BALB/c mice were from Bommice (Ry, Denmark) and C57BL/6BomTac mice (C57BL/6) from Taconic Bioscience, Inc. (Hudson, NY, USA). Mice were age and sex matched within each experiment (both males and females were used) 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 the Uppsala Animal Research Ethics Committee.

Polyclonal IgG^b anti-SRBC was prepared from hyperimmune C57BL/6 serum, and polyclonal IgG^a anti-SRBC was prepared from hyperimmune BALB/c serum. IgG was purified by affinity chromatography using a Protein A Sepharose column (Amersham Pharmacia Biotech, Uppsala, Sweden) (25). Isolated IgG anti-SRBC was dialyzed against PBS, sterile filtered and stored at −20°C until use. SRBC in sterile Alsever’s solution were purchased from Håtunalab AB (Håtunaholm, Sweden) and stored at 4°C until use. SRBC were washed three times in PBS prior to use.

Mice were immunized with 200 µl SRBC ± 200 µl IgG anti-SRBC, both in PBS, in one of their lateral tail veins. Ten micrograms of IgG anti-SRBC were administered 30 min prior to 5 × 10⁶ SRBC. Controls received 5 × 10⁶ SRBC alone, 10 µg IgG anti-SRBC alone, or were left unimmunized. Blood was collected from the ventral tail artery. Details regarding IgG-allotypes are given in figure legends.

To distinguish SRBC-specific IgG, actively produced in the immunized mice, from the passively administered IgG anti-SRBC an allotype-specific protocol was used (20). Briefly, ELISA plates were coated with 100 µl 0.25% SRBC and blocked with 5% dry milk in PBS. Serum samples were added, followed by a 1:1 mixture of biotinylated anti-mouse IgG1^a (clone 10.9) and IgG2a^a (clone 8.3) (BD Pharmingen, San Jose, CA, USA). Plates were developed using alkaline phosphatase conjugated to streptavidin (BD Pharmingen) and p-nitrophenylphosphate as substrate (Sigma-Aldrich). Absorbance at 405 nm was measured and analyzed using SoftMax software (Molecular Devices, Sunnyvale, CA, USA). The results are given as OD_405nm values, and serum dilutions are chosen so that the highest values do not reach plateau levels.

To detect IgG^a specific for IgG^b, 96-well plates (Sigma-Aldrich) were coated with 100 µl 50 µg/ml IgG^b anti-SRBC in PBS overnight at 4°C and blocked with 5% dry milk in PBS for 2 h in room temperature prior to use. The remaining steps were performed as for the allotype-specific SRBC ELISA described above.

Single cell suspensions were prepared from spleens as described (26). Cells were resuspended in FACS buffer (2% fetal bovine serum in PBS) and incubated with Fc-block (anti-CD16/32; BD Biosciences) for 10 min on ice. GC B cells and DZ/LZ polarization were evaluated by staining with anti-B220-Pacific blue (clone RA3-6B2), anti-CD95-PECy7 (clone Jo2), anti-CD38-Alexa fluor 647 (clone 90), anti-CD83-PE (clone Michel-19) (BD Biosciences, San Jose, CA, USA), and anti-CD86-Biotin (clone GL1) (BD Pharmingen) for 30 min at 4°C. After washing twice in FACS buffer, Streptavidin-FITC (eBioscience, San Diego, CA, USA) was added, and samples incubated for 30 min at 4°C. T_FH/T_FR populations were analyzed by surface staining with anti-CD4-PECy5 (clone GK1.5), anti-CXCR5-PECy7 (clone SPRCL5), and anti-PD1-Biotin (clone J43) (eBioscience). CXCR5 staining was performed at 37°C for 30 min and CD4 and PD1 at 4°C for 30 min. After washing twice in FACS buffer, Streptavidin-FITC (eBioscience) was added, and samples incubated for 30 min at 4°C. Intracellular staining of Foxp3 was performed using the Foxp3-PE kit from eBioscience according to the manufacturer’s suggestions. Cells were resuspended in 300 µl FACS buffer, and data were acquired using an LSR Fortessa cytometer (BD Biosciences) at the BioVis platform, Uppsala, Sweden and analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA).

Spleen sections for confocal microscopy were prepared as described (21). GCs were visualized by staining the sections with anti-B220-Pacific Blue (clone RA3-6B2) (BD Biosciences), anti-CD169 (MOMA)-FITC (clone MOMA-1) (Bio-Rad antibodies, Raleigh, NC, USA), and biotinylated peanut agglutinin (PNA; Vector Laboratories, Burlingame, CA, USA) for 1 h in room temperature. After washing twice in PBS, PE-conjugated streptavidin was added, and the slides were incubated for 1 h in room temperature and washed twice before mounting with Fluoromount G (Southern Biotech, Birmingham, AL, USA). Tile-scan images of immunofluorescence of whole spleen sections were acquired with an LSM 700 confocal microscope (Carl Zeiss, Thornwood, NY, USA) using Zen 2009 software (Carl Zeiss). Images were processed and analyzed 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 follows: ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.

Germinal center formation is important for affinity maturation and the generation of long-lived plasma cells and memory B-cells (27, 28). Specific IgG limits the amount of SRBC reaching the spleen (21, 24), and it seemed possible that IgG would also limit the induction of splenic GCs. To assess this, BALB/c mice were immunized with SRBC ± IgG^b anti-SRBC, and the GC response was followed for 4–14 days after immunization. Half of each spleen was analyzed by confocal laser scanning microscopy (Figures 1A–C). Six days after immunization, clearly visible PNA⁺ GCs had developed in both groups although at this time point the GC responses were higher in the groups immunized with SRBC alone than with IgG and SRBC (Figures 1A–C). Surprisingly, from day 8 to 10, the GC frequency (Figure 1B) and GC size (Figure 1C) were equally strong in both groups (Figures 1A–C). Immunization with IgG anti-SRBC alone did not induce GC formation (Figures 1A–C).

The other half of each spleen was analyzed by flow cytometry, and the gating strategy for CD95^hiPNA^hi GC B-cells is shown (Figure 1D). After 6 days, the number of CD95^hiPNA^hi GC B-cells in mice immunized with SRBC alone peaked and was fourfold to fivefold higher than in mice immunized with IgG anti-SRBC + SRBC (Figure 1E). Importantly, at all other time points GCs developed to the same extent in both groups (Figure 1E). No GC B-cells were detected in mice immunized with IgG anti-SRBC alone (Figure 1E). The same results were obtained when the percentage of GC B-cells was analyzed using other gating strategies, defining GC B-cells as CD95^hiCD38⁻, GL7^hiCD95^hi, PNA^hiCD95^hi, or PNA^hiGL7^hi cells (data not shown).

The administration of IgG^b anti-SRBC to BALB/c mice, with the Ig allotype a, constitutes an allogeneic situation (Figure 1). To investigate whether GC formation occurred to a similar extent in a syngeneic situation, BALB/c mice were immunized with SRBC alone or with SRBC together with either syngeneic IgG^a or allogeneic IgG^b anti-SRBC. The GC response was analyzed 6 and 10 days after immunization. After 6 days, PNA⁺ GCs had developed in all three groups (Figures 2A–C). At this time, the GC frequency (Figure 2I), the GC size (Figure 2J), and the number of GC B-cells (Figure 2L) were higher in mice immunized with SRBC alone than in mice immunized with either IgG^a or IgG^b together with SRBC. After 10 days, the GC response was similar in all three groups (Figures 2E–L).

To confirm that IgG indeed suppressed the anti-SRBC response, in spite of the strong GC reactions observed, additional mice from the groups analyzed for GC responses (Figures 2A–L) were bled and tested in ELISA. In BALB/c mice immunized with IgG^b anti-SRBC, the actively produced IgG^a anti-SRBC could be discriminated from the passively administered IgG^b anti-SRBC in an allotype-specific ELISA. In BALB/c mice immunized with IgG^a anti-SRBC, the ELISA detection antibodies will also detect the passively administered IgG. To compensate for this, a group of BALB/c mice was immunized with IgG^a anti-SRBC alone, and the OD_405nm values in their sera were subtracted from the OD_405nm values in mice immunized with IgG^a anti-SRBC together with SRBC. As expected, passively administered IgG severely suppressed the IgG anti-SRBC response, both in the allotype-specific ELISA (Figure 2M) and in an ELISA measuring total IgG anti-SRBC antibodies (Figure S1 in Supplementary Material).

In summary, GCs develop in mice where the primary SRBC-specific antibody response is severely suppressed either by syngeneic or allogeneic IgG anti-SRBC. Six days after immunization, the groups immunized with either IgG^a or IgG^b anti-SRBC and SRBC had a lower GC response than mice immunized with SRBC alone, but at other time points the GC responses were equally strong in all groups.

Upon maturation, GCs polarize into two anatomically and functionally distinct zones, the DZ and the LZ (27, 28). The GC B-cells in these zones can be phenotypically distinguished: DZ B cells are characterized as CXCR4^lo, CD86^lo, CD83^lo, and LZ B cells as CXCR4^hi, CD86^hi, and CD83^hi (29). To evaluate whether the passively administered IgG anti-SRBC affected the polarization of splenic GCs, spleens from the mice described in Figure 2 were analyzed. DZ and LZ CD95^hiCD38⁻ GC B-cells were gated based on their relative surface expression of CD86 and CD83 as shown in (Figure 3A). Six days after immunization, IgG induced a small but clear shift from LZ to DZ phenotype, seen as a reduction in CD86^hiCD83^hi LZ GC B cells (Figure 3C) and an increase in CD86^loCD83^lo DZ GC B cells (Figure 3B). However, 10 days after immunization, IgG induced a shift from DZ to LZ B cells, seen as an increase in CD86^hiCD83^hi LZ GC B cells (Figure 3C), and a reduction in the CD86^loCD83^lo DZ GC B cells (Figure 3B). Thus, in IgG-suppressed mice, LZ GC B cells increased with time while DZ GC B cells decreased.

During the GC response, two effector subsets of CD4⁺ T cells play an essential role in regulating the response. T follicular helper cells (T_FH) facilitate B cell selection and stimulate antibody production by providing limiting help to cognate GC B cells (27, 30–32). In contrast, T follicular regulatory cells (T_FR) inhibit the GC reaction by suppressing cytokine production by T_FH as well as antibody production and class-switch recombination by B cells (33–35). How T_FR affect antibody responses is not completely understood, but the ratio between T_FH and T_FR seems to be of importance (36). Here, we sought to evaluate whether the proportion of these subsets was altered in IgG-suppressed mice, in spleens obtained from the mice described in Figure 2. All groups had comparable levels of CD4⁺ T cells (Figure 4B). In all three groups, CXCR5⁺PD-1⁺ cells comprised 7–13% of the CD4⁺ T cells (Figure 4C). Of these, slightly more than 90% were identified as T_FH (CD4⁺CXCR5⁺PD-1⁺Foxp3⁻) (Figures 4A,D), and slightly less than 10% as T_FR (CD4⁺CXCR5⁺PD-1⁺Foxp3⁺) (Figures 4A,E). In summary, administration of specific IgG together with SRBC does not alter the T_FH and T_FR proportions from those observed in mice immunized with SRBC alone.

Administration of IgG anti-SRBC together with SRBC causes an almost complete suppression of the SRBC-specific antibody response (Figure 2M) but allows for a potent GC response (Figures 1 and 2). These apparently paradoxical observations raised the question of which antigen induces the GCs. One possibility is that the passively administered IgG antibodies, bound to the SRBC surface, are immunogenic. Indeed, BALB/c mice which had a suppressed SRBC response owing to immunization with IgG^b anti-SRBC together with SRBC (Figure 2M), had a potent IgG^a anti-IgG^b response (Figure 5). No anti-IgG^b response was seen in mice immunized with SRBC or IgG^b anti-SRBC alone (Figure 5). Thus, administration of allogeneic IgG anti-SRBC together with SRBC causes a potent antibody response against IgG (Figure 5) although no antibody response against SRBC is detected in the same sera (Figure 2M).