C57BL/6JBomTac mice (C57BL/6) were from Taconic Bioscience, Inc. (Hudson, NY, USA) and BALB/c mice from Bommice (Ry, Denmark). Fc receptor gamma chain (FcRγ) KO founders were a gift from Ravetch et al. (26) and were backcrossed to BALB/c for 10 generations. FcRγ KO mice lack the common FcRγ-chain and thereby all activating FcγRs associated with this chain (FcγRI, FcγRIII, and FcγRIV). Mice were age and sex matched within each experiment and were bred and maintained in the animal facilities of the National Veterinary Institute (Uppsala, Sweden). This study was carried out in accordance with the recommendations of the Uppsala Animal Research Ethics Committee and the protocol was approved by this committee.

Polyclonal IgG^a anti-SRBC and polyclonal IgG^a anti-NP were prepared from hyperimmune BALB/c serum and polyclonal IgG^b anti-SRBC from hyperimmune C57BL/6 serum. IgG was purified by affinity chromatography over a Protein-A Sepharose column (Amersham Pharmacia Biotech, Uppsala, Sweden) (27), dialyzed against PBS, sterile filtered and stored at −20°C until use. IgG^a anti-NP was biotinylated using 0.18 mg EZ-Link Sulfo-NHS-LC-LC-Biotin (sulfosuccinimidyl-6-[biotinamido]-6-hexanamido hexaonate) (Thermo Scientific, Waltham, MA, USA) per 2 mg IgG according to manufacturer’s recommendations. The reaction was performed at room temperature for 30 min and free biotin was removed by dialysis against PBS. IgG^a anti-NP-biotin was sterile filtered and stored at 4°C until use. SRBC were acquired from Håtunalab AB (Håtunaholm, Sweden) and stored in sterile Alsever’s solution at 4°C. SRBC were washed three times in PBS before use. 4-hydroxy-3-nitrophenylacetic-e-aminocaproyl-OSu (NP-ε-Aminocaproyl-OSu) (Biosearch Technologies, Petaluma, CA, USA) was conjugated to SRBC as described before (16). Briefly, NP-ε-Aminocaproyl-OSu was dissolved in dimethylformamide at a concentration of 7.5 mg/ml. Dissolved NP-ε-Aminocaproyl-OSu were then added to 5% SRBC suspensions in conjugation buffer (0.1 M NaHCO₃ with 0.15 M NaCl, pH 8.5) to a final concentration of 250 μg/ml (in Figure 3, a final concentration of 1 mg/ml NP-ε-Aminocaproyl-OSu was used in order to achieve higher coupling ratio and facilitate visualization in sections) and incubated for 1 h at room temperature. After the conjugation reaction, cells were washed three times in PBS before use.

All mice were immunized with SRBC or NP-SRBC in one of their lateral tail veins in 200 μl PBS. Antigen-specific IgG was always administered in 200 μl PBS 30 min prior to antigen, also via the lateral tail veins. Controls received antigen alone or antigen-specific IgG alone. Secondary immunizations were done with SRBC alone. Further details of doses are given in the figure legends. The “default” doses were 30–50 μg IgG and 5 × 10⁷ (NP-)SRBC, both of which are known cause 90–99% suppression of antibody responses to this amount of erythrocytes. In studies of NP-specific antibody-secreting cells (Figure 3), the higher dose 5 × 10⁸ NP-SRBC had to be used to induce detectable numbers of NP-specific cells and the dose of IgG was correspondingly increased to 100 μg. In immunological memory experiments, the lower priming dose 5 × 10⁶ SRBC together with 10 μg IgG was used (Figure 5A) in addition to the default doses (Figure 5B) to test whether the strength of priming affected whether IgG was able to suppress memory induction or not. Blood was collected from the ventral tail artery.

Enzyme-linked immunospot assay was used for measuring SRBC-specific IgG-secreting cells (28). Briefly, 96-well plates (Costar 96-well enzyme immunoassay/RIA; Sigma-Aldrich, St. Louis, MO, USA) were coated with 0.25% SRBC. Spleen cells were diluted in cell culture medium (DMEM with 0.5% FCS), 100 μl was added to each well, and the plates were incubated at 37°C for 2.5 h. SRBC-specific antibodies produced by the single cells were detected after addition of goat anti-mouse IgG-alkaline phosphatase (Jackson ImmunoResearch Laboratories, Inc., West Baltimore Pike, Media, PA, USA) for 3 h at room temperature, followed by addition of the precipitating substrate 5-bromo-4-chloro-3-indolyl phosphate (Sigma-Aldrich) for 30 min at room temperature. The plates were washed three times in PBS and samples were counted blindly under a stereomicroscope.

Also NP- and SRBC-specific antibodies in sera were measured by ELISA (16). Briefly, 96-well plates (Costar 96-well enzyme immunoassay/RIA; Sigma-Aldrich) were coated with either 100 μl 0.25% SRBC or 100 μl NP₂₀-BSA (Biosearch Technologies) (50 μg/ml) in PBS with 0.05% NaN₃. The plates were then blocked with 5% dry milk at 4°C overnight. After washing three times with PBS, serum samples were serially diluted, added to the plates, and incubated overnight at 4°C. When measuring secondary IgG responses, starting serum dilution for titer determination was 1:5. Cutoff for titers was set as mean OD_405nm + 2× SD for a group of eight individual sera from naive BALB/c mice diluted 1:5 (29). When measuring the IgG response, depending on the allotype of injected antibodies, either biotinylated anti-IgG1^b (clone B68-2) and anti-IgG2a^b (clone 5.7) or biotinylated anti-IgG1^a (clone 10.9) and anti-IgG2a^a (clone 8.3) (all from BD Pharmingen, San Jose, CA, USA) were mixed 1:1, added to each well, and incubated overnight at 4°C. After washing, alkaline phosphatase-conjugated streptavidin (BD Pharmingen, San Jose, CA, USA) was added and incubated for 3 h at room temperature. After washing, plates were developed using the substrate (p-nitrophenylphosphate; Sigma-Aldrich). Absorbance at 405 nm was measured and data analyzed using SoftMax software (Molecular Devices, Sunnyvale, CA, USA).

Spleen cells were prepared as described before (30) and re-suspended in FACS buffer (PBS with 2% fetal bovine serum). Samples were treated with Fc block (anti-CD16/32; BD Biosciences, San Jose, CA, USA) for 10 min on ice, then stained with anti-B220-Alexa700 (clone RA3-6B2), anti-GL7-BV421 (clone GL7), anti-CD95-PEcy7 (clone Jo2), anti-λ1-biotin (clone R11-153) (all from BD Biosciences) at 4°C for 30 min. After washing twice in FACS buffer, samples were stained with streptavidin-FITC and NP-PE at 4°C for 30 min and re-suspended in 300 μl FACS buffer after washing in FACS buffer. For each sample, 2–3 million events were acquired on a LSR Fortessa cytometer (BD Biosciences) at the BioVis platform, SciLifeLab, Uppsala, Sweden. Data were analyzed with FlowJo software (Tree Star Inc., Ashland, OR, USA).

Spleens were harvested, embedded in optimal cutting temperature embedding compound (VWR international, Radnor, PA, USA), flash-frozen in liquid nitrogen, and stored at −80°C. Eight-micrometer sections were cut using Thermo Scientific CryoStar NX70 Cryostat (Thermo Scientific, Waltham, MA, USA), thaw-mounted on frost plus microscope slides (Menzel-Gläser, Braunschweig, Germany), air-dried and stored at −80°C until use. Prior to staining, slides were fixed in 4% paraformaldehyde (Merck, Darmstadt, Germany) in PBS (pH 7.8) for 15 min or in 50% acetone for 30 s followed by 100% acetone (Sigma-Aldrich) for 5 min. Slides were then rehydrated in PBS and blocked with 5% horse serum (Sigma-Aldrich) in PBS for 30 min. Slides were stained with primary antibodies for 1 h. After washing twice in PBS, fluorochrome-conjugated streptavidin was added and the slides were incubated for 1 h and washed twice in PBS prior to mounting with Fluoromount G (Southern Biotech, Birmingham, AL, USA). For detection of NP-specific B cells, slides were stained with anti-IgD-Alexa 488 (clone 11-26C.2a, BioLegend, San Diego, CA, USA) and NP-PE (Biosearch Technologies). Localization of SRBC to splenic marginal zones (MZs) were investigated using anti-B220-Pacific blue (clone RA3-6B2, BD Biosciences), anti-CD169 (MOMA)-FITC (clone MOMA-1, AbD Serotec, Raleigh, NC, USA), and NP-SRBC detection by an in-house produced polyclonal IgG anti-NP-biotin followed by streptavidin-PE (eBioscience, San Diego, CA, USA). Images of immunofluorescence were acquired blindly and randomized with a LSM 700 confocal microscope (Carl Zeiss, Thornwood, NY, USA) using Zen 2009 software (Carl Zeiss). Tile-scan images of whole spleen sections were acquired using Zen 2009 software. Images were analyzed blindly with ImageJ software (NIH, Bethesda, MD, USA).

Statistical differences between groups were determined by the two-tailed Student’s t-test. Statistical significance levels were set as: ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.

Specific IgG administered together with SRBC reduces the localization of SRBC in the MZ of the spleen and increases clearance of SRBC from the blood (16). However, after 10 min, the levels of SRBC in the circulation were undetectable, regardless of whether IgG had been co-administered or not (16). The correlation between suppression and antigen localization was not directly assessed nor was the importance of FcγRs studied. To investigate this, BALB/c and FcRγ KO mice were immunized with IgG anti-SRBC + 5 × 10⁷ NP-SRBC, 5 × 10⁷ NP-SRBC alone, or with 1 × 10⁷ NP-SRBC alone. Spleens were harvested after 10 min and the amount of extracellular NP-SRBC was determined by confocal laser scanning microscopy (Figures 1A–R; enhanced images are shown in Figure S1 in Supplementary Material). In parallel, the serum IgG anti-SRBC response was followed in groups of mice immunized at the same time (Figure 1S). Because the passively administered IgG was obtained from mice with a different IgG allotype than the recipient mice, the actively produced endogenous IgG antibodies could be distinguished by an allotype-specific ELISA (21). Administration of IgG significantly reduced the amount of SRBC in the MZ of wild-type BALB/c mice (Figures 1B,F vs. Figures 1A,E; Figure 1Q) as well as in the entire spleen section (Figure 1R; Figure S2 in Supplementary Material). BALB/c mice immunized with 1 × 10⁷ NP-SRBC alone, or with the fivefold higher amount of NP-SRBC together with IgG anti-SRBC, had comparable amounts of NP-SRBC in their spleens (Figures 1C,G,R). In spite of this, mice immunized with 1 × 10⁷ NP-SRBC alone mounted a potent antibody response while mice immunized with IgG anti-SRBC + 5 × 10⁷ NP-SRBC had a suppressed antibody response (Figure 1S). Some spleens were also analyzed for NP-SRBC after 1 h and 24 h, but, as expected, very little or no antigen could be detected at these time points [Ref. (31) and data not shown]. In FcRγ KO mice, IgG was unable reduce the amount of SRBC in the MZ (Figures 1J,N,Q) and in the entire spleen (Figure 1R). In spite of the similar amounts of SRBC in the spleens, the antibody response was suppressed in the IgG group (Figure 1S).

In summary, in BALB/c mice, equal (low) levels of antigen in the spleen can result in high (1 × 10⁷ NP-SRBC-group) or suppressed (IgG + 5 × 10⁷ NP-SRBC-group) antibody responses. In FcγR KO mice, equal (high) levels of antigen can result in high (5 × 10⁷ NP-SRBC) or suppressed (IgG + 5 × 10⁷ NP-SRBC-group) antibody responses. This lack of correlation between suppression and the amount of NP-SRBC detected in the spleen is hard to reconcile with clearance of antigen as the major explanation for IgG-mediated suppression. Moreover, the observations confirm that administration of SRBC-specific IgG reduces the amount of antigen localized in the spleen (16) and demonstrate that the reduction is dependent on activating FcγRs while the suppression of antibody responses is not.

Next, we sought to determine whether IgG only suppresses responses against the epitopes to which it binds (epitope-specific suppression) or whether it, in addition, suppresses responses against other epitopes on the same antigen (non-epitope specific suppression). To this end, C57BL/6 mice were immunized with NP-SRBC alone or together with either IgG anti-NP or IgG anti-SRBC. The mice were bled every 2 weeks and their antibody responses against NP and SRBC were analyzed. IgG anti-NP suppressed the NP- but not the SRBC-specific IgG-responses (Figures 2A,B) while IgG anti-SRBC suppressed the SRBC- but not the NP-specific IgG response (Figures 2C,D). These observations were highly reproducible (4/4 experiments with IgG anti-NP and 2/2 with IgG anti-SRBC) and demonstrate that IgG is able to suppress IgG responses in an epitope-specific way.

In previous studies, the ability of IgG to suppress the overall serum antibody levels or splenic antibody-secreting cells was analyzed. Here, we investigated which B cell subpopulations were suppressed by IgG, using a system in which the BCRs of the antigen-specific B cells could be stained directly. The antibody response against NP in C57BL/6 mice is genetically restricted, mainly comprising λ1 light chains and the V186.2 segment of the VHJ558 gene family (32). Therefore, staining for λ1 and NP can be used to identify NP-specific B cells. C57BL/6 mice were immunized with NP-SRBC ± IgG anti-NP and negative controls with unconjugated SRBC or left unimmunized. Six days later, spleens were analyzed for NP-binding cells with flow cytometry and confocal microscopy (Figure 3). In mice immunized with NP-SRBC alone, a small population of λ1⁺NP⁺B220⁺ cells were identified both among the GL7^high CD95^high GC B cells (Figure 3B bottom right panel; Figure 3C) and among the GL7^lowCD95^low non-GC B cells (Figure 3B bottom left panel; Figure 3C). The NP-specific cells (GC B cells and non-GC B cells together) constituted approximately 0.05% of the total B220⁺ population and ~1/5 were GC B cells and ~4/5 non-GC B cells (Figure 3C). Administration of IgG anti-NP together with NP-SRBC abolished the induction of NP-specific GC and non-GC B cells (Figures 3A,C). In fact, the levels in these mice were equally low as in mice immunized with unconjugated SRBC or left untreated (Figure 3C).

The other half of the spleens were analyzed in confocal microscopy (Figures 3D–I). NP-binding cells were readily detected in extrafollicular foci in mice immunized with NP-SRBC (Figure 3E) while very few NP-binding cells were detected in mice immunized with IgG anti-NP + NP-SRBC (Figure 3D) and in the negative controls (Figures 3F,G). Although it cannot be excluded that some of the NP-binding cells in extrafollicular foci had exited GCs, the majority are presumably true extrafollicular cells owing to the early time point at which they were analyzed. Importantly, all NP-specific cells outside the follicles, regardless of their origin, are absent in mice immunized with IgG anti-NP + NP-SRBC. NP-specific cells could also be detected in some GCs (Figure 3H). GC B cells stained less brightly than the extrafollicular NP-specific cells possibly due to a wider span in affinity (33) and lower Ig expression. To confirm that the unstained areas close to the IgD⁺ areas are indeed GCs, co-staining with PNA was performed (Figure S3 in Supplementary Material). In summary, IgG anti-NP can suppress the generation of both extrafollicular NP-specific antibody-secreting cells and NP-specific GC B cells.

To determine whether IgG suppressed the development of long-lived plasma cells, mice were immunized with SRBC ± IgG anti-SRBC and the number of SRBC-specific IgG-secreting cells in the spleen and bone marrow was analyzed 5–70 days after immunization (Figure 4). Cells secreting IgG anti-SRBC were detected in both organs at all times, but were severely suppressed in the groups receiving IgG together with SRBC. As expected, the number of antibody-producing cells in the spleen was highest early after immunization and then decreased, while the cell number in the bone marrow increased during the entire time period. Thus, administration of IgG anti-SRBC efficiently suppressed the specific IgG-secreting cells at all times, both in spleen and bone marrow.

To test the ability of IgG to regulate memory induction, BALB/c mice were immunized in two regimes: 5 × 10⁷ SRBC ± 50 μg IgG anti-SRBC or 5 × 10⁶ SRBC ± 10 μg IgG anti-SRBC. The IgG anti-SRBC responses were followed during 63 days after priming, confirming efficient suppression (Figures 5A,B). And, 70 days after priming, all mice were boosted with a suboptimal number of SRBC and, in addition, a group of naive mice received the same “booster” dose. Because of the great differences in antibody levels between the groups, we analyzed the antibody responses as endpoint titers, starting at a serum dilution of 1:5. In mice that had been primed with SRBC alone and boosted, a very high secondary antibody response was observed (Figures 5A,B). In mice primed with IgG + SRBC and boosted, secondary responses were low (Figures 5A,B). A minute priming effect was, however, visible also in the IgG groups, evidenced by the fact that their antibody titers were higher than those in naive and “boosted” mice (Figures 5A,B). Naive mice, primed on day 70 with the suboptimal dose 5 × 10⁵ SRBC, produce low titers of IgG antibodies [mean titers: 30 (Figure 5A) and 1,100 (Figure 5B)], which are barely visible in the figures. The relative suppression of primary responses and induction of memory appeared to be equal. In the low dose experiment, IgG suppressed 96% of the IgG-response on day 21 and 99% on day 91 (21 days after boost) (Figure 5A). In the high dose experiment, suppression was 97% both 21 days after priming and 21 days after boost (Figure 5B). Thus, when IgG-mediated suppression of primary antibody responses is very efficient, priming for secondary antibody responses seems to follow along the same lines.