CD40, but Not CD154, Expression on B Cells Is Necessary for Optimal Primary B Cell Responses ¹

The production of Ab in response to infection or immunization is regulated by a complex array of cellular and molecular interactions that take place between Ag and cells of the innate and adaptive immune system. One of the most important interactions for humoral immune responses is the engagement of CD40 by its ligand, CD154 (1, 2). CD40 is a member of the TNF receptor family of proteins (3) that is expressed on APCs such as B cells, dendritic cells (DCs),³ and monocytes (4, 5, 6, 7), as well as some endothelial cells and fibroblasts (4, 8, 9, 10). Signaling through CD40 is essential for many aspects of the immune response because engagement of CD40 triggers the activation of DCs and other APCs (4, 6, 11, 12). CD40-activated APCs up-regulate expression of B7 and MHC class II molecules (2, 13, 14, 15) and produce inflammatory cytokines, such as IL-12 (16, 17, 18). In turn, these molecules facilitate T cell activation and direct T cell differentiation (19). CD40 is also an important costimulatory molecule for B cells. CD40 signaling facilitates T-dependent B cell activation (20, 21, 22, 23, 24), T-dependent B cell proliferation (25, 26, 27, 28), germinal center formation (29, 30, 31, 32), and Ig isotype switching (30, 33, 34). Furthermore, depending on the duration and level of engagement of CD40 on B cells, CD40 signaling can either promote or prevent B cell terminal differentiation and Ab production (25, 26, 35, 36, 37, 38). The importance of CD40 signaling in humoral immune responses is most dramatically illustrated in CD40^−/− and CD154^−/− mice, which are unable to form germinal centers, generate memory B cells, or produce high-affinity, isotype-switched Ab to T-dependent Ags (30, 39).

Although CD40/CD154 interactions are clearly required for humoral immune responses in vivo, it is unclear whether the engagement of CD154 on Th cells with CD40 on B cells is the primary interaction that controls humoral immune responses in vivo. A recent study suggests that the lack of isotype switching and Ab secretion after immunization in CD40^−/− mice is a B cell-intrinsic defect (40). However, it is also possible that defects in the humoral immune response of CD40^−/− mice may be the result of an inability of CD40-deficient non-B cell APCs to prime T cells (41, 42, 43). In support of this hypothesis, anti-CD28 treatment was demonstrated to restore germinal center formation and Ab production in CD154^−/− mice (44). Furthermore, a recent study showed that CD40 expression on non-B cell APCs is necessary for CD4⁺ T cells to acquire a follicular homing phenotype (42), suggesting that insufficient or inappropriate T cell priming may be responsible for the poor humoral immune responses observed in CD40^−/− and CD154^−/− mice. Finally, there are several reports demonstrating that CD40 can be functionally expressed on both CD4 and CD8 T cells (45, 46, 47, 48). Thus, it is not entirely clear whether CD40 expression on B cells, other APCs, or even T cells is required for optimal humoral immune responses.

Similar uncertainty surrounds the role of CD154 in the immune response. Initial studies suggested that CD154 is inducibly expressed on activated T cells, particularly on CD4 Th cells (49, 50, 51). However, more recent studies have suggested that the expression of CD154 is not limited to T cells. In fact, CD154 is expressed by mast cells and basophils (52), as well as NK cells (53), DCs (54), and even B cells (55, 56, 57, 58, 59). The expression of CD154 by non-T cells implies that non-T cells have the ability to facilitate B cell responses in vivo (52, 56). In particular, the expression of CD154 by B cells suggests that interactions between B cells may facilitate B cell activation, isotype switching, proliferation, and Ab production (55, 56, 57). For example, expression of CD154 on the plasma membrane and in the cytoplasm of activated murine B cells has been observed and has been shown to promote B cell proliferation in vitro (57). In addition, the expression of CD154 on activated human B cells as well as human germinal center B cell subsets appears to promote homotypic B cell aggregation, proliferation, and Ab production in vitro (55, 56), suggesting that interactions between CD154-expressing B cells in the germinal center may promote germinal center B cell survival and expansion. Thus, it is possible that CD154 expression by B cells could be necessary for some aspects of the humoral immune response.

To clarify whether CD40 expression by B cells is necessary for optimal humoral immune responses and to determine whether CD154 expression by B cells plays any role in humoral immunity, we generated mixed bone marrow (BM) chimeras in which the B cells were unable to express either CD40 or CD154, while the majority of cells in all other lineages were competent to express CD40 or CD154. Upon immunization with a protein Ag or infection with influenza virus, we found that CD40 expression by B cells was required for B cell expansion, germinal center formation, isotype switching, and the long-term production of Ab, even when the other APCs were able to express CD40. Furthermore, although we found that T cell expansion and homing were normal in mice with CD40-sufficient APCs, the CD40-deficient Ag-specific B cells were still unable to expand and differentiate. In contrast, CD154 expression on B cells was not required for B cell expansion, germinal center formation, isotype switching, or the long-term production of Ab. In fact, CD154 expression exclusively on Ag-specific CD4 T cells was sufficient for optimal humoral responses. These results suggest that the interaction of CD154 on activated CD4 T cells with CD40 on B cells is the primary pathway to achieve B cell activation and differentiation, and that CD154 expression on B cells does not noticeably facilitate B cell activation and differentiation.

C57BL/6 and CD154^−/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME). CD40^−/−, OTII TCR transgenic mice (60) that express enhanced green fluorescent protein (EGFP-OTII mice) and μMT mice were bred and maintained in the Trudeau Institute Animal Breeding Facility. All mice were on the C57BL/6 genetic background. Recipient μMT mice were irradiated with 950 rad from a ¹³⁷Cs source at 93 rad/min and reconstituted with a total of 1 × 10⁷ whole BM cells. Mice were allowed to reconstitute for at least 6 wk before infection or immunization. Mice were immunized i.p. with 100 μg (4-hydroxy-3-nitrophenyl)-acetyl₍₁₅₎-chicken OVA (NP-OVA) adsorbed to alum or intranasally infected with 1000 egg infectious units (EIU) of influenza A/PR8/34 (PR8) in 100 μl PBS. All experimental procedures involving animals were approved by the Trudeau Institute Institutional Animal Care and Use Committee and performed according to the guidelines set by the National Research Council.

CD4 T cells were purified from EGFP-OTII TCR transgenic mice by positive selection. Briefly, single cell suspensions were obtained from the spleens and lymph nodes of 4- to 5-wk-old EGFP-OTII mice; erythrocytes were lysed with 150 mM NH₄Cl, 1 mM KHCO₃, 0.1 mM EDTA pH 7.3; and cells were incubated with FcR blocking Ab, 2.4G2, at 1 μg/10⁷ cells for 15 min on ice. Anti-CD4 magnetic beads (Miltenyi Biotec, Auburn, CA) were added at 25 μl/1 × 10⁸ total cells in 100 μl final volume and incubated on ice for an additional 15 min on ice. Cells were washed, filtered, and passed over a type LS⁺ magnetic column (Miltenyi Biotec). After washing, the magnetically bound cells were eluted from the column with the supplied plunger. This procedure routinely resulted in 95–98% pure CD4 T cells, of which ∼90% expressed the TCR transgene. Purified EGFP-OTII cells (2 × 10⁶/mouse) were adoptively transferred to recipient chimeric mice the day before immunization.

Total B cells were identified using anti-CD19 (1D3) from BD PharMingen (San Diego, CA), and NP-specific B cells were identified by binding to NP₍₃₀₎-allophycocyanin. Phenotypic analysis of B cells was performed using fluorochrome-conjugated peanut agglutinin (PNA) from Sigma-Aldrich (St. Louis, MO) and anti-IgM (AF6-78), anti-IgG1 (A85-1), anti-IgG2 (R2-40), anti-FAS (Jo2), and anti-syndecan (CD138) (281-2) from BD Phar-Mingen. CD40 expression on B cells or DCs from chimeric mice was confirmed using anti-CD40 (1C10), anti-B220 (RA3-6B2), and anti-CD11c (HL3) from BD PharMingen. EGFP-OTII T cells were identified by intrinsic green fluorescence and characterized using anti-CD4 (GK1.5) and anti-CD62L (MEL-14) from BD PharMingen. Flow cytometry was performed on a dual-laser FACSCalibur from BD Biosciences (San Jose, CA).

Plates were coated with either NP₍₁₆₎-BSA or NP₍₂₎-BSA at 1 μg/ml, with proteins from disrupted influenza virions at 1 μg/ml or with anti-IgE (R35-72 from BD PharMingen) at 1 μg/ml and blocked with 10 mg/ml BSA. Serum samples were initially diluted 100-fold in PBS with 10 mg/ml BSA and 0.1% Tween 20 and then serially diluted seven additional times in 3-fold steps in the same buffer and applied to the coated plates. Bound Ig was detected with HRP-conjugated goat anti-mouse IgM, goat anti-mouse IgG, or goat anti-mouse IgA (all from Southern Biotechnology Associates, Birmingham, AL). NP-specific IgE was detected using biotinylated NP₍₅₎-BSA, followed by HRP-conjugated streptavidin (Southern Biotechnology Associates). Ab titers are defined as the dilution required to reduce a positive signal to 3-fold above background.

Spleens of NP-OVA-immunized chimeric mice were fixed in 4% Formalin, 10% sucrose, and 7% picric acid for 30 min and washed in PBS before being embedded in OCT (Tissue-Tec) and frozen over liquid N₂. The fixation is required for the preservation of EGFP in the EGFP-OTII T cells (61). Thin sections (7 μm) were cut with a cryomicrotome, fixed with acetone, and probed with biotinylated PNA (Vector, Burlingame, CA), goat anti-mouse λ (R2B-46), biotinylated anti-CD3ε (145-2C11), or biotinylated anti-B220 (RA3-6B2), all from BD PharMingen. EGFP fluorescence was visualized directly, while PNA and B220 were visualized using streptavidin Alexa 350 (Molecular Probes, Eugene, OR), and λ and CD3 were visualized with streptavidin 594 (Molecular Probes). Images were captured using a Zeiss (Oberkochen, Germany) Axioplan 2 microscope at ×10 original magnification utilizing a Zeiss Axiocam digital camera.

To determine whether CD40 or CD154 expression on B cells was required for humoral immune responsiveness, we generated BM chimeric mice, in which the CD40 locus or CD154 locus was selectively disrupted in the B cell compartment. As shown in Fig. 1,A, μMT mice were lethally irradiated and reconstituted with a mix of 75% μMT BM and 25% C57BL/6 BM. Because the BM of both donors is CD40^+/+, all cells derived from the donor BM in the chimeric recipients will be CD40^+/+. These mice are referred to as wild-type (WT) chimeras. As shown in the dot plots of Fig. 1,A, the B220⁺ B cells of WT chimeras all express CD40. Likewise, the majority of CD11c⁺ DCs also express CD40. The lack of CD40 expression on some DCs is most likely due to the immature phenotype of a portion of the CD11c⁺ DCs, which have not yet up-regulated CD40 expression (Fig. 1,A) (62). In the second group of chimeras (Fig. 1,B), lethally irradiated μMT mice were reconstituted with 100% CD40^−/− BM. Because all the hemopoietic cells in these chimeras will be derived from CD40^−/− BM, these mice are referred to as CD40-knockout (KO) chimeras. As shown in the dot plots of Fig. 1,B, none of the splenic B cells or DCs express CD40 in these mice. Although there are a small number of cells that appear to coexpress B220 and CD40 in these animals, this is due to the background staining with the anti-CD40 Ab and is observed even in fully CD40-deficient nonchimeric mice (data not shown). We also reconstituted lethally irradiated μMT mice with 75% μMT BM and 25% CD40^−/− BM (Fig. 1,C). Because the μMT BM is genetically unable to produce B cells, all the B cells in the recipient mice are derived from the CD40^−/− donor BM. In contrast, 75% of all other hemopoietic cells in these mice are derived from the μMT donor BM and will be CD40^+/+. Because the B cells are selectively CD40 deficient in these chimeras, they are referred to as CD40-B cell KO (BKO) chimeras. The phenotype of these chimeras is shown in the dot plots of Fig. 1,C, in which all of the B220⁺ B cells are CD40 deficient, while the majority of the CD11c⁺ DCs express CD40. Although the ratio of CD40^+/+ DCs to CD40^−/− DCs should theoretically be 75:25, a portion of the DCs is not fully mature (62) and is not yet competent to express CD40; thus, the predicted ratio is slightly reduced (Fig. 1,C). In the final group, we reconstituted lethally irradiated μMT mice with a mix of 75% μMT BM and 25% CD154^−/− BM (Fig. 1,D). Because the μMT BM does not produce B cells, all the B cells in the chimeric mice are derived from the CD154^−/− BM. In contrast, the majority of all other hemopoietic cells are derived from the CD154^+/+ μMT donor. These mice are referred to as CD154-BKO chimeras. Because neither resting B cells nor resting T cells express CD154 (38, 57), we were unable to demonstrate the actual ratio of CD154^−/− cells by flow cytometry. However, both the B cells and DCs express the expected ratios of CD40 in these mice (Fig. 1 D).

To determine how the chimeric mice responded to immunization, it was important to track Ag-specific B cells as well as Ag-specific CD4 T cells. To do so, we adoptively transferred CD4 T cells from EGFP-expressing OTII TCR transgenic mice to the chimeric recipients. Because the transgenic T cells express EGFP as well as an Ag receptor specific for OVA, the CD4 response to immunization with haptenated OVA could be easily tracked by following green cells using flow cytometry or fluorescence microscopy (61). Furthermore, by using NP as the hapten, NP-specific B cells could be visualized by flow cytometry due to the ability of NP-specific B cells to bind NP-haptenated fluorochromes, such as NP-allophycocyanin (63). In the first experiment, we adoptively transferred 2 × 10⁶ EGFP-OTII cells to the chimeric mice and immunized them with NP-OVA. Splenocytes were obtained at the days indicated, and a total of 1 × 10⁶ cells was analyzed for the presence of NP-binding B cells. As shown in Fig. 2,A, a large increase in the frequency of NP-specific B cells was observed on day 7 postimmunization in the WT and CD154-BKO chimeras; however, very few NP-specific B cells were observed in the CD40-BKO chimeras. In addition, although the frequency of NP-specific B cells declined after day 7 in WT and CD154-BKO mice, these cells were maintained at much higher levels than in CD40-BKO mice (Fig. 2,A). These results are expressed quantitatively for the full course of the experiment in Fig. 2 B.

To test whether the chimeric mice responded to NP-OVA in a manner similar to intact (nonchimeric) mice, we performed a similar experiment, in which we transferred 2 × 10⁶ OTII cells to nonchimeric C57BL/6, CD40^−/−, and CD154^−/− mice, immunized them with NP-OVA, and determined the frequency and total numbers of NP-binding B cells. Again, 1 × 10⁶ total splenocytes were analyzed; however, only the NP-binding cells were saved in the data file. As shown in Fig. 2, C and D, the frequency and number of NP-specific B cells in NP-OVA-immunized C57BL/6 and CD154^−/− (nonchimeric) mice that had received EGFP-OTII cells were nearly identical with that of WT and CD154-BKO chimeric mice (Fig. 2, A and B). Similarly, the NP-specific B cell response of CD40^−/− (nonchimeric) mice (Fig. 2, C and D) was nearly identical in kinetics and magnitude with that of CD40-BKO chimeric mice (Fig. 2, A and B). Together, these data demonstrate that the presence of CD40 on B cells is essential for the expansion of Ag-specific B cells after immunization and that the expression of CD40 on other APCs in the CD40-BKO chimeras does not facilitate this expansion. Furthermore, expression of CD154 by B cells is not required to any detectable degree for Ag-specific B cell expansion. In fact, the expression of CD154 exclusively on the transferred OTII cells was sufficient for normal NP-specific B cell expansion.

We next determined the phenotype of the NP-specific B cells responding to immunization in chimeric mice as well as intact control mice. As shown in the dot plots of Fig. 3,A, a large proportion of the NP-specific B cells in WT and CD154-BKO chimeras expressed both FAS and PNA on day 7, consistent with a germinal center phenotype (64, 65). In contrast, exceedingly few NP-specific B cells with a germinal center phenotype were present in CD40-BKO chimeras at this time (Fig. 3,A). Although very few NP-specific germinal center B cells were present in CD40-BKO mice, the numbers of these cells did peak at day 7 postimmunization, similar to that observed in WT or CD154-BKO chimeric mice (Fig. 3,A, graph). These results are very consistent with those obtained using nonchimeric mice (Fig. 3,B). Again, robust NP-specific germinal center responses were observed in C57BL/6 and CD154^−/− mice, while only a few NP-specific germinal center cells were observed in CD40^−/− mice. Thus, CD40 expression by B cells, but not CD154 expression by B cells, is required for germinal center responses. In fact, when CD154 expression was limited exclusively to the transferred OTII T cells, completely normal germinal center responses were observed (Fig. 3 B).

We also examined the generation of NP-specific IgM- and IgG-expressing plasma cells. As shown in Fig. 3,C, both WT and CD154-BKO chimeras generated large numbers of NP-specific IgM-producing plasma cells by day 7. In contrast, the CD40-BKO chimeras generated 10- to 100-fold fewer NP-specific IgM-producing plasma cells than did WT and CD154-BKO chimeras. These results were mirrored by the production of large numbers of NP-specific IgM-producing plasma cells in C57BL/6 and CD154^−/− nonchimeric mice, but not in CD40^−/− mice (Fig. 3 D).

Similar to the production of NP-specific IgM-producing plasma cells, the production of NP-specific IgG1-secreting plasma cells is robust in WT and CD154-BKO chimeras and is nearly 1000-fold reduced in CD40-BKO chimeras (Fig. 3,E). Again, the production of NP-specific IgG-producing plasma cells in C57BL/6 mice and CD154^−/− mice (Fig. 3,F) is nearly identical with those in WT and CD154-BKO chimeras (Fig. 3,E), while the production of NP-specific IgG-producing plasma cells in CD40^−/− nonchimeric mice (Fig. 3,F) is very reduced, similar to that in CD40-BKO chimeras (Fig. 3 E). Together, these results demonstrate that CD40 expression on B cells is required for robust Ag-specific B cell expansion and differentiation, even when other APCs are CD40 sufficient, as seen in the immunized CD40-BKO chimeric mice. In contrast, the expression of CD154 exclusively on Ag-specific CD4 T cells is sufficient to mediate normal B cell expansion and differentiation.

We next determined whether the expansion and differentiation of NP-specific B cells correlated with the production of NP-specific Ab. As shown in Fig. 4,A, all three groups of chimeric mice produced detectable levels of NP-specific IgM, although NP-specific IgM titers were higher in the WT and CD154-BKO chimeras compared with the CD40-BKO chimeras. These results correlated well with what we observed using C57BL/6 and CD154^−/− nonchimeric mice, in which NP-specific IgM Ab titers were similar to those observed in WT and CD154-BKO chimeras, while NP-specific IgM levels in CD40^−/− nonchimeric mice were similar to those in CD40-BKO chimeric mice (Fig. 4,B). In contrast to our observations with NP-specific IgM responses, NP-specific IgG responses appear to be very dependent on CD40 signaling. As shown in Fig. 4,C, we observed robust and sustained production of NP-specific IgG in WT and CD154-BKO chimeras. Furthermore, we observed nearly identical production of NP-specific IgG in nonchimeric C57BL/6 mice and CD154^−/− mice (Fig. 4,D), demonstrating that CD154 expression on transferred OTII cells is sufficient for robust responses and that CD154 expression on B cells or other non-T cells is not important for IgG Ab responses. However, IgG responses were severely curtailed in CD40-BKO chimeras (Fig. 4,C) and were also very reduced in CD40^−/− nonchimeric mice (Fig. 4,D). Interestingly, the CD40^−/− mice, but not the CD40-BKO mice, produced a transient wave of IgG that peaked at day 10 postimmunization and rapidly declined (Fig. 4, C and D). These results demonstrate that CD40 expression on B cells is necessary for robust IgG responses and that even the presence of CD40-expressing APCs in the CD40-BKO mice was insufficient to mediate an IgG response.

IgE production is also thought to be dependent on CD40 signaling (66, 67, 68). Therefore, in a separate experiment, we transferred 2 × 10⁶ purified OTII T cells to chimeric mice and immunized them with NP-OVA adsorbed to alum to determine whether isotype switching to IgE requires CD40 or CD154 expression on B cells or other APCs. As shown in Fig. 4,E, high titers of NP-specific IgM were produced in WT and CD154-BKO mice within 10 days after immunization. In contrast, only modest increases in NP-specific IgM were observed in CD40-KO or CD40-BKO chimeric mice after immunization. These results confirm what we observed in Fig. 4,A. Similarly, high titers of NP-specific IgE were observed in WT and CD154-BKO mice within 10 days after immunization (Fig. 4 F). In contrast, peak titers of NP-specific IgE were reduced at least 100-fold in CD40-KO or CD40-BKO chimeric mice after immunization. Thus, isotype switching to IgE is highly dependent on CD40 expression on B cells, but is independent of CD154 expression on B cells.

Affinity maturation is also an important component of humoral immunity and is thought to occur in germinal centers through the positive selection of B cells with high affinity Ag receptors (69, 70). To determine the extent of affinity maturation in each of the chimeras after immunization, we assayed the titers of total IgG by ELISA on plates coated with NP₍₁₆₎-BSA and the titers of high-affinity IgG on plates coated with NP₍₂₎-BSA. As shown in Fig. 4,G, the titers of total NP-specific IgG peaked rapidly and were sustained for up to 40 days in WT and CD154-BKO chimeras. In contrast, total NP-specific IgG responses were greatly reduced in CD40-BKO chimeras and CD40-KO chimeras (Fig. 4,G). As we observed above (Fig. 4, C and D), the CD40-KO mice, but not the CD40-BKO mice, produced a transient peak of NP-specific IgG around day 10. Interestingly, this transient peak of NP-specific IgG is entirely composed of low-affinity IgG, as it is undetectable using NP₍₂₎-BSA-coated plates (Fig. 4,H). In contrast, the high titers of IgG produced in WT and CD154-BKO chimeric mice are composed primarily of high-affinity IgG, particularly after day 10 (Fig. 4 H). Thus, CD40 expression, but not CD154 expression, on B cells is required for affinity maturation, regardless of whether other APCs express CD40.

Based on published reports, we expected that the presence of CD40-expressing APCs in CD40-BKO chimeras would facilitate T cell priming (39, 71) and promote follicular homing (42). Therefore, we next examined the expansion and phenotype of the EGFP-OTII cells responding to OVA in the chimeric mice. As shown in Fig. 5,A, naive EGFP-OTII cells were observed at similar frequencies in all chimeras on day 0 before immunization. Furthermore, the majority of these cells expressed high levels of CD62L, consistent with their resting phenotype (Fig. 5,A). The frequency and total number of these cells increased by day 3 postimmunization in all groups of chimeras (Fig. 5, A and B), and many of these cells expressed reduced levels of CD62L (Fig. 5,A), suggesting that the T cells were also activated. However, the transferred OTII cells expanded the least in the NP-OVA-immunized CD40-KO chimeras (Fig. 5, A and B) and fewer of the OTII cells had reduced CD62L expression in these mice (Fig. 5, A and C). These trends are consistent throughout the experiment, and were also observed in nonchimeric C57BL/6, CD40^−/−, and CD154^−/− mice (data not shown). These data demonstrate that even though CD40 signaling on non-B cell APCs facilitates CD4 T cell expansion in CD40-BKO mice, Ag-specific CD4 T cells were clearly capable of expanding and differentiating in all groups of chimeras.

To test whether the transferred T cells were capable of homing to B cell follicles, spleens of immunized chimeric mice were harvested on day 7 postimmunization and sections were probed with Abs to B220 and CD3 to determine the placement of EGFP-expressing OTII cells. As shown in Fig. 6,A, the majority of transferred T cells (green) were found in the T cell areas (red) in the spleens of WT chimeric mice. However, some of the transferred OTII T cells (green) as well as endogenous T cells (red) could also be found in the B cell follicles (blue) of WT chimeras. Similar placement of T cells was observed in CD40-BKO and CD154-BKO chimeras (Fig. 6, C and D). In contrast, relatively few OTII cells were observed in the B cell follicles of CD40-KO mice (Fig. 6,B), even though these cells could be easily observed in the T cell areas and at the border between the B and T cell areas in these mice (Fig. 6,B). Thus, as described previously (42), the presence of CD40 on non-B cell APCs helps T cells efficiently acquire a follicular homing phenotype. We also probed sections (serial to those in Fig. 6, A–D) with PNA (blue) and anti-λ L chain (red) to confirm the presence of NP-specific germinal centers (Fig. 6, E–H). PNA⁺ germinal centers containing λ⁺ B cells could be easily observed in WT chimeras (Fig. 6,E, circled area) as well as CD154-BKO chimeras (Fig. 6,H, circled area). However, we did not find any germinal centers in either CD40-KO or CD40-BKO chimeras (Fig. 6, F and G). Thus, despite the clear presence of Ag-specific T cells in the B cell follicles of CD40-BKO mice (Fig. 6,C), germinal centers were not formed (Fig. 6 G). Therefore, the lack of B cell responses in CD40^−/− mice is most likely due to a B cell intrinsic defect rather than defects in T cell expansion or homing.

The previous experiments demonstrated that the humoral immune response to an inert Ag was clearly dependent on CD40 expression by B cells, but was independent of CD154 expression by B cells. However, it is clear that replicating pathogens, such as viruses, stimulate the immune system in very different ways (18, 62, 72) and may be able to bypass the requirement for CD40 on B cells in mice that express CD40 on other APCs (CD40-BKO mice). Therefore, we next determined the ability of the various chimeric mice to generate a humoral immune response to influenza. Cohorts of chimeric mice were infected with influenza, and splenocytes from infected animals were analyzed at the times indicated for the presence of germinal center B cells. As shown in Fig. 7 A, CD19⁺FAS⁺PNA⁺ germinal center B cells were easily detectable at day 14 postinfection in WT and CD154-BKO chimeras, but were very infrequent in CD40-KO or CD40-BKO chimeras. This was consistent throughout the course of infection, and was also observed in cells from the draining lymph node (data not shown). Thus, the expression of CD40, but not CD154, on B cells is important for the expansion and maintenance of the germinal center compartment in response to influenza, just as it is in the response to NP-OVA.

We also examined the ability of the chimeric mice to make plasma cells secreting IgM, IgG2, and IgA in response to influenza. As shown in Fig. 7,B, CD19⁺CD138⁺ IgM-expressing plasma cells could be observed in all groups of mice at day 14 postinfection; however, the frequency of IgM plasma cells was somewhat reduced in the CD40-KO and CD40-BKO chimeric mice relative to that in WT and CD154-BKO chimeric mice. This difference was maintained throughout the course of the experiment, and the numbers of IgM plasma cells in WT and CD154-BKO chimeras were consistently 2- to 5-fold higher than those in CD40-KO and CD40-BKO chimeras (data not shown). In contrast, there was a much bigger difference in the frequency and total numbers of IgG2-producing plasma cells. As shown in Fig. 7,C, the frequency of IgG2-expressing plasma cells is similar in WT and CD154-BKO chimeric mice, but these cells were essentially undetectable in CD40-KO and CD40-BKO mice at day 14 postinfection. The difference in the total number of IgG2-expressing plasma cells in CD40-BKO mice compared with CD154-BKO mice was consistently ∼10-fold higher for the CD154-BKO mice throughout the course of the experiment (data not shown). Similar results were observed for plasma cells producing IgG1 and IgG3; however, the total numbers of these cells were much lower than the number of cells producing IgG2 (data not shown). Finally, although we observed that IgA-secreting plasma cells in WT and CD154-KO chimeric mice were ∼10-fold higher in frequency and total number on day 14 postinfection relative to those in CD40-KO and CD40-BKO mice (Fig. 7 D), this difference was not consistent at all time points. Thus, CD40 expression on B cells is most important for germinal center formation and production of IgG, but is less important for the production of IgM and IgA.

Although there were clear differences between the abilities of each of the chimeras to make germinal center B cells and plasma cells in response to influenza, these assays measured total germinal center B cells and total plasma cells rather than those specific for influenza. Therefore, we next assayed influenza-specific serum Ab levels by ELISA. As shown in Fig. 8, all groups of chimeric mice produced influenza-specific IgM with similar kinetics and similar titers, although the CD40-KO chimeras produced slightly higher titers of IgM at day 9. This is consistent with the ability of all groups of chimeric mice to generate IgM plasma cells (Fig. 7,A) and is consistent with normal (or elevated) IgM levels in CD40^−/− mice and CD154^−/− humans (30, 73, 74). In contrast, although WT and CD154-BKO chimeras rapidly generated high titers of influenza-specific IgG and maintained these titers for up to 100 days postinfection, the CD40-KO and CD40-BKO mice generated much lower levels of influenza-specific IgG, and even these reduced titers were not maintained for extended periods (Fig. 8). Interestingly, although the CD40-KO mice generated a transient spike in IgG titers, similar to that observed in CD40^−/− mice upon viral infection (41, 75), this transient increase in IgG production was not observed in the CD40-BKO mice (Fig. 8). This difference between influenza-specific IgG responses in CD40-KO and CD40-BKO mice is very similar to what we observed for NP-specific IgG production in these same groups of chimeric animals (Fig. 4, C and D). Finally, we observed that all groups of chimeric mice produce influenza-specific IgA that peaked on day 9. Surprisingly, the CD40-BKO chimeras made higher titers of influenza-specific IgA than either WT or CD40-KO chimeras. Thus, even though the CD40-BKO chimeras were less able to produce influenza-specific IgG than CD40-KO chimeras, they were actually slightly better at producing influenza-specific IgA at the early time point of infection. However, the flu-specific IgA titers were not maintained long-term in the CD40-BKO chimeras (or CD40-KO chimeras) compared with the WT or CD145-BKO chimeras, indicating that CD40 expression on B cells is necessary for sustained influenza-specific IgA responses.

Interactions between CD40 and its ligand, CD154, are important for B cell isotype switching, germinal center formation, memory B cell generation, and the long-term production of Ab (30). However, it is unclear whether defects in humoral immunity in CD40^−/− and CD154^−/− mice are intrinsic to B cells, or result from the lack of CD40 or CD154 expression on other cell types. In particular, the inability of non-B cell APCs to be activated through CD40 signaling may lead to impaired CD4 T cell priming and result in reduced B cell responses (4, 12, 76, 77, 78). We directly tested this possibility by generating chimeric mice that lack CD40 expression exclusively on B cells. We found that B cell expansion, isotype switching to IgG, and germinal center formation remain severely impaired in the absence of CD40 expression on B cells, regardless of whether CD40 is expressed on non-B cell APCs or even T cells. Furthermore, CD40 expression is required on B cells even when chimeric mice were challenged with an intrinsically inflammatory Ag, such as influenza virus. Thus, even in the presence of stimuli that activate DCs independently of CD40 (72), the humoral immune response is still severely impaired in the absence of CD40 signaling on B cells.

Based on other reports, we expected that Ag-specific T cells transferred to CD40^−/− mice would fail to expand normally after immunization and would be unable to home to B cell follicles (39, 42). In fact, we did observe slightly reduced CD4 T cell expansion in CD40-KO hosts; however, the reduction was not as severe as previously reported (43). Furthermore, we also observed that Ag-specific CD4 T cells were less able to home to B cell follicles when activated in mice that completely lack CD40 expression on BM-derived cells. However, both T cell expansion and follicular homing were returned to normal levels in mice that expressed CD40 on non-B cell APCs, but not on B cells. This is consistent with other reports using similar chimeric mice (42). However, even the restoration of normal T cell expansion and follicular homing was unable to drive Ag-specific B cell expansion, germinal center formation, or isotype switching to IgG in the absence of CD40 signaling directly to B cells. Thus, we think that the lack of B cell responses in CD40^−/− or CD154^−/− mice is primarily the result of poor B cell expansion and differentiation rather than poor T cell priming or defects in T cell homing.

These results are in contrast to those showing that germinal center formation and Ab production are restored upon the administration of anti-CD28 to CD154^−/− mice (44). This study suggested that CD40 was primarily required to induce the expression of B7 on APCs and that T cell activation could be restored by providing a direct signal to CD28 via an anti-CD28 Ab (44). These results implied that CD4 T cell help is the limiting factor in B cell responses. However, our results suggest that CD4 T cell help is not the limiting factor and that B cells directly require a CD40 signal to expand and differentiate. Our data are more consistent with a recent study showing that Ag-specific IgG production in mixed BM chimeras that contain equal numbers of CD40^+/+ and CD40^−/− B cells was derived almost exclusively from the CD40^+/+ B cells (40). Presumably, both types of B cells in the chimeras had access to exactly the same CD4 Th cells; therefore, the inability of the CD40^−/− B cells to respond to immunization must be an intrinsic defect. Although this study did not examine the ability of CD40^−/− B cells to participate in germinal center reactions (40), our results clearly show that CD40^−/− B cells are unable to form germinal centers, switch to IgE, or undergo affinity maturation, despite the restoration of T cell expansion and follicular homing. Again, we conclude that the inability of CD40^−/− mice to make normal humoral immune responses is related to the lack of B cell expansion and differentiation, rather than limiting T cell help.

Given that we expected that the B cell response would be improved to some degree in CD40-BKO mice due to increased CD4 T cell expansion and follicular homing, we were surprised to find that the transient production of Ag-specific IgG normally observed in CD40-KO mice was almost entirely lost in CD40-BKO mice. This effect was observed in response to influenza infection (a type I response) as well as in response to immunization with NP-OVA adsorbed to alum (a type II response). Because the B cells in both of these chimeras lack CD40 expression, the difference in IgG production must be due to the activity of CD40 expressed on non-B cells. We believe that the CD40-bearing cell type that is responsible for the difference in switching to IgG is most likely an APC, such as a DC. CD40-activated DCs may influence B cell isotype switching directly through physical interactions with B cells (79, 80, 81) or indirectly by controlling the activities of T cells. Alternatively, T cells have been shown to express CD40 (45, 46, 47, 48, 82). Thus, CD40^+/+ T cells in CD40-BKO mice may have different activities than CD40^−/− T cells in CD40-KO mice. However, all chimeras received CD40^+/+ OTII T cells before immunization; therefore, it is unlikely that differences in CD40 expression in the T cell compartment are directly responsible for the alterations in IgG production. Instead, the ability of the transferred OTII T cells to promote switching to IgG must be regulated indirectly by the activity of endogenous T cells or by the activity of endogenous APCs. Although the mechanism behind the differential switching to IgG in CD40-KO and CD40-BKO mice is not yet understood, we do know that it is not due to an inability of transferred OTII T cells to home to the follicle, as shown in Fig. 6. This will be an area of future investigation.

Our studies also tested whether the expression of CD154 on B cells facilitates the humoral immune response in any way. Although CD154 was originally described to be expressed exclusively on activated CD4 T cells (49, 50) and was proposed to provide the bulk of B cell help by signaling through CD40 on B cells (23), recent data suggest that CD154 is also expressed on activated human and murine B cells (55, 57), as well as human B cell lymphomas (83). In addition, CD154 expression has been described on subsets of germinal center B cells isolated from human tonsils (55), suggesting that CD154 expression on B cells may play an autocrine role in maintaining the germinal center or facilitating germinal center B cell survival. In agreement with this hypothesis, CD154 expression by B cells was shown to facilitate B cell activation, proliferation, and survival in vitro (55, 56, 57). Although it was clear from experiments using CD154^−/− and CD40^−/− mice and CD40-blocking studies that CD40 signaling is necessary for the generation and maintenance of germinal centers (1, 30, 31, 32), it was not clear which cell types must express CD154 in the germinal center. Our data clearly demonstrate that CD154 expression by B cells is not required for Ag-specific B cell expansion, for the formation of germinal centers, or for terminal differentiation of B cells to Ab-producing plasma cells. Furthermore, even when CD154 expression is limited to Ag-specific CD4 T cells, the induction of germinal centers and robust Ab responses occur normally. Thus, it is clear that CD154 expression on B cells plays little, if any, role in normal humoral immune responses in vivo.

It is possible, however, that CD154 expression on B cells plays a more important role under conditions of chronic inflammation or autoimmunity. For example, several studies have shown that abnormally high expression of CD154 is observed on B cells from autoimmune BXSB mice (58), on B cells from humans with the autoimmune disease, systemic lupus erythematosus (59), and some human B cell lymphomas (83). In addition, studies demonstrating the functional role of CD154 expression on B cells used human germinal center B cells isolated from tonsils, a site of chronic inflammation (55). However, it is not clear whether the expression of CD154 on B cells in these situations is part of a normal physiological response or whether CD154 expression on these types of B cells is aberrant and is part of the pathological process. Although our data do not directly assess whether CD154 is expressed on B cells during humoral responses to NP-OVA or influenza, our data clearly demonstrate that CD154 expression on B cells does not play any functional role in these responses. Thus, the in vivo significance of CD154 expression on B cells remains enigmatic.

Together, these data support the original model of T:B collaboration, in which activated CD4 T cells express CD154 and provide help to B cells through CD40 (84). Although CD40 signaling on non-B cell APCs does appear to facilitate Th expansion and their ability to home to B cell follicles (39, 42), the limiting step in the humoral immune responses in CD40^−/− or CD154^−/− mice is the lack of CD40 signaling to B cells rather than the lack of T cell priming. Furthermore, despite the recently described expression of CD154 on B cells themselves, the ability of B cells to express CD154 does not facilitate humoral immune responses to any appreciable degree. Finally, the expression of CD154 exclusively on CD4 T cells is sufficient for robust humoral immune responses. Thus, the expression of CD40, but not CD154, on B cells is necessary for optimal B cell responses.