Amplified B lymphocyte CD40 signaling drives regulatory B10 cell expansion in mice

Abstract

Background: Aberrant CD40 ligand (CD154) expression occurs on both T cells and B cells in human lupus patients, which is suggested to enhance B cell CD40 signaling and play a role in disease pathogenesis. Transgenic mice expressing CD154 by their B cells (CD154(TG)) have an expanded spleen B cell pool and produce autoantibodies (autoAbs). CD22 deficient (CD22(-/-)) mice also produce autoAbs, and importantly, their B cells are hyper-proliferative following CD40 stimulation ex vivo. Combining these 2 genetic alterations in CD154(TG)CD22(-/-) mice was thereby predicted to intensify CD40 signaling and autoimmune disease due to autoreactive B cell expansion and/or activation.

Methodology/principal findings: CD154(TG)CD22(-/-) mice were assessed for their humoral immune responses and for changes in their endogenous lymphocyte subsets. Remarkably, CD154(TG)CD22(-/-) mice were not autoimmune, but instead generated minimal IgG responses against both self and foreign antigens. This paucity in IgG isotype switching occurred despite an expanded spleen B cell pool, higher serum IgM levels, and augmented ex vivo B cell proliferation. Impaired IgG responses in CD154(TG)CD22(-/-) mice were explained by a 16-fold expansion of functional, mature IL-10-competent regulatory spleen B cells (B10 cells: 26.7×10(6)±6 in CD154(TG)CD22(-/-) mice; 1.7×10(6)±0.4 in wild type mice, p<0.01), and an 11-fold expansion of B10 cells combined with their ex vivo-matured progenitors (B10+B10pro cells: 66×10(6)±3 in CD154(TG)CD22(-/-) mice; 6.1×10(6)±2 in wild type mice, p<0.01) that represented 39% of all spleen B cells.

Conclusions/significance: These results demonstrate for the first time that the IL-10-producing B10 B cell subset has the capacity to suppress IgG humoral immune responses against both foreign and self antigens. Thereby, therapeutic agents that drive regulatory B10 cell expansion in vivo may inhibit pathogenic IgG autoAb production in humans.

Figure 1. B cell development in CD154^TGCD22^−/− and CD154^TG mice.

(A) B cell CD154 expression in WT, CD22^−/−, CD154^TG and CD154^TGCD22^−/− mice. Blood and spleen B220⁺ B cells were assessed for CD154 expression by immunofluorescence staining with flow cytometry analysis. Dashed lines delineate borders between CD154⁺ and CD154⁻ cells as determined using WT lymphocytes. (B) Ectopic CD154 expression is B cell-restricted in CD154^TGCD22^−/− mice. CD154 expression by blood and spleen B220⁺ and B220⁻ mononuclear cells was assessed by immunofluorescence staining with flow cytometry analysis. Results represent those obtained in 6 pairs of mice. (C) Increased CD5 expression by B cells from CD22^−/− and CD154^TGCD22^−/− mice. Spleen B cells were assessed for CD5 expression by immunofluorescence staining. Gates indicate percentages of CD5⁺ B cells among total B220⁺ cells. (D) Spleen CD1d^hiB220⁺ B cell localization. Tissue sections from WT, CD22^−/−, CD154^TG, and CD154^TGCD22^−/− mice were stained with B220 (FITC, green) and CD1d (PE, red) mAbs. Merged images highlight CD1d^hiB220⁺ cells (yellow). (A–C) Results are representative of ≥3 mice of each genotype. (E) Spleen CD1d^hiB220⁺ B cell numbers in WT, CD22^−/−, CD154^TG and CD154^TGCD22^−/− mice as determined by immunofluorescence staining. Results represent the mean (±SEM) from ≥3 mice of each genotype. (F) CD154^TGCD22^−/− B cells are hyper-responsive to CD40 signals. Spleen B cells from WT, CD22^−/−, CD154^TG, and CD154^TGCD22^−/− mice were cultured with mitogenic CD40 mAb or anti-IgM Ab for 72 h, with [³H]-thymidine incorporation assessed during the final 18 h of culture. Values represent means (±SEM) of triplicate cultures. (E–F) Means significantly different from WT values are indicated by asterisks (**p<0.01), and between other genotypes by crosses (††p<0.01). (G) CD154^TGCD22^−/− B cell proliferation in response to LPS stimulation. Cell division of CFSE-labeled B cells from WT, CD22^−/−, CD154^TG and CD154^TGCD22^−/− mice following LPS stimulation was quantified after 72 h in culture by flow cytometry. (F–G) Results are representative of 2 independent experiments with similar results. (H) Enhanced survival by CFSE-labeled B cells from CD154^TG and CD154^TGCD22^−/− mice after 10 days in culture without mitogenic stimulation as assessed by flow cytometry. Results represent 3 independent experiments producing similar results.

Figure 2. Impaired IgG and GC responses in CD154^TGCD22^−/− mice.

(A) Serum IgM and IgG levels of 4 and 12 mo-old WT, CD22^−/−, CD154^TG, and CD154^TGCD22^−/− mice. Symbols represent serum concentrations for individual mice as determined by ELISA, with means indicated by horizontal bars. (B) Serum autoAbs reactive with dsDNA, ssDNA, or histone proteins in 12 mo-old mice. ELISA OD values for IgM (upper panels) and IgG (lower panels) autoAbs are shown for individual mice, with means indicated by horizontal bars. Sera from 2 mo-old WT C57BL/6 and 6 mo-old MLRlpr mice were used as negative and positive controls, respectively. (C) Impaired IgG responses to a TD Ag. WT (n = 3), CD22^−/− (n = 4), CD154^TG (n = 6), and CD154^TGCD22^−/− (n = 8) mice were immunized with DNP-KLH in adjuvant on day 0, and boosted on day 21. The graph shows mean (±SEM) DNP-specific IgG levels as determined by ELISA. Images on the right represent immunofluorescence staining of frozen spleen sections from all genotypes harvested 7 days after the boost phase of DNP-KLH immunization. Merged images show the presence of B220⁺ B cells (red) and GC GL7⁺B220⁺ B cells (yellow). Enlarged regions from these sections indicate typical GC structures present within the follicles of WT and CD22^−/− mice, and detectable GL7⁺B220⁺ B cells within the follicles of CD154^TG mice, but not in CD154^TGCD22^−/− mice (representative regions are shown for comparison). (A–C) Means significantly different from WT are indicated by asterisks (*p≤0.05, **p≤0.01), and between other indicated groups by crosses (†p<0.05, ††p<0.01). (D) Reduced GC B cells in CD154^TGCD22^−/− mice as quantified by flow cytometry analysis. Mice were immunized with NP-CGG in alum, with spleens analyzed for GL7⁺B220⁺ B cells on day 10. Contour plots show mean (±SEM) GL7⁺ cell frequencies among total B220⁺ cells from n = 3 mice of each genotype. Bar graphs show mean (±SEM) GL7⁺ B cell numbers from naive (open bars) and immunized (filled bars) mice. In the contour plots, mean B10 cell frequencies significantly lower than those of WT mice are indicated by asterisks (**p≤0.01). In the bar graphs, means significantly different between naïve and immunized mice of the same genotype are indicated by asterisks (*p≤0.05); crosses for CD154^TGCD22^−/− mice indicate that mean GL7⁺ cell numbers were significantly reduced relative to all other genotypes (††p<0.01).

Figure 3. Spleen B10 and B10pro cells expand dramatically in CD154^TGCD22^−/− mice.

(A) Blood, spleen, inguinal LN, and peritoneal cavity B10 cell numbers in WT, CD22^−/−, CD154^TG, and CD154^TGCD22^−/− mice. Leukocytes from the indicated tissues were cultured with LPS, PMA, and ionomycin for 5 h to induce IL-10 production, with monensin included in the cultures to block IL-10 secretion (L+PIM stimulation). The cells were then stained for cell surface CD19 and cytoplasmic IL-10 with flow cytometry analysis. Monensin treated cells served as negative controls for IL-10 expression. Representative contour plots gated on CD19⁺ cells are shown with the percentage of IL-10⁺ cells indicated. Bar graphs represent mean (±SEM) B10 cell frequencies (%) and numbers (#) from ≥3 mice of each genotype. (B) Similar B10 cell surface phenotypes among mouse genotypes. Splenocytes were cultured with L+PIM for 5 h and then stained for cell surface molecules and cytoplasmic IL-10. Histograms indicate cell surface molecule expression by IL-10⁺ (thick lines) and IL-10⁻ (thin lines) B cells. Dashed histograms represent isotype-matched control mAb staining. Results are representative of ≥3 mice of each genotype analyzed. (C) Spleen B10+B10pro cell frequencies and numbers. Splenocytes were cultured with an agonistic CD40 mAb, with L+PIM added during the final 5 h of 48 h cultures. Cytoplasmic IL-10⁺ B cells were identified as in (A). Bar graphs show means (±SEM) from ≥3 mice of each genotype. (A,C) Significant differences from WT mice are indicated: *p<0.05, **p<0.01. Means significantly different between CD154^TGCD22^−/− and CD154^TG mice are indicated: †p<0.05, ††p<0.01.

Figure 4. CD154^TGCD22^−/− B10 cells are regulatory.

(A) B10 cells in CD154^TGCD22^−/− mice are found predominately within the spleen CD1d^hiCD5⁺ B cell population. Splenocytes from WT and CD154^TGCD22^−/− mice were cultured for 5 h with L+PIM to induce IL-10 expression, with the cells analyzed for cell surface CD19, CD1d and CD5, and intracellular IL-10 expression. CD19⁺IL-10⁺ and CD19⁺IL-10⁻ B cells (left panels, gated regions) were further gated to show relative CD1d and CD5 expression (merged dot plots, right panels). Bar graphs show the mean (±SEM) B10 cell frequencies within the indicated populations for ≥4 mice of each genotype based on the gated regions indicated in the merged dot plots. Mean B10 cell frequencies significantly different between cell populations from the same genotype are indicated by asterisks (**p≤0.01), and for the same population between genotypes by crosses (††p<0.01). (B) Analysis of CD154^TGCD22^−/− spleen B cell purity within the CD1d^hiCD5⁺ (B10-rich) and CD1d^loCD5⁻ populations following cell sorting of splenocytes stained for CD19, CD1d and CD5. These cells were subsequently used for the adoptive transfer experiments described in (C). (C) B10 cells from CD154^TGCD22^−/− mice reduce EAE disease severity. Purified CD1d^hiCD5⁺ or CD1d^loCD5⁻ spleen B cells from naïve CD154^TGCD22^−/− mice (B) were either adoptively transferred into WT recipient mice immediately (non-activated) or were cultured with agonistic CD40 mAb for 48 h with LPS added during the final 5 h of culture before transfer. Other recipient mice received PBS alone (Control). One day after cell transfers, EAE was induced by MOG immunization. Values represent mean (±SEM) clinical EAE scores from 5 mice per group, with significant differences from PBS control mice indicated: *p<0.05.

Figure 5. Endogenous B10 cells regulate IgG isotype switching in normal mice.

(A–B) CD22 mAb treatment preferentially depletes CD1d^hiCD5⁺ and B10 cells. Seven days after WT mice were given CD22 or control mAb, splenocytes were isolated and cultured in the presence of L+PIM for 5 h followed by cell surface CD19, CD1d and CD5, and intracellular IL-10 staining, with flow cytometry analysis. IL-10^−/− mice given control mAb were also evaluated for comparison. (A) CD22 mAb treatment preferentially depletes CD1d^hiCD5⁺ B cells. Representative CD1d^hiCD5⁺ B cell frequencies are indicated. (B) CD22 mAb treatment preferentially depletes B10 cells. Representative cytoplasmic IL-10⁺ B cell frequencies among spleen CD19⁺ cells are indicated. Bar graphs show mean (±SEM) frequencies and numbers of CD1d^hiCD5⁺ B cells (A) or IL-10⁺ B10 cells (B) from control (open bars) and CD22 mAb treated (closed bars) mice. (C–D) B10 cells regulate IgG Ab responses. WT mice were given CD22 or control mAb (n = 3 per group) on day 0, and immunized with DNP-KLH without adjuvant on days 0 and 21. (C) Serum DNP-specific Abs were quantified by ELISA. (D) The frequency of B cells secreting DNP-specific IgG was determined by ELISPOT analysis of spleen cells harvested on day 28. (A–D) Significant differences between control and CD22 mAb-treated mice (≥3 mice for each treatment group) are indicated: *p<0.05; **p<0.01.