Bystander B cells rapidly acquire antigen receptors from activated B cells by membrane transfer

Abstract

The B cell antigen receptor (BCR) efficiently facilitates the capture and processing of a specific antigen for presentation on MHC class II molecules to antigen-specific CD4⁺ T cells (1). Despite this, the majority of B cells are thought to play only a limited role in CD4⁺ T cell activation because BCRs are clonotypically expressed. Here, we show, however, that activated B cells can, both in vitro and in vivo, rapidly donate their BCR to bystander B cells, a process that is mediated by direct membrane transfer between adjacent B cells and is amplified by the interaction of the BCR with a specific antigen. This results in a dramatic expansion in the number of antigen-binding B cells in vivo, with the transferred BCR endowing recipient B cells with the ability to present a specific antigen to antigen-specific CD4⁺ T cells.

The activation of antigen-specific CD4⁺ T cells relies on the capacity of antigen-presenting cells (APCs) to internalize and present an antigen, as peptide fragments, on MHC class II (MHC-II) molecules. Dendritic cells (DCs) are regarded as the most potent APC (2), but B cells expressing a B cell antigen receptor (BCR) specific for a particular antigen are also extremely proficient at capturing and presenting an antigen to antigen-specific CD4⁺ T cells (1, 3, 4). This arises from the remarkable efficiency of the BCR in capturing a specific antigen and facilitating antigen delivery to MHC-II-rich compartments that specialize in the generation of MHC-II-peptide complexes (1). However, because BCRs are clonotypically produced within a large repertoire of B cells, it has been generally accepted that the majority of B cells play only a minor role in presenting antigen to T cells (3, 4).

Recent reports have demonstrated that a number of cell types, including pheochromocytoma cell lines (5), DCs (6–8), T cells (9), natural killer cells (10), and B cell lymphomas (11) have the capacity to transfer membrane and cellular components. In the case of B cell lymphomas, membrane exchange is thought to be due to transfer of membrane components during direct cell contact (11), and possibly via membrane nanotubes (also referred to as cytonemes) (5, 10, 12), in processes that are enhanced by BCR cross-linking. Because the most unique difference between antigen-specific B cells is their BCR, membrane exchange may allow the exquisite antigen-binding and processing capacity of one B cell to be transferred to another via BCR transfer. By this means, it can be postulated that BCR transfer from antigen-specific B cells to bystander B cells during immune responses could enhance the capacity of the local B cell pool to specifically bind and present antigen, a process that may greatly enhance the development of CD4⁺ T cell responses. In the present study, we examined the potential of B cells to exchange their BCRs and how this affects their capacity to present a specific antigen.

Results

B Cells Share BCRs After Activation.

The demonstration of membrane exchange between B cell lymphomas upon BCR cross-linking (11) suggests that BCRs may be transferred between B cells upon cell activation. To assess this possibility, we cultured splenocytes from hens-egg lysozyme (HEL)-specific BCR-transgenic (Tg) MD4 mice (13) with splenocytes from C57BL/6 (B6).CD45.1 mice in the presence of the B cell mitogen, LPS. B cells from each of the two populations were discriminated on the basis of CD45 allotypic differences and BCR exchange determined by changes in the cell-surface expression of Ig (Ig)M allotypes as B6.CD45.1 B cells express the “b” allotype of IgM (IgM^b), whereas the MD4 B cells carry the “a” allotype of IgM (IgM^a). Using this approach, we observed substantial transfer of IgM^a to B6.CD45.1 B cells and a concomitant transfer of IgM^b to MD4 B cells (Fig. 1A), with the acquisition of donor IgM steadily increasing over 3 days of culture. Transfer could also be observed by confocal microscopy, which showed the acquired BCR to be distributed evenly over the cell surface [supporting information (SI) Fig. 6]. Negligible BCR transfer was observed in the absence of LPS, but similar BCR transfer was observed when a specific antigen in the presence of antigen-specific T helper cells, or T helper cell signals, were used to activate B cells (SI Fig. 7 A and B). BCR transfer was found to be restricted to B cells because B220-negative cells acquired negligible amounts of donor IgM (Fig. 1B).

Fig. 1.

B cells exchange BCRs after prolonged activation by LPS. (A) B6.CD45.1 and MD4 spleen cells were cocultured in the presence of LPS. Dot plots show expression of IgM^a and IgM^b on live (Hoechst 33258^low) CD45.1⁺ (red) and CD45.1⁻ (black) B220⁺ cells over 3 days. Values are percentages of B6.CD45.1 B cells expressing IgM^a (red) or MD4 B cells expressing IgM^b (black) relative to background expression on B cells cultured independently (represented by quadrant gates). (B) Transfer of IgM^a to B6.CD45.1 B220⁺ cells and B220⁻ cells after 3 days of culture with MD4 cells (red open histograms). Gray filled histogram represents background IgM^a staining of B6.CD45.1 splenocytes cultured alone, and black open histogram depicts IgM^a expression on MD4 B cells.

BCR Transfer Is Rapid, Is Not Due to Secreted Ig, and Is Enhanced by Specific Antigen.

The rate of BCR transfer between B cells was assessed by mixing whole cultures of LPS-activated MD4 and B6.CD45.1 splenocytes and determining IgM^a transfer to B6.CD45.1 B cells over short periods of coculture (Fig. 2A). At 37°C, substantial BCR transfer occurred within 10–20 min, and by 60 min, B6.CD45.1 B cells had gained ≈60% of the IgM^a levels that they had acquired after 3 days of continuous culture with MD4 splenocytes (Fig. 2A). Similar IgM^a transfer was observed with washed activated MD4 splenocytes, whereas low IgM^a transfer was observed with the MD4 culture supernatant (CSN), this residual activity being completely depleted when CSN was passed through an 800-nm filter (Fig. 2A). Intriguingly, rapid BCR transfer also occurred after a 1-h incubation at 4°C, a condition that normally inhibits endocytic and exocytic processes, including Ig secretion (14) (Fig. 2A). Indeed, we have found that activated B cells from μ_s^−/− mice, which express surface IgM^a but have essentially no capacity to secrete IgM^a (15), have a similar capacity to transfer their BCRs relative to wild-type IgM^a secretors (Fig. 2B). These results indicate that IgM transfer is not due to IgM secretion.

Fig. 2.

Activated B cells rapidly transfer BCRs to both resting and activated B cells through nonsoluble factors in a process enhanced by specific antigens. (A) Fractions of day-3 LPS-activated MD4 spleen cell cultures were added to 3-day LPS-activated B6.CD45.1 spleen cells and incubated at 37°C for up to 2 h or at 4°C for 1 h. The percentage of B6.CD45.1 B cells expressing IgM^a was assessed as in Fig. 1A and IgM^a transfer represented as a proportion (%) of IgM^a transfer over 3 days of continuous coculture. MD4 cell culture fractions included whole culture, MD4 cells washed to remove soluble factors, MD4 culture supernatant (CSN) depleted of cells by centrifugation, and the cell-depleted CSN passed through an 800-nm filter. (B) The capacity of B cells from μ_s^−/− mice to transfer IgM^a to B6 B cells after continuous coculture for 3 days at 37°C or after brief (3 h) coincubation at 4°C. For 3-day incubations, B6 spleen cells (IgM^b) were cocultured with spleen cells from μ_s^−/− mice (IgM^a) or 129sv (μ_s^+/+) mice (the founder strain for the IgM^a expressed in μ_s^−/− mice but that can secrete IgM^a) in the presence of LPS. For brief incubations, 3-day LPS-activated spleen cells were incubated together for ≈3 h. B6 B cells were then assessed for IgM^a surface expression. Numbers without parentheses are the percentages of B6 B cells expressing IgM^a after coculture with donor B cells relative to IgM^a expression by B6 B cells cultured alone. Numbers in parentheses are mean fluorescence intensities of the respective histograms. The histogram depicts the concentration of IgM in the culture supernatants of either μ_s^−/− or μ_s^+/+ splenocytes cultured for 3 days with LPS. (C) Day-3 LPS-activated MD4 spleen cells were incubated with either 3-day LPS-stimulated (activated) or freshly isolated (resting) B6.CD45.1 spleen cells for 1 h at 4°C. B6.CD45.1 B cells were then assessed for IgM^a surface expression (Left) and MD4 B cells assessed for IgM^b surface expression (Right). Data are expressed as the percentage of B cells expressing nonendogeneous IgM. (D) Day-3 LPS-activated MD4 spleen cells were incubated with LPS-activated B6.CD45.1 spleen cells in the presence of HEL or OVA for 1 h at 4°C. B6.CD45.1 B cells were then assessed for IgM^a surface expression represented as a proportion (%) of IgM^a transfer over 3 days of continuous coculture. Data in A, C, and D are representative of at least three independent experiments.

Freshly isolated B cells were unable to transfer their BCRs, whereas LPS-activated B cells could readily donate BCRs to both nonactivated and activated B cells (Fig. 2C), although only viable B cells mediated transfer (SI Fig. 8). Furthermore, the presence of a specific antigen (HEL), but not an unrelated antigen [ovalbumin (OVA)], substantially increased the transfer of HEL-specific IgM^a to both LPS-activated and freshly isolated bystander B cells (Fig. 2 D and SI Fig. 9). BCR transfer, therefore, appears to depend on donor B cell activation, does not require recipient B cell activation, and can be further enhanced by BCR-specific antigen.

BCR Donation to Bystander B Cells Involves Membrane Transfer.

It appears likely that BCR transfer is mediated by membrane donation. Consistent with this, we observed that CD45.2 molecules and cell-surface molecules covalently labeled with fluorescein were also transferred from MD4 cells to B6.CD45.1 B cells, and those B cells that had acquired higher levels of these surface molecules were also the same B cells that gained higher levels of IgM^a, suggesting cotransfer of the molecules (Fig. 3A). Furthermore, by labeling the activated MD4 B cells with the membrane intercalating dye, PKH-26, we could directly show membrane transfer in parallel with transfer of IgM^a (Fig. 3B). This was further enhanced by the addition of BCR-specific antigen over a wide concentration range (Fig. 3B). These results suggest that membrane donation between B cells is responsible for BCR transfer.

Fig. 3.

BCR transfer occurs concomitantly with transfer of other membrane components. (A) (Upper) Simultaneous expression of IgM^a and CD45.2 on B6.CD45.1 B cells cultured with MD4 B cells for 3 days with LPS. The percentage of B6.CD45.1 B cells positive for IgM^a and CD45.2 after coculture with MD4 B cells is shown adjacent to the respective axis (and arrow) for each marker. The boxed area represents background staining for CD45.2 and IgM^a of B6.CD45.1 spleen cells cultured alone. (Lower) Simultaneous transfer of IgM^a and prelabeled MD4 surface molecules to resting B6.CD45.1 B cells. Day-3 LPS-activated MD4 spleen cells were covalently cell-surface labeled with fluorescein and prelabeled with antibodies to IgM^a and CD45.2. After extensive washing, cells were incubated for 1 h at 37°C with freshly isolated B6.CD45.1 spleen cells prelabeled with antibodies to B220 and CD45.1 and immediately analyzed by flow cytometry. The percentage of B6.CD45.1 B cells positive for IgM^a and fluorescein is shown as in Upper. The boxed area represents background staining for IgM^a and background fluorescein fluorescence of freshly isolated B6.CD45.1 spleen cells. (B) Day-3 LPS-activated MD4 spleen cells, labeled with PKH-26, were incubated with day-3 LPS-activated B6.CD45.1 spleen cells for 1 h at 4°C in the presence of various concentrations of HEL. B cells were then assessed for expression of PKH-26 and IgM^a. Numbers are the percentage of B6.CD45.1 B cell expressing each marker. (C) Day-3 LPS-activated MD4 spleen cells were stained with PKH-26 and an anti-B220 mAb, incubated on a cooling stage set at 4°C with anti-B220 mAb-labeled day-3 LPS-activated EGFP-Tg spleen cells, and analyzed by confocal microscopy. Red, PKH-26; green, EGFP. (Scale bars, 5 μm.) (i and ii) Green and red channel merging of simultaneous confocal images of B cells, showing many membrane adhesions between the two B cell populations within 60 min of coincubation (i, 59 min; ii, 17 min). Arrows show overlap of EGFP and PKH-26 fluorescence within the same focal plane. Data in A–C are representative of at least three independent experiments.

Confocal microscopy was used to directly visualize membrane exchange between PKH-26-labeled activated MD4 B cells and activated EGFP⁺ B cells (Fig. 3C). PKH-26⁺ and EGFP⁺ B cells rapidly formed cell aggregates (Fig. 3 Ci and Cii), and diffusion of PKH-26 into the membranes of recipient EGFP⁺ B cells was evident, suggesting that PKH-26-labeled membranes were redistributing into the membranes of adjacent EGFP⁺ B cells. Analysis of the interaction between LPS-activated lymphocytes and bystander resting lymphocytes by transmission and scanning electron microscopy (TEM and SEM, respectively) revealed close association between plasma membranes after cell mixing (SI Fig. 10), with membranous nanotube-like extensions also connecting the cells. Intriguingly, attempts to interfere with most of the known molecular processes involved in membrane fusion and exchange (recently reviewed in ref. 16) had no inhibitory effect on BCR transfer (SI Fig. 11).

Bystander B Cells That Acquire an Antigen-Specific BCR Gain the Ability to Present Antigen to CD4⁺ T Cells.

To test whether the donated BCR could enhance antigen presentation, we assessed the capacity of bystander B cells, after they had acquired the BCR from antigen (HEL)-specific B cells, to stimulate CD4⁺ T cells specific for the same antigen. Purified B cells from CBA/H (H-2^k) and EGFP⁺ MD4 mice were cocultured, under activating conditions, to allow transfer of the HEL-specific BCR of the MD4 B cells to the bystander CBA/H B cells. The different B cell populations were separated by flow cytometry and, after pulsing with HEL, assessed for their ability to stimulate CFSE-labeled HEL-peptide/I-A^k complex-specific CD4⁺ T cells from TCR-Tg 3A9 mice (17). This revealed that CBA/H B cells that had acquired the HEL-specific BCR stimulated, over a 1,000-fold antigen dose range, a substantial proportion (up to 70%) of the HEL-specific CD4⁺ T cells to up-regulate CD69 and proliferate based on CFSE dilution (Fig. 4 and SI Fig. 12A). In contrast, the CBA/H B cells that were not cocultured with the MD4 B cells induced only a low proportion of the HEL-specific T cells to up-regulate CD69 and proliferate. The enhanced T cell responses induced by CBA/H B cells could not be attributed to contaminating MD4 B cells because the sorted CBA/H B cell population was of high purity (<1% EGFP⁺ cells, data not shown), and MD4 B cells lacked the appropriate I-A^k MHC-II for effective antigen presentation.

Fig. 4.

Bystander B cells that acquire an antigen-specific BCR gain the ability to present antigen to CD4⁺ T cells. B cells purified from CBA/H (H-2^k) spleen were cultured with LPS in either the absence or presence of purified splenic B cells from MD4/EGFP-Tg (H-2^b) mice. After 3 days of culture, EGFP⁺ MD4 B cells were depleted from CBA/H-MD4 B cell cocultures by flow cytometry. CBA/H B cells purified from the cocultures (CBA/H +/- MD4), as well as CBA/H and MD4 B cells cultured alone, were then pulsed with HEL for 20 min at 4°C and washed and equal numbers (1.5 × 10⁵) cultured with 1 × 10⁵ purified CFSE-labeled CD4⁺ 3A9 TCR-Tg T cells specific for a HEL-peptide presented by I-A^k. After 0.5 days, CD4⁺ T cells were assessed for CD69 expression. Data are representative of three independent experiments.

These results raise the important issue of whether bystander B cells could be activated through the acquisition of an antigen-specific BCR, with the consequent induction of nonspecific antibody responses. However, B6.CD45.1 B cells that had acquired the MD4 BCR displayed no measurable intracellular Ca²⁺ flux above control B6.CD45.1 B cells after cross-linking with an IgM^a-specific antibody (SI Fig. 12B). In contrast, these B cell populations gave a strong intracellular Ca²⁺ response when exposed to an antibody that cross-linked their endogenous IgM^b. Therefore, donated BCR are able to enhance specific antigen presentation to CD4⁺ T cells without delivering normal signaling events associated with BCR-mediated B cell activation.

Antigen-Specific B Cells Can Donate Their BCR to Bystander B Cells During Antigen-Specific Immune Responses in Vivo.

To assess whether BCR transfer occurs in vivo, an antigen-specific immune response was initiated in mice. Freshly isolated CFSE-labeled MD4 spleen cells, mixed with a HEL-OVA antigen conjugate and LPS, were injected i.v. into B6.CD45.1 recipient mice containing adoptively transferred CFSE-labeled OVA-peptide/I-A^b complex-specific TCR-Tg OT-II lymphocytes (18) to provide cognate T cell help. The IgM^a expression on all CD45.2⁺ (donor) and CD45.2⁻ (host) leukocytes was monitored over a 7-day period (Fig. 5Ai). This showed that at day 3 after adoptive transfer, the MD4 B cells had begun to donate their BCR to bystander B cells, with ≈28% of the host B cell pool expressing IgM^a at this time point. This correlated with extensive MD4 B cells proliferation (Fig. 5Aii). The MD4 B cells did not proliferate much further over the next 4 days, but the number of host B cells acquiring IgM^a increased to 39% on day 4 and was maintained at approximately one-third of the B cell pool over the next 3 days (Fig. 5 Ai and Aii). Detection of BCR transfer could not be attributed to nonspecific antibody binding, because antibody isotype controls showed no binding above background, and control experiments with mice injected with LPS, antigen, and OT-II cells, without the presence of MD4 cells, showed antibody binding levels nearly identical to those of the isotype control (Fig. 5 and data not shown). As observed in vitro, in vivo BCR sharing was essentially limited to B220⁺ cells (Fig. 5Ai). BCR transfer was also considerably enhanced 1 h after challenging the animals with the BCR-specific antigen (i.e., 63% of the bystander B cells acquired the donor BCR, Fig. 5B). CD45.2 transfer to host B6.CD45.1 B cells was also evident in vivo, a finding consistent with membrane transfer (Fig. 5C). Indeed, costaining for both IgM^a and CD45.2 revealed that host B cells appeared to have acquired both markers in parallel, and this coincidental transfer increased not only after in vivo challenge with HEL but also after ex vivo exposure of B cells for 1 h at 4°C to a wide range of HEL concentrations. Recipient B cells that had acquired the HEL-specific BCR also gained the ability to present antigen to CD4⁺ T cells (SI Fig. 13).

Fig. 5.

Antigen-specific B cells transfer their BCR to bystander B cells in vivo during antigen-specific immune reactions. (A) CFSE-labeled MD4 spleen cells, together with 10 μg of HEL–OVA and 10 μg of LPS, were injected i.v. into B6.CD45.1 mice 2 h after the i.v. injection of CFSE-labeled OT-II lymphocytes. Spleen cells from the mice were analyzed over the next 7 days by flow cytometry. (i) Dot plots show B220 and IgM^a expression on host (CD45.2⁻, black events) and donor (CD45.2⁺, red events) cell populations, with numbers referring to the percentage of splenocytes (boxed region) that are MD4 B cells (B220⁺, CD45.2⁺, IgM^ahigh). Histograms show IgM^a expression on host B cells (B220⁺, CD45.2⁻, CFSE⁻) and MD4 B cells (B220⁺, CD45.2⁺, IgM^ahigh) and isotype control mAb binding to host B cells, with numbers showing the percentage of host cells expressing IgM^a relative to the isotype control. (ii) Viable MD4, OT-II, and recipient (host) B cells were assessed for CFSE expression. (B) An identical experiment to that described in A, except that on day 3 after commencement of the experiment, one animal was challenged i.v. with a 10-μg bolus of HEL 1 h before spleen cell harvest and marker analysis. (C) As in A, except that on day 4 after commencement of the experiment, one animal was challenged i.v. with a 10-μg bolus of HEL 1 h before spleen harvest. In addition, spleen cells from a non-HEL-injected animal were incubated at 4°C for 1 h in the presence of various amounts of HEL before measurement of marker expression. Analysis involved assessing expression of IgM^a and CD45.2 on B6.CD45.1 host B cells compared with autofluorescence and isotype control mAb binding (boxed region). Numbers are the percentages of B6.CD45.1 B cells expressing CD45.2 or IgM^a above background levels, with MD4 B cells that express high levels of IgM^a and CD45.2 being gated out of the analysis.

Discussion

In this report, we describe a mechanism by which bystander B cells can acquire antigen-specific BCRs from activated B cells and gain the ability to capture and present specific-foreign antigen, thereby increasing the effective APC pool. BCR transfer is mediated by direct membrane donation, demonstrating an important role for membrane transfer between antigen-specific and non-antigen-specific B cells during immune responses. Recently, there have been several reports indicating that many cell types have the capacity to transfer membrane and cellular components as a form of intercellular communication (9–11), although the functional significance of this phenomenon is unclear. It should be noted, however, that DCs (7, 8) and B cells (19) have been reported to acquire antigen tethered to cell surfaces and efficiently present these to antigen-specific T cells. In these cases, antigen could be acquired from multiple cell types, but uptake appeared to be restricted to DCs (7, 8) or B cells bearing an antigen-specific BCR (19) and hence appeared to be recipient-, but not donor-, driven as we have reported here. As with these reports, and despite our own extensive studies (SI Fig. 11), the molecular basis of membrane exchange is uncertain. However, our confocal and electron microscopy studies suggest that plasma membranes from activated and bystander B cells may coalesce via short membranous extensions resembling short membrane nanotubes. In this regard, it should be noted that membrane nanotubes can transfer material between cells at temperatures as low as 0.7°C (5), and similarly, BCR transfer occurs quite efficiently at 4°C. Furthermore, it has been reported that the formation of membrane nanotube-like extensions, referred to as cytonemes, are increased upon BCR stimulation (12), which in our study significantly enhances BCR transfer.

An intriguing aspect of this study is the speed and magnitude of BCR transfer between B cells, even at 4°C. Once appropriately activated, B cells can, within minutes, share their BCR with adjacent B cells, this process being substantially enhanced after BCR engagement by a specific antigen. Furthermore, in vivo studies revealed that up to two-thirds of the splenic B cells in recipient animals gained the HEL-specific BCR of the transferred Tg B cells once the Tg B cells had been specifically activated by antigen and CD4⁺ T helper cells. This represents at least a 9- to 16-fold expansion in the number of B cells that can bind significant levels of specific antigen. Additional studies revealed that the bystander B cells that acquired the HEL-specific BCR could very efficiently present HEL to HEL-specific TCR Tg T cells, these B cells being able to stimulate antigen-specific CD4⁺ T cell responses with >1,000 times less antigen than bystander B cells that have not acquired specific BCRs. Thus, based on these data, BCR sharing results in a rapid expansion in the number of B cells that can present specific antigen to T cells. Significantly, a number of studies have identified an important role for B cells (20–23) and, in particular, B cells bearing antigen-specific BCRs (22), in CD4⁺ T cell responses. Therefore, we postulate that BCR transfer is an important mechanism by which B cells can help facilitate the amplification and development of antigen-specific CD4⁺ T cells during an immune response.

Methods

Animals.

Mice were obtained from the Animal Services Division, Australian National University and from the Australian Phenomics Facility and were bred under specific pathogen-free conditions. Mouse strains used were B6, CBA/H, and B6.CD45.1 (B6 congenic for CD45.1). Tg mouse strains were MD4 [BCR-Tg expressing HEL-specific-IgM^a and IgD^a on a B6 background (13)], OT-II [TCR-Tg specific for I-A^b-OVA_323–339 peptide on a B6 background (18)], and 3A9 [TCR-Tg specific for I-A^k-HEL_46–61 peptide on a B10.BR background (17)]. The Rosa-EGFP Tg (EGFP-Tg) mice were generated by crossing a Rosa26 stop/flox-EGFP mouse (kindly provided by Martyn Goulding, Department of Neurobiology, Salk Institute, University of California at San Diego, La Jolla, CA) with a generalized Cre recombinase-expressing mouse TNAP Cre (24), to activate expression of EGFP. Double-Tg (MD4/EGFP-Tg) mice were also used and were generated by crossing MD4 mice with EGFP-Tg mice. Secretory IgM^a deficient B6 mice [μ_s^−/− (15)] and 129sv (IgM^a) mice were generously provided by Michael R. Ehrenstein (Department of Rheumatology, University College, London, U.K.) and tested for IgM secretion by ELISA. Mice were used at 4–20 weeks of age.

Cell Preparation and Purification.

Leukocytes were obtained from spleen and/or lymph nodes as described (25). Leukocyte subsets were purified by magnetic cell separation in LS columns (Miltenyi Biotec) using streptavidin-conjugated MicroBeads (Miltenyi Biotec) to target biotin-conjugated mAb-labeled cells. CD4⁺ T cells were enriched from pooled lymph nodes and spleen, and B cells were enriched from spleen. The cells were incubated with biotin-conjugated mAbs (Pharmingen) specific for unwanted cell populations with mAbs used for CD4⁺ T cell enrichment being specific for CD8 (53-6.7), CD11b (M1/70), CD11c (HL3), and B220 (RA3-6B2) and with mAbs used for B cell enrichment being specific for CD4 (GK1.5), CD8 (53–6.7), CD11b (M1/70), CD11c (HL3), and CD90.2 (53-2.1). Negatively selected B cell and T cell populations were found to be 90–98% pure, as assessed by flow cytometry.

Fluorescent Dye and Covalent Labeling of Cells.

Lymphocytes were labeled with the intracellular dye, CFSE (Molecular Probes) and were cell-surface labeled with LC-N-hydroxysuccinimidyl-fluorescein (Pierce), as described for CFSE labeling (25). PKH-26 (Sigma) labeling was performed according to the manufacturer's instructions. Hoechst 33258 (Calbiochem) was used to discriminate among viable, dead, and apoptotic cells (26). Cells (1–5 × 10⁶ per milliliter) were labeled with 1 μg/ml of Hoechst 33258 for 7 min at 37°C before flow cytometry.

Preparation of HEL–OVA Conjugates.

Maleimide-activated OVA (Sigma) was conjugated to sulfhydrylated-HEL (Sigma) to generate stable HEL–OVA conjugates. OVA was maleimide activated with a 25-fold molar excess of succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; Pierce). HEL was incubated with a 7-fold molar excess of N-succinimidyl S-acetylthioacetate (SATA; Pierce), and conjugated SATA was deacetylated in 0.5 M hydroxylamine/25 mM EDTA in PBS (pH 7.4). Activated proteins were then conjugated with a 4-fold molar excess of HEL to OVA before being dialyzed against PBS in a 10-kDa cut-off dialysis bag.

Cell Culture, Cell Culture Fractionation, and Antigen-Presentation Assays.

Activated B cells were generated by culturing purified B cells or RBC-depleted splenocytes (1 × 10⁶ per milliliter) in supplemented DMEM (sDMEM; GIBCO–BRL) containing 10% FCS as described (25) in six-well plates (Nunc) in the presence of 10 μg/ml LPS (Sigma) for up to 3 days at 37°C in 5% CO₂. In some experiments, cultures were harvested and cells readjusted to a final concentration of 1 × 10⁶ cells per milliliter for reincubation in 96-well U-bottomed plates (Nunc) either at 37°C in 5% CO₂ or at 4°C (on ice) for various times. To remove CSN components, cells were washed at least three times with 10 ml of PBS. CSN was harvested by pelleting cells at 300 × g for 5 min and aspirating off 3/4 of the uppermost supernatant. Full removal of cell debris was accomplished by filtering CSN through 800-nm-cut-off cellulose filters (Millipore).

For antigen presentation assays, purified CFSE-labeled 3A9 TCR-Tg CD4⁺ T cells (1 × 10⁵) were cultured with or without 1.5 × 10⁵ purified B cells, in a total of 200 μl of sDMEM/10%FCS in 96-well U-bottomed plates (Nunc). B cells were pulsed with HEL on ice for 20 min, washed, and cultured with CD4⁺ T cells. Cultures were incubated for 15–18 h or for 3 days, at which time, cells were analyzed by flow cytometry for CD69 expression and CFSE content.

In Vivo Experimentation.

RBC-depleted MD4 splenocytes (3 × 10⁷), with 10 μg of HEL–OVA and 10 μg of LPS, were adoptively transferred via the lateral tail vein into B6.CD45.1 mice in 200 μl of PBS. B6.CD45.1 hosts had received i.v. 3 × 10⁷ RBC-depleted CFSE-labeled lymphocytes from OT-II lymph nodes and spleen 2 h before MD4 adoptive transfer to provide cognate T cell help. Spleen cells from host mice were harvested at various times, depleted of RBC by lysis, and adjusted to 1–4 × 10⁷ cells per milliliter, ready for antibody staining and flow cytometry. In some experiments, animals were challenged i.v. with an additional bolus 10 μg of HEL 1 h before spleen harvesting.

Antibody Staining.

Cells for flow cytometry analysis were stained on ice with specific mAbs and secondary fluorochrome conjugates as described (25). MAbs used are described in SI Table 1. The specificity of antibody binding and secondary reagents was monitored through the use of isotype-matched control antibodies (eBioscience).

Flow Cytometry.

Analytical flow cytometry was performed in an LSR flow cytometer (Becton Dickinson), and cell sorting to purify B cell populations was performed by using a FACStar plus (Becton Dickinson) or a FACSVantage with the DiVa option (Becton Dickinson). Postacquisition gating was used to analyze cell subsets (doublet discrimination is detailed in SI Text). Unless otherwise indicated, statistical analysis of the proportion of a cell population expressing a marker was based on histogram profiling using Overton subtraction (27) as applied by FlowJo software (Tree Star).

Confocal Microscopy.

Confocal microscopy was performed with a Radiance Confocal Microscope (Bio-Rad). For live-cell imaging, 1 × 10⁵ LPS-activated MD4 splenic B cells, prelabeled with PKH-26 and a Cy-Chrome-conjugated B220-specific mAb, were added to 5 × 10⁵ Cy-Chrome-conjugated B220-specific mAb-labeled LPS-activated EGFP-Tg splenic B cells on a cooling stage set at 4°C and images taken for 60 min.