Germinal center B cells recognize antigen through a specialized immune synapse architecture

Abstract

B cell activation is regulated by B cell antigen receptor (BCR) signaling and antigen internalization in immune synapses. Using large-scale imaging across B cell subsets, we found that, in contrast with naive and memory B cells, which gathered antigen toward the synapse center before internalization, germinal center (GC) B cells extracted antigen by a distinct pathway using small peripheral clusters. Both naive and GC B cell synapses required proximal BCR signaling, but GC cells signaled less through the protein kinase C-β-NF-κB pathway and produced stronger tugging forces on the BCR, thereby more stringently regulating antigen binding. Consequently, GC B cells extracted antigen with better affinity discrimination than naive B cells, suggesting that specialized biomechanical patterns in B cell synapses regulate T cell-dependent selection of high-affinity B cells in GCs.

Figure 1

Large-scale imaging shows subset-specific differences in B cell synapses. Gating in image datasets on B cell subsets using fluorescence intensities of membrane markers and image galleries of B cell synapses after 20 min incubation with antigen (anti-Igκ) on PLBs. (a) Gates show T1 (IgM⁺ CD23^-) and T2-3 (IgM^+/int CD23⁺) cells. (b) Gates show marginal zone (MZ) (CD21⁺ CD23^-) and follicular (FO) (CD21^lo CD23⁺) cells. (c) Quantification of synaptic features. Data are means and s.e.m. from n = 92 T1, 161 T2-3, 4504 FO, and 459 MZ cells from one out of three experiments. Scale bar, 5 μm.

Figure 2

GC B cells exhibit unique synaptic architecture. (a) Plot of antigen (anti-Igκ) centralization in synapses of B220-positive splenic cells from immunized mice after 20 min on PLBs. Cells with low centralization (red rectangle) are shown in a Fas versus CD38 versus dotplot. (b) Galleries showing synapses of Fas⁺ CD38^- GC B cells and Fas^- CD38⁺ follicular naive B cells. (c) Quantification of synaptic features of naive and GC B cells. (d) Left, galleries of synapses of pre-enriched GC and naive B cells showing F-actin localization in the synapse plane. Right, correlation of actin and antigen fluorescence in the synapse. (e) Antigen centralization and maximum projection images of synapses of naive (B220⁺ Fas^- CD38⁺ IgM⁺ IgG1^-), pre-enriched GC B cells (B220⁺ Fas⁺ CD38^- IgM⁺ IgG1^- or B220⁺ Fas⁺ CD38^- IgM^- IgG1⁺, and IgG-memory B cells (B220⁺ Fas^- CD38⁺ IgM^- IgG1⁺) imaged on the same PLB. Data are from one experiment representative of two (d), three (a,b,c) or four (e) independent experiments. Means and s.e.m from n = 1596 (c), 453 (d) naïve and 28 (c) or 65 (d) GC b cells. **** = p < 10^-4, ** = p < 0.005 in unpaired t-tests. Scale bars, 5 μm.

Figure 3

Live cell imaging reveals the dynamics of synapse formation in naive and GC B cells. (a) Kymographs of antigen (anti-Igκ) in the synapses on PLBs. (b) Fluorescence intensities of antigen clusters tracked from their appearance to fusion with other clusters or their disappearance. (c) Growth rates of antigen clusters in the first two minutes of synapse formation. Data are from one cell each from one out of three experiments (a,b), or means and s.e.m. of n = 29 naive and 24 GC cell antigen clusters from three experiments (c). **** = p < 10^-4 in an unpaired t-test. Scale bar, 5 μm.

Figure 4

Signaling in naive and GC B cell synapses with PMSs. (a) Dot plot of mean synaptic antigen and phospho-BLNK intensities in each cell. Gating shows cells outside (synaptic antigen intensity^lo) and on PMSs. (b) Images of phospho-BLNK and antigen (anti-Igκ) in the plane of the synapse and in a sideview reconstruction. (c) Quantification of intracellular signaling in naive and GC B cells. (-), cells outside of PMSs, +, cells on PMSs. Synapses were analysed after 10 min (p-Syk) or 20 min (all other staining). Data are normalized to the signal in naive cells outside of PMSs. (d) Intracellular calcium levels in cells interacting with antigen on PLBs. Left, individual live cell traces of Fluo-4/Fura-Red ratios from one experiment. Right, calcium area under the curve (AUC). Error bars represent mean and s.e.m from n = 3 experiments. * = p < 0.05 between GC cells and the corresponding naive cell groups in paired t-tests. Scale bars, 5 μm.

Figure 5

Different activities of PKCs and NFκB, in naive and GC B cells. (a-f) Quantification of staining with antibodies against phospho-PKCαβ (a-c) and PKCδ(d-f). (a, d) Levels of synaptic staining analyzed as in Fig. 4. (b, e) Recruitment of PKCs to the synapse calculated as the ratio of synaptic to total cell intensity. Data are from the same experiments as in (a, d). (c, f) Membrane recruitment of PKCs in cells stimulated with PMA plus ionomycin (+) compared to unstimulated cells (-). (g) Nuclear translocation of p50. Top, representative images of p50 and DAPI staining in naive and GC B cells unstimulated, or stimulated with glass-attached antigen or CD40L. Bottom, large-scale image quantification of p50 nuclear to cytoplasm ratios. All data are means and s.e.m. from n = 3 (a-c, f-g) or 4 (d, e) experiments. * = p < 0.05 in paired t-tests between the GC and corresponding naive cell groups or between the bracketed groups. Scale bar, 5 μm.

Figure 6

Antigen internalization by naive and GC B cells from synapses with PMSs. (a) Quantification of total extracted antigen per cell after 20 min on PMSs. For visualization, cells with zero antigen extraction were assigned a random background value. The box shows cells that interacted with PMSs and therefore were considered for analysis. FU, fluorescence units. (b) Features of extracted antigen clusters. (c) Sideview sections through the middle of cells showing degradation of a DNA sensor attached to antigen. (d) Percent of extracted antigen that was degraded (Deg.) after 30 min on PMSs. Nondeg, antigen labeled with a nondegradable sensor. (e) Sideview sections through the middle of cells showing surface staining for antigen. (f) Percent of extracted antigen that was internalized after 30 min. (g) Effect of inhibition of signaling and myosin contractility on antigen extraction. Data are means and s.e.m of n = 982 naïve, 854 GC cells from one out of five experiments (b), n = 465-982 cells from two experiments (d,f). Data in (g) show means and s.e.m. normalized to internalization in DMSO-treated naive B cells, n = 3 experiments. **** = p < 10^-4, *** = p < 10^-3, * = p < 0.05 where p is significance in nonparametric tests (b, d, f) or paired t-tests (g). Scale bars, 5 μm.

Figure 7

Regulation of antigen binding by mechanical forces in naive and GC B cell synapses. (a) Images of sensor fluorescence and the fluorescence ratio indicating 9 pN sensor opening in synapses with PLBs. (b) 9 pN force sensor fluorescence and ratios in a tracked BCR-antigen cluster. (c) Sensor opening ratios as a function of the sensor opening force. The control sensor lacks the force-sensitive hairpin. (d) Mean antigen (anti-Igκ) fluorescence in synapses of cells with or without blebbistatin treatment, normalized by antigen fluorescence outside of cells. Data are means and s.e.m from n = 33-49 (b), n = 12-19 (d) naive, and 16-59 (b), 11-12 (d) GC synapses from two (b) or three (d) experiments. **** p < 10^-4 in unpaired t-tests. Scale bar, 5 μm.

Figure 8

Affinity discrimination by naive and GC B cells. Naive, CTV-labeled B1-8 B cells were imaged on NP₁₅- or NIP₁₅-containing PMSs together with donor-enriched B cells from mice that received B1-8 B cells and were immunized with SRBC-NIP. (a) Levels of phospho-BLNK and (b) antigen extraction by B1-8 naive (CTV⁺, Fas^-) and GC (CTV^-, Fas⁺) B cells. (c) Ratio of NIP₁₅ to NP₁₅ antigen extracted from PMSs by the B1-8 naive and GC B cells. Data show mean and s.e.m. normalized to values of naive B cells on NIP15 from n = 1693 and 1517 naive, and 81 and 84 GC B cells aggregated from five experiments (a, b), or mean and s.e.m from n = 5 experiments (c). **** = p < 10^-4 in nonparametric tests (b) or in a paired t-test (c).