PD-1 Suppresses B-1b Cell Expansion and Long-lived IgG Production in Response to T cell Independent Type 2 Antigens

Abstract

B1b cells play a key role in producing antibodies (Ab) against T cell-independent type 2 (TI-2) antigens (Ags). However, the factors regulating Ab production by this unique B cell subset are not well understood. In this study, a detailed analysis of the B cell response to TNP-Ficoll was performed using normal mice. TNP-Ficoll delivered i.p. or i.v. induced rapid Ag-specific B-1b cell activation, expansion, isotype switching and plasmablast/plasma cell differentiation. Ag-specific B-1b cell numbers peaked at day 5 and then gradually declined in the spleen but remained elevated in the peritoneal cavity beyond 40 days post-immunization. In addition to expressing CD43, CD44, and CD86, Ag-activated B-1b cells transiently expressed PD-1, which functionally suppressed BCR-induced B-1b cell in vitro proliferation when additional costimulatory signals were lacking. Inhibiting PD-1:PD-1 ligand (PDL) interactions during TNP-Ficoll immunization significantly enhanced Ag-specific B-1b cell expansion and the frequency of IgG isotype switching and plasmablast/plasma cell differentiation. Remarkably, PD-1 mAb blockade during the first week following immunization resulted in significantly increased numbers of both splenic and bone marrow Ag-specific IgG3, but not IgM, secreting cells at both early (day 5) and late (week 6) timepoints. Moreover, Ag-specific serum IgG3, as well as IgG2c, IgG2b, and IgA levels remained significantly elevated in PD-1 mAb-treated relative to control Ab-treated mice for at least 6 weeks post-immunization. Thus, PD-1:PDL interactions occurring shortly after initial TI-2 Ag encounter play a critical role in suppressing Ag-specific B-1b cell expansion and the development of long-term IgG-producing bone marrow and spleen cells.

Introduction

Humoral immune responses to T cell independent type 2 (TI-2)² antigens (Ag) are critical for protective immunity to encapsulated bacteria such as Streptococcus pneumoniae, an important cause of localized and systemic life-threatening infections (1). TI-2 Ags, such as pneumococcal polysaccharides, are often carbohydrate structures consisting of repeating epitopes that extensively crosslink Ag-specific B cell receptors (BCR) and induce Ab production in the absence of major histocompatibility complex class II-restricted T cell help (2). Numerous pathogens are known (3–9) or suspected (4) to display TI-2 Ags. Ab responses to TI-2 Ags are elicited rapidly, yield persistent titers, and may offer significant protection. For example, the Pneumovax vaccine composed of 23 pneumococcal polysaccharides provides significant protection against invasive pneumococcal disease in adults with titers lasting for ~10 years (10). Nonetheless, TI-2 Ags can often present unique challenges to vaccine development, including modest Ag-specific IgG production and impaired Ab responses in neonates (11–13). Thus, a better understanding of the mechanisms regulating TI-2 Ab production is necessary to develop enhanced TI-2 Ag-based vaccines.

Ab responses to TI-2 Ags differ in multiple respects to those elicited by T cell dependent (TD) Ags. Importantly, the B cell subsets producing Ab in response to TI-2 versus TD Ags differ. B-1b cells produce Ab responses to classical carbohydrate TI-2 Ags including pneumococcal polysaccharide (14), NP-Ficoll (15), and α1-3 dextran (16), as well as protein-based TI Ags present on clinically relevant pathogens (17–20). B-1a and marginal zone B cells also contribute to TI Ab production (21–23). This is in contrast to TD Ab responses in which follicular B cells largely contribute to Ab production. The accessory signals required for optimal TD and TI-2 Ab responses also differ. TI-2 Ab responses can ensue in the absence of cognate T cell help, whereas TD Ab responses are dependent on T-cell derived signals. As these signals drive somatic hypermutation, class switching, and B memory cell formation, TI-2 Ags, as well as TI-1 Ags (supplying additional activating signals), induce limited affinity maturation and isotype switching (IgG3 in mouse and IgG2 in human), and an unconventional type of memory (15–17, 24). Hence, the factors modulating TI-2 Ag-dependent B cell activation, proliferation, isotype switching and differentiation may differ from those involved in TD Ag-dependent B cell responses.

Humoral responses to TI-2 Ags also rely heavily on distinct BCR signaling pathways (25, 26) as well as key regulators of these pathways. Numerous cell surface receptors that regulate BCR signaling, including Programmed cell death 1 (PD-1), have been implicated in regulating TI-2 Ab responses. PD-1, a member of the B7/CD28 family, is expressed by Ag-specific B cells shortly after TI-2 Ag immunization (27), and is well-documented to negatively regulate Ag receptor signaling on both B and T cells following engagement of its ligands, PD-L1 and PD-L2 (28, 29). PD-1^−/− mice generate enhanced IgG3 production in response to the TI-2 Ag, DNP-Ficoll, and exhibit multiple immune abnormalities including moderate myeloid and lymphoid hyperplasia, hyperresponsive B cells, and decreased CD5 expression on peritoneal B-1 cells that may be due to dysregulated CD5 expression and/or increased B-1b cell numbers (29). It is unclear whether increased TI-2 Ab responses in PD-1^−/− mice are due to one or more of these preexisting abnormalities or due to PD-1 regulatory effects that occur at the time of immunization. Thus, the role of PD-1:PD-1 ligand (PDL) interactions in regulating TI-2 Ab responses remain unknown.

Studies investigating factors regulating TI-2 Ag responses have employed the use of pathogen-derived Ags, including pneumococcal polysaccharides, as well as synthetic Ags such as haptenated Ficoll, which in contrast to pathogen-derived Ags, are free of contaminating pathogen-associated molecular patterns (PAMPs) that can supply additional immunomodulatory signals. The synthetic Ag, 2,4,6-trinitrophenol-Ficoll (TNP-Ficoll), an inert copolymer of sucrose and epichlorohydrin conjugated to TNP, has been used for decades as a prototypic TI-2 Ag. Recent studies using knockout mice with deficiencies in select B cell populations have suggested that marginal zone B cells play a key role in the humoral immune response to TNP-Ficoll (30, 31). Nonetheless, since mice lacking this subset remain able to produce anti-TNP-Ficoll Ab responses (27), alternative populations may participate in humoral responses to this commonly used Ag.

In this study, Ag-specific B cell activation, expansion, differentiation, and Ab production in response to TNP-Ficoll were examined using normal mice. In contrast to BCR transgenic mice, normal mice are advantageous in that they express a broad Ab repertoire and unaltered B cell subset distribution, both of which may be important factors in shaping TI-2 Ab responses. Importantly, in the current study, Ag-specific B-1b cells were found to be a major B cell population that responded to TNP-Ficoll, regardless of immunzation route. In response to immunization, Ag-specific B-1b cells selectively increased in number, expressed multiple markers of activation, including PD-1, and underwent isotype switching and expressed CD138, a marker of plasmablast/cell differentiation. Data generated using a mAb to block PD-1 from interacting with its ligands at the time of immunization provides evidence that PD-1:PDL interactions suppress Ag-specific B-1b cell expansion, isotype switching, and overall B cell Ab production in response to TI-2 Ags. Collectively, the results of this study support a key role for PD-1 in regulating B-1b cell responses and long-lived Ab production against TI-2 Ags.

Materials and Methods

Mice

Experiments were performed on 2–3 month-old wild type C57BL/6 mice (Jackson Laboratories, Bar Harbor, ME) or CD19^−/− mice (14) housed under specific pathogen free conditions. All studies and procedures were approved by the Wake Forest University Animal Care and Use Committee.

Immunizations, ELISAs, and ELISPOTs

TNP-Ficoll-specific B cell expansion and PD-1 upregulation experiments were performed on mice immunized intraperitoneally (i.p.) or intraveneously (i.v.) with 50 µg TNP₆₅-Ficoll (Biosearch Technologies, Novato, CA). For serum Ab analyses, mice were immunized i.p. 25 µg of TNP₆₅-Ficoll. Ags were diluted in sterile PBS and injected in a final volume of 200 µl. In some experiments, mice were administered PD-1 mAb (RMP1-14; low endotoxin/no azide format; Biolegend, San Diego, CA) or control rat IgG (Southern Biotechnology Associates, Birmingham, AL) in 200 µl sterile PBS via i.p. injection with 200 µg mAb on d0 and 100 µg mAb on d3 and d5.

ELISAs were as described (27, 32). Serum samples were diluted in TBS containing 1% BSA (Sigma Chemical, St. Louis, MO). TNP-specific Ab levels were measured by adding diluted serum samples to plates that had been coated with 5 µg/ml TNP-BSA (Biosearch Technologies) in 0.1 M borate buffered saline overnight at 4°C. AP-conjugated polyclonal goat anti-mouse IgM, IgG1, IgG2c, IgG2b, IgG3, IgG, and IgA Abs (all from Southern Biotechnology Associates) and pNPP (Sigma) were used to detect Ag-specific Ab.

ELISPOTs were performed on total splenocytes and bone marrow cells. ELISPOT 96-well plates (Immobilon P, Millipore, Billerica, MA) were precoated with TNP-BSA (5 µg/ml) in PBS overnight at 4°C, washed 2 times with PBS, and blocked 1 hr at 37°C with cRPMI containing 10% FCS (Gibco BRL). Cells were plated at a concentration ranging from 10⁶–10⁷ cells/ml in cRPMI containing 10% FCS and cultured 18 hrs. AP-conjugated polyclonal goat anti-mouse IgM and IgG3 Abs (Southern Biotechnology Associates) were used in conjunction with NBT/BCIP substrate (Promega, Madison, WI), according to manufacturer instructions. Membranes were dried and spots were enumerated.

Abs and Immunofluorescence Analysis

Blood collected in heparin and spleen homogenate pellets were lysed in RBC lysis buffer (0.15 M NH₄Cl/0.01M KHCO₃). Peritoneal cells were isolated by lavaging the peritoneal cavity with 10 ml PBS. Single cell blood, spleen, lymph node and peritoneal cavity leukocyte suspensions (2 × 10⁷/ml) were incubated in PBS containing 2% bovine calf serum (BCS) with 20 µg/ml TNP₃₀-Ficoll-Fluorescein₈ (Biosearch Technologies) for 30 min. at room temperature, followed by subsequent staining with fluorochrome-labeled mAbs on ice for 25 min. Biotinylated- or fluorochrome-conjugated antibodies and secondary detection reagents used included: anti-mouse IgM and IgG3 (Southern Biotechnology Associates, Inc.); or Abs reactive with mouse CD1d (1B1), CD5 (53-7.3), B220 (RA3-6B2), CD11b (M1/70), CD23 (B3B4), CD19 (6D5), CD44 (IM7), and CD86 (GL1) (all from Biolegend); CD19 (ID3), CD80 (16-10A1), CD21/35 (7G6), and PD-1 (J43) mAb (all from eBioscience), and CD138 (BD Biosciences). Cells were analyzed using FACSCalibur and FACSCantoII flow cytometers (Becton Dickinson, San Jose, CA). Positive and negative cell populations were determined using unreactive isotype-matched Abs (Biolegend and eBioscience) and data was analyzed using FlowJo analysis software (Treestar).

B cell proliferation assays

CD5⁻ B cell subsets were purified from peritoneal cavity lavage by a negative depletion procedure. Macrophages were removed by plate adherence in RPMI containing 5% FCS (1 hr at 37°C, 5% CO₂). Nonadherent cells were depleted of Thy1.2⁺ cells using magnetic bead depletion (Dynal). Thy1.2⁻ cells were further depleted using biotinylated F4/80, GR1, DX5, and CD5 mAbs (Biolegend) in conjunction with magnetic depletion using Biotin binder beads (Dynal). In some experiments, CD11b⁺ B cells were further purified using Miltenyi bead purification. Purities were typically ~85–95% B cells. Purified peritoneal B cells were CFSE-labeled (0.6 µM) using Vybrant’s CFDA SE Cell Tracer Kit (Invitrogen), according to manufacturer instructions. Cells (2 × 10⁶/ml) were cultured in complete RPMI 1640 medium containing 10% FCS (Gibco Certified serum, Invitrogen) for 4 days in medium alone or in the presence of 1 µg/ml biotinylated F(ab')₂ anti-mouse IgM Ab from Jackson Immunoresearch (West Grove, PA). In some cultures, 2 µg/ml biotinylated PD-1 mAb (J43, eBioscience) or biotinylated Armenian hamster IgG (eBioscience) was added, along with 5 µg/ml streptavidin (Sigma). LPS (Escherichia coli 0111:B4; Sigma) and anti-mouse CD40 (HM40-3; BD Biosciences) were also used. Cells were harvested on d4, stained with fluorochrome-labeled mAbs against CD11b and B220, as well a 7AAD and Annexin V-PE (BD Biosciences). An equal number of CD11b⁺B220⁺ events were collected using a FACSCalibur instrument and data was analyzed using FlowJo analysis software.

Statistical analysis

Data are shown as means ± SEM. Differences between sample means were assessed using Student’s t test.

Results

TNP-specific B cell activation and expansion following TNP-Ficoll immunization

As early as 3 days post TNP-Ficoll immunization, significant increases in both the frequency and number of TNP-specific (B220⁺) B cells were observed in both the peritoneal cavity and spleen (Fig. 1A), as previously demonstrated (27). Five days post immunization, Ag-specific B cell frequencies and numbers peaked in the spleen (Fig. 1A). However, by 35 days post immunization splenic Ag-specific B cell numbers were only ~20% increased over numbers in naïve animals (Fig. 1A). In contrast, elevated Ag-specific peritoneal B cell frequencies and numbers did not decrease following the day 5 timepoint, but remained significantly increased over naïve levels beyond 35 days post immunization. The increases observed in TNP-Ficoll binding B cells following immunization was likely due to Ag-specific binding as opposed to Fc receptor binding of TNP-specific Ab, since stripping B cells of any Fc receptor-bound Ab by 3 minute incubation with 50 mM glycine buffered saline, pH=3 (33) did not significantly alter the frequency of TNP-Ficoll binding B cells (99 ± 4% of no treatment control, n=4). Thus, TNP-Ficoll immunization rapidly increases Ag-specific B220⁺ B cell numbers in the spleen and peritoneal cavity, with numbers gradually contracting in spleen but remaining elevated in the peritoneal cavity 5 weeks beyond immunization.

Figure 1. Ag-specific B cell phenotype, activation, differentiation, and expansion kinetics in response to TNP-Ficoll immunization.

A–G, Flow cytometric analysis and enumeration of TNP₃₀-Fl-Ficoll-binding (Ag-specific) cells from naïve and immune mice (50 µg TNP₆₅-Ficoll administered i.p.). A, Representative flow cytometric analysis (left panels) of Ag-specific B220⁺ splenic and peritoneal B cells in naïve and immune (d3) mice. Frequencies and numbers of splenic and peritoneal Ag-binding B cells from days 0 to 35 post-immunization are indicated (n≥5 mice per group). B, Phenotype of splenic and peritoneal Ag-specific B cells in naïve (shaded histogram) and immune (d5; thick line) mice. Isotype control binding for Ag-specific B cells from immune mice is indicated (dashed line). C, B220 and CD11b expression by blood, spleen, peritoneal cavity, and lymph node Ag-specific cells at d0 and 5 post-immunization. D, CD5 and CD11b expression on spleen Ag-specific B cells at d0 and 35 post-immunization. E, Frequencies and numbers of Ag-specific splenic and peritoneal B-2 (B220⁺CD11b⁻CD5^neg), B-1b (B220⁺CD11b⁺CD5^lo-neg), and B-1a (B220⁺CD11b⁺CD5⁺) cells in naïve and immune mice. A through G, Activation marker (CD43, CD44, CD80, and CD86; F) and IgG3 (d5; G) expression by Ag-specific CD11b⁺ and CD11b⁻ B cells in spleen (F, G) and peritoneal cavity (G) following TNP-Ficoll immunization. Isotype control binding is shown for Ag-specific cells from immune mice (F). A–G, Results are representative of data obtained with ≥3 mice per group. Values in (A) and (E) represent means (±SEM), with significant differences between naive and immune mice indicated by asterisks; *, p<0.05.

The phenotype, activation, and differentiation status of Ag-specific B cells was assessed following TNP-Ficoll immunization. Relative to Ag-binding B cells in naïve mice, Ag-specific splenic and peritoneal B cells in immune mice (day 5) had increased FSC, SSC, and increased CD86, CD44, and CD43 expression, indicative of activation (Fig. 1B and data not shown). In addition, Ag-specific splenic B cells in immune mice had unchanged CD1d^int expression, but reduced levels of CD21/35, CD23, and B220, and increased levels of CD19. A fraction of Ag-specific B220⁺ B cells in the spleen had also undergone isotype switching to IgG3 and expressed CD138, indicative of plasmablast differentiation (Fig. 1B and data not shown). Thus, TNP-Ficoll immunization induces Ag-specific peritoneal and splenic B cell activation and differentiation, with a substantial population of splenic Ag-specific B cells expressing a B220^loCD19^hiCD1d^intCD23^loCD21/35^lo phenotype.

Ag-specific B-1b cells are a major B cell population responding to TNP-Ficoll immunization

The B220^loCD19^hiCD1d^intCD21/35^lo phenotype is common to B-1 cells. Thus, the expression pattern of additional B-1 markers, CD11b and CD5, was examined for Ag-specific cells. In naïve mice, Ag-specific peritoneal B cells were mostly CD11b⁺ and expressed either intermediate (CD5^int) or very low levels of CD5 (CD5^neg-lo), characteristic of B-1a and B-1b cells, respectively (Fig. 1B), while naïve splenic TNP-specific B cells were CD5^neg and CD11b^neg. Following i.p. immunization, CD11b⁺ Ag-specific B cell frequencies increased in the peritoneal cavity, spleen, blood, and lymph nodes (Fig. 1B–C). These cells had decreased B220 expression levels. In addition, increases in Ag-specific B cells in the spleen and peritoneal cavity expressing a CD5^neg-lo phenotype were observed following immunization (Fig. 1B). Importantly, the level of CD5 expressed by Ag-specific B cells from immune mice was lower than that expressed by peritoneal B-1a cells and was comparable to levels present on peritoneal B-1b cells (Fig. 1B and Supplemental Fig. 1A). Interestingly, Ag-specific B cells expressing CD5 were not found in naïve CD19^−/− mice, but were identified in TNP-Ficoll-immunized CD19^−/− mice (Supplemental Fig. 1B), which lack B-1a cells (14). Thus, B cells responding to TNP-Ficoll express a CD11b⁺CD5^neg-lo phenotype.

The phenotype of Ag-specific B cells (CD11b⁺B220^loCD19^hiCD1d^intCD21/35^loCD5^neg-lo) responding to TNP-Ficoll was similar to that of peritoneal B-1b cells. Therefore, changes in Ag-specific B-1b cell frequencies, activation, isotype switching, and differentiation were specifically evaluated following TNP-Ficoll immunization. Relative to naïve numbers, Ag-specific B-1b cell numbers were significantly increased in spleen (16-fold) and peritoneal cavities (3-fold) 5 days following immunization, in contrast to Ag-binding B-1a and B-2 cells, which were not significantly altered (Fig. 1E). Ag-specific peritoneal B-1b cell numbers were increased further (12-fold) at day 35. Although CD11b expression gradually diminishes on B-1 cells that have migrated out of the serosal cavities, Ag-specific CD11b-expressing CD5^neg cells were still detectable in the spleen 35 days post immunization (Fig. 1D), and were still significantly higher in frequency and number (2.5-fold) than naïve mice at this timepoint (Fig. 1E). Ag-specific CD11b⁺ (B-1b) B cells were the predominant subset activated by immunization as evidenced by increased levels of CD43, CD44, CD86, and CD80 expression on this subset relative to CD11b⁻ B cells (Fig. 1F). Moreover, Ag-specific B-1b cells selectively underwent isotype switching to IgG3 and these cells were found in both the spleen and peritoneal cavity (Fig. 1G). Finally, Ag-specific CD138⁺CD11b⁺ plasmablasts were observed in the spleen, but not peritoneal cavity. Nonetheless, a large fraction of Ag-specific B220^loCD138⁺ cells were CD11b⁻ (data not shown), consistent with the loss of this marker upon differentiation to Ab-secreting cells (34–36). Thus, Ag-specific B-1b cells become activated, expand, undergo class switching, and differentiate into splenic Ab-secreting cells in response to TNP-Ficoll immunization.

Ag-specific B-1b cells are the major B cell population responding to TNP-Ficoll immunization, regardless of the route of Ag delivery

To assess whether the Ag-specific B cell responses to TNP-Ficoll were dependent on the route of Ag delivery, mice were immunized with TNP-Ficoll by i.v., s.c., or i.p. injection. Five days post-immunization, Ag-specific peritoneal and spleen B cell frequencies, numbers, and phenotypes were analyzed. Increases in total Ag-specific splenic B cell frequencies over naïve mice (0.47 ± 0.02%; p<0.05) were not significantly different between mice immunized i.v. (1.15 ± 0.13 %) and i.p. (1.00 ± 0.04%; data not shown). Similarly, the increases observed in total Ag-specific peritoneal B cell frequencies over naïve mice (0.75 ± 0.08%; p<0.05) were not significantly different between mice immunized i.v. (1.60 ± 0.22 %) and i.p. (1.67 ± 0.31%; data not shown). Subcutaneous immunization elicited weaker, but nonetheless significant increases in Ag-specific splenic B cell frequencies (0.63 ± 0.11% vs. 0.47 ± 0.02% for naïve, p<0.05) and increases in Ag-specific peritoneal B cell frequencies (1.05 ± 0.12% vs. 0.75 ± 0.08% for naive). Thus, i.p. and i.v. immunizations elicit similar increases in Ag-specific splenic and peritoneal B cell frequencies 5 days following immunization, whereas s.c. immunization elicits a weaker response.

To assess whether the Ag-specific B cell subset(s) responding to TNP-Ficoll were altered by the route of Ag delivery, Ag-specific B cell phenotypes were assessed. As shown in Fig. 2A–B, the frequencies of CD21^intCD1d^int, CD21^loCD1d^int, CD21^lo-intCD1d^lo, and CD21^hiCD1d^hi Ag-binding splenic B cells were examined. Significant increases in CD21^loCD1d^int Ag-specific B cells (>4-fold) and CD21^intCD1d^lo Ag-specific B cells (≥2-fold) were observed relative to naïve mice (Fig. 2B). These increases were comparable between i.p. and i.v. immunized mice (p>0.05). Ag-specific CD21^intCD1d^int (mainly follicular B cells) and CD21^hiCD1d^hi (MZ B cells) frequencies were not significantly changed 5 days following i.p. or i.v. TNP-Ficoll immunization. Moreover, CD11b⁺ Ag-specific cells appeared in CD21^loCD1d^int and CD21^lo-intCD1d^lo populations following immunization, regardless of whether TNP-Ficoll was delivered by i.p. or i.v. route (Fig. 2C). In contrast, Ag-specific CD21^intCD1d^int and MZ B cell populations exhibited minimal CD11b expression following immunization. Notably, no significant difference was observed in the frequencies of Ag-specific CD11b⁻ splenic B cells in mice immunized i.p. versus i.v. (data not shown). Thus, i.p. and i.v. TNP-Ficoll immunization elicits similar increases in CD11b⁺ Ag-specific splenic B cells that coexpress a CD21^loCD1d^int or CD21^lo-intCD1d^lo phenotype.

Figure 2. Ag-specific B-1b cells participate in the response to TNP-Ficoll regardless of whether Ag is delivered i.p. or i.v.

A–F, Flow cytometric analysis and enumeration of splenic and peritoneal Ag-specific B cells from naïve and immune mice (50 µg TNP₆₅-Ficoll administered i.p. or i.v.; d5). A–B, Representative flow cytometric analysis (A) and frequencies (B) of Ag-specific splenic B cells in naïve and immune mice expressing a CD21^intCD1d^int, CD21^loCD1d^int, CD21^lo-intCD1d^lo, and CD21^hiCD1d^hi (MZ B) phenotype. C, CD11b expression by CD21^intCD1d^int, CD21^loCD1d^int, CD21^lo-intCD1d^lo, and CD21^hiCD1d^hi Ag-specific cells. D, Frequencies and numbers of splenic and peritoneal Ag-specific B-1b (CD19⁺CD11b⁺CD5^lo-neg) cells in naïve and immune mice. E, IgG3 expression by Ag-specific splenic B (CD19⁺) cells (upper panels) and CD11b expression by IgG3⁺ Ag-specific cells (middle panels). Frequencies of CD19⁺ Ag-specific cells (“Total”) and CD19⁺CD11b⁺Ag-specific cells (“CD11b⁺”) expressing IgG3 are shown for immune mice. F, CD138 expression by Ag-specific splenic B cells (upper panels) and CD11b and B220 expression by CD138⁺ Ag-specific cells (middle panels). Frequencies of B220⁺ Ag-specific cells (“Total”) and B220⁺CD11b⁺Ag-specific cells (“CD11b⁺”) expressing CD138 are shown for immune mice. Results are representative of data obtained with ≥3 mice per group. Values (B, D–F) represent means (±SEM), with significant differences between naive and immune mice indicated by asterisks; *, p<0.05.

Consistent with the results above, i.v. and i.p. immunization resulted in similar increases in Ag-specific splenic B-1b (CD19⁺CD11b⁺CD5^neg/lo) cell frequencies and numbers (Fig. 2D; >10-fold) over naïve mice. Subcutaneous immunization also increased Ag-specific splenic B-1b cell frequencies and numbers over naïve mice, albeit to a lesser extent (data not shown). Remarkably, i.v. and i.p. immunization induced similar increases in Ag-specific peritoneal B-1b cell frequencies, although numbers were slightly higher in i.p. immunized mice (Fig. 2D). As shown in Fig. 2E, similar frequencies of the Ag-specific splenic B cell pool had undergone isotype switching to IgG3 in i.p. and i.v. immunized mice. Moreover, regardless of the route of immunization, the majority of Ag-specific IgG3⁺ B cells coexpressed CD11b⁺ (Fig. 2D). In addition, similar frequencies of the Ag-specific splenic B cell pool expressed CD138 in i.p. and i.v. immunized mice (Fig. 2F), and no differences were observed the frequencies of Ag-specific CD138⁺ cells coexpressing CD11b. Thus, i.p. and i.v. TNP-Ficoll immunizations stimulated similar increases in Ag-specific B-1b cell frequencies and numbers in multiple tissues and induced similar degrees of IgG3 isotype switching and plasmablast/cell differentiation, with comparable participation by Ag-specific B-1b cells.

PD-1 expression is induced on Ag-specific B-1b cells in vivo

Ag-specific B cells express PD-1 3 days following TNP-Ficoll immunization (27). As shown in Fig. 3, PD-1 upregulation is largely confined to CD11b-expressing (B-1) cells. PD-1 upregulation on Ag-specific peritoneal B cells was observed as early as 1 day post-immunization and appeared on blood and spleen B-1b cells by 2 days post-immunization. PD-1 expression levels were highest between days 2–3, decreased by day 5, and were undetectable by day 9 post-immunization. A similar trend was observed for CD44 expression (data not shown). Thus, PD-1 is selectively and transiently induced on Ag-specific B-1b cells following TNP-Ficoll immunization.

Figure 3. PD-1 expression is induced on Ag-specific B-1b cells in vivo.

B220⁺ Ag-specific peritoneal, spleen and blood CD11b⁻ and CD11b⁺ cells were evaluated for PD-1 expression (thick line) by flow cytometry on days 0, 1, 2, 3, 5, 9, and 14 post-immunization with 50 µg TNP₆₅-Ficoll. Isotype control binding by Ag-binding cells is indicated by the shaded histogram.

PD-1-BCR coengagement suppresses BCR-induced B-1b cell proliferation

Similar to in vivo expression kinetics, PD-1 is induced on cultured purified peritoneal B-1b cells and spleen B cells between days 1 and 2 post BCR activation, with peak expression observed on day 3 (Fig. 4A). CD5 expression was similarly induced on BCR-activated spleen B cells and B-1b cells purified by negative bead selection (Fig. 4B, Supplemental Fig. 1C, and data not shown) or FACS sorting (Supplemental Fig. 1D–E). CD5 expression was also induced on FACS-purified CD19^−/− peritoneal B-1b cells (Supplemental Fig. 1E). Culturing cells in the presence of 5 µg/ml LPS had little effect on BCR-induced PD-1 expression levels in B-1b cells, and only slightly reduced expression on splenic B cells (Fig. 4C). Similar to these results with LPS, TNFR superfamily members (ie., BlyS receptors and CD40) do not modulate BCR-induced PD-1 upregulation on B-1b cells (27). In contrast, LPS suppressed CD5 upregulation on BCR-activated B-1b cells (Fig. 4B), consistent with that previously reported for spleen B cells (37). Thus, BCR signaling induces PD-1 expression on B-1b cells, with secondary signals supplied by LPS and TNFR family members having little effect on PD-1 upregulation.

Figure 4. PD-1-BCR coengagement suppresses B-1b cell proliferation but not survival.

A, BCR-induced PD-1 surface expression on purified peritoneal B-1b cells and spleen B cells. B cells were cultured with 5 µg/ml goat anti-mouse IgM F(ab’)₂ for the indicated number of days, harvested and stained with PD-1 mAb and analyzed by flow cytometry. The shaded histogram indicates isotype control binding by activated B cells (d2). PD-1 expression by unstimulated (d3) medium (Med) cultured B cells is shown in the left-most panel. B, BCR-induced CD5 (B) and PD-1 (C) upregulation on B-1b cells (B–C) and splenic B cells (C) activated with 5 µg/ml goat anti-mouse IgM F(ab’)₂ in the presence or absence of 5 µg/ml LPS. The shaded histograms indicate isotype control staining in (C). D–F, Proliferation and survival in BCR-activated B-1b cells (D, F) and spleen B cells (E) in the presence and absence of BCR-PD-1 co-crosslinking. Purified CD5⁻ peritoneal B or spleen B cells were labeled with 0.5 µM CFSE and cultured in the presence of 1 µg/ml biotinylated goat anti-mouse IgM F(ab’)₂ along with 2 µg/ml biotinylated Armenian hamster anti-mouse PD-1 (J43) or biotinylated Armenian hamster IgG control. Streptavidin (5 µg/ml) was added to cultures as a crosslinking agent. At d4, cells were harvested, stained with fluorochrome-labeled mAbs against CD11b and B220, with CFSE loss assessed in the B220⁺CD11b^int peritoneal population (D, F) or splenic B220⁺ population (E). Apoptosis was assessed using 7AAD and Annexin V-PE staining. Mean (±SEM) division indices were calculated from ≥3 experiments using FlowJo analysis software (D, E). Significant differences between means are indicated by asterisks; *, p<0.05. G–H, Proliferation for CFSE-labeled peritoneal B-1b (G) and spleen B cells (H) cultured as in D–E along with anti-CD40 (0.5 µg/ml; HM40-3) or LPS (1 µg/ml). In (H), symbols represent division indices for individual mice.

To determine the functional consequences of PD-1 engagement on B-1b cell proliferation induced by BCR signaling, a biotinylated PD-1 mAb was used in combination with streptavidin to crosslink PD-1 independently or with the BCR using biotinylated F(ab’)₂ goat anti-mouse IgM. Independent PD-1 crosslinking during BCR activation had little effect on B-1b cell or spleen B cell proliferation elicited either by anti-IgM or LPS (data not shown). However, co-crosslinking PD-1 with IgM during B cell activation significantly reduced B-1b cell and spleen B cell proliferation as measured by reduced division indices (average number of cell divisions of entire population) relative to cultures in which a biotinylated isotype control mAb was used in place of PD-1 mAb (Fig. 4D–E). The frequencies of B-1b cells characterized as viable (Annexin V⁻/7AAD⁻), early (Annexin V⁺/7AAD⁻), or late (Annexin V⁺/7AAD⁺) apoptotic cells were similar between cultures in which PD-1 was cocrosslinked with the BCR compared to cultures in which biotinylated isotype control mAb was used in place of PD-1 mAb (Fig. 4F). Thus, reduced proliferation was not due to decreased survival, as the viabilities of B-1b cell cultures (as well as spleen B cell cultures) subjected to anti-IgM-PD-1 mAb crosslinking versus anti-IgM-control mAb crosslinking were not significantly different within experiments (B-1b cell cultures: 45.6 ± 16% vs. 41.9 ± 13.8%, n=3 experiments, p> 0.05, paired t-test; spleen B cell cultures: 22.4 ± 4.7% vs. 23.6 ± 6%, n=5; p> 0.05, paired t-test). Finally, co-stimulation supplied by either CD40 (Fig. 4G–H) or LPS (Fig. 4H) prevented PD-1 inhibitory effects on BCR-induced proliferation in both B-1b (Fig. 4G and data not shown) and spleen cells (Fig. 4H). Thus, PD-1 co-engagement with the BCR exerts an inhibitory effect on BCR-induced B cell proliferation, but not survival, that can be overcome by costimulation.

PD-1 mAb blockade significantly increases Ag-specific B-1b cell numbers and differentiation following TI-2 Ag immunization

Given the expression pattern of PD-1 by Ag-specific B-1b cells following TNP-Ficoll immunization (Fig. 3) and its inhibitory effects on primary B cell proliferation (Fig. 4 and ref.(38)), the effect of blocking PD-1 from interacting with its ligands during TI-2 Ag immunization was assessed. This was accomplished using the PD-1 blocking mAb, RMP1-14. Following TNP-Ficoll immunization, mice receiving rat IgG control Ab had significantly increased Ag-specific peritoneal B cell frequencies (1.7-fold) and numbers (2.6-fold) relative to naïve mice (Fig. 5A). As expected, this increase was largely attributed to significantly increased Ag-specific B-1b cells, as significant increases were not observed in B-1a or B-2 subsets (Fig. 5B). However, mice receiving the PD-1 blocking mAb following immunization had significantly higher increases in Ag-specific peritoneal B cell frequencies (2.4-fold) and numbers (4.4-fold) relative to control mice (Fig. 5A). PD-1 mAb treatment significantly increased Ag-specific peritoneal B-1b cells over mice treated with control Ab 5 days post immunization, but had no effect on other Ag-binding B cell subsets (Fig. 5B). Whereas Ag-specific B-1b cell frequencies increased by 2.6-fold in control immune mice, they were increased 4-fold in mice receiving PD-1 mAb. As observed in earlier experiments (Fig. 1E), increases in Ag-specific peritoneal B-1b cells were still observed out to 40 days following immunization (Fig. 5C). However, mice receiving PD-1 blocking mAb (at d1, 3, and 6) exhibited significantly higher increases in Ag-specific peritoneal B-1b cell frequencies and numbers relative to mice receiving control Ab (Fig. 5C). Notably, PD-1 mAb treatment did not influence overall total peritoneal B-1b cell frequencies or numbers at day 5 and 40 timepoints (Fig. 5D and data not shown). Finally, PD-1 mAb treatment resulted in significantly increased Ag-specific IgG3⁺ peritoneal B cell frequencies and numbers 5 days post immunization (Fig. 5F), nearly all of which expressed CD11b (Fig. 5E). Thus, blocking PD-1 interactions with its ligands during TNP-Ficoll immunization significantly and selectively increased Ag-specific peritoneal B-1b cell numbers and IgG3⁺ B cells.

Figure 5. Blocking PD-1:PD-L interactions during TI-2 Ag immunization increases the frequency of Ag-specific B-1b cells, IgG3⁺ cells, and CD138⁺ B cells.

A–H, Ag-specific B220⁺ frequencies and numbers at days 5 and 40 in mice immunized with 50 µg TNP₆₅-Ficoll i.p. and administered PD-1 (RMP1-14; black bars) or rat IgG control (gray bars) Abs (200 µg on d1 and 100 µg on d3; 100 µg was also given on d5 for the 40-day experiment). Ag-specific cells and CD11b, CD5, IgG3, and CD138 expression were assessed by flow cytometry. A–B, Peritoneal Ag-specific B220⁺ B cell frequencies and numbers (A) and Ag-specific peritoneal B-1a, B-1b, and B-2 subset frequencies (B) in naïve mice (white bars) and 5 days post immunization in mice treated with PD-1 mAb (black bars) or rat IgG control (gray bars). C, Ag-specific peritoneal B-1b cell frequencies and numbers in naïve and immune mice 40 days post immunization. D, Total peritoneal B-1b cell frequencies at days 0, 5, and 40 post immunization. E–F, Ag-specific IgG3⁺ CD11b⁺ peritoneal B cell frequency plots (E) and Ag-specific B220⁺IgG3⁺ cell frequencies and numbers 5 days post immunization (F). G–H, Blood (G) and spleen (H) Ag-specific B-1b cell frequencies and numbers at days 0 and 5 post immunization. I, CD138⁺B220⁺ cell frequencies within the Ag-specific B cell population. The first plot demonstrates the gating strategy for Ag-binding cells. Values represent means (±SEM), with significant differences between conditions indicated by asterisks; *, p<0.05 (n≥3–4 mice/group).

Immunization-induced increases in Ag-specific total B, and B-1b, cell frequencies and numbers were not significantly altered in blood or spleen by PD-1 mAb treatment relative to control Ab-treated mice 5 or 40 days post immunization (Fig. 5G–H, and data not shown). However, 5 days post immunization, CD138 (marking plasmablast/plasma cell differentiation) was expressed by a significantly higher frequency of Ag-specific B cells in mice that received PD-1 mAb compared to mice that received control Ab (Fig. 5I). Approximately one-third of these Ag-specific CD138⁺ plasmablasts expressed CD11b in both treatment groups at day 5 (data not shown). Moreover, although overall Ag-specific splenic B cell frequencies were not altered by treatment, the frequency of remaining Ag-specific splenic B cells that were IgG3⁺ at 40 days post immunization was significantly increased (2-fold) in mice that had received PD-1 mAb blockade (0.41 ± 0.06 %, n=4) versus control Ab (0.2 ± 0.03 %, n=4; p<0.05, data not shown). Thus, blocking PD-1:PDL interactions significantly increased isotype switching in Ag-specific B cells and the frequency of Ag-specific B cells committed to producing Ab.

PD-1 mAb blockade significantly increases IgG production against TI-2 Ags

Since blocking PD-1 interactions with its ligands significantly promoted increases in Ag-specific B-1b cell numbers, IgG switching, and differentation following immunization, the effect of PD-1 mAb blockade on TI-2 Ab production was assessed. As shown in Fig. 6, PD-1 blockade had no effect on TNP-specific IgM levels following TNP-Ficoll immunization. However, PD-1 mAb blockade significantly enhanced Ag-specific IgG levels. IgG3 production, the dominant isotype produced in response to TNP-Ficoll, was significantly increased by PD-1 mAb blockade, as were Ag-specific IgG1, IgG2b, IgG2c, and IgA levels. Notably, Ag-specific IgG3 and IgG2b levels remained significantly augmented at 6 weeks post immunization in mice that had received transient early PD-1 mAb blockade. Thus, blocking PD-1:PDL interactions during the first week following immunization significantly enhanced the production of isotype-switched Ag-specific Abs, with IgG3 and IgG2b levels remaining increased up to 6 weeks post immunization.

Figure 6. Blocking PD-1:PD-L interactions during TI-2 Ag immunization significantly increases the production of isotype-switched Ab.

TNP-specific serum IgM, IgG, IgG1, IgG2b, IgG2c, IgG3 and IgA levels in mice immunized with 25 µg TNP-Ficoll and administered PD-1 or rat IgG control Abs as in Figure 5. Values represent means (±SEM), with significant differences between mice receiving control Ab (open circles) and PD-1 mAb (filled squares) indicated by asterisks; *, p<0.05 (n≥3–4 mice/group). Similar results were obtained in an independent immunization experiment.

PD-1 mAb blockade significantly increases both early and long-term splenic and bone marrow IgG3-producing cells in response to TI-2 Ag

Evidence suggests that TI-2 Ag-activated, as well as TI-pathogen activated, B-1b cells may predominantly secrete long-term Ab as plasmablasts within the spleen (15, 16, 20, 39), although B-1b cells may also differentiate into long-lived bone marrow plasma cells (40). To assess Ab production by splenic and bone marrow plasma(blast) cells at both early and late timepoints following TNP-Ficoll immunization, ELISPOTs were performed. At 5 days post immunization, TNP-secreting IgM and IgG3 ASC spleen and bone marrow numbers were significantly increased over naïve mice (Fig. 7A). Relative to day 5 ASC numbers, day 40 IgM and IgG3 splenic ASC numbers were diminished, whereas bone marrow ASC numbers were increased (Fig. 7B). Given previous estimation of total bone marrow cellularity in mice (1.5 × 10⁸) (41) relative to spleen (1 × 10⁸), bone marrow ASC therefore likely make a substantial contribution to TNP-specific Ab levels. Notably, at both timepoints, IgG3 ASC accounted for ≥75 % of splenic IgG ASC numbers and ~50% of bone marrow IgG ASC, indicating this was the major IgG isotype produced in response to TNP-Ficoll (data not shown). Thus, TNP-Ficoll immunization induces rapid induction of splenic ASCs and long-term production of IgM and IgG3 Ab by both splenic and bone marrow ASC.

Figure 7. Transient blockade of PD-1:PD-L interactions increases both early and long-term Ag-specific IgG3-secreting cell numbers in spleen and bone marrow.

A–B, TNP-specific IgM and IgG3-secreting cells (ASC) in spleen and bone marrow 5 (A) and 40 (B) days post-immunization as determined by ELISPOT. C, TNP-specific IgM, IgG, and IgG3 levels secreted by total splenocytes (harvested 5 days post-immunization) cultured in cRPMI + 10% FCS for 7 days. Culture supernatants were diluted 1:3 in TBS containing 1% BSA and assessed for TNP-specific Abs by ELISA. Mice were immunized and administered PD-1 or rat IgG control Abs as in Figure 5. Values represent means (±SEM), with significant differences between mice receiving control Ab (gray bars) and PD-1 mAb (black bars) indicated by asterisks; *, p<0.05 (n≥3–4 mice/group). Although not indicated, in all cases Ag-specific ASC numbers and anti-TNP Ig levels were significantly higher in immune mice compared to naïve mice. N.D., not detected.

The effect of blocking PD-1:PDL interactions during the first week following immunization on ASC numbers at both early (d5) and late (d40) timepoints was assessed. PD-1 mAb treatment during the first week following TNP-Ficoll immunization significantly increased (≥2-fold) IgG3-secreting cells in spleen and bone marrow at both d5 and d40 timepoints relative to control Ab-treated mice (Fig. 7A–B). In contrast, PD-1 mAb blockade did not significantly alter Ag-specific splenic IgM producing cell numbers relative to control Ab-treated mice at either timepoint (Fig. 7A–B), although there was a trend towards increased IgM-secreting bone marrow cells at both timepoints. Moreover, anti-TNP-specific IgG and IgG3, but not IgM, levels were significantly increased in 7-day cultures of spleen cells from PD-1 mAb treated mice relative to control Ab-treated mice (Fig. 7C), consistent with increased IgG-secreting cell frequencies found in these mice. Thus, blocking PD-1:PDL interactions during the early stages of TI-2 Ag encounter significantly increases the number of IgG3-secreting splenic and bone marrow ASC found at both early (d5) and late (d40) timepoints following immunization.

Discussion

In this study, nontransgenic normal mice were used to analyze the activation, phenotypic alterations, and expansion kinetics of Ag-specific B cells in response to a defined prototypic TI-2 Ag. B cells responding to TNP-Ficoll expressed a phenotype consistent with activated B-1b cells, as was true for isotype-switched Ag-specific B cells (Fig. 1). Importantly, immunization induced Ag-specific B-1b cells to selectively upregulate PD-1 (Fig. 3). PD-1 has been shown in multiple studies to suppress Ag-specific T cell expansion and function (42), and in the current study, is shown to similarly inhibit Ag-receptor-induced proliferation of B-1b cells in vitro (Fig. 4). PD-1 expression by Ag-specific B-1b cells may similarly contribute to suppression of Ag-specific B-1b cell expansion and/or antibody production in vivo, as PD-1 mAb blockade significantly enhanced Ag-specific B-1b cell numbers, IgG3 switching and Ab production (Figs. 5–7). Collectively, the results of this study demonstrate a key role for the PD-1:PDL regulatory axis in controlling B-1b cell responses and IgG production to TI-2 Ags.

The division of labor among B cell subsets in Ab production against TI-2 Ags is not completely clear, with B-1b cells, B-1a cells, marginal zone B cells, and even follicular B cells (43) under certain circumstances implicated in producing TI-2 Ab responses. Work by multiple labs has demonstrated a key role for B-1b cells in producing Abs in response to defined TI-2 Ags, including PPS-3 (14, 40), NP-Ficoll(15), and α1-3;3-dextran (16), additional TI Ags (17–19), and the Gal α1-3Galbeta1-4GlcNAc (Gal) carbohydrate epitope involved in transplant rejection (34). Nonetheless, marginal zone B cells and B-1a cells may also contribute to Ab production against these and other TI-2 Ags (16) (30, 31), including phosphorylcholine when either displayed on bacteria (21) or on Ficoll as a TI Ag (14). These subsets may differentially respond to Ag and/or additional cues to produce Abs (23, 44). The present study demonstrates that the Ag-specific B cells that rapidly become activated, expand and/or mobilize in response to TNP-Ficoll are CD11b⁺CD19^hiCD21^loCD44⁺CD86⁺CD80⁺CD1d^intCD23^negB220^loCD43⁺PD-1⁺CD5^−/lo (Fig. 1), consistent with an activated B-1b cell phenotype (Fig. 4 and ref. (27)). Importantly, CD5 was expressed by Ag-specific B-1b cells at lower levels than that found for B-1a cells (Fig. 1B and Supplemental Fig. 1A), was present on Ag-specific B cells from immunized (but not naïve) CD19^−/− mice (Supplemental Fig. 1B) which lack B-1a cells and but produce near normal Ab responses to TNP-Ficoll (14, 27). Increased CD5 expression is observed on BCR-activated spleen B cells (Fig. 4B and ref. 37) and bead-purified B-1b cells (Fig. 4B and Supplemental Fig. 1C). As CD5^lo B cells could have possibly remained following bead selection, wild type and CD19^−/− CD5⁻ B-1b cells were also FACS-sorted to high purity and were similarly found to express increased CD5 levels with BCR activation (Supplemental Fig. 1D–E). Thus, low CD5 expression on Ag-specific B-1b cells following immunization may be indicative of Ag receptor-mediated B-1b cell activation. Alternatively, it remains possible that TNP-Ficoll induces expansion of Ag-specific CD5^lo B-1 cells as opposed to increasing expression of CD5 on Ag-activated B-1b cells. Interestingly, Francisella tularemia LPS (Ft-LPS), an extremely weak TLR4 agonist, elicits a similar Ag-specific TI response to that observed with TNP-Ficoll, with i.p. immunization leading to the appearance of a population of Ag-specific IgD^loCD21^lo-intCD23^negCD138^+/−CD5⁺ cells in the spleen (22). Whether the B cell populations responding to these two different antigens are related or distinct is presently unclear. CD5 expression by Ag-activated B-1b cells may be transient as CD5 expression was not found on the Ag-specific B-1b cells remaining in the spleen 5 weeks post immunization (Fig. 1D) and previous work by Hayakawa et al. demonstrated that anti-TNP ASC in the spleen are CD5⁻ following TNP-Ficoll immunization (45). Similarly, Cole et al. reported diminished CD5 expression as Ft-LPS specific B cells differentiated to plasma cells (22). Notably, CD5 expression on Ag-specific B-1b cells may be suppressed in the context of infection, as PAMPs such as LPS may inhibit CD5 upregulation on Ag-activated B-1b cells (Fig. 4B) as observed for spleen B cells (Fig. 4B and ref. 37). Finally, Ag-specific CD11b⁺ B cells expressed IgG3 and CD138 following immunization, demonstrating the potential for B-1b cells to undergo isotype switching and produce Ab in response to signals from a synthetic TI-2 Ag alone, consistent with previous findings employing adoptive transfer experiments (14). Thus, B-1b cells become activated, expand, undergo isotype switching, and differentiate into Ab-producing cells in response to TNP-Ficoll.

In addition to examining the phenotype of responding B cells, the dynamics of expansion, differentiation, and contraction of Ag-specific B cells in response to TNP-Ficoll was investigated. Significant increases in CD11b⁺ B cell frequencies and numbers responding to TNP-Ficoll were observed in the peritoneal cavity, spleen, blood, and lymph nodes, with peak increases typically observed at d5. This is likely explained by mobilization of responding peritoneal B-1b cells as well as expansion of these and other non-peritoneal B-1b cells, found in tissues such as lymph nodes (14, 15). Indeed, that i.p. and i.v. immunization elicited similar increases in Ag-specific B-1b cells (Fig. 2), highlights a potential role for non-peritoneal B-1b cells in TI-2 Ab responses, as supported by other findings (19). Following peak expansion, Ag-specific B cell numbers and frequencies gradually declined in the spleen, although frequencies remained elevated over naïve mice and Ag-specific B-1b cells expressing low levels of CD11b were still present 5 weeks post-immunization (Fig. 1A, D–E). This is not unexpected as CD11b expression diminishes on B-1 cells outside of the peritoneum (35, 36). It is possible that these remaining cells are long-lived plasmablasts, memory cells, or cells continuing to participate in the primary response to TNP-Ficoll, which may resist degredation in vivo. In a previous study, Maclennan and colleagues used immunohistochemistry to demonstrate that NP-reactive B cells can persist in splenic tissue of Rag-1^−/− mice as either plasma cells or plasmablasts for at least 2 months in response to NP-Ficoll, which is known to involve B-1b cells (15), (46). Infection models using Enterobacter cloacae (16) and Borrelia hermsii (17) similarly demonstrate that B-1b cells may yield a form of unconventional memory that may be attributed to the presence of long-lived splenic B-1b plasmablasts. Nonetheless, it is evident from our study that although splenic ASC produce the majority of TI-2 Ab in the early stages of the immune response, bone marrow ASC make a substantial contribution to persistent TI-2 Ab levels. Importantly, the degree to which B-1b cells contribute to bone marrow ASC in response to TNP-Ficoll is not resolved by the current study. Notably, B-1b cells have been shown to give rise to BM ASC in response to PPS-3 (40) and B-1 BLIMP-1⁺ B cells can seed the long-lived ASC compartment (47). However, whether B-1b cell-derived BM ASC represent plasmablasts or fully differentiated BM plasma cells is not clear, given a recent report demonstrating that plasmablasts also reside in the BM (48). In contrast to results with spleen B-1b cells, Ag-specific peritoneal B-1b cell numbers remained elevated at peak levels up to ~6 weeks following immunization (Fig. 1E and 5C). This long-term maintenance of peritoneal Ag-specific B-1b cells following immunization has not been observed (49) or examined (50, 51) in previous studies using BCR transgenic mouse strains to examine B cell expansion in response to purified TI-2 Ags, although it was reported for α-1,3 dextran specific B-1b cells in V_HJ558 Tg mice following Enterobacter cloacae challenge (16). Moreover, accumulation and maintenance of Ft-LPS-specific peritoneal B-1a cells 2 months following Ft-LPS immunization was reported by Cole et al. (22). The explanation and significance of this finding is presently unclear, but studies examining the functionality and role of these peritoneal cells are underway. In summary, Ag-specific B-1b cell numbers significantly increase in multiple tissues during the first week following TI-2 Ag immunization and with the exception of the peritoneal B cells, gradually decline thereafter. Further studies aimed at examining the signals controlling Ag-specific B-1b cell expansion, contraction, and differentiation, and long-term maintenance are warranted.

The results of this study demonstrate a significant role for PD-1:PDL interactions in regulating B-1b cell function. First, BCR crosslinking induced PD-1 expression on B-1b cells and PD-1 significantly inhibited BCR-induced proliferation when co-crosslinked with the BCR without enhancing apoptosis (Fig. 4), in a manner similar to that observed for splenic B cells (Fig. 4 and ref. (38)) Second, Ag-specific B-1b cells in spleen, peritoneal cavity, and blood were specifically induced to express PD-1 following immunization, with the highest levels observed between days 2–5. Finally, interfering with PD-1:PDL interactions by administering a PD-1 blocking mAb between days 1–6 post immunization significantly increased Ag-specific B-1b cell numbers, the frequency of Ag-specific B-1b cells switching to IgG3, the frequency of Ag-specific splenic CD138⁺ cells, and Ag-specific IgG production by both splenic and bone marrow Ab-secreting cells (Fig. 5). These increases were accompanied by significant increases in Ag-specific serum IgG, as well as IgA, levels (Fig. 6). While it is possible that anti-TNP BM and spleen AFC are largely derived from Ag-specific B-1b cells, whether this is the case, and whether increased BM or spleen AFC observed with PD-1 mAb blockade are due to expanded peritoneal B-1b cells, remains unresolved. That transient PD-1 mAb blockade applied during the first week of immunization resulted in significantly increased Ag-specific B-1b cell numbers and IgG-producing ASC at both early and late timepoints suggests that 1) PD-1:PD-L interactions play a critical role in inhibiting the early immune response to TI-2 Ags and 2) that the splenic and BM IgG ASC generated during the early response to TI-2 Ags are long-lived.

PD-1:PD-L interactions may limit Ag-specific B-1b cell expansion and IgG production by several mechanisms. Since in vitro proliferation assays support an inhibitory role for PD-1 in regulating BCR-driven proliferation in B-1b cells, simultaneous interaction between a TI-2 Ag-activated (PD-1-expressing) B-1b cell and a PD-L-expressing cell provides a likely mechanism by which PD-1:PD-L interactions may limit Ag-receptor signals that drive B-1b cell proliferation in vivo. Importantly, whether TNP-specific IgG production by additional B cell subsets is modulated by PD-1 is not clear. As B cell division is required for isotype switching (52, 53), PD-1 inhibitory signaling in TI-2 Ag-activated B cells would thereby limit both clonal expansion and isotype switching. Indeed this is what is observed. Notably, if TI-2 Ags are associated with PAMPs or elicit T cell help via protein association, costimulatory signals (eg., LPS, CD40L) may help to overcome PD-1 inhibitory signals (Fig. 4G–H) or upregulation in certain cases (38). Although PD-1 effects have most often been studied in the context of Ag receptor signaling, PD-1 has effects on other non-Ag-receptor bearing cell types and must therefore regulate additional signaling pathways. Thus, PD-1 may modulate B cell survival, proliferation, and/or differentiation independently of its effects on Ag-receptor signaling. Interestingly, PD-1 blockade in macaques during chronic SIV infection has been proposed to increase SIV Env–binding Ab titers (54) as well as other humoral memory responses by preventing deletion of activated PD-1⁺ memory B cells (55). Although the in vitro results in this study (Fig. 4F) demonstrate that PD-1-BCR co-ligation suppresses proliferation as opposed to survival, it remains possible that PD-1 influences Ag-specific B-1b cell survival in vivo following activation. Finally, it remains possible that PD-1 expressed by some other cell type regulates B-1b cell responses to TI-2 Ags. For example, PD-1 expression by T follicular helper cells, as opposed to B cells, plays a significant role in promoting (as opposed to suppressing) germinal center B cell responses and plasma cell formation, and hence Ab production, in response to T cell dependent antigens (56). Future experiments using PD-1 conditional knockouts or mixed bone marrow chimeras will be required to test whether PD-1 expression by B-1b cells contributes to suppression of TI-2 Ab responses.

PD-1 shares functional similarity to other B cell inhibitory receptors expressed on B-1b cells, including CD5 and CD22, which recruit and activate SHP-1 to dampen BCR signaling. Although their expression patterns differ, these receptors are likely involved in maintaining B cell tolerance as mice deficient in any of these receptors have hyperresponsive B cells that produce autoantibodies (57–66). Interactions between these B-cell expressed inhibitory receptors and host cell-surface ligands likely function in suppressing Ab responses against self-Ags. Nonetheless, PD-1, CD5, and CD22 may have distinct roles in regulating Ab responses to immunogenic TI-2 Ags versus self-Ags, as TI-2 Ab responses are augmented in PD-1^−/− mice (29), normal in CD5^−/− mice (67), and decreased in CD22^−/− mice (63–65, 68). However, recent work by Nemazee and colleagues indicates that CD22 promotes tolerance to TI-2 Ags that decorated with sialic acid (self) ligands (68). Thus, PD-1 and other B cell immunoinhibitory receptors may play complex roles in regulating B cell responses against self versus foreign TI-2 Ags.

In summary, this study reveals a immunoinhibitory role for PD-1:PD-L interactions in regulating B-1b cell responses to TI-2 Ags, with particular significance in suppression of long-lasting Ag-specific IgG production. These findings may have important implications for current strategies targeting PD-1:PD-L interactions as treatment modalities for multiple diseases and conditions. More importantly, these findings may have significance for TI-2 Ag-based vaccine development. TI-2 Ab responses in mice parallel those in humans (11) and a human B-1 cell counterpart has recently been identified (69). Thus, it is possible that PD-1:PD-L interactions similarly suppress TI-2 IgG responses in humans. In contrast to IgM, IgG is produced in limiting quantities against most TI-2 Ags. Nonetheless, as IgG may elicit enhanced protection against carbohydrate-bearing pathogens relative to IgM, future strategies transiently targeting the PD-1:PD-L pathway may provide an opportunity to elicit enhanced IgG-mediated protection against TI-2 Ag-bearing pathogens.