IgG2a-Mediated Enhancement of Antibody and T Cell Responses and Its Relation to Inhibitory and Activating Fcγ Receptors¹

Immunoglobulin G-containing immune complexes (ICs)³ are capable of inducing inflammatory reactions, which, to a large extent, are mediated via a family of activating FcγRs (reviewed in Ref. 1). FcγRI and FcγRIII contain immunoreceptor tyrosine-based activation motifs (ITAM). Cross-linking of these receptors by their ligand, IgG/Ag complexes, induces intracellular signaling and various effector responses such as Ab-dependent cellular cytoxicity, mast cell degranulation, and phagocytosis (2). FcγRIIB is an inhibitory receptor with an immunoreceptor tyrosine-based inhibition motif in the cytoplasmic tail. When FcγRIIB is co-cross-linked with ITAM-containing receptors such as the B cell receptor (BCR), FcγRI, FcγRIII, and FcεRI, it inhibits ITAM-mediated signals (3). The in vivo role of FcγRs has been extensively studied in gene-targeted mice. Generally, inflammatory reactions, including models for autoimmune diseases, are exacerbated in mice lacking FcγRIIB (FcγRIIB^−/−) and impaired in mice lacking functional activating FcγRI and FcγRIII (owing to deletion of the common FcRγ chain, FcRγ^−/−) (4, 5).

The ligands for FcγRs, IgG Abs, obviously play a crucial role in initiating the effector functions of these receptors. IgG Abs are also potent regulators of the Ab-production per se. This takes place via a mechanism called Ab-mediated feedback regulation (6, 7). IgG, forming complexes in vivo with large particulate Ag, completely suppresses Ab responses to this Ag, a phenomenon that does not seem to require FcγRs (8, 9). IgG complexed to soluble protein Ags instead enhances the Ab response to this Ag by several hundredfold (10, 11, 12). Enhancement of Ab responses by the subclasses IgG1 and IgG2a requires the presence of activating FcγRs on a bone-marrow-derived cell type (11, 12) and, unlike enhancement by IgG3 (13), takes place in the absence of a functioning C system (14, 15). ICs containing IgG1 and IgG2a also bind to FcγRIIB which, in addition to its inhibitory effects, can mediate endocytosis of IgG/Ag complexes by e.g., dendritic cells and therefore, hypothetically cause IgG-mediated enhancement. However, the observation (11) that IgG-mediated enhancement was impaired in FcRγ^−/− animals (expressing normal levels of FcγRIIB) made it unlikely that FcγRIIB is responsible for IgG-mediated enhancement. Attempting to directly confirm this conclusion in FcγRIIB^−/− mice, we noted that IgG1, IgG2a, and IgG2b could induce enhancement which was not only functional, but was markedly higher than in wild-type (WT) animals (11).

Given the central role of IgG/Ag complexes in immune responses, both in the early phases regulating Ab responses and in later phases as effector molecules inducing inflammatory reactions, we found it worthwhile to clarify how IgG2a up-regulates Ab-responses to soluble protein Ags. A working hypothesis was that IgG/Ag complexes are efficiently taken up by FcγR⁺ APCs and presented to Th cells which then, via cognate help to B cells, induce augmented Ab responses. Using a transgenic system where the expansion of Ag-specific T cells could be monitored, we now visualize for the first time that IgG2a-mediated enhancement of Ab responses in vivo is indeed preceded by a FcRγ chain-dependent proliferation of Ag-specific T cells. We have also confirmed and extended our previous analysis of the influence of FcγRIIB on IgG2a-mediated feedback enhancement. The ability of IgG2a to enhance Ab responses, development of germinal centers (GCs), and immunological memory was much more pronounced in FcγRIIB^−/− mice than in WT controls. The difference in the magnitude of Ab-responses between the two strains was as big as 56-fold, thereby revealing one of the most potent down-regulatory effects of FcγRIIB observed. Thus, IgG2a has a dual role in regulating immune responses to soluble Ags. Targeting Ag to activating FcRγ chain-containing receptors enhances Ab as well as T cell responses while ligation of FcγRIIB down-regulates enhancement.

Mice carrying the H-2^b haplotype have an I-A^b-linked low responsiveness to IgE/BSA-2,4,6-trinitrophenyl (TNP) and IgG/BSA-TNP complexes (16, 17). To achieve responder phenotypes, male FcγRIIB-deficient mice (H-2^b) (18), a gift from Dr J. V. Ravetch (Rockefeller University, New York, NY), were backcrossed to CBA/J mice (H-2^k) (Bommice, Ry, Denmark) for 5 or 10 generations. Offspring from the 5th or 10th generation were intercrossed, and homozygous mutant H-2A^k animals were identified by PCR analysis of tail DNA. Offspring from these mice were used in the experiments. As WT controls for the 10th generation, CBA/J mice from Bommice were used. Mice homozygous for the WT FcγRIIB alleles were selected from the intercrosses of the 5th-generation backcross and their offspring used as WT controls for 5th-generation mutant mice. PCRs to identify the H-2 haplotype (H-2^k or H-2^b) and FcγRIIB genotype have been described (9).

In vivo T cell activation studies were performed on mice with a BALB/c background. As a source of OVA-specific T cells, DO11.10 mice, a gift from L. Westerberg (Karolinska Institute, Stockholm, Sweden) with the permission of Prof. K. Murphy (Washington University School of Medicine, St. Louis, MO), were used. DO11.10 mice carry a construct containing rearranged TCRα and TCRβ genes encoding a TCR specific for OVA 323-339 bound to I-A^d class II molecules (19), and have been backcrossed to BALB/c for >15 generations. Recipient mice were BALB/c (Bommice) and FcRγ^−/− (20) backcrossed to BALB/c for 12 generations (Taconic Farms, Germantown, NY). All animals were bred and maintained in the animal facilities at Department of Genetics and Pathology, Uppsala University (Uppsala, Sweden).

OVA, BSA, keyhole limpet hemocyanin (KLH), and TNP were obtained from Sigma-Aldrich (St. Louis, MO). TNP in the form of picrylsulfonic acid hydrate (P-2297; Sigma-Aldrich) was conjugated to OVA and BSA in 0.28 M cacodylate buffer, pH 6.9, as described (21). After a 45- to 70-min incubation at room temperature, the reaction was stopped by an excess of glycyl-glycin (1 mg/ml; Merck, Darmstadt, Germany). Proteins were dialyzed against PBS, sterile-filtered, and stored at 4°C. The number of TNP residues per BSA or OVA was determined as described earlier (21). Conjugates with 12 TNP/BSA and 1.3 or 3 TNP/OVA (both yielding similar results) were used.

mAbs were derived from B cell hybridomas producing IgG2a anti-TNP (C4007B4, 7B4) (10) and IgE anti-TNP (IGELb4) (22). The hybridoma cell lines were cultured in DMEM with 5% FCS. IgG2a was purified on a protein A-Sepharose column (Amersham Pharmacia Biotech, Uppsala, Sweden) according to manufacturer’s recommendations. IgE was purified by affinity chromatography on a Sepharose column conjugated with monoclonal rat anti-mouse κ (187.1.10) (23). Bound Ab was eluted with 0.1 M glycine-HCl buffer, pH 2.8. Abs were dialyzed against PBS, sterile-filtered, and stored at −20°C. Protein concentrations were determined by absorbance at 280 nm, assuming that an absorbance of 1.5 equals 1 mg/ml Ab. For flow cytometry, we used a PE-labeled IgG2a rat anti-mouse CD4 mAb (KH-CD4 or RM4–5; ImmunoKontact, Bioggio, Switzerland or BD PharMingen, San Diego, CA, respectively). The transgenic TCR was detected with a FITC-labeled murine IgG2a mAb specific for this particular TCR heterodimer (KJ1-26; Caltag Laboratories, Burlingame, CA) (24).

Mice were immunized in the tail veins with 0.2 ml of BSA-TNP or BSA-TNP/Ab complexes (and OVA as a specificity control) in PBS or with 0.1 ml of OVA-TNP or OVA-TNP/Ab complexes (and BSA or KLH as specificity controls) in PBS. The complexes were formed by incubating Ag with TNP-specific IgG2a or IgE mAbs for 1 h at 37°C immediately before immunization. For studies of the secondary response, mice were boosted with 20 μg of BSA in PBS s.c. in the flank.

Single cell suspensions from mouse spleens were prepared in ice-cold PBS by pressing cells through a mesh screen. RBCs were removed by hypotonic lysis, incubating spleen cells in 5 ml of ACK lysing buffer (0.15 M NH₄Cl (Merck), 1.0 mM KHCO₃ (Sigma-Aldrich), 0.1 mM Na₂EDTA (Sigma-Aldrich), pH 7.3) for 5 min at room temperature. The spleen cells were washed twice in PBS, resuspended in 5 ml of PBS, and counted using a hemacytometer. Fluorescence staining was performed at 4°C in 100 μl of PBS containing 5 × 10⁵ cells and predetermined optimal amounts of PE-labeled anti-CD4 (KH-CD4) and FITC-labeled anti-TCR (KJ1-26) mAbs. To reduce unspecific binding, 15 μl of normal BALB/c serum was added to the cell suspension before adding KJ1-26. After 30 min, the cells were washed twice with PBS containing 1% BSA (Sigma-Aldrich) and 0.1% NaN₃ (Sigma-Aldrich). From each stained aliquot of cells, 40,000 events acquired by gating on all cells with the forward- and side-scatter properties of lymphocytes were collected on a FACSort flow cytometer (BD Biosciences, Mountain View, CA). Dead cells were excluded based on propidium iodide absorption. The collected data was analyzed using CellQuest version 3.3 software (BD Biosciences). DO11.10 T cells were identified as CD4⁺KJ1-26⁺ events. Absolute numbers of DO11.10 T cells were calculated by multiplying the percentage of CD4⁺KJ1-26⁺ events with the total number of live spleen cells.

To study T cell activation in vivo, an adoptive transfer system developed in Dr. M. Jenkins’ group (25) was used. Spleen-cell suspensions from DO11.10 mice, containing 3 × 10⁶ CD4⁺ KJ1-26⁺ cells, were adoptively transferred i.v. to BALB/c mice 1–2 days before immunization. In experiments where FcRγ^−/− (and their WT controls) were the recipients of DO11.10 cells (see Fig. 2), only CD4⁺ cells were transferred (to avoid “contamination” with FcγR⁺ cells). CD4⁺ cells were prepared using Dynabeads Mouse CD4 (L3T4; Dynal Biotech, Oslo, Norway) and DETACHaBEAD Mouse CD4 (Dynal Biotech) according to manufacturer’s recommendations. Flow cytometry revealed that >99% of the cells were CD4⁺, and 2–2.5 × 10⁶ CD4⁺KJ1-26⁺ cells were adoptively transferred i.v. to recipients.

Mice were bled from the tails and sera tested in ELISA for BSA-, OVA-, or KLH-specific IgG. Microtiter plates (Immunolon 2 HB; Dynex Technologies, Chantilly, VA) were coated overnight (o.n.) at 4°C with 100 μl OVA (10 μg/ml), BSA (50 μg/ml), or KLH (10 μg/ml) in PBS with 0.05% NaN₃ (Bio-Rad, Richmond, CA). Fifty microliters of serial dilutions of serum samples in PBS-0.05% Tween 20 (PBS-Tw; Bio-Rad) was added to plates and incubated o.n. at 4°C. After washing, 50 μl of alkaline phosphatase-conjugated sheep anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA), diluted 1/1000 in PBS-Tw, was added and plates were incubated for 3 h at room temperature. After washing, 100 μl of substrate (p-nitro-phenylphosphate; Sigma-Aldrich) diluted in diethanolamine buffer was added and the absorbance at 405 nm was measured after 30 min incubation at room temperature. For BSA-specific ELISA, a polyclonal BSA-specific standard sera (affinity purified on BSA-Sepharose) with a starting concentration of 5.35 μg/ml was used. For OVA-specific ELISA, a polyclonal OVA-specific standard serum with a starting concentration of 3.47 μg/ml was used. Absorbance values at 405 nm were used to compare KLH-specific responses. Construction of standard curves and calculations were made by computer (Softmax; Molecular Devices, Menlo Park, CA). Stimulation indices (SI) were calculated as geometrical mean (anti-log₁₀) of the experimental group divided by the geometrical mean of the control group.

The number of B cells secreting BSA-specific IgG was determined using an ELISPOT assay (26). Spleens were removed and cell suspensions were prepared in DMEM with 0.5% normal mouse serum. One hundred-microliter cell suspensions were applied to BSA-coated microtiter plates (see ELISA) and incubated at 37°C, 5% CO₂ for 3.5 h. Plates were washed and incubated at 4°C o.n. with 50 μl of alkaline phosphatase-conjugated sheep anti-mouse IgG diluted 1/100 in PBS-Tw. Spots were developed for 1 h at room temperature in 50 μl of 5-bromo-4-chloro-3-indolyl phosphate (Sigma-Aldrich) and counted under a stereomicroscope.

Spleens were removed and divided into three pieces. Each piece was mounted onto a cork plate with an embedding medium (Tissue-Tek OCT 4583 compound; Sakura Finetechnical, Tokyo, Japan) and frozen in isopropanol in liquid nitrogen. Six-micrometer cryosections were mounted onto a SuperFrost Plus microscope slide (Menzel-Gläser, Braunschweig, Germany). The sections were fixed in ice-cold acetone and rehydrated in PBS. GCs were stained with 10 μg/ml biotinylated peanutagglutinin (L-6135; Sigma-Aldrich). The bound biotinylated peanutagglutinin was detected using an ABComplex/HRP kit (DAKO, Glostrup, Denmark). GCs were counted “blindly” under a microscope in 15 sections taken from different parts of the spleen.

Mice were injected intradermally with 10 or 100 μg of BSA emulsified in CFA (Difco, Detroit, MI) at the root of the tail in a total volume of 50 μl. Seven days after primary immunization, mice were challenged with 10 μg of BSA in 10 μl of PBS in the right ear, or 10 μl of PBS alone in the left ear. Ear thickness was measured before, as well as 1, 2, 3, and 5 days after challenge with a micrometer, accurate to 0.01 mm (Oditest; Kroeplin, Schlüchtern, Germany). Ear swelling due to BSA-specific delayed-type hypersensitivity (DTH) is expressed as the micrometer swelling in the BSA-challenged ear subtracted from the micrometer swelling in the PBS-challenged ear.

Statistical differences between the control and the experimental groups were determined by Student’s t test: NS, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; or ∗∗∗, p < 0.001.

We first addressed the question of whether IgG2a/Ag complexes could induce detectable proliferation of Ag-specific Th cells in vivo. Transgenic OVA-specific DO11.10 CD4⁺ T cells were transferred to BALB/c mice. Recipients were immunized with OVA-TNP or preformed IgG2a anti-TNP/OVA-TNP complexes, and the number of OVA-specific T cells in the spleen was determined in flow cytometry (Fig. 1 a). BALB/c mice transferred with DO11.10 cells but not immunized, or immunized with IgG2a anti-TNP alone, had small but detectable OVA-specific T cell populations (0.24 and 0.18%, respectively) which were absent in normal BALB/c mice (0.03%). Immunization with OVA-TNP increased the population to 0.36% and immunization with IgG2a anti-TNP/OVA-TNP complexes increased the population even further (1.56%). The 4.3-fold increase in population size after immunization with IgG2a-complexed Ag is representative; based on 12 independent experiments, the expansion ratios varied from 3.2- to 9.5-fold.

The kinetics of the expansion of specific CD4⁺ T cells and the correlation to serum titers of OVA-specific IgG were then analyzed in the same system (Fig. 1 b). As early as 2 days postimmunization, the number of OVA-specific T cells was significantly increased in mice immunized with IgG2a anti-TNP/OVA-TNP. The T cell response peaked at day 3, and the titers of OVA-specific IgG in serum around day 9. Thus, immunization of mice with IgG2a/Ag results in proliferation of specific CD4⁺ T cells, which precedes the enhanced Ab response.

As mentioned, IgG2a-mediated enhancement of Ab responses is severely impaired in FcRγ^−/− mice (11). To examine whether the IgG2a-induced T cell expansion is FcRγ-dependent, transgenic OVA-specific DO11.10 CD4⁺ T cells were transferred to WT and FcRγ^−/− mice. Recipients were immunized with OVA-TNP or preformed IgG2a anti-TNP/OVA-TNP complexes and the number of Ag-specific T cells in spleens was determined. Whereas T cell expansion was 5-fold in WT mice, no significant expansion of T cells was seen in FcRγ^−/− mice, neither in the experiment shown in Fig. 2,a, nor in the three other experiments performed. Four mice per group were saved and bled at various time points, confirming that FcγR^−/− mice have significantly reduced Ab responses to IgG2a/Ag as compared with WT mice (Fig. 2 b).

To characterize the role of FcγRIIB in regulating Ab responses, FcγRIIB^−/− and WT mice were immunized with preformed complexes of IgG2a anti-TNP/BSA-TNP or BSA-TNP alone, and the Ab response was analyzed. During the entire test period, FcγRIIB^−/− mice given IgG2a/Ag complexes had a much higher production of IgG anti-BSA than WT animals (Fig. 3 a). The SI at the peak of the response (days 21 and 28) were 9.6 and 10.4 in WT animals, whereas the corresponding SI in FcγRIIB^−/− mice were 246 and 476. Thus, IgG2a caused an “enhanced enhancement” of the primary Ab response in the absence of the inhibitory FcγRIIB.

To confirm that only the presence of the ligand for FcγRIIB, IgG, in an IC would influence the Ab response, control mice were immunized with complexes of IgE anti-TNP/BSA-TNP or with BSA-TNP alone. IgE-mediated enhancement of Ab responses is dependent on the low affinity receptor for IgE (FcεRII, CD23) (27, 28), which does not contain an ITAM, and therefore is not expected to be down-regulated by FcγRIIB. In line with this, there was no difference at any time point between the responses of FcγRIIB^−/− and WT mice to immunization with IgE/Ag-complexes (Fig. 3 b).

To investigate the effect of FcγRIIB on immunological memory, mice were primed with IgG2a anti-TNP/BSA-TNP i.v. and boosted with one s.c. injection of 20 μg of BSA in PBS 64 days later. Although secondary (as well as primary) Ab responses were markedly enhanced by IgG2a in WT mice, the effect was much more dramatic in FcγRIIB^−/− mice (Fig. 4).

The augmented serum IgG response in FcγRIIB^−/− mice could be due to increased affinity, decreased clearance of IgG, increased number of specific Ab-producing B cells, or to a combination of these factors. To investigate whether the number of specific Ab-producing B cells differed between FcγRIIB^−/− and WT mice, both strains were immunized with IgG2a anti-TNP/BSA-TNP and their spleens were analyzed in an ELISPOT assay (Fig. 5). In FcγRIIB^−/− mice, B cells producing BSA-specific IgG were found already at day 6 with a peak at day 14 (3576 spots per spleen). The response was sustained at least until day 22 when the experiment was terminated. WT mice had fewer BSA-specific IgG-producing B cells over the entire observation period, with a maximal number (435 spots per spleen) 10 days after immunization. Sera from the mice analyzed in ELISPOT assays were also analyzed in ELISA, confirming that the levels of IgG anti-BSA were higher in FcγRIIB^−/− than in WT mice (data not shown).

Next, we examined whether the elevated numbers of specific IgG-producing B cells in FcγRIIB^−/− mice were reflected in an increased GC formation. Spleens from FcγRIIB^−/− and WT mice were removed 3, 7, 10, and 14 days after immunization with IgG2a/BSA-TNP-complexes. Cryosections were stained with peanut hemagglutinin and 15 nonconsecutive sections of each spleen were examined (Table I). Three days after immunization, only a few GCs were detected in the two strains. Seven days after immunization, there were almost three times as many GCs in FcγRIIB^−/− mice as in WT mice. With time, however, the difference in GC numbers between WT and FcγRIIB^−/− mice appeared to even out, although the Ab response in the latter remained higher.

DTH is a cell-mediated inflammatory response induced by primed Th1 cells (29). Mice lacking either Abs (μMT) or activating FcγRs (FcRγ^−/−) have a reduced ability to mount DTH reactions (30), suggesting that IgG formed during priming with Ag in CFA, is important for initiating DTH responses. We reasoned that this experimental system could be used to investigate whether FcγRIIB negatively regulates the FcRγ-dependent T cell priming required for a DTH reaction. FcγRIIB^−/− and WT mice were immunized with BSA in CFA and challenged with soluble BSA. Subsequent measuring of ear swelling demonstrated that both strains developed similar DTH reactions (Fig. 6, a and b). Interestingly, FcγRIIB^−/− mice given the highest dose of BSA had a higher IgG anti-BSA response than WT mice, although this was not mirrored in an enhanced DTH reaction (Fig. 6 a).

We have studied the involvement of FcγRs in B and T cell responses to IgG2a/Ag complexes in vivo. The fact that IgG2a can ligate both activating and inhibitory FcγRs is clearly reflected in the response to IgG2a/Ag complexes. Our main findings are that IgG2a, complexed to its specific Ag, induces 1) a potent FcRγ chain-dependent proliferation of Ag-specific Th cells and 2) very high primary and secondary Ab responses in mice lacking the inhibitory FcγRIIB.

The molecular mechanism behind the ability of IgG2a to enhance Ab responses via FcγRs has been postulated to be increased uptake of IgG2a/Ag complexes by FcγR⁺ APCs, followed by efficient presentation of Ag to CD4⁺ Th cells. The ability of IgG to potentiate Ag presentation has been demonstrated in vitro (31, 32, 33, 34, 35) and ex vivo (34, 35), and has been implied in vivo (30, 36, 37). Here, we are able to directly visualize in vivo that IgG2a-mediated feedback enhancement of Ab responses to OVA is preceded by a rapid expansion of specific CD4⁺ T cells. The impairment of both T cell expansion and enhancement of Ab responses in FcRγ^−/− animals show that these functions are dependent on the presence of activating FcγRs. It is remarkable that i.v. immunization with rather low doses of native OVA (20 μg per mouse), complexed to IgG2a but without administration of conventional adjuvants, gives rise to a robust proliferation of T cells specific for one single OVA-peptide. We observed a 3.2- to 9.5-fold expansion of the splenic Ag-specific T cell population in mice immunized with IgG2a/Ag compared with mice immunized with Ag alone. Using the same adoptive transfer system, other workers observed a similar magnitude of T cell expansion in lymph nodes of mice immunized s.c. with native OVA together with LPS (38). However, to reach these levels of T cell proliferation, very high doses of Ag (2 mg per mouse) were required. Our results support the idea that IgG2a enhances Ab responses by increasing uptake of Ag by FcγR⁺ APCs, leading to enhanced presentation of Ag to Th cells, which provide help to specific B cells. The observations also suggest that capture of Ag by IgG2a is an extremely efficient way to induce immune responses when the concentration of Ag is limited and no adjuvants are present, as could be expected during biological immune responses.

FcγRIIB^−/− mice are more susceptible to autoimmune diseases such as Goodpasture’s syndrome (39), collagen-induced arthritis (40, 41), and lupus (42) than WT animals, and also have augmented IgG-mediated anaphylactic (18) and Arthus (43) reactions. Only a limited number of reports are available on the role of FcγRIIB in regulating Ab responses, i.e., the afferent limb of the inflammatory response. FcγRIIB^−/− mice immunized with SRBC or with KLH-TNP in adjuvant, produced ∼5-fold higher amounts of Abs compared with WT mice (18). The difference in Ab production was not evident until after ∼1 wk, probably reflecting the fact that endogenous IgG must first be produced to ligate FcγRIIB. Administration of preformed IgG/Ag complexes is a shortcut to studying the regulatory effects of FcγRIIB on Ab responses, without having to wait for endogenous IgG production. Using this approach, we have characterized a striking hyperresponsiveness to IgG-containing ICs in FcγRIIB^−/− mice. Although IgG2a caused enhancement of the primary Ab response, the priming for a secondary Ab response, the number of specific IgG-producing B cells, and the induction of GCs in WT mice, the responses were higher in FcγRIIB^−/− mice. As an example, we observed a 56-fold higher primary Ab response to IgG-complexed protein Ags than in WT mice, which is the most pronounced effect of FcγRIIB on Ab responses reported so far. Interestingly, FcγRIIB seems to negatively regulate Ab production by setting the upper level for Ab production rather than by completely preventing it. This is evidenced by the fact that IgG does indeed enhance Ab responses in WT animals although FcγRIIB is present (Refs. 10 and 11 ; Figs. 3,a and 4).

There are several hypothetical pathways by which FcγRIIB can negatively regulate responses to IgG/Ag complexes. On the level of T cell priming, it may inhibit FcγRI-mediated uptake of IgG/Ag or inhibit IC-induced dendritic cell maturation (34, 44, 45). FcRγ^−/− mice have reduced DTH responses and Th cell proliferation, most likely due to impaired IgG-mediated Ag presentation (30). The fact that FcγRIIB does not inhibit DTH responses (Fig. 6) therefore argues against the idea that this receptor down-regulates responses to IgG2a/Ag complexes by inhibiting Ag presentation. In a similar system as the one described in Fig. 2, we have tested directly whether FcγRIIB influences T cell priming. FcγRIIB^−/− mice on a BALB/c background were transferred with OVA-specific transgenic T cells from DO11.10 mice and immunized with OVA-TNP or IgG2a/OVA-TNP-complexes. In two of the experiments, we found no significant difference in T cell proliferation, whereas in two experiments, a 2-fold higher T cell proliferation in FcγRIIB^−/− mice was detected. Although this suggested that no major differences existed, interpretation of the data was complicated by our finding that Ab responses after immunization with IgG2a/OVA-TNP were only 3-fold higher in FcγRIIB^−/− mice on a BALB/c background than in WT animals. The reason for the less dramatic impact of FcγRIIB on Ab responses in BALB/c mice compared with CBA/J mice is not known. Noteworthy is that while lack of FcγRIIB in C57BL/6 mice led to spontaneous autoimmune disease, as well as to production of anti-nuclear Abs, no signs of autoimmunity were seen in BALB/c mice lacking the same receptor (42), again implying that FcγRIIB in BALB/c mice plays a less significant inhibitory role on Ab responses than it does in other mouse strains.

Another hypothesis is that FcγRIIB negatively regulates B cells when IgG2a/Ag complexes co-cross-link the BCR and FcγRIIB. In vitro, this leads to inhibition of B cell signaling (46, 47, 48, 49), B cell apoptosis (50), and inhibition of BCR-mediated Ag uptake and presentation (51, 52). Support for the existence of FcγRIIB-mediated inhibition of B cell activation in vivo comes from studies in different mouse strains, showing a correlation between reduced levels of FcγRIIB expression on GC B cells (as a result of FcγRIIB promoter polymorphisms) and higher Ab responses (53, 54). To date, there are no reports demonstrating that FcγRIIB negatively regulates Ag-presentation via activating FcγRs. Our present finding that DTH reactions take place normally in FcγRIIB^−/− animals as well as recent data from Dr. T. Takai’s laboratory (55) argue against an influence of FcγRIIB on FcγR-mediated Ag-presentation. Additionally, an abundance of data (mentioned above) show that FcγRIIB does indeed negatively regulate B cells. Therefore, we favor the idea that FcγRIIB inhibits Ab responses to IgG-complexed Ag by coligating the BCR. To definitely prove this point, studies in conditional knockout mice, selectively lacking FcγRIIB on B cells, will probably be required.

The presented data illustrates the intricate immunoregulatory role of IgG Abs. They enhance Ab responses via activating FcγRs, and in contrast, they down-regulate the same Ab response via FcγRIIB. The net result is seen in WT mice as enhanced responses to IgG-complexed soluble Ags. The effect of the positive regulation alone is seen in FcγRIIB^−/− mice as extremely high responses to IgG-complexed soluble Ags. The balance between IgG-mediated up- and down-regulatory effects is most likely a delicate one and it is easy to envisage how perturbations may lead to autoimmune disease, reported to occur with increased frequency in FcγRIIB^−/− mice (39, 40, 41, 42).

We thank I. Brogren for skilful technical assistance, Dr. S. Kleinau for critical reading of the manuscript, Dr. J. V. Ravetch for the gift of the FcγRIIB^−/− founder animals, Drs. K. Murphy and L. Westerberg for DO11.10 founders, Dr. M. Wabl for IgE, and Dr. P. Coulie for IgG hybridomas.