IKK-induced NF-κB1 p105 proteolysis is critical for B cell antibody responses to T cell–dependent antigen

To investigate the role of IKK-induced p105 proteolysis in B cell development and homeostasis, we analyzed B cell subpopulations in the BM, secondary lymphoid organs (spleen and lymph nodes), and peritoneal cavity of Nfkb1^SSAA/SSAA (Nfkb1^SSAA) and WT mice by flow cytometry. B cell development in the BM was similar between both genotypes, with similar proportions and absolute numbers of pro–B cells, pre–B cells, and immature B cells (Fig. 1 A and Fig. S1 A). The numbers of mature recirculating B cells in the BM were also unaffected by Nfkb1^SSAA mutation. Total numbers and proportions of B cells in spleen and peripheral lymph nodes, respectively, were equivalent in Nfkb1^SSAA/SSAA and WT mice (Fig. 1, B and C; and Fig. S1, B and C). Numbers of splenic FM B cells were also similar. In contrast, marginal zone (MZ) B cells number in the spleen were substantially reduced by Nfkb1^SSAA mutation, whereas the numbers of transitional type 1 (T1) and type 2 (T2) B cells were fractionally increased (Fig. 1 B). The proportion of B1a cells in the peritoneal cavity was also significantly reduced in Nfkb1^SSAA/SSAA mice compared with WT (Fig. 1 D and Fig. S1 D). However, the fractions of peritoneal B1b and B2 cells were similar between genotypes.

In summary, mutation of the IKK target serines on NF-κB1 p105 did not alter the overall number of peripheral mature B cells, in marked contrast to deletion of IKK2 or NEMO in the B cell lineage that results in substantially reduced numbers of such cells (Pasparakis et al., 2002). However, MZ B cell numbers were significantly reduced by Nfkb1^SSAA mutation, in line with the known importance of NF-κB for the production of this subset of B cells in the spleen (Pillai and Cariappa, 2009). Mixed BM chimeras demonstrated that the reduction in MZ B cells resulted from a B cell–autonomous effect of the Nfkb1^SSAA mutation and confirmed that Nfkb1^SSAA mutation did not affect FM B cell numbers (Fig. 1 E).

Measurement of steady-state levels of Ig in serum revealed significantly reduced IgM (Fig. 1 F), consistent with the reduction in MZ B cells in the spleen and B1a cells in the peritoneum (Zhang, 2013). Serum titers of IgA and both IgG1 and IgG2b were also significantly decreased in unimmunized mice, raising the possibility that one or several aspects of antibody responses might also be impaired by Nfkb1^SSAA mutation. Because we were interested in investigating the role of NF-κB1 p105 proteolysis in B cell function in immunity, we focused our subsequent experiments on the effects of Nfkb1^SSAA mutation on TD antibody responses. It has previously been shown that RBP-J–deficient mice, which lack MZ B cells, are still able to generate normal antibody responses to chicken gammaglobulin (CGG; Tanigaki et al., 2002). The reduction in MZ B cells in Nfkb1^SSAA/SSAA mice, therefore, should not contribute significantly to TD antibody responses, although this might decrease antibody responses to T cell–independent antigens. Consequently, any detected effects of Nfkb1^SSAA mutation on antibody responses to a TD antigen were likely to result from alterations in the function of FM B cells (Oracki et al., 2010), as their numbers were normal in Nfkb1^SSAA/SSAA mice (Fig. 1 B).

To investigate the B cell–intrinsic effects of Nfkb1^SSAA mutation on FM B cell function, we generated mixed BM chimeras. BM cells from μMT^−/− mice, lacking B cells (Kitamura et al., 1991), were mixed at 4:1 ratio with BM cells from WT or Nfkb1^SSAA/SSAA mice. These were then transferred into irradiated Rag1^−/− mice, producing chimeric mice in which all B cells had a WT (Nfkb1^+/+) or Nfkb1^SSAA/SSAA genotype, whereas 80% of the other hematopoietic cells had a WT genotype. After reconstitution for 8 wk, chimeric mice were then injected with NP₂₇-CGG precipitated in alum or alum-only control. Nfkb1^SSAA mutation totally blocked the production of serum anti-NP IgM and IgG1 at 7 and 14 d (Fig. 2 A), although the numbers of splenic B cells were equivalent between WT and Nfkb1^SSAA/SSAA chimeric mice (Fig. 2 B, left). Nfkb1^SSAA mutation also prevented the generation of anti–NP-specific plasmablasts (Fig. 2 B, right) and anti–NP-specific GC B cells (Fig. 2 C). Furthermore, immunofluorescence analysis of spleen sections demonstrated the absence of GC structures in Nfkb1^SSAA/SSAA chimeras after NP₂₇-CGG-alum immunization (Fig. 2 D), which is consistent with a reduction in total GC B cell numbers (Fig. 2 C, left). Analysis of allotype-marked mixed BM chimeras confirmed that the effect of Nfkb1^SSAA mutation on generation of anti-NP–specific plasmablasts was B cell intrinsic (Fig. 2 E).

The results in this section demonstrated an essential role for IKK-induced proteolysis of NF-κB1 p105 within FM B cells for their participation in TD antibody responses.

Our earlier work demonstrated that blockade of IKK-induced p105 proteolysis by Nfkb1^SSAA mutation can impair the activation of both TPL-2 and NF-κB in macrophages (Yang et al., 2012). Therefore, the inhibitory effects of Nfkb1^SSAA mutation on TD antibody responses could have resulted from reduced TPL-2 and/or NF-κB activation in B cells. To investigate whether TPL-2 had important B cell–intrinsic functions in TD antibody responses, mixed BM chimeras were generated in which B cells lacked expression of TPL-2 (Dumitru et al., 2000). Immunization of these chimeric mice with NP₂₇-CGG revealed that TPL-2–deficient B cells produced a TD antibody response equivalent to that of WT controls (Fig. 3). This suggested that the failure of Nfkb1^SSAA/SSAA FM B cells to mount an antibody response to NP₂₇-CGG did not result from defective TPL-2 signaling, and was likely due to altered NF-κB activation.

To investigate this possibility, we determined the effect of Nfkb1^SSAA mutation on NF-κB activation in purified splenic FM B cells in vitro. Steady-state protein levels of RelA, RelB, and c-Rel were comparable between Nfkb1^SSAA/SSAA and WT FM B cells (Fig. 4 A). However, similar to our previous findings with macrophages (Yang et al., 2012), Nfkb1^SSAA mutation increased the steady-state concentration of NF-κB1 p105, while NF-κB1 p50 was reduced. Nfkb1 mRNA levels were comparable between Nfkb1^SSAA/SSAA and WT FM B cells (unpublished data), implying that these changes in p105 and p50 protein levels resulted from Nfkb1^SSAA mutation blocking constitutive p105 phosphorylation by IKK. Consistent with that hypothesis, unstimulated WT FM B cells had small amounts of phosphorylated p105, that increased after stimulation with anti-IgM, CD40L and LPS (Fig. 4 B and not depicted). In addition, steady-state levels of NF-κB2 p52 were increased in Nfkb1^SSAA/SSAA B cells compared with WT, whereas levels of p100 were fractionally reduced (Fig. 4 A), suggesting that Nfkb1SSAA mutation increased constitutive activation of the NF-κB alternative pathway.

Nuclear translocations of RelA and p50, after stimulation with anti-IgM for 1 h, were clearly decreased in Nfkb1^SSAA/SSAA FM B cells compared with WT cells (Fig. 4 C). Accordingly, NF-κB ELISAs demonstrated that Nfkb1^SSAA mutation impaired BCR activation of p50 and RelA binding activity at 1 h (Fig. 4 D). Although nuclear translocation of c-Rel was not obviously affected after 1h BCR stimulation, c-Rel binding activity was fractionally reduced by Nfkb1^SSAA mutation (Fig. 4, C and D). However, Nfkb1^SSAA mutation had a more pronounced effect on the second wave of c-Rel activation that results from prolonged BCR stimulation (Damdinsuren et al., 2010), detected at 12 h (Fig. 4 D). Both p50 and RelA nuclear translocations and binding activities were reduced by Nfkb1^SSAA mutation after 4-h CD40L stimulation. In contrast, CD40L-induced nuclear translocations and NF-κB binding activities of c-Rel, RelB, and NF-κB2 p52 were similar between Nfkb1^SSAA/SSAA and WT FM B cells. Thus, the major effects of Nfkb1^SSAA mutation in FM B cells were to reduce the early activation of canonical NF-κB subunits p50 and RelA, after stimulation of either BCR or CD40, and the late activation of c-Rel after prolonged BCR cross-linking. p50, RelA and c-Rel each co-immunoprecipitated with p105, suggesting that Nfkb1^SSAA mutation reduced activation of these Rel subunits, in part, by p105SSAA retention in the cytoplasm (Fig. 4 E).

Optimal TD antibody responses depend on the clonal expansion and survival of antigen-stimulated B cells (Oracki et al., 2010). Furthermore, in vitro studies with purified splenic B cells from Rel-deficient mice have suggested that NF-κB promotes the cell growth, proliferation and survival of mature B cells after BCR stimulation (Gerondakis and Siebenlist, 2010a; Kaileh and Sen, 2012). As NF-κB activation was impaired by Nfkb1^SSAA mutation, we next compared the growth, proliferation and survival of Nfkb1^SSAA/SSAA FM B cells with WT cells. Nfkb1^SSAA mutation did not affect FM B cell growth (Fig. 5 A), and only modestly decreased the fraction of FM B cells that divided following anti-IgM stimulation (Fig. 5 B). However, Nfkb1^SSAA mutation prevented BCR ligation increasing FM B cell survival (Fig. 5 C). CD40-induced proliferation and survival were similar between Nfkb1^SSAA/SSAA and WT FM B cells (Fig. 5, B and C). The survival defect of Nfkb1^SSAA/SSAA FM B cells after BCR cross-linking was rescued by co-stimulation with CD40L (Fig. 5 D).

NF-κB was previously shown to maintain the viability of BCR-activated B cells by inducing the expression of Bcl-xl and A1 (Bfl-1) genes (Grumont et al., 1999; Köntgen et al., 1995). Expression of Mcl-1, which is essential for survival of mature B cells (Hettmann et al., 2003) and plasma cells (Peperzak et al., 2013), may also be regulated by NF-κB (Buggins et al., 2010). To gain further insight into how IKK-induced p105 proteolysis regulates B cell survival, we determined the expression of these genes in purified FM B cells from Nfkb1^SSAA/SSAA and WT FM mice. Nfkb1^SSAA mutation significantly reduced BCR up-regulation of Bcl-xl, A1, and Mcl-1 mRNAs (Fig. 5 E). CD40 induction of Bcl-xl and Mcl-1 mRNAs was also reduced in Nfkb1^SSAA/SSAA FM B cells compared with WT. In contrast, CD40 up-regulation of A1 mRNA was similar between genotypes, which may explain why CD40-induced FM B cell survival was not altered by Nfkb1^SSAA mutation.

Together these results suggested that failure of the BCR to up-regulate Bcl-2 family proteins in Nfkb1^SSAA/SSAA FM B cells increased apoptosis, and might explain the block in TD antibody responses caused by Nfkb1^SSAA mutation. To investigate this hypothesis, we crossed Nfkb1^SSAA/SSAA mice with a transgenic mouse strain that overexpresses Bcl-xl in the B cell lineage (Takahashi et al., 1999). In vitro experiments demonstrated that Bcl-xl normalized Nfkb1^SSAA/SSAA FM B cell survival after BCR stimulation (Fig. 6 A). Nevertheless, chimeric mice in which the B cell compartment was generated from the BM of Bcl-xl Nfkb1^SSAA/SSAA mice were still unable to mount an IgG1 antibody response to NP₂₇-CGG, while the IgM anti-NP response was substantially reduced compared with WT Bcl-xl chimera controls (Fig. 6 B). The recovery of transplanted FM B cells from MD4 Nfkb1^SSAA/SSAA mice, which express a transgenic BCR specific for HEL (hen egg lysozyme; Goodnow et al., 1988), was increased by Bcl-xl expression (Fig. 6 C). However, MD4 Bcl-xl Nfkb1^SSAA/SSAA B cells did not differentiate into plasmablasts or GC B cells after immunization with HEL-SRBC, in contrast to MD4 Bcl-xl Nfkb1^+/+ control B cells under the same conditions (Fig. 6 C and Fig. S2). Thus, Nfkb1^SSAA mutation did not block TD antibody responses solely by reducing BCR-induced FM B cell survival.

The antigen-specific interaction between B cells and helper T cells in secondary lymphoid organs is an essential step in TD humoral immune responses (Cyster, 2005). T cells provide a second signal via CD40 ligand, which promotes B cell survival and proliferation, and prevents B cell paralysis following BCR stimulation (Goodnow et al., 2010). Our observation that Nfkb1^SSAA mutation did not affect CD40L induction of FM B cell proliferation or survival (Fig. 5, B and C), and that CD40 co-stimulation rescued the BCR-induced survival defect of Nfkb1^SSAA/SSAA FM B cells in vitro (Fig. 5 D) raised the possibility that these cells failed to receive a signal from CD4⁺ T cells after NP₂₇-CGG immunization.

After binding antigen, efficient FM B cell interaction with CD4⁺ T cells is facilitated by BCR up-regulation of CCR7, which induces FM B cell migration to the B cell zone-T cell zone boundary of the follicle in response to the T zone-expressed CCR7 ligand, CCL21 (Pereira et al., 2010). CCR7 is an NF-κB–regulated gene (Krappmann et al., 2004), but Nfkb1^SSAA mutation did not impair the up-regulation of CCR7 expression on FM B cells following BCR ligation (Fig. 7 A). Furthermore, anti-IgM-induced mRNA expression of Epstein-Barr virus-induced molecule 2 (EBI-2), another NF-κB–regulated gene (Glynne et al., 2000) encoding a cytokine which facilitates the initial uniform distribution of B cells along the B/T border (Gatto et al., 2009), was also unaffected by Nfkb1^SSAA mutation (Fig. 7 B). Consistent with these data, anti-IgM-stimulated CD45.2⁺ Nfkb1^SSAA/SSAA FM B cells migrated to the follicle-T cell zone border after transfer into wild-type CD45.1⁺ recipients, similar to their WT counterparts (Fig. 7 C). These results suggest that a failure to co-localize with T cells did not explain the defective function of Nfkb1^SSAA/SSAA B cells in TD antibody responses. However, we cannot exclude some contribution of the NF-κB1 signaling pathway in this process, as in vitro stimulation before cell transfer may not completely replicate the in vivo environment.

For productive T cell-B cell cooperation in TD antibody responses, B cells must capture external antigens and present them as peptide fragments on MHC class II molecules to CD4⁺ T cells (Yuseff et al., 2013). It was therefore possible that while Nfkb1^SSAA mutation did not inhibit T cell–B cell co-localization, the antigen presenting function of FM B cells was impaired. A model system was used to assay this (Huntington et al., 2006), in which OVA was targeted to the BCR by treating Nfkb1^SSAA/SSAA and WT FM B cells sequentially with biotinylated anti-IgM and antibiotin OVA-FITC. Antigen-pulsed B cells were then co-cultured with CTV-labeled OTII CD4⁺ T cells, which recognize OVA-derived peptides presented on I-Ab class II molecules. Flow cytometric analysis of OTII CD4⁺ T cell proliferation indicated that Nfkb1^SSAA/SSAA FM B cells were able to process and present OVA peptides in an active process requiring viable cells as efficiently as WT cells (Fig. 7 D). In accordance with these data, BCR up-regulation of CD86 and MHC class II was equivalent between Nfkb1^SSAA/SSAA and WT FM B cells (unpublished data). The absence of a TD antibody response in Nfkb1^SSAA/SSAA chimeric mice, therefore, did not result from a block in the ability of FM B cells to present antigen.

After antigen exposure and T cell help, proliferating B cells differentiate into either short-lived extrafollicular plasma cells or GC B cells, which subsequently generate plasma cells that secrete high-affinity antibody and persist for a lifetime. Although GC did not develop after NP₂₇-CGG immunization of Nfkb1^SSAA/SSAA chimeric mice (Fig. 2 C), the lack of anti-NP IgM and IgG1 at 7d suggested that the inhibitory effect of Nfkb1^SSAA mutation might precede GC formation. The complete lack of NP₂₇-specific IgM plasma cells in Nfkb1^SSAA/SSAA chimeric mice raised the possibility that Nfkb1SSAA mutation might affect plasmablast differentiation of FM B cells.

Plasma cell differentiation depends on high expression of the transcription factors BLIMP-1 (B lymphocyte-induced maturation protein 1) and IRF4 (interferon regulatory factor 4). The transcription of Prdm1 (encoding BLIMP-1) and Irf4 genes is regulated by NF-κB (Grumont and Gerondakis, 2000; Saito et al., 2007; Morgan et al., 2009), although the role of NF-κB activation in plasma cell differentiation has not been investigated. The failure to generate antigen-specific plasmablasts in Nfkb1^SSAA/SSAA chimeric mice suggested that Nfkb1SSAA mutation might prevent antigen-stimulated FM B cells differentiating into plasmablasts by affecting the expression of these transcription factors. To initially investigate this, qRT-PCR was used to determine Prdm1 and Irf4 mRNA levels in cultured FM B cells. We found that Nfkb1^SSAA mutation significantly reduced the mRNA levels of both genes induced by stimulation with anti-IgM (Fig. 8 A). LPS induction of Irf4 mRNA was also decreased in Nfkb1^SSAA/SSAA FM B cells compared with WT controls, which correlated with a defect in LPS activation of NF-κB (unpublished data). In accordance with these latter data, flow cytometric analyses demonstrated that Nfkb1^SSAA mutation prevented FM B cells (from mice transgenic for Bcl-xl to maintain cell survival) from up-regulating IRF4 protein and differentiating into B220^loCD138⁺ plasmablasts secreting IgM after stimulation with LPS plus IL-4 (Fig. 8, B and C). Nfkb1^SSAA mutation also prevented LPS plus IL-4 induced switching to IgG1 (Fig. 8 D), consistent with the critical role of IRF4 in class switch recombination (Sciammas et al., 2006). Importantly, the inhibitory effect of Nfkb1^SSAA/SSAA mutation on plasmablast differentiation was confirmed in vivo. Purified FM B cells from MD4 Bcl-xl Nfkb1^SSAA/SSAA mice did not differentiate into plasmablasts after transfer to WT recipient mice and concomitant immunization with HEL-SRBC, which contrasted markedly with the MD4 Bcl-xl Nfkb1^+/+ control B cells under the same conditions (Fig. 6 C and Fig. S2). This defect in MD4 Bcl-xl Nfkb1^SSAA/SSAA B cells correlated with impaired up-regulation of IRF4 (Fig. 8 E and Fig. S3 A) and down-regulation of IgD (Fig. S3 A). Transient expression of Irf4 is essential for up-regulation of Bcl6 and generation of GC B cells (Ochiai et al., 2013). In line with this, Nfkb1^SSAA mutation blocked antigen-induced up-regulation of BCL6 (Fig. 8 F and Fig. S3 B) and the formation of GC B cells (Fig. 6 C and Fig. S2 B).

To determine whether the reduction of Irf4 and/or Prdm1 mRNA levels caused by Nfkb1^SSAA mutation was responsible for the block in plasma cell differentiation, WT and Nfkb1^SSAA/SSAA FM B cells were transduced with recombinant retroviruses encoding BLIMP-1 or IRF4. Restoring IRF4 expression increased the fraction of Nfkb1^SSAA/SSAA FM B cells that differentiated into plasmablasts sixfold after restimulation with LPS and IL-4 in vitro, whereas BLIMP-1 expression had little effect (Fig. 8 G). Furthermore, retroviral expression of IRF4 in purified FM B cells from MD4 Bcl-xl Nfkb1^SSAA/SSAA mice rescued antigen-induced generation of GC B cells after transfer to WT recipient mice (Fig. 8 H and Fig. S4). Insufficient plasmablasts were generated in chimeric mice for quantitation. Together, these results suggested that the inhibitory effect of Nfkb1^SSAA mutation on TD antibody responses resulted partly from reduced BCR-mediated up-regulation of Irf4 mRNA, which blocked differentiation of FM B cells to plasmablasts and GC B cells.

Our in vitro experiments indicated that a major effect of Nfkb1^SSAA mutation with respect to NF-κB was to reduce the steady-state levels and activation of NF-κB1 p50 in FM B cells. To investigate the role of reduced p50 levels in the function of Nfkb1^SSAA/SSAA FM B cells, Nfkb1^SSAA/SSAA mice were crossed with Nfkb1^ΔCT/ΔCT mice, which express elevated amounts of p50 compared with WT cells and no p105 (Ishikawa et al., 1998), as described previously (Yang et al., 2012). For comparison with these Nfkb1^SSAA/ΔCT mice, Nfbk1^+/− and Nfkb1^SSAA/− strains were also generated. Nfkb1^SSAA/ΔCT FM B cells expressed higher amounts of p50 compared with Nfkb1^+/− cells, whereas p50 levels were very low in Nfkb1^SSAA/− FM B cells (Fig. 9 A). The steady-state concentrations of p105SSAA were similar between Nfkb1^SSAA/− and Nfkb1^SSAA/ΔCT FM B cells. NF-κB ELISAs revealed reduced BCR activation of p50, RelA, and c-Rel in Nfkb1^SSAA/− FM B cells compared with Nfkb1^+/− cells (Fig. 9 B). However, increasing p50 levels with the Nfkb1DCT allele normalized BCR activation of all three Rel subunits.

BCR-induced survival was reduced in Nfkb1^SSAA/− FM B cells in comparison with Nfbk1^+/− controls (Fig. 9 C), as expected. In contrast, survival of Nfkb1^SSAA/ΔCT FM B cells after BCR cross-linking was similar to Nfbk1^+/− controls. BCR-induced proliferation was similar between the three genotypes (unpublished data), in line with the small inhibitory effect of homozygous Nfkb1^SSAA mutation on anti-IgM–induced cell division (Fig. 5 B). Nfkb1^SSAA mutation reduced both survival and proliferation of FM B cells after TLR4 stimulation, and these inhibitory effects were rescued by the Nfkb1^ΔCT allele (Fig. 9 C and not depicted). The impairment by Nfkb1^SSAA mutation of IRF4 up-regulation, differentiation to CD138⁺ B220^lo plasmablasts and switching to IgG1 after LPS and IL-4 stimulation of FM B cells were also overcome by the Nfkb1^ΔCT allele (Fig. 9, D–F). These results showed that the reduced survival, plasmablast differentiation and class switch recombination of FM B cells in vitro caused by Nfkb1^SSAA mutation were rescued by increasing p50 levels with the Nfkb1^ΔCT allele, which normalized NF-κB activation.

Finally, to investigate the importance of p50 in the inhibitory effect of Nfkb1^SSAA mutation on TD antibody responses, we generated mixed BM chimeras and then immunized with NP₂₇-CGG. The titers of anti-NP IgM and IgG1 in the serum, as well as the number of anti-NP IgG1 plasma cells and GC B cells in the spleens of chimeras reconstituted with Nfkb1^SSAA/− BM, were significantly reduced compared with Nfbk1^+/− control chimeras (Fig. 9 G and not depicted). In contrast, no significant differences were detected between chimeras generated with Nfbk1^+/− and Nfkb1^SSAA/ΔCT BM cells. Therefore, rescuing NF-κB activation by increasing p50 levels with the Nfkb1DCT allele overcame the inhibitory effect of Nfkb1^SSAA mutation on the FM B cell antibody response to a TD antigen.

Analyses of conditional knockout mouse strains lacking IKK2 or NEMO expression in the B cell lineage have demonstrated that canonical NF-κB activation is critical for mature B cell generation or homeostasis (Pasparakis et al., 2002; Li et al., 2003). However, this has prevented using such mouse strains for investigation of the potential role for canonical NF-κB activation in B cells for antibody production during an immune response. In the present study, we used Nfkb1^SSAA/SSAA mice, in which the development and homeostasis of peripheral FM B cells were normal, to investigate the role of NF-κB activation in FM B cells in humoral immunity. Nfkb1^SSAA mutation selectively inhibits the IKK signaling pathway that regulates NF-κB activation by inducing NF-κB1 p105 proteolysis (Sriskantharajah et al., 2009). Our results demonstrate for the first time that IKK-induced p105 proteolysis is required for optimal BCR activation of NF-κB in B cells, and that this is essential for generation of a TD antibody response, specifically regulating the survival of FM B cells and their differentiation to plasmablasts and GC B cells after BCR stimulation.

Nfkb1^SSAA mutation reduced activation of p50 and RelA after both BCR and CD40 stimulation. Activation of c-Rel was also significantly reduced in Nfkb1^SSAA/SSAA FM B cells compared with WT, particularly after prolonged BCR stimulation. Immunoprecipitation and immunoblotting demonstrated that p50, RelA, and c-Rel were able to interact with p105. The inhibitory effect of Nfkb1^SSAA mutation on the activation of these Rel subunits, therefore, probably resulted from their retention in the cytoplasm by p105^SSAA. In addition, steady-state levels of p50 were reduced in Nfkb1^SSAA/SSAA FM B cells compared with WT controls, suggesting that Nfkb1^SSAA mutation reduced p105 processing to p50, similar to the situation we previously described in macrophages (Yang et al., 2012). The Nfkb1^ΔCT allele overcame the inhibitory effect of Nfkb1^SSAA mutation on BCR activation of p50, RelA, and c-Rel, presumably by increasing p50 expression to a degree that the concentration of p50 exceeded p105 (p105:p50 [Nfkb1^SSAA/ΔCT] = 0.63 ± 0.03; p105:p50 [Nfkb1^SSAA/−] = 4.04 ± 0.3). Consequently, p105^SSAA became saturated with p50 monomers and homodimers, with which it preferentially binds (Moorthy and Ghosh, 2003), and no longer able to block BCR activation of RelA and c-Rel monomers and p50/RelA and p50/c-Rel heterodimers. Remarkably, the Nfkb1^ΔCT allele also normalized FM B cell survival and plasmablast differentiation in vitro, and TD antibody responses in vivo, demonstrating that the inhibitory effects of Nfkb1^SSAA mutation on FM B cell function resulted from impaired NF-κB activation.

Nfkb1^SSAA mutation did not affect RelB and p52 NF-κB–binding activity in CD40L-stimulated cells, presumably because these subunits are predominantly associated with NF-κB2 p100 (Solan et al., 2002). Inability to inhibit activation of the alternative NF-κB pathway probably explains why Nfkb1^SSAA mutation did not affect A1 mRNA up-regulation and the survival of FM B cells after CD40 stimulation, which activates both canonical and alternative NF-κB signaling pathways (Rickert et al., 2011). CD40 co-stimulation overcame the BCR-induced survival defect of Nfkb1^SSAA/SSAA FM B cells in vitro. However, although these cells could co-localize and presumably interact with CD40 on T cells in the spleen after BCR cross-linking, they still failed to mount an antibody response to NP₂₇-CGG. This suggests that defective survival alone did not block the TD antibody response of Nfkb1^SSAA/SSAA FM B cells. In line with this, transgenic overexpression of Bcl-xl, which restored in vitro BCR-induced survival of Nfkb1^SSAA/SSAA FM B cells to normal, did not rescue antibody responses to a TD antigen.

Nfkb1^SSAA mutation blocked the late activation of c-Rel after BCR cross-linking with anti-IgM, and impaired BCR induction of Irf4 mRNA, a c-Rel–regulated gene (Grumont and Gerondakis, 2000). The up-regulation of IRF4 protein expression in MD4 B cells by HEL antigen in vivo was also prevented by Nfkb1^SSAA mutation, confirming the physiological significance of these findings. IRF4 is a critical transcription factor for the induction of plasma cell differentiation and Ig class switch recombination (Sciammas et al., 2006; De Silva et al., 2012). Consistent with this, Nfkb1^SSAA mutation reduced the ability of LPS plus IL-4 to induce IRF4 expression in FM B cells, impairing differentiation to IgM-secreting plasmablasts and switching to IgG1. The induction of plasmablast differentiation of MD4 B cells in vivo by HEL was also prevented by Nfkb1SSAA mutation. In addition, retroviral expression of IRF4 rescued plasma cell differentiation of Nfkb1^SSAA/SSAAv FM B cells stimulated with LPS and IL-4, confirming a functional effect of reduced IRF4 in these cells. These experiments establish Irf4 as a physiologically important transcriptional target of the NF-κB1 signaling pathway in B cell antibody responses, and demonstrate for the first time the critical role of NF-κB activation in plasmablast differentiation.

CD19-Cre Irf4^fl/fl mice have a substantially reduced GC response, and residual GCs form only with B cells in which the Irf4 allele has not been deleted (Ochiai et al., 2013). IRF4, therefore, plays an essential and B cell autonomous role in GC formation, regulating the expression of the key GC transcription factor Bcl6 (Goodnow et al., 2010; Ochiai et al., 2013). Antigen-stimulated Nfkb1^SSAA/SSAA FM B cells did not up-regulate Bcl6 or differentiate into GC B cells during a TD immune response. NF-κB signaling activity is very low in GC B cell centroblasts (Shaffer et al., 2001; Basso et al., 2004). However, IRF4 is required for the initiation, but not maintenance of the GC state (Klein et al., 2006; Ochiai et al., 2013). This suggests that Nfkb1^SSAA mutation may have prevented the initiation of GC formation by inhibition of NF-κB–dependent transcriptional up-regulation of Irf4. Consistent with this hypothesis, retroviral expression of IRF4 significantly increased antigen-induced differentiation of MD4 Bcl-xl Nfkb1^SSAA/SSAA B cells. The failure to induce IRF4 in Nfkb1^SSAA/SSAA B cells was a feature of agonist of the BCR and TLR4, but not CD40 (unpublished data). As TLR signaling is not required for antibody responses to antigens delivered in alum (Gavin et al., 2006), this suggests that the immune response defects detected were primarily caused by impaired BCR-mediated signaling.

In this study, we investigated the role of IKK-induced NF-κB1 p105 proteolysis in mature B cells, using Nfkb1^SSAA/SSAA mice in which this IKK signaling pathway is blocked (Sriskantharajah et al., 2009). Nfkb1^SSAA mutation prevented FM B cells from mounting an antibody response to a TD antigen by selectively blocking BCR-induction of BCL-2 family proteins and IRF4, which impaired FM B cell survival and antigen-induced differentiation of FM B cells to plasmablasts and GC B cells, respectively. Our study demonstrates that the regulation of NF-κB activation in mature B cells by IKK-induced NF-κB1 p105 proteolysis is essential for TD humoral immune responses.

Mouse strains were bred in the specific pathogen–free animal facility of the National Institute for Medical Research (NIMR; London, UK). 7–12-wk-old mice were used for all experiments. Rag1^−/− (Mombaerts et al., 1992), μMT^−/− (Kitamura et al., 1991), and HEL-specific BCR transgenic mice (MD4; Mason et al., 1992) and Nfkb1^SSAA/SSAA (Sriskantharajah et al., 2009), Nfkb1^−/− (Sha et al., 1995), and Map3k8^−/− (Dumitru et al., 2000) mouse strains have been previously described and were all fully backcrossed on to a C57BL/6 background. Nfkb1^+/−, Nfkb1^SSAA/- and Nfkb1^SSAA/ΔCT (Ishikawa et al., 1998) mice were generated by intercrossing Nfkb1^SSAA/SSAA, Nfkb1^{ΔCT/ ΔCT}, and Nfkb1^−/− strains. Nfkb1^SSAA/SSAA.Bcl-xL and Nfkb1^+/+.Bcl-xL mice experimental mice were generated by intercrossing Bcl-xL transgenic mice (Takahashi et al., 1999) with Nfkb1^SSAA/SSAA and C57BL/6 mouse strains, respectively. These were subsequently intercrossed with MD4.Rag1^−/− mice to generate MD4.Rag1^+/−.Nfkb1^SSAA/SSAA.Bcl-xL and MD4.Rag1^+/−.Nfkb1^+/+.Bcl-xL mice.

This study was conducted following authorization by the UK Home Office, under relevant Project Licence authority.

BM cells were harvested from femora and tibiae of the indicated donor mice, and resuspended in ACK lysis buffer (155 mM NH₄Cl, 10 mM KHCO₃, 100 µM EDTA) to lyse red blood cells. Cells (4–5 × 10⁶ per mouse) were then transferred by intravenous injection into sublethally irradiated Rag1^−/− mice (500 rads; Cs source). For competition experiments, BM cell suspensions from CD45.1⁺ C57BL/6 were mixed with either Nfkb1^SSAA/SSAA or wild-type Nfkb1^+/+ CD45.2⁺ BM cells, at a 1:1 ratio. For mixed BM chimeras, sublethally irradiated Rag1^−/− mice were reconstituted with Nfkb1^SSAA/SSAA or wild-type Nfkb1^+/+ BM cells mixed at 1:4 ratio with μMT^−/− BM cells. Chimeric mice were immunized 8 wk after reconstitution.

To analyze B cell positioning in the spleen after BCR stimulation, 8–10 × 10⁶ splenic purified B cells from either Nfkb1^SSAA/SSAA or wild-type mice were cultured in vitro with or without anti-IgM (10 µg/ml; Jackson ImmunoResearch Laboratories) for 1 h (Pereira et al., 2009). Cells were then transferred into CD45.1⁺ C57BL/6 hosts by intravenous injection, and spleens removed 5.5 h later for immunofluorescence analysis.

To monitor plasma cell differentiation and GC B cell formation in vivo, 8–10 × 10⁶ splenic purified B cells from MD4.Rag1^+/−.Nfkb1^SSAA/SSAA.BcL-xL and MD4.Rag1^+/−.Nfkb1^+/+.Bcl-xL were cotransferred with 0.5–1.10⁹ HEL-conjugated or control sheep red blood cells into CD45.1⁺ C57BL/6 hosts by intravenous injection (Reif et al., 2002; Kelly et al., 2011). 6 d after transfer, recipient mice were culled and spleens were removed for flow cytometric analysis.

For TD antibody responses, mixed BM chimeras were immunized i.p. with 50 µg NP₂₇-CGG in PBS (Biosearch Technologies) precipitated in alum (Imject; Thermo Fisher Scientific; 3:1 ratio). Alum mixed with PBS was used as a control. The production of NP-specific and total immunoglobulins (IgM and IgG1) was monitored by ELISA, using 96-well Maxisorp ELISA plates pretreated with 50 µl/well of 5 µg/ml NP_14.-BSA (Biosearch Technologies). Antibody isotypes and subclasses were detected with horseradish peroxidase–conjugated mouse IgM and subclass IgG-specific antibodies (Southern Biotech).

Cryosections (7–10 µM) of spleen were prepared, dried, and fixed with acetone before immunofluorescent analysis (Turner et al., 1997). GC B cells were stained using GL7-FITC mAb (Ly77, BD), revealed with rabbit anti-FITC (Alexa Fluor 488 Signal-Amplification kit; Invitrogen), IgD-PE (11-26c; eBioscience), and B220-Alexa Fluor 647 (RA3-6B2, BD). For B cell migration assays, CD45.2-FITC (104, eBioscience), CD4-Alexa Fluor 647 (ABD Serotec), and IgD-PE (11-26c; eBioscience) were used to identify transferred B cells. Spleen sections were analyzed by immunofluorescence microscopy using either a Leica confocal microscope TSP-SP2 (Leica) or a DeltaVision RT Imaging System (Applied Precision Inc.).

Single-cell suspensions were generated from peripheral lymph nodes, spleen, BM, or peritoneal wash via gentle homogenization through nylon mesh filters (70 µM; BD). Erythrocytes in BMs and spleens were lysed with ACK lysis buffer before staining, and cell concentrations were determined using a Casy Counter (Scharfe Instrument Systems). CD1d-PE and CD1d-bio (clone 1B1); CD21-FITC (eBio8D9); CD23-PE (B3B4); B220-FITC, B220-PE, B220-APC, and B220-eFluor450 (RA3-6B2); CD93-APC (AA4.1); IgM-PECy7 (II/41); IgD-FITC, IgD-eFluor450, IgD-PE, and IgD-bio (11–26); Ly5.1-FITC, and Ly5.1-APC (A20); Ly5.2-FITC, and Ly5.2-APC-eFluor780 (104), CD43-bio (eBioR2/60), IgG1-APC (X56), Gr1-bio (RB6-8C5), CD11b-Bio (M1/70); CD11c-Bio (N418); CD5-PE (53–7.3); CD19-FITC (MB19-1); TCRβ-FITC (H57-597); and CCL19-Fc fusion protein (plus anti–human Fc-PE secondary to stain CCR7) were obtained from eBioscience. CD4-PerCP (RM4-5), B220-PcP (RA3-6B2), and CD8-PerCP (53–6.7) were purchased from BioLegend; CD2-PE (RM2-5), CD9-bio (KMC8), CD21-PE (7G6), Streptavidin-PerCP (554064), CD138-PE (281–2), and GL7-FITC (Ly77) were obtained from BD; PNA-Bio (FL-1075) was purchased from Vector Laboratories; and CD19-APC (RM705) and CD23-APC (MCD2305) were purchased from Life Technologies. For thymus-dependent (TD) antibody response assays, NP₂₇-BSA (Biosearch Technologies) was labeled with the lynx rapid rpe or percp antibody conjugation kits (AbDserotec) according to the manufacturer’s protocol. Intracellular IRF4 was detected using IRF4-PE (eBRG1), after cell permeabilization (Foxp3 Fixation/Permeabilization kit; eBioscience). Live cells (Live/dead Fixable Neat-IR Dead Cell Stain kit; Life Technologies) and extracellular markers were identified by prior staining of intact cells.

Stained cells were analyzed by flow cytometry on a CyAn ADP Analyzer (Beckman Coulter) or FACSCanto II Analyzer (BD). Data analyses were performed with FlowJo 887 software (Tree Star).

Splenic follicular B cells were isolated by sequential hypotonic lysis of RBC and magnetic bead separation using a mixture of biotinylated anti-CD43 and anti-CD1d antibodies and Streptavidin-coupled Dynabeads (Life Technologies). Purity of the resulting FM B cell population (B220⁺CD23^hiCD21^lo) was >96% as assessed by flow cytometry. Cells were cultured at a density of 2 × 10⁶ cells/ml in complete medium (DMEM plus 10% heat inactivated FCS [Lonza], 100 U/ml Penicillin, 100 µg/ml Streptomycin, 100 µM nonessential amino acids, 100 mM Na-Pyruvate, and 50 µM 2-mercaptoethanol), and stimulated with anti-IgM (10 µg/ml; Jackson ImmunoResearch Laboratories), CD40L (0.5 µg/ml; R&D Systems), or LPS (1–10 µg/ml; Salmonella minnesota R595; Alexis Biochemicals) ± IL-4 (20 ng/ml; PeproTech).

For in vitro survival assays, 3 × 10⁵ B cells were cultured in a 96-well tissue-culture plate in complete medium (200 µl) with or without stimulation (see above). Cells were harvested 2 d later, stained for a B cell marker (B220 or CD19) resuspended and 7-amino-actinomycin D (7-AAD; Sigma-Aldrich). Cell division was monitored by flow cytometry after labeling of FM B cells with 2 µM CFSE (CellTrace; Invitrogen) in Dulbecco’s PBS (Invitrogen). The frequency of the precursor population triggered into cell division was determined using FlowJo V8.5 proliferation analysis software (Tree Star). To monitor the antigen-presenting function of FM B cells, OVA was targeted to the BCR by treating purified B cells sequentially with biotinylated anti-IgM (0.5 µg/ml; Jackson ImmunoResearch Laboratories) and anti-biotin OVA-FITC (Miltenyi Biotec). Antigen-pulsed B cells (2 × 10⁶/ml) were then co-cultured with Cell Trace Violet-labeled OTII CD4⁺ T cells (5 µM; Life Technologies; 1 × 10⁶/ml) for 72 h. T cell activation was determined by estimating the ability of T cells to divide. To demonstrate that OVA presentation was actively mediated, antigen-pulsed B cells were fixed in 3% paraformaldehyde before co-culture with T cells.

To monitor plasma cell differentiation, FM B cells (1.5 × 10⁶/ml) were cultured with LPS (40 µg/ml) plus IL-4 (20 ng/ml) for 4 d. Plasma cells (B220^loCD138⁺) were identified by flow cytometry. In vitro IgM production was analyzed by harvesting cells after 4-d culture with LPS and IL-4, and replating washed live cells at 10⁶ cells/ml in complete medium. Supernatants were collected after an additional 1-d culture, and analyzed by ELISA. The yields of Nfkb1^SSAA/SSAA and WT FM B cells were similar after reculture.

pMSCV-IRES-GFP, pMSCV-IRF4-IRES-GFP, and pMSCV-Blimp-IRES-GFP plasmids (Kallies et al., 2007) were a gift from A. Kallies (The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia). Retroviruses were prepared as previously described using Plat-E packaging cells (Robinson et al., 2007). FM B cells were cultured for 24 h with CD40L and IL-4 before retroviral infection. To analyze the effect of ectopic IRF4 expression on GC B cell formation, purified MD4.Rag1^+/−.Nfkb1^SSAA/SSAA.Bcl-xL and MD4.Rag1^+/−.Nfkb1^+/+.Bcl-xL FM B cells were transduced twice with retrovirus, prior to cotransfer with 0.5 to 1.10⁹ HEL-conjugated or control sheep red blood cells (SRBC) into CD45.1⁺ C57BL/6 hosts by intravenous injection. Splenic B cells were analyzed by flow cytometry after 6 d.

For immunoblots of total lysates, FM B cells (3 × 10⁶ cells per point) were lysed RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 2 mM EDTA, 50 mM NaF, 2 mM Na₄P₂O₇, 5 mM dithiothreitol, 1 mM Na₃VO₄, protease inhibitors [Roche], 0.5% deoxycholate, 1% Triton X-100, and 0.1% SDS). After centrifugation, lysates (20–25 µg) were mixed with an equal volume of 2x SDS-PAGE sample buffer. For immunoprecipitation, cells (5 × 10⁷ per point) were lysed in IP buffer (10 mM Hepes, pH 7.6, 1 mM EGTA, 10 mM KCl, 1 mM dithiothreitol, 20 mM NaF, 1 mM Na₄P₂O₇, protease inhibitors, and 0.6% Nonidet-P40). Immunoprecipitations (150 µg of protein extract) were performed as previously described (Kabouridis et al., 1997).

To analyze nuclear translocation of Rel subunits, washed cells were lysed with buffer N (10 mM Hepes, pH 7.6, 0.1 mM EGTA, 10 mM KCl 1.5 mM MgCl2, 1 mM DTT, 20 mM NaF, 1 mM Na-pyrophosphate, protease inhibitors, and 0.2% NP-40). After 2 min, nuclei were pelleted by centrifugation, washed twice in lysis buffer, and nuclear proteins extracted with RIPA buffer. Rel proteins in nuclear extracts (20 µg) were quantified by immunoblotting and laser densitometry, normalizing to SAM68 content.

The p105C antibody to the p105 C terminus, used for p105 immunoprecipitation, has been described previously (Salmerón et al., 2001). p50 antibody was obtained from (Delta Biolabs) and phospho S933 p105 antibody from Cell Signaling Technology. Antibodies to other Rel family proteins, SAM68 and IκBα, were purchased from Santa Cruz Biotechnology, Inc. Tubulin was detected with TAT-1 α-tubulin mAb (provided by K. Gull, University of Oxford, Oxford, England, UK) and used as loading control protein on immunoblots of total cell lysates.

Nuclear extracts were prepared as above, and activation of NF-κB subunits assayed (2–4 µg per point) using a commercial ELISA kit (TransAm NF-κB family kit; Active Motif). To increase sensitivity, c-Rel binding was assayed with anti–c-Rel (1 µg per well; sc-71X; Santa Cruz Biotechnology, Inc.). Data were normalized against SAM68 content of nuclear extracts.

RNA from stimulated and unstimulated FM B cells (5 × 10⁶/ml) was isolated using the RNeasy kit, and contaminating DNA removed using RNase-free DNase set (QIAGEN) according to manufacturer’s instructions and DNase treated (Roche). cDNA was produced using the SuperScript III First-Strand Synthesis SuperMix for qRT-PCR (Life Technologies) and standard protocols. Expression of target genes was determined by real time PCR using a Perkin Elmer ABI Prism 7000 Sequence Detection System, and commercial FAM labeled probes (Applied Biosystems) with the TaqMan Gene Expression Master Mix (Applied Biosystems). Target gene mRNA levels were normalized against Hprt mRNA levels.

All data analyses were performed using GraphPad software (GraphPad Software Inc.). In vitro data were compared using Student’s t test (two tailed and unpaired test). For in vivo experiments, all statistical comparisons were performed using the nonparametric two-tailed Mann-Whitney test. Statistically significant differences are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Fig. S1 shows gating strategies for Fig. 1. Fig. S2 shows gating strategies for Fig. 6. Fig. S3 shows gating strategies for Fig. 8 (E and F). Fig. S4 shows gating strategies for Fig. 8 H.

We thank the NIMR Photographics department, NIMR Confocal Microscopy, NIMR Biological Services, NIMR flow cytometry service, Harald Hartweger (NIMR), Abduelhakem Ben-Addi, and other members of the Ley laboratory for help during the course of this work. We are also grateful to Anne O’Garra (NIMR), Ben Seddon (NIMR), and Steve Gerondakis (Monash University, Australia) for their helpful comments on the manuscript.

This work was funded by the UK Medical Research Council, a Leukaemia and Lymphoma Research Project Grant (LRF:06050), and an award from the Bettencourt-Schueller Foundation to E. Jacque.

The authors declare no competing financial interests.