B cell-intrinsic requirement for WNK1 kinase in T cell-dependent antibody responses

Migration and adhesion play critical roles in B cells, regulating recirculation between lymphoid organs, migration within lymphoid tissue and interaction with CD4+ T cells. However, there is limited knowledge of how B cells integrate chemokine receptor and integrin signaling with B cell activation to generate efficient humoral responses. Here we show that the WNK1 kinase, a regulator of migration and adhesion, is essential in B cells for T-dependent antibody responses. We demonstrate that WNK1 transduces signals from the BCR, CXCR5 and CD40, and using intravital imaging we show that WNK1 regulates migration of naive and activated B cells, and their interactions with T cells. Unexpectedly, we show that WNK1 is required for BCR- and CD40-induced proliferation, acting through the OXSR1 and STK39 kinases, and for efficient B cell-T cell collaboration in vivo. Thus, WNK1 is critical for humoral immune responses, by regulating B cell migration, adhesion and T cell-dependent activation. Summary The WNK1 kinase is essential in B cells for T-dependent antibody responses because it is activated by signaling from BCR, CXCR5 and CD40 and regulates B cell migration, adhesion, T-dependent activation, and differentiation into germinal center B cells and plasma cells.


Introduction
Migration and adhesion play critical roles in B cell physiology. Naïve B cells migrate between lymphoid organs via the bloodstream, entering lymph nodes by extravasating across the endothelium of high endothelial venules using LFA-1 integrin-mediated adhesion triggered by chemokine signals (Girard et al., 2012). Once inside lymphoid tissue, B cells migrate into and within follicular areas under the influence of chemokines such as CXCL13 (Schulz et al., 2016). This continuous movement of B cells allows the cells to locate their cognate antigen and is thus essential for recruitment of B cells into the adaptive immune response.
Migration and cell-cell interactions are also vital during T-dependent B cell activation and subsequent generation of high affinity antibodies after microbial challenge or vaccination (Cyster and Allen, 2019). Binding of cognate antigen to the B cell antigen Hayward et al receptor (BCR) results in internalization of the antigen-BCR complex, degradation of the antigen and, for protein-based antigens, presentation of resulting peptides on MHC class II molecules (Akkaya et al., 2020). Furthermore, binding of antigen to the BCR induces signals that cause movement of B cells from the follicles towards the border of the T cell zone where they interact with CD4 + T cells bearing T cell antigen receptors (TCR) specific for the peptide-MHC complex on B cells (Tangye et al., 2013).
Subsequent two-way signaling between the B and T cells causes both cell types to become activated and to divide (Crotty, 2015;Petersone et al., 2018). B cell-T cell couples migrate back into the follicle and differentiate into germinal center B (GCB) cells and T follicular helper (Tfh) cells, thereby establishing germinal centers (Mesin et al., 2016;Victora and Nussenzweig, 2012). Within these structures somatic hypermutation of immunoglobulin genes drives affinity maturation, and GCB cells differentiate into antibody-secreting plasma cells and memory B cells. Thus, migration and adhesion play essential roles in B cells during T-dependent activation and need to be closely integrated with activation signals that jointly result in high affinity antibody responses.
WNK1, a member of the WNK-family of protein kinases, is best characterized in kidney epithelial cells where it regulates salt uptake from the urine (McCormick and Ellison, 2011;Shekarabi et al., 2017;Wilson et al., 2001). WNK1 phosphorylates and activates the related OXSR1 and STK39 kinases, which in turn phosphorylate members of the SLC12A-family of ion co-transporters causing net influx of Na + , K + and Cl - (Alessi et al., 2014;Thastrup et al., 2012;Vitari et al., 2005;Vitari et al., 2006). Surprisingly, we discovered that in CD4 + T cells, WNK1 is activated by signaling from the TCR and CCR7 and that it is a positive regulator of CCR7-induced migration, via OXSR1, STK39 and SLC12A2, and a negative regulator of LFA-1-mediated adhesion (Köchl et al., We hypothesized that WNK1 may regulate migration and adhesion of B cells, and thus play an important role in humoral immune responses. We show that WNK1 is activated by signals from the BCR and CXCR5 and is a positive regulator of chemokine-induced migration and a negative regulator of integrin-mediated adhesion. Unexpectedly, we also show that WNK1 is activated by CD40 signaling and that it transduces BCR and CD40 signals via OXSR1 and STK39 that lead to B cell division. Using intravital imaging we find that WNK1 is required for the migration of naive and activated B cells and their interactions with T cells, and show that B cell-intrinsic loss of WNK1 impairs B cell-T cell collaboration, differentiation into GCB cells, and severely reduces Tdependent antibody responses. Thus, WNK1 is critical for humoral immunity because it regulates B cell migration, adhesion, activation and division.

B lineage cells express members of the WNK pathway
To evaluate whether WNK1 may regulate B cell migration and adhesion, we first assessed if members of the WNK1 pathway were expressed in B lineage cells. Using our previously generated RNAseq data from 8 different B cell subsets , we found that Wnk1 is the only WNK-family member expressed in any B cell subset ( Figures S1A-D). Furthermore, we found that Oxsr1 and Stk39, encoding the related WNK1-activated OXSR1 and STK39 kinases, are expressed in all B cell subsets, albeit expression of Oxsr1 is higher than Stk39 ( Figure S1E, S1F). Of the seven genes encoding the SLC12A-family of co-transporters, expression of Slc12a2, Slc12a3, Slc12a4, Slc12a6 and Slc12a7 was also detected (Figure S1G-M). Thus, B cells express multiple members of the WNK1 pathway.

Hayward et al
To study the role of WNK1 in B cells, we used three different strategies. Firstly, since constitutive inactivation of Wnk1 results in embryonic lethality (Zambrowicz et al., 2003), we bred mice with a loxP-flanked (floxed) allele of Wnk1 (Wnk1 fl ), a tamoxifeninducible Cre recombinase expressed from the ROSA26 locus (ROSA26 CreERT2 , RCE) and either a wild type or constitutively deleted allele of Wnk1 (Wnk1 + or Wnk1 -) and used bone marrow from the resulting Wnk1 fl/+ RCE or Wnk1 fl/-RCE mice to reconstitute irradiated RAG1-deficient mice ( Figure S2A). Treatment of these animals with tamoxifen led to efficient deletion of the floxed Wnk1 allele in B cells 7 d later, hence generating mice with WNK1-deficient B cells and control WNK1-expressing B cells ( Figure S2B).
Secondly, to extend the analysis to the function of WNK1 kinase activity we used a Wnk1 allele expressing a kinase-inactive WNK1-D368A (Wnk1 D368A ) (Köchl et al., 2016), reconstituting RAG1-deficient mice with bone marrow from either Wnk1 fl/+ RCE or Wnk1 fl/D368A RCE mice ( Figure S2A). Treatment of these chimeras with tamoxifen resulted in generation of B cells expressing only WNK1-D368A or control B cells expressing wild-type WNK1. In both cases, loss of WNK1 or its kinase activity caused a drop in the number of mature B cells in the spleen and lymph nodes (Figures S2C,S2D), suggesting that WNK1 activity may be important for B cell survival. Nonetheless, sufficient numbers of WNK1-deficient or WNK1-D368A expressing B cells remained for functional analysis.
Thirdly, it is possible that loss of WNK1 or its kinase activity may cause changes in B cells during the 7 d following the start of tamoxifen treatment, resulting in phenotypes that are not due to an acute requirement for WNK1. To overcome this limitation, we used WNK463, a highly selective WNK kinase inhibitor (Yamada et al., 2016), which allowed investigation of the effect of acute inhibition of WNK1 kinase activity.
Hayward et al 6 WNK1 transduces signals from both the TCR and CCR7 (Köchl et al., 2016), thus we investigated whether, by analogy, WNK1 is activated downstream of the BCR and CXCR5 in B cells. We used phosphorylation of Ser325 on OXSR1 as a readout, as this residue is a direct WNK1 target (Vitari et al., 2005). We found that stimulation of B cells with anti-IgM or CXCL13, the ligand for CXCR5, resulted in a rapid increase in phosphorylated OXSR1 (pOXSR1), which peaked at 10-20 min in response to BCR stimulation and at 2 min in response to CXCL13, with the latter giving a larger response . This increase in pOXSR1 was eliminated by loss of WNK1 or its kinase activity, or by treatment of wild-type B cells with WNK463, demonstrating that both BCR and CXCR5 signaling activates WNK1. Notably, the loss of WNK1 or its kinase activity reduced levels of pOXSR1 to below those detected in unstimulated B cells.
Thus, WNK1 kinase has basal activity in resting B cells which is rapidly increased in response to stimulation through the BCR or CXCR5.

BCR and CXCR5 signaling activate WNK1 via PI3K and AKT
TCR and CCR7 signaling activates WNK1 via phosphoinositide 3-kinase (PI3K) and AKT (Köchl et al., 2016), thus we investigated if this is also true for BCR and CXCR5 signaling in B cells. Stimulation of B cells with anti-IgM or CXCL13 in the presence of PI-103 and MK2206, inhibitors of PI3K and AKT respectively, resulted in decreased levels of phosphorylated AKT and PRAS40, confirming that the inhibitors were functional ( Figures 1G, 1H). Both inhibitors also reduced the BCR-and CXCR5induced increase in pOXSR1. Thus, both BCR and CXCR5 transduce signals via PI3K and AKT that lead to WNK1 activation.

WNK1 is a negative regulator of B cell adhesion to ICAM-1
As WNK1 is a negative regulator of LFA-1-mediated adhesion in CD4 + T cells (Köchl et al., 2016), we investigated if WNK1 plays a similar role in B cells stimulated via the BCR or CXCR5. We found that in response to stimulation by anti-IgM or CXCL13, B Hayward et al cells deficient in WNK1 or its kinase activity showed substantially higher binding to ICAM-1 and VCAM-1, ligands for the LFA-1 and VLA-4 integrins respectively (Figures 2A,2B,S2E). Thus, WNK1 is a negative regulator of LFA-1-and VLA-4-mediated adhesion in B cells. Furthermore, loss of WNK1 also induced higher binding to ICAM-1 following treatment of B cells with MnCl2 which induces a conformational change in LFA-1 resulting in a high affinity for ICAM-1, demonstrating that WNK1 also regulates this outside-in mode of LFA-1 activation. Analysis of surface proteins showed that WNK1-deficient B cells had normal levels of IgM, slightly increased levels of CD11a, a subunit of LFA-1, and slightly decreased levels of CXCR5 ( Figure S2F). The increase in CD11a is much smaller than the increase in adhesion, so it is unlikely to account for the majority of the increased adhesion.
To investigate if OXSR1 or STK39, two well-characterized substrates of WNK1 (Vitari et al., 2005), transduce WNK1 signals that regulate LFA-1 mediated adhesion, we used mice containing floxed alleles of Oxsr1 (Oxsr1 fl ), reconstituting irradiated RAG1deficient mice with bone marrow from Oxsr1 +/+ RCE and Oxsr1 fl/fl RCE mice ( Figure   S2G). Treatment of these chimeras with tamoxifen resulted in efficient loss of OXSR1, but no change in numbers of mature B cells (Figures S2H,S2I). Analysis of OXSR1deficient B cells showed that they also had increased ICAM-1 binding in response to treatment with anti-IgM, CXCL13 or MnCl2 ( Figure 2C). However, the increase was not as large as that seen in the absence of WNK1, suggesting that other WNK1 substrates may be involved, such as STK39. To investigate this potential redundancy between OXSR1 and STK39, we combined the floxed Oxsr1 allele with an allele of Stk39 expressing a mutant STK39-T243A protein that cannot be phosphorylated and activated by WNK1 (Stk39 T243A ). Reconstitution of irradiated RAG1-deficient mice with bone marrow from Oxsr1 +/+ Stk39 +/+ RCE and Oxsr1 fl/fl Stk39 T243A/T243A RCE mice ( Figure   S2G) and subsequent tamoxifen treatment resulted in efficient loss of OXSR1 in the double mutant mice, and again no change in B cell numbers ( Figures S2J, S2K).
Analysis of B cells deficient in OXSR1 and expressing STK39-T243A showed that they also had increased binding to ICAM-1 in response to stimulation with anti-IgM, CXCL13 or MnCl2, and that the increase was larger than that seen following loss of OXSR1 alone ( Figure 2D). These results are consistent with the hypothesis that WNK1 transduces signals from the BCR and CXCR5 via OXSR1 and STK39, leading to negative regulation of LFA-1-mediated adhesion.

WNK1 is a positive regulator of CXCL13-induced B cell migration in vitro
WNK1 regulates migration in different cell types (Köchl et al., 2016;Sun et al., 2006;Zhu et al., 2014). Thus, we hypothesized that WNK1 may also regulate CXCR5induced migration in B cells. Using a Transwell assay, we found that WNK1-deficient B cells migrated through the Transwell filter much less efficiently in response to CXCL13 ( Figure 2E). Furthermore, imaging of CXCL13-induced migration showed that loss of WNK1 resulted in reduced B cell speed and displacement ( Figures 2F, 2G).
Thus, WNK1 is a positive regulator of CXCL13-induced chemotaxis and chemokinesis.

WNK1 is required for efficient homing of B cells to lymphoid organs and migration within them
In view of the altered adhesion and migration of WNK1-deficient B cells in vitro, and the importance of these processes in B cell trafficking in vivo, we examined the ability of the mutant B cells to home efficiently to lymphoid organs. We injected a mixture of control and WNK1-deficient B cells intravenously into mice and 1 h later analyzed their distribution by flow cytometry ( Figure 3A). The results showed that fewer WNK1deficient B cells entered lymph nodes and spleen compared to control B cells, whereas more mutant cells remained in the blood, indicating that WNK1 is required for efficient homing of B cells to lymphoid organs ( Figure 3B). Imaging of unsectioned lymph nodes confirmed that fewer mutant B cells had entered the tissue and, in comparison to control B cells, WNK1-deficient B cells were more likely to be located in the lumen of blood vessels or in perivascular areas, and less likely to have entered the parenchyma ( Figures 3C, S3A, Video S1). B cells that traffic to the spleen first enter the red pulp and then migrate into the lymphocyte-containing white pulp in response to CXCL13 (Bronte and Pittet, 2013). Histological analysis of the spleen showed that 1 h after transfer, fewer WNK1-deficient B cells were found in both the red and white pulp compared to control B cells, with a relatively larger reduction in the white pulp ( Figures   3D, S3B). Taken together these results show that WNK1 is required for B cells to home efficiently to lymphoid organs and to enter the regions of these tissues harboring lymphocytes.
Once inside the follicles of lymphoid tissue, B cells continue to migrate in response to CXCL13, a process that allows them to scan cells for presentation of cognate antigen (Akkaya et al., 2020;Cyster and Allen, 2019). To evaluate if WNK1 also regulates this mode of migration, we transferred control or WNK1-deficient B cells into wild-type mice, waited 24 h, and used multi-photon intravital microscopy (MP-IVM) to measure their migration within follicles ( Figure 3E). We found that in the absence of WNK1, B cells migrated more slowly, exhibited larger turning angles with reduced track straightness (lower motility coefficient and meandering index) and stopped moving more frequently (increased arrest coefficient) (Figures 3F-3I, S3C, S3D, Video S2).
Thus, WNK1 regulates B cell migration within lymphoid tissue.

WNK1 is required for activation of B cells in vitro
Since WNK1 is activated by BCR signaling, it may also contribute to BCR-induced activation, beyond its role in migration and adhesion. To test this possibility, we stimulated control and WNK1-deficient B cells for 3 days with anti-IgM and measured changes in levels of cell surface proteins (activation markers), cytokine secretion, cell division and cell survival. We extended this analysis to include stimulation of the cells through CD40 and TLR4, by treating cells with CD40L and LPS respectively.
In response to BCR stimulation, both control and WNK1-deficient cells upregulated CD69, CD71, CD80, CD86, CD95 and MHC class II ( Figure S4A). However, in the absence of WNK1 the upregulation of CD80 and CD95 was higher on day 3 of the culture, whereas upregulation of MHC class II was reduced on days 2 and 3. In response to CD40L stimulation, both control and WNK1-deficient B cells upregulated CD69, CD80, CD86, CD95, ICOSL and MHC class II, but the mutant cells had higher levels of CD69, CD80, CD86 and CD95 on days 1-3, and reduced levels of ICOSL and MHC class II on days 2 and 3 ( Figure S4B). Analysis of cytokine secretion showed that in response to anti-IgM WNK1-deficient B cells secreted normal levels of VEGF-A, but higher levels of IL6, IL10 and TNFa, whereas in response to CD40L stimulation secretion of all four cytokines was unaffected ( Figure S4C). Thus, in the absence of WNK1, B cells can still be activated by signaling through both the BCR and CD40 but show altered levels of activation markers and cytokines.
In contrast, analysis of cell proliferation showed that WNK1-deficient B cells have greatly reduced cell division in response to either anti-IgM or CD40L stimulation, but normal division in response to LPS ( Figures 4A, 4B). WNK1-deficient B cells express unaltered levels of IgM and CD40, so this cannot account for the defects ( Figure S2F).
We observed reduced recovery of live WNK1-deficient B cells after 3 days of stimulation with anti-IgM, CD40L or LPS compared to control B cells ( Figure 4C). To distinguish if this decrease in cell numbers was due to reduced cell division or poorer survival, we calculated numbers of remaining B cells after taking out the effect of cell division. This showed that fewer WNK1-deficient B cells survived in cultures with all three agonists ( Figure 4D). Taken together, these results show that WNK1 is required for normal cell division in response to BCR and CD40 but not TLR4 stimulation, and is required for cell survival during B cell activation.
We extended this analysis to expression of Ki-67, which is induced when B cells move from the G0 to G1 phase of the cell cycle. Both anti-IgM and CD40L induce Ki-67 expression in B cells with > 80% cells becoming Ki-67 + after 3 days of culture ( Figure   4E). WNK1-deficient B cells also induce Ki-67 expression, albeit with slower kinetics, and by day 3 of anti-IgM and CD40L stimulation around 50% and 75% of B cells are Ki-67 + , respectively ( Figure 4E). Thus, in the absence of WNK1, B cells are delayed in their entry into G1. In a recent study we showed that WNK1-deficient thymocytes fail to proliferate in response to pre-TCR signals due to defective upregulation of MYC, a transcription factor that induces expression of many proteins required for cell division (Köchl et al., 2020). To investigate if WNK1 plays a similar role in B cells, we measured MYC levels in control and WNK1-deficient B cells in response to anti-IgM and CD40L stimulation. We found that in both cases WNK1-deficient B cells upregulate MYC at least as much as control B cells, suggesting that in B cells WNK1 does not regulate cell division by controlling levels of MYC ( Figures S4D, S4E).
To further investigate the mechanism by which WNK1 regulates cell division, we analyzed if WNK1 kinase activity was required for this process. We found that B cells expressing kinase-inactive WNK1-D368A again had greatly reduced cell division in response to anti-IgM and CD40L stimulation ( Figures 4F, 4G). Thus, WNK1 kinase activity is required for this activation response. Next, we investigated whether the WNK1 substrates, OXSR1 and STK39 are also involved. Analysis of OXSR1-deficient B cells showed that they divided normally in response to anti-IgM and CD40L ( Figures   4H, 4I). In contrast, B cells lacking OXSR1 that also had the STK39-T243A mutation had a reduction in anti-IgM and CD40L-induced cell division, although the defect was not as large as in WNK1-deficient B cells ( Figures 4J, 4K). Taken together these results suggest that BCR and CD40 transduce signals via WNK1 to OXSR1 and STK39, which are required for B cells to divide.

CD40 signaling activates WNK1
Since WNK1 is required for CD40-induced cell division, we investigated if signaling from this receptor activates WNK1. Stimulation of B cells through CD40 resulted in a rapid induction of pOXSR1, peaking around 10 min, which was eliminated by loss of WNK1 or its kinase activity and was also inhibited by WNK463 ( Figures 4L-4N). In contrast, treatment of B cells with LPS did not cause any change in pOXSR1 ( Figure   S4F). Thus, signals from CD40, but not TLR4, induce WNK1 activation. Since both BCR-and CXCR5-induced activation of WNK1 requires PI3K and AKT, we analyzed if the same was true for CD40. Inhibition of PI3K caused no change in CD40-induced pOXSR1, whereas inhibition of AKT resulted in a partial reduction of pOXSR1 ( Figure   S4G). We conclude that, unlike BCR and CXCR5, CD40 does not activate WNK1 via PI3K, and is only partially dependent on AKT.

WNK1-deficient B cells activate T cells less efficiently in vitro
In view of the defective BCR-and CD40-induced cell division and altered levels of cell surface proteins that are involved in B cell-T cell communication (CD80, CD86, ICOSL and MHC class II), we investigated if WNK1 was required for cognate B cell-T cell interaction and activation. We treated control or WNK1-deficient B cells with beads coated with anti-IgM antibodies and Ovalbumin (OVA), washed out unbound beads and cultured the B cells with OVA-specific OT-II CD4 + T cells for 1-3 days. In this assay, crosslinking of the BCR results in the internalization of anti-IgM and OVA-coated beads and the B cells present the OVA 323-339 peptide (OVA323-339) on surface MHC class II molecules. These peptide-MHC complexes bind to the TCR on OT-II T cells, triggering the formation of B-T conjugates, and activation and division of the T cells ( Figure 5A).
In control cultures where the B cells were given no beads ( Figure 5B), or beads coated in just anti-IgM or just OVA, or no B cells were present (data not shown), the OT-II T cells did not respond, as measured by upregulation of activation markers (CD25, CD44, CD69) and division. In contrast, in cultures containing WNK1-expressing B cells given beads coated with anti-IgM and OVA, T cells upregulated activation markers and started to divide, demonstrating that T cell activation in this assay requires cognate antigen-driven B cell-T cell interaction ( Figure 5B). Notably, in cultures containing WNK1-deficient B cells, the T cells showed lower induction of activation markers, and fewer started to divide. Thus, in the absence of WNK1 in B cells, B-T collaboration is severely impaired.
A key event in B-T collaboration is the formation of conjugates in response to presentation of cognate antigen by the B cells to the T cells, a process which is dependent on integrin-mediated adhesion. In view of the role of WNK1 in regulating adhesion, we investigated if conjugation might be perturbed. However, we found no difference in the efficiency of conjugate formation between control and WNK1-deficient B cells ( Figure 5C). Once antigen is bound by the BCR, it is internalized by the B cell, degraded, and presented on MHC class II molecules. Analysis of these steps showed that loss of WNK1 did not adversely affect antigen internalization or degradation but caused a partial decrease in the amount of antigen presented on class II molecules ( Figure 5D-F). Thus, the defective collaboration of WNK1-deficient B cells with T cells may be due in part to reduced antigen presentation.

WNK1-deficient B cells fail to mount T-dependent antibody responses
To evaluate the effect of loss of WNK1 on T-dependent antibody responses, we reconstituted irradiated RAG1-deficient mice with a mixture of Ighm µMT/µMT bone marrow (80%) which is unable to generate B cells (Kitamura et al., 1991) and either Wnk1 fl/+ RCE or Wnk1 fl/-RCE bone marrow (20%). Treatment of the resulting mixed radiation chimeras with tamoxifen results in mice with either WNK1-expressing or WNK1-deficient B cells, whereas most other cells express WNK1. Seven days after the start of tamoxifen treatment, the mice were immunized with NP-conjugated chicken gamma globulin (NP-CGG) in alum, a T-dependent antigen, and blood and splenocytes were analyzed over the following 28 days ( Figure 6A). WNK1-deficient B cells were severely impaired in their ability to form NP-specific GCB cells and plasma cells and generated greatly reduced levels of NP-specific IgM and IgG1 in the serum ( Figures 6B, 6C, S5A). Thus, WNK1 is required in B cells for T-dependent antibody responses.
Since the loss of WNK1 results in lower numbers of B cells ( Figure S2C), this could contribute to the reduced antibody response. To evaluate the role of WNK1 without this reduced number of B cells, we used the Aicda-Cre transgene to delete Wnk1, since this is only induced in B cells once they are activated. We immunized Wnk1 fl/+ Aicda-Cre or Wnk1 fl/-Aicda-Cre mice with CGG or NP-CGG in alum and analyzed the resulting immune response ( Figure 6D). We found that loss of WNK1 in activated B cells resulted in a large reduction of NP-specific GCB cells and plasma cells ( Figures 6E, S5B). Notably, around 70% of the few remaining GCB cells did not express human CD2, a marker of expression of the Aicda-Cre transgene, implying that these were cells that had most likely not deleted the Wnk1 gene, further emphasizing the critical role of WNK1 in the differentiation of naïve mature B cells into GCB cells (Figures 6F,S5B). Moreover, in the absence of WNK1 in activated B cells, the levels of NP-specific IgG1 were reduced ( Figure 6G), although there was no effect on the levels of NP-specific IgM, most likely because this is generated early in the immune response before the B cells lost WNK1. These results show a cell-intrinsic requirement for WNK1 in B cells for differentiation into GCB cells and plasma cells during a Tdependent immune response.

B cells require WNK1 to collaborate efficiently with T cells in vivo
To understand the role of WNK1 in the early phases of the antibody response, we treated control or WNK1-deficient B cells with beads coated with anti-IgM antibodies, earlier with OVA in alum to provide activated cognate T cells ( Figure  and hence their inability to support T-dependent antibody responses.

Discussion
We have discovered a previously unknown signaling pathway in B cells involving the kinase WNK1. Using a wide variety of in vitro and in vivo techniques, we found that WNK1 is a critical regulator of multiple aspects of B cell physiology. WNK1 is rapidly activated by signaling from the BCR, CXCR5 and CD40, regulates B cell migration, adhesion, antigen presentation, proliferation and survival, and in its absence from B cells, T-dependent antibody responses are strongly inhibited.
Signaling from the BCR and CXCR5 activates WNK1 via PI3K and AKT, and CD40induced WNK1 activation is partially AKT-dependent. AKT may directly regulate WNK1, and indeed AKT phosphorylates WNK1 on Thr60 (Jiang et al., 2005;Vitari et al., 2004). However, this phosphorylation does not affect WNK1 in vitro kinase activity, or its cellular localization. Thus, it remains unclear how signaling from AKT regulates WNK1 activity, and the identity of other regulators of WNK1 remain unknown.
We found that WNK1 is a positive regulator of CXCR5-induced migration of naïve B cells and that the WNK1-OXSR1-STK39 pathway negatively regulates BCR-and CXCR5-induced activation of adhesion through LFA-1 and VLA-4 integrins. WNK1deficient T cells have higher levels of active RAP1, a GTPase which is involved in inside-out integrin activation (Alon and Feigelson, 2012). WNK1 may also regulate RAP1 in B cells. It is unknown how the WNK1-OXSR1-STK39 pathway regulates RAP1 activity, but it may act directly or indirectly on RAP1 guanine nucleotide exchange factors or RAP1 GTPase activating proteins.
Unexpectedly, we discovered that WNK1 is required for BCR-and CD40-induced B cell proliferation. This requirement for WNK1 in B cell proliferation is dependent on its kinase activity and is most likely transduced via OXSR1 and STK39 since mutations in these two related kinases resulted in a phenotype similar to WNK1-deficiency.
Previous cell line studies reported that WNK1 regulates cell division (Sun et al., 2006;Tu et al., 2011) and our own work in thymocytes showed that WNK1 is required for pre-TCR-induced upregulation of MYC and hence proliferation of thymocytes at the bselection checkpoint (Köchl et al., 2020). However, the defective proliferation of WNK1-deficient B cells must be due to another mechanism since they upregulate MYC normally.
Some of the best characterized substrates of the WNK-OXSR1-STK39 pathway are the SLC12A-family of ion co-transporters (Shekarabi et al., 2017). Activation of WNK1 leads to phosphorylation and activation of OXSR1 and STK39, which directly phosphorylate multiple SLC12A-family proteins resulting in activation of cotransporters that allow influx of Na + , K + and Cland inhibition of co-transporters that mediate efflux of K + and Clions. It is possible that the requirement for the WNK1-OXSR1-STK39 pathway in BCR-and CD40-induced B cell activation stems from a need for regulated ion import into the cell, or subsequent water movement, especially as division and migration both require changes, either globally or locally, in cell volume. This is an interesting area for future studies, but these will be challenging given that B cells express five members of the SLC12A-family, which are likely to have redundant functions.
Using an in vitro antigen-specific co-culture system, we showed that loss of WNK1 impairs the ability of B cells to activate cognate T cells, potentially due to reduced antigen presentation. While WNK1-deficient B cells internalized and degraded antigen normally, they presented less antigenic peptide on cell surface MHC II molecules. This implies that loss of WNK1 may affect the loading of antigenic peptides onto MHC II molecules in the late endosomal-lysosomal antigen-processing compartment, or subsequent trafficking of peptide-MHC II complexes to the surface (Roche and Furuta, 2015). Several studies have shown that WNK1 regulates vesicle exocytosis and trafficking of membrane proteins to the cell surface (Henriques et al., 2020;Kim et al., 2018;Lee et al., 2004;Mendes et al., 2010;Oh et al., 2007). It will be interesting to investigate if WNK1 regulates exocytosis of peptide-loaded MHC II molecules.
Strikingly, we have shown a B cell-intrinsic requirement for WNK1 in T-dependent antibody responses. WNK1 is required for B cells to receive T cell help and differentiate into GCB cells and plasma cells, because WNK1 regulates B cell migration to the B-T border and communication between B cells and CD4 + T cells. WNK1 may also be a critical regulator of GCB cell function as migration and cell-cell interactions are required for GCB differentiation to memory B and plasma cells, both of which provide protection during reinfection. It will be important to further explore how WNK1 regulates humoral responses as this may reveal novel pathways that could be manipulated to improve vaccine responses or suppress generation of pathogenic autoantibodies.

Radiation chimeras
To generate radiation chimeras, bone marrow cells were harvested from Wnk1 fl/+ RCE, Oxsr1 fl/fl RCE, Oxsr1 fl/fl Stk39 T243A/T243A RCE, Ighm µMT/µMT or OT-II mice, or fetal livers were harvested from E14.5 embryos generated by intercrossing or Slc12a2 +/mice. RAG1-deficient animals (5-8 weeks of age) were irradiated with 5Gy using a 137 Cssource, and then reconstituted intravenously with at least 1x10 6 bone marrow cells/recipient. The sex of the donor and recipient mice was always matched. All chimeric animals received Baytril (0.02%, Bayer Healthcare) in their drinking water for at least 4 weeks post-transplantation. If required, 8-20 weeks after reconstitution chimeric mice were injected intraperitoneally with 2 mg/day of tamoxifen (20 mg/ml in corn oil, Sigma) for 3 successive days and analyzed either 7 d after start of tamoxifen treatment for mice containing a Wnk1 fl allele or treated for 5 successive days and analyzed 21 d after start of tamoxifen treatment for mice containing a Oxsr1 fl allele.

Flow Cytometry
Flow cytometry was carried out using standard techniques with pre-titered antibodies.
Antibodies for flow cytometry and cell isolation against the following proteins (all mouse unless otherwise indicated) were obtained from BioLegend, eBioscience, Invitrogen, Q-PCR for Wnk1 mRNA 5x10 5 splenic B cells from control or WNK1-deficient animals were isolated as described above, total RNA was extracted with an RNAEasy Plus Micro Kit (Qiagen) and cDNA was synthesized with a SuperScript TM VILO TM cDNA Synthesis kit (Thermo Fisher Scientific). Samples were analyzed on a QuantStudio3 Real-time PCR system (Thermo Fisher Scientific) using a TaqMan gene expression assay spanning exon1 and exon 2 of Wnk1 (Mm01184006_m1, Thermo Fisher Scientific). Data was normalized to Hprt (Mm03024075_m1, Thermo Fisher Scientific) and analyzed using the comparative threshold cycle method.

Migration assays
B cell migration was assessed in 96-well Transwell plates, containing polycarbonate filters (5 µm pore size, Corning). Transwell filters were coated overnight with mouse ICAM1-Fc (500 ng/ml in PBS) and blocked with PBS, 2% BSA for over 1 h. The B cells were rested in AB IMDM, 5% FCS for 1 h at 37°C. The receiver plate was filled with RPMI, 0.5% BSA, containing CXCL13 (1 µg/ml) or no chemokine, and 8x10 4 splenic B cells in RPMI, 0.5% BSA were added to each well of the filter plate. After 3 h at 37°C the filter plate was removed, EDTA was added to each well (40 mM final concentration) and the cells were transferred to 96-well V-bottom plates, spun, resuspended in PBS, 0.5% BSA, and cell numbers determined by flow cytometry. Percentage migration was calculated by dividing the number of cells that migrated through the filter by the total number of cells that had been added to each well. CXCL13 (1 µg/ml) was added to the chamber and three videos (60 frames at 3 frames/min) were generated for each chamber using Eclipse Ti2 and µManager (https://micro-manager.org). Videos were analyzed using the TrackMate (Tinevez et al., 2017) plugin in FIJI (Schindelin et al., 2012). Analysis was limited to cells that appeared in all frames and had a displacement ³ 5 µm.
After 1 h, blood, spleen and lymph nodes were harvested, stained with antibodies against B220 and analyzed by flow cytometry to determine the ratio between CMFDAand CTV-labelled B220 + B cells. Technologies). Sections were also counter-stained with DAPI. Images were acquired using an Zeiss Axio Scan.Z1 slide scanner and analyzed by drawing regions of interest around the spleen and the white pulp areas and cells were counted using FIJI.
Numbers of cells were first normalized to the red and white pulp area, respectively and a ratio of cell density in the white pulp to red pulp was then calculated.

Intravital Microscopy of Naïve B cells
Splenic B cells were isolated using a mouse B cell isolation kit (STEMCELL Technologies) according to the manufacturer's protocol. B cells of each genotype were labelled either with 5 µM CMTMR or 20 µM CMAC for 20 min at 37°C, 5% CO2. and mixed at a 1:1 ratio. The mixture was transferred intravenously into sex-matched C57BL/6J mice. 24 h after transfer, the popliteal LN of recipient mice was surgically exposed for two-photon microscopy (2PM) as described (Moalli et al., 2014).
Immediately prior to imaging, HEVs were labelled by intravenous injection of Alexa  (Vladymyrov et al., 2016). Cell migration analysis was performed using Imaris 9.1.2 software (Bitplane). Median speeds, meandering index and turning angles (defined as the angle between the two velocity vectors before and after a measurement time point) were calculated from the (x,y,z) coordinates of the transferred B cell centroids. The arrest coefficient was calculated as the percentage of time per track that a cell moved < 4 μm/min, using a custom script for Matlab (MathWorks) (Moalli et al., 2014). The motility coefficient was calculated from the slope of a graph of displacement against time 1/2 .

Activation, Proliferation and Survival Assays
Splenic B cells were purified by negative depletion as described above and

Luminex Cytokine Secretion
A 4-Plex ProcartaPlex kit (Thermo Fisher Scientific, Inc.) was used to measure IL-6, IL-10, TNF-a and VEGF-A using Luminex-based technology according to the manufacturer's instructions on a Bio-Plex 200 (Bio-Rad Laboratories).

Antigen Presentation to OT-II T Cells
0.2 µm streptavidin coated polystyrene microspheres (Bangs Laboratories) were incubated at 37°C for 1 h with biotinylated anti-IgM at a final concentration of 0.561 µg/10 9 microspheres and biotinylated OVA (Nanocs, Inc.) at a final concentration of 0.219 µg biotin-OVA/10 9 microspheres. Conjugated microspheres were counted using a ZE5 cell analyzer (Bio-Rad Laboratories, Inc.) and resuspended in DMEM+ at a bead concentration of 10-25 x 10 8 /ml. Splenic B cells were isolated as described before and naïve splenic CD4 + T cells were isolated from OT-II mice were isolated using biotinylated antibodies against B220, CD8, CD11b, CD11c, CD19, CD25 and CD44. T cells were labelled with 1 μM CFSE. Splenic B cells were incubated with conjugated microspheres at a ratio of 10-25 beads per B cell for 30 min at 37°C. Excess microspheres were washed away. Co-cultures of 2 x 10 5 B cells and 10 5 OT-II CD4 + T cells were cultured in 96-well U-bottom plates at 37°C for 24, 48 and 72 h. Flow cytometry was used to determine proliferation and upregulation of activation markers.

B-T Conjugation
Splenic B cells were isolated using negative depletion as described above, and naïve splenic and lymph node CD4 + T cells from OT-II bone marrow chimeras were isolated using negative depletion as described above. The B cells were incubated with the indicated concentrations of OVA323-339 for 30 min at 37°C in FCS-free medium. The peptide-loaded B cells were mixed with the OT-II T cells at a 1:1 ratio and incubated at 37°C for 1 h, fixed with 4% PFA and pipetted 12 times to remove non-specific conjugates. Flow cytometry was used to determine conjugation frequency defined as the frequency of B cells in a conjugate with a T cell.
Goat F(ab')2 anti-mouse Igk (Southern Biotech) was biotinylated with 20-fold molar excess of NHS-LC-LC-biotin (Thermo Fisher Scientific, Inc.) and labelled with Cy3 monoreactive dye pack (GE Healthcare) in sodium carbonate buffer (biotinylated antik-Cy3). Excess dye was removed using Zeba 7K MWCO desalting columns (Thermo Fisher Scientific, Inc.). Splenocytes were incubated with biotinylated anti-k-Cy3 for 30 min on ice. Samples were incubated at 37°C for the times indicated before fixation with 2% PFA on ice for 20 min. Cells were stained with fluorescently-labelled streptavidin to label the biotinylated anti-k-Cy3 on the surface. Mice were injected i.p. with 50 µg of either chicken gamma globulin (CGG) or NP-CGG (BioSearch Technologies) with 25% Alum (Thermo Fisher Scientific, Inc.) in PBS. The immune response was assessed by either withdrawing blood from superficial veins of the mouse tail or the mice were sacrificed at the indicated time points and spleens were used for flow cytometric analysis.

ELISA
Blood harvested from the superficial tail vein was left to clot for at least 30 min before centrifugation at 17000 g for 10 min at room temperature. The supernatant was transferred to a new tube and centrifuged again at 17000 g for 10 min. Finally, the supernatant was transferred to a new tube and stored at -80°C until the ELISA was performed. 96-well Maxisorp Immunoplates (Nalge Nunc International Corporation) were coated with 5 µg/ml NP20-BSA (Santa Cruz Biotechnology, Inc.) overnight at 4°C and then washed five times with PBS, 0.01% Tween-20. Plates were blocked with 3% BSA in PBS for 2 h at room temperature and washed twice with PBS, 0.01% Tween-20. Serum was diluted 1:100 in PBS for the IgM ELISA and 1:1000 for the IgG1 ELISA and a further serial dilution was performed with 2-fold dilution steps. Diluted sera were added to the coated plates and incubated overnight at 4°C, followed by three washes with PBS, 0.01% Tween-20. Biotin-conjugated goat anti-mouse IgM (1.6 µg/ml, CALTAG laboratories, Thermo Fisher Scientific, Inc.) or biotin-XX conjugated goat anti-mouse IgG1 (4 µg/ml, Invitrogen; Therrmo Fisher Scientific, Inc.) were added to their respective plates and incubated at room temperature for 2 h. After three washes, plates were incubated with peroxidase labelled streptavidin (1 µg/ml, Vector Laboratories) for 2 h at room temperature. Plates were washed five times before incubation with 1x TMB ELISA substrate solution (eBioscience; Thermo Fisher Scientific, Inc.) and the reaction was stopped with 1 M H2SO4. Signals were quantified using absorption at 450 nm on a SpectraMax 190 (Molecular Devices LLC). Data was analyzed using Microsoft Excel and GraphPad Prism 8 (GraphPad). Values in the linear portion of the response curve were used to calculate concentrations of NPspecific antibodies in arbitrary units.

In Vivo Activation and Proliferation
(B6 x B6.SJL)F1 mice were injected i.p. with 50 µg of ovalbumin (Invivogen) with 25% Alum (Thermo Fisher Scientific, Inc.) in PBS. Three days later, splenic B cells were isolated by negative depletion, labelled with CMFDA (1 μM, Thermo Scientific, Inc.) and incubated with microspheres conjugated with either OVA, anti-IgM or both as described above. These cells were then injected intravenously into the pre-immunized (B6 x B6.SJL)F1 mice at 500,000 cells per mouse. Mice were sacrificed three days later and spleens were analyzed by flow cytometry.

HEL-OVA conjugation
HEL and OVA (both from Sigma) was mixed in a 1:1 ratio (w/w) in PBS containing 0.023% glutaraldehyde and stirred at room temperature for 1 h. After removing precipitate by centrifugation, the solution was dialysed against PBS (Dialysis tubes 8 kDa molecular weight cutoff Mini Dialysis kit, Cytiva) overnight at 4°C as described (Allen et al., 2007). HEL-OVA conjugation was confirmed by SDS-PAGE and silver staining. Conjugated HEL-OVA solution was further purified from unconjugated proteins using a 50 kDa Amicon filter tube (Merck).

Statistical Analysis
All statistical comparisons were carried out using the nonparametric two-tailed Mann-Whitney test or a Two-Way ANOVA test as detailed in Figures (Prism 8, GraphPad).

Author contributions
DH, LV, SW, SS, HH carried out experiments and analyzed data, DH and VLJT designed the study and wrote the manuscript, RK, JVS and VLJT supervised the work.

Acknowledgments
We thank Edina Schweighoffer for technical advice. We thank Pavel Tolar, George

Declaration of Interests
The authors declare no competing interests.          . Expression is measured as transcripts per million (TPM). Sample size, n=5.

Video S2
A time-lapse video showing migration of B cells in lymph node follicles analyzed by two-photon intravital microscopy 24 h after transfer of control (white) and WNK1deficient (blue) B cells into C57BL/6J mice from the experiment described in Figure   3E. Analysis of B cell tracks from these videos is shown in Figures 3F-3I and S3C and S3D.

Video S3
Video from two-photon intravital microscopy of the experiment described in Figure 7E

Video S4
Video from two-photon intravital microscopy of the experiment described in Figure 7E