The early involvement of marginal zone (MZ) B lymphocytes in T-independent immune responses is well established. In this study we compared the abilities of MZ and follicular (FO) B cells to collaborate with T cells. After immunization with soluble hen egg lysozyme, both MZ and FO B cells captured Ag and migrated to T cell areas in the response to hen egg lysozyme. MZ B cells were far superior to FO B cells in inducing CD4+ T cell expansion both in vitro and in vivo. MZ, but not FO, B cells, after interaction with T cells, differentiated into plasma cells, and in addition they stimulated Ag-specific CD4+ T cells to produce high levels of Th1-like cytokines upon primary stimulation in vitro. These results indicate that MZ B cells rapidly and effectively capture soluble Ag and activate CD4+ T cells to become effector T cells. The enhanced capacity of MZ B cells to prime T cells in this study appeared to be intrinsic to MZ B cells, as both MZ and FO B cell populations express an identical Ag receptor.
An initial Ag-specific T cell activation requires interactions between T cells and APCs, including B cells (1, 2, 3, 4). Reciprocal interactions between T and B cells are initiated by Ag capture through the B cell receptor (BCR),3 followed by Ag processing and presentation of peptide-MHC class II complexes, along with costimulatory signals to Ag-specific T cells (5, 6). These Ag-specific T cells then, in turn, help B cells for Ab production through T cell-derived cytokines and direct physical contacts (1, 2, 3, 4). Together with the orchestrated set of responses by other cell types, such as dendritic cells (DCs), B and T cells then proliferate and differentiate into mature effector cells. Subsets of B cells also respond to T-independent (TI) Ags, and this response depends on interactions with subpopulations of DCs (7, 8).
The splenic marginal zone (MZ) is critical in the first line of defense against blood-borne particulate pathogens (9, 10, 11). MZ B cells residing in this area differ in phenotype and function from follicular (FO) B cells, which are in the majority in the spleen. Accumulating evidence points to a major role of MZ B cells in Ab responses against TI Ag (8, 9, 10, 11, 12, 13, 14, 15, 16, 17). In contrast to the established role of MZ B cells in TI immune responses, the function of MZ B cells in the primary TD immune response has not been well studied. Previous studies showed that bona fide memory B cells in the MZ were able to generate large numbers of Ab-forming cells in secondary responses to haptenated proteins (18). However, it is not known whether naive MZ B cells residing in this area are able to efficiently activate T cells. Recently, we have shown that freshly isolated MZ B cells from naive animals exhibit high levels of B7.1 and B7.2 indicative of previous antigenic experience. Additionally, upon activation with LPS or anti-CD40, activated MZ B cells induce vigorous alloreactive T cell responses, suggesting that MZ B cells also play a role in TD immune responses (19).
In this study we demonstrate that activated MZ B cells are potent protein Ag presenters to CD4+ T cells and have the ability to induce Ag-specific T cell clonal expansion both in vitro and in vivo. In addition, MZ B cells provide signals for CD4+ T cells to produce polarized cytokines after primary stimulation. In contrast, under the same conditions of primary stimulation, FO B cells were poor inducers of CD4+ T cell proliferation and cytokine production. However, these FO B cell-primed CD4+ T cells were competent to produce effector cytokines after secondary stimulation in vitro. Collectively, our findings provide evidence that in addition to the involvement of MZ B cells in the initial response to TI Ag, they mount rapid and efficient primary responses to soluble protein Ag and have an extraordinary ability to promote T cell proliferation and cytokine production after immunization with protein Ag.
Materials and Methods
Eight- to 12-wk-old MD4 Ig transgenic mice with B cells specific for hen egg lysozyme (HEL) (20), 3A9 TCR transgenic mice specific for HEL peptide 46–61 bound to I-Ak (21), and nontransgenic C57BL/6 or B10.BR mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were bred and maintained in the animal facility at University of Alabama (Birmingham, AL).
Flow cytometric analysis and cell sorting
Four-color surface staining and analysis were performed as previously described (19). Data from stained cell samples were acquired using FACSCalibur and CellQuest software (BD Biosciences, Mountain View, CA) and were analyzed with WinList 6.0 (Verity Software House, La Jolla, CA) or Win MDI 2.0 (email@example.com) software programs.
FO and MZ B cells were separated using anti-CD19, anti-CD23, and anti-CD21 mAbs as previously described after B cell enrichment from RBC-depleted spleen cells by treatment with anti-CD43, anti-CD11b, and anti-CD11c magnetic beads and negative selection by using AutoMACS (Miltenyi Biotec, Auburn, CA) (12, 22).
CD4+ T cells were obtained from spleens of 3A9 TCR transgenic or littermate mice after depletion with biotinylated anti-CD8α, anti-CD11b, anti-CD19, and anti-I-Ab, or anti-I-Ak and streptavidin-conjugated magnetic beads, which yield 95–98% purity. In some experiments CD4+ T cells were further purified by staining with anti-CD4-PE and separated on a MoFlo cell sorter (Cytomation, Ft. Collins, CO).
CD11c+ cells were prepared and enriched as previously described by treatment with anti-CD11c-conjugated magnetic beads and passed through the AutoMACS (Miltenyi Biotec). CD11c+ CD8α+ and CD11c+ CD8α− DCs were further purified using a MoFlo (Cytomation) (23).
In vivo Ag priming and in vitro Ag presentation
MD4 or nontransgenic littermate mice were immunized i.v. with 1 mg/mouse of HEL or OVA. MZ and FO B cells were isolated at 4 or 8 h after Ag priming, treated, and stimulated as previously described (24). Fifty microliters of supernatant was collected at 42 h for a CTLL-2 assay, and cultures were pulsed with 1 μCi of [3H]thymidine for 6 h. Cells were harvested, and [3H]thymidine incorporation was measured in a scintillation counter (Wallac, Gaithersburg, MD).
B cells obtained from mice primed with Ag 8 h previously were irradiated and preincubated at 4°C with 2.5 μg/ml blocking reagents, anti-B7.1, anti-B7.2, a combination of anti-B7.1 and B7.2 mAbs (BD PharMingen, San Diego, CA), or recombinant mouse CTLA-4/Fc chimera (R&D Systems, Minneapolis, MN). The percentage of inhibition was calculated as follows: [1 − (cpm of culture with blocking reagent/cpm of control culture)] × 100.
Cell migration and immunofluorescence analysis of tissue sections
After MD4 transgenic mice received 100 μg/mouse HEL-Alexa-488 i.v., spleens were collected at different times. Frozen sections were processed and stained as previously described (19). Spleen sections were stained with MOMA-1 developed with goat anti-rat IgG-7-aminomethylcoumarin (AMCA) and goat anti-mouse IgM-rhodamine isothiocyanate Abs.
In cell migration and homing studies, 2 × 106 sorted MZ or FO B cells from MD4 transgenic mice were adoptively transferred i.v. into nontransgenic recipients. After 24 h, recipient mice were injected i.v. with 100 μg/mouse HEL. Eight hours after cell transfer, spleens were collected, and frozen sections were prepared as previously described.
MD4/B10.BR-F1 mice were immunized i.v. with 1 mg/mouse HEL or OVA. After 8 h, MZ or FO B cells were isolated and cultured separately in 200 μl of complete medium together with 3A9 CD4+ T cells in 96-well, round-bottom plates. On days 3, 5, and 7, supernatant was collected, and IFN-γ, IL-12, IL-4, IL-5, and IL-10 levels were measured using a double-sandwich ELISA (Quantikine M ELISA kit; R&D Systems) according to the manufacturer’s instructions.
For secondary stimulation, T cells from primary cultures were recovered, washed, and replated onto anti-CD3-coated, 48-well, flat-bottom plates (5 μg/ml) in 500 μl of complete medium. On day 3, supernatant was collected and measured in triplicate for IFN-γ, IL-12, IL-4, IL-5, and IL-10 as previously described.
In vivo Ag presentation and T cell clonal expansion
MD4/B10.BR-F1 mice were immunized i.v. with HEL (1 mg/mouse). After 8 h, 2–3 × 106 purified HEL-primed MZ or FO B cells were transferred i.v. together with 2–3 × 106 CD45.1+3A9 CD4+ T cells loaded with 1.5 μM 5-chloromethylfluorescein diacetate (CMFDA) (25) into C57BL6/B10.BR-F1 nontransgenic recipients. Control recipients received MZ, FO, or T cells alone. At 2, 3, 5, and 7 days after adoptive transfer, recipients were sacrificed, and spleens, lymph nodes, and lungs were analyzed for transferred cells by flow cytometry and immunohistochemistry as previously described.
MZ B cells activate naive T cells more rapidly and efficiently than FO B cells
The ability of HEL-primed-MD4 MZ and FO B cells to activate T cells was compared by coculturing them with syngeneic naive CD4+ T cells in the presence of 1 μg/ml of anti-CD3 mAb (26). After 4 h of in vivo priming, T cell proliferation and IL-2 production were much greater with primed MZ than with FO B cells, and this ability increased with time after priming (Fig. 1,B). At higher densities of primed MZ B cells, T cell proliferation declined, most likely due to T cell overexpansion and susceptibility to activation-induced cell death (27). OVA-primed MZ and FO B cells from MD4 and nontransgenic littermate immunized with HEL or OVA stimulated only low levels of T cell proliferation and no IL-2 production (Fig. 1, A and B, and data not shown). It has been previously shown that this nonspecific B cell activation is due to nonspecific BCR-mediated Ag uptake (28).
Thus, both MZ and FO B cells acquired T cell stimulatory capacity that peaked at 8 h after interaction with Ag. However, MZ B cells were much more effective than FO B cells in activating Ag-specific naive T cells.
B7 molecules on MZ B cells are involved in T cell activation
As shown in Fig. 2 A, the basal level of B7.2 expression was higher on MZ than on FO B cells, and after 4 and 8 h of in vivo priming, both MZ and FO B cells up-regulated B7.2 expression, with higher levels being achieved by MZ B cells (mean fluorescence intensity of MZ cells, 85; vs 34 for FO cells). Similar to previous studies, B7.1 was not up-regulated in response to BCR-mediated signals (29).
Anti-B7.1 and anti-B7.2 mAbs inhibited T cell proliferation more in cultures stimulated with primed FO B cells (Fig. 2,B), but both mAbs together produced almost complete inhibition of IL-2 production in cultures stimulated with either MZ or FO B cells (Fig. 2,C). Blocking of both B7.1 and B7.2 with mouse CTLA-4-Ig or anti-B7.1 and anti-B7.2 mAbs produced additive inhibition of T cell proliferation and IL-2 production (Fig. 2, B and C, and data not shown).
Although MZ B cells have a higher initial basal expression of B7.2 than FO B cells, further activation enabled them to costimulate naive CD4+ T cells to proliferate and secrete IL-2. Although FO B cells can also capture Ag through the BCR, they appear to be much less efficient in up-regulating B7 molecules, resulting in a lower ability to trigger T cell responses.
Activated MZ B cells present Ag to Ag-specific T cells more rapidly and efficiently than FO B cells.
We next investigated the ability of MZ and FO B cells to present Ag to HEL-specific CD4+ T cells in vitro after i.v. immunization with soluble HEL. This immunization strategy had previously been shown to successfully prime naive Ag-specific T cells both in vitro and in vivo (24, 30, 31).
MZ B cells were much more potent than FO B cells in the induction of Ag-specific T cell proliferation and IL-2 secretion (Fig. 3, A and B). OVA-primed MZ and FO B cells from MD4 mice cultured under the same conditions did not induce any T cell responses (data not shown). These results indicate that in vivo-primed MZ B cells not only respond to Ag by expressing higher levels of B7.2 than FO B cells, but they are also capable of engaging the TCR on naive CD4+ T cells, leading to effective T cell activation and cytokine production.
Comparison of activated MZ B cell and DC abilities to present Ag
As DCs are the most competent APCs of the immune system (32), we directly compared the Ag-presenting capabilities of DC and B cell subsets isolated from HEL-primed mice. Two DC populations can be distinguished in the spleens of mice; CD8α+DEC-205+CD11b− (lymphoid) DCs are generally located in the T cell areas, whereas CD8α−DEC-205−CD11b+ (myeloid) DCs are located in the MZ areas (33, 34, 35).
As previously demonstrated by others (36), in vivo-primed CD11c+ CD8α− DCs were the most potent at priming naive CD4+ T cells in our system (Fig. 4), although at higher numbers the differences became less pronounced. CD11c+CD8α− and CD11c+ CD8 α+ DCs were equally efficient in inducing Ag-specific T cell proliferation, but at higher cell densities CD11c+CD8α− DCs were better than CD11c+CD8α+ DCs in stimulating IL-2 production. However, CD11c+CD8α+ DCs were less effective than MZ B cells in the induction of IL-2 production from CD4+ T cells.
MZ B cells induce Th1-like cytokine production
As shown in Fig. 5, MZ B cells were more efficient than FO B cells in inducing Ag-specific T cells to produce a Th1-like cytokine profile with high levels of IFN-γ and low levels of IL-4, IL-5, and IL-10. The levels of IFN-γ were almost 10–100 times greater than those of Th2 cytokines. HEL-primed FO B cells induced only low levels of IFN-γ and IL-10 and undetectable levels of IL-4 and IL-5 in primary cultures. No cytokine secretion was detected in T cells cultured with OVA-primed B cells or in the absence of B cells (data not shown).
Additionally, T cells previously stimulated with HEL-primed MZ B cells produced robust amounts of IFN-γ with relatively low levels of IL-4, IL-5, and IL-10 after restimulation (Fig. 5), and the differences between MZ and FO B cells were less pronounced than in the primary stimulation. These results show that not only do MZ B cells readily respond to Ag resulting in priming of naive T cells, they also provided signals for Th1 development in the primary response.
MZ B cells rapidly capture Ag and migrate toward the T cell areas
Alexa-488-conjugated HEL was seen mostly on the surface of MZ B cells and to a much lower degree on FO B cells and minimally on DCs and macrophages at 30 min after i.v. injection of fluorescent HEL into MD4 mice (Fig. 6, A and E). However, at 4 and 8 h after immunization, both MZ and FO B cells had captured comparable amounts of HEL as indicated by the bright staining with red anti-IgM and green HEL. MZ B cells were depleted from the MZ area by 4 and 8 h (Fig. 6, B and C), and the majority of intact HEL had been cleared from the spleen after 24 h as indicated by the absence of green staining except for a few isolated cells.
When purified MZ or FO B cells from unimmunized MD4 transgenic mice were adoptively transferred into nontransgenic recipients and immunized 24 h later with 100 μg of HEL, the majority of transferred MZ and FO B cells were detected in T cell areas (Fig. 7, B and D), suggesting that both populations responded to Ag and were able to migrate to the T-B border. In contrast, donor cells in unimmunized recipients were scattered throughout the MZ and FO areas.
MZ B cells activate Ag-specific CD4+ T cell clonal expansion in vivo
T cells from 3A9 × CD45.1 F1 mice were transferred into unmanipulated C57BL/6 × B10.BR F1 mice together with MZ or FO B cells previously primed for 8 h in vivo. As expected, CD4+CD45.1+Vβ8+ 3A9 T cells were detected only in the spleens of recipients that received 3A9 donor T cells (Fig. 8,A). These cells were large and expressed high levels of CD44 when cotransferred with B cells. This activated phenotype was not observed in donor 3A9 T cells recovered from recipients that received T cells alone or by CD4+ CD45.1− cells of the recipients (Fig. 8, A and B, and data not shown). The accumulation of 3A9 T cells in spleens of recipients receiving cotransferred MZ B and T cells peaked on day 3 when there was ∼3-fold more expansion than in those transferred with FO B cells. The higher T cell numbers in the MZ B-T cell transfer was due to a higher rate of T cell division, as the majority of T cells had undergone four to six cell divisions compared with three or four divisions in the FO B-T cell cotransfer (Fig. 8,B). As the differences in the numbers of cell division were minimal, activated MZ B cells might also play a role in supporting T cell survival in vivo. By day 5 after cell transfer, the transferred T cell population in the spleen had contracted and remained constant until at least day 7 (Fig. 8,C). This observation correlated with results from a previous report showing that T cell clonal expansion in response to soluble protein Ags in an adjuvant-free system is transient (38). Secondary Ag administration into recipients prevented this clonal loss and sustained T cell numbers until at least day 7 after cell transfer (data not shown). Preferential T cell migration was unlikely to be the cause of the accumulation of donor T cells in the spleens because the numbers of T cells cotransferred with MZ B cells were higher in both lymph nodes and lung than those cotransferred with FO B cells (Table I). Thus, both primed MZ and FO B cells were able to activate T cells, prime MZ B cells, and induce Ag-specific clonal expansion in vivo more efficiently than FO B cells.
|.||% of Transferred T Cells .||.|
|.||Lymph nodes .||Lung .|
|Naive T cells||0.48 ± 0.01a||0.08 ± 0.02b|
|MZ-T cotransfer||1.2 ± 0.21ac||0.24 ± 0.09bd|
|FO-T cotransfer||0.57 ± 0.03ac||0.15 ± 0.03bd|
|.||% of Transferred T Cells .||.|
|.||Lymph nodes .||Lung .|
|Naive T cells||0.48 ± 0.01a||0.08 ± 0.02b|
|MZ-T cotransfer||1.2 ± 0.21ac||0.24 ± 0.09bd|
|FO-T cotransfer||0.57 ± 0.03ac||0.15 ± 0.03bd|
p < 0.001.
p < 0.05.
p < 0.005.
p > 0.05.
MZ B cells differentiate into plasma cells in vivo
Having shown that primed MZ and FO B cells differentially induce Ag-specific T cell clonal expansion in vivo, we next analyzed the reciprocal activities of these B cell subpopulations after interaction with T cells in vivo. As shown in Fig. 9, A and B, only small numbers of FO and MZ HEL-binding B cells were detected in the spleens of recipients that received MD4-donor B cells. As expected, this population was not detected in nontransferred controls or in those mice that received T cells alone. On day 5 after cell transfer, when transferred B cell numbers peaked, only MZ B cells that were cotransferred with CD4+ T cells had differentiated into plasma cells, as detected by syndecan-1 expression (Fig. 9, C and D). These results show that interactions with CD4+ T cells were able to promote MZ, but not FO, B cell-derived plasma cell generation.
The development of an effective Ab response to TD Ag requires collaboration between T and B cells. Several lines of evidence support the idea that the ability of B cells to induce both naive T cell priming and tolerance depends on their resting vs activated state (28, 29, 30, 39, 40, 41, 42, 43, 44). In agreement with previous results, we found that both resting MZ and FO B cells were unable to induce T cell responses in vivo and in vitro. However, within a few hours of in vivo Ag priming, MZ B cells were far superior to FO B cells in the activation of naive CD4+ T cells in vitro. MD4 B cell populations have been shown to be a little different in the phenotype of their subsets, in that IgM on FO B cells tended to persist at higher levels than on FO B in normal mice. However, we sorted FO and MZ B cell subsets based on their expression of CD23 and CD21. In this respect the MZ B cells are clearly distinguishable from their FO counterparts. Despite this anomaly in MD4 transgenic mice MZ and FO B cells showed the same differences in Ag-presenting capacities in MD4 mice as well as in a variety of other normal strains (not shown). It is unlikely that the high IgM level of FO B cells contributes to the low ability of FO B cells to present Ag.
Similar to previous studies, higher levels of B7.2 are expressed on the cell surface of MZ compared with FO B cells after Ag stimulation (28). Blocking of B7-CD28 interactions inhibited CD4 T cell proliferation, suggesting that the higher expression of B7 molecules on activated MZ B cells than on FO B cells was responsible. The outcome of blocking CD28-B7 interactions is similar to the failure of CD28−/− T cells to expand upon immunization (45). However, the CD28 pathway has been shown to be required for initiation of T cell expansion in response to Ag (45, 46, 47, 48, 49, 50). Therefore, involvement of other pathways downstream from TCR and CD28 stimulatory pathways such as CD40-CD40 ligand and participation of other B7 superfamily-ligand interactions cannot be excluded in the CD4+ T cell activation observed in our study.
The ability of MZ B cells to present Ag to CD4+ T cells was not the result of their proximity to blood-born Ag entering via the spleen marginal sinuses. Thus, the different abilities of MZ and FO B cells to activate T cells are probably due to intrinsic differences in the signaling strength through BCR between MZ and FO B cells. The recruitment of Syk tyrosine kinase by the Ig-α subunit of the BCR has been shown to be critical for the MHCII-restricted Ag presentation by B cells (51, 52). Furthermore, signals through IgM have been shown to mediate phosphorylation of Syk in MZ cells to a greater degree than in FO B cells (53). Therefore, phosphorylation of Syk upon BCR engagement may result in enhanced Ag-presenting capabilities of MZ B cells.
In relB−/− mice, deficient in myeloid and functional lymphoid DCs, CD4+ T cells can be primed in response to soluble protein Ag, suggesting that B cells play a role in T cell priming in the absence of DC functions (54). In mice lacking B cells, primary CD4+ or CD8+ T cell priming can be generated (55). In a direct comparison of the Ag-presenting capabilities of in vivo-primed MZ B cells and DCs, we showed that MZ B cells were ∼2- to 4-fold less efficient at inducing T cell proliferation and IL-2 production than were CD11c+CD8α− DCs, which have been shown to be the cells most efficient at presenting soluble protein Ag to CD4+ T cells (34). T cells primed in the absence of B cells failed to provide help to B cells to produce Ag-specific IgG (55, 56) and in a comparison of the Ag-presenting capabilities of DCs and B cells, it was shown that Ag-presenting efficiency is influenced by numerous factors, such as cell numbers, Ag dose and form, and intact protein vs peptides (27, 42, 57, 58, 59). Besides higher expression of MHC class II, adhesion, and co-stimulatory molecules (60, 61), DCs process and present a limited immunodominant peptide, whereas B cells present a heterogeneous set of peptide-MHC complexes, supporting a more diverse T cell response (62). Because of the monoclonal nature of the T cells used this study, the possibility that MZ B cells present subdominant peptides to T cells was not examined. In addition, MZ B cells at high cell concentrations induced more IL-2 from Ag-specific T cells than did CD11c+CD8 α+ DCs, which have been hypothesized to be involved in tolerance induction (61, 63, 64). These results indicate that effective T cell responses to protein Ags may depend on cells residing in the splenic MZ, including Ag-specific MZ B cells and CD11c+CD8α− DCs, which have been shown to migrate toward T cell areas upon Ag encounter (34, 35, 65, 66). B cells have also been shown to transfer Ag to CD8α+ DCs in vivo (67). Therefore, these two cell types may act in concert to modulate Ag-specific T cell responses to further engage in B cell responses.
The precise role of B cells in promoting cytokine production toward a Th1 or Th2 profile is not clear. Numerous studies both in vitro and in vivo suggest that B cells are involved in Th2 responses (68, 69, 70). However, B cells activated with oligodeoxynucleotide (CpG)-conjugated Ag promote Th1 differentiation from unprimed T cells (71). Effector B cell subsets that produce distinct sets of cytokines, named Be1 and Be2, have been demonstrated to facilitate Th1 and Th2 polarization, respectively (72). Although there were no qualitative differences in our studies, both MZ and FO B cells were able to induce more Th1 cytokine (IFN-γ) than Th2 cytokines (IL-4, IL-5, and IL-10) in vitro when T cells were primed in the absence of exogenous cytokines. However, MZ B cells were much more effective than FO B cells in stimulating Ag-specific T cells to secrete IFN-γ, IL-4, and IL-10 upon primary and secondary stimulations. CD4+ T cells stimulated with FO B cells were capable of secreting high amounts of IFN-γ only after secondary stimulation.
Th1/Th2 polarization is regulated by Ag dose and form (73, 74), costimulatory molecules (75, 76), cytokine environment (75, 77), and type of APCs (76, 78, 79, 80, 81). The data presented in this study suggest that Ag uptake and presentation through surface BCR together with sufficient costimulatory signals facilitated the production of Th1 cytokines from naive T cells with high efficiency when MZ B cells were used as APCs. The cytokine polarization observed in this study may be due in part to the strength of TCR stimulation and the presence of costimulatory molecules provided by MZ B cells in the primary stimulation. IL-12, mostly secreted from DCs, which has been shown to be the major cytokine involved in Th1 polarization (81) did not appear to be involved as DCs were not included, and no IL-12 was detected in the cultures. Therefore, IFN-γ produced from T-MZ B cell coculture via activation of T-bet (T box transcription factor) might positively augment the expression of T-bet, thus further biasing T cell responses toward Th1 commitment (78, 82, 83, 84). MZ B cells also stimulate T cells to secrete IL-10, which might function as a negative regulator to suppress the proliferation of T cells and to prevent overexpansion.
By adoptive transfer, we confirmed in vivo the potent ability of MZ B cells to activate T cells in vivo. In the absence of adjuvant or further Ag administration, Ag-specific T cell clonal expansion was most likely initiated by primed B cells, because we eliminated the direct involvement of DCs, which are thought to provide the initial activation of naive CD4+ T cells in vivo. However, the transfer of Ag to CD8α+ DCs in vivo by B cells could be used indirectly to activate naive CD4+ T cells (67). Through direct or indirect involvement in Ag presentation, MZ B cells are far more efficient than FO B cells at promoting Ag-specific clonal expansion in vivo. These differences were not due to the migration of activated Ag-specific T cells to other organs, because the numbers of Ag-specific T cells were correlated in every organ measured. The higher number of Ag-specific T cells detected in the spleen was unlikely to be due to T cell survival alone, because T cells that were cotransferred with MZ B cells progressed through more cell divisions than those that were cotransferred with FO B cells.
It has been shown that T cell responses to a model soluble protein Ag, OVA, give rise to a transient clonal expansion, followed by deletion of Ag-specific T cells in vivo. The residual Ag-specific T cell population is long lasting, but is defective in the ability to proliferate and secrete cytokines upon secondary challenge (85, 86). This anergic state of Ag-experienced T cells was shown to require Ag persistence in the immune host environment (86, 87) In our study the secondary administration of HEL was able to induce T cell response at the same magnitude observed in the primary HEL response. This nonanergic state of the secondary T cell response may result from the existence of HEL on B cells in the absence of exogenous HEL, thus preserving the naive environment of the recipients.
Although the number of donor B cells detected in the recipients was small by day 5 after cell transfer, Ag-specific T cells had supported plasma cell generation from residual MZ B cells. In TI immune responses against particulate Ag, MZ and B1 B cells have been shown to produce the early wave of IgM plasma cells within 3–4 days after Ag encounter (9). The results in this study clearly demonstrate that not only do MZ B cells participate in TI immune responses, but they contribute to the generation of plasma cells in the primary TD immune responses.
Infectious agents are complex and composed of both TD and TI epitopes. Taking into account that MZ B cell subpopulations are likely to be heterogeneous, multireactive, and react with a wide variety of bacterial associated and self-Ags (88, 89, 90), it is conceivable that MZ B cells rapidly activate T cells and/or differentiate into short-lived plasma cells in response to both TI and TD Ags depending on their BCR specificity. In contrast, accumulation of abnormal self-reactive MZ B cell clones may exacerbate autoimmune diseases, as MZ B cells are highly reactive and appear to have a low threshold for activation.
We thank Lisa Jia for invaluable technical help, and Dr. G. Larry Gartland for FACS sorting, Woong-Jai Won and To-Ha Thai for critical reading of this manuscript, and Ann Brookshire for manuscript preparation.
This work was supported by National Institutes of Health Grants CA13148 and AI14782.
Abbreviations used in this paper: BCR, B cell Ag receptor; DC, dendritic cell; FO, follicular; HEL, hen egg lysozyme; MZ, marginal zone; TI, T independent; AMCA, 7-aminomethylcoumarin; CMFDA, 5-chloromethylfluorescein diacetate.