Research ArticleImmunology

Diacylglycerol kinase ζ promotes actin cytoskeleton remodeling and mechanical forces at the B cell immune synapse

See allHide authors and affiliations

Science Signaling  14 Apr 2020:
Vol. 13, Issue 627, eaaw8214
DOI: 10.1126/scisignal.aaw8214

Balancing B cell signals

Diacylglycerol kinases (DGKs) catalyze the conversion of the lipid signaling molecule DAG into phosphatidic acid (PA), both of which have distinct effects. Merino-Cortés et al. found that the DGKζ isoform was critical in promoting signaling required for the optimal function of B cells. Compared with wild-type B cells, DGKζ-deficient B cells exhibited decreased mechanical forces at the plasma membrane, impairing the ability of these cells to form stable contacts with antigen-presenting cells, extract antigen, and present it to T cells. As a result, mice with DGKζ-deficient B cells showed diminished antibody responses in germinal centers. These data suggest that DGKζ regulates the balance in signaling between DAG and PA that is required for optimal B cell function.

Abstract

Diacylglycerol kinases (DGKs) limit antigen receptor signaling in immune cells by consuming the second messenger diacylglycerol (DAG) to generate phosphatidic acid (PA). Here, we showed that DGKζ promotes lymphocyte function–associated antigen 1 (LFA-1)–mediated adhesion and F-actin generation at the immune synapse of B cells with antigen-presenting cells (APCs), mostly in a PA-dependent manner. Measurement of single-cell mechanical force generation indicated that DGKζ-deficient B cells exerted lower forces at the immune synapse than did wild-type B cells. Nonmuscle myosin activation and translocation of the microtubule-organizing center (MTOC) to the immune synapse were also impaired in DGKζ-deficient B cells. These functional defects correlated with the decreased ability of B cells to present antigen and activate T cells in vitro. The in vivo germinal center response of DGKζ-deficient B cells was also reduced compared with that of wild-type B cells, indicating that loss of DGKζ in B cells impaired T cell help. Together, our data suggest that DGKζ shapes B cell responses by regulating actin remodeling, force generation, and antigen uptake–related events at the immune synapse. Hence, an appropriate balance in the amounts of DAG and PA is required for optimal B cell function.

INTRODUCTION

Diacylglycerol kinases (DGKs) convert the lipid diacylglycerol (DAG) into phosphatidic acid (PA), shaping the pools of both second messengers. There are 10 mammalian DGK isoforms, which are classified into five subgroups based on their distinct regulatory domains. DGKs are soluble enzymes that translocate to specific cellular locations to regulate the relative amounts of DAG and PA (1, 2). Enrichment of DAG or PA at the plasma membrane supports the localized recruitment of effector proteins. DAG-dependent effectors include conventional protein kinase C (PKC), PKD, and Ras guanyl nucleotide–releasing protein (RasGRP), which drive the activation of nuclear factor κB (NF-κB) and extracellular signal–regulated kinase 1 and 2 (ERK1/2) signaling cascades and subsequent gene expression. PA acts as a lipid anchor for distinct effectors through the binding of its negatively charged region to cationic regions on those proteins. PA-binding proteins, such as the Rac activator dedicator of cytokinesis protein 1 (DOCK1), Rho GDP-dissociation inhibitor (RhoGDI), atypical PKCζ, and partitioning defective protein 3 (Par3), are involved in cytoskeletal remodeling and cell polarity (1, 2).

In immune cells, DGKs are well known for limiting the intensity of DAG-regulated signals downstream of antigen receptor stimulation. DGKα and DGKζ are the most studied isoforms, both of which are expressed in B and T cells (1, 3). Recognition by lymphocytes of antigen on the surface of antigen-presenting cells (APCs) triggers the formation of the immune synapse at the interface between the lymphocyte and the APC. Establishment of the immune synapse requires actin cytoskeletal remodeling and protein segregation into two concentric regions: the central supramolecular activation cluster (cSMAC), which is characterized by the central accumulation of antigen-bound antigen receptors together with certain signaling molecules, and the peripheral ring-shaped domain [peripheral SMAC (pSMAC)], which is enriched in the integrin lymphocyte function–associated antigen 1 (LFA-1) [bound to its ligand intercellular adhesion molecule-1 (ICAM-1)], filamentous actin (F-actin), and other proteins involved in adhesion and cytoskeletal rearrangements, including vinculin, talin, and Wiskott-Aldrich syndrome protein (WASP) (47). In T cells, phospholipase C-γ (PLC-γ) generates a localized DAG pool at the immune synapse to trigger downstream signaling (8). Both DGKα and DGKζ translocate to the immune synapse to regulate DAG abundance, thus decreasing the intensity of T cell receptor (TCR) signaling (9, 10). DGKζ-deficient B cells show enhanced activation of the Ras-ERK1/2 pathway after stimulation of the B cell receptor (BCR), leading to increased B cell responses (3).

In nonimmune cells, DGKs participate in actin cytoskeletal rearrangements, cell polarity, and integrin recycling. DGKα-mediated PA generation at the plasma membrane recruits PKCζ, which phosphorylates RhoGDI. This promotes the release and activation of Rac1 and, thus, actin polymerization for the generation of invasive protrusions by epithelial cells (11). Similarly, DGKζ-produced PA facilitates Rac1 activation through the p21-activated kinase 1 (PAK1)–mediated phosphorylation of RhoGDI in neuronal and skeletal muscle cells (12, 13). PA generation by DGKs stimulates integrin recycling and tumor invasiveness through a Rab11-dependent pathway (14). PA also targets phosphatidylinositol-4-phosphate 5-kinase I (PIP5KI), promoting its lipid kinase activity to produce phosphatidylinositol-4,5-bisphosphate (PIP2) at the plasma membrane (15, 16); PIP2 is a substrate for PLC-γ and phosphatidylinositol 3-kinase (PI3K), and promotes adhesion and actin dynamics (17). Actomyosin reorganization, integrin clustering, and polarized membrane trafficking all occur at the immune synapse. DGKs are linked to T cell polarization events because microtubule-organizing center (MTOC) translocation and polarized secretion at the immune synapse are impaired in the absence of DGKα or DGKζ (18, 19). Nonetheless, PA-related DGK functions at the immune synapse are largely unknown.

Here, we investigated the roles of DGKs in the assembly of the B cell immune synapse. We used primary mouse B cells deficient in DGKζ (DGKζ−/−) or DGKα (DGKα−/−) or treated with a DGK inhibitor. In addition, we used a B cell line overexpressing green fluorescent protein (GFP)–tagged DGKζ constructs. We found that DGKζ promoted LFA-1–mediated adhesion and F-actin accumulation at the immune synapse mainly through PA generation and that the DOCK2 and PAK1 regulation of Rac activity was also involved. Furthermore, we used traction force microscopy (TFM) and micropipette force probe (MFP) technique to study single-cell force generation at the immune synapse (20, 21). We detected decreased mechanical forces for DGKζ−/− B cells and inhibitor-treated B cells. Forces are critical to acquire antigen at the B cell immune synapse (22). These results, together with the finding of impaired myosin activation and MTOC translocation to the immune synapse, correlated with the diminished ability of DGKζ-defective B cells to extract antigen and present it to T cells in vitro. In immunocompetent mice, DGKζ−/− B cells exhibited reduced germinal center (GC) responses compared with those of wild-type (WT) B cells. Our data suggest pivotal functions for DGKζ in cytoskeletal remodeling, mechanical force generation, and antigen uptake at the immune synapse to determine B cell responses.

RESULTS

DGKζ stimulates LFA-1–mediated adhesion and F-actin accumulation at the B cell immune synapse

We first analyzed the relative amounts of DGKα and DGKζ in B cells by Western blotting. Both isoforms were detected in WT B cells (fig. S1A), which is consistent with previous findings at the RNA level (3). Treatment with the DGK inhibitor R59949 (R59) had no substantial effect on the abundance of either DGK (fig. S1A). We investigated the ability of DGKζ−/−, DGKα−/−, or R59-treated B cells to trigger immune synapse formation and maturation compared with that of WT B cells. For inhibitor experiments, the B cells were pretreated with 10 μM R59 for 30 min at 37°C and washed before use. We used artificial planar lipid bilayers that contained the glycosylphosphatidylinositol (GPI)–linked adhesion molecule ICAM-1, various densities of tethered surrogate antigen [su-Ag; anti-κ LC antibody (Ab)], and were coated with the chemokine CXCL13. This system mimics an APC surface and was used to evaluate immune synapse formation by confocal microscopy (4). Splenic B cells were isolated by negative selection (<90% CD19+). WT and DGKζ−/− B cells showed similar cell surface amounts of immunoglobulin M (IgM) and IgD, whereas DGKα−/− B cells displayed slightly increased amounts of IgM (fig. S1, B and C). B cells were left in contact with the lipid bilayer for 10 min at 37°C and then were imaged. The frequency of B cells that formed an immune synapse was analyzed on the basis of two criteria: the formation of a central su-Ag cluster (cSMAC) and of a cell contact with the substrate (immune synapse contact area), which were estimated by su-Ag–associated fluorescence and by interference reflection microscopy (IRM), respectively. At a density of su-Ag of 20 molecules/μm2, we found a small increase in the percentage of DGKζ−/− B cells that formed an immune synapse compared with controls (Fig. 1, A and B). R59-treated B cells showed a similar tendency (Fig. 1, A and B). In immune synapse–forming B cells, contact areas (estimated by IRM) were statistically significantly reduced in DGKζ−/− and R59-treated B cells compared with those for WT B cells (Fig. 1, C and D). In contrast, the area and the total quantity of su-Ag accumulated at the immune synapse (both estimated by fluorescence) were comparable between DGKζ−/−, R59-treated, and WT B cells (Fig. 1, C and D). Similar results were obtained when lower su-Ag densities were used at the lipid bilayer (fig. S1, D to I).

Fig. 1 DGKζ dysfunction alters LFA-1–mediated adhesion and F-actin content at the B cell synapse.

(A to H) B cells were allowed to settle on ICAM-1–coated and CXCL13-coated planar lipid bilayers loaded with su-Ag (20 molecules/μm2) for 10 min before being imaged or fixed for immunofluorescence analysis. (A) Differential interference contrast (DIC), IRM, and fluorescence su-Ag images at the contact plane of representative immune synapse–forming WT, DGKζ−/−, and R59-treated B cells. (B) Percentages of the indicated cells that exhibited immune synapse formation. (C and D) Contact area (left), su-Ag central cluster area (cSMAC, middle), and total su-Ag fluorescence (FL) in arbitrary units (AU, right) for (C) DGKζ−/− B cells and (D) R59-treated B cells with established immune synapses compared with WT B cells. Each dot in (B) represents a single image field, whereas each dot in (C) and (D) represents a single cell. (E) DIC and FL images of F-actin (white) for representative immune synapse-forming WT, DGKζ−/−, and R59-treated B cells, which were fixed at 10 min. (F) Values of total F-actin FL at the immune synapse in each case and in the presence of distinct su-Ag densities (20 and 5 molecules/μm2). (G) DIC and FL images of vinculin (green) and su-Ag (red) for representative immune synapse-forming WT, DGKζ−/−, and R59-treated B cells, which were fixed at 10 min. (H) Values of total vinculin FL at the immune synapse in each case and in the presence of distinct su-Ag densities (20 and 5 molecules/μm2). Each dot in (F) and (H) represents a single cell. Data in (B) and (H) are pooled from two experiments from a total of four experiments. Data are representative of three (C and D) or four (F) experiments. Scale bars, 2.5 μm. *P < 0.05, **P < 0.01, ****P < 0.0001 by two-tailed, unpaired Student’s t test.

Because the contact area is the sum of the su-Ag central cluster area (cSMAC) and the surrounding region of LFA-1 interactions with ICAM-1 (pSMAC), these data imply that impaired DGKζ function caused pSMAC defects. We analyzed other pSMAC features, namely, vinculin and F-actin content, at the lipid bilayers by immunofluorescence. DGKζ−/− B cells had less vinculin and F-actin at the pSMAC than did controls for the su-Ag densities tested, and the reductions were greater for R59-treated B cells (Fig. 1, E to H). In contrast, DGKα−/− B cells showed unimpaired immune synapse formation (pSMAC/cSMAC; vinculin/F-actin content) (fig. S2). We then centered our study on the DGKζ isoform. Lack of DGKζ did not affect the abundance of DGKα protein (fig. S1A). We also determined that a su-Ag density of 20 molecules/μm2 was optimal for immune synapse formation, so this concentration was used going forward.

We set out to study the effect of an excess of DGKζ activity on immune synapse formation. A20 B cells were transiently transfected with plasmids encoding GFP-tagged DGKζ-WT or a kinase-deficient mutant (DGKζ-KD). GFP-expressing and nontransfected (GFPneg) A20 cells were included as controls (fig. S3A). Using the aforementioned experimental approach, we found that the percentage of cells expressing either DGKζ construct that formed an immune synapse was decreased (fig. S3, B and C). Whereas the immune synapse contact areas were larger in the DGK-overexpressing cells, there were no differences in su-Ag cluster area or the total quantity of su-Ag (fig. S3, D and E). By immunofluorescence microscopy, we detected a statistically increased amount of F-actin at the pSMAC of A20 B cells overexpressing DGKζ-WT but not at the pSMAC of cells expressing DGKζ-KD (fig. S3F). For vinculin, we observed increased frequency of A20 cells overexpressing either DGKζ-WT or DGKζ-KD with a well-formed ring as well as increased vinculin abundance (fig. S3G). These data, thus, suggest a role for DGKζ in mediating LFA-1–mediated adhesion, vinculin recruitment, and increased F-actin content at the B cell immune synapse.

DGKζ-derived PA shapes LFA-1–mediated adhesion and the DOCK2–Rac–F-actin pathway at the B cell immune synapse

We next investigated whether an excess of PA could rescue the defects in LFA-1–mediated adhesion and F-actin abundance caused by impaired DGKζ function. To do so, we allowed WT and DGKζ−/− B cells, untreated or treated with R59, to form an immune synapse and then added 0.1 mM PA to the medium. After 30 min of PA exposure, we imaged the cells and detected larger immune synapse contact areas in all instances (Fig. 2, A and B). The su-Ag area values and total quantities of su-Ag at the immune synapse were reduced after PA treatment (fig. S4, A and B). By immunofluorescence microscopy, we detected increased F-actin content at the immune synapse of PA-exposed B cells (Fig. 2, C and D). Increasing the abundance of PA, thus, resulted in enhanced LFA-1–mediated adhesion and actin polymerization at the immune synapse and altered su-Ag central cluster dynamics.

Fig. 2 PA generation promotes LFA-1–mediated adhesion and F-actin polymerization.

(A to D) The indicated B cells were in contact for 10 min with ICAM-1–containing and CXCL13-containing planar lipid bilayers loaded with su-Ag (20 molecules/μm2) to establish immune synapses, imaged, and then exposed to 0.1 mM PA for 30 min and either imaged or fixed for immunofluorescence. (A) DIC and IRM images for representative WT and DGKζ−/− B cells that were left untreated or were treated with R59 before (none) and after PA exposure. (B) Contact areas for B cells from the experiments shown in (A). (C) DIC and FL images of F-actin for WT and DGKζ−/− B cells that were left untreated or were treated with R59 before (none) and after PA exposure. (D) Values of total F-actin FL at the immune synapse in the indicated cells from the experiments shown in (C). (E) Left: DIC and FL images of F-actin for representative WT and DGKζ−/− B cells, which were left untreated or were treated with LY294002 and then fixed 10 min after contact with planar bilayer as described in (A). Right: Values of total F-actin FL at the immune synapse in the indicated cells. (F) Left: DIC and FL images of F-actin for representative WT and PI3Kδ KD B cells, which were left untreated or were treated with R59. Right: Values of total F-actin FL at the immune synapse in the indicated cells. (G) Left: DIC and FL images of DOCK2-GFP for representative immune synapse–forming DOCK2-GFP knock-in B cells, which were left untreated or were treated with R59, after 10 min in contact with a su-Ag–loaded (20 molecules/μm2) planar bilayer. Right: Values of total DOCK2-GFP FL at the immune synapse in the indicated cells. (H) Top: Untreated and R59-treated (1 hour) WT and DGKζ−/− B cells as well as transfected A20 B cells sorted for the nontransfected (GFPneg; none) or for the expression of GFP-DGKζ-WT (DGKζ-WT) were analyzed by Western blotting with specific antibodies against the indicated proteins. For primary B cells, lysates from three mice of each genotype are shown. Bottom: Quantification of phosphorylated-PAK1 (p-PAK1) and p-PAK2 band intensities, which were normalized to that of α-tubulin (α-Tub), which was used as a loading control. Each dot in (B) to (G) represents a single cell. Data in (B) and (D) pooled from two experiments from a total of four experiments. Data are representative of three (E and F) and two (G) experiments. Data in (H) are means ± SD of three mice and of three A20 cell transfection experiments. Scale bars, 2.5 μm. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed, unpaired Student’s t test.

Previous studies noted the relevance of PI3K-derived phosphatidylinositol-3,4,5-trisphosphate (PIP3) in regulating F-actin ring assembly at the T cell immune synapse. PIP3 recruits DOCK2 to the periphery of the immune synapse, which promotes actin polymerization through the Rac guanosine 5′-triphosphatases (GTPases) (23). Thus, we addressed the interplay between PIP3 and DGKζ-derived PA in B cell immune synapse formation. PI3K activity in WT and DGKζ-impaired (knockout and R59-treated) B cells was assessed by measuring Akt phosphorylation after BCR stimulation with anti-mouse IgM Ab-coated plates. We found no statistically significant differences in the abundance of Akt phosphorylated at Ser473 (Akt-pSer473) between the three cell types (fig. S5A). We treated WT and DGKζ−/− B cells with 10 μM LY294002 (a PI3K inhibitor) for 30 min at 37°C and then allowed them to settle on lipid bilayers for immune synapse formation. Treatment with this inhibitor caused a reduction in immune synapse area and in su-Ag cluster size in both cell types (fig. S5, B to D), as well as reducing total F-actin content (Fig. 2E). Within class I PI3Ks, the PI3Kδ isoform (which contains the p110δ catalytic subunit) was previously identified as the major contributor for PIP3 production to regulate F-actin remodeling at the immune synapse (23). Thus, we isolated splenic B cells from knock-in mice expressing a kinase-deficient p110δ catalytic subunit (PI3Kδ KD), treated them with or without R59, and then evaluated immune synapse formation. PI3Kδ KD B cells displayed reduced immune synapse contact area and F-actin content compared with WT control B cells (fig. S5, E and F, and Fig. 2F). We also detected increased su-Ag aggregation in PI3Kδ KD B cells (fig. S5, E and G). Treatment of PI3Kδ KD B cells with R59 further decreased immune synapse area and F-actin content compared with that in untreated cells, without modifying su-Ag clustering (fig. S5, E to G, and Fig. 2F).

We studied DOCK2 recruitment to the immune synapse in splenic B cells isolated from DOCK2-GFP knock-in mice, which were left untreated or were treated with R59. We observed a ring-shaped DOCK2-GFP structure at the immune synapse, and quantification of total DOCK2-GFP fluorescence at the immune synapse plane revealed a statistically significant reduction in B cells that were treated with R59 (Fig. 2G). DGKζ-dependent PAK1 activation promotes RhoGDI/Rac dissociation and, thus, Rac activation (24). We assessed PAK1/2 activation by measuring the relative amounts of phosphorylated PAK1/2 (p-PAK1/2) by Western blotting. We found that DGKζ−/− and R59-treated B cells had lower amounts of p-PAK1/2 than that of controls, whereas the overexpression of DGKζ-WT in A20 B cells resulted in increased p-PAK1/2 abundance (Fig. 2H). PA did not increase the amount of p-PAK1/2 in WT B cells (fig. S4C). This suggests that DGKζ associates with the PAK/RhoGDI complex and promotes its activation, as was previously described for fibroblasts (24). Hence, our data suggest that DGKζ promotes actin polymerization at the B cell immune synapse by increasing Rac function in a DOCK2- and PAK1/2-dependent manner.

DGKζ stimulates mechanical forces at the B cell immune synapse

Several studies revealed the relevance of mechanical forces at the immune synapse for B and T cell effector function (20, 25). The robust actin polymerization and remodeling at the immune synapse induces force generation. LFA-1 and antigen receptors act as mechanosensitive proteins because their function and signaling properties are shaped by these mechanical forces. DGKζ−/− B cells and R59-treated WT B cells had defects in LFA-1–mediated adhesion and F-actin abundance at the immune synapse. We asked whether these defects affected the mechanical forces generated at the synapse of these B cells compared with those of WT B cells. To do that, we used two complementary methods: TFM and MFP. We used TFM to measure the forces exerted by B cells when they were in contact with polyacrylamide (PAA) hydrogels loaded with su-Ag alone or in combination with ICAM-1-Fc. Displacements of the fluorescent microbeads embedded on the hydrogel, monitored over time, enabled us to calculate the magnitude of the applied forces (fig. S6A) and the cell strength on the substrate at each time point. Traction energy values were statistically significantly greater in the presence of ICAM-1 at the substrate compared with the values in the presence of su-Ag alone (Fig. 3, A and B, and movies S1 and S3), pointing to the importance of LFA-1–mediated adhesion for force generation at the immune synapse. DGKζ−/− B cells exhibited reduced traction forces compared with those of control B cells in the presence of both su-Ag and ICAM-1 (Fig. 3, A and B, and movies S2 and S4).

Fig. 3 Mechanical force generation at the B cell immune synapse is mediated by DGKζ.

(A) B cells were allowed to settle on PAA gels coated with su-Ag alone or with ICAM-1–Fc and then were monitored for up to 15 min. Time-lapse color maps of stress (in pascal) for representative WT and DGKζ−/− B cells on PAA gels under the indicated conditions are shown. (B) Left: Average values of synaptic traction forces (in joules) over time for WT and DGKζ−/− B cells under the indicated conditions. Each solid line corresponds to the mean of 25 to 30 measured cells; dotted lines represent ± SD (confidence interval). Right: Average value of synaptic traction forces per cell over time. Each dot represents a single cell. Data are pooled from three experiments. (C to J) B cells were monitored by MFP while in contact with silica beads that were coated with ICAM-1–containing lipid bilayers and either unloaded (none) or loaded with su-Ag (100 molecules/μm2). (C) Bright-field microscopy images (processed using a high-pass filter for better visualization using ImageJ software) for representative WT and DGKζ−/− B cells that were activated by a su-Ag–loaded bead. In both examples, the cells are submitted to an oscillatory force of 50-pN average, 25-pN amplitude, and 1-Hz frequency. (D) Time trace of the su-Ag–loaded bead position (Xbead, in micrometers) for WT and DGKζ−/− B cells. Each line corresponds to the average value of 10 cells. Data are from a single experiment that is representative of three independent experiments. (E) Values of pushing speed (in micrometers per second) per cell when in contact with the su-Ag–loaded bead. Each dot represents a single cell. Data are from a single experiment that is representative of three independent experiments. (F to G) Mechanical changes during activation. (F) Values of Young’s modulus (in pascal) per cell when contacting beads under the indicated conditions (none, in the absence of tethered su-Ag). Each dot represents a single cell. Data are pooled from three experiments. (G) Left: Time evolution of cell stiffness K′ (in nanonewtons per micrometer) averaged over cells. Each line corresponds to the average value of 20 cells. Data are pooled from two experiments (n = 3). Right: Cell stiffness K′ averaged over 250 s after cell contact with a bead. Each dot represents a single cell. Data are pooled from three experiments. (H) Bright-field microscopy images processed as described for (C) for representative untreated and R59-treated WT B cells activated by a su-Ag–loaded bead, as described in (C). (I) Values of Young’s modulus were determined as described in (F) but for untreated and R59-treated (R59) WT B cells. (J) Left: Time evolution of cell stiffness K′ (in nanonewtons per micrometer) averaged over untreated and R59-treated WT B cells. Each line corresponds to the average value of 30 cells. Data are pooled from three experiments. Right: Cell stiffness K′ averaged over 250 s after cell contact with a bead. Each dot represents a single cell. Data are pooled from three experiments. Data in (I) and (J) are pooled from three experiments. Scale bars, 5 μm. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed, unpaired Student’s t test.

To define the three-dimensional components of the forces involved, we delineated B cell mechanical behavior and quantified the forces generated at the immune synapse over time by MFP (21). In this technique, a bead coated with stimulatory ligands is aspirated at the tip of a flexible micropipette used as a sensitive force transducer, and brought in contact with the cell, which is aspirated at the tip of another micropipette (fig. S6B). We used silica beads (5-μm diameter) coated with lipid bilayers containing GPI–ICAM-1 and tethered su-Ag. After contact with the stimulatory bead, WT B cells pushed it away during the first 40 s (positive values of bead displacement, Xbead, relative to initial bead position) (Fig. 3, C and D, fig. S6C, and movies S5 and S6) at a pushing speed of 0.025 ± 0.010 μm/s (Fig. 3E). DGKζ−/− B cells showed a reduced pushing phase that correlated with lower pushing speed values (0.015 ± 0.005 μm/s; Fig. 3, C to E, and movie S7). After the pushing phase, WT B cells pulled on the bead (Xbead reached negative values) and formed a cup-like structure on it (Fig. 3, C and D, fig. S6C, and movie S6), whereas the pulling ability of the DGKζ−/− B cells was decreased (Fig. 3D). MFP also enables the measurement of cell rigidity just upon cell contact with the bead, before the pushing phase begins (Young’s modulus parameter) (21). We found that values were higher with su-Ag compared with nonantigen (Fig. 3F), indicating increased cell stiffness after BCR stimulation. Young’s modulus values for DGKζ−/− B cells were lower than for WT in presence of su-Ag (Fig. 3F). To measure cell mechanical changes at the immune synapse, we monitored B cell elastic properties by quantifying cell stiffness through the K′ parameter. Compared with WT B cells, DGKζ−/− B cells had lower K′ values (Fig. 3G), indicating an impaired ability to undergo cytoskeletal remodeling, which changed their mechanical properties upon immune synapse formation. Similarly, we detected mechanical defects in R59-treated B cells (Fig. 3, H to J, and movie S8). Therefore, data obtained from both TFM and MFP experiments suggest that DGKζ is required for the mechanical properties and force generation at the B cell immune synapse.

DGKζ activity limits immune synapse–triggered B cell activation

A previous study reported that the lack of DGKζ enhances activation of DAG-dependent pathways upon B cell stimulation with soluble antigen in vitro, with ERK1/2 activation, CD69 abundance, and cell proliferation being increased (3). We investigated the effects of impaired DGKζ function for B cell activation with regard to the immune synapse. We loaded WT, DGKζ−/−, and R59-treated B cells with a Ca2+-sensitive fluorescent probe and monitored Ca2+ influx during immune synapse formation by real-time fluorescence microscopy. Peak and sustained Ca2+ influx were enhanced in DGKζ−/− and R59-treated B cells compared with those in control cells (Fig. 4, A and B), which might be due to increased stimulation of DAG-dependent Ca2+ channels (26). We evaluated ERK1/2 activation by measuring the relative amounts of phosphorylated ERK1/2 (p-ERK) at the immune synapse by immunofluorescence (Fig. 4C). We found that DGKζ absence or inhibition resulted in increased p-ERK abundance (Fig. 4, C and D). DGKζ−/− B cells had similar amounts of p-ERK at the immune synapse as did R59-treated WT B cells. This finding supports the major role of the DGKζ isoform in limiting DAG-related signaling downstream of the BCR, as was previously reported (3). The lack of DGKα in B cells did not modify p-ERK abundance at the immune synapse compared with that of WT B cells (fig. S6D). The increased amount of p-ERK in R59-treated DGKζ−/− B cells compared with that in untreated cells implies the contribution of another DGK isoform in absence of DGKζ.

Fig. 4 DGKζ diminishes BCR-dependent B cell activation in the context of the immune synapse.

(A) Fluo-4FF–labeled WT, DGKζ−/−, and R59-treated B cells were monitored for Ca2+ influx at early times of immune synapse formation on ICAM-1–containing and CXCL13-containing planar lipid bilayers loaded with su-Ag (20 molecules/μm2). Fluorescence Fluo-4FF images of representative B cells over time are shown. (B) Left: Values of total Fluo-4FF FL (in AU) over time. Data are means ± SD of 30 B cells per condition. Right: To statistically compare the Fluo-4FF FL data, we calculated the area under the curve (AUC) per B cell and per condition. Each dot represents a single cell. (C) WT and DGKζ−/− B cells, which were left untreated (none) or were treated with R59, were in contact with su-Ag–loaded (20 molecules/μm2), ICAM-1–containing, and CXCL13-containing planar lipid bilayers for 10 min and then were fixed for immunofluorescence. DIC and FL images of phosphorylated ERK1/2 (p-ERK, green) for representative immune synapse–forming B cells are shown. (D) Values of total p-ERK FL at the immune synapse for the indicated cells. Each dot represents a single cell. (E and F) B cells were cultured on ICAM-1–containing and CXCL13-containg planar lipid bilayers, which were either unloaded (none) or loaded with su-Ag [20 molecules/μm2; su-Ag (20)], for 20 hours and then were collected for flow cytometry analysis. (E) Representative profiles of CD69, CD25, and CD86 staining for the indicated cells. (F) Percentages of B cells expressing CD69, CD25, or CD86 (left) and mean fluorescence intensity (MFI) values for these markers (right) in each condition and for each B cell type. (G and H) CFSE tracer–labeled WT and DGKζ−/− B cells, which were left untreated or were treated with R59, were cocultured with pseudo-APCs (silica beads coated with ICAM-1–containing and CXCL13-containing lipid bilayers), which were unloaded (none) or loaded with su-Ag [1000 molecules/μm2, su-Ag (1000)] at ratios of 1:1 and 1:5 in the presence of IL-4 for 72 hours. (G) Representative profiles of CFSE tracer for the indicated conditions. (H) Percentages of dividing B cells (as determined by monitoring CFSE dilution) in each condition and for each B cell type. Data in (D) are from a single experiment and are representative of two experiments. Each dot represents a cell. Data are pooled from two (B), four (F), and six (H) experiments. Scale bars, 2.5 μm. *P < 0.05, **P < 0.01, ****P < 0.0001 by two-tailed, unpaired Student’s t test.

We incubated B cells in contact with planar lipid bilayers, which were unloaded or su-Ag loaded, for 20 hours and then analyzed the cell surface expression of the activation markers CD69, CD25, and CD86 by flow cytometry. DGKζ−/− B cells expressed more of those markers at the cell surface than did WT B cells, although the increase was statistically significant only for CD69, and treatment with R59 had a similar effect (Fig. 4, E and F). To evaluate cell proliferation, we modified the experimental approach (fig. S7A) by using silica beads (5-μm diameter) coated with lipid bilayers containing GPI–ICAM-1, a CXCL13 coating, and tethered su-Ag, because these beads were suitable for longer coculture periods. We refer to these beads as pseudo-APCs. We increased the su-Ag density (1000 molecules/μm2) at the pseudo-APC surface to promote greater B cell proliferation, thus facilitating detection. WT, DGKζ−/−, and R59-treated B cells were stained with CFSE (carboxyfluorescein succinimidyl ester) and cocultured with pseudo-APCs at different ratios (1:1 and 1:5) in the presence of inerleukin-4 (IL-4) for 96 hours. Compared with WT B cells, DGKζ−/− B cells showed increased proliferation, although treatment of these cells with R59 did not lead to further changes (Fig. 4, G and H; for the gating strategy, see fig. S7A).

DGKζ deficiency diminishes the antigen presentation capacity of B cells in vitro

B cell immunity against T cell–dependent antigens entails antigen acquisition, degradation, and presentation to T cells in the form of antigenic peptides by the major histocompatibility complex (MHC) class II complex. In this process, B cells receive T cell help, mainly through CD40 stimulation, which triggers B cell survival, proliferation, and class switching. We investigated the role of DGKζ in the molecular events related to antigen acquisition, processing, and presentation. MTOC polarization to the immune synapse supports the membrane trafficking needed for these events (27). We incubated WT, DGKζ−/−, and R59-treated B cells with pseudo-APCs, unloaded or loaded with su-Ag, at a ratio of 1:1 for 30 min at 37°C, and then fixed the cells and analyzed MTOC location by γ-tubulin staining. The distance of the MTOC from the immune synapse for each B cell was measured and normalized to the cell diameter. We found that su-Ag promoted MTOC relocalization in most WT B cells (70%), whereas this was reduced in DGKζ-defective cells (20% in DGKζ−/− B cells; 30% in R59-treated B cells) (Fig. 5, A and B). The nonmuscle motor protein myosin II is involved in antigen extraction at the B cell immune synapse (22). Therefore, we used Western blotting to analyze phosphorylation of the regulatory subunit myosin light chain (MLC) after BCR triggering with Ab-coated plates and found that DGKζ−/− and R59-treated B cells had impaired MLC activation compared with that of control cells (Fig. 5C).

Fig. 5 MTOC translocation, myosin activation, and antigen presentation ability are reduced in DGKζ−/− and R59-treated B cells.

(A) The indicated B cells were mixed with unloaded (none) or su-Ag–loaded [20 molecules/μm2; su-Ag (20)] pseudo-APCs at a 1:1 ratio, cultured for 30 min on poly-l-lysine–coated coverslips, and then fixed for immunofluorescence. DIC and FL γ-tubulin (γ-tub; green) images are shown for representative B cell–pseudo-APC conjugates for each indicated condition. Dashed circle, pseudo-APC. Scale bar, 2.5 μm. (B) Percentages of B cells from the experiments shown in (A) in the specified polarity index (PI) groups. The PI per B cell was estimated as the ratio of “a” and “b” distances (left). Data are means ± SD of 40 B cells in each case. (C) Left: WT, DGKζ−/−, and R59-treated B cells stimulated on Ab-coated plates for the indicated times were analyzed by Western blotting with specific antibodies against the indicated proteins. Right: Quantification of phosphorylated-MLC (p-MLC) band intensity was normalized to that of the loading control α-tubulin (α-Tub). (D and E) WT, DGKζ−/−, and R59-treated B cells were cultured for 2 hours in contact with ICAM-1–containing and CXCL13-containing planar lipid bilayers, which were coated with Alexa Fluor 647–streptavidin (strep) and were either unloaded (none) or loaded with su-Ag/OVA (su-Ag/OVA; 2500 molecules/μm2). The B cells were then collected, treated with trypsin for 5 min, and analyzed by flow cytometry for strep fluorescence as a readout of su-Ag/OVA extraction. (D) Representative profiles of strep for each case. (E) Percentages of strep+ B cells (left) and mean fluorescence intensity (MFI) strep values (right) in the presence of su-Ag/OVA for each B cell type. Data are pooled from three experiments (four WT and four DGKζ−/−mice were used in total). (F to I) Experiments were performed as described for (D), but, after 2 hours of incubation with the planar bilayers, the B cells were collected, washed, and cocultured with CFSE-labeled OTII CD4+ T cells at a ratio 1:1 for 72 hours. Cell culture medium was then collected to analyze secreted IL-2, and the cells were analyzed by flow cytometry. (F) Left: Representative CD25 and CD4 dot plots for CD4-gated OTII T cells in each case and for each condition: B cells exposed to unloaded planar bilayers, none; B cells exposed to su-Ag/OVA-loaded bilayers, su-Ag/OVA. Right: Representative profiles of CFSE for CD4+CD25+ T cells in each case and for each condition. (G) Percentages of CD4+CD25+ T cells and (H) of dividing CD25+ T cells in each case. (I) Amounts of T cell–secreted IL-2 in the medium of the indicated cell cocultures. Data in (G) to (I) are means ± SD of three experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed, unpaired Student’s t test.

We evaluated B cell–mediated antigen extraction and presentation to T cells in vitro. We prepared planar lipid bilayers containing GPI–ICAM-1 and CXCL13 that were left unloaded or were loaded with a mixture of su-Ag and Alexa Fluor 488–conjugated ovalbumin (OVA) protein (see Materials and Methods). To quantify BCR-mediated antigen extraction, we measured the fluorescence intensity of the Alexa Fluor 647–conjugated streptavidin (strep) used to tether su-Ag and OVA to the lipid bilayer (see Materials and Methods). We also assessed OVA acquisition by monitoring Alexa Fluor 488 fluorescence intensity. We incubated WT B cells in the absence or presence of su-Ag and OVA at different densities (ranging from 20 to 2500 molecules/μm2) for 2 hours at 37°C, followed by collecting the cells, treating them with trypsin, and analyzing them by flow cytometry. We detected strep/OVA extraction at densities of 500 and 2500 molecules/μm2 (~30 and 70% strep+ B cells, respectively), which was dependent on BCR stimulation by tethered su-Ag (fig. S7B). We then evaluated the antigen extraction ability of DGKζ−/− and R59-treated B cells using the highest density to improve the detection of streptavidin. Strep+ B cell frequencies were similar, but the mean fluorescence intensity values were lower, for DGKζ−/− and R59-treated B cells compared with those of WT B cells (Fig. 5, D and E), which is suggestive of reduced antigen acquisition, although the difference was not statistically significant. The addition of PA did not modify the ability of WT B cells to extract antigen (fig. S7C).

To assess T cell antigen presentation, we incubated WT, DGKζ−/−, and R59-treated B cells in contact with unloaded or su-Ag/OVA–loaded planar lipid bilayers for 2 hours at 37°C. The B cells were then collected and cocultured with CFSE-labeled OTII CD4+ T cells at a 1:1 ratio (fig. S7D). The TCR of OTII CD4+ T cells recognizes OVA-derived peptides (residues 323 to 339) in the context of MHC class II (I-Ab) on the B cell surface, triggering T cell activation. After 72 hours, we evaluated the cell surface expression of CD25 as a marker of T cell activation, T cell proliferation (for the gating strategy, see fig. S7D), and the amount of IL-2 secreted into the cell culture medium. Using this system, we detected increased CD4+ CD25+ T cell frequencies (up to 25%) in presence of su-Ag/OVA compared with the control condition for WT B cells (Fig. 5, F and G), whereas the frequencies were statistically significantly reduced for DGKζ−/− and R59-treated B cells (Fig. 5, F and G). T cell proliferation and IL-2 production were reduced in the context of DGKζ−/− and R59-treated B cells (Fig. 5, F, H, and I). These results suggest that DGKζ stimulates antigen presentation by mediating antigen acquisition– and antigen processing–related molecular events at the immune synapse.

The absence of DGKζ in B cells impairs the GC response in vivo

We investigated whether the antigen presentation defects found in B cells with altered DGKζ function limited GC responses to T-dependent antigens in vivo. We isolated WT or DGKζ−/− B cells (CD45.2+) and adoptively transferred them to CD45.1+ immunocompetent recipient mice. One day later, the mice were immunized with 4-hydroxy-3-iodo-5-nitrophenylacetyl (NIP)–OVA embedded in alum, and the splenic GC response was evaluated by flow cytometry at day 7 after immunization (for the gating strategy used, see Fig. 6A for CD45.2+ B cell analysis). The frequency of NP-specific GC (GL7+Fas+NP+) B cells was statistically significantly less in mice that received DGKζ−/− B cells compared with that in mice that received WT B cells (Fig. 6, B and C). We then determined the frequency of plasma cells (PCs; CD138+) and IgG1+ B cells within the CD45.2+ B cell population. The transferred DGKζ−/− B cells showed reduced frequencies of both populations in comparison to transferred control B cells (Fig. 6, D and E). Transferred DGKζ−/− B cells exhibited preferential generation of IgM+ PCs as opposed to IgG1+ PCs when compared with transferred WT B cells (Fig. 6, F and G). The memory-like CD138IgG1+ B cell subset was reduced for DGKζ−/− B cells compared with that for WT B cells, although this was not statistically significant (Fig. 6, H and I). As expected, the recipient CD45.1+ B cell response was comparable between animals that received CD45.2+ WT B cells or CD45.2+ DGKζ−/− B cells (fig. S8). These results suggest that DGKζ−/− B cells have a competitive disadvantage for T cell help, which results in diminished GC responses.

Fig. 6 DGKζ−/− B cells exhibit a decreased GC response in vivo.

(A) Experimental design for comparing the in vivo responses of DGKζ−/− and WT CD45.2+ B cells in immunocompetent CD45.1+ mice. Right: Gating strategies to analyze CD45.2+ B cells (CD45.2+ CD19+ or CD45.2+ B220+) isolated from the spleen. (B) Representative strategy to measure by flow cytometry the percentages of total GC (GL7+ Fas+) and NP-specific GC (GL7+ Fas+ IgDneg NP+) CD45.2+ B220+ B cells generated 7 days after immunization with the T cell–dependent antigen Nip-OVA with Alum. Representative density plots for WT and DGKζ−/− B cells are shown. The percentages of the gated cells are indicated. (C) Percentages of total GC B cells (left) and of NP-specific GC B cells (right) in the CD45.2+ B220+ WT or DGKζ−/− B cell populations in the spleen. Each dot represents a single mouse. (D) Representative density plots of PC (CD19+ CD138+) and class-switched IgG1 B cell (CD19+ IgG1+) generation for adoptively transferred CD45.2+ WT or DGKζ−/− B cells. The percentages of the gated cells are indicated. (E) Percentages of PC and IgG1+ B cells in the CD45.2+CD19+ WT or DGKζ−/− B cell populations in the spleen. Each dot represents a single mouse. (F) Representative density plots of IgM and IgG1 surface expression on PCs (gated as CD19+CD138+; PC) for transferred CD45.2+ WT or DGKζ−/− B cells. The percentages of the gated cells are indicated. (G) Percentages of IgG1+ PCs (left) and IgM+ PCs (right) for the indicated conditions. (H) Representative profile of IgG1 surface expression on CD19+CD138 B cells [memory-like cells (MC)] for transferred CD45.2+ WT or DGKζ−/− B cells. The percentages of the gated cells are indicated. (I) Percentages of CD138IgG1+ MCs for the indicated conditions. Each dot in (G) to (I) represents a single mouse. Data in (C), (E), (G), and (I) are pooled from two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by two-tailed, unpaired Student’s t test.

DISCUSSION

This study reports a pivotal role for DGKζ in the regulation of actin polymerization and LFA-1–mediated adhesion at the B cell immune synapse and, consequently, in the generation of mechanical forces at the immune synapse. Impaired MTOC translocation to the immune synapse also suggests that DGKζ mediates cell polarity–related events in this context. Traction forces and cell polarization are necessary for the acquisition and processing of antigen at the B cell immune synapse (22, 28, 29). The ability of B cells to present antigenic peptides determines the chances of receiving costimulatory T cell help and the subsequent fate of the B cell response. Data from in vitro and in vivo assays support a role for DGKζ in shaping the ability of B cells to extract antigen from the APC surface and, thus, to receive T cell help and facilitate an appropriate GC response.

Immune synapse formation induces robust actin polymerization and the assembly of a peripheral F-actin ring, which provides a framework for signaling events, membrane trafficking, and adhesion support. In T cells, PI3K-mediated PIP3 production at the periphery of the immune synapse plays a major role in the maintenance of the F-actin ring (23). PIP3 recruits DOCK2, promoting Rac activation, and thus, actin polymerization. Our data support a similar role for PIP3 in regulating F-actin ring formation at the B cell immune synapse. DGKζ−/− B cells and DGK-inhibited WT or PI3Kδ KD B cells showed reduced amounts of F-actin at the immune synapse, whereas treatment with PA or overexpression of DGKζ increased the size and content of the F-actin ring. DGKζ and its product, PA, are thus involved in the regulation of F-actin ring formation at the B cell immune synapse. Previous reports of nonimmune cells showed the importance of DGKζ in connecting lipid signaling with actin reorganization through its kinase and scaffold activities. DGKζ associates with the PAK1-RhoGDI-Rac1 complex and promotes Rac activation, a process that requires both its scaffold and kinase functions (24). DGKζ also stimulates RhoA activation through a scaffolding mechanism, forming a complex with PKCα and RhoGDI (30). We propose that DGKζ affects Rac function and actin polymerization at the immune synapse by mediating DOCK2 recruitment and PAK1/2 activation. Our data from experiments with the DGKζ KD construct, the addition of exogenous PA addition, and the reduction in DOCK2-GFP abundance in R59-treated B cells provide evidence of a major contribution of DGKζ-kinase activity. In addition, the lack of increase in p-PAK1/2 abundance upon PA exposure suggests the involvement of DGKζ scaffold properties. PAK1/2 are also targets of active Rac and coordinate actin cytoskeleton remodeling (31, 32); however, whether they are involved in immune synapse assembly downstream of Rac requires further study. In neutrophil migration, DOCK2 dynamics at the plasma membrane is sequentially regulated by PIP3 and PA. Upon stimulation, PIP3 rapidly recruits DOCK2 to the plasma membrane, whereas the PA that is generated stabilizes DOCK2 at the membrane promoting its local accumulation and Rac activation (33). We propose that, in a similar fashion, the sequential actions of PIP3 and PA determine DOCK2 dynamics and, thus, F-actin ring maintenance at the B cell immune synapse. BCR signaling initially leads to PI3K activation and PIP3 production, whereas the PLC-γ2–mediated degradation of PIP2 produces the DAG that activates classical PKC. In turn, PKC phosphorylates DGKζ, driving its activation and relocation to the plasma membrane, where it produces PA (fig. S9).

DGKζ−/− and R59-treated B cells exhibited increased BCR-dependent Ca2+ influx. Ca2+ influx downstream of the BCR is mainly driven by the PLC-γ2–dependent and IP3-dependent activation of store-operated Ca2+ channels (SOCs). Nonetheless, B cells express several members of the family of Ca2+-permeable transient receptor potential channels (TRPCs) (34). TRPC activation seems to be Ca2+ store independent and DAG sensitive (3537). Previous work in the DT40 B cell line showed that DAG-dependent Ca2+ influx by TRPC3 enhances Ca2+ signaling downstream of the BCR and that TRPC3 promotes PLC-γ2 translocation to the plasma membrane and activation, maintaining IP3 and DAG production (26, 38). In addition, active TRPC3 retains PKCβ at the plasma membrane through a direct interaction, which sustains ERK1/2 activation (26). The lack of DGKζ activity enables DAG accumulation, which might amplify Ca2+ influx and signaling downstream of the BCR by inducing TRPC3 activation.

Cell polarization at the B cell immune synapse orchestrates the membrane-trafficking events that are required for antigen processing and presentation to T cells. MHC class II–containing lysosomes translocate together with the MTOC to the immune synapse, where their local secretions promote antigen extraction (29). The polarization of the MTOC and lysosomes depends on the GTPase Cdc42 and its effector, the atypical PKCζ, which is part of the Par polarity complex together with Par3 and Par6 (29). Par3 is enriched at the B cell immune synapse and is involved in the transport of the MTOC and lysosomes to the immune synapse interface (39). DGK-derived PA promotes PKCζ location and activity in nonimmune cells (11). In DGKζ−/− OTI CD8+ cytotoxic T cells, impaired MTOC recruitment to the immune synapse correlates with reduced amounts of active (phosphorylated) PKCζ (18). DAG accumulates at the CD4+ T cell immune synapse and establishes an intracellular gradient that drives MTOC polarization, which involves three novel PKC isoforms (ε, η, and θ) and the motor protein dynein (40, 41). DGKα has a major role in shaping the DAG gradient at the immune synapse and establishing T cell polarity (19). A report showed that Arp2/3-dependent F-actin nucleation at the MTOC connects this organelle with the nucleus in resting B cells and that BCR stimulation reduces F-actin content at the MTOC, enabling its detachment from the nucleus and polarization to the immune synapse (42). The defects in MTOC translocation in DGKζ−/− and DGK-inhibited B cells suggest that DGKζ promotes cell polarization events. More studies are necessary to relate DGKζ with the PKCζ/Par3 axis, F-actin nucleation at the MTOC, or the DAG gradient in B cells.

Actin cytoskeleton remodeling drives force generation in cells. Actin polymerization per se generates pushing forces, whereas F-actin, in combination with the contractile activity of myosin II, produces pulling forces. To generate and exert forces against the extracellular matrix, or another cell, cells connect protrusive and contractile F-actin dynamics to adhesion structures (43). Lymphocytes link actin dynamics to LFA-1–mediated adhesion at the immune synapse, which involves vinculin and talin. Our data point to DGKζ being a regulator of force generation at the B cell immune synapse, which may be achieved by influencing LFA-1–mediated adhesion, actin polymerization, and myosin II activity downstream of the BCR. In T cells, F-actin flow and mechanical forces are important for LFA-1 activity at the immune synapse (44). DGKζ might affect LFA-1–mediated adhesion through the activities of Rac, myosin, or both. In addition, DGK-produced PA promotes PIP5KI activity and the subsequent generation of PIP2, which recruits to the plasma membrane proteins that are involved in actin polymerization and adhesion site dynamics, such as vinculin, talin, and WASP (4, 17, 45). Although further investigations are required to dissect the underlying mechanisms, DGKζ appears to use the DOCK2-PAK1-Rac and PIP5K-PIP2 axes to support mechanical force generation at the immune synapse. Our studies using the MFP technique showed sequential pushing, pulling, and cup-like stages during B cell immune synapse formation, similar to T cells (21, 46). Experiments involving inhibitor treatment of T cells indicate a main role of the actin cytoskeleton in generating pushing forces, whereas myosin activity is needed for the pulling/contractile stage, with PI3K-DOCK2 signaling also participating in the pulling phase (47). The reduced pushing and pulling forces values that we observed in DGKζ−/− B cells correlated with impaired actin polymerization at the immune synapse and reduced myosin activation.

Two mutually nonexclusive mechanisms support antigen acquisition from the APC surface by B cells: local secretion of lysosomes at the immune synapse interface, which liberates proteases to facilitate antigen extraction (29), and myosin II–mediated pulling forces that promote the internalization of antigen-BCR complexes (22). Myosin II–derived forces enable discrimination of BCR affinity for the antigen, which is crucial for the T cell–dependent selection of high-affinity GC B cells (28). Our results suggest that the DGKζ-mediated regulation of mechanical forces and MTOC translocation at the B cell immune synapse facilitate antigen acquisition and presentation to T cells. In our in vitro system, OVA was present with su-Ag at the planar lipid bilayer without the two components being physically attached to each other. This implies that lysosomal secretion and/or strong forces able to detach a piece of artificial membrane are required for OVA acquisition and degradation.

In a competitive in vivo environment, we demonstrated that the lack of DGKζ resulted in reduced GC B cell activity and diminished numbers of antigen-specific GC B cells, PCs, and IgG1 class-switched B cells. A previous study addressed, the role of DGKζ in the B cell response with experiments involving the immunization of DGKζ−/− mice with NP-Ficoll, a T-independent, type 2 antigen (3). This study reported increased numbers of antigen-specific IgM- and IgG3-secreting PCs and increased serum concentrations of IgM and IgG3 in DGKζ−/− mice compared with those in WT mice, suggesting that DGKζ limits the early PC response. In addition, the previous report used MD4 BCR (HEL-specific) transgenic B cells, WT (CD45.1) and DGKζ−/− (CD45.2) mixed at 1:1 ratio, for adoptive transfer into immunocompetent mice. Mice were then immunized with HEL mutants of low and medium affinity conjugated to SRBC, a T cell–dependent antigen. The authors found increased numbers of GC B cells and IgM+ PCs generated by DGKζ−/− B cells compared with those generated by WT B cells at the beginning of the Ab response (day 5) (3). Their model precluded analysis at later time points because MD4 B cells are unable to undergo class switching. The authors suggested that DGKζ limits early PC generation in T cell–dependent responses by promoting antigen affinity discrimination by DAG signaling. Note that we evaluated later stages of the GC response. In our system, the reduced production of IgG1 B cells, supported by the in vitro data on antigen acquisition and presentation, provides evidence that DGKζ determines the ability of B cells to acquire antigen and compete for T cell help. Nevertheless, the increased number of IgM+ PCs might reflect the enhanced early generation of PCs by DGKζ-deficient B cells, which was previously described (3). In addition, our data showed that in vitro DAG-related activation (as determined by measuring Ca2+ influx, ERK1/2 activation, and CD69/CD25/CD86 expression) and proliferation in DGKζ−/− B cells are enhanced when compared with WT B cells after BCR stimulation, as also reported in that study (3). The later timing of analysis and impaired ability to receive T cell help might account for the reduced GC frequencies in our model compared with those in the previous study. B cell clone frequency and antigen affinity determine B cell recruitment to and interclonal competition at the GC (48). The affinity values reported for HEL3X (low-affinity mutant) and NP are similar (KD values are in the micromolar range), but distinct B cell precursor frequency might also explain the differences in results. B cell competition for antigen is likely lower in the previously published model than in the model use here, as all (and only) the transferred B cells recognize the antigen (HEL) used for immunization, which was not the case in our experimental approach. The method used for antigen administration (NIP-OVA in alum versus HEL on SRBCs) may also account for some of the differences between the two studies.

DGKs are currently considered as therapeutic targets to manipulate T cell function in autoimmune diseases and to subvert tumor immunosuppression. Increased amounts of DGKα and DGKζ correlate with reduced effector function in tumor-infiltrating lymphocytes (49, 50). Pharmacological intervention to manipulate DGKs focuses on the capacity of drugs to limit DAG-mediated signals and subsequent gene transcription. Our study underlines the relevance of DGKζ functions pertaining to PA generation for B cell function. The described roles for DGKζ in organizing the B cell:APC interface might also apply to other immune cell interactions and should be considered when targeting DGKs therapeutically.

MATERIALS AND METHODS

Mice and B cell isolation

Primary B lymphocytes were isolated from the spleens of adult (10- to 20-week-old) WT, DGKζ−/− (51), DGKα−/− (52), PI3Kδ kinase-deficient knock-in [provided by D. F. Barber, Centro Nacional de Biotecnología (CNB)–Consejo Superior de Investigaciones Científica (CSIC), Spain; (53)], and DOCK2-GFP knock-in [provided by J. Stein, University of Bern, Switzerland, and Y. Fukui, Kyushu University, Japan; (54)] mice, all of in which are on a C57BL/6 genetic background. Splenic B cells were purified by negative selection with mouse pan-T Dynabeads (DynaI Biotech, Invitrogen) after a Lympholyte step (Cedarlane Laboratories); we enriched to >90% B cells. Primary OTII CD4+ T cells were isolated from the spleens of adult OTII transgenic (OVA 323-339–specific TCR) mice (55) by negative selection using a CD4+ T cell isolation kit [magnetic-activated cell sorting (MACS), Miltenyi Biotec; purity, >90% CD4+ T cells]. Animal procedures were approved by the CNB-CSIC Bioethics Committee and conform to institutional, national, and European Union (EU) regulations. The A20 mouse B cell line was transiently transfected with plasmids encoding for GFP or Cherry fluorescent protein alone, or GFP- or cherry-tagged DGKζ-WT or -DGKζ-kinase–deficient (KD) constructs (9) by electroporation (260 mV, 950 μF) and were used 20 hours later. Cells were cultured in complete RPMI (10 mM Hepes, 2 mM l-glutamine, and 50 μM β-mercaptoethanol) supplemented with 10% fetal calf serum (FCS).

Real-time microscopy on planar lipid bilayers

Artificial planar lipid bilayers were assembled in FCS2 chambers (Bioptechs) as described previously (56). Briefly, unlabeled murine GPI-linked ICAM-1–containing 1,2-dioleoyl-PC (DOPC) liposomes and DOPC liposomes containing biotinylated lipids were mixed with DOPC liposomes at distinct ratios to obtain specified molecular densities (ICAM-1 at 200 molecules/μm2; biotin lipids, as indicated in the figure legends). Artificial planar lipid bilayers were assembled on sulphochromic solution–treated coverslips in FCS2 closed flow chambers (Bioptechs) and blocked with phosphate-buffered saline (PBS)/2% FCS for 1 hour at room temperature. su-Ag was tethered to membranes by incubation with Alexa Fluor 647– or Alexa Fluor 555–conjugated streptavidin (Molecular Probes), which was followed by monobiotinylated rat anti-κ light chain monoclonal Ab (mAb; clone187.1). Monobiotinylation was achieved by labeling the Ab (0.5 mg/ml; 1 ml) with NHS-LC-LC-biotin (1 μg/ml; 30 min, room temperature, in PBS; Pierce), followed by dialysis and analysis by flow cytometry. We estimated the number of molecules per square micrometer of GPI–ICAM-1 or anti-κ Ab at the lipid bilayers by immunofluorometric assay with anti–ICAM-1 or anti–rat IgG antibodies, respectively. We obtained the standard values from microbeads with distinct calibrated IgG-binding capacities (Bangs Laboratories). Before imaging, membranes were coated with murine recombinant CXCL13 (100 nM, Peprotech) for 20 min at room temperature. Lipids stock in chloroform were obtained from Avanti Polar Lipids Inc. WT and genetically modified B cells (4 × 106) were coinjected into the warmed chamber (37°C) for imaging. To distinguish them, one cell type was violet tracer labeled (0.1 μM, 10 min, 37°C; Molecular Probes). Where indicated in the figure legends, B cells were pretreated with the pan-DGK inhibitor R59949 [10 μM, 30 min, 37°C; half maximal inhibitory concentration (IC50), 3.3 μM; Sigma] or with the PI3K inhibitor LY294002 (10 μM, 30 min, 37°C; Sigma) and washed before use. Confocal FL (1-μm optical section), DIC, and IRM images were acquired every 30 s for 10 to 20 min. Consecutive videos were acquired when needed. Similarly, transfected A20 B cells (2 × 106) were injected and imaged. For Ca2+ flux measurements, B cells were labeled with Fluo-4FF (1 μM, Molecular Probes) for 30 min at room temperature, injected into the warmed FCS2 chamber, and imaged every 10 s for 15 min at low quality to speed up acquisition. Assays were performed in chamber buffer [PBS, 0.5% FCS, d-glucose (0.5 g/liter), 2 mM MgCl2, and 0.5 mM CaCl2]. For exogenous PA assays, we used freshly prepared 10 mM PA stock [in 10 mM tris-HCl (pH 8.0) and 150 mM NaCl]. The cells were left in contact with the lipid bilayers for 10 min to form the immune synapse and then were imaged. At the 15-min time point, 0.1 mM PA (1 ml) was injected into the chamber buffer. After 30 min, the B cells were imaged. Images were acquired on an Axiovert LSM 510 META inverted microscope with a 40× oil immersion objective (Zeiss).

Immunofluorescence

Primary B cells or transfected A20 B cells were in contact with ICAM-1/CXCL13 lipid bilayers containing tethered su-Ag for 10 min, fixed with 4% paraformaldehyde for 10 min, at 37°C, permeabilized with PBS/0.1% Triton X-100 for 5 min at room temperature, blocked with PBS/2% FCS/2% bovine serum albumin (BSA) overnight at 4°C, and stained with Alexa Fluor 647–conjugated phalloidin (Molecular Probes) and the following antibodies: rabbit anti–phospho-ERK1/2 (Cell Signaling) with Alexa Fluor 488–conjugated goat anti-rabbit IgG (Southern Biotechnology), mouse anti-vinculin (clone hVIN-1; Sigma) with fluorescein isothiocyanate (FITC)–conjugated goat anti-mouse IgG1 (BD Biosciences). For PA assays, B cells were fixed at 30 min after PA exposure and stained for phalloidin as described earlier. For MTOC analysis, B cells were mixed with unloaded or su-Ag–loaded pseudo-APCs (20 molecules/μm2) at a 1:1 ratio, cultured for 30 min at 37°C on poly-l-lysine–coated coverslips, fixed, permeabilized, and blocked as described earlier, stained with rabbit anti–γ-tubulin (T5192, Sigma) and Alexa Fluor 488–conjugated goat anti-rabbit IgG, and mounted using Fluoromount (Southern Biotech). FCS2 chambers and coverslips were imaged by confocal fluorescence microscopy as previously described.

Cell conjugates and activation assays

For cell activation assays, freshly isolated B cells (2 × 105) were cocultured with unloaded or su-Ag–loaded (20 molecules/μm2) planar lipid bilayers, assembled in glass-bottom p96-size wells for 20 hours, and then collected and analyzed by flow cytometry. To prepare pseudo-APCs, silica beads (5 × 106; 5-μm diameter; Bangs Laboratories) were washed in distilled water (2600g, 1 min, room temperature), incubated with 20 μl of DOPC liposomes containing GPI-linked ICAM-1 (200 molecules/μm2) and biotin lipids (20 or 1000 molecules/μm2) for 10 min at room temperature, washed twice with chamber buffer, blocked with PBS/2% FCS for 30 min, washed twice, incubated with the monobiotinylated su-Ag and 10 nM CXCL13 for 20 min, washed twice, and counted. All incubations were done in a rotary shaker at room temperature. For cell proliferation assays, B cells were labeled with CFSE tracer (0.1 μM, 10 min, 37°C), washed with complete RPMI/10% FCS, and cocultured with unloaded or su-Ag–loaded pseudo-APCs at specified ratios and with recombinant murine IL-4 (10 ng/ml; Peprotech) in flat-bottom p96 wells for 96 hours. The cells were collected, stained with APC-conjugated CD19, and analyzed for CFSE tracer dilution with a FACSCalibur cytometer (BD Biosciences). To assess antigen acquisition and T cell presentation, freshly isolated B cells (5 × 105) were cultured for 2 hours in contact with ICAM-1/CXCL13 planar lipid bilayers and assembled in glass-bottom p96-size wells. These planar bilayers contained distinct densities of biotinylated lipids (20, 100, 500, or 2500 molecules/μm2) and were loaded with a mixture of monobiotinylated su-Ag (5 μg/ml) and monobiotinylated Alexa Fluor 488–conjugated OVA (10 μg/ml; Molecular Probes) by previous coating with Alexa Fluor 647–conjugated streptavidin (Molecular Probes). OVA monobiotinylation was performed as described earlier for su-Ag. B cells were then collected; one-half of them (2.5 × 105) was treated with trypsin for 5 min at 37°C, washed with complete RPMI/10% FCS, and analyzed by flow cytometry for streptavidin and OVA fluorescence signals to measure antigen extraction. The other half of the B cells (2.5 × 105) was washed with complete RPMI/10% FCS and cocultured with CFSE tracer–labeled CD4+ OTII T cells (2.5 × 105) at 1:1 ratio in round-bottom p96 wells. After 72 hours, the culture medium was collected for IL-2 detection by enzyme-linked immunosorbent assay (ELISA) kit (IL-2 ELISA Max 413005, BioLegend), and the cells were collected, stained with Pacific Blue–conjugated rat anti-mouse B220, APC-conjugated rat anti-mouse CD4, and PE (phycoerythrin)–Cy7–conjugated rat anti-mouse CD25, and analyzed for CFSE tracer dilution and CD25 expression in the CD4+ T cell population with a fluorescence-activated cell sorting (FACS) LSR-II cytometer (BD Biosciences). When required, B cells were pretreated with the DGK inhibitor R59949 for 30 min before adding them to the planar lipid bilayers. The inhibitor was kept during the antigen extraction time (2 hours). When indicated in the figure legends, B cells were exposed to 0.1 mM PA for the 2 hours of antigen extraction.

Imaging data analysis

The frequency of immune synapse formation per imaged field was estimated as [number of B cells showing a central su-Ag cluster and IRM contact/total number of B cells (estimated by DIC)] × 100, using FiJi [National Institutes of Health (NIH)] software. Confocal images (1-μm optical sections) were acquired at the contact plane or immune synapse plane. We used the IRM confocal image to focus on the B cell–artificial membrane contact plane and to define the immune synapse plane. Imaris 7.0 software (Bitplane) was used for the qualitative and quantitative analyses of fluorescence signals, as well as for cell contact area (IRM area) and su-Ag cluster area measurements. To set up the background of the fluorescence intensity signal, we used the fluorescence signal of the lipid bilayer in each case. To apply statistical analysis to the Ca2+ influx curves, we calculated the area under de curve (AUC) for each cell in each condition (WT, DGKζ−/−, and R59 treated) and then compared the obtained AUC values with those of the control condition (untreated WT). To obtain the AUC value per cell, we sectioned the area in three trapezoids and calculated the area of each one; the AUC is the sum of the three trapezoid areas.

Western blotting analysis

Freshly isolated primary B cells (5 × 106) were cultured on a p48 plate in depletion medium (0.5 ml of complete RPMI) for 1 hour in the presence of 10 μM R59 when needed and then stimulated in an Ab-coated (goat anti-mouse IgM, μ-specific; Jackson ImmunoResearch) p48 plate for the times indicated in the figure legends. Wells were precoated with the appropriate Ab (5 μg/ml) in PBS for 1 hour at 37°C, washed, and used for analysis. To detect DGK isoforms or phosphorylated PAK1/2, isolated B cells (5 × 106) were cultured in complete RPMI, 10% FCS, without or with 10 μM R59 or 0.1 mM PA for 1 hour, and then were collected. Cells were lysed in lysis buffer [10 mM tris-HCl (pH 7.4), 150 mM NaCl, and 1 mM EDTA]/1% Triton X-100 with protease and phosphatase inhibitors (Roche) for 30 min at 4°C. Lysates were centrifuged at 20,000g for 30 min at 4°C, and the supernatants were collected and stored at −80°C. Total protein was quantified with the Micro BCA Protein Assay Kit (Thermo Scientific). Proteins were resolved by SDS–polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene difluoride membranes (Bio-Rad). Blots were blocked with TBS-T [10 mM tris-HCl (pH 8.0), 150 mM NaCl, and 0.1% Tween 20]/5% BSA for 1 hour at room temperature and incubated with rabbit anti-DGKζ (ab105195; Abcam), anti-DGKα (11547-1-AP; Proteintech), anti–phospho (S473)–Akt (Cell Signaling), anti-phospho (Thr18/Ser19)–MLC (Cell Signaling), anti–phospho (S144)–PAK1/(S141)-PAK2 (Cell Signaling), or loading control mouse anti–α-tubulin (clone DM1A; Sigma) or rabbit anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (FL305, Santa Cruz Biotech) overnight at 4°C. The blots were then incubated for 1 hour at room temperature with horseradish peroxidase–conjugated secondary antibodies (DAKO). Signals were detected with the enhanced chemiluminescence (ECL) detection system (GE Healthcare). Signal intensity values in arbitrary units (AU) for each protein were quantified with FiJi (NIH) software and were normalized to that of tubulin.

Immunization

Freshly isolated CD45.2+ WT or DGKζ−/− B cells (5 × 106 to 8 × 106) were adoptively transferred to CD45.1+ immunocompetent mice by intravenous injection. Twenty-four hours later, the mice were immunized intraperitoneally with NIP-OVA (200 μg; N-5041-10, Biosearch Technology) complexed with Alum (100 μl; 77161, Thermo Scientific) diluted 1:1 in PBS (0.2 ml of final volume). Seven days after immunization, the spleens were harvested and processed for CD45.2+ B cell population analysis by flow cytometry.

Flow cytometry

B cells were stained with fluorochrome-conjugated (FITC, PE, or APC) rat anti-mouse IgD, CD19, CD25, CD69, or CD86 (BioLegend), and DyLight-649–conjugated Fab fragment goat anti-mouse IgM, μ-specific (Jackson ImmunoResearch) for 30 min at 4°C. Samples were acquired with a FACSCalibur cytometer (BD Biosciences). Splenocytes obtained from the immunizations were stained with the following fluorochrome-conjugated Ab mixes: rat anti-mouse CD45.1 (APC-Cy7), CD45.2 (APC), B220 (V450), CD95 (PE-Cy7), and GL7 (FITC), and with PE-conjugated NP(36) (N-5070-1, Biosearch Technology) for GC B cell analysis; rat anti-mouse CD45.1 (APC-Cy7), CD45.2 (Per-CP5.5), CD19 (PE-Cy7), CD138 (APC), IgG1 (PE), IgD (V450), and IgM (biotin) together with FITC-conjugated streptavidin for PC and IgG1+ B cell analysis; all antibodies were obtained from BD Biosciences. Samples were acquired in a FACSCanto II cytometer. Data were analyzed with FlowJo software (BD Biosciences).

Traction force microscopy

PAA gels were produced in 35-mm FD35 fluorodishes (World Precision Instruments Inc.). Dishes were first treated by ultraviolet (UV) irradiation for 2 min and then with 3-aminopropyltrimethoxysilane for 5 min and, lastly, washed thoroughly in distilled water before PAA gel preparation. Hydrophobic coverslips were prepared by incubation in Sigmacote (Sigma-Aldrich) for 3 min, which is followed by thorough washing and drying. A 500-Pa gel was prepared by diluting 40% PAA and 2% bis-acrylamide solutions to obtain stock solutions of 12% PAA/0.1% bis-acrylamide. We sonicated 167 μl of this solution with 1% of 0.2-μm-diameter carboxilated fluorescent (660/680) beads (Thermo Fisher Scientific) and then added 0.2 μl of TEMED and 1% ammonium persulfate and mixed vigorously to initiate polymerization. A volume of 9 μl of the PAA mixture was immediately pipetted onto the surface of the Fluorodish, and a Sigmacote-activated coverslip was carefully placed on top. Fluorodishes were immediately inverted to bring the beads to the surface of the gel. Polymerization was completed in 45 min. The top coverslip was then slowly peeled off and the gel was immediately immersed in PBS. Sulfo-SANPAH (Sigma-Aldrich), a surface functionalizing reagent with an amine-binding group and a photoactivable azide group, was used to crosslink molecules to the surface of the gel. Sulpho-SANPAH [150 μl of stock (0.5 mg/ml) in 10 mM Hepes] was attached to the gel surface through UV light activation for 2 min. The gels were then washed with PBS, and the process was repeated. The gel was washed thoroughly with PBS and coated with 100 μl of su-Ag (10 μg/ml; rat anti-κ light chain mAb; clone187.1; BD Biosciences) alone or mixed with recombinant mouse ICAM-1–Fc (10 μg/ml; BioLegend) by overnight incubation at 4°C. Freshly isolated WT and DGKζ−/− B cells were mixed at 1:1 ratio (1 × 106), with one set of cells labeled with CFSE tracer to distinguish between them, and then added to the gels and imaged. Assays were performed in complete RPMI/10% FCS medium. All TFM movies were acquired at 37°C/4.5% CO2 on an inverted spinning disk confocal microscope (Eclipse Ti Nikon/Roper spinning head) with a 60×/1.4 numerical aperture (NA) oil immersion objective (pixel size, 108 nm) with MetaMorph software (Molecular Device, France) and a HQ2 Coolsnap Photometric camera. Time lapse was typically at a frame rate of one image per 5 s and lasted for 15 min. The traction force algorithm was based on that used by Butler et al. (57) and modified by Mandal et al. (58). Force reconstruction was conducted with the assumption that the substrate is a linear elastic half space, using Fourier transform traction cytometry with Tikhonov regularization (regularization parameter was set to 5 × 10−19). The bead position in the reference image and the deformed one was measured using the multi-target tracking (MTT) algorithm (59). The problem of calculating the stress field from the displacement was solved in Fourier space and then inverted back to real space. The final stress field was obtained on a grid with 0.432-μm spacing (four pixels). All calculations and image processing were performed in MATLAB. The mask of the cell (defined by the user based on fluorescence or bright field images) increased by 10% (dilation of the binary image using MATLAB morphological tools) and was used as domain of integration for the energy. Given the B cell size, the density of beads, and the magnitude of displacement, some parameters needed optimization for the analysis, in particular for the detection algorithm (MTT): search window size (5 pixels), particle radius (2.5 pixels), and maximum distance for nearest neighbor (4 pixels). The same parameters were applied for noise detection by measuring force in a nonstressed area not too far from the cell. Further calculations based on the output of the algorithm were performed to extract the total strain energy (scalar product force by displacement integrated over the entire cell area). Noise greater than a certain threshold (chosen at 3 × 10−17 J) indicated poorly acquired data (for example, due to defocus); the corresponding frames were, thus, eliminated from the analysis.

Micropipette force probe

MFP (21) uses a flexible glass micropipette as a cantilever to measure pushing and pulling forces generated by a single cell. We added supplementary, single-cell rheometer capabilities to measure the mechanical properties of the cell during its activation. Micropipettes were prepared as described previously (21, 47, 60, 61) by pulling borosilicate glass capillaries (Harvard Apparatus) with a P-97 micropipette puller (Sutter Instruments), cutting them with an MF-200 microforge (World Precision Instruments) and bending them at a 45° angle with an MF-900 microforge (Narishige). Micropipettes were held by micropipette holders (IM-H1, Narishige) placed at a 45° angle relative to a horizontal plane, so that their tips were in the focal plane of an inverted microscope under bright-field illumination (TiE, Nikon Instruments) equipped with a 100× oil immersion, 1.3 NA objective (Nikon Instruments), and placed on an air suspension table (Newport). The flexible micropipette was linked to a nonmotorized micropositioner (Thorlabs, Newton, NJ, USA) placed on top of a single-axis stage controlled with a piezo actuator (TPZ001; Thorlabs). The bending stiffness k of the flexible micropipette (about 0.2 nN/μm) was measured against a standard microindenter previously calibrated with a commercial force probe (model 406A; Aurora Scientific). The flexible micropipette aspirates a GPI–ICAM-1–containing lipid-coated bead with tethered su-Ag (100 molecules/μm2), while a second (rigid) micropipette holds a B cell at its tip. The B cell is brought in an adequate position using a motorized micromanipulator (MP-285; Sutter Instruments). Experiments were performed in glass-bottom petri dishes (Fluorodish, WPI) filled with about 5 ml of complete RPMI/10% FCS at room temperature. Images were acquired using a Flash 4.0 complementary metal-oxide semiconductor (CMOS) camera (Hamamatsu Photonics). To perform rheological experiments, the setup automatically detects, at a rate of 400 to 500 Hz, the position of the bead at the tip of the force probe (Xbead) and imposes the position of the base of the flexible micropipette by regulating the position of the piezo stage. The deflection of the force probe is the difference between the position of the bead and the position of the piezo stage. The force applied to the cell is the product of this deflection by the bending stiffness k. A retroaction implemented in MATLAB (MathWorks) regulating both the camera by the Micromanager software (Edelstein 2014) and the piezo stage moves the latter in reaction to the measurement of the bead position to maintain a desired deflection of the cantilever. In this way, a controlled force is applied to the cell at any given time. The experiment was decomposed in two phases. During a first phase, the base of the force probe was translated at constant velocity v = 1 μm/s toward the cell, leading to an increasing compressive force until a maximum compressive force of 240 pN was reached. Young’s modulus of the cell was obtained by postprocessing the recordings made during this phase, as previously described (60). Then, the algorithm automatically switched to a second phase, during which an oscillatory force F was applied to the cell with an average force F0 = 60 pN, an amplitude ∆F = 20 pN, and at a frequency f = 1 Hz. Knowing the position of the bead, we could deduce the changes in cell length (L) over time. L was approximatively sinusoidal with an average value L0, an amplitude ∆L, and a phase lag φ relative to the imposed force. This phase lag results from the fact that the cell is not purely elastic but also viscous. In this study, we focused on the variations of the elastic properties of the cell, which we quantified with the stiffness K′ of the cell that is expressed as K′ = (∆F/∆L) cos φ (see the Supplementary Materials for K′ parameter quantification). The average length L0 evolves over time, and its measurement enables monitoring of cell growth or shrinkage.

Statistical analysis

Graphs and statistical analyses were performed using GraphPad Prism 6.0f software. Two-tailed unpaired Student’s t tests were applied. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/13/627/eaaw8214/DC1

K′ parameter quantification

Fig. S1. DGKζ mediates LFA-1–dependent adhesion at the B cell immune synapse.

Fig. S2. DGKα−/− B cells have no defects in immune synapse formation.

Fig. S3. DGKζ overexpression enhances LFA-1–mediated adhesion as well as vinculin and F-actin content at the A20 B cell immune synapse.

Fig. S4. Exogenous PA modifies the su-Ag central cluster at the B cell immune synapse.

Fig. S5. DGKζ-derived PA in combination with PI3K-generated PIP3 mediates F-actin polymerization at the B cell immune synapse.

Fig. S6. TFM and MFP experimental setup and analysis of relative p-ERK abundance at the immune synapse of WT and DGKα−/− B cells.

Fig. S7. Experimental setup for the evaluation of antigen extraction by B cells and presentation to OTII CD4+ T cells.

Fig. S8. GC response of recipient CD45.1+ B cells.

Fig. S9. Model for DGKζ roles and lipid signaling interplay at the B cell immune synapse.

Movie S1. Stress maps generated by WT B cells contacting su-Ag–coated substrates.

Movie S2. Stress maps generated by DGKζ−/− B cells contacting su-Ag–coated substrates.

Movie S3. Stress maps generated by WT B cells contacting su-Ag– and ICAM-1-Fc–coated substrates.

Movie S4. Stress maps generated by DGKζ−/− B cells contacting su-Ag– and ICAM-1-Fc–coated substrates.

Movie S5. MFP assay of a WT B cell in the absence of su-Ag.

Movie S6. MFP assay of a WT B cell in the presence of su-Ag.

Movie S7. MFP assay of a DGKζ−/− B cell in the presence of su-Ag.

Movie S8. MFP assay of an R59-treated WT B cell in the presence of su-Ag.

Reference (62)

REFERENCES AND NOTES

Acknowledgments: We thank M. Mellado and C. R. Jimenez-Saiz (CNB-CSIC) for the critical review of the manuscript, and M. Balland (Liphy, Grenoble) for providing tools and suggestions for TFM experiments. We acknowledge the PICT-IBiSA imaging platform at Institut Curie, a member of the French National Research Infrastructure France-BioImaging (ANR-10-INBS-04). Funding: S.V.M.-C. is supported by an FPI contract from the Spanish Ministry of Economy (MINECO; BES-2014-068006). This work was supported by grants from the MINECO (BFU2013-48828-P), from the Worldwide Cancer Research (WCR; grant reference number 15-1322), and from MCIU/AEI/FEDER EU (RTI2018-101345-B-I00) to Y.R.C. J.H. has benefited from the financial support of the LabeX LaSIPS (ANR-10-LABX-0040-LaSIPS), managed by the French National Research Agency under the “Investissements d’avenir” program (no. ANR-11-IDEX-0003-02), and from PEPS CNRS funding. Author contributions: S.V.M.-C. designed parts of the study, performed experiments, analyzed the data, and assisted in manuscript preparation; S.R.G. and S.R.-G. performed some experiments, analyzed the data, and assisted in manuscript preparation; A.M.-R. and B.A. performed in vitro antigen presentation and in vivo immunization experiments, assisted in the data analysis, and provided input into the project; R.L. and I.M. provided DGKζ−/− and DGKα−/− mice, DGKζ constructs, and input into the project; J.P., A.-M.L.D., and P.P. performed and analyzed the TFM experiments and provided input into the project; J.H. performed and analyzed the MFP experiments and provided input into the project; and Y.R.C. designed and supervised all aspects of the work and wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article