Critical Role of SAP in Progression and Reactivation but Not Maintenance of T Cell-Dependent Humoral Immunity

Abstract

Signaling lymphocytic activation molecule (SLAM)-associated protein (SAP) is a small adaptor molecule mutated in X-linked lymphoproliferative disease, a human immunodeficiency. SAP plays a critical role in the initiation of T cell-dependent B cell responses leading to germinal center reaction, the production of high-affinity antibodies, and B cell memory. However, whether SAP has a role in these responses beyond their initiation is not known. It is important to address this matter not only for mechanistic reasons but also because blockade of the SAP pathway is being contemplated as a means to treat autoimmune diseases in humans. Using an inducibly SAP deficient mouse, we found that SAP was required not only for the initiation but also for the progression of primary T cell-driven B cell responses to haptens. It was also necessary for the reactivation of T cell-dependent B cell immunity during secondary immune responses. These activities consistently correlated with the requirement of SAP for full expression of the lineage commitment factor Bcl-6 in follicular T helper (T_FH) cells. However, once memory B cells and long-lived antibody-secreting cells were established, SAP became dispensable for maintaining T cell-dependent B cell responses. Thus, SAP is pivotal for nearly all phases, but not for maintenance, of T cell-driven B cell humoral immunity. These findings may have implications for the treatment of immune disorders by targeting the SAP pathway.

INTRODUCTION

Signaling lymphocytic activation molecule (SLAM)-associated protein (SAP; also known as SH2D1A) is a Src homology 2 (SH2) domain-only intracellular adaptor expressed in T cells, natural killer (NK) cells, and some transformed B cells (1–3). It does not appear to be expressed in normal B cells, including germinal center (GC) B cells (4). SAP is mutated in X-linked lymphoproliferative (XLP) disease, a human immunodeficiency. Studies of immune cells from XLP patients and genetically engineered SAP-deficient mice have shown that SAP plays a critical role in multiple immune cell functions, including follicular T helper (T_FH) cell polarization, T cell-dependent antibody production, memory B cell generation, T helper 2 (T_H2) cytokine production, NK-T cell development, CD8⁺ T cell-mediated cytotoxicity, and NK cell-mediated cytotoxicity. These functions reflect the ability of SAP to control the signals emanating from SLAM family receptors, a group of self-associating immune cell-specific receptors. Most of the functions of SAP are dependent on its capacity to bind and activate the Src-related protein tyrosine kinase Fyn (5–10). However, this is not the case for T_FH cell functions, which are largely Fyn independent (10–12).

T cell-dependent B cell immunity leads to the generation of high-affinity antibodies, memory B cells, and long-lived antibody-secreting cells (ASCs) against protein antigens (13). These responses are crucial for protection against many pathogens and for responsiveness to vaccination. When excessive, they can lead to autoimmune diseases. Accumulating evidence indicates that T cell-dependent B cell responses are mediated largely by the ability of a subset of CD4⁺ T cells, the T_FH cells, to initiate GC reactions in lymphoid follicles (14–19). When contacted by antigen-specific T_FH cells, GC B cells sharing the same antigen specificity as the T cells undergo maturation, isotype switching, and somatic hypermutation. These modifications enable B cells to produce high-affinity antibodies against the antigen. GC B cells also differentiate into memory B cells and long-lived ASCs, which provide long-term immunity. Once antigen exposure is resolved, some T_FH cells can persist as memory T_FH cells, which are reactivated upon secondary exposure to an antigen and are more efficient at initiating secondary B cell responses (20–22).

SAP is essential for GC reaction and T cell-dependent antibody production (11, 23, 24). It appears to enable these processes by stabilizing the formation of a conjugate between antigen-specific T_FH cells and GC B cells. In a previous study using a conditionally SAP deficient mouse, we showed that this was due to a role of SAP in T cells, not in B cells (4). This activity is also mediated by the SLAM family receptors Ly108 and CD84, which are expressed both on T_FH cells and on GC B cells. Adoptive transfer experiments showed that SAP is not needed for early T_FH cell differentiation, which depends primarily on the induced T cell costimulator (ICOS) (22, 25–27). Rather, SAP acts at a later stage of T_FH cell polarization. A recent report using a viral infection model showed that SAP enables T_FH cells to express full amounts of B cell lymphoma 6 (Bcl-6), a lineage commitment factor necessary for T_FH cell functions (25). Bcl-6 is also highly expressed in GC B cells, and this expression is a prerequisite for GC B cell differentiation.

Key issues remain to be addressed regarding the role of SAP in T cell-dependent B cell immunity. While analyses of constitutively SAP deficient mice have indicated that SAP expression in T_FH cells is required for the initiation of normal T cell-dependent B cell immunity, these studies did not address the question of whether SAP also plays a role in the progression or maintenance of these responses. Moreover, it is not established whether SAP is required for the reactivation of previously generated memory T_FH cells in the context of secondary responses to an antigen. Here we address these issues using a broad range of immunization and genetic approaches. Our studies document that SAP is necessary not only for the initiation, but also for the progression, of T cell-dependent B cell responses to a hapten. SAP is also required for the reinduction of T_FH cell polarization, as well as for the reexpansion of GC B cells and memory B cells, during secondary responses to the hapten. In contrast, SAP is not needed for the maintenance of memory B cells, long-lived ASCs, and antibodies once these T cell-dependent responses are established.

MATERIALS AND METHODS

Mice.

C57BL/6 mice were obtained from Harlan Sprague Dawley (Montreal, QC, Canada). SAP-deficient mice (Sh2d1a^−/), conditionally SAP deficient mice (Sh2d1a^fl/), and mice expressing a tamoxifen (TAM)-inducible fusion protein between the human estrogen receptor (ER) and Cre recombinase (Cre-ERT2) under the control of the ubiquitin C (UBC) promoter (Ubc-CreERT2) have been described previously (4, 8, 28–31). These mice were maintained in the C57BL/6 background. For mice expressing constitutive or inducible Cre, hemizygous Cre-positive Sh2d1a^fl/ males were used for experimentation. Control mice were usually Cre-positive Sh2d1a^+/ mice. In a few cases, Cre-negative Sh2d1a^fl/ or Sh2d1a^+/ mice were used. No differences in immune responses were observed between the various controls. Animal experimentation was carried out according to the guidelines of the Canadian Council of Animal Care and was approved by the Animal Care Committee of the Clinical Research Institute of Montreal. All mice were maintained in a specific-pathogen-free animal facility.

Tamoxifen-induced deletion of SAP.

Mice were fed TAM (200 μg/g of body weight; product no. T5648; Sigma-Aldrich, St. Louis, MO) by gavage daily for 5 consecutive days. Ten days after the first feeding, 50 μl of blood was obtained, and SAP expression was analyzed by flow cytometry of peripheral blood lymphocytes using an anti-SAP monoclonal antibody, as described previously (4).

Hapten immunization.

Mice were injected with a 200-μl suspension containing a 50% volume of a hapten (at 1 mg/ml) and a 50% volume of alum (catalog no. 77161; Thermo Scientific, Rockford, IL). Two haptens were used: nitrophenyl (NP)(13)-ovalbumin (OVA) (catalog no. N-5051-100; Biosearch Technology, CA) and phosphorylcholine (PC)(3)-OVA (catalog no. PC-5051-10; Biosearch Technology). Experiments were typically performed at least 3 times, and data from all experiments were pooled to generate the figures.

ELISA.

Levels of antigen-specific antibodies in serum were measured by enzyme-linked immunosorbent assays (ELISA), as described previously (4). To detect hapten-specific Ig, NP(4)-bovine serum albumin (BSA) (catalog no. N-5050-10; Biosearch Technology) or PC(5)-BSA (catalog no. PC-1011-10; Biosearch Technology) was used as the coating antigen. High-affinity antigen-specific antibodies were detected as described earlier (32). Briefly, after the incubation of antigen-coated plates with diluted serum, plates were washed twice with phosphate-buffered saline (PBS) and were incubated at room temperature for 15 min with 150 μl of the chaotropic agent sodium thiocyanate (NaSCN; 1.0 or 1.5 M). After five additional washes with PBS, plates were processed for ELISA.

Enzyme-linked immunosorbent spot (ELISPOT) assays.

MultiScreen IP filter plates (MSIPS4W10; Cedarlane, Burlington, ON, Canada) were precoated overnight with 100 μl of NP(4)-BSA (at 2 μg/ml) and were then blocked in the presence of RPMI 1640 medium supplemented with 10% fetal bovine serum. Spleen cells or bone marrow cells from immunized or control nonimmunized mice were then seeded in a complete culture medium onto the plate at 1 × 10⁵, 2 × 10⁵, or 4 × 10⁵ cells/well and were incubated overnight at 37°C. After extensive washing, wells were incubated for 2 h at room temperature with horseradish peroxidase (HRP)-conjugated anti-mouse IgG1. After washing, 100 μl of the substrate 3-amino-9-ethylcarbazole (AEC) (catalog no. 551951; BD Biosciences, Mississauga, ON, Canada) was added to each well, and filters were incubated for 20 min. The spots, which represent cells secreting NP-specific IgG1, were counted using a microscope.

Flow cytometry.

Anti-CD3 (monoclonal antibody [MAb] 145-2C11), anti-CD4 (MAb RM4.5), anti-CD19 (MAb MB19-1), anti-CD45R (MAb RA3-6B2), anti-CD49b (MAb DX5), anti-CD5 (MAb 53-7.3), anti-CD38 (MAb 90), anti-CD90.2 (MAb 30-H12), anti-CD93 (MAb AA4.1), anti-NK1.1 (MAb PK136), anti-Ly6G (MAb RB6-8C5), anti-F4/80 (MAb BM8), anti-TER119 (MAb TER119), anti-IgM (MAb eB121-15F9), anti-IgD (MAb 11-26C), streptavidin-eFluor 450, and isotype-matched control antibodies were obtained from eBioscience (San Diego, CA). Anti-T- and B-cell activation antigen (MAb GL-7), anti-CXCR5 (MAb 2G8), anti-Bcl-6 (MAb K112-91), and anti-mouse IgG1 (MAb A85-1) were from BD Biosciences. A monoclonal antibody directed against SAP (MAb 1A9) was generated in our laboratory (4). 4-Hydroxy-3-iodo-5-nitrophenyl acetyl (NIP)(5)-BSA (catalog no. N-5040-10; Biosearch Technology) was coupled to Alexa Fluor 647 by using the Alexa Fluor 647 protein labeling kit (catalog no. A20173; Life Technologies Inc., Burlington, ON, Canada). This conjugate was then used to detect NP-specific B cells by flow cytometry. Most flow cytometry data were collected using a CyAn ADP flow cytometer (Beckman Coulter, Mississauga, ON, Canada) and were analyzed with FlowJo (Ashland, OR). In some cases, data were collected with a FACSCalibur flow cytometer (BD Biosciences) and were analyzed with CellQuest (BD Biosciences). NIP-specific GC and memory B cells were stained and analyzed as described previously (33, 34). Bcl-6 expression was analyzed as outlined elsewhere (25).

Statistical analyses.

Data were analyzed using the Student t test. P values equal to or less than 0.05 were considered to be statistically significant.

RESULTS

Generation and characterization of an inducibly SAP deficient mouse.

To ascertain the role of SAP in the progression, reactivation, and maintenance of T cell-dependent B cell responses, a mouse carrying a conditional (floxed) allele of the SAP-encoding gene (Sh2d1a^fl/ mouse) was bred with a mouse expressing tamoxifen (TAM)-inducible Cre (UBC-Cre-ERT2 mouse) (4, 29) (Fig. 1A). Whereas Cre-ERT2 is normally cytosolic and inactive, it becomes active and translocates to the nucleus upon treatment with TAM. This enables induced removal of SAP expression. Since the SAP-encoding gene is X-linked, hemizygous Sh2d1a^fl/ males, as well as wild-type (WT) control male littermates, were used for experimentation.

Fig 1.

Generation and characterization of inducibly SAP deficient mice. (A) Mice expressing the conditional (“inducible”) allele of the SAP-encoding gene (Sh2d1a^fl/) were bred with mice expressing the tamoxifen (TAM)-responsive Cre recombinase under the control of the ubiquitin C promoter (UBC-creERT2). The transgene used to generate the UBC-creERT2 mice contained a cytomegalovirus (CMV) enhancer, a human UBC promoter, a gene encoding a fusion protein consisting of Cre recombinase and a mutant form of the human estrogen receptor (Cre-ERT2) that can be activated only in the presence of TAM, and a woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) to enhance mRNA stability. Treatment of mice with TAM results in Cre-mediated deletion of exon 2 of the SAP-encoding gene, causing loss of SAP protein expression. (B) Sh2d1a^fl/; UBC-creERT2 mice (inducible knockout [iKO] mice) were treated with TAM for 5 consecutive days to delete SAP. SAP expression was then determined on the indicated days using intracellular staining with anti-SAP antibodies and flow cytometry. (C) The expression of SAP in CD3⁺ CD4⁺ peripheral blood lymphocytes on the indicated days (D) after the start of TAM treatment is shown for representative mice. Green lines, conventional SAP-deficient mouse; red lines, TAM-treated inducibly SAP deficient mouse; blue lines, control mouse. (D) The mean fluorescence intensity (MFI) of SAP expression is shown for multiple mice at the indicated times after the start of TAM treatment. Time 0 represents control mice. Symbols represent values for individual mice, and horizontal bars indicate the means. (E) The expression of SAP at various times was examined by intracellular staining and flow cytometry using peripheral blood lymphocytes gated on CD3⁺ CD4⁺ cells. Purple, SAP-deficient mouse; green, mice with the indicated genotypes. Mice were or were not fed TAM. The number of days (D) after the start of TAM treatment is given at the top right. (F) The expression of SAP was determined 10 days after the start of TAM treatment. Percentages of SAP-negative CD3⁺ CD4⁺ cells in individual mice are shown. Only mice with >95% SAP-negative cells were used for functional studies. PBLs, peripheral blood lymphocytes. (G) The efficiency of SAP deletion in various subsets of CD4⁺ T cells from the spleen after immunization with NP-OVA is shown for representative mice. CXCR5⁺ PD-1⁺ cells correspond to T_FH cells. Green lines, constitutively SAP deficient mouse; red lines, TAM-treated inducibly SAP deficient mouse; blue lines, control mouse. (H) Efficiency of SAP deletion in T_FH cells and non-T_FH CD4⁺ T cells. Results are shown for 3 mice immunized with NP-OVA. The percentages of remaining SAP-positive cells are shown.

Feeding TAM to Sh2d1a^fl/; UBC-CreERT2 mice for 5 consecutive days resulted in a time-dependent loss of SAP expression, as judged by intracellular staining of peripheral blood CD4⁺ T cells with anti-SAP antibodies (Fig. 1B to D). Loss of SAP expression began within 1 day of TAM treatment and was maximal after 7 days. Ablation of SAP was stable for as long as 100 days after TAM treatment (Fig. 1E; also data not shown). Under these conditions, >95% of CD4⁺ T cells became SAP deficient (Fig. 1F). This was verified for all mice used for experimentation. Mice showing lower efficiencies of deletion were not used. In contrast, feeding TAM to WT mice with had no impact on SAP expression (Fig. 1E).

We also ensured that SAP was efficiently deleted in various subpopulations of CD4⁺ T cells, in particular T_FH cells (Fig. 1G and H). To this end, CD4⁺ T cells were stained with antibodies against CXCR5 and PD-1, two markers highly expressed on T_FH cells. T_FH cells are typically defined as CD4⁺ CXCR5⁺ PD-1⁺ cells. Efficient and equal deletion of SAP was observed in all gated subsets of CD4⁺ T cells, including CD4⁺ CXCR5⁺ PD-1⁺ cells.

Thus, this model provided an efficient and reliable way to ablate SAP expression inducibly in T cells, including T_FH cells.

SAP is necessary for the progression of T cell-dependent B cell responses to a hapten.

To address whether SAP is required for the progression of T cell-dependent B cell responses and GC reaction, we used a hapten immunization protocol in which SAP has been shown to be critical for the initiation of the responses (4, 10, 23). Sh2d1a^fl/; UBC-CreERT2 mice were immunized with the T cell-dependent antigen nitrophenyl (NP)-ovalbumin (OVA) prior to treatment with TAM (Fig. 2A). In this system, the ability of B cells to respond to the hapten, NP, is dependent on a T_FH cell response to the antigen, OVA. Then, at day 4, when the induction of T_FH cell differentiation by the antigen was well under way (17, 25, 26), SAP expression was eliminated by TAM treatment.

Fig 2.

SAP expression during the ongoing GC reaction is required for maximal antibody production. (A) Sh2d1a^fl/; UBC-creERT2 mice (iKO) were immunized with NP-OVA. Four days later, they were treated with TAM to delete SAP. Analyses were performed 14 or 28 days later. (B and C) NP-specific IgM and IgG1 were measured at day 14 and day 28. Symbols show values for individual mice, whereas horizontal lines represent the means. CTL, control mice. (D) High-affinity anti-NP IgG1 was measured by performing ELISA in the presence of 1.5 M NaSCN.

Elimination of SAP during an ongoing T cell response resulted in a ∼50-to-60% reduction of the levels of anti-NP IgM and IgG1, which were measured at day 14 and day 28 after immunization (Fig. 2B and C). This was also the case for the abundance of high-affinity anti-NP IgG1, which was measured by performing ELISA in the presence of the chaotropic agent NaSCN (Fig. 2D). Notably, however, the magnitude of the defect caused by SAP deficiency was not more marked when high-affinity antibodies were measured, suggesting that a lack of SAP affected the production, but not the affinity maturation, of anti-NP antibodies.

Previous studies have shown that the antibody deficiency seen in constitutively SAP deficient mice is due to a defect in the maturation of T_FH cells (11, 23, 24). To address the impact of the loss of SAP during an ongoing T cell response on T_FH cells, we first quantitated total T_FH cells in immunized mice (Fig. 3A). There was no difference in the total number of T_FH cells, identified as CD4⁺ CXCR5⁺ PD-1⁺ cells, between control and SAP-deficient mice.

Fig 3.

SAP expression during an ongoing GC reaction is required for maximal accumulation of Bcl-6^hi GC T_FH cells. The experiment was performed as described for Fig. 2. (A) Gating protocol and abundance of total T_FH cells (CD4⁺ CXCR5⁺ PD-1⁺ cells) in the spleen. The experiment was conducted at day 14 after immunization. Symbols show values for individual mice, whereas horizontal lines represent the means. SS:Lin, side scatter:linear. (B) Gating protocol and abundances of GL-7⁻ and GL-7⁺ T_FH cells. (C) The expression of Bcl-6 in the indicated populations was determined by intracellular staining and flow cytometry. CTL, control mice. (D) The increase in the mean fluorescence intensity (ΔMFI) of Bcl-6 was examined for total, GL-7⁻, and GL-7⁺ T_FH cells. (E) Absolute numbers of Bcl-6^hi GL-7⁺ T_FH cells. Bcl-6^hi was defined as an MFI above 200. SPC, spleen cells.

An earlier study using protein immunization and viral infection models showed that a large number of cells with T_FH cell markers do not reside in GCs (27). In an attempt to identify specific markers for T_FH cells located in GCs (so-called GC T_FH cells), these authors found that GC T_FH cells are positive for the activation marker GL-7, whereas other T_FH cells are not. Thus, the impact of induced SAP deficiency on GC T_FH cells was also examined (Fig. 3B). Although loss of SAP had no effect on the abundance of GL-7⁻ T_FH cells, it caused a significant reduction (by ∼50%) of the accumulation of GL-7⁺ T_FH cells. This finding implied that loss of SAP during an ongoing GC reaction specifically affected the accumulation of GC T_FH cells.

We also tested the impact of induced SAP deficiency on the expression levels of the lineage commitment factor Bcl-6 in T_FH cells (Fig. 3C and D). Using intracellular staining for Bcl-6, we found that the levels of Bcl-6 in total T_FH cells and GL-7⁻ T_FH cells were not affected by loss of SAP expression. However, the presence of Bcl-6^hi T_FH cells, which were detected only among GL-7⁺ T_FH cells, was compromised in SAP-deficient mice. These cells represent the most polarized subset of T_FH cells (25). Enumeration of Bcl-6^hi GC T_FH cells showed that this subpopulation was diminished by ∼65% in mice depleted of SAP (Fig. 3E).

In these experiments, we also noted that SAP-depleted mice exhibited a diminution of the number of Bcl-6^hi GL-7⁺ cells lacking the CD4 marker (Fig. 4A). Further studies showed that these cells were positive for the B cell antigen CD19, indicating that they were B cells (Fig. 4B). This Bcl-6^hi GL-7⁺ B cell population was absent in constitutively SAP deficient mice (25) (Fig. 4B). Thus, these cells likely represented GC B cells, which are known to express high levels of Bcl-6 and are absent in SAP-deficient mice. More importantly, the number of these cells was reduced by ∼80% in mice with inducible deletion of SAP (Fig. 4C).

Fig 4.

SAP expression during an ongoing GC reaction is required for maximal accumulation of GC B cells. (A) Gating protocol for CD4⁻ GL-7⁺ Bcl-6^hi cells. SS:Lin, side scatter:linear. (B) Control mice (CTL) and constitutively SAP deficient mice (KO) were immunized and analyzed as described for Fig. 2 and 3. Results for a pair of mice, representative of three independent pairs tested, is shown. Splenocytes were stained as for the experiment for which results are shown in Fig. 3B and B; they were also stained for CD19. All CD4⁻ GL-7⁺ Bcl-6^hi cells were CD19⁺, indicating that they were B cells. (C) Absolute numbers of Bcl-6^hi GC B cells. Symbols show values for individual mice, and horizontal bars indicate the means. SPC, spleen cells.

These data showed that sustained SAP expression during an ongoing GC reaction was needed for full expansion of Bcl-6^hi GC T_FH cells and GC B cells, as well as for maximal antibody production, in response to a hapten. However, SAP expression did not seem to be required for the affinity maturation of the antibodies.

SAP is needed for the reexpansion of fully polarized T_FH cells and antigen-specific B cells.

Next, we ascertained whether SAP was necessary for the reinduction of T_FH cell polarization and the expansion of antigen-specific B cells during secondary immune responses (Fig. 5). To this end, Sh2d1a^fl/; UBC-CreERT2 mice were immunized with NP-OVA, and at day 29, SAP was eliminated by treatment with TAM (Fig. 5A). After 140 days, mice were immunized again with NP-OVA. Seven days later, the accumulation of T_FH cells was quantitated (Fig. 5B). This time point was chosen because it is within the time frame needed for maximal expansion of T_FH cells in response to immunization. Removal of SAP had no effect on the number of GL-7⁻ T_FH cells or GC T_FH cells after secondary immunization (Fig. 5B). However, it caused a reduction in the levels of Bcl-6 in GC T_FH cells (Fig. 5C and D) and, consequently, a diminution (by ∼75%) of the abundance of Bcl-6^hi GC T_FH cells (Fig. 5E).

Fig 5.

SAP is required for the reinduction of Bcl-6^hi GC T_FH cells after secondary immunization. (A) Sh2d1a^fl/; UBC-creERT2 mice (iKO) were immunized with NP-OVA in the presence of alum. After 29 days, SAP was deleted by treatment of mice with TAM. Then, 140 days after immunization, mice were reimmunized with NP-OVA. Mice were analyzed 7 days later. (B) T_FH cells (GL-7⁻ and GL-7⁺) were quantitated in the spleen 7 days after secondary immunization. GL-7⁺ T_FH cells represent GC T_FH cells. CTL, control mice; SPC, spleen cells. (C) The expression of Bcl-6 in the indicated T_FH cell subpopulations was determined by intracellular staining and flow cytometry. Results for representative mice are shown. (D) The increase in the mean fluorescence intensity (ΔMFI) of Bcl-6 was examined for GL-7⁻ and GL-7⁺ T_FH cells. Symbols show values for individual mice, and horizontal bars indicate the means. (E) Absolute numbers of Bcl-6^hi GL-7⁺ T_FH cells. Bcl-6^hi was defined as an MFI above 200.

To address the impact of SAP deficiency on the reexpansion of antigen-specific B cells, antigen-specific B cells were quantitated 14 days after the second immunization (Fig. 6). This later time point is most appropriate for monitoring B cell expansion. NP-specific B cells were detected by staining with 4-hydroxy-3-iodo-5-nitrophenyl acetyl (NIP), which binds NP-specific antigen receptors (33, 34). The abundance of NP-specific GC B cells, which were identified as NIP⁺ B220⁺ IgG1⁺ CD38⁻ cells, was significantly reduced (by ∼75%) in mice depleted of SAP (Fig. 6A and B). This was also the case for cells displaying markers of memory B cells, which were identified as NIP⁺ B220⁺ IgG1⁺ CD38⁺ cells. Surprisingly, however, the reinduction of anti-NP antibodies, including high-affinity antibodies, was not affected (Fig. 6C and D). The latter finding suggested that boosting of the antibody response by secondary immunization was SAP independent, perhaps because it was secured by long-lived ASCs and memory B cells and was T_FH cell independent (see below).

Fig 6.

SAP is required for the reinduction of antigen-specific GC B cells after secondary immunization. Mice were immunized as described for Fig. 5 and were analyzed 14 days later. (A) Scheme for the identification of NP-specific GC B cells and memory B cells (Mem. B). NP-specific GC cells are NIP⁺ B220⁺ IgG1⁺ CD38⁻, whereas NP-specific memory B cells are NIP⁺ B220⁺ IgG1⁺ CD38⁺. Results for representative mice are shown. CTL, control mice; iKO, Sh2d1a^fl/;UBC-creERT2 mice. (B) NP-specific GC B cells and memory B cells were quantitated in several individual mice. Symbols show values for individual mice, and horizontal bars indicate the means. SPC, spleen cells. (C and D) Levels of NP-specific IgG1 (C) and high-affinity NP-specific IgG1 (D) were measured before (2D0) or 14 days after (2D14) secondary immunization.

Thus, SAP is needed for the reinduction of Bcl-6^hi GC T_FH cells, antigen-specific GC B cells, and memory B cells during secondary immunization with a hapten.

SAP is necessary for T cell-dependent B cell immunity against a second hapten.

The findings shown in Fig. 5 and 6 suggested that SAP was needed for the repolarization of T_FH cells and for their ability to drive recall T cell-dependent B cell responses. To better ascertain the facts, a variant of the experiment for which results are shown in those figures was performed (Fig. 7). Sh2d1a^fl/; UBC-CreERT2 mice were first immunized with phosphorylcholine (PC)-OVA (Fig. 7A). In this protocol, PC is the hapten. After 29 days, SAP was eliminated as described for Fig. 5. Then, at day 70, mice were immunized with a different hapten, NP, also coupled to OVA. The production of antibodies against the two haptens, PC and NP, was monitored over time (Fig. 7B to D).

Fig 7.

SAP is required for the generation of antibodies against a second hapten. (A) Sh2d1a^fl/; UBC-creERT2 mice (iKO) were immunized with PC-OVA in the presence of alum. After 29 days, SAP was deleted by treatment of mice with TAM. Then, at day 70, mice were immunized with NP-OVA. CTL, control mice. (B to D) PC-specific and NP-specific IgG1 (B), IgM (C), and IgE (D) were measured at the indicated times. The time of immunization with PC-OVA is indicated by a black arrow, whereas the times of TAM treatment and immunization with NP-OVA are indicated by red and blue arrows, respectively. Means and standard deviations are shown. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (E) Absolute numbers of NP-specific GC B cells and memory B cells in the spleen were measured 14 days after the second immunization. NP-specific GC B cells are B220⁺ NIP⁺ IgG1⁺ CD38⁻, whereas NP-specific memory B cells are B220⁺ NIP⁺ IgG1⁺ CD38⁺. Symbols show values for individual mice. Horizontal bars indicate the means. SPC, spleen cells.

In agreement with the data in Fig. 6, loss of SAP had no influence on the persistence of antibodies against the first hapten, PC (Fig. 7B to D). However, it severely compromised the ability to produce antibodies against the second hapten, NP. Seven days after immunization with NP-OVA, the production of anti-NP antibodies was diminished by ∼95%, ∼99%, and ∼50% for IgG1, IgE, and IgM, respectively. This was accompanied by a reduction in the numbers of NP-specific GC B cells and memory B cells, which were quantitated 14 days after immunization with the second hapten (Fig. 7E).

These data implied that SAP was required for the ability of previously generated antigen-specific T_FH cells to trigger B cell responses against a new hapten coupled to an already encountered antigen.

SAP is not required for maintaining long-term T cell-dependent B cell immunity.

The data shown in Fig. 5 and 6 also indicated that SAP was not required for the persistence of antibodies against haptens once these antibodies were induced. This suggested that SAP was not necessary for the maintenance of already established T cell-driven B cell responses. To address this idea more clearly, Sh2d1a^fl/; UBC-CreERT2 mice were immunized with NP-OVA and, 29 days after immunization, were fed TAM to eliminate SAP (Fig. 8A). Their ability to maintain anti-NP antibodies in the absence of any additional immunization was then examined over time (Fig. 8B). Removal of SAP had no impact on the persistence of antibodies against NP. This was true for as long as 140 days after immunization. At day 140, SAP deficiency also had no effect on the presence of NP-specific IgG1-producing ASCs in the spleen and bone marrow (Fig. 8C and D). In addition, it had no influence on the persistence of NP-specific memory B cells (Fig. 8E and F).

Fig 8.

SAP is not required for the maintenance of antigen-specific antibodies, memory B cells, or long-lived antibody-secreting cells. (A) Sh2d1a^fl/; UBC-creERT2 mice (iKO) were immunized with NP-OVA in the presence of alum. After 29 days, SAP was deleted by treatment of mice with TAM. CTL, control mice. (B) NP-specific IgG1 was measured at the indicated times after immunization. The red arrow indicates the time of treatment with TAM. (C and D) The absolute numbers of anti-NP antibody-secreting cells (ASCs) in the spleen (C) and bone marrow (D) were determined 140 days after immunization (112 days after SAP deletion). Symbols show values for individual mice. Horizontal bars indicate the means. SPC, spleen cells; BMC, bone marrow cells. (E) Scheme for identification of NP-specific memory B cells in the spleen. NP-specific memory B cells are B220⁺ NIP⁺ IgG1⁺ CD38⁺. (F) The absolute numbers of NP-specific memory B cells in the spleen were determined 140 days after immunization (112 days after SAP deletion).

Combined with the results presented above, these data provided evidence that SAP was not required for long-term maintenance of antibodies, antigen-specific memory B cells, or ASCs after immunization with a T cell-dependent antigen. It is unlikely that this was due to incomplete deletion of SAP, given that the similar experimental protocol used in Fig. 5 and 6 yielded functional defects in de novo T cell-dependent responses.

DISCUSSION

SAP plays a critical role in the initiation of GC reactions leading to antibody production, as well as in the generation of memory B cells and long-lived plasma cells, in response to T cell-dependent protein antigens (1–3). By use of a tissue-specific SAP-deficient mouse, we showed previously that this is due to a function of SAP in T cells, not in B cells (4). To ascertain whether SAP also has a role beyond the initiation of these responses, we used a hapten immunization system in which T cell-driven B cell responses are SAP dependent (4, 10, 23). We also took advantage of a mouse with a TAM-inducible SAP deficiency, in which SAP could be eliminated at various times after hapten immunization.

To address the role of SAP in the progression of GC reactions, mice were immunized with NP-OVA, and SAP expression was eliminated ∼4 to 9 days later, at a time when the priming of T cells by the antigen is believed to be complete (12, 17, 25). Analyses of these mice showed that removal of SAP during an ongoing GC reaction compromised the production of anti-NP antibodies. There was also reduced generation of GC T_FH cells, Bcl-6^hi GC T_FH cells, and GC B cells. Bcl-6^hi GC T_FH cells are the most polarized and active subset of T_FH cells (22, 25). In keeping with an earlier report (27), these cells are restricted to the GL-7⁺ subpopulation of T_FH cells. Therefore, sustained SAP expression was needed during an ongoing GC reaction in order to allow maximal T cell-dependent B cell responses.

Surprisingly, however, we failed to obtain any evidence that SAP was needed for the affinity maturation of the antibodies. This result could be due to the crudeness of the assays used to measure affinity maturation. Alternatively, it could indicate that once isotype switching has occurred, the affinity maturation of the antibodies does not require further contacts between T_FH cells and GC B cells. Sustained contacts between T_FH cells and GC B cells are known to be dependent on SAP (11, 23, 24). Further experiments addressing the latter possibility are warranted.

We also examined the role of SAP in the reactivation of T cell-dependent B cell responses. This was achieved by first immunizing mice with NP-OVA and then eliminating SAP ∼4 weeks later, when the GC reaction was finished and immune memory was established. Then, at ∼140 days, mice were immunized again with NP-OVA. Deletion of SAP after primary immunization had no impact on the total number of T_FH cells or GC T_FH cells after secondary immunization. However, it compromised the induction of Bcl-6^hi GC T_FH cells after the second immunization. In agreement with the notion that reactivation of T_FH cell functions was compromised by a lack of SAP, there was also reduced expansion of NP-specific GC B cells and memory B cells.

A variant of this experiment further cemented the idea that SAP was needed in order for previously generated antigen-specific T_FH cells to induce new B cell responses. In this study, mice were first immunized with PC-OVA, followed by SAP deletion and then by immunization with NP-OVA. The ability to generate antibodies against the second hapten, NP, was severely compromised by SAP deletion. This was also the case for the generation of NP-specific GC B cells and memory B cells. Thus, SAP was required for the reexpansion of fully polarized T_FH cells during recall responses.

Memory T_FH cells have been documented both in humans and in mice (20–22). These cells exhibit some of the markers of T_FH cells (CD4⁺ CXCR5⁺), but have lower levels of other markers, such as ICOS and Bcl-6. While some debate exists regarding the location and plasticity of memory T_FH cells, they have the ability to respond more strongly and faster to repeated antigenic exposure. Our finding that SAP was required for the expansion of Bcl-6^hi GC T_FH cells and antigen-specific GC B cells during recall responses implied that SAP was required for the reactivation of these memory T_FH cells.

In contrast to the effects of SAP on T_FH cell polarization and GC reactions, the studies with repeated immunization indicated that SAP was not required for the reinduction or maintenance of antibody levels once B cell memory was established. To address this finding further, mice were immunized with NP-OVA, followed by SAP deletion and monitoring of anti-NP antibody levels. SAP was not required for the maintenance of anti-NP antibodies over time. More importantly, it was also not required for the persistence of NP-specific memory B cells and ASCs. This is not because the function of fully polarized T_FH cells became independent of SAP. Indeed, the experiment with immunization with PC-OVA, followed by immunization with NP-OVA, showed that SAP deletion after primary immunization seriously compromised the production of antibodies, GC B cells, and memory B cells against the second hapten. Thus, SAP was not required for the persistence of antigen-specific memory B cells and long-lived ASCs. This is presumably because this process was independent of T_FH cells.

How does SAP promote the progression and reactivation, but not the maintenance, of T cell-dependent B cell immunity? It was proposed previously that the ability of SAP to support the initiation of GC reactions, antibody production, and memory B cell generation relates to its capacity to stabilize the formation of a conjugate between T_FH cells and GC B cells (11, 23, 24). Since interactions between T_FH cells and GC B cells also occur during the progression and reactivation of GC reactions, we propose that a similar mechanism likely explains the involvement of SAP in these phases of T cell-dependent B cell responses. In all likelihood, close contacts between cognate T_FH cells and GC B cells are needed for the full polarization and expansion of GC T_FH cells and GC B cells during these various phases. In contrast, SAP is not needed for the maintenance of memory B cells and long-lived ASCs, presumably because these cells are no longer positioned in functional GCs and their functions have become independent of T_FH cells.

The molecular mechanism by which SAP enables T_FH cell functions still remains to be fully clarified. Previous studies showed that this function was independent of the ability of SAP to bind Fyn (10, 11). Moreover, a recent report indicated that in the absence of SAP, the SLAM family receptor Ly108 becomes inhibitory, suppressing T_FH cell functions (24). This appears to be due to enhanced coupling of Ly108 to the protein tyrosine phosphatase SHP-1 in the absence of SAP. In support of this idea, those investigators also showed that breeding of SAP-deficient mice with mice lacking Ly108 restored the antibody defect seen in SAP-deficient animals. Hence, it is likely that a significant component of the function of SAP in T_FH cells is the prevention of inhibitory signals mediated by SLAM family receptors, in particular Ly108.

It was suggested previously that pharmacological interference with the SAP pathway might be helpful in the treatment of certain human autoimmune diseases mediated by hyperactive or overabundant T_FH cells, such as systemic lupus erythematosus (SLE), dermatomyositis, and perhaps rheumatoid arthritis (RA) (23). Pharmacological inhibitors of SAP or, more realistically, blocking antibodies against SAP-associated SLAM family receptors expressed on T_FH cells and B cells, in particular NTB-A (the human equivalent of Ly108) and CD84, could be used for such interference. However, given our observation that SAP was not required for maintaining established T cell-dependent B cell immunity and autoimmunity, one has to wonder whether these therapies alone would be successful at treating clinically apparent autoimmunity. Additional therapies aimed at eliminating memory B cells and long-lived ASCs would likely be necessary to consolidate the effects of blockade of the SAP pathway. Nonetheless, human antibody-mediated autoimmune diseases, in which the period of exposure to the antibodies may be difficult to pinpoint, might still benefit by targeting the SAP pathway.

In summary, our data, coupled with previous findings (11, 23, 24), showed that SAP is critically important for several stages of T_FH cell polarization and GC reaction. At the beginning of a primary immune response, SAP is required for full polarization of T_FH cells, leading to the generation of Bcl-6^hi T_FH cells. The latter are most efficient at providing help to B cells and at initiating a GC reaction. During an ongoing GC reaction, SAP is also needed for full expansion of Bcl-6^hi T_FH cells and GC B cells. This enables the completion of GC formation and maximal antibody production. And during a secondary immune response, SAP is needed for previously generated T_FH cells, possibly memory T_FH cells, to undergo full repolarization and reexpansion, thereby reinducing the GC reaction. These pleiotropic effects further underscore the key role of SAP in T_FH cell-dependent B cell responses, as well as, perhaps, in some autoimmune conditions mediated by T_FH cells.