MyD88 signaling is not essential for induction of antigen-specific B cell responses but is indispensable for protection against Streptococcus pneumoniae infection following oral vaccination with attenuated Salmonella expressing PspA antigen

Abstract

Toll-like receptors (TLRs) directly induce innate host defense responses, but the mechanisms of TLR-mediated adaptive immunity remain subject to debate. In this study, we clarified a role of TLR-mediated innate immunity for induction of adaptive immunity by oral vaccination with a live recombinant attenuated Salmonella enteric serovar typhimurium vaccine (RASV) strain expressing Streptococcus pneumoniae surface protein A (PspA) antigen. Of note, oral or intranasal vaccination with RASV expressing PspA resulted in identical or even significantly higher levels of PspA-specific IgG and IgA responses in the systemic and mucosal compartments of MyD88^−/− mice of either BALB/c or C57BL/6 background when compared with those of wild-type mice. Although PspA-specific CD4⁺ T cell proliferation in the MyD88^−/− mice was minimal, depletion of CD4⁺ T cells abolished PspA-specific IgG and IgA responses in the MyD88^−/− mice of BALB/c background. Of most interest, MyD88^−/− mice that possessed high levels of PspA-specific IgG and IgA responses but minimal levels of CD4⁺ T cell responses died earlier than non-vaccinated and vaccinated wild-type mice following intravenous or intranasal challenge with virulent S. pneumoniae. Taken together, these results suggest that innate immunity activated by MyD88 signals might not be necessary for antigen-specific antibody induction in both systemic and mucosal sites but is critical for protection following oral vaccination with attenuated Salmonella expressing PspA.

Introduction

Innate and adaptive immunity are host defense mechanisms against invading pathogens (1, 2). Innate immune recognition is mediated by toll-like receptors (TLRs) that recognize specific microbial signature molecules termed pathogen-associated molecular patterns (PAMPs) at either the cell surface or lysosome/endosome membranes. It includes bacterial lipoproteins and lipoteichoic acids (detected by TLR2), lipopolysaccharide (LPS) (detected by TLR4), and flagellin (detected by TLR5) (1, 2). After microbial ligand binding, TLR signaling is mediated by four main adaptor molecules: myeloid differentiation primary-response protein 88 (MyD88), TIR-domain-containing adaptor protein inducing IFN-β (TRIF), TIR-domain-containing adaptor protein (TIRAP), and TRIF-related adaptor molecule (TRAM) (1, 3). These adaptor molecules trigger the activation of various transcriptional factors and then induce inflammatory and antimicrobial responses, which not only lead to stimulation of innate immunity but also help to initiate the development of antigen-specific adaptive immunity (4).

Recent studies have indicated that TLRs sense the microbial infection and engage multiple mechanisms that control the initiation of adaptive T and B cell responses (4). However, the essential role of innate immunity for control of adaptive immunity remains controversial. One previous study demonstrated that activation of TLRs is essential to generate T cell-dependent antigen-specific B cell responses in addition to CD4⁺ T cell help (5). More recently, innate immune sensing of peptidoglycan by a nucleotide binding oligomerization domain (Nod)-1 was reported to be essential for onset of antigen-specific T and B cell responses (6). In contrast, several typical adjuvants elicited robust T cell-dependent antigen-specific B cell responses in the absence of both MyD88 and TRIF signals, indicating that TLR signal is not a key factor for induction of B cell activation (7). To resolve these controversial issues, we investigated the role of TLR-mediated innate immunity for induction of adaptive immunity by mucosal (oral and intranasal) vaccination with a live attenuated Salmonella vaccine strain that potentially has several TLR agonists such as lipoprotein, LPS, and flagellin.

Mucosal surfaces that serve as boundaries with the exterior environment are covered with special epithelial layers that act as barriers against exogenous challenges by pathogens and soluble antigens (8, 9). The mucosal immune system, which is functionally independent of the systemic immune apparatus, has developed its own highly organized immunological tissues (8, 9). These tissues maintain homeostasis in the vast mucosa by mounting specialized anti-inflammatory immune defenses such as the production of secretory IgA (SIgA) antibody and the induction of tolerance against innocuous soluble substances as well as commensal bacteria. Furthermore, due to the migration of IgA antibody-secreting cells (ASCs), local mucosal immunization leads to antigen-specific IgA production at distant mucosal sites (10).

Modified virulent genes in bacteria have potential as mucosal vaccines and antigen carrier vehicles (11, 12). Mucosally administered attenuated Salmonella typhimurium expressing recombinant antigen from other pathogens elicits primarily a Th1-type dominant immune response to both recombinant and Salmonella antigens (13, 14). We previously reported that oral administration of recombinant attenuated S. enteric serovar typhimurium vaccine (RASV) strains expressing pneumococcal surface protein A (PspA) antigen resulted in high levels of PspA-specific IgG responses and efficient protection against challenge with virulent Streptococcus pneumoniae (15). Because Salmonella organisms express a variety of TLR agonists both on their surface and internally (e.g., lipoprotein, LPS, and flagellin), all known as strong adjuvants for enhancement of antigen-specific T and B cell responses (16–18), we questioned whether the TLR agonists in bacterial cell components are involved in inducing complementary and synergistic effects that modulate adaptive immunity following oral immunization with RASV strain expressing PspA antigen. In the present study, we found that a T cell-dependent antigen-specific B cell response was normally induced in MyD88^−/− mice while antigen-specific CD4⁺ T cell responses were minimal. Of interest, MyD88^−/− mice did not have efficient protection against S. pneumoniae infection following oral vaccination with attenuated Salmonella expressing PspA antigen. Thus we conclude that TLR-mediated MyD88 signaling is not crucial for induction of antigen-specific adaptive immunity but is indispensable for protection against bacterial infection.

Materials and Methods

Mice

Wild-type BALB/c and C57BL/6 mice were purchased from Charles River Laboratories (Orient Co., Sungnam, Korea). pIgR^−/− mice of BALB/c background, MyD88^−/− and MyD88^−/− TRIF^−/− mice of both BALB/c or C57BL/6 background were kindly provided by Drs. Masanobu Nanno (Yakult Central Institute for Microbiological Research, Japan) and Shizuo Akira (Research Institute for Microbial Diseases, Osaka Univ., Osaka, Japan), respectively. To generate the PP-null mice, pregnant BALB/c mice were injected intravenously (i.v.) with 600 μg of anti-IL-7Rα mAb on gestational day 14 (19). All mice used in experiments were between 6 and 12 weeks of age. Mice were maintained under pathogen-free conditions in the experimental facility at the International Vaccine Institute (Seoul, Korea), where they received sterilized food and water ad libitum. The experiments were approved by the Institutional Animal Care and Use Committee of the International Vaccine Institute.

Bacterial strains

Salmonella enterica serovar Typhimurium (χ9241 ΔpabA1516 ΔpabB232 ΔasdA16 ΔaraBAD23 ΔrelA198::araCPBADlacI(ATG)TT containing pYA3802) expressing rPspA-RX1 (15, 20) or wild-type UK-1 strain was grown at 37°C in Luria-Bertani (LB) broth or LB agar. Recombinant S. typhimurium BRD 847 strain, a double aroA aroD mutant that expresses the nontoxic, immunogenic 50-kDa ToxC fragment of tetanus toxin from plasmid pTETnir15 under the control of the anaerobically inducible nirB promoter (rSalmonella-ToxC), was cultured in LB broth or LB agar containing ampicillin (21). Bacterial suspensions for mouse immunization were prepared in PBS from LB broth. For the protection assay, S. pneumoniae virulent capsular type 3 strain WU2 was cultured on Trypticase^™ Soy agar containing 5% sheep blood (Becton Dickinson, Sparks, NV, USA) or Todd Hewitt broth plus 0.5% yeast extract (Becton Dickinson) (22). WU2 strain was grown in 100 ml of THY medium until late log phase, adjusted with 3% glycerol, and frozen in 1-ml aliquots containing about 10⁹ CFU/ml. For inoculation, a fresh aliquot was thawed and appropriately diluted (about 1,000-fold) for injection. The actual number of CFU injected was confirmed by plating on the 5% sheep blood (Becton Dickinson).

Immunization

Mice were immunized intragastrically with a single 100-μl aliquot of bacterial suspension (equivalent to 10⁹ CFU) via an intubation needle (Fuchigami, Japan). For intranasal vaccination, mice were immunized under anesthesia by pipette with bacterial suspension (1×10⁸ CFU in a 20-μl PBS aliquot). All immunizations were performed at 2-week intervals.

ELISA and ELISPOT

Serum and fecal extracts were obtained, and antigen-specific antibody titers were determined by ELISA as described elsewhere (23). Endpoint titers were expressed as the reciprocal log₂ of the last dilution giving an optical density at 450 nm of 0.1 greater than background. To assess the number of antibody-secreting cells (ASC), mononuclear cells were obtained from the spleen, lung, and small and large intestines as previously described (23). Antigen-specific ASC were measured by an ELISPOT assay according to an established protocol (24) and were counted with the aid of a stereomicroscope (SZ2-ILST; Olympus, Tokyo, Japan).

Antigen-specific CD4⁺T cell proliferation

Mesenteric lymph node (MLN) and spleen were isolated from wild-type and MyD88^−/− mice of BALB/c background 7 days after the final oral dose of RASV-PspA, and CD4⁺ T cells from each tissue sample were isolated after staining with anti-CD4 microbeads (clone L3T4; Miltenyi Biotec, Auburn, CA) by MACS (Miltenyi Biotec). Dendritic cells (DCs) from spleen of naïve BALB/c mice were sorted by the FACS Aria system (Becton Dickinson) and used as professional antigen-presenting cells. For CD4⁺ T cell proliferation, different numbers of CD4⁺ T cells from each tissue and DCs (1×10⁵) were co-cultured in the absence and presence of PspA antigen (1 μg/ml) for 72 hr. Incorporation of [³H] thymidine after a 16-hr culture period was determined. For neutralizing experiments, anti-CD4 mAb was injected three times per week throughout the immunization period.

Streptococcus pneumoniae challenge

Groups of mice were challenged i.v. (2×10⁶ or 2×10⁷ CFU in 200 μl of PBS) or intranasally (5×10⁷CFU in 50 μl of PBS) with virulent S. pneumoniae WU2 strain. Survival was monitored after challenge.

Neutralizing antibodies

Hybridomas producing anti-CD4 (GK1.5) antibody were purchased from ATCC (Rockville, MD) and mAb was obtained from ascites. Mice were injected intraperitoneally (i.p.) with 150 μg of depleting anti-CD4 or isotype control rat IgG (Sigma-Aldrich, St. Louis, MO) antibodies 1 day before vaccination and every 3 or 4 days thereafter. Almost 95% of CD4⁺ T cells were depleted in the spleen after treatment of anti-CD4 neutralizing antibody.

Adoptive transfer

To reconstitute BALB/c wild-type mice with sera, CD4⁺ T cells, or B cells obtained from RASV-immunized MyD88^−/− mice of BALB/c background, MyD88^−/− mice were orally administered 10⁹ CFU of RASV twice at 2-week intervals. The CD4⁺ or B220⁺ cells were isolated from the spleen and MLN of the immunized MyD88^−/− mice using microbead-conjugated mAbs (Miltenyi Biotec) according to the manufacturer’s instructions. Groups of BALB/c wild-type mice were i.v. transferred with 500 μl of sera or 2 × 10⁷ cells of CD4⁺ T cells or B220⁺ cells, and challenged the next day with the virulent S. pneumoniae WU2 strain.

Data and statistical analysis

The Kaplan-Meier method was used to determine the statistical significance of differences in survival time. We performed the Log-Rank test (Mantel-Cox), using SPSS 12.0K for Windows. To compare the differences between two experimental groups, we used the Student t-test. To compare multiple groups, we carried out one-way ANOVA followed by the Tukey-HSD post hoc test. Each experiment was repeated at least three times using 5 to 10 mice per group. P values of <0.05, <0.01, and <0.001 were assumed to be statistically significant.

Results

Oral immunization of RASV expressing PspA antigen elicited high levels of antigen-specific IgG and SIgA responses in serum and intestine

We first investigated the immunogenicity of RASV strain expressing PspA antigen in systemic (i.e., serum) and mucosal (i.e., fecal extract) compartments after priming and boosting via the oral route. At day 14 after the first vaccination (priming), high levels of PspA-specific IgG antibody were detected in serum of wild-type BALB/c mice (Fig. 1a). The levels of PspA-specific IgG and IgA antibodies in serum and IgA antibodies in fecal extract were much higher 14 days after boosting than after priming (Fig. 1a). To determine whether IgA antibodies in fecal extracts are in secretory form, we used polymeric Ig receptor (pIgR)^−/− mice lacking the IgA secretion pathway (25). These mice showed complete loss of IgA responses in fecal extracts but had levels of IgG and IgA antibodies comparable to those in wild-type mice (Fig. 1b). To assess the protective efficacy of oral vaccination with RASV strain expressing PspA antigen, groups of wild-type and pIgR^−/− mice with or without vaccination were challenged intratracheally (i.t.) with a lethal dose of virulent S. pneumoniae WU2 strain and their survival was monitored (26). As expected, non-vaccinated wild-type mice died within 2 days following i.t challenge with S. pneumoniae (Fig. 1c). Interestingly, all vaccinated wild-type BALB/c mice survived whereas vaccinated pIgR^−/− mice of BALB/c background died within 5 days after i.t. challenge (Fig. 1c). This result strongly suggests that oral vaccination with RASV expressing PspA antigen elicits effective protection against a lethal dose of S. pneumoniae and further that SIgA is essential for protection.

FIGURE 1.

Levels of antigen-specific antibody responses in wild-type (WT) (BALB/c or C57BL/6 background) mice (a) or polymeric immunoglobulin receptor (pIgR)^−/− mice of BALB/c background (b) were determined following oral vaccination with recombinant attenuated Salmonella typhimurium vaccine (RASV) strains expressing pneumococcal surface protein A (PspA) antigen. Groups of mice were immunized orally with RASV expressing PspA on days (D) 0, 14, and 28. Serum and fecal extracts were obtained on days 14 (after priming) and 35 (after boosting) for measurement of PspA-specific IgG and IgA in serum and intestinal IgA in fecal extracts, respectively. No PspA-specific IgG and IgA responses were detected in serum and fecal extracts isolated from non-vaccinated WT mice (data not shown). (c) WT and pIgR^−/− mice of BALB/c background were orally immunized with RASV-PspA on days 0, 14, and 28, groups of mice were injected intratracheally with 5×10⁷ CFU of Streptococcus pneumoniae WU2 strain at day 14 after final vaccination and survival was monitored. (d) Levels of PspA-specific antibody responses in Peyer’s patch (PP) null mice of BALB/c background were also determined following oral vaccination with RASV strains expressing PspA antigen. * p<0.05, **p<0.01, ***p<0.001 vs. after priming (a), vs. WT mice (b and d), vs. WT-PBS (c) groups.

Both Peyer’s patches (PP) and lamina propria in the small intestine are main entry sites for bacterial invasion (27, 28). In order to identify antibody inductive sites by oral RASV-PspA vaccination, PP-null mice were produced. The levels of PspA-specific IgA responses both in serum and fecal extracts were significantly impaired in PP-null mice after oral doses of RASV-PspA (Fig. 1d). In contrast, PP-null and wild-type mice of BALB/c background had identical levels of PspA-specific IgG in serum (Fig. 1d). These data suggest that PP play an indispensable role for induction of PspA-specific IgA in both systemic and mucosal compartments following oral administration of RASV-PspA.

Antigen-specific B cell response was elicited in MyD88 ^−/− mice after oral vaccination of a bacteria-based vaccine strain expressing exogenous protein antigen

We used MyD88^−/− mice to look for a role of innate immunity in development of B cell response following oral vaccination with RASV expressing PspA. Since we found no defect in function of colonization by oral administration with RASV expressing PspA in the MyD88^−/− mice (data not shown), we further addressed inductive levels of T cell-dependent antigen-specific antibody in both serum and fecal extracts after administration of oral RASV expressing PspA. Both wild-type and MyD88^−/− mice of BALB/c background were orally immunized three times 2 weeks apart with RASV expressing PspA. Unexpectedly, we found brisk levels of PspA-specific IgG and IgA in serum and SIgA in fecal extract in MyD88^−/− mice (Fig. 2a). The levels of PspA-specific antibodies were much higher in the MyD88^−/− mice than in wild-type mice (Fig. 2a). In support of these results, high numbers of PspA-specific IgG and IgA ASCs were detected in the spleen and small and large intestines of MyD88^−/− mice. Levels were significantly higher in MyD88^−/− mice than in wild-type mice (Fig. 2a). Further, the ratio of IgG2a (Th1) to IgG1 (Th2) antibody in serum was lower in the MyD88^−/− mice than in the wild-type BALB/c mice, indicating a Th2-dominant condition in the MyD88^−/− mice following oral vaccination with RASV expressing PspA (Fig. 2a). Although Th2-dominant responses were elicited in the MyD88^−/− mice, no detectable levels of PspA-specific IgE antibody were found in their sera (Fig. 2a). We also assessed the effect of intranasal inoculation for induction of antigen-specific B cell responses by RASV expressing PspA in wild-type and MyD88^−/− mice. Again, intranasal administration of RASV expressing PspA resulted in significant levels of PspA-specific IgG and IgA antibodies in systemic and mucosal compartments in MyD88^−/− mice (Fig. 2b). Of note, i.n. administration was much more effective than the oral route for producing IgG and IgA ASCs in the spleen and small intestine, respectively. Overall, innate immunity initiated by MyD88 signal is dispensable for induction of antigen-specific B cell response induced by bacteria-based vaccine expressing T cell-dependent exogenous antigen.

FIGURE 2.

Comparison of PspA-specific antibody responses of MyD88^−/− and wild-type (WT) mice after oral (a) or intranasal (b) vaccination. WT and MyD88^−/− mice of BALB/c background mice were immunized with RASV-PspA on days (D) 0, 14, and 28 by oral (1×10⁹) or intranasal (1×10⁸) routes. Titers of PspA-specific IgG and IgA in serum and intestinal IgA in fecal extract were measured at day 14 after priming and day 7 after final boosting. PspA-specific antibody-secreting cells (ASCs) were determined from spleen (SP), small (SI) and large (LI) intestines, and lung 7 days after the last immunization. (c) BALB/c background WT and MyD88^−/− mice were immunized orally with recombinant S. typhimurium expressing tetanus toxoid (TT). TT antigen-specific IgG and IgA in serum and intestinal IgA in fecal extract were measured at day 14 after priming and day 7 after boosting two times. Titers of TT-specific IgG subclasses in serum were measured at day 14 after boosting. * p<0.05, **p<0.01, ***p<0.001 vs. WT mouse group.

We next queried whether normal induction of B cell responses occurred in MyD88^−/− mice given other attenuated Salmonella strains. For this experiment, we used an attenuated Salmonella vaccine strain expressing the C fragment of the tetanus toxin (rSalmonella-Tox C) (14, 29). Interestingly, identical levels of tetanus toxoid-specific IgG and IgA responses were found in serum and fecal extracts in wild-type and MyD88^−/− mice following oral vaccination with rSalmonella-Tox C strain (Fig. 2c). Consistent with results obtained with the RASV expressing PspA (Fig. 2a and b), dominant Th2-type responses (high IgG1) were found in the MyD88^−/− mice after oral administration with rSalmonella-Tox C strain (Fig. 2c). We also checked for C57BL/6 background in MyD88^−/− mice at same time because antigen-specific antibody induction might be related to mouse genetic background (Fig. 3a). Oral administration of RASV expressing PspA resulted in predominant levels of antigen-specific IgG and IgA both in serum and fecal extract of C57BL/6 background MyD88^−/− mice (Fig. 3a). In addition, identical levels of PspA-specific IgG and IgA both in the serum and fecal extract of C57BL/6 background MyD88^−/− TRIF^−/− mice were elicited following oral vaccination with RASV expressing PspA (Fig. 3b). Of note, high levels of both IgG1 (Th2 type) but reduced levels of IgG2c (Th1 type) antibodies were determined in the serum of C57BL/6 background MyD88^−/− mice when compared to those of C57BL/c wild-type mice (Fig. 3). Overall, MyD88- and TRIF-dependent innate immunity does not involve development of T cell-dependent antigen-specific B cell responses.

FIGURE 3.

Comparison of PspA-specific antibody responses of MyD88^−/− TRIF^−/− and wild-type (WT) mice. WT and MyD88^−/− (a) and MyD88^−/− TRIF^−/− mice (b) of C57BL/6 background were immunized with RASV-PspA on days (D) 0, 14, and 28 by the oral route (1×10⁹). Titers of PspA-specific IgG and IgA in serum and intestinal IgA in fecal extract were measured at day 14 after priming and day 7 after final boosting. * p<0.05 vs. WT mice group.

A critical role of CD4⁺ T cells in MyD88^−/− mice for induction of antigen-specific antibody

Several previous studies found impaired proliferation of MyD88 signal-deficient CD4⁺ T cells both in vivo and in vitro (5, 30). Thus, we next examined antigen-specific proliferation of CD4⁺ T cells in wild-type and MyD88^−/− mice following oral administration with RASV expressing PspA antigen. CD4⁺ T cells were isolated from vaccinated wild-type and MyD88^−/− mice and co-cultured with DCs from spleens of naïve mice with or without PspA antigen in vitro. Of note, PspA-specific CD4⁺ T cell proliferation was significantly decreased in MyD88^−/− mice compared to those in wild-type mice but minimal levels of proliferation were maintained in MyD88^−/− mice (Table 1). To further confirm a role of CD4⁺ T cells for induction of B cell responses, CD4⁺ T cells were depleted by neutralizing anti-CD4 mAb during the vaccination period (Fig. 4a). In wild-type mice, PspA-specific IgG and IgA levels were dramatically decreased while intact levels of IgA response remained in fecal extract after depletion of CD4⁺ T cells. Interestingly, no detectable levels of PspA-specific IgG and IgA responses were found in serum and fecal extracts after depletion of CD4⁺ T cells in the MyD88^−/− mice (Fig. 4a).

Table 1.

Proliferation of CD4⁺ T cells following oral administration of RASV- PspA antigen.

Mice	Tissue	CD4+ T cells alone (+PspA)	Ratio of DCs to CD4+ T cells (+PspA)
			1/20	1/10
BALB/c	Spleen	3797.7 ± 640.6	17882.6 ± 3008.5**	23864.9 ± 3436***
MyD88−/−		677.1 ± 43.5	2045.8 ± 637.4*	2596.3 ± 543.9**
BALB/c	MLN	1572.1 ± 527.6	3975.9 ± 1115.6*	10104.9 ± 127.1***
MyD88−/−		914.5 ± 294.4	1412.2 ± 431.5	2405.3 ± 1313.6

Different numbers of CD4⁺ T cells from each tissue of BALB/c or MyD88^−/− mice of BALB/c background and splenic dendritic cells (DCs) from naïve BALB/c mice as antigen-presenting cells (1×10⁵) were co-cultured in presence of PspA antigen (1 mg/ml) for 5 days.

^*p<0.05,

^**p<0.01,

^***p<0.001 vs. CD4⁺ T cell alone group.

FIGURE 4.

Effectiveness of CD4⁺ T cell depletion on antibody responses and protective efficacy of MyD88^−/− and wild-type (WT) mice after oral vaccination with RASV-PspA strain. (a) BALB/c background WT and MyD88^−/− mice were orally administered RASV expressing PspA on days 0 and 14. Serum and fecal extracts were obtained on days 13 (after priming) and 27 (after boosting). Neutralizing antibody for depletion of CD4⁺ T cells was injected intraperitoneally during vaccination two times per week. **p<0.01, ***p<0.001 vs. rat IgG isotype control group. (b) Protective efficacy by CD4⁺ cells isolated from RASV-vaccinated BALB/c mice. Groups of mice were injected intravenously with 2×10⁶ CFU of Streptococcus pneumoniae WU2 strain 1 day after reconstitution; survival was monitored. **p<0.01, ***p<0.001 vs. WT-PBS group.

Since previous studies suggested an important role of CD4⁺ T cells for protection against colonization of S. pneumoniae (31, 32), we next looked for a direct role of CD4⁺ T cells for protection against S. pneumoniae infection. First, CD4⁺ T cells from RASV-vaccinated mice were isolated and reconstituted in naïve BALB/c mice. Then, recipient mice were challenged i.v. with a lethal dose of virulent S. pneumoniae WU2 strain and their survival was monitored (Fig. 4b). Recipient mice reconstituted with CD4⁺ T cells of RASV-vaccinated BALB/c mice died at an early time point following i.v. challenge whereas RASV-vaccinated mice had 100% survival. Taken together, even though antigen-specific CD4⁺ T cell proliferation was impaired in the MyD88^−/− mice compared with wild-type mice, it seems likely that minimal CD4⁺ T cell function might be involved in enhancing antigen-specific B cell responses in the MyD88^−/− mice. Further, CD4⁺ T cells are indispensable for B cell activation but might not directly manipulate protection against S. pneumoniae infection.

MyD88 signal plays an indispensable role for protection against bacterial infection in the absence or presence of antigen-specific B cell responses

Our next question was whether enhanced levels of acquired immunity in the systemic and mucosal tissues of MyD88^−/− mice are sufficient and efficient for protection against S. pneumoniae infection. To clarify this issue, orally vaccinated wild-type and MyD88^−/− mice were challenged with virulent S. pneumoniae via i.v. (sepsis model) or i.t. (lung pneumonia model) routes. As expected, PBS-vaccinated BALB/c mice died within 4 days following challenge with S. pneumonia (Fig. 5a). Importantly, MyD88^−/− mice that possessed high levels of PspA-specific IgG and IgA antibodies but minimal CD4⁺ T cell responses died about 4–6 days following i.v. or intranasal (i.n.) challenge with virulent S. pneumoniae while vaccinated wild-type BALB/c mice had a 100% survival rate (Fig. 5a). These results suggest that innate immunity activated by MyD88 signals might not be necessary for antigen-specific antibody induction but is critical for protection against S. pneumoniae infection following oral vaccination with attenuated Salmonella expressing PspA.

FIGURE 5.

Role of MyD88 signal in survival after challenge with Streptococcus pneumoniae. (a) MyD88^−/− mice are more susceptible to challenge with S. pneumoniae WU2 strain than are wild-type (WT) mice. WT and MyD88^−/− mice were orally immunized with RASV-PspA on days 0 and 14. At 2 weeks after boosting, groups of mice were challenged intravenously (i.v., sepsis) or intratracheally (pneumonia) with 2×10⁶ CFU or 5×10⁷ CFU of S. pneumoniae WU2 strain, respectively, and survival was monitored. *p<0.05, **p<0.01 vs. WT-PBS group. (b) Neutralizing effect of sera isolated from MyD88^−/− mice. Recipient naive BALB/c mice were reconstituted with sera, CD4⁺ T cells, or B220⁺ B cells isolated from RASV-vaccinated MyD88^−/− mice. Groups of mice were injected i.v. with 2×10⁶ CFU of S. pneumoniae WU2 strain 1 day after reconstitution and survival was monitored. ***p<0.001 vs. Nil group.

To directly assess the neutralizing effects of PspA-specific antibodies that induce the MyD88 deficient condition, sera or CD4⁺ T cells or B220⁺ cells from RASV-vaccinated MyD88^−/− mice of BALB/c background were isolated and reconstituted in naïve BALB/c mice. Then, recipient BALB/c mice were i.v. challenged with a lethal dose of S. pneumonia 1 day after reconstitution (Fig. 5b). Of note, the recipient mice reconstituted with sera of RASV-vaccinated MyD88^−/− mice had 100% survival (Fig. 5b) and no loss of body weight (data not shown). In addition, recipient BALB/c mice reconstituted with B220⁺ cells of RASV-vaccinated MyD88^−/− mice had approximately 50% survival whereas mice reconstituted with CD4⁺ T cells died at early time points following challenge with S. pneumonia. These results clearly demonstrate that PspA-specific antibodies induced by MyD88^−/− B cells possess normal neutralizing function even though RASV-vaccinated mice died earlier than non-vaccinated BALB/c mice.

Discussion

To clarify the role of innate immunity for induction of antigen-specific adaptive immunity following mucosal vaccination with bacteria-based vaccine possessing several TLR ligands, we adopted attenuated Salmonella vaccine strains expressing T cell-dependent antigen (e.g., PspA or tetanus toxoid). We found that antigen-specific B cell responses, including systemic IgG and intestinal SIgA antibodies, were induced after oral administration through PP-dependent, CD4⁺ T cell-dependent, and MyD88/TRIF signaling-independent pathways. Of interest, innate immunity mediated by MyD88 signal was not essential for production of PspA-specific antibodies, but was indispensable for survival against S. pneumoniae infection.

The necessity of innate immunity is controversial, specifically the role of TLR signals as an inducer to provoke antigen-specific adaptive immunity in which T and B cells become armed against bacteria/virus infection (33). It is thought that to induce an effective immune response, microorganisms must stimulate complex sets of pattern-recognition receptors (PRRs), which exist both within and outside the cells. It has been suggested that both TLRs on B cells and DCs are required for optimal antibody responses to T cell-dependent antigen (5). However, a recent study indicated that TLRs are not essential to link between innate and adaptive immunity following vaccination with a protein antigen with different kinds of adjuvants (7). Gavin et al. observed mainly IgG2b and IgG2c responses in MyD88^−/− TRIF^−/− mice given antigen with alum or complete Freund’s adjuvant (CFA) (7). Our results support those findings. We found that TLR-mediated innate immunity is dispensable for induction of antigen-specific acquired immunity induced by bacteria-based vaccine (Fig. 2). Previous studies demonstrated that MyD88^−/− mice have a profound defect in the activation of antigen-specific Th1 but not Th2 immune responses (34–36). Although mucosal vaccination with RASV strain enhanced Th1 dominant responses in innate immunity intact mice, MyD88^−/− mice had a Th2-dominant condition (IgG1>IgG2a or IgG2c) following oral vaccination with RASV expressing PspA. In this regard, DCs from wild-type mice induced allogenic T cells to secrete IFN-γ but DCs from MyD88^−/− mice skewed T cells to secrete IL-4 instead of IFN-γ in response to LPS (37). When MyD88^−/− mice were immunized with KLH-CFA via the systemic route, predominant levels of IL-4 but not IFN-γ were produced by CD4⁺ T cells isolated from draining lymph node in the presence of antigen-presenting cells (37). Possibly Th2-dominant responses support stronger humoral immune responses in MyD88-deficient mice.

It has been postulated that there are alternative pathways through which the innate immune response influences an adaptive immune response—TLR-dependent or -independent (38–40). Two different families of PRRs are expressed in epithelial cells; Nod proteins are located intracellularly whereas many TLRs are typically expressed on the cell surface. Recent findings indicate that once Salmonella is taken up in the cytoplasmic area, intracellular innate immune recognition systems other than surface molecules contribute to the onset of antigen-specific T and B cell responses (38, 41). The relevance of Nod1 and Nod2 for bacterial infections was clearly demonstrated for microbial infections such as Helicobacter pylori and Listeria monocytogenes (41). Ipaf, caspase recruitment domain (CARD)-containing nuclear localizing domain-leucine-rich repeat protein, is responsible for S. typhimurium-induced caspase-1 inflammasome activation (42). In addition, caspase-1 inflammation activation can also be triggered by cytosolic microbial and host DNA in an ASC (apoptosis-associated speck-like protein containing a CARD) - or NALP (also known as cryopyrin) -3-dependent manner (43). Since caspase-1 is responsible for the secretion of active proinflammatory cytokines such as IL-1β, IL-18, and IL-33, which are known to be involved in generating Th2-dominant immune responses (44), the Th2 responses in Myd88^−/− mice after RASV vaccination can be ascribable to the activation of caspase-1 inflammasome by attenuated Salmonella vaccine strain. Thus, it might be necessary to clarify the role of intracellular signals by NOD-like receptors including NOD, Ipaf, and NALP-3 signals for induction of antigen-specific acquired immunity following oral administration with attenuated Salmonella vaccine strain although so far there is no evidence for this hypothesis.

Bacteria such as Shigella or Salmonella or Yersinia can invade M cells on PP and the villous epithelium and DCs on the lamina propria of the small intestine and subsequently result in the acquired phase of antigen-specific immune responses (28, 45, 46). A recent study that used different S. typhimurium strains that targeted the two mechanisms of bacterial entry—M cells in PP (InvA⁺) and DCs in lamina propria of the small intestine (invA⁻)—showed high numbers of M cell-penetrating invasive S. typhimurium in PP, leading to fecal IgA induction (47). In contrast, both strains of bacteria (InvA⁺ or invA⁻) were equally capable of reaching the MLN and the spleen and inducing IgG responses (47). A recent study also showed that PP but not isolated lymphoid follicle in the small intestine is required for antigen-specific intestinal SIgA induction of oral attenuated Salmonella expressing tetanus toxoid (29). Our present results support the finding that PP is essential for induction of PspA-specific IgA but not for IgG responses in the serum and fecal extracts following oral administration of RASV expressing PspA antigen (Fig. 1d). Further, PP-null MyD88^−/− mice did not mount intestinal IgA responses but had intact levels of systemic IgG responses after oral administration of RASV expressing PspA (data not shown). Overall, these data suggest that M cell entry into PP is crucial for development of antigen-specific IgA antibody but not for IgG response following oral vaccination with attenuated bacteria-based vaccine candidates as shown by the independent manner of MyD88-mediated innate immunity.

Of special interest was our finding that MyD88^−/− mice vaccinated with RASV expressing PspA, which possessed brisk levels of neutralizing PspA-specific IgG and IgA antibodies and minimum activation of PspA-specific CD4⁺ T cells, died early and at a time point similar to that of non-vaccinated wild-type mice following lethal challenge with S. pneumoniae (Fig. 5). These results provide convincing evidence that the MyD88 signal is more essential for induction of effective protection against S. pneumoniae infection than the enhanced PspA-specific B cell response. In this regard, MyD88^−/− mice are more likely to have bacteremia, gangrenous mucosal necrosis, severe colitis, and death following infection with microbials than MyD88-intact mice (48–50). The terminal point of signal through the TLR pathway can activate the level of NF-κB which causes production of pro-inflammatory cytokines such as IFN-γ, TNF-α, and IL-6. (1, 2). Our data support previous findings that IL-6 but not TNF-α in the serum of MyD88^−/− mice is completely impaired following oral vaccination with RASV expressing PspA (data not shown). On the other hand, another study suggests a crucial role of MyD88 in killing of gram-negative bacteria (e.g., Escherichia coli and attenuated S. typhimurium) due to diminished NADPH oxidase activity in phagocytic cells, such as macrophages in MyD88^−/− mice (51). Taken together, our results imply that although an adaptive response occurs, dysfunction of pro-inflammatory cytokines and/or phagocytosis mediated by MyD88-dependent signaling might be one factor that explains insufficient clearance of pathogens such as S. pneumoniae in MyD88^−/− mice.

Khan et al. reported that MyD88^−/− mice are markedly defective in their ability to induce multiple proinflammatory cytokine- and chemokine-specific mRNAs in the spleen after i.p. immunization with heat-killed S. pneumoniae capsular type 14 (36). However, prior i.p immunization of MyD88^−/− mice with heat-killed S. pneumoniae type 14 resulted in efficient protection although naïve MyD88^−/− mice were more susceptible than were naïve wild-type mice against otherwise-lethal challenge with live S. pneumoniae type 14 (36). The challenge dose of live S. pneumoniae in the study by Khan et al. was moderated (no naïve wild-type mice were killed) compared to that used in our present study (100% of naïve wild-type mice were killed within four days after challenge). Taken together, whereas low amounts of live S. pneumoniae can be protected by MyD88-independent antibodies, MyD88 signal might be indispensable for protection against infection by large quantities of live S. pneumoniae.

In other work, we demonstrated that LPS-specific adaptive immunity is normally induced in MyD88^−/− mice following oral administration of RASV strain (manuscript in submission). However, orally vaccinated MyD88^−/− mice with high levels of neutralizing antibodies in both systemic and mucosal compartments died earlier than non-vaccinated wild-type mice when challenged orally with virulent Salmonella. Thus, we can conclude that, regardless of T cell-dependent or –independent antigens, MyD88 signal is dispensable for induction of adaptive immune responses but essential for protection against bacterial infection that mainly invaded via the mucosal surface.

Vaccination is the most effective means of preventing infectious diseases. The information learned in this study will greatly facilitate the rational design of novel bacteria-based vaccines against a multitude of infections common in developing countries.

Abbreviation used

TLRs

Toll-like receptors

LPS

lipopolysaccharide

RASV

recombinant attenuated Salmonella enteric serovar typhimurium vaccine

PspA

Pneumoniae surface protein A

MyD88

myeloid differentiation primary-response protein 88

ASCs

antibody-secreting cells

SI

small intestine

LI

large intestine

SIgA

secretory IgA