Figure S2F). This is consistent with the decrease in total spleen cell number and indicates that the reduced B cell number in p38γ/δ^–/– mouse spleens is due to a decreased size of the organ.

Conventional mature B cells are classified into two subsets in the mouse, follicular (FO) and marginal zone (MZ). Recirculating FO B cells are usually found in lymphoid follicles of the spleen and lymph nodes, while MZ B cells reside primarily around the periphery of spleen lymphoid nodules. Using cell staining and flow cytometry techniques, we analysed the two main spleen B cell subpopulations separately, FO B cells (CD21⁺CD23⁺) and MZ B cells (CD21^highCD23^–) (Figure 2E). FO B cell frequency was not markedly affected by the absence of p38γ and p38δ in either the whole mouse or specifically in B cells (Figure 2F), whereas absolute numbers were lower in p38γ/δ^–/– mice as expected (Figure 2G). Both, MZ B cell frequency and absolute numbers were markedly reduced in p38γ/δ^–/– spleens compare to WT (Figures 2H,I); however, there were not differences between MZ B cell frequency and absolute numbers from CD19^KI/+-p38γ/δ^f/f and p38γ/δ^f/f mice (Figures 2H,I). The lower MZ B cell frequency observed in p38γ/δ^f/f mice compared to WT mice could be due to the reduced levels of p38γ and p38δ protein in those mice (Supplementary Figure S1). p38γ^–/– and p38δ^–/– mice also had lower MZ B cell frequencies and total cell number than WT mice, although the differences were not statistically significant (Supplementary Figures S2G,H). FO B cell frequency and absolute cell number were similar in p38γ^–/–, p38δ^–/– and WT mice (Supplementary Figures S2G,H). The MZ B cell compartment in p38γ/δ^–/– mice was almost half the size found in WT counterparts (Figure 2I), which implicated p38γ and p38δ in MZ B cell differentiation in spleen.

MZ B cells are located around the follicles in the spleen. To examine the white pulp structure and MZ B cell localisation, in situ, we performed H&E and immunofluorescence staining in cryostat sections of spleens from p38γ/δ^–/– and WT mice. Spleen architecture was maintained in p38γ/δ^–/– compared to WT mice, and white and red pulps were similar (Figure 2J). Staining with anti-IgM and -IgD revealed a normal B cell distribution in the white pulps of p38γ/δ^–/– spleen (Figure 2K). To visualise the MZ, we stained with anti-MOMA-1, which stains the metallophilic macrophages located at the boundary between the FO and the MZ. Despite the B cell deficiencies in p38γ/δ^–/– mice, we detected no alterations in the formation of B cell follicles (IgM⁺IgD⁺) or MZ B cell (IgM⁺IgD^–) location near MOMA-1⁺ metallophilic macrophages (Figure 2K).

Chemokine-triggered B cell migration is essential for the B cell response. It facilitates homing to secondary lymphoid organs and scanning of the follicular stromal cell network where B cells find the antigen. To evaluate if the diminished B cell numbers at spleen of p38γ/δ^–/– mice might be related with impaired migration and homing, we studied B cell motility in response to chemokine stimulation by two different approaches. We analysed the effect of the lack of p38γ and p38δ on the B cell response to CXCL13 or CCL21 chemokine gradients using Boyden chambers (see methods); both chemokines are involved in B cell migration in vivo. The chemotactic response of p38γ/δ^–/– B cells was equivalent to that of WT for both chemokines (Figure 3A). We next addressed adhesion-dependent B cell motility using a two-dimensional model based on artificial planar lipid bilayers, combined with real-time interference reflection microscopy (IRM) and confocal fluorescence microscopy; this model mimics in vivo B cell motility within the follicles (Saez de Guinoa et al., 2011). We isolated WT and p38γ/δ^–/– B cells and assayed their migratory capacity on CXCL13-coated, intracellular adhesion molecule (ICAM)-1-containing membranes. p38γ/p38δ deficiency did not seem to cause any impairment in either B cell migration or adhesion frequency (Figure 3B); moreover, both mean speed and directionality (measured as straightness index) were similar in p38γ/δ^–/– and WT splenic cells (Figure 3C).

BCR recognition of antigen leads to immune synapse (IS) formation at the B cell/antigen-presenting cell (APC) interface. The IS framework supports signalling and polarised membrane trafficking to reach B cell activation and antigen extraction from the APC. We evaluated the role of p38γ and p38δ on B cell IS formation by tethering su-Ag (anti-κ light chain antibody; su-Ag) to ICAM-1-containing artificial planar lipid bilayers. We monitored IS formation in the presence of different su-Ag density using IRM and confocal fluorescence microscopy, to measure the B cell contact with the artificial bilayer (adhesion; IS area) and su-Ag central clustering, respectively (Figure 3D). At 20 molecules/μm² su-Ag density, the frequency of B-cell adhesion and su-Ag clustering is almost 100%, whereas at 1 molecules/μm² su-Ag density, B-cell adhesion was approximately 60% and su-Ag central cluster was hardly detected (Figures 3D,E). These results were similar with those previously described (Saez de Guinoa et al., 2011). The frequency values of p38γ/δ^–/– B cells were equivalent to those for WT B cells (Figures 3D,E). Additionally, either IS contact areas with the planar bilayer, detected by IRM, or total su-Ag aggregation at the IS were not affected by the lack of p38γ/p38δ (Figure 3F). Overall, our data indicated that p38γ and p38δ are not involved in CXCL13/CCL21-triggered B cell migration or BCR-promoted IS formation.

Since p38γ and p38δ have a role in the regulation of immune cell signalling and gene expression (Cuenda and Sanz-Ezquerro, 2017) we analysed the responsiveness of B cells from WT and p38γ/δ^–/– mice to both T-independent and -dependent stimuli. B cell activation by anti-BCR, LPS or anti-CD40 leads to upregulation of several cell surface proteins, including CD69 and the costimulatory molecule CD86. We performed BCR and TLR4 stimulation experiments with total splenocytes or purified B cells, and obtained similar B cell activation and proliferative responses (data not shown); for practical reasons, the following assays were done using splenocytes, and B cells were gated as CD19⁺ cells. To determine the effect of p38γ and p38δ deficiency on CD69 and CD86 expression, we cultured splenocytes from WT or p38γ/δ^–/– mice alone or with different doses of soluble su-Ag [F(ab’)₂ anti-μ heavy chain antibodies] for B cells; 24 h later, we analysed B cells for surface expression of the activation markers. At a 1 μg/ml su-Ag dose, p38γ/δ^–/– B cell activation was reduced compared to WT B cells (Figures 4A–C); p38γ^–/– B cells also had BCR-triggering defects at this antigen dose, suggesting a prominent role for p38γ than for p38δ downstream of BCR signalling (Supplementary Figures S3A–C). We nonetheless observed no defects in BCR-triggered B cell activation at a higher su-Ag dose (10 μg/ml; Figures 4A–C). The results indicate that, above a BCR signalling threshold, B cells counteract the effect of p38γ and p38δ deficiency in early B cell activation events (CD69 and CD86 upregulation).

We then evaluated later events in B cell activation using in vitro proliferation assays. Violet tracer-labelled splenocytes were stimulated with two doses of su-Ag in the presence of IL-4, and analysed by flow cytometry 72 h later for violet tracer signal. CD19⁺ gated B cells from p38γ/δ^–/– mice proliferated significantly less than those from WT mice, independently of the antigen dose (Figures 4D,E). Neither p38γ^–/– nor p38δ^–/– B cells showed defects in BCR-triggered proliferation rates (Supplementary Figure S3D). These data indicated that both p38γ and p38δ participate in the BCR signalling cascade that drives B cell activation and proliferation. Whereas p38γ appears to be more important in early activation events (Supplementary Figures S3A–C), each isoform compensates for the other in antigen-mediated B cell proliferation; only the absence of both p38γ and p38δ decreased proliferation in response to su-Ag.

In in vitro activation assays using the TLR4 ligand LPS as a stimulus, we did not detect any significant variation in p38γ/δ^–/– B cell activation frequency (CD86⁺CD69⁺) compared to WT; CD69 levels were, however, reduced (Supplementary Figures S3E,F). Lack of p38γ alone also affected CD69 upregulation in LPS stimulated B cell activation, indicating the involvement of this isoform in LPS-induced B cell activation. To evaluate the need for p38γ and p38δ in TLR4-induced proliferation, we stimulated splenocytes with LPS. LPS-triggered CD19⁺-gated B cell proliferation was similar in p38γ/δ^–/–, p38γ^–/–, p38δ^–/– and WT B cells (Supplementary Figure S3G), indicating that in these conditions, p38γ and p38δ do not have a central role in the TLR4 signalling cascade that drives B cell proliferation.

CD40 activation of B cells causes the expression of CD25 as well as high levels of CD69 and the co-stimulatory molecule CD86. We did not find any significant variation in the expression of CD25 in p38γ/δ^–/– B cell compared to WT after CD40 stimulation (Figures 4F,G); CD69 and CD86 levels and CD86⁺CD69⁺ frequency were however reduced (Figures 4F,G), indicating the involvement of these p38MAPK isoforms in CD40-induced B cell activation. CD40-mediated proliferation was slightly, but significantly, reduced in p38γ/δ^–/– B cells (Figure 4H), suggesting that in these conditions, p38γ and p38δ regulate the CD40 signalling cascade that drives B cell proliferation.

Since the MZ B cell compartment in p38γ/δ^–/– spleens is reduced compared to WT (Figures 2H,I), we examined if the activation of either MZ B cells or FO B cells was affected by the lack of p38γ and p38δ. We flow-sorted MZ and FO B cells from both WT and p38γ/δ^–/– mice (Supplementary Figure S4A), and performed comparative analysis of CD69, CD86 and CD25 expression in response to different stimuli (anti-BCR, anti-CD40 and LPS) (Supplementary Figures S4B,C). We found that CD69, but not CD86 expression was decreased in both p38γ/δ^–/– MZ and FO B cells compared to WT MZ and FO B cells, whereas CD25 expression in response to anti-CD40 stimulation was not affected (Supplementary Figures S4B,C). These results indicate that p38γ and p38δ regulate the activation of both MZ B and FO B cells.

p38γ and p38δ regulate TLRs and Dectin-1 signalling in macrophages by impairing ERK1/2 pathway activation (Risco et al., 2012; Alsina-Beauchamp et al., 2018), which is central for cellular proliferation; we thus examined BCR and CD40 signalling using phosphospecific antibodies to detect activated ERK1/2 and also p38MAPK by immunoblot. Both BCR and CD40 stimulation led to ERK1/2 activation in WT cells, and lack of p38γ and p38δ impaired the activation of ERK1/2 induced by CD40, but not by anti-BCR (Figures 4I,J). CD40 activation of ERK1/2 is mediated by the upstream MEK kinase, TPL2 (Eliopoulos et al., 2003; Papoutsopoulou et al., 2006); p38γ/p38δ deletion decrease TPL2 protein levels in B cells (Figure 4I), therefore, it is possible that the impaired ERK1/2 activation in p38γ/δ^–/– B cell is due to the lack of the upstream kinase TPL2, similar to what happen in TLR- and Dectin-1-mediated ERK1/2 activation (Risco et al., 2012; Alsina-Beauchamp et al., 2018). In contrast, CD40-induced p38α, JNK and IKK pathways activation were not affected in p38γ/δ^–/– B cell (Figure 4I). This finding suggests a role for ERK1/2 pathway in the decreased proliferation in p38γ/p38δ-deficient B cells in response to CD40, not in response to BCR activation. Since Akt pathway is also implicated in B-cell proliferation and it is one of the main pathways implicated in BCR signalling (Rickert, 2013), we measured Akt activation in response to BCR, and found that the phosphorylation of Akt and its downstream target GSK3 were similar in WT and p38γ/δ^–/– B cell (Figure 4J). p38α phosphorylation was not impaired in p38γ/δ^–/– B cell in response to BCR activation (Figure 4J). These results suggest that, although p38γ and p38δ do not control the activation of ERK1/2 and Akt pathways in response to BCR, they could be directly regulating the activity of other downstream protein essential for B cell proliferation.

To explore the in vivo function of p38γ and p38δ in the antibody production, we immunised p38γ/δ^–/– and WT mice with either T-independent (TI) or T-dependent (TD) antigen. We monitored the humoral immune response to the TD antigen NP-KLH in alum and the TI antigen TNP-Ficoll measuring the levels of NP- or TNP-specific antibodies in the serum at different days post-immunisation by ELISA (Figure 5A and Supplementary Figure S5A). When mice were immunised with NP-KLH in alum, we found that p38γ/δ^–/– had a significantly lower IgG2 antibody response (IgG2a, 50% reduction; IgG2b, 25%) than WT mice. IgM, IgG1, IgG3 and IgA production were similar in both genotypes (Figure 5A). Analysis of the TI antibody response showed no major differences between the genotypes (Supplementary Figure S5A). These results suggest that p38γ and p38δ play a role in B cell antibody responses to T-dependent (TD), but not to T-independent (TI) antigens. The data also show that the reduction in the MZ B cell compartment in p38γ/δ^–/– mice has no major effect on antibody responses to T-independent antigens. Additionally, the ratio of high-affinity (NP₃)/low-affinity (NP₃₀) IgG2 antibodies was decreased in p38γ/δ^–/– compared to WT mice (Figure 5B), indicating that the production of high-affinity (NP₃) IgG2a and IgG2b antibodies were reduced in p38γ/δ^–/– mice and suggesting that p38γ and p38δ are positive regulators of IgG2 antibody production to TD antigens in vivo.

We extended our analysis and, upon immunisation with a TD antigen, examined: GC B cells (CD95⁺GL7⁺), centrocytes and centroblasts; spleen and BM plasma cells (PC) (B220^low CD138^high, IgM⁺ and IgG1⁺) and T follicular helper (Tfh) cells (CXCR5⁺PD-1⁺). We found that both p38γ/δ^–/– and WT mice generate similar frequency of GC B cells, PC and Tfh (Figures 5C–F). GC and PC analysis based on CD38 and PNA expression levels showed similar results (Supplementary Figure S6B).

To determine whether the specific IgG2a and IgG2b reduction in p38γ/δ^–/– mice resulted from an intrinsic B cell defect, we immunised CD19-Cre^KI/+-p38γ/δ^f/f and p38γ/δ^f/f mice with NP-KLH in alum and found that the production of IgM, IgG1, IgG2a, IgG2b, IgG3 and IgA antibodies was similar between genotypes (Supplementary Figure S5B). Only a slight effect was detected for IgG2b production in CD19-Cre^KI/+-p38γ/δ^f/f mice compare to controls; however, this was smaller than the those found in p38γ/δ^–/– mice (Supplementary Figure S5B). Additionally, we examined the ability of p38γ/δ^–/– B cells to class switch to IgG subclasses in vitro. In experiments using purified WT and p38γ/δ^–/– B cells with anti-CD40 plus IL-4 to mimic a TD response, we found no differences between these cells in class switch to the IgG isotypes analysed (Figure 5G). The data also suggest that the reduced in vivo antibody response to TD antigens is not B cell-autonomous. In vitro class switching experiments were also carried out using LPS to mimic a TI response and purified B cells; results were similar for WT and p38γ/δ^–/– B cells, with no detectable defects in class switch or antibody production due to lack of p38γ and p38δ (Figure 5H). The in vitro antibody production confirmed the in vivo results showing p38γ/δ^–/– B cell responses to T-independent antigen.