Type I IFN Receptor Signaling on B Cells Promotes Antibody Responses to Polysaccharide Antigens

This article is featured in Top Reads, p.

Pneumococcal vaccines consist of Streptococcus pneumoniae–derived capsular polysaccharides (PPS) that behave as T cell–independent type 2 Ags (TI-2 Ags). Pneumovax23 (Merck), composed of 23 types of PPS, provides protection against invasive pneumococcal disease in adults for up to 10 y, with an estimated efficacy of 60–70% (1). Adjuvants for native polysaccharide vaccines have not been used in the clinic. Alum does not boost primary Ab responses to TI-2 Ags (2) and poorly boosts responses to PPS conjugates (3). Earlier work with Ribi adjuvant, consisting of Salmonella typhimurium monophosphoryl lipid A (MPL) and Mycobacterium cord factor in squalene, elicited increased primary PPS-specific Ab responses in mice (4, 5), but Ribi adjuvant is not suitable for use in humans because of its toxicity. We recently demonstrated that low-toxicity Salmonella minnesota MPL and synthetic cord factor analog trehalose-6,6′-dicorynomycolate (TDCM) emulsified in squalene significantly increased Ag-specific IgM and IgG levels in response to PPS and haptenated Ficoll (6). The mechanisms by which this adjuvant carries out its effects have not been fully elucidated, but they are important for informed design of adjuvants for use with polysaccharide Ags in humans.

We previously reported MPL/TDCM-induced type I IFN production in vivo (7), likely through the effects of MPL on TLR4–Toll/IL-1R domain-containing adapter inducing IFN-β (TRIF) activation of type I IFN production (8). Importantly, a previous study demonstrated that the TLR3 agonist polyinosinic-polycytidylic acid [poly(I:C)] increased Ab responses to (4-hydroxy-3-nitrophenyl)-acetyl (NP)-Ficoll via a mechanism dependent upon IFNAR signaling and follicular B cells (9). To further understand the mechanisms by which MPL/TDCM increases Ab responses to TI-2 Ags, we investigated the importance of type I IFN and B cell subsets in adjuvant effects. We unexpectedly found a role for B cell–intrinsic IFNAR expression in supporting B cell responses to nonadjuvanted TI-2 Ags. B-1b cells, which express significantly higher levels of IFNAR than B-2 cells, were particularly dependent on type I IFN signaling for optimal activation, survival, expansion, and differentiation in response to TI-2 Ags. However, B-1b cell development did not appear to be impacted by IFNAR deficiency. In contrast to reported adjuvant effects of poly(I:C) on follicular (CD23⁺) B cell responses to TI-2 Ags, we found that MPL/TDCM adjuvant effects were carried out largely through innate B cells (B-1b and CD23⁻ splenic B cells). However, type I IFNAR signaling was not required for MPL/TDCM adjuvant effects. These findings highlight the importance of type I IFN in supporting B-1b cell Ab responses to nonadjuvanted TI-2 Ags, but they reveal that B cell–activating adjuvants may activate alternative pathways supportive of activation, survival, and differentiation that ultimately overcome the requirement for IFNAR-derived signals.

Wild-type (WT), mumt (Ighm^tm1Cgn), CD19^Cre(B6.129P2(C)-Cd19^tm1(cre)Cgn/J), and B1-8hi IgH knock-in (V_HB1-8 Tg) mice were on a C57BL/6 background (The Jackson Laboratory). Ifnar1^−/− and floxed Ifnar1^fl/fl (Ifnar1^tm1Uka) mice were on a C57BL/6 background and were kindly provided by Dr. Erik Barton (originally from Drs. Herbert Virgin and Ulrich Kalinke, respectively) and used as previously described (7), with Cd19^tm1(cre)Cgn/J mice crossed with Ifnar1^fl/fl mice to generate heterogeneous cd19-cre^+/−/Ifnar1^fl/+ breeders. These breeders were used to generate male mice deficient for Ifnar1 on B cells (Cd19-Cre^+/−/Ifnar1^fl/fl), male mice expressing heterozygous Ifnar1 levels on B cells (Cd19-Cre^+/−Ifnar1^fl/+), and male littermate controls (Cd19-Cre^+/−) for experiments. Homozygous WT and Ifnar1^−/− breeders were used to generate mice for experiments, the majority of which used male mice. Studies including female mice used sex matching among groups. Studies used age-matched mice housed in a specific pathogen-free facility and were approved by the Wake Forest School of Medicine Animal Use Committee.

Mice were immunized with 10 μg 2,4,6-Trinitrophenyl (TNP)₆₅-Ficoll, 1–5 μg NP-Ficoll, or the equivalent of 0.125–1 μg each PPS within Pneumovax23. In one case, a mixture of PPS (type 3, 4, 6B, 8, 9N, 12F, 14, 19F, 23F; American Type Culture Collection) was used to generate Pneumovax9 because of a Pneumovax23 vaccine shortage. MAR1-5A3 mAb mouse IFNAR-1 (BioXCell) or control mouse IgG (Jackson ImmunoResearch) was administered i.p. (200 μg on day [d]0 and 100 μg on d2 and d4 of immunization) to block IFNAR1 signaling where indicated. Adjuvant containing 20 μg S. minnesota MPL and 20 μg TDCM in 0.5% squalene/0.05% Tween 80 or 2% squalene/0.2% Tween 80 (Sigma-Aldrich) for i.p. and i.m. injections, respectively, was mixed with Ag prior to injection. The ELISAs were performed as previously described (6, 10, 11). TNP- and PPS-specific Ig levels were estimated using a standard curve generated using anti-mouse Ig (H+L)-coated wells in conjunction with mouse IgM and IgG isotype standards (Southern Biotechnology Associates). NP-specific Ig concentrations were estimated using NP-specific IgM and IgG standard curves as previously described (11).

Peritoneal cells were harvested using 10 ml of Dulbecco’s PBS to lavage the peritoneal cavity. Blood was collected in heparin, with PBMCs purified using a 1083 Ficoll density gradient (Sigma-Aldrich). Spleen and bone marrow suspensions were lysed in ammonium chloride lysis buffer, followed by washing in staining buffer (PBS containing 2% newborn calf serum). Cells were resuspended in staining buffer and preincubated with 0.5 µg/ml Fc Block (eBioscience) and stained with fluorochrome-conjugated mAbs (BioLegend, eBioscience, and BD Biosciences): CD19 (1D3), CD11b (M1/70), CD5 (53-7.3), CD23 (B3B4), CD21/35 (7E9), CD138 (281-2), CD86 (GL1), CD19 (1D3), CD11b (M1/70), CD45R/B220 (RA3-6B2), rat anti-mouse IgG (pooled rat anti-mouse IgG1, IgG2b, IgG2a, IgG3; Southern Biotechnology Associates), and Live/Dead Fixable Aqua stain. CountBright beads (Thermo Fisher) were included for enumeration. After staining, samples were washed and fixed in 1.5% buffered formaldehyde. For TNP- and NP-specific B cell staining, cells were incubated with TNP(65)-FL-AECM (aminoethyl carboxy methyl)-Ficoll (20 μg/ml; Biosearch Technologies) or NP₄₀-allophycocyanin and anti-mouse CD45.1 (12). Intracellular Ki-67 (solA15) and rabbit anti-active caspase 3 (BD Biosciences) staining was performed according to the manufacturer’s instructions (eBioscience fix/permeabilization kit). Cells were analyzed using a BD LSR Fortessa X-20 (Becton Dickinson), and data were analyzed using FlowJo analysis software (BD Biosciences).

WT recipient mice were lethally irradiated (950 rad; single dose [¹³⁷Cs] γ-irradiator) and reconstituted i.v. with 10⁷ total bone marrow cells consisting of muMT bone marrow mixed with either WT or Ifnar1^−/− bone marrow in a 90:10 ratio 5–6 h later, as previously described (13). Recipient mice were maintained on Septra 1 wk prior to irradiation and 2 wk afterward. Mice were rested for 4 wk prior to immunization.

Naive splenic and peritoneal B cells were purified from V_HB1-8 Tg mice using CD43 bead depletion (Dynal) because CD43 is largely lacking on naive NP-specific B1b, marginal zone (MZ), and follicular B cells from these mice (11). CD23⁺ and CD23⁻ B cells were further purified using Miltenyi Biotec beads. B-1b cells were further enriched using CD11b Miltenyi Biotec beads. V_HB1-8 Tg B cells were transferred i.v. into CD45.2⁺ mice unless otherwise indicated. Naive spleen B cells were purified from WT mice in a similar manner. WT peritoneal B cells were obtained using EasySep pan–B cell isolation (STEMCELL Technologies) combined with biotinylated anti-F4/80 and CD23 mAbs. A second purification step was used for B-1b cell enrichment using biotinylated Abs against CD5, F4/80, and GR1 in conjunction with streptavidin Dynabeads (>90% CD5⁻CD19⁺ B cell purity; 70–85% CD11b⁺). Recipient muMT mice were reconstituted with cells i.v.

Naive CD43⁻ V_HB1-8 Tg spleen (1 × 10⁷) and peritoneal (5 × 10⁵) B cells were purified as described above and adoptively transferred into Ifnar1^−/− recipient mice. Mice were immunized with 25 μg NP₄₀-Ficoll the next day. Splenocytes were harvested and placed in culture with media alone or with type I IFN-α/β (600 U/ml; murine IFN-α/β, NR-3082 obtained through BEI Resources, National Institute of Allergy and Infectious Diseases, National Institutes of Health). Twenty-four hours later, cells were harvested and stained.

Data are shown as mean ± SEM. Unless indicated otherwise, differences between sample means were assessed using Student t test. Where indicated, comparisons among multiple groups were made using one-way ANOVA with Tukey’s post hoc test, and comparisons over time used repeated measures two-way ANOVA (Prism version 7 software).

Before investigating the role of type I IFN signaling in MPL/TDCM adjuvant effects, we first investigated whether type I IFN played a role in regulating Ab responses to TI-2 Ags. As shown in (Fig. 1A, mice lacking IFNAR1 and thus functional IFNAR (Ifnar1^−/−) exhibited a significant reduction (∼30%) in TNP-specific IgM and IgG levels 7 d after TNP-Ficoll immunization, although responses approximated WT levels at later time points. However, in response to Pneumovax23 immunization, Ifnar1^−/− mice produced significantly less (50%) IgM and IgG to serotype 3 polysaccharide (PPS3) found within the vaccine as well as significantly less IgM and IgG reactive with total Pneumovax23, and these differences were sustained (Fig. 1B, 1C). Administration of an IFNAR1 blocking mAb to WT mice at the time of immunization was also found to significantly reduce PPS3 and Pneumovax23-specific IgM responses 5 d after immunization, and PPS-specific IgG levels, although low, were also reduced (Fig. 1D). Thus, IFNAR supports optimal TI-2 Ab responses and is particularly important for optimal IgM and IgG responses to pneumococcal polysaccharides.

Because alterations in innate B cell subsets can lead to defective TI-2 Ab responses, we assessed B cell subsets in Ifnar1^−/− mice. As shown in (Fig. 2, we found total B cell frequencies and numbers in the spleen and peritoneal cavity were similar between WT and IFNAR-deficient mice. Furthermore, we did not find differences in B cell subsets. Peritoneal B-2 cell and spleen follicular B cell frequencies and numbers were similar in WT and Ifnar1^−/− mice, as were innate peritoneal B-1a and B-1b and splenic MZ B cell frequencies and numbers (Fig. 2). Splenic B-1a cell frequencies and numbers were also similar between WT and Ifnar1^−/− mice (Supplemental Fig. 1A). Thus, B-1 and B-2 B cell subset development and maintenance appear to be normal in Ifnar1^−/− mice.

IFNAR expression is ubiquitous, and, hence, altered TI-2 Ab responses in Ifnar1^−/− mice could be due to lack of IFNAR on one or more cell types. We investigated the importance of B cell–expressed IFNAR in TI-2 Ab responses by crossing Ifnar1^fl/fl mice with Cd19-Cre transgenic mice to generate mice lacking IFNAR only on CD19-expressing B cells (Cd19-Cre^+/−Ifnar1^fl/fl) or expressing heterozygous IFNAR on B cells (Cd19-Cre^+/−Ifnar1^fl/+). As shown in (Fig. 3, Cd19-Cre^+/− mice produced lower Ab responses to TNP-Ficoll and PPS relative to WT mice and were therefore used to draw comparisons with Cd19-Cre^+/−Ifnar1^fl/fl mice. In response to TNP-Ficoll, Cd19-Cre^+/−Ifnar1^fl/fl mice produced significantly less TNP-specific IgG relative to Cd19-Cre^+/− and Cd19-Cre^+/−IFNAR^fl/+ mice (Fig. 3A). In response to Pneumovax23, Cd19-Cre^+/−Ifnar1^fl/fl mice produced significantly less PPS3-specific IgM, and PPS3- and Pneumovax23-specific IgG levels were 50% decreased relative to Cd19-Cre^+/− mice (Fig. 3B, 3C). Interestingly, Cd19-Cre^+/−Ifnar1^fl/+ mice expressing heterozygous IFNAR levels on B cells produced TI-2 Ab levels similar to Cd19-Cre^+/− mice, suggesting that heterozygous IFNAR expression on B cells is sufficient to support B cell responses to TI-2 Ags, whereas selective deficiency of IFNAR on B cells results in impaired Ab responses to TI- 2 Ags.

We complemented the above studies by generating radiation bone marrow chimeras. Although bone marrow chimeras may not support complete B cell reconstitution (e.g., fetal-derived B-1 cells), bone marrow progenitors support the eventual development of splenic MZ B cell and peritoneal B-1b cell populations and, to a much lesser extent, B-1a cells (14, 15) (Supplemental Fig. 1B). PPS-specific Ab responses are therefore reconstituted in bone marrow chimeras, albeit to a lesser extent than elicited in nonirradiated animals (6) (Fig. 3). Given the reconstitution of TI-2–specific Ab responses in bone marrow chimeras, we assessed responses in chimeras constructed by reconstituting irradiated WT mice with bone marrow from B cell–deficient mumt mice mixed with either bone marrow from WT mice or Ifnar1^−/− mice (90:10 ratio). As shown in (Fig. 3D, chimeras with B cells lacking IFNAR produced significantly less IgG (30% reduced) in response to TNP-Ficoll than chimeras reconstituted with WT B cells. IgM and IgG responses to PPS3 were also decreased three- to fourfold in chimeras with IFNAR-deficient B cells, and in Pneumovax-specific ELISAs, IgG was barely detectable in these chimeras (Fig. 3E, 3F). Thus, mice with selective IFNAR deficiency on B cells exhibit impaired IgM and IgG responses to TI-2 Ags to a degree similar to that in mice lacking IFNAR on all cells. Finally, reconstitution of Ifnar1^−/− mice with B cells from Ifnar1^+/+ V_HB1-8 Tg mice in which a fraction of B cells (5–10%) coexpress the λ1 chain with the V_HB1-8 transgene to yield a high-affinity receptor specific for NP (11) yielded NP-specific IgG responses that were similar to those in WT recipients (Fig. 3G), indicating that loss of non–B cell IFNAR expression did not noticeably impact this TI-2 Ab response. Thus, B cell–expressed IFNAR is required for optimal responses to TI-2 Ags.

We assessed the early B cell response to TNP-Ficoll in WT and Ifnar1^−/− mice to determine if defects in activation, expansion, and/or differentiation could be identified. As shown in (Fig. 4A, naive WT and Ifnar1^−/− mice had similar frequencies of TNP-specific CD19⁺ B cells in the spleen (representative gating strategy shown in Supplemental Fig. 1C). However, 5 d after immunization, TNP-specific CD19⁺ B cell frequencies expanded threefold in WT mice but only twofold in Ifnar1^−/− mice (Fig. 4A). We noted similar decreases (∼30%) in the frequencies of TNP-specific CD138⁺ and IgG⁺CD138⁺ plasmablasts as well as the frequencies of dividing (Ki67⁺) TNP-specific B cells, IgG⁺ B cells, and CD138⁺ B cells (Fig. 4A).

Alterations in early responses to TI-2 Ags may be attributed to changes in innate B cell responses. CD11b expression is present on B1 cells several days after entry into the spleen (16, 17). Consistent with this, V_HB1-8 Tg B-1b cells adoptively transferred into WT recipients express CD11b in the spleen up to 7 d after transfer (6 d after NP-Ficoll immunization), in contrast to adoptively transferred V_HB1-8 Tg MZ and follicular B cells (Supplemental Fig. 1D). Thus, CD11b expression may be used to distinguish hapten-specific B-1 cells from MZ and follicular B cells during the early splenic response to TI-2 Ag, although it may not capture all B-1 cells. As shown in (Figure 4B, the frequency of TNP-specific CD19⁺CD11b⁺ B cells in the spleen was significantly reduced (twofold) in Ifnar1^−/− relative to WT mice 5 d after immunization. In contrast, no significant difference was observed for frequencies of TNP-specific CD19⁺CD11b⁻ B cells after immunization (Fig. 4B). Cell numbers of splenic TNP-specific CD19⁺CD11b⁺ B cells were similarly reduced (greater than twofold) in Ifnar1^−/− mice (Supplemental Fig. 1E). The frequencies of Ki67⁺, CD138⁺, and IgG⁺ TNP-specific CD19⁺CD11b⁺ cells among splenocytes was significantly decreased in Ifnar1^−/− mice due to decreased overall TNP-specific CD19⁺CD11b⁺ frequencies (Fig. 4B). Overall splenic frequencies of Ki67⁺, CD138⁺, and IgG⁺ TNP-specific CD19⁺CD11b⁻ B cells were not different between Ifnar1^−/− and WT mice. Thus, IFNAR deficiency results in reduced participation of CD11b⁺ B cells in the response to haptenated Ficoll.

CD11b⁺ B cells in peritoneal cavity and spleen express significantly higher (twofold) levels of IFNAR1 relative to CD11b negative (largely B-2) B cells (Fig. 5A). Given our data supporting normal numbers of peritoneal B-1b cells in Ifnar1^−/− mice (Fig. 2) and a role for B cell–intrinsic IFNAR expression in promoting TI-2 Ab responses (Fig. 3), we next sought to determine the extent to which IFNAR expression on B-1b cells regulates early TI-2 Ab responses. To do this, we isolated and adoptively transferred CFSE-labeled peritoneal B-1b cells (CD19⁺CD11b⁺CD23⁻) from IFNAR-sufficient V_HB1-8 Tg mice into Ifnar1^−/− mice and treated recipient mice with a control or IFNAR1 blocking Ab at the time of NP-Ficoll immunization. As shown in (Fig. 5B, IFNAR blockade significantly decreased frequencies of CD45.1⁺ NP-specific B-1b cells in the peritoneal cavity, spleen, blood, inguinal lymph node, and bone marrow on d5 relative to control Ab-treated mice (representative gating strategy shown in Supplemental Fig. 1F). These differences were reflected in decreased CD45.1⁺ NP-specific cell yields in mice that had received IFNAR1 mAb blockade. However, we did not detect differences in division as measured by CFSE loss (data not shown). IFNAR blockade also significantly decreased CD86 expression levels on CD45.1⁺ NP-specific peritoneal (fourfold), blood (twofold), and spleen (1.5-fold) B-1b cells (Fig. 5C). In addition, IFNAR blockade significantly decreased frequencies of IgG⁺ and CD138⁺ CD45.1⁺ NP-specific B-1b cells in the peritoneal cavity, spleen, and blood relative to control IgG-treated mice (Fig. 5C). Thus, IFNAR expression by B-1b cells supports activation, expansion, and differentiation in response to TI-2 Ag.

To investigate direct effects of type I IFN on B cells responding to TI-2 Ags, we adoptively transferred V_HB1-8 Tg B-1b and splenic B cells into Ifnar1^−/− mice, and, 5 d after immunization, we assessed the effects of type I IFN on these cells in vitro. After 24 h of culture, the overall frequencies of total, IgG⁺, and CD11b⁺ CD45.1⁺ NP-specific B cells was not changed by addition of type I IFN (Fig. 5D). However, the frequency of CD138⁺ cells was significantly increased (threefold), whereas the frequency of active caspase-3⁺ staining was significantly decreased, among CD45.1⁺ NP-specific B cells in type I IFN–treated cultures (Fig. 5D). We assessed active caspase 3 staining for CD11b⁺ (B-1b) and CD11b⁻ cells and found that type I IFN significantly decreased the frequency of staining among CD11b⁺, but not CD11b⁻, CD45.1⁺ NP-specific B cells (Fig. 5E).

Type I IFN significantly increased the frequency of CD138⁺ cells among CD11b⁺CD45.1⁺ NP-specific B cells but had no effect on the frequency of IgG⁺ cells among CD11b⁺CD45.1⁺ NP-specific B cells (Fig. 5F). However, type I IFN significantly decreased caspase 3 staining in both CD138⁺ and IgG⁺CD11b⁺CD45.1⁺ NP-specific B cell populations. Significantly increased viable CD138⁺ and IgG⁺ CD11b⁺CD45.1⁺ NP-specific cell counts in these cultures corroborated these findings (Fig. 5F). Thus, type I IFN increases differentiation of TI-2 Ag–activated B cells to CD138⁺ Ab-secreting cells (ASCs) and promotes survival of B-1b cells responding to TI-2 Ags.

MPL/TDCM induces type I IFN production in vivo and does not optimally enhance TI-2 Ab responses in mice lacking the TRIF adapter, which is an important inducer of IFN-β (8, 18). This, along with results from a previous study demonstrating type I IFN–mediated signaling on B cells, was critical for poly(I:C)-mediated adjuvant effects in increasing IgG responses to NP-Ficoll (9) and raised the possibility that type I IFN signaling was also critical for MPL/TDCM adjuvanticity. Given that IFNAR deficiency resulted in a more significant impairment in Ab responses to PPS relative to haptenated Ficoll and previous work showing that MPL/TDCM has more significant effects on increasing Ab responses to PPS relative to haptenated Ficoll (6), we assessed MPL/TDCM effects on responses of Ifnar1^−/− mice to Pneumovax23 delivered i.m. As shown in (Fig. 6, Ifnar1^−/− mice exhibited significantly impaired PPS3- and Pneumovax23-specific IgM and IgG responses to Pneumovax23 delivered i.m. (Fig. 6), similar to results obtained for i.p. immunization (Fig. 1). However, inclusion of MPL/TDCM yielded significantly increased primary and secondary IgM and IgG responses to PPS3 and Pneumovax23, although IgG responses were variable. Two-way repeated measures ANOVA with vaccine response (Pneumovax23 or Pneumovax23 + adjuvant) and time point as factors indicated that both factors and their interaction were significant for WT and Ifnar1^−/− groups (p < 0.05), and these analyses specifically support the conclusion that Ifnar1^−/− mice receiving adjuvant had significantly higher overall levels of Pneumovax23- and PPS3-specific IgG and IgM relative to their counterparts that had received Pneumovax23 without adjuvant. Thus, IFNAR is not required for MPL/TDCM adjuvant effects in the context of PPS-specific Ab responses.

Given the above findings and previous work demonstrating that a type I IFN–activating adjuvant elicited its IFNAR-dependent effects on Ab responses to NP-Ficoll through activating follicular B cells (9), we questioned whether MPL/TDCM enhanced TI-2 Ab responses by supporting Ab production by distinct B cell subsets. We found that splenic and peritoneal NP-specific V_HB1-8 Tg B cells, particularly CD11b⁺ peritoneal B-1b cells, were responsive to MPL/TDCM effects on ASC differentiation in the context of TI-2 Ag activation in vitro, as evidenced by increased BLIMP1 expression by these cells in the presence of adjuvant (Supplemental Fig. 1G, 1H).

To determine whether MPL/TDCM elicited its effects through select subsets in vivo, we performed adoptive transfer experiments using B cells from V_HB1-8 Tg mice (11). Naive CD43⁻ peritoneal B-1b cells and CD23⁺ (enriched for follicular B cells) and CD23⁻ spleen B cells (enriched for MZ B cells) from V_HB1-8 Tg mice were adoptively transferred i.v. into WT recipients, which were then immunized with NP-Ficoll (i.p.). Importantly, λ1-expressing (NP-specific) B cells are similarly represented among these subsets (11). B-1b cells produced the highest level of IgM and IgG in response to NP-Ficoll (Fig. 7A). Inclusion of MPL/TDCM also significantly increased NP-specific IgM^a (produced by Tg cells) and IgG (fivefold) by B-1b and CD23⁻ spleen B cells. However, MPL/TDCM had no effect on Ab production by CD23⁺ spleen cells.

We next assessed whether i.m. immunization would impact Ab responses by distinct subsets. In response to NP-Ficoll delivered i.m., B-1b cells produced the highest amount of NP-specific IgM^a, whereas CD23⁻ B cells produced the most IgG (Fig. 7B). MPL/TDCM significantly increased NP-specific IgM^a production from B-1b cells (1.5-fold) over recipient mice that received adjuvant alone. Increases in IgM^a in recipients of CD23⁻ and CD23⁺ cells were moderate. MPL/TDCM increased NP-specific IgG production in recipients of each cell type, with d7 levels increased greater than sixfold in B-1b and ninefold in CD23⁺ recipients but less than twofold in CD23⁻ B cell recipients. Thus, in the NP-Ficoll immunization system using high-affinity V_HB1-8 Tg B cells, B-1b and CD23⁻ spleen cells preferentially differentiate into ASCs, regardless of immunization route, and MPL/TDCM further potentiates their level of Ab production, although there is nonetheless some responsiveness to MPL/TDCM in the V_HB1-8 Tg CD23⁺ B cell population after i.m. immunization.

We next examined effects of MPL/TDCM on responses of B cell subpopulations to the Pneumovax23 pneumococcal vaccine. In muMT mice reconstituted with high numbers of spleen cells (3 × 10⁷), IgM and IgG levels against Pneumovax23 as well as PPS3 were not significantly different among recipients that received no immunization, Pneumovax only, or MPL/TDCM only (i.p.), whereas mice receiving Pneumovax plus MPL/TDCM i.p. produced significantly increased Pneumovax-specific IgM and IgG (Fig. 7C). PPS3-specific IgM and IgG were also increased, although not significantly. Thus, there is a PPS-specific spleen B cell population that responds to MPL/TDCM adjuvant effects.

When examining sorted B cell populations, we found that recipients of B-1b cells produced the highest level of Pneumovax23- and PPS3-specific IgM and PPS3-specific IgG in response to Pneumovax23 i.p. (Fig. 7C). MPL/TDCM significantly increased total Pneumovax23-specific IgM in B-1b recipients, but not in CD23⁻ and CD23⁺ B cell recipients. Similarly, MPL/TDCM significantly increased PPS3-specific IgM production in B-1b cell recipients and, to a lesser extent, in CD23⁻ but not CD23⁺ B cell recipients. PPS3- and Pneumovax-specific IgG was low, but MPL/TDCM produced moderate increases in some recipients.

In response to Pneumovax23 plus adjuvant delivered i.m., recipients of high-dose splenocytes produced moderately increased Pneumovax23- and PPS3-specific IgM but not IgG (Fig. 7D). Recipients of B-1b cells and splenic CD23⁻ B cells produced significantly increased PPS3- and Pneumovax23-specific IgM in response to adjuvant (Fig. 7D). MPL/TDCM significantly increased PPS3-specific IgG as well as increased Pneumovax23-specific IgG selectively in B-1b cell recipients. CD23⁺ B cell recipients were not very responsive. Thus, MPL/TDCM adjuvant effects observed in the context of Pneumovax23 immunization are carried out largely through innate B cell populations.

Ab responses to polysaccharide Ags are critical for protection against encapsulated pathogens. However, our understanding of the factors regulating these responses remains incomplete. Our present study, to our knowledge, reveals several novel findings regarding regulation of TI-2 Ab responses. First, we demonstrate type I IFN signaling on B cells, and in particular on B-1b cells, promotes optimal Ab responses to TI-2 Ags, including pneumococcal polysaccharides and does so via support of early B cell activation, survival, and ASC differentiation. Second, we demonstrate an MPL-based adjuvant that significantly augments primary and secondary responses to TI-2 Ags (6), induces type I IFN production in vivo (7), and requires TRIF signaling for optimal adjuvant effects (6) but does not require type I IFN for its effects. Finally, we demonstrate that MPL/TDCM elicits its adjuvant effects largely by promoting Ab production by innate-like B cells (i.e., B-1b and CD23⁻ B cells). Collectively, our results provide valuable information regarding the role of type I IFN in supporting polysaccharide-specific Ab responses as well as critical information that may be leveraged for designing strategies to improve vaccines targeting polysaccharide Ags in the future.

The role of IFNAR in regulating B cell responses is complex, based on findings reported for Ifnar1^−/− mice and in vitro studies of B cells. We show that IFNAR deficiency on a C57BL/6 background does not impact peritoneal or splenic B cell subset frequencies or numbers. An earlier study reported Ifnar1^−/− mice on a 129sv background had normal absolute B cell numbers in bone marrow and spleen; however, it indicated that the repertoire of immature bone marrow B cells was altered, and, furthermore, it showed that type I IFN could modulate the sensitivity of activated B cells to BCR-dependent inhibition of terminal differentiation resulting from LPS or CD40 + IL-4 stimulation (19). In contrast, another study reported that type I IFN in conjunction with IL-6 promotes plasma cell differentiation (20). Nonetheless, type I IFN protects B cells against apoptosis and has been reported to enhance BCR-mediated activation and proliferation in some studies, but it has the opposite effect in others (21–24). Thus, both stimulatory and inhibitory effects of IFN-α/β on in vitro B cell proliferation and Ig production have been reported. In some cases this may be explained by the fact that high-dose type I IFN is inhibitory, whereas low-dose IFN is stimulatory, for Ig production by activated B cells (25). Its positive role in supporting TNP- and NP-Ficoll–specific B cell activation and expansion in vivo in the absence of strong type I IFN production is consistent with this notion.

On the basis of our present findings using mAb blockade in both WT mice and Ifnar1^−/− V_HB1-8 Tg B cell recipient mice, as well as those obtained using mice selectively lacking type I IFN receptor on B cells (bone marrow chimeras and cre-flox mice), type I IFN regulates B cell TI-2 IgM and IgG responses directly, as opposed to impacting the development of the B cell populations that give rise to these responses. Work on the effects of type I IFN on TI Ab responses has been limited and largely restricted to antiviral responses. The effect of type I IFN on B cell responses to vesicular stomatitis virus (VSV), which displays highly ordered VSV-G and induces TI IgM responses, has been examined in two studies. One study demonstrated that IFN-α enhanced VSV-specific B cell IgM production in response to UV-inactivated VSV in vitro and showed that B cell–expressed IFNAR was required for optimal IgM but not IgG responses to live VSV in vivo (26). In contrast, in response to noninfective VSV-like particles (VLPs), IFNAR deficiency significantly reduced T cell dependent IgG responses but only slightly reduced IgM responses. This regulation was through B cell–extrinsic IFNAR expression, and the study reported little effect of B cell–expressed IFNAR on the TI Ab response to VSV-VLP (27).

B-1 cells express significantly higher levels of type I IFN receptor than do B-2 cells. Thus, it is perhaps not surprising that hapten-specific CD11b⁺ B-1 cells lacking IFNAR exhibited greater defects in activation, expansion, and survival than CD11b⁻ B cells. Indeed, IFNAR blockade on NP-specific B-1b cells in vivo significantly decreased their activation and expansion in all tissues examined except mesenteric lymph nodes. Interestingly, with regard to innate B1 cell production of Abs, IFNAR is required for B1 cell lymph node trafficking and IgM production in response to influenza virus infection due to its role in regulating CD11b conformational changes (28). However, IFNAR is not required for MPL-induced tumor-reactive IgM production by B1a cells (7), despite its dependence on the TLR4-TRIF activation pathway (29). We did not find differential representation of NP-specific B-1b cells among tissues (peritoneal cavity, spleen, blood, inguinal lymph node, mesenteric lymph node, or bone marrow) after IFNAR mAb blockade, and thus we have no evidence to support that trafficking is impaired in the absence of IFNAR signaling in the haptenated Ficoll model system. When taken together, our in vivo and ex vivo experimental results suggest that IFNAR supports increased activation of B-1b cells in the context of TI-2 Ag activation and promotes their expansion, likely through activating survival pathways (21). Moreover, our results indicate that type I IFN supports differentiation of TI-2 Ag–activated B-1b cells into ASCs. Our work does not rule out a role for type I IFN in supporting TI-2 Ab responses by other B cell subsets. Furthermore, although our results suggest that IFNAR regulates B cell responses to TI-2 Ag primarily through its expression on B cells, we did find that IgG responses were more consistently decreased relative to IgM responses when IFNAR was solely lacking on B cells, suggesting a potential role for IFNAR-derived signals on other cell types in supporting IgM production and/or a possible role for IFNAR in supporting class switching through its survival-promoting effects. Although type I IFN induction of class switching is possible, Stat1 signaling specifically facilitates IgG2a/c switching (30), which generally comprises a minor subclass produced in response to nonadjuvanted TI-2 Ags (6, 11). Interestingly, augmentation of IgG responses to T cell dependent Ag adjuvanted by IFN-α also requires IFNAR expression on B cells (31), whereas VSV-displaying VLPs induce IgG production via a mechanism dependent on IFNAR expression on non-B cells (27). Currently, it is not clear whether basal levels of type I IFN contribute to regulation of Ab responses to unadjuvanted TI Ags or if TI-2 Ags induce low levels of type I IFN that may further drive this regulation. It is certainly possible that the cell wall components that contaminate the pneumococcal polysaccharide vaccine induce type I IFN through TLR2 signaling (32). Further work is required to determine whether this is the case.

Although IFNAR1 deficiency yields a distinct phenotype in humans relative to mice at least with regard to mucosal viral infections (33), whether human B cells share a similar dependency on type I IFN for optimal TI-2 Ab responses remains unknown. Interestingly, a small sample study reported slightly higher geometric mean titer ratios for Pneumovax23 serotype-specific responses in patients with multiple sclerosis treated with type I IFN compared with patients treated with dimethyl fumarate (34). Future work with analysis of larger cohorts of patients either treated with IFN therapeutics or who have developed autoAbs against IFNs, along with studies defining the effects of mutations involved in altered type I IFN pathways (33, 35–37), will likely shed light on this.

We investigated the role type I IFN plays in the adjuvant effects of MPL/TDCM on TI-2 Ab responses in the present study because of (1) its dependency on TRIF signaling (6), (2) results of a previous study demonstrating the requirement for B cell–expressed IFNAR in the adjuvant effects of poly(I:C) on TI-2 Ab responses (9), and (3) the dependency of nonadjuvanted TI-2 Ab responses on B cell–expressed IFNAR identified herein. On the basis of adjuvant effects observed in Ifnar1^−/− mice, MPL/TDCM works independently of type I IFN signaling. The TLR4-dependent effects of MPL on activating both MyD88 and TRIF pathways in B cells (6), the latter of which activate NF-κB in addition to type I IFN production, may explain these findings. Our work demonstrates that MPL/TDCM significantly enhances Ab responses to polysaccharide Ag largely through its effects on innate-like B cells (i.e., B-1b and CD23⁻ spleen B cells). MPL/TDCM adjuvant effects on significantly increased PPS-specific IgG production by B-1b cells to Pneumovax23 delivered i.m. was particularly notable. B-1 and MZ B cells are most often found producing Abs against TI-2 Ags in the absence of adjuvant, so it is perhaps not surprising that this adjuvant further promotes the ability of these B cells to produce more IgM, and in some cases IgG, against these polysaccharide Ags (11, 38). In addition to being less responsive to TI-2 Ags relative to MZ and B-1 cells, follicular B cells are also less responsive to TLR agonists (39). In contrast to the mechanism defined for poly(I:C) involving B cell–expressed IFNAR (9), the MPL-based adjuvant functions through direct activation of TLR4 on B cells (6). This supports the possibility that combinations of distinct pattern recognition receptor agonists that work through different pathways may be optimized to significantly improve B cell responses to polysaccharide Ags in humans in the future.

The authors have no financial conflicts of interest.