Uncoupling CD21 and CD19 of the B-cell coreceptor

Abstract

Complement receptors (CRs) CD21 and CD35 form a coreceptor with CD19 and CD81 on murine B cells that when coligated with the B-cell receptor lowers the threshold of activation by several orders of magnitude. This intrinsic signaling role is thought to explain the impaired humoral immunity of mice bearing deficiency in CRs. However, CRs have additional roles on B cells independent of CD19, such as transport of C3-coated immune complexes and regulation of C4 and C3 convertase. To test whether association of CR with CD19 is necessary for their intrinsic activation-enhancing role, knockin mice expressing mutant receptors, Cr2^Δ/Δgfp, that bind C3 ligands but do not signal through CD19 were constructed. We found that uncoupling of CR and CD19 significantly diminishes survival of germinal center B cells and secondary antibody titers. However, B memory is less impaired relative to mice bearing a complete deficiency in CRs on B cells. These findings confirm the importance of interaction of CR and CD19 for coreceptor activity in humoral immunity but identify a role for CR in B-cell memory independent of CD19.

B cells, like T cells, are regulated in large part by signals from the innate immune system (1–3). For example, engagement of complement (C3d)-coated antigen (Ag) by mature cognate B cells ligates the coreceptor (i.e., CD19, CD21, and CD81) and the B-cell receptor (BCR), enhancing BCR signaling by several orders of magnitude (1, 4).

In mice, complement receptors CD21 and CD35 (CRs) are encoded at the Cr2 locus by splicing of message, and they are coexpressed primarily on B cells and follicular dendritic cells (FDCs) (5, 6). CD35 and CD21 bind similar split products of C3, but in addition CD35 binds C3b and C4b. CD35 also interacts with CD19 to form a B-cell coreceptor (7). Systemic blocking of CR by antibody (8–10) or soluble receptor (11) results in impaired humoral immunity. Likewise, mice deficient in the receptors (Cr2^−/−) bear impaired B-cell immunity to T-independent (12–14) and T-dependent Ags (15–17), infectious bacteria (18), and viruses (19, 20). Studies using chimeric mice expressing CRs on either B cells (16) or FDCs (21, 22) indicate that overall humoral immunity is dependent on the presence of the receptors on both cell types. Thus, intrinsic B-cell signaling by the coreceptor and retention of Ags on FDCs are both important in B-cell immunity.

Recent studies demonstrate that CR expression on B cells is important in the transport of immune complexes (ICs). Bloodborne complexes of complement-coated Ags are bound rapidly by marginal zone (MZ) B cells and are transported into the splenic follicles (12, 13, 23), and Ag is offloaded to FDCs (24). An analogous role for CRs on follicular B cells in peripheral lymph nodes (LNs) was identified. ICs draining via afferent lymph are trapped by subcapsular sinus (SCS) macrophages and “transferred” to noncognate B cells in the underlying follicular region in a CR-dependent manner (25, 26). The complement-coated ICs are then transported to FDCs in a manner similar to that proposed for MZ B cells. Thus, CRs have at least two intrinsic roles on B cells: coreceptor on cognate B cells, and transport of ICs by noncognate B cells. A third, less studied role for CRs on B cells is complement regulatory activity. All mammalian cells express complement regulatory proteins (CRPs), such as Decay Activating Factor (DAF) and membrane cofactor protein (MCP), which act to protect the host from complement activation on the cell surface (27). In the mouse, DAF and Crry are the major CRPs expressed on host nucleated cells (28–30); however, CD35 also has complement regulatory activity. Thus, its deficiency could contribute to a loss of CRP activity that leads to activation of complement and an alteration in the local environment (31).

To determine the functional importance of the interaction between CR and CD19, a gene-targeting approach was used to generate the mutant mouse line Cr2^Δ/Δgfp, where the transmembrane (Tm) and cytoplasmic tail (CyT) of HLA class I α chain were substituted for the analogous region of CR. Based on studies in a B-cell line expressing a similar CR-HLA chimeric receptor, it was predicted that mutant B cells would lack functional coreceptor signaling but retain the ability to bind and transport C3d-coated ICs and maintain CRP activity (32).

Characterization of Cr2^Δ/Δgfp B cells in vitro and in vivo confirmed their ability to bind C3d-coated specific Ag without activation of CD19, demonstrating functional uncoupling of CR and CD19. Comparison of humoral responses to T-dependent Ags in chimeric mice bearing WT FDCs and B cells derived from Cr2^Δ/Δgfp, Cr2^−/−, or WT mice identified an intermediate response by the Cr2^Δ/Δgfp relative to WT and Cr2^−/−, and survival of germinal center (GC) B cells was reduced in both the Cr2^−/− and Cr2^Δ/Δgfp chimeric mice. These findings confirm the importance of interaction of CR and CD19 for coreceptor activity in humoral immunity but identify additional roles for CRs in maintenance of B-cell memory.

Results

Generation of Cr2^Δ/Δgfp Mice.

To uncouple CR and CD19, a strategy developed by Matsumoto et al. (32) was used to substitute the Tm and CyT regions of HLA class I α chain for the analogous region of CD21/CD35 (SI Text and Fig. S1 A and B). The mutant receptor, Cr2^Δ/Δgfp, developed normally, and CR2 expression, detected by GFP fluorescence, was restricted to FDCs and B cells (Fig. S1C). Moreover, Cr2^Δ/Δgfp B cells maintained the capacity to bind C3d (Fig. S1C). Therefore, mutant receptor expression and levels were comparable to levels of CR in WT mice. Furthermore, the mutant CD21/CD35 receptors were functionally intact with regard to binding ligand (Fig. S1). B-cell surface levels of CD19 are dependent on expression of CD21/CD35, and the levels are increased by ≈30% in mice deficient in CD21/CD35 (18, 33). To examine whether surface levels are similarly increased in Cr2^Δ/Δgfp, splenic B cells were isolated and analyzed by flow cytometry. The results identify an increase in cell surface expression of CD19 on B cells prepared from both the mutant and deficient mice relative to WT (Fig. S1D). Thus, in the absence of interaction between CD21/CD35 and CD19, the levels of CD19 were increased, as found in Cr2^−/− mice.

Signaling via Mutant CR.

Coligation of CD21 and CD19 with subthreshold BCR stimulus augments B-cell activation (16, 34, 35). To test functional coreceptor signaling in primary B cells, the three strains of mice (WT; Cr2^−/−, ref. 17; and Cr2^Δ/Δgfp) were crossed with the B6.MD4 line, which bears a classical Ig transgene (Tg) encoding anti-hen egg lysozyme (HEL) antibody (36). Splenocytes were treated with an optimal amount of anti-IgM or a subthreshold level of duck egg lysozyme (DEL) alone or combined with C3d (rDEL-C3d₃). MD4 B cells from each mouse strain mobilized intracellular Ca²⁺ in response to cross-linking with 10 μg/mL anti-IgM (Fig. 1A Top). As reported previously (34), stimulation of MD4 B cells with DEL in amounts exceeding 100 nmol was sufficient to induce release of intracellular Ca²⁺ independently of coreceptor (Fig. 1A Middle). However, response to 1 nmol of DEL required both coreceptor and BCR signaling, as evidenced by lack of threshold response by WT MD4 B cells (Fig. 1A Middle). MD4 B cells from all three lines were treated with 1 nmol of rDEL-C3d₃ and, as expected, WT MD4 B cells but not Cr2^−/− B cells were responsive. Notably, the mutant B cells were also unresponsive to 1 nmol of rDEL-C3d₃, demonstrating that the coreceptor was not intact in the Cr2^Δ/Δgfp mice (Fig. 1A Bottom).

Fig. 1.

Functional responses of B cells with uncoupled CD21-CD19 coreceptor. (A) Early activation of Tg B cells stimulated with (Top) anti-IgM, (Middle) DEL, and (Bottom) rDELC3d₃ as measured by Ca²⁺ mobilization. B cells from WT (MD4; thick black lines), Cr2-deficient (Cr2^−/− MD4; thin black lines), and mutant Cr2 (Cr2^Δ/Δgfp MD4; thick gray lines) mice were loaded with Indo-1 am. The indicated stimuli were added after collecting baseline for 30 s. Results are representative of two independent experiments with at least three mice per genotype. (B) Proliferative responses in B cells from WT (Top), Cr2^−/− (Middle), and Cr2^Δ/Δgfp (Bottom) after stimulation with anti-CD40 antibody and either 10 μg of anti-IgM (optimal anti-IgM; Left) or with 1 μg of C3dg tetramers containing suboptimal anti-IgM (Right).

As a further test, functional coreceptor activity was also evaluated in vitro by using a carboxyfluorescein succinimidyl ester (CFSE) proliferation assay. Splenocytes from the three lines of mice were labeled with CFSE before culture with CD40 antibody and optimal anti-IgM (Fig. 1B). To test coreceptor signaling, mixed tetramers, composed of a suboptimal level of anti-IgM and C3d, and anti-CD40 were cultured with B cells for 3 days (Fig. 1B) (37). WT B cells proliferated in response to mixed tetramers and anti-CD40, as demonstrated previously (37), whereas Cr2^−/− and Cr2^Δ/Δgfp B cells were unresponsive. Therefore, CR binding of C3d tetramers and signaling via CD19 are functionally uncoupled in the Cr2^Δ/Δgfp B cells, leading to an impaired proliferation in vitro.

In Vivo Transport of ICs.

Previous studies have reported that naive B cells bind ICs through the CR and transport them into the B-cell follicle (25, 26). To determine whether transport of ICs into the follicle is dependent on coupling of CR to CD19, WT, Cr2^−/−, and Cr2^Δ/Δgfp mice were passively immunized with anti-B-phycoerythrin (PE) antibody and 24 h later were injected s.c. in the hind flanks with 10 μg of B-PE. The draining inguinal LNs were collected 8 h later and analyzed by FACS. As expected, Cr2^−/− B cells bound negligible PE-ICs. By contrast, both WT and mutant B cells bound appreciable levels of PE-ICs (Fig. 2 A and B). Thus, Cr2^Δ/Δgfp B cells, although uncoupled from CD19, retain the ability to transport PE-ICs into the follicle.

Fig. 2.

CR-mediated transport and uptake of ICs by B cells and FDCs is independent of CD19. (A) WT, Cr2^−/−, and Cr2^Δ/Δgfp were passively immunized with rabbit anti-B-PE and 24 h later were injected s.c. in the hind flank with 10 μg of B-PE. In vivo uptake of PE-ICs by naive polyclonal B cells was assessed 8 h after PE injection by FACS. Exogenous cells were added during processing to control for ex vivo capture of PE-ICs. Dot plots are representative of at least three LNs, and results are compiled into averages ± SEM in the graph (B). (C) Inguinal LNs were analyzed 24 h after PE injection for the degree of PE-IC deposition on FDCs by confocal microscopy by using anti-CD35 (blue) to label FDCs. (D) TEL-ICs are stably deposited on motile dendrites of FDCs. Preformed ICs containing 1 μg of Alexa 633-labeled TEL (red) were injected into a Cr2^Δ/Δgfp mouse, in which FDCs (green) express GFP, and were analyzed 24 h later. Multiphoton intravital microscopy analysis with illumination at 880 nm allowed suboptimal yet simultaneous excitation of both Alexa 633 and GFP. Second harmonic signals from collagen fibers are shown in blue. Macrophages (Mac) are identified by their yellow autofluorescence. The smaller images on the right highlight that TEL remains colocalized (Bottom), with motile FDC dendrites in the highlighted subregion of the main panel. Times are given in minutes and seconds. (Scale bar: 25 μm.)

To compare the retention of PE-ICs on FDCs in mice with intact CR/CD19 versus the mutant Cr2^Δ/Δgfp receptor, the inguinal LNs of the passively immunized mice were collected 24 h after PE injection. Confocal microscopic analysis identified a similar level of PE-ICs colocalized with FDCs in both WT and Cr2^Δ/Δgfp mice (Fig. 2C). Together, transfer of PE-ICs by naive B cells to FDCs and retention of PE-ICs on FDCs are independent of CD19.

To examine further the uptake of ICs in real time by FDCs in Cr2^Δ/Δgfp mice, we took advantage of the endogenous expression of GFP. Note that GFP levels were not sufficiently intense for imaging in vivo of the mutant follicular B cells, whereas the levels were ≈10-fold higher on FDCs. Cr2^Δ/Δgfp mice were injected in the footpad with 1 μg of preformed ICs comprising Alexa 633-labeled turkey egg lysozyme (TEL) and rabbit anti-lysozyme. In a previous report, we found that labeled TEL alone injected in the footpad rapidly filled the SCS of the popliteal LNs and drained into the follicles via discrete conduits (26). In the current study, TEL–IgG complexes were used to activate C3 and enhance uptake on FDCs via CD21/CD35 receptors. Subsequently, the draining popliteal LNs were surgically exposed and imaged by multiphoton intravital microscopy. TEL Ag (red) colocalized with the FDC dendritic processes (green) (Fig. 2D; see also Movie S1). Follicular conduits (identified by second harmonic signals; blue), which were shown recently to channel small Ags from the SCS into the FDC area, colocalized with the FDCs (26). Thus, the mutant line of mice could be used in future studies to track uptake of labeled Ag onto FDC by intravital microscopy and to address whether uptake via CD21/CD35 receptors induces signaling.

Humoral Immunity in CR Mutant Mice.

To examine the intrinsic effect of uncoupling of CR and CD19 on B cells in vivo, a bone marrow (BM) transplantation approach was used. The chimeric mice bore CR⁺ FDCs from the WT C57BL/6 recipients, but the B-cell compartment was derived from WT, Cr2^−/−, or Cr2^Δ/Δgfp donor BM. BM chimeric mice were immunized with either a high or low dose of 4-hydroxy-3-nitrophenyl conjugated to keyhole limpet hemocyanin (NP-KLH) at 0 and 3 weeks, and antibody responses were analyzed (Fig. S2). A potential disadvantage of the chimeric approach is increased variation among individual mice. However, the overall advantage is that the FDCs are consistently Cr2^+/+.

In response to high-dose Ag (50 μg i.v.), both WT and Cr2^Δ/Δgfp mice responded comparably (Fig. 3A Right). Further, the frequency of NP-specific antibody-secreting cells (B_ASCs) was similar in the BM of both WT and Cr2^Δ/Δgfp (Fig. 3B Right). Although the numbers of B_ASCs in the spleens of Cr2^Δ/Δgfp mice were dramatically higher than Cr2^−/−, they were only about one-half that of WT. As expected, Cr2^−/− mice failed to generate significant serum antibody titer to the hapten NP, and they also showed a markedly reduced frequency of NP-specific B_ASCs (Fig. 3). Therefore, humoral responses to high-dose soluble Ag were comparable between WT and Cr2^Δ/Δgfp B cells.

Fig. 3.

Effects of uncoupling CR–CD19 interactions on humoral immune responses. (A) Anti-NP IgG titers after immunization and challenge with 50 μg (Right) and 10 μg (Left) of NP-KLH i.v. Immunized BM chimeric mice were bled at day 0 and then 1 and 7 weeks after boost, and titers were assessed by ELISA (34). (B) Frequencies (per 10⁶ splenocytes) of NP-specific B_ASCs in WT, Cr2^−/−, and Cr2^Δ/Δgfp mice were determined by ELISPOT after immunization with 50 μg (Right) and 10 μg (Left) of NP-KLH. Asterisks denote significant differences relative to WT (*, P < 0.02; **, P < 0.04, t test).

Complement-dependent antibody responses are most apparent with lower amounts of Ag. Thus, humoral responses were measured 10 days after challenge with a low dose of Ag (10 μg i.v.). WT chimeras yielded a significantly greater titer than Cr2^−/− chimeras (Fig. 3A Left). Importantly, the IgG anti-NP titers were also reduced in Cr2^Δ/Δgfp relative to WT chimeras (Fig. 3A). Further, after 7 weeks, the decay in circulating NP-specific antibody was greater in the Cr2^−/− and mutant mice than Cr2⁺ chimeras. Although it correlated directly with serum anti-NP IgG titers, the frequency of B_ASCs was markedly higher in immunized Cr2⁺ chimeras compared with Cr2^−/− and Cr2^Δ/Δgfp cohorts (Fig. 3B Left). Thus, uncoupling of CR and CD19 substantially reduced humoral immunity relative to WT, but mutant CRs contributed to an enhanced response relative to Cr2^−/−.

GC Responses in CR Mutant Mice.

Given the reduced secondary response to low-dose Ag observed in the Cr2^−/− and Cr2^Δ/Δgfp mice, GC B cells were compared among the three groups of immunized mice. Ten days after secondary immunization with low-dose NP-KLH, the numbers of peanut agglutinin (PNA)^highB220⁺ B cells were quantitated. WT, Cr2^−/−, and Cr2^Δ/Δgfp BM chimeric mice did not differ statistically (Fig. S3 and Table 1). By contrast, a significant reduction was observed in the number of Cr2^−/− NP⁺ B cells relative to WT that had undergone class switch recombination (CSR) (IgM⁻) (Table 1). Similarly, the number of PNA^highNP⁺IgM⁻ B cells prepared from Cr2^Δ/Δgfp mice was reduced relative to WT (Fig. 4 A and B, and Table 1). Thus, efficient CSR of the NP⁺ B cells required an intact CR CD19 coreceptor. This could be explained by a requirement for coreceptor signaling for efficient Ag presentation to cognate T cells (38) and/or for enhanced up-regulation of survival genes, such as Bcl-2, cFLIP, or Bcl-xl (39, 40). Interestingly, CSR was more efficient in the Cr2^Δ/Δgfp than Cr2^−/− B cells, suggesting that binding of C3-coated Ag (or CRP activity) could also contribute to CSR and survival.

Table 1.

Comparison of GC and memory B cells isolated from WT, Cr2 mutant, and Cr2^−/− mice

Chimeras	PNAhighB220+IgM+ B cells, mean ± SD	NP+PNAhigh B cells 1–2 weeks after immunization, mean ± SD		NP+ BMEM cells 6–8 weeks after immunization, mean ± SD, IgM−
		IgM+	IgM−
WT	856,331 ± 139,023	273,869 ± 141,777	8,157 ± 2,050	1,169 ± 462
Cr2Δ/Δgpf	763,728 ± 337,073	439,864 ± 167,570	1,784 ± 1,074*	656 ± 437†
Cr2−/−	663,821 ± 332,473	368,816 ± 186,873	880 ± 1,275‡	Undetectable

Efficient class switch recombination and generation of B-cell memory are dependent on intact CR–CD19 coreceptors. GC B cells were enumerated in chimeric mice containing WT-derived FDCs and either WT, Cr2^−/−, or Cr2^Δ/Δgfp BM-derived B cells after immunization with 10 μg (low dose) of NP-KLH. Total numbers of GC B cells (second column), NP-specific B cells expressing surface IgM (third column), or NP-specific B cells that had undergone class switch recombination (fourth column) were assessed at 10 days after immunization by multiplying their frequency, as determined by flow cytometry, by the total splenocyte count. At 7 weeks after immunization, NP-specific memory B cells (B_MEMs) were quantitated in WT vs. Cr2^−/− and Cr2^Δ/Δgfp chimeras (fifth column).

*, P < 0.003.

†, P < 0.05.

‡, P < 0.001.

Fig. 4.

Identification of NP-specific B cells in GC from spleens of immunized BM chimeric mice. (A) Representative FACS plots from immunized (10 μg i.v.) BM chimeric mice with WT (Left), Cr2^−/− (Middle), and Cr2^Δ/Δgfp (Right) B cells. B cells were gated as described in Fig. S3. (B) Scatter plot showing numbers of NP-binding GC cells (three experiments; n = 6) in WT (black squares), Cr2^−/− (open circles), and Cr2^Δ/Δgfp (gray triangles) chimeras 7–14 days after rechallenge with 10 μg of NP₅-KLH i.v. (C) MFI for IgD levels on the surface of NP-binding IgM cells within GCs of chimeric mice. Shown are representative plots for NP⁺IgM⁺ from WT (solid gray), Cr2^−/− (dotted black line), and Cr2^Δ/Δgfp (solid gray line) BM chimeric mice. Mean values are indicated by bars. Statistical differences (t test) are indicated (*, P ≤ 0.001; **, P ≤ 0.003).

Maturation of GC B cells is suggested to occur stepwise from IgM⁺IgD⁺ > IgM⁺ IgD⁻ > IgG⁺ (41). To examine the stage of differentiation in which coreceptor signaling is required, Ag-specific, PNA^high B cells were analyzed at each of the three stages. Interestingly, no statistically significant difference in the number of NP⁺IgM⁺ GC B cells was observed between WT, Cr2^−/−, and Cr2^Δ/Δgfp (Table 1). However, when the NP⁺IgM⁺ population was examined for surface IgD levels, the highest mean fluorescence intensity (MFI) for IgD was consistently NP⁺IgM⁺ Cr2^−/− and Cr2^Δ/Δgfp B cells (Fig. 4C).

B-Cell Memory in Mutant Mice.

GCs serve to produce long-term protective humoral immunity via production of Ag-specific memory cells (B_MEMs) and B_ASCs (42, 43). To determine whether B-cell memory correlates with the block in isotype switch observed in PNA^high cells, two approaches were taken using splenic B cells from immunized BM chimeric mice: (i) cells were analyzed by FACS 7 weeks after challenge with low-dose soluble Ag, and the number of NP-binding B cells was quantitated, and (ii) B cells were adoptively transferred along with T cells from KLH-primed mice into Rag-1^−/− mice, and recipients were subsequently challenged with 50 μg of Ag in adjuvant, and antibody titers were determined (Fig. S4).

GC responses began to dissipate by 3 weeks after Ag encounter. To evaluate the production of B_MEM cells, BM chimeric mice were analyzed by FACS 7 weeks after immunization with low-dose Ag. As predicted, the number of NP-binding B cells was reduced over the time period analyzed for all three strains (Table 1). Interestingly, Cr2^Δ/Δgfp NP⁺ B cells showed an intermediate number of B_MEM cells relative to WT and Cr2^−/− (Fig. 5A and Table 1). Thus, an intact coreceptor is important for efficient differentiation of B_MEM, but CRs appear to have an additional role or roles.

Fig. 5.

Generation of B_MEMs from BM chimeric mice. (A) Spleens were analyzed 7 weeks after Ag challenge for the presence of NP⁺IgM⁻ (B220⁺PNA⁻IgD⁻CD3⁻CD4⁻CD11c⁻) B_MEM cells from WT (black squares), Cr2^−/− (open circles), and Cr2^Δ/Δgfp (gray triangles) chimeras (from three experiments). Statistical analyses are indicated: *, P ≤ 0.035; **, P ≤ 0.05. Total splenic B cells isolated 7 weeks after boost were adoptively transferred into Rag1^−/− recipients along with carrier-primed T cells from WT mice. After 2 weeks, Rag1^−/− chimeras were immunized with 50 μg of NP₅-KLH in alum, and IgG anti-NP titers (B) and ASCs (C) were measured as described. Shown is a summary from two experiments; statistical significance is indicated (*, P ≤ 0.03; **, P ≤ 0.04).

As an additional measure of B-cell memory, a Rag-1^−/− chimeric mouse model was used (Fig. S4). Rag-1^−/− mice were reconstituted with a mixture of hapten-primed splenic B cells and carrier-primed T cells. Rag-1^−/− chimeras receiving B cells from immunized WT mice had a mean IgG anti-NP titer significantly greater than recipients of Cr2^−/− B cells (Fig. 5B). B cells from immunized Cr2^Δ/Δgfp mice produced an intermediate response compared with WT and Cr2^−/− donor-derived B cells (Fig. 5B). ELISPOT results were consistent with antibody titers, because Cr2^Δ/Δgfp donor splenocytes produced an intermediate frequency of IgG, NP-specific B_ASCs compared with Rag-1^−/− chimeras receiving either WT or Cr2-deficient B cells (Fig. 5C). Thus, coreceptor signaling is a critical factor in formation and maintenance of long-term B_MEM, but CRs appear to play additional role/roles in memory that are independent of CD19.

Discussion

In the current study, we introduced a mutation within the murine Cr2 locus to uncouple the functional interaction between complement receptors CD21 and CD35 (CRs) from CD19 based on earlier studies in a human B-cell line (32). The mutant receptor, Cr2^Δ/Δgfp, binds C3 ligands and includes the CD35 domain that has complement regulatory activity (N-terminal domain) but impaired coreceptor signaling. To evaluate the intrinsic effect on B cells, a BM chimeric approach was used. Characterization of the Cr2^Δ/Δgfp mice identified an intermediate (relative to WT and Cr2^−/−) humoral response to low-dose, T-dependent Ag. The overall reduced antibody response correlated with a partial block at the IgM⁺IgD⁺ to IgM⁺IgD⁻ stage within the GC similar to that observed in Cr2^−/− B cells. Moreover, the mutant line had an impaired development of long-term B_ASCs and B_MEMs.

These results are consistent with the interpretation of findings from earlier studies in Cr2^−/− mice and Cr2^−/− BM chimeric mice that coreceptor signaling is critical for efficient humoral immunity to T-dependent Ags (15, 16, 44). Importantly, the new results identify the stage in GC differentiation in which coreceptor signaling is required (i.e., IgM⁺IgD⁺ to IgM⁺IgD⁻). Wang and Carter (45) identified a block in differentiation at a similar stage in GC B cells in CD19^−/− and CD19^−/− transgenic mice that express mutant CD19 (Y482F/Y513F). In their study, analysis of GC by BrdU labeling and histology identified arrest in differentiation of mutant B cells in the FDC light zone. This region of the GC is thought to promote clonal selection of B cells and includes cognate T cells as well as C3d-coated Ag.

One explanation for the signaling requirement of coreceptor within the GC light zone could be efficient presentation of Ag to cognate T cells to ensure costimulation and up-regulation of activation-induced cytidine deaminase. Pierce and colleagues (46) reported that coupling of C3d to HEL enhanced localization of BCR and coreceptor (CD19/CD21/CD81) in lipid rafts, resulting in prolonged signaling and presentation of Ag to T cells. The tetra-span protein CD81 interacts with CD19 and facilitates localization of the coreceptor and BCR in lipid rafts, and this enhances signaling via Ig α and vav (47). CD81-deficient mice have reduced responses to certain T-dependent Ags and increased responses to type II T-independent Ags (48). Moreover, the deficient mice have reduced numbers of B-1 cells, probably resulting from impaired expression of CD19 (49). Similarly, CD19^−/− mice have a block in GC response to T-dependent Ags, as discussed above, and a deficiency in B-1 cells (50).

This stage of differentiation is also critical for up-regulation of survival factors, such as Bcl-2 and Bcl-xl, and overexpression of these survival genes leads to excess survival of low-affinity GC B cells (39, 40). Thus, cross-linking of coreceptor enhances up-regulation of Bcl-2 (51) and Bcl-xl (52) and would enhance continued B-cell differentiation within the GC.

Not all T-dependent Ags require CR for development of humoral immunity, because the nature of the Ag and dose are important factors. For example, Cr2^−/− mice appear to respond to infectious vesicular stomatitis virus similarly to WT controls (14). Similarly, Cr2^−/− mice immunized with bacteriophage (QB phage) developed an apparently normal GC response, with a frequency of B memory cells and early antibody titers similar to those of WT controls; however, the antibody response failed to persist, as had been observed in responses to haptenated proteins (53). Interestingly, it was found that transcription factors Blimp-1 and XBP-1 were not efficiently up-regulated by post-GC plasma cells in the Cr2^−/− mice and that the defect was intrinsic to the B cells. Although the current study did not evaluate expression of Blimp-1 and XBP-1, identification of a general block in B-cell differentiation within the GC would explain the actual reduction in the number of Ag-specific B_ASCs and B_MEMs.

Although the humoral response of Cr2^Δ/Δgfp mice was impaired relative to WT mice, it was less severe than that of Cr2^−/− mice. This finding might be explained by a low level of functional interaction between mutant CR and CD19. Although we cannot rule out this possibility, negligible coreceptor activity was observed by using two sensitive assays in vitro. Moreover, the observation of a similar increase in expression of CD19 on B-cell surfaces in both the Cr2^Δ/Δgfp and Cr2^−/− mice supports a physical separation between CD19 and the mutant receptor. By contrast, functional activity involving binding of C3d or transport of C3d-coupled ICs appeared normal in the mutant mice.

The less severe defect in Cr2^Δ/Δgfp relative to Cr2^−/− mice supports CD19-independent role/roles of CR. For example, mutant B cells retain the ability to bind and transport C3d-coated ICs into the B-cell follicles. Recent studies highlight the importance of B-cell transport of Ag in both the spleen and peripheral LNs (23, 25, 26). In addition, Shlomchik and Rossbacher demonstrated that binding of Ag by cognate B cells in Ig-deficient mice could activate C3 on the cell surface, resulting in coupling of C3b to the specific Ag (54). Moreover, focusing of Ag–C3d complexes on the surface of cognate B cells could enhance “presentation” to neighboring B cells or transport to FDCs (55). Thus, the responsiveness and differentiation of B cells in mutants relative to Cr2^−/− could be explained in part by increased efficiency of cognate B cells in presentation of C3d-coated Ags to neighboring B cells (or deposition on FDC).

Finally, although a requirement for complement regulatory activity by CD35 on B cells has not been reported, it is possible that the less impaired response of mutant B cells could be explained in part by the expression of CR. Thus, complete absence of CR on Ag-binding B cells might increase their sensitivity to complement injury. By contrast, B cells in WT and mutant mice retain full CRP activity.

In summary, characterization of a novel line of mutant mice in which CRs are uncoupled from CD19 on B cells confirms the importance of coreceptor signaling for efficient response to T-dependent Ags. However, the finding of a less impaired humoral response relative to Cr2^−/− mice suggests a role for CRs independent of CD19.

Materials and Methods

Antibody Titer.

Serum was collected from individual mice, and NP-specific antibody titers were determined by sandwich ELISA, as described previously (56).

C3d Tetramers.

Functional binding of C3d on B cells was confirmed by using the approach of Henson et al. (57). C3d was kindly provided by David Karp (University of Texas Southwestern School of Medicine, Dallas, TX). Mixed tetramers of F(ab)₂′ goat anti-mouse IgM and C3d were prepared as described and were purified by gel filtration (37).

Chimeric Rag1^−/− Mice.

The generation of chimeric Rag1^−/− mice is described in SI Text.

Flow Cytometric Analysis.

Antibodies and reagents used for flow cytometry are described in SI Text.

Histology.

Histological analyses are described in SI Text.

Immunogens and Immunizations.

NP-KLH was used at two doses for i.v. immunizations: 10 μg (low dose) or 50 μg (high dose). Mice were analyzed 2 to 3 weeks after primary immunization and 1 to 6 weeks after boost. The formation of ICs in vivo and analysis of IC uptake was described previously (26).

Mice and BM Chimeras.

Mice were housed at the Immune Disease Institute (IDI) and Harvard Medical School (HMS) in specific pathogen-free facilities. MD4 HEL Ig Tg mice were maintained on a C57BL/6 background, with either WT (36), Cr2-deficient (Cr2^−/−) (17), or mutant (Cr2^Δ/Δgfp) Cr2 locus. A BM chimera approach was used to localize CR mutation to the B-cell population, as described previously (16). All animal procedures were Institutional Animal Care and Use Committee-approved at IDI and HMS.