Presentation of high antigen-dose by splenic B220(lo) B cells fosters a feedback loop between T helper type 2 memory and antibody isotype switching

Abstract

Effective humoral immunity ensues when antigen presentation by B cells culminates in productive cooperation with T lymphocytes. This collaboration, however, remains ill-defined because naive antigen-specific B cells are rare and difficult to track in vivo. Herein, we used a defined transfer model to examine how B lymphocytes, as antigen-presenting cells, shape the development of T-cell memory suitable for generation of relevant antibody responses. Specifically, we examined how B cells presenting different doses of antigen during the initial priming phase shape the development of CD4 T-cell memory and its influence on humoral immunity. The findings indicate that B cells presenting low dose of antigen favour the development of T helper type 1 (Th1) type memory, while those presenting a high antigen dose yielded better Th2 memory cells. The memory Th2 cells supported the production of antibodies by effector B cells and promoted isotype switching to IgG1. Moreover, among the B-cell subsets tested for induction of Th2 memory, the splenic but not peritoneal B220(lo) cells were most effective in sustaining Th2 memory development as well as immunoglobulin isotype switching, and this function involved a tight control by programmed death 1-programmed death ligand 2 interactions.

Keywords: B cells; T/B-cell collaboration; antigen presentation; co-stimulatory molecules.

Figure 1

Schematic representation of the animal model used to investigate the cooperation of memory T‐cell generation and humoral immunity. Splenic CD4⁺ T cells from adult DO11.10 mice are plated with irradiated (3000 × rads) purified BALB/c B cells and the culture is stimulated with ovalbumin peptide (OVAp). The T cells are then transferred into MHC II ^−/− BALB/c mice. After 2–4 months of parking, some of the hosts were given MHC II ⁺ dendritic cells (DCs) (to serve as antigen‐presenting cells), immunized with a suboptimal dose of OVAp in complete Freund's adjuvant (CFA) and used to evaluate memory responses. Other hosts were given MHC II ⁺ DCs and naive B cells (to serve as antibody‐producing cells), immunized with a mixture of OVAp and native OVA (nOVA) in CFA and used to evaluate isotype switching.

Figure 2

Low antigen dose presented by B cells leads to the development of greater T helper type 1 (Th1) effector response. (a) CFSE‐labelled naive DO11.10 CD4⁺ T cells were incubated with naive BALB/c splenic B cells and the culture was stimulated with the indicated amounts of ovalbumin peptide (OVAp) (in μm). The plots show the division pattern of T cells as analysed by CFSE dilution on day 4 of incubation. The data are representative of three independent experiments. (b) shows the per cent of CD3⁺ T cells producing intracellular interferon‐γ (IFN‐γ; left panel) or interleukin‐5 (IL‐5; right panel) from the culture stimulated with low (0·01 μm) and high (0·5 μm) doses of OVAp. Each bar shows the mean ± SE of nine independent experiments. *P < 0·05. IL‐5 was chosen as a signature cytokine for Th2 because IL‐4 is re‐absorbed by the cells and requires blockade of IL‐4 receptor to obtain accurate measurement of the cytokine.

Figure 3

High antigen dose presented by B cells leads to the development of a greater T helper type 2 (Th2) memory response, which supports isotype switching to IgG1. Effector T cells from wild‐type (a) or scid (b) DO11.10TCR transgenic mice that were stimulated with B cells loaded with either low (0·01 μm) or high (0·5 μm) doses of ovalbumin peptide (OVAp) were transferred into MHC II ^−/− hosts and parked for 4 months. Unstimulated (Naive) T cells were also transferred into MHC II ^−/− hosts to serve as control. The mice were then given 1 × 10⁶ dendritic cells [DCs; to serve as antigen‐presenting cells (APCs)] and 24 hr later immunized with a suboptimal dose (20 μg) of OVAp in complete Freund's adjuvant (CFA). Five days post immunization, the lymph node (LN) and spleen (SP) cells were harvested, stimulated with OVAp or the control haemagglutinin (HA) peptide, and memory interferon‐ γ (IFN‐γ) and interleukin‐5 (IL‐5) responses were analysed by ELISA. Each point represents the mean ± SD. The data are representative of three independent experiments. The bar graphs represent optimal cytokine production obtained with 1 μm OVAp stimulation in vitro from each group. (c) Hosts recipient of naive or effector T cells as in (a) were parked for 4 months. The mice were then given DCs (1 × 10⁶ cells per mouse) to serve as APCs and splenic naive B lymphocytes (30 × 10⁶ cells per mouse) to serve as antibody‐producing cells. After 24 hr the hosts were immunized with a suboptimal dose (20 μg) of OVAp (for stimulation of memory T cells) together with 300 μg of native OVA (nOVA) (for induction of antibody response) in CFA. Mice were bled on day 7 after immunization and the isotypes of antibodies specific for nOVA were measured using SBA clonotyping system‐HRP. Each bar represents the mean ± SD of triplicate wells after subtraction of the background obtained with HA peptide. The data are representative of three independent experiments. *P < 0·05, **P < 0·01, and ***P < 0·001.

Figure 4

Expression profile of co‐stimulatory and inhibitory molecules on B and T cells during antigen stimulation. Purified naive BALB/c splenic B and T cells were assessed for expression of activation and inhibitory molecules before and after co‐culture in the presence of either low (0·01 μm) or high (0·5 μm) dose of ovalbumin peptide (OVAp). Marker expression is shown for (a) 7AAD ⁻ CD4⁺ CD3⁺ T cells and for (b) 7AAD ⁻ CD19⁺ B cells. The plots are representative of four independent experiments.

Figure 5

Programmed death 1 (PD‐1) controls the generation of T helper type 2 (Th2) memory through interaction with programmed death ligand 2 (PD‐L2). CD4⁺ T cells from DO11.10 TCR transgenic mice were stimulated with B cells loaded with a high (0·5 μm) dose of ovalbumin peptide (OVAp) in the presence of 10 μg/ml of anti‐PD‐1, anti‐PD‐L1, anti‐PD‐L2 antibody or isotype control for 4 days. After extensive washing to remove the excess of OVAp and blocking antibodies, the cells were then transferred into MHC II ^−/− hosts and parked for 2 months. The mice were then given 1 × 10⁶ dendritic cells (DCs) (to serve as antigen‐presenting cells) and 24 hr later were immunized with a suboptimal dose (20 μg) of OVAp in complete Freund's adjuvant. Five days post immunization, the lymph node (LN) and spleen (SP) cells were harvested, stimulated with 10 μm OVAp or the control haemagglutinin (HA) peptide and memory (a) interferon‐γ (IFN‐γ) and (b) interleukin‐5 (IL‐5) responses were analysed by ELISA. Each bar represents the mean ± SD. The data are representative of two independent experiments. *P < 0·05.

Figure 6

B220^lo B cells presenting high‐dose ovalbumin peptide (OVAp) drive the development of T helper type 2 (Th2) memory. (a) Naive BALB/c splenic B cells were isolated on anti‐CD19 antibody coupled microbeads and stained with anti‐B220, anti‐CD21 and anti‐CD23 antibodies (left panel). The cells were then sorted as B220^lo CD21⁻ CD23⁻ (B220^lo), B220^hi CD21^int CD23⁺ (FO), B220^hi CD21^hi CD23^int (MZ), or B220^hi CD21⁻ CD23⁻ (NF) (right panel). The four subsets were then loaded with high‐dose OVAp (0·5 μm) and incubated with naive DO11.10 T cells. After 4 days of culture, the cells were separated from the supernatant and transferred into MHC II ^−/− mice for parking for analyses of memory responses. The supernatant, however, was used to measure in vitro primary interferon‐γ (IFN γ) and interleukin‐5 (IL‐5) cytokine responses as shown in (b). Each bar represents the mean ± SE of four independent experiments. (c) After 2‐month parking, the mice recipient of T cells were given 1 × 10⁶ bulk dendritic cells (DCs) and 24 hr later were immunized with a suboptimal dose (20 μg/mouse) of OVAp in complete Freund's adjuvant (CFA). Mice that received unstimulated T cells were used as controls (Nil). Five days later the lymph node (LN) and spleen (SP) cells were harvested and stimulated with 10 μm OVAp or the control haemagglutinin (HA) peptide and memory IFN‐γ and IL‐5 responses were determined by ELISA. The bars represent the amounts of cytokines after subtraction of the background obtained with the control HA peptide. Each bar represents the mean ± SEM of four independent experiments. *P < 0·05.

Figure 7

B220^lo but not peritoneal B cells drive differentiation of T helper type 2 (Th2) cells despite displaying similar B1‐like phenotypes. (a) Splenic and peritoneal B cells were stained for CD19, B220, 7‐AAD and one of the indicated markers simultaneously. Expression of the markers was then analysed on CD19⁺ 7‐AAD ⁻ B220^lo or CD19⁺ 7‐AAD ⁻B220^hi cells. The numbers indicate the mean fluorescence intensity (MFI) of the markers after subtraction of the isotype. The results are representative of three independent experiments. (b) Sorted B220^lo, newly formed (NF), and peritoneal B‐cell subsets (5 × 10⁵ cells/well) were loaded with high‐dose ovalbumin peptide (OVAp) (0·5 μm) and incubated with DO11.10 T cells (5 × 10⁵ cells/well) for 4 days. Subsequently, the culture was stained for surface CD3 and intracellular interferon‐γ (IFN‐γ) and interleukin‐5 (IL‐5). The percentages of IFN γ‐ and IL‐5‐producing CD3⁺ T cells were determined by flow cytometry.

Figure 8

Splenic but not peritoneal B220^lo B cells are capable of inducing T helper type 2 (Th2) memory that fosters isotype switching to IgG1. CD4⁺ T cells from DO11.10 mice were stimulated by sorted splenic or peritoneal B220^lo B cells loaded with high dose of ovalbumin peptide (OVAp). Effector T cells were then transferred into MHC II ^−/− hosts and parked for 2 months to test for memory T‐cell responses (a) or isotype switching (b). Unstimulated (Naive) T cells were also transferred into MHC II ^−/− hosts to serve as controls. (a) For analysis of memory T‐cell responses, the mice were given 1 × 10⁶ dendritic cells (DCs) [to serve as antigen‐presenting cells (APCs)] and 24 hr later immunized with 20 μg of OVAp in complete Freund's adjuvant (CFA). Five days post immunization, the lymph node (LN) and spleen (SP) cells were harvested, stimulated with OVAp or the control hameagglutinin (HA) peptide, and memory interferon‐γ (IFN‐γ; Th1) and interleukin‐5 (IL‐5; Th2) responses were analysed by ELISA. Each bar represents the mean ± SD. The data are representative of two independent experiments. (b) For analysis of isotype switching other hosts were given dendritic cells (DCs; 1 × 10⁶ cells per mouse) to serve as antigen‐presenting cells and splenic naive B lymphocytes (30 × 10⁶ cells per mouse) to serve as antibody‐producing cells. After 24 hr the hosts were immunized with 20 μg of OVAp (for stimulation of memory T cells) together with 300 μg of native OVA (nOVA) protein (for induction of antibody response) in CFA. Mice were bled on day 7 and the isotype as well as total nOVA‐specific antibodies were measured. The data are representative of two independent experiments. *P < 0·05, **P < 0·01, ***P < 0·001.