Research ArticleImmunology

Profiling the origin, dynamics, and function of traction force in B cell activation

See allHide authors and affiliations

Sci. Signal.  07 Aug 2018:
Vol. 11, Issue 542, eaai9192
DOI: 10.1126/scisignal.aai9192

B cells use the force

B cells recognize antigens through membrane-bound antibodies that are part of the B cell receptor (BCR). Antigen recognition stimulates BCR-dependent intracellular signaling that is required for B cell activation. Noting that B cells first spread over antigen-presenting surfaces before contracting, Wang et al. used traction force microscopy to measure the displacing forces exerted by B cells on fluorescent beads coated onto antigen-containing gel surfaces of different stiffness values. The authors found that memory B cells generated greater traction forces than did naïve B cells and that the strength of these forces was correlated with increased BCR microcluster mean fluorescent intensity. In addition, B cells from patients with rheumatoid arthritis exerted greater traction forces than did B cells from healthy donors, which may play a role in the enhanced activation of autoreactive B cells observed in these patients.

Abstract

B lymphocytes use B cell receptors (BCRs) to recognize membrane-bound antigens to further initiate cell spreading and contraction responses during B cell activation. We combined traction force microscopy and live-cell imaging to profile the origin, dynamics, and function of traction force generation in these responses. We showed that B cell activation required the generation of 10 to 20 nN of traction force when encountering antigens presented by substrates with stiffness values from 0.5 to 1 kPa, which mimic the rigidity of antigen-presenting cells in vivo. Perturbation experiments revealed that F-actin remodeling and myosin- and dynein-mediated contractility contributed to traction force generation and B cell activation. Moreover, membrane-proximal BCR signaling molecules (including Lyn, Syk, Btk, PLC-γ2, BLNK, and Vav3) and adaptor molecules (Grb2, Cbl, and Dok-3) linking BCR microclusters and motor proteins were also required for the sustained generation of these traction forces. We found a positive correlation between the strength of the traction force and the mean fluorescence intensity of the BCR microclusters. Furthermore, we demonstrated that isotype-switched memory B cells expressing immunoglobulin G (IgG)–BCRs generated greater traction forces than did mature naïve B cells expressing IgM-BCRs during B cell activation. Last, we observed that primary B cells from patients with rheumatoid arthritis generated greater traction forces than did B cells from healthy donors in response to antigen stimulation. Together, these data delineate the origin, dynamics, and function of traction force during B cell activation.

INTRODUCTION

B lymphocytes mediate antibody responses arising from the recognition of antigens by the surface expressed B cell receptor (BCR) (1). The BCR contains a membrane-bound immunoglobulin (mIg) and a heterodimer of Igα and Igβ subunits (2, 3). The mIg is mainly responsible for the recognition of antigens, whereas the Igα and Igβ heterodimer stimulates transmembrane signaling through immunoreceptor tyrosine activation motifs in the cytoplasmic domains (4, 5). Antigen binding–induced activation of BCR signaling is efficiently regulated by the presentation of variable forms of antigens that B cells encounter in vivo (5, 6). These antigen characteristics include, but are not limited to, antigen density (7, 8), antigen affinity (7, 8), antigen valency (914), the Brownian mobility feature of the antigen (1517), the mechanical forces delivered to the BCRs by the antigens (18, 19), and the stiffness feature of the substrates presenting the antigen (20, 21). All of these studies suggest that the BCR is a versatile receptor, which can efficiently sense both the chemical and physical features of an antigen ligand and convert them into a cytosolic signal to determine the fate of the cell.

Early biochemical studies extensively investigated how the chemical cues from the antigen determine the strength of the signaling cascade mediated by the BCR (4, 5). However, reports have revealed that physical cues are also important layers of external information that are delivered to the BCRs by the antigens (22). For example, it was reported that, when B cells encounter membrane-bound antigens, they first exhibit a substantial spreading response over the antigen-presenting surface, which is followed by a marked contraction response (8). Thus, dynamic traction forces are generated between the B cells and the antigen-presenting surface during the B cell spreading and contraction responses. However, the traction force in B cell activation is poorly characterized.

Here, we used a traction force microscopy system to profile the origin, dynamics, and function of traction force generation within the B cell immunological synapse during B cell activation. We found that B cells generated a total traction force of 10 to 20 nN when they encountered antigens presented by substrates with stiffness values of 0.5 to 1 kPa, which mimic the rigidity of antigen-presenting cells in vivo (2325). The traction force generation in B cells after BCR-mediated antigen recognition relied on the remodeling of polymerized microfilament actin (F-actin) and motor proteins, including dynein and myosin. We also observed a positive correlation between the strength of the traction forces and the mean fluorescence intensity (MFI) of the BCR microclusters. Moreover, the requirement of membrane-proximal BCR signaling molecules and adaptor molecules linking BCR microclusters and motor proteins to sustain traction force generation was also revealed. Furthermore, we demonstrated that, during the initiation of immune activation, isotype-switched memory B cells, which express IgG-BCRs, generated greater traction force than that of naïve B cells, which express IgM-BCRs. These differences are likely due to the increased amounts of motor proteins in memory B cells and the formation of prominent IgG-BCR microclusters, which is mediated by the evolutionarily conserved cytoplasmic region of the heavy chain of mIgG-BCR. Furthermore, our findings have clinical relevance because we found that primary B cells from patients with rheumatoid arthritis (RA) generated excessive amounts of traction forces in comparison with B cells from healthy controls during B cell activation. Together, these data prompted us to propose a three-step model, manifesting the anchoring rivet and traction force transmitter functions of BCR microclusters, to explain the molecular mechanism for the generation of traction forces during B cell activation. This model may provide useful information to better understand the function of mechanical forces in B cell activation, which may contribute to the development of better vaccines and therapies for autoimmune diseases.

RESULTS

Dynamic traction forces are generated by B cells on antigen-containing substrates

The physical nature of the traction forces that B cells apply to antigen-containing substrates was measured using traction force microscopy (Fig. 1A). In this approach, B cells were placed on polyacrylamide (PA) gels that were precoated with anti-BCR surrogate antigens according to our published protocols (20). We also anchored fluorescent beads to the surface of the PA gel substrates to accurately track and measure the lateral deformation changes of the PA gels as a reflection of the traction forces that could be applied by the B cells to the antigen-containing substrates. These fluorescent beads served as tracking markers in imaging experiments. Time lapse of both cell phase-contrast and fluorescent bead images was recorded (at 12-s intervals) by confocal fluorescence microscopy to capture and investigate the spatiotemporal dynamics of the traction forces in the B cells. We performed the imaging starting from the earliest time points before the cells established an interaction with the substrates and continued recording up to 30 min. Substrate deformations were measured by computer-aided tracking of the fluorescent beads using a digital image correlation (DIC) algorithm, as reported previously (26). The fluorescence image of the beads after the B cells were detached from the PA gel was used as a reference image. An improved Fourier transform traction cytometry method was used to calculate the traction stress map from the measured substrate deformations (27, 28). We first examined the response of laboratory DT40 cells, a chicken B cell line. There was no detectable displacement of the fluorescent beads at a distance beyond ~1 μm from these B cells (Fig. 1B). This indicates that bead displacements around the B cell reflect the deformation of the PA substrates that were induced by the traction forces generated by the B cells. The stiffness of the PA gel used in these experiments was about 1 kPa, which is similar to the stiffness of antigen-presenting cells (2325).

Fig. 1 Schematic diagram of traction force microscopy.

(A) Schematic representation of the traction force microscopy method used in this study. In this approach, B cells were placed on PA gels that were precoated with anti-BCR surrogate antigens (Ags). To accurately track and measure the lateral deformation changes in the PA gels that are exerted by the B cells, we also anchored fluorescent beads to the surface of the PA gel substrates. (B) Representative images of DT40 cells under the following conditions: B cell spreading on the PA gel (phase contrast), fluorescent beads linked to the PA gel surface (fluorescent beads), the displacement map (displacement) of the substrate calculated from the lateral deformation of the PA gels by DIC, and the corresponding traction stress map (traction stress) computed from the displacement map. Scale bar, 5 μm.

Using this methodology, we acquired a B cell traction stress map consisting of different time points (Fig. 2, A and B). As detailed in Materials and Methods and in our previous studies (27, 28), the total traction force at each individual time point was calculated by integrating the absolute cellular traction stress with the area of cell-substrate interactions. The distribution of the traction force was highly dynamic and exhibited large variations over time at different regions within the interface of the cell’s contact with the antigen-containing surfaces (Fig. 2C). These data revealed the unexpected highly variable spatiotemporal dynamics of the traction forces that were applied to the antigen-containing substrates by the B cells. Further analyses showed that the traction force curve in the total time course correlated with a two-exponential function (regression fitting showed an R value as high as 0.9715; fig. S1A). We calculated the slope of the two-exponential function over time, which represented the rate of the growth of the total traction force (fig. S1B). The value of the slope was very high within the first 5 min, demonstrating the marked increase in the traction forces within this period of time (Fig. 2C and fig. S1B). However, the increase in the traction forces was only mild starting 10 min, as demonstrated by the lack of changes of the value of the slope over time from 10 to 30 min. Thus, in the present study, most of the traction force data generated by the B cells were measured at 20 min after B cell contact with the PA gel substrates, unless otherwise stated. Here, we also quantified the traction work of B cells by multiplying the total traction force by the distance changes, rendering this parameter a unit of joule. We used the traction work done by B cells to accurately quantify the total amount of work that B cells exerted on the antigen-presenting surfaces during B cell activation and found that B cells exerted an average total traction work on the order of 1.00 × 10−15 to 1.59 × 10−15 J during this dynamic process (fig. S2, A to C).

Fig. 2 Traction forces generated by DT40 cells and B6 primary B cells.

(A and B) Representative time-lapse, phase-contrast images of DT40 cells on antigen-coated PA substrates with a stiffness of 1 kPa (A) and the corresponding traction stress map generated by the same cell (B). Scale bar, 5 μm. (C) Total traction force exerted by DT40 cells varied with time up to 30 min. Red lines represent the selected example cells (n = 15 cells), whereas the blue line displays the average of the total number of tested cells (n = 49 cells). (D) Total traction forces generated by DT40 cells incubated on substrates coated with goat anti-chicken IgM (Ag) or neutravidin (NC) for 20 min. Data are means ± SEM of the total traction force calculated from at least 15 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test. (E and F) Representative phase-contrast and fluorescence images of B6 primary B cells incubated on antigen-coated PA substrates with a stiffness of 0.5 kPa (E) and analysis of the total traction forces exerted by B6 primary B cells incubated on PA substrates coated with antigen (Ag) or neutravidin (NC) for 20 min (F). Data are means ± SEM of the total traction force calculated from at least 25 cells in one experiment that is representative of three independent experiments. Scale bar, 10 μm. ***P < 0.001 by two-tailed t test. (G) Total traction forces exerted by DT40 WT cells without (control) or with (RGD) pretreatment with the integrin inhibitor RGD peptide upon stimulation by substrates coated with goat anti-chicken IgM antibodies. Also provided as NC are the total traction forces generated by DT40 WT cells upon stimulation by substrates coated with irrelevant goat anti-mouse IgM antibodies (anti-mouse IgM). Data are means ± SEM of the total traction force calculated from at least 21 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test. (H) Total traction forces exerted by WT primary B6 B cells without (control) or with (RGD) pretreatment with the integrin inhibitor RGD peptide and primary B cells from CD11a KO mice (CD11a-KO) upon stimulation by substrates coated with anti-mouse IgM antibodies. Also provided as NC are the total traction forces generated by WT primary B6 B cells upon stimulation by substrates coated with irrelevant goat anti-chicken IgM antibodies (anti-chicken IgM). Data are means ± SEM of the total traction force calculated from at least 34 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test.

To verify that the observed traction forces were specific to antigen recognition, traction forces were also examined for DT40 cells in contact with PA gel substrates lacking antigen [referred to as negative control (NC)]. On antigen-free substrates, DT40 cells did not spread, and the total traction forces were statistically significantly less than those generated on antigen-containing surfaces (anti-BCR antigen–coated, referred to as Ag) (Fig. 2D). We also repeated these experiments with primary B cells isolated from wild-type (WT) C57BL/6J mice (referred to as B6 B cells) and confirmed that primary B cells generated much higher total traction forces on antigen-containing substrates than on antigen-free substrates (Fig. 2, E and F). To further confirm that these traction forces were induced by the BCR-antigen bonds to the substrate, rather than because antigen recognition by the BCR triggered other B cell surface molecules, such as integrins, to bind to the substrate to exert traction force, we performed the following control experiments. First, we pretreated B cells with a short synthetic Arg-Gly-Asp acid (RGD peptide), which inhibits cell adhesion by binding to integrins (29), before loading the B cells onto the antigen-presenting PA substrates. The results from these experiments showed that blocking integrin function with RGD peptide did not affect the generation of traction forces by either DT40 cells or primary mouse B cells (Fig. 2, G and H). Second, we found that similar amounts of traction forces were generated by primary B cells from WT mice and B cells derived from CD11a (ITGAL) knockout (KO) mice (Fig. 2H), which are deficient in the integrin lymphocyte function–associated antigen 1 (30, 31), the major type of integrin found on B cells (32). Third, as NC, we placed mouse primary B cells on substrates precoated with an irrelevant antibody (goat anti-chicken IgM) that cannot recognize mouse IgM-BCR, or conversely, we placed DT40 cells on substrates precoated with goat anti-mouse IgM as an irrelevant antibody (Fig. 2, G and H). Traction force microscopy results showed that in both types of cells, the values of traction force generated on the substrates presenting irrelevant antibodies were similar to those on antigen-free substrates (NC), both of which were markedly less than the values of the traction forces that were generated on stimulating surrogate antigens (Fig. 2, F and G). Together, these results indicate that the traction force microscopy method algorithm is a reliable method to examine the generation of traction forces from B cells. These results also showed that the traction forces from B cells exposed to antigen-containing substrates were mainly induced by BCR-antigen recognition, although it is also evident that there are very low basal traction forces that are exerted by B cells in the presence of antigen-free substrates.

Myosin and dynein motor proteins are involved in the generation of traction forces

The highly dynamic nature of the traction forces that were applied to the antigen-containing substrates by B cells prompted us to examine the molecular machinery accounting for their generation. Our primary targets were the motor proteins, including dynein and myosin IIA, which are highly abundant in B cells and play important roles in antigen-triggered B cell activation responses (18, 33). To investigate the effects of these two motor proteins on the generation of traction forces in B cells, we used hedgehog pathway inhibitor 4 (HPI4), a specific inhibitor of dynein (34), blebbistatin (referred to as BLEB), a specific inhibitor of the adenosine tripohosphatase activity of myosin IIA (35), and ML7, a myosin light chain kinase inhibitor (36), to disrupt the function of these two types of motor proteins. The inhibitors were used according to the common protocols for HPI4 (incubation at 37°C overnight), BLEB (incubation at 37°C for 30 min), and ML7 (incubation at room temperature for 20 min), as detailed in Materials and Methods. The generation of traction forces by the inhibitor-treated cells was compared with that of the corresponding control cells [which were pretreated with the vehicle dimethyl sulfoxide (DMSO)]. We also examined the MFI of the BCRs within the contact interface between the B cells and the antigen-containing substrates because the BCR MFI value has been widely used as a parameter to characterize the extent of B cell activation (21, 37). The quantification of the MFI of the BCRs demonstrated that the addition of myosin IIA inhibitors statistically significantly decreased the BCR MFI as compared to that of the control cells (DMSO), and the dynein inhibitor had a similar, but relatively, mild effect (Fig. 3A). When examining the generation of the traction forces and traction work, we found that both types of myosin inhibitor, BLEB and ML7, markedly impaired the dynamics of traction force generation and the traction work (Fig. 3B and fig. S3A, respectively) compared to those of the DMSO-treated control cells (Fig. 3B). Furthermore, the total traction forces and traction work were markedly inhibited in BLEB- or ML7-treated cells (20 min after B cell contact with the antigen-presenting substrates; Fig. 3C and fig. S3B, respectively) in comparison to those of the DMSO-treated control cells. Similar to the experiments inhibiting myosin IIA, experiments with the dynein inhibitor showed that dynein is involved in traction force generation (Fig. 3, B and C). These observations were also confirmed in experiments with B6 primary mouse B cells (Fig. 3D and fig. S3C). These results demonstrate that both the myosin and dynein motor proteins are involved in the generation of traction forces from B cells.

Fig. 3 Myosin and dynein are involved in BCR activation and traction force generation.

(A) DT40 cells were pretreated with DMSO (vehicle control) or with the indicated inhibitors before being incubated on PA surfaces coated with goat anti-chicken IgM. The MFIs of the BCRs from each indicated group of DT40 cells were quantitated as the fluorescence intensity of the labeled BCRs averaged over the area of the B cell. Data are means ± SEM of the MFIs analyzed from at least 36 cells in one experiment that is representative of three independent experiments. **P < 0.01 and ***P < 0.001 by two-tailed t test. (B) Time-lapse analysis of the average total traction force generated by DT40 cells treated with DMSO (black), HPI4 (red), BLEB (ochre), and ML7 (blue). Data are from at least 14 cells in one experiment that is representative of three independent experiments. (C and D) Total traction force exerted by DT40 (C) or B6 primary B cells (D) that were pretreated with DMSO (vehicle) or the indicated motor protein inhibitors and then incubated for 20 min on PA substrates coated with goat anti-chicken IgM or goat anti-mouse IgM, respectively. Data are means ± SEM of the total traction force calculated from at least 21 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test.

Dynamic F-actin remodeling regulates traction force, whereas the MFI of BCR microclusters determines the strength of traction force

BCR microclusters are the basic platform in the initiation of B cell activation, and their formation is highly dependent on the remodeling of the F-actin cytoskeleton (18, 38). Thus, we analyzed the codistribution of the traction force with both F-actin and BCR molecules within the contact interface between the B cells and the antigen-presenting substrates. To this end, we detected F-actin with a Lifeact-mCherry–derived labeling strategy in the B cells that were prestained with an Alexa Fluor 647–conjugated Fab anti-chicken IgM antibody. We then imaged F-actin, IgM-BCR, and traction stress by traction force microscopy at 2, 5, 10, 15, 20, 25, and 30 min after placing the B cells on antigen-presenting and fluorescein isothiocyanate (FITC) bead–containing substrates (Fig. 4A). We found that the response of the F-actin cytoskeleton was pertinent to the behavior of BCR during B cell activation because the F-actin cytoskeleton was remodeled during the progress of B cell activation. First, F-actin was diffusely distributed in the contact interface between B cell and substrate surface in the early stage of the B cell activation progress (the first 5 min), whereas F-actin was then remodeled to form an integrative and stable architecture in the peripheral area of the contact interface between the B cell and substrate surface in the middle to late stage of B cell activation progress (refers to the following 5 to 10 min). Disrupting the polymerization of F-actin by treating the cells with latrunculin-B (39) impaired the generation of traction force during B cell activation (Fig. 4B). These results demonstrate that F-actin remodeling is required for the generation of traction force during B cell activation.

Fig. 4 The correlation between the BCR MFI and the strength of the traction forces.

(A) Representative time-lapse traction stress map (top) and the corresponding fluorescence images of F-actin (middle) and the BCR (bottom) of DT40 cells incubated for up to 30 min on PA substrates coated with goat anti-chicken IgM. Scale bars, 5 μm. (B) Total traction force exerted by DT40 cells that were pretreated with either DMSO or latrunculin-B (Lat-B) and then incubated for 20 min on PA substrates coated with goat anti-chicken IgM. Data are means ± SEM of the total traction force calculated from at least 40 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test. (C) Correlation between the strength of traction stress and the MFI of F-actin or BCR at 10 min after incubation for the cells shown in (A). Left: Three representative ROIs (a, b, and c) in the three images represent the same ROIs that were used to calculate the MFIs of F-actin and the BCR and the strength of the traction stress. Scale bars, 5 μm. Right: Correlation shows the linear regression analysis between the MFI of both F-actin and BCR with the traction stress value. Data are from at least 12 cells (28 ROIs per cell) in one experiment that is representative of three independent experiments. (D) Average R value of the linear regression (left) and the Pearson correlation coefficient (PCC; right) between F-actin and traction stress of 12 tested cells varied with time up to 30 min. Red lines represent the selected example cells (n = 12 cells), the ochre line represents the cell shown in (A), and the blue line displays the average of the total tested cells. (E) Average R value of linear regression (left) and the PCC (right) between BCR and traction stress of 12 tested cells varied with time up to 30 min. Red lines represent the selected example cells (n = 12 cells), the ochre line represents the cell shown in (A), and the blue line displays the average of total tested cells.

Subsequently, we acquired the MFI of both F-actin and BCR and the value of the traction stress within the same region of interest (ROI) (Fig. 4C; see the three indicated ROIs). With these data, we profiled the spatiotemporal dynamics of the correlation of traction force with either F-actin or BCR at each time point by the methods of both linear correlation and the Pearson correlation coefficient (PCC) (Fig. 4, D and E). The results showed that there was only weak correlation between traction force and F-actin because both the R value of linear regression and the PCC value were low and failed to exhibit dynamic changes at each time point (Fig. 4D). In contrast, the correlation between traction force value and the MFI of BCR microclusters was initially low in both the linear regression and PCC analysis at the 2-min time point; however, both correlations markedly increased to maximal values at 10 min and then decreased at later time points (Fig. 4E). Together, these results suggest that the MFI of the BCR microclusters positively correlates with the strength of the traction force.

The generation of traction forces is dependent on the activation of membrane-proximal BCR signaling molecules

Having identified motor proteins and BCR microclusters as the source of traction force generation, we continued to investigate the molecular signaling pathways that regulate traction force generation through the BCR complex. It has been reported that the recognition of membrane-tethered antigen by the BCR results in the rearrangement of molecules proximal to the plasma membrane of B cells (6, 40). Subsequently, these rearranged molecules, such as Lck/Yes novel tyrosine kinase (Lyn), spleen tyrosine kinase (Syk), Bruton’s tyrosine kinase (Btk), phospholipase C–γ2 (PLC-γ2), B cell linker (BLNK), and the guanine nucleotide exchange factor Vav3, are required for the initiation and regulation of B cell spreading after stimulation with membrane-tethered antigen. Therefore, we selected a panel of DT40 cell KOs for each of these key signaling molecules and analyzed their ability to generate traction force on antigen-containing substrates. The results showed that KO of most of the BCR signaling molecule, including Lyn, Syk, Btk, PLC-γ2, BLNK, and Vav3, markedly impaired the ability of DT40 cells to generate traction forces and traction work in comparison to WT DT40 cells (Fig. 5A and fig. S4A). To further validate the conclusion that the generation of the traction forces was dependent on these signaling molecules but not on other off-target effects during the generation of these KO B cells, we cloned the complementary DNAs encoding Vav3 and PLC-γ2 from the mRNA of WT DT40 cells and performed a rescue experiment by constructing plasmids encoding either Vav3 or PLC-γ2 fused to a green fluorescent protein (GFP) tag and using these constructs to transfect the Vav3-KO or PLC-γ2–KO DT40 cells, respectively (Fig. 5, B and C). We found that exogenous expression of these molecules in the corresponding KO DT40 cells statistically significantly restored the generation of traction force and traction work (Fig. 5D and fig. S4B). Thus, these data indicate that the initiation of membrane-proximal BCR signaling is required for the sustained generation of traction force.

Fig. 5 Membrane-proximal BCR signaling molecules and adaptor molecules linking BCR microclusters and motor proteins are required for the generation of traction forces.

(A) Scatter diagrams of the total traction forces exerted by WT DT40 cells and indicated KO DT40 cells when incubated for 20 min on PA substrates coated with goat anti-chicken IgM. Data are means ± SEM of the total traction force calculated from at least 28 cells in one experiment that is representative of three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 by two-tailed t test. (B and C) Representative DT40 cell phase-contrast and fluorescence images of Vav3 KO DT40 cells reconstituted with Vav3-GFP (B) and PLC-γ2 KO DT40 cells reconstituted with PLC-γ2–GFP (C) after incubation for 20 min on PA substrates coated with goat anti-chicken IgM. Scale bar, 5 μm. (D) Scatter diagrams showing the total traction forces generated by Vav3-KO DT40 cells, Vav3 KO DT40 cells reconstituted with Vav3-GFP, PLC-γ2–KO DT40 cells, and PLC-γ2 KO DT40 cells reconstituted with PLC-γ2–GFP after incubation as described in (C). Data are means ± SEM of the total traction force calculated from at least 30 cells in one experiment that is representative of three independent experiments. ***P < 0.001 by two-tailed t test. (E and F) Fluorescence intensity (FI) of BCR microclusters (E) and scatter diagrams of total traction forces (F) exerted by WT DT40 cells and the indicated KO DT40 cell lines after incubation for 20 min on PA substrates coated with goat anti-chicken IgM. Data are means ± SEM of the fluorescence intensity calculated from at least 2311 BCR microclusters analyzed from at least 37 cells (E) and the total traction force calculated from at least 45 cells (F) in one experiment that is representative of three independent experiments, respectively. **P < 0.01 and ***P < 0.001 by two-tailed t test.

The generation of traction force requires the loading of BCR microclusters onto motor proteins

Next, we investigated the function of essential bridging molecules that load BCR microclusters onto motor proteins during the generation of traction force. BCRs are transported in a retrograde manner to the center of the B cell immunological synapse along cytoskeletal tracks through motor proteins. A report using mass spectrometry identified three key bridging molecules that load BCR microclusters onto motor proteins (33). Whereas B cells deficient in casitas B cell lymphoma (Cbl), growth factor receptor–bound protein 2 (Grb2), or the third member of the Dok (docking protein) family (Dok-3) can produce BCR microclusters at the peripheral region of the B cell immunological synapse, these BCR microclusters are not transported to the central region of the B cell immunological synapse because of the lack of linkage with motor proteins. Thus, we compared side by side both BCR microcluster formation and the generation of traction forces in WT DT40 cells versus DT40 cells deficient in Cbl, Grb2, or Dok-3. We found that a deficiency in each of these molecules did not affect the formation of BCR microclusters because there was no statistically significant difference in the MFI of BCR microclusters between WT and KO cells in response to antigen stimulation (Fig. 5E). However, loss of any of these molecules markedly impaired the ability of the cells to generate traction force on antigen-presenting substrates (Fig. 5F). Among these, Grb2-KO B cells showed a more marked defect in force generation compared to the Cbl-KO and Dok-3-KO cells. This result is consistent with the reports showing that the recruitment of both Dok-3 and Cbl to BCR microclusters depends on Grb2 (33). Furthermore, Grb2 is a key signaling molecule that mediates multiple cell functions, including remodeling of F-actin by interacting with molecules such as Sos1 (son of sevenless homolog 1), Rho (Ras-like guanosine 5′-triphosphate–binding protein), and Vav3 (4143). Thus, Grb2-KO cells would be expected to exhibit more severe effects than would Cbl-KO and Dok-3–KO B cells in the generation of traction forces in response to antigen stimulation. These results suggest that the loading of BCR microclusters onto motor proteins and the retrograde movement of BCR microclusters are both essential for the generation of traction force.

IgG-BCR–expressing memory B cells generate more traction forces than do IgM-BCR–expressing naïve B cells

B cells use different isotypes of BCRs to recognize antigens during the initiation of B cell activation. Mature naïve B cells use IgM-BCRs, whereas memory B cells mainly use class-switched IgG-BCRs. Memory B cells, but not mature naïve B cells, are mainly responsible for the highly effective antigen recall antibody responses upon vaccine immunization (4, 44). However, there has been a lack of investigation into quantifying traction forces during the activation of IgG-BCR–expressing memory B cells compared to IgM-BCR–expressing naïve B cells. Here, we addressed this question by using primary mature naïve B cells expressing the B1-8–IgM-BCR from IgH B1-8/B1-8 Igκ−/− transgenic (Tg) mice (45) and B1-8 primary B cells expressing memory IgG-BCRs that were derived from an in vitro class-switch response according to our published protocol (46). To achieve an unbiased comparison, we prelabeled the IgM-BCR–expressing mature naïve B cells with the DyLight 649–conjugated Fab fragment from goat anti-mouse IgM (heavy chain–specific) antibodies, whereas the IgG-BCR–expressing cells were prelabeled with the Alexa Fluor 488–conjugated Fab fragment from goat anti-mouse IgG (heavy chain–specific) antibodies. We placed a mixture of both types of prelabeled B cells into the traction force–measuring chambers with substrates coated with 4-hydroxy-3-nitrophenylacetyl antigen, which is the specific antigen for the B1-8–BCRs. We observed that the IgG-BCR–expressing memory B cells generated more traction forces and traction work than did the IgM-BCR–expressing naïve B cells (Fig. 6A and fig. S5A). Next, we investigated the mechanism potentially responsible for the different responses in the naïve and memory B cells. This mechanism is especially intriguing because both the IgG-BCR and the IgM-BCR use the exact same signaling initiation component: the Igα and Igβ heterodimer. Thus, we focused on the BCR component that recognizes the antigens in these two types of BCRs, which are mIgG-BCR and mIgM-BCR (2, 3). mIgG and mIgM differ greatly in the cytoplasmic domains of their respective heavy chains. The heavy chain of mIgM has only 3 amino acids in its cytoplasmic tail (KVK), whereas the heavy chains of all mIgG subtypes have 28 amino acids in their cytoplasmic tails, which are conserved across species (7, 47, 48). Early mouse model studies using biochemical assays and live-cell imaging demonstrated that the cytoplasmic tail of the heavy chain of mIgG is both necessary and sufficient to confer the enhanced activation of IgG-BCR–expressing B cells compared to that of IgM-BCR–expressing B cells (46, 4955). Because the generation of traction forces is dependent on the activation of membrane-proximal BCR signaling, as we demonstrated earlier, it is of interest to examine the contribution of the cytoplasmic tail of the mIgG heavy chain in generating more traction forces in IgG-BCR–expressing B cells compared to IgM-BCR–expressing B cells. To explore this phenomenon, we took advantage of the two types of hen egg lysozyme (HEL)–specific BCR Tg mice (51): (i) Tg mice with mature naïve B cells expressing HEL-specific IgM-BCRs (referred to as IgM) and (ii) Tg mice with mature naïve B cells expressing HEL-specific IgM-BCRs with the cytoplasmic tail of the heavy chain of mIgM swapped with a mIgG cytoplasmic tail (referred to as IgMG). By using a similar traction force–measuring system with substrates presenting HEL antigens, we confirmed that the cytoplasmic tail of the heavy chain of mIgG was sufficient to mediate the enhanced generation of traction force and traction work during the initiation of B cell activation (Fig. 6B and fig. S5B). Together, these results suggest that IgG-BCR–expressing memory B cells generate more traction forces than do IgM-BCR–expressing naïve B cells and, furthermore, that this enhanced effect is likely mediated by the evolutionarily conserved cytoplasmic tail of mIgG.

Fig. 6 Traction forces generated by IgM-BCR–expressing naïve B cells and isotype-switched IgG-BCR–expressing memory B cells.

(A) Quantification of the traction forces generated by IgM-BCR–expressing and isotype-switched, IgG-BCR–expressing B1-8 primary B cells after incubation for 20 min on PA substrates coated with NP8-BSA. Data are means ± SEM of the total traction force calculated from at least 25 cells in one experiment that is representative of three independent experiments. **P < 0.01 by two-tailed t test. (B and C) Quantification of the traction force generated (B) and the FI of the BCR microclusters (C) from mature naïve B cells expressing HEL-specific IgM-BCRs (IgM) and from mature naïve B cells expressing HEL-specific IgM-BCRs with the mIgG cytoplasmic tail (IgMG) after incubation for 20 min on PA substrates coated with HEL. Data are means ± SEM of the total traction force calculated from at least 48 cells (B) and means ± SEM of the fluorescence intensity calculated from at least 3042 BCR microclusters analyzed from at least 50 cells (C) in one experiment that is representative of three independent experiments. **P < 0.01 by two-tailed t test. (D) Representative original (top rows), pseudocolored (middle rows), and 2.5-dimensional Gaussian images (bottom rows) of typical BCR microclusters from mature naïve B cells expressing HEL-specific IgM-BCRs (top) and naïve B cells expressing HEL-specific IgM-BCRs with the mIgG cytoplasmic tail (bottom) tested in (C). Scale bar, 1.5 μm.

Our previous studies demonstrated that the cytoplasmic tail of the IgG-BCR heavy chain promotes the formation of more prominent BCR microclusters of a higher MFI than that in the case of IgM-BCR and, as a consequence, enhance the strength of BCR signaling (46). We speculate that these features may facilitate the generation of traction forces, because we have provided evidence here showing that the MFI of the BCR microclusters is positively correlated to the strength of the traction force. We therefore compared the MFI of the BCR microclusters in primary B cells from these two HEL-specific Tg mice when placed on HEL-containing lipid bilayers according to our previously published protocol (37, 46). We found that cells expressing an mIgM harboring the cytoplasmic tail of mIgG formed statistically significantly larger BCR microclusters than those formed by cells expressing WT mIgM (Fig. 6, C and D). Furthermore, our previous data revealed that, after antigen stimulation, both dynein and myosin IIA were required for B cells to generate traction force. Thus, the abundances of these two motor proteins in B cells would be another key factor determining the magnitude of the traction force generation from B cells. According to the data sets acquired by ImmGen (www.immgen.org/databrowser/index.html) and previous studies (5658), memory B cells have more myosin light and heavy chains and more dynein light and intermediate chains compared to mature naïve B cells (fig. S6). These data suggest that the increased amounts of motor proteins in memory B cells may account for their ability to generate greater traction forces than those of naïve B cells.

B cells from RA patients generate excessive amounts of traction force compared to those of healthy controls

Having shown that the generation of traction force was dependent on the activation of BCR signaling molecules and that the MFI of the BCR microcluster positively correlated with the strength of these traction forces, we were curious to investigate the generation of traction forces in B cells from RA patients because numerous studies have demonstrated that B cells from RA patients exhibit enhanced BCR signaling and produce autoreactive autoantibodies [such as rheumatoid factors (RFs)] (5961). Thus, we compared the generation of traction force in B cells from RA patients with that in B cells from healthy individuals by placing these human primary B cells on the same antigen-containing substrates. To reduce intersample and interbatch variations, we chose three age- and gender-matched pairs of healthy controls and RA patients. In each batch of the experiment, we only compared one pair of samples, a healthy control versus an RA patient. We prelabeled peripheral blood B cells from the paired samples with an Alexa Fluor 647–conjugated Fab fragment from an anti-human IgM constant region (Fig. 7A) and placed these cells on the PA gel substrates presenting anti-human Igκ and anti-human Igλ antibodies, which functioned as the surrogate antigens. The cells were in contact with the antigen-coated PA gel for 20 min. The MFIs of the BCRs (Fig. 7B) were measured, and the traction forces (Fig. 7C) and traction work (fig. S7) were calculated as described earlier. The results showed that in all three paired groups, the BCR MFI and the traction force derived from RA patient B cells were statistically significantly greater than those of the B cells from healthy donors. These findings suggest that B cells from RA patients generate an excessive amount of traction force as compared to those from healthy donors.

Fig. 7 Traction forces generated by B cells from healthy donors and patients with RA.

(A) Representative phase-contrast and BCR fluorescence images of B cells from a healthy control and an RA patient. Insets show contrast-enhanced, magnified views of the respective primary B cells in the dashed boxes. Scale bar, 10 μm. (B and C) BCR MFIs (B) and total traction forces (C) of primary B cells from three pairs of healthy human controls and RA patients after incubation and spreading for 20 min on PA substrates coated with goat anti-human Igκ and Igλ light chain. Data are means ± SEM of BCR MFI (B) or the total traction force (C) calculated from at least 26 cells (per sample) from three pairs of donors. **P < 0.01 and ***P < 0.001 by two-tailed t test.

DISCUSSION

Mechanical forces are thought to be essential for mediating the activation of antigen receptors, including the T cell receptor (TCR) and the BCR, according to our previous studies and those of others (18, 20, 6264). Here, we assessed the traction force exerted by B cells, which was transmitted to elastic substrates through the BCR microclusters at the contact interface between the B cells and the antigen-containing substrates. By using traction force microscopy, we reported here that B cells generated an average total traction force on the order of 10 to 20 nN, which varied with time in response to antigen stimulation. In marked contrast, only very low traction forces were detected in B cells that were placed on antigen-free substrates. We propose that the extremely low extent of traction force generation may be correlated with antigen-searching events by B cells, which are facilitated by membrane-probing behaviors, as reported in our previous study (37). Here, we also quantified the energy consumption during the execution of the traction work by B cells, which is termed as traction work done by B cells. Mechanical potential is generally believed to be one major type of free energy that a cell must exert when functioning (65). However, the mechanism of mechanical energy production and the physiological role of mechanical energy consumption and transmission are not known in immune cells, let alone B cells, mainly because of the lack of a suitable measurement method (66). Changes in energy define the direction for spontaneous changes in chemistry and physics. The gradients of chemical factors, such as chemokines, can induce chemotaxis, suggesting that overcoming the chemical concentration gradient requires work in the form of expenditure of energy (67, 68). Here, we quantified the traction work of B cells by multiplying total traction force with the distance changes, rendering this parameter a unit of joule to accurately quantify the total amount of work that B cells exerted to the antigen-presenting surfaces during B cell activation.

Cell traction force is mainly generated by the actin-myosin cytoskeleton and is exerted on the underlying substrate or the extracellular matrix (ECM) (69). This distinguishes various physiological and pathological behaviors and functions of suspension cells, such as immune cells (6264, 66), and adherent cells, such as fibroblasts, muscle cells, cancer cells, and mesenchymal stem cells (7072). In suspension cells, especially immune cells, traction force generation is used to recognize antigens to further activate immune cells and kill target cells (62). In adherent cells, traction force generation is also used to sense signals from the cellular microenvironment, which is essential for cell morphology maintenance, cell migration, cardiomyocyte contraction, and mesenchymal stem cell differentiation (7072). Traction force can stimulate downstream signaling pathways, which further regulates cell behavior and function (73). On the basis of this model, we propose that traction force is generated to initially enable the B cell to search and capture antigens through the BCRs. It is likely that traction forces can further enhance B cell activation by recruiting more BCRs and membrane-proximal BCR signaling molecules to form larger BCR microclusters, which, in turn, help B cells exert more traction force on the underlying substrates or the ECM, resulting in more efficient antigen acquisition. This positive feedback mechanism, which originates from the cell interior and proceeds to the cell exterior and then back to the cell interior again, is similar to a mechanism reported in adherent cells (73). Taking these findings together, we propose that the higher the traction force that a B cell can generate in response to antigen stimulation, the stronger the B cell activation would be.

Cell contraction is also generally mediated by the actomyosin cytoskeletal network through the sliding of myosin IIA on actin filaments and the moving of dynein along microtubules (33, 74, 75). The motor proteins contribute to the actomyosin cytoskeletal system by generating the contraction force, which is necessary for controlling cell morphology, which, in turn, regulates cell functions. In B cell studies, it is well documented that B cells use dynein to induce retrograde BCR microcluster movement into the center of the B cell immunological synapse, whereas myosin IIA is used by B cells to disrupt the interaction between the BCR and antigens (18). Thus, we investigated the contribution of dynein and myosin IIA to B cell activation and traction force generation. We observed that blocking myosin IIA– and dynein-mediated contractility substantially reduced the generation of traction force and impaired B cell activation.

BCR microclusters are first formed on the cell periphery and are later transported toward a central aggregate through actin filament polymerization and contraction (8). Simultaneously, it is observed that F-actin is also mainly distributed at the periphery of the spreading B cells (76). We assessed whether the F-actin cytoskeleton colocalized with the traction force. Unexpectedly, F-actin did not correlate with the traction force very well, but instead, we observed that the MFI of the BCR microclusters strongly correlated with the traction force. These results lead us to speculate that the BCR microclusters might have a similar function to that of focal adhesion molecules or integrins. Both integrins and the focal adhesion molecules serve multiple functions, including the regulation of adhesion and migration in many adherent cells (77), the regulation of T cell activation (66, 78, 79), and the transmission of traction force from cells to their substrates (64). These possible parallels in function between BCR microclusters and integrins or the focal adhesion proteins may help uncover at least part of the mechanisms that regulate B cell activation and B cell traction force generation.

We found that F-actin contractility was needed for the generation of traction force. This conclusion is supported by the observation that dynamical F-actin remodeling from a diffusely distributed structure in the early stage of B cell activation (before 5 min; Fig. 4A) to an integrative and stable architecture mainly located in the peripheral area of the contact interface between the B cell and the substrate surface in the middle to late stage of B cell activation (5 to 10 min; Fig. 4A). Moreover, we found that disrupting the polymerization of F-actin by latrunculin-B impaired the generation of traction force during B cell activation (Fig. 4B), although there was no strong spatiotemporal correlation between F-actin and force. We speculate that motor proteins are the generator of the traction forces, a process that is performed on the track of actin filaments. It was reported that the motor proteins myosin IIA and dynein interact with their cargo BCR microclusters through the major histocompatibility complex class II invariant chain II (80) and the E3 ubiquitin ligase Cbl and adaptors Grb2 and Dok-3 (33), respectively. We found that the generation of traction forces required the linking of BCR microclusters with motor proteins and the pulling of BCR microclusters on the tracks of the cytoskeleton toward the central region of the B cell immunological synapse. In this case, traction force was transmitted to the force-calculating beads through bonds between the BCR microclusters and the antigen on the substrate. As expected, the traction force was mainly applied to the force-transmitting site that is the circular-shaped BCR microcluster, showing as a positive correlation between the strength of the traction force and the MFI of the BCR microclusters (Fig. 4C). However, it is reported that there is lack of correlation between F-actin filaments and BCR microclusters and that BCR microclusters are mainly located at the F-actin–poor region, with F-actin filaments usually forming a coral-like structure outside of BCR microclusters (81, 82). Thus, it is reasonable to observe that the linear correlation between F-actin and traction force is poor.

This speculation is also supported by the fact that there are no studies supporting the high colocalization between traction force and F-actin in conventional mechanosensing biology studies in other types of cells (83). As for the loss of correlation between BCR MFI and traction stress (decrease of R value) in later stages of activation (after 10 min; Fig. 4D), we think that it could be ascribed to the retrograding movement of BCR microclusters to the center of the B cell immunological synapse (Fig. 4A) because there were BCR microclusters in the central area, whereas there was almost no traction stress. The measured displacements of fluorescent beads in the cell center and horizontal direction parallel to the contact interface between the cell and the substrate were especially experimentally small, which would lead to the absence of local traction forces during the retrograde transport of the BCR to the center of the B cell. Traction force is mainly located at the peripheral region of the contact between the cell and activating substrates according to various studies (28, 8486). Thus, in our model, the correlation between BCR microclusters and traction forces would decrease in the later stage of the time course when the BCR microclusters translocate to the central region through the movement of motor proteins.

On the basis of these findings, we propose a three-step model, manifesting the anchoring rivets and traction force–transmitter function of BCR microclusters, to explain the molecular mechanism for the generation of traction forces during B cell activation (Fig. 8). These three steps include the following: (i) the formation and growth of BCR microclusters in the peripheral area of the contact interface between the B cell and the antigen-presenting substrate surface, which function as the anchoring rivets and the traction force transmitters (Fig. 8A); (ii) the remodeling of the F-actin structure and the establishment of stable interactions between BCR microclusters and motor proteins, which is the step of preparing the “track” and loading the cargo (BCR microclusters) onto the motor proteins for the generation of traction forces (Fig. 8B; magnified region indicates the loading of BCR microclusters onto motor proteins); and (iii) the retrograde movement of BCR microclusters to the center of the B cell immunological synapse on the tracks of F-actin and microtubules by related motor proteins. Our data also indicate that the cytoskeleton- and motor protein–mediated BCR microcluster movement is of fundamental importance for the production and maintenance of traction forces that have been experimentally measured in traction force microscopy (Fig. 8C; red arrows in the magnified region indicate the generation of traction forces during the retrograde movement of BCR microclusters to the center of B cell immunological synapse through the motor proteins).

Fig. 8 Proposed molecular mechanism of traction force generation during B cell activation.

(A) Formation and growth of BCR microclusters in the peripheral area of the contact interface between the B cell and the antigen-presenting substrate surface. (B) Remodeling of F-actin structures and the loading of cargo (BCR microclusters) onto the motor proteins. (C) Retrograde movement of BCR microclusters to the center of the B cell immunological synapse along the tracks of F-actin and microtubules. Red arrows in the magnified regions indicate the generation of traction forces during the retrograde movement of BCR microclusters to the center of the B cell immunological synapse by motor proteins.

Another observation was made when comparing the traction forces delivered by IgM- and IgG-BCRs. As we delineated, more traction forces were generated by IgG-BCR–expressing memory B cells than were generated by IgM-BCR–expressing naïve B cells. Similarly, it is found that, compared to naïve B cells, germinal center B cells apply more persistent and stronger tensile forces on the BCR, which inhibits antigen binding by using myosin II contractility to achieve more strict affinity discrimination when extracting antigens from immune synapses (87). We found that the evolutionarily conserved cytoplasmic tail of mIgG was likely responsible for the enhanced generation of traction force. Mechanistically, we propose the following hypotheses to explain the increased traction force generation in IgG-BCR–expressing memory B cells compared to IgM-BCR–expressing mature naïve B cells. First, the IgG-BCR microclusters are much larger than the IgM-BCR microclusters because the cytoplasmic tail of mIgG promotes IgG-BCR microcluster formation, strengthens the initiation of signaling, and consequently reinforces the generation of traction force, all of which are supported by our published studies (21, 37). Second, memory B cells have greater amounts of motor proteins compared to mature naïve B cells, and these motor proteins promote traction force exertion. A report on T cells also revealed the key role of myosin light chain phosphorylation on the dynamics of microtubule and F-actin structures, which are essential for the generation of traction force (88). We compared the abundances of myosin and dynein mRNAs in mature naïve B cells and memory B cells from humans and mice. We found that the murine myosin light chain 6 mRNA was 1.6-fold more abundant in memory B cells than in mature naïve B cells. Furthermore, the abundance of myosin light chain 9 mRNA was 1.4- and 5.3-fold greater in memory B cells than in mature naïve B cells from mice and humans, respectively. In addition, the mRNA abundances of several murine dynein light chains, including LC-8 (light chain–8 kDa), roadblock, and Tctex (T complex–associated testis–expressed 1–like), were 2.2- to 2.8-fold greater in memory B cells than in mature naïve B cells. Together, these data suggest that memory B cells have increased amounts of motor protein mRNAs compared to those of mature naïve B cells (fig. S6). Thus, the increased production of the traction forces mediated by these motor proteins may lower the threshold of IgG-BCR activation and further promote antigen acquisition by IgG-BCR–expressing memory B cells during the initiation of B cell activation. These data enable a better understanding of the relatively more potent activation of memory B cells.

It is well known that contractile prestress (that is, cell traction force) carried by the actomyosin cytoskeleton can be transmitted to the underlying substrate and the ECM in most adherent cells through the structure of focal adhesions (69). Focal adhesions are also regulated by myosin, which functions as a mechanical output regulator and also a mechanosensor, because myosin is linked to focal adhesions through F-actin (75, 89). Myosin can also regulate the formation of focal adhesions during the mechanotransduction process. On the other hand, myosin participates in the formation of micro-adhesion rings to induce the formation of TCR microclusters, which are essential for initial T cell activation through the outside-in signaling of integrins (90). On the basis of these findings and our experimental results, we propose that myosin may also function as a mechanical output regulator and a mechanosensor in B cells. Thus, myosin may participate in the positive feedback loop, as described earlier. Various myosin isotypes, especially the light chains, are activated by phosphorylation to generate and sense cytoskeletal mechanics. Thus, the increased or decreased amounts of these myosin isotypes can lead to enhanced or reduced activation of such a positive feedback loop to further stimulate or inhibit the activation of different subsets of B cells, respectively.

In terms of the potential clinical relevance of our findings, we found that B cells from RA patients generated greater traction force during activation than did B cells from healthy individuals, which may help, at least in a part, to uncover the pathological mechanism of the production of autoreactive antibodies in RA patients. RA is a chronic autoimmune disease with complex pathological mechanisms involving the interplay of multiple cell types and the cross-linking of multiple signaling pathways. B cells play several critical roles in the pathogenesis of RA because B cells can both respond to and produce the chemokines and cytokines that assemble at the sites of inflammation (59, 91). Furthermore, B cells produce RFs, which are autoantibodies specific for the constant region of self-IgG antibodies (59, 91). The dysfunctional molecular signaling pathways involved in B cell activation, the changes in expression profiles of genes important for B cell function and behavior (59, 60), and the abnormality of B cell spreading– and force generating–related molecules (such as focal adhesion kinase families) are all responsible for the progression of RA (92). B cells may respond to the changes in the biochemical and biophysical properties of the cartilage because of the remodeled ECM microenvironment during the autoimmune reactions that occur during RA progression (93). This scenario is similar to the aforementioned positive feedback loop. Autoantigen-reactive B cells are naturally present in both healthy individuals and patients with autoimmune disease; however, it is not completely understood why these autoreactive B cells can remain quiescent in healthy individuals but are activated in autoimmune patients. Our data support the hypothesis that the changed biophysical and biochemical properties of the cartilage ECM microenvironment in RA patients may potentially induce B cells to generate increased traction forces. This increase in traction force may be greater than the force barrier that determines whether B cells remain quiescent or become activated to commence the production of autoantibodies, such as RFs. The mechanism of CD8+ T cell–mediated killing of target cells is coupled to a concept known as “mechanopotentitation,” which predicts that mechanical force modulates information flow out of the cell by potentiating cytotoxicity (62). Consistent with our observations here, the enhanced activation of autoreactive B cells from RA patients is defined by excessive BCR microcluster formation, downstream signaling initiation, and further differentiation of autoantibody-producing plasma cells. Therefore, we propose that excessive traction force generation may serve as another index of the dysregulated activation of autoreactive B cells from RA patients. Excessive traction force generation may play a role in the enhanced activation of autoreactive B cells from RA patients. Together, these findings suggest a three-step working model as the molecular mechanism to define the origin, spatiotemporal dynamics, and function of traction force generation during B cell activation. Furthermore, these studies revealed that B cell traction force generation is markedly increased in the physiological condition of class-switched IgG-BCR–expressing memory B cells and in the pathological condition of primary B cells from RA patients.

MATERIALS AND METHODS

Cells and reagents

All of the chicken DT40 cell lines were provided by T. Kurosaki (RIKEN, Japan). All human primary naïve B cells isolated from the peripheral blood of RA patients and healthy donors, as well as mouse primary naïve B cells, were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 50 μM β-mercaptoethanol (Sigma-Aldrich), and penicillin/streptomycin antibiotics (Invitrogen). The WT DT40 cell line and all of the DT40 KO cell lines used in this study were maintained at 37°C in RPMI 1640 medium supplemented with 1% chicken serum. Biotin-conjugated goat anti-chicken IgM antibody was purchased from Rockland Inc. Biotin-conjugated goat anti-human Igκ and Igλ light chains were purchased from Southern Biotech. DyLight 649–conjugated Fab anti-mouse IgM constant region antibody and biotin-conjugated goat F(ab)2 anti-mouse IgM antibody were purchased from Jackson ImmunoResearch. The labeling of mouse anti-chicken IgM antibody (clone M1) with Alexa Fluor 647, the labeling of HEL with biotin, and the digestion of the Fab fragment of mouse anti-chicken IgM antibody (clone M1) were performed as previously described (46).

Immunostaining

To evaluate the abundance and distribution of the BCR, DT40 cells, B6 mouse cells, or human primary B cells were prestained with an appropriate anti-BCR antibody conjugated with a specific fluorochrome (100 nM) on ice for 5 min before extensive washing, as outlined earlier, respectively. The MFIs of BCRs within the contact interface between the B cell and the antigen-tethered gel were processed and analyzed with ImageJ (National Institutes of Health) or MATLAB (MathWorks) software, as described previously (20).

PA gel preparation and surface conjugation of fluorescent beads

To measure weak traction forces more accurately, we improved the traditional traction force microscopy method using surface chemical modifications to fabricate a PA gel with fluorescent beads on surface, as described previously (94, 95). Briefly, the glass bottoms of 35-mm dishes were pretreated with bind-silane to ensure PA gel attachment. Here, gels with acrylamide/bis ratios of 3:0.03 (for a stiffness of 0.5 kPa) and 3:0.1 (for a stiffness of 1 kPa) were prepared to examine the traction force generation of primary human and mouse B cells or DT40 cells, respectively. After polymerization, the PA gel surface was completely covered by 0.2-μm-diameter fluorescent beads (rhodamine/FITC carboxylate-modified, diluted at 1:400 in water) for 18 min, and then, the beads were covalently linked to the gel surface with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide, hydrochloride [Invitrogen; 3.8 mg/ml in 2-(N-morpholino)ethanesulfonic acid (MES; pH 5.5; Sigma-Aldrich)] and hydroxy-2,5-dioxopyrrolidine-3-sulfonic acid [Sigma-Aldrich; 7.6 mg/ml in MES (pH 5.5)] solution for 2 hours and then in phosphate-buffered saline (pH 7.4) for 2 hours at room temperature, as previously described (94). The gel substrates were activated with sulfo-SANPAH (Pierce), were subsequently coated overnight at 4°C with neutravidin, and were then incubated with biotin-conjugated goat anti-chicken IgM for DT40 cells, biotin-conjugated goat F(ab)2 anti-mouse IgM for B6 mouse primary B cells, and biotin-conjugated goat anti-human Igκ and Igλ light chains for human primary B cells for 2 hours at 37°C. Before measuring the traction force, the gels were blocked with 1% casein for 30 min at room temperature.

Measurement of traction forces

To measure the traction force generated by B cells, movies of live cells and fluorescent beads were acquired every 12 s during B cell spreading for up to 30 min in phase contrast for the cells, in the 488 or 561 channels for the beads, and in the 640 channel for the BCR with an epifluorescence microscope (Ti-E, Nikon) combined with a spinning-disk laser confocal scanning microscope (PerkinElmer). For the traction force experiments, all of the data were acquired at 40× magnification. For the correlation experiments, the data were recorded at 60× and 100× magnifications, together with prestaining of the BCR with Alexa Fluor 647–conjugated mouse Fab anti-chicken IgM and staining of F-actin with Lifeact-mCherry. During image acquisition, the dishes were kept at 37°C by means of a live-cell station. After image acquisition, fluorescence images of fluorescent beads were acquired as reference images after the B cells were detached from the gel. A series of dynamic fluorescence images of the substrates were first recorded by the microscopy system during the cell-substrate interactions. Sample drift was corrected for by tracking the displacement of beads located at the marginal area, which were far away from any cell, and the displacement field of the gel exerted by the B cells was measured by analyzing the positions of the fluorescent beads with DIC before and after cell detachment (26). On the basis of the substrate displacement field and Young’s moduli of substrates, which had been measured in advance, we then reconstructed cellular traction stress fields (in units of pascal) by the optimal filtering approach founded on Fourier transform traction cytometry implemented in MATLAB, as previously described (27, 28). In this context, we obtained the total traction force Ftotal (in nanonewton) by means of the following expression:Ftotal=A|T|dA(1)where T denotes the local cellular traction stress (in pascal), and A is the area of the cell-substrate interaction. Accordingly, we further defined mechanical traction work (in joule) during the cell-substrate interplay as:Wm=A(Td)dA(2)where d is the displacement field of the substrate. In experiments with inhibitors, both DT40 cells and B6 mouse primary B cells were pretreated with inhibitors as follows: 20 μM RGD peptide to block integrin at 37°C for 2 hours, 30 μM HPI4 at 37°C overnight, 50 μM BLEB to block dynein at 37°C for 30 min, and 10 μM ML7 to block myosin IIA and myosin light chain kinase at room temperature for 20 min. As a control, DT40 cells and B6 mouse primary B cells were pretreated with DMSO at 37°C overnight as the control for HPI4, at 37°C for 30 min as the control for BLEB, and at room temperature for 20 min as the control for ML7. In experiments to disrupt the polymerization of F-actin, the DT40 cells were treated with 1 μM latrunculin-B for 10 min.

Data analysis

For all assays, the values shown in the figures are means ± SEM. Statistically significant differences were determined by one-way analysis of variance (ANOVA), followed by a t test for multiple comparisons between groups, and two-exponential function regression, together with curve slopes, was calculated with GraphPad Prism software.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/542/eaai9192/DC1

Fig. S1. Total traction force regressed to a two-exponential function of time.

Fig. S2. Traction work done by DT40 cells and B6 primary B cells when exerting traction forces.

Fig. S3. Myosin and dynein are involved in the production of traction work in B cells.

Fig. S4. Membrane-proximal BCR signaling molecules are required for sustained traction work.

Fig. S5. Traction work done by IgM-BCR–expressing naïve B cells and isotype-switched IgG-BCR–expressing memory B cells.

Fig. S6. Myosin and dynein mRNA abundances in IgM-BCR–expressing naïve B cells and IgG-BCR–expressing memory B cells from mice and humans.

Fig. S7. Traction work exerted by B cells from healthy controls and RA patients.

REFERENCES AND NOTES

Acknowledgments: We thank S. K. Pierce (National Institute of Allergy and Infectious Diseases, NIH), K. Rajewsky (Immune Regulation and Cancer, Max Delbrück Center for Molecular Medicine), M. Shlomchik (University of Pittsburgh), C. Goodnow (Garvan Institute of Medical Research), and T. Kurosaki and H. Shinohara (World Premier International Immunology Frontier Research Center, Osaka University) for providing experimental materials. Funding: This work was supported by funds from the Ministry of Science and Technology of China (2014CB542500 to W.L. and 2013CB933702 to C.X.), the National Natural Science Foundation of China (81422020 and 81621002 to W.L. and 11472013 to C.X.). Author contributions: J.W. and F.L. designed, carried out, and analyzed all experiments and wrote the manuscript. Z.W., Y.L., and Y.Z. commented on the study to assist with its improvement. J.H. calculated the traction work. X.S., W.Z., and Z.L. provided patient blood samples. F.W. and Y.-H.C. helped with the revisions of the manuscript. Y.S. provided the Tg mouse cells. W.L. and C.X. supervised the study and helped with the writing and editing of 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.
View Abstract

Navigate This Article