Structurally identical capsular polysaccharide expressed by intact Group B Streptococcus versus Streptococcus pneumoniae elicits distinct murine polysaccharide-specific IgG responses in vivo

Abstract

We previously reported distinct differences in the murine in vivo Ig polysaccharide (PS)-specific responses to intact Streptococcus pneumoniae compared with Neisseria meningitidis, and that in each case, the bacterial sub-capsular domain markedly influences the Ig response to the associated PS. In light of potentially unique contributions of biochemically distinct capsular PS and/or their characteristic attachments to the underlying bacterium, it remains unresolved whether different bacterial sub-capsular domains can exert differential effects on PS-specific Ig responses to distinct bacterial pathogens. In this report, we utilized a mutant strain of group B Streptococcus (S. agalactiae) type III (GBS-III) that expresses desialylated IIIPS, biochemically identical to type 14 capsular PS (PPS14) of Streptococcus pneumoniae (Pn14), in order to directly compare the in vivo PPS14-specific IgG responses to two distinct Gram-positive bacteria. Although both GBS-III and Pn14 elicited relatively rapid primary PPS14-specific IgG responses dependent on CD4+ T cells, B7-dependent costimulation, and CD40/CD40L interactions, only GBS-III induced a highly boosted ICOS-dependent PPS14-specific IgG response following secondary immunization. Of note, priming with Pn14 and boosting with GBS-III, although not isolated PPS14, elicited a similar boosted PPS14-specific IgG response that was dependent on CD4+ T cells during secondary immunization, indicating that Pn14 primes for memory but, unlike GBS-III, fails to elicit it. The inability of Pn14 to elicit a boosted PPS14-specific IgG response was overcome by co-immunization with unencapsulated GBS-III. Collectively, these data establish that structurally identical capsular PS expressed by two distinct Gram-positive extracellular bacteria can indeed elicit distinct PS-specific IgG responses in vivo.

Introduction

Extracellular bacteria expressing polysaccharide (PS) capsules such as Streptococcus pneumoniae (Pn), Streptococcus agalactiae (group B Streptococcus [GBS]), and Neisseria meningitidis are leading causes of morbidity and mortality due to sepsis, pneumonia and meningitis (1–3). The PS capsule is a major virulence factor and vaccines targeting these pathogens are thus designed to elicit protective PS-specific IgG (4). In contrast to proteins, non-zwitterionic PS are not enzymatically processed within the APC for association with MHC-II molecules (5, 6) and therefore do not recruit CD4+ T cells. Thus, isolated PS are T cell-independent (TI) antigens that fail to elicit robust germinal center formation and immunologic memory (7). However, PS are comprised of repeating sugar units that effect multivalent BCR crosslinking resulting in efficient and robust B cell proliferation (8), and in the presence of second signals, such as TLRs and/or cytokines, the induction of Ig secretion and class switching (9).

The view of proteins and PS as T-cell dependent (TD) and TI antigens, respectively have largely come from studies based on immunization with isolated proteins and haptenated or non-haptenated PS. However during natural infections, the host encounters intact PS-encapsulated bacteria, in which proteins and PS are co-expressed in a single particulate structure along with multiple moieties like TLR, NLR, and scavenger receptor ligands that engage the innate immune system (10). Thus, the non-covalent co-expression of PS and proteins by the intact bacterium could potentially elicit bacterial peptide-specific CD4+ T cell help for PS-specific B cells, leading to the generation of enhanced PS-specific primary IgG responses and induction of memory. Indeed, an earlier study using intact, heat-killed Streptococcus pyogenes (group A streptococcus) reported lower IgM and IgG3 anti-group A streptococcal carbohydrate responses in T cell-deficient mice (11). In this regard, our previous studies using intact Streptococcus pneumoniae, capsular type 14 (Pn14), a Gram-positive (GP) bacterium, provide strong evidence that intact PS-encapsulated extracellular bacteria may represent unique immunogens, exhibiting both TI and TD characteristics of the associated PS-specific IgG response. Thus the primary PPS14 (pneumococcal PS type 14)-specific IgG response to intact Pn14 is dependent on CD4+ T cells, B7-dependent co-stimulation and CD40/CD40L interaction, and comprise all 4 IgG isotypes (TD-like). However, the primary PPS14-specific IgG response develops rapidly as an apoptosis-prone extrafollicular response that is more dependent on BCR signaling than the protein-specific IgG response. In addition the PPS14-specific IgG response is ICOS-independent and fails to generate a boosted IgG response after secondary immunization (TI-like) (12–16).

In contrast, our more recent studies using the Gram negative (GN) extracellular bacterium, Neisseria meningitidis type C (MenC), have shown that the primary IgG MCPS (Meningococcal type C PS)-specific IgG response develops more slowly, with a highly boosted MCPS-specific IgG response following secondary immunization. The primary MCPS-specific IgG response is TI, whereas the boosted MCPS-specific response is dependent on CD4+ T cells, B7-dependent co-stimulation, CD40/CD40L and ICOS/ICOSL interactions (17). Collectively, these studies utilizing intact Pn14 and MenC demonstrate that the bacterial sub-capsular domain markedly influences the Ig response to the associated PS, relative to that observed using isolated PS (18). However, they leave unresolved whether differences observed in PS-specific Ig responses between distinct intact extracellular bacteria reflect intrinsic differences in their biochemically unique capsular PS, and/or the nature of the PS attachment to, or composition of, the underlying sub-capsular bacterial domain.

In this regard, capsular PS expressed by Gram-positive bacteria are covalently attached to an underlying thick peptidoglycan cell wall to which a number of immunogenic proteins are also covalently linked (19, 20), whereas in Gram-negative bacteria, PS are covalently attached to the acyl glycerol moiety of the outer membrane that contains LPS and immunogenic proteins (21, 22). PS may also exhibit intrinsic features that potentially impact on the elicited immune response. Thus, PPS14 can bind to SIGN-R1, a scavenger receptor present on marginal zone macrophages (23), while MCPS has unique immunomodulatory properties that can impact on the MCPS-specific Ig response (24). In addition, the physical and chemical nature of the capsular PS such as the molecular weight (25), charge (26) or sialic acid content (27) could also influence the induction of PS-specific Ig. In light of these potentially unique contributions of biochemically distinct capsular PS and/or unique bacterial attachments, to the PS-specific Ig response to an intact bacterium, it remains of interest whether different bacterial sub-capsular domains themselves can exert differential effects on PS-specific Ig responses between distinct bacteria.

In the present study, we used intact heat-killed group B Streptococcus type III (GBS-III), a Gram-positive bacterium, to initially determine the nature of the murine in vivo IIIPS-specific Ig response relative to that observed using isolated IIIPS antigen. Further, to directly assess the potential contribution of the sub-capsular bacterial domain on the associated PS-specific responses between two distinct intact bacteria, we took advantage of the previous demonstration that the core IIIPS antigen of the IIIPS capsule expressed by GBS-III is biochemically identical to PPS14 expressed by Pn14, differing from the native IIIPS in lacking the terminal sialic acid (28). Thus, immunization with either conjugated or unconjugated IIIPS vaccines produce IgG that cross-react with PPS14 (29). In particular, we utilized a mutant GBS-III strain (COH1-11) that expresses the desialylated IIIPS that is identical to PPS14. In this manner, we were able to directly determine whether two distinct Gram-positive bacteria (Pn14 and GBS-III) expressing identical PS with similar attachments to the underlying sub-capsular bacterial domain, potentially elicit distinct PS-specific IgG responses in vivo. This report demonstrates that differences in underlying sub-capsular bacterial domains can indeed differentially regulate PS-specific Ig responses in vivo.

Materials and methods

Mice

BALB/c mice and athymic nude (BALB/c background) mice were purchased from the National Cancer Institute (Frederick, MD). Female mice were used between 7 and 10 weeks of age. These studies were conducted in accordance with the principles set forth in the Guide for Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Research Council, revised 1996), and were approved by the Uniformed Services University of the Health Sciences Institutional Animal Care and Use Committee.

Bacterial strains

Group B Streptococcus (Streptococcus agalactiae) type III (GBS-III) [strain M781] was obtained from American Type Culture Collection [Manassas, VA] (ATCC BAA-22). Another GBS strain COH1 expressing the same capsular PS type III and two isogenic COH1 mutants designated as COH1-13 lacking a capsule and COH1-11 expressing the desialylated type III capsule that is similar to the PPS14 capsule of Pn14, were also used in this study (30). All three COH strains were a kind gift of Dr. Craig Rubens of Children’s Orthopedic Hospital (Seattle, WA). Streptococcus pneumoniae, capsular type 14 (strain R614) was prepared as described previously (31). The isogenic non-encapsulated mutant of D39 (Pn, capsular type 2), [strain R36A]) was provided by Dr. David Briles (University of Alabama at Birmingham, Birmingham, AL). Lyophilized or frozen stocks of bacteria were grown overnight on BBL blood agar plates (VWR International, Bridgeport, NJ). Isolated colonies on blood agar were grown in Todd Hewitt Broth media (BD Biosciences, San Jose, CA) to mid-log phase, collected, and heat-killed by incubation at 65°C for 2 h or inactivated by overnight U.V. irradiation. Sterility was confirmed by subculture on blood agar plates. After extensive washings, the bacterial suspension was adjusted with PBS to give an absorbance reading at 650 nm of 0.6 which corresponded to 10⁹ CFU/ml. Bacteria were then aliquoted at 10¹⁰ CFU/ml and frozen at −20°C until their use as immunogens for mouse immunizations.

Reagents

Purified capsular PS from GBS-III (IIIPS) was isolated and purified from strain M781 (32). The conjugates IIIPS-HSA and III-rAlp3 were made respectively, as previously described (29, 33). Purified S. pneumoniae capsular polysaccharide type 14 (PPS14) was purchased from ATCC. Rat IgG2b anti-mouse CD4 mAb (clone GK1.5) was purified from ascites by ammonium sulfate precipitation and passage over a protein G column. Purified polyclonal rat IgG was purchased from Sigma (St. Louis, MO). Hamster IgG anti-mouse CD40L mAb (clone MR1), polyclonal hamster IgG, rat IgG2a anti-mouse CD275 mAb (clone HK5.3), rat IgG2a isotype control (clone 2A3), hamster IgG2 anti-mouse CD80 (B7-1) mAb [clone 16-10A1], and rat IgG2a anti-mouse CD86 (B7-2) mAb [clone GL-1] were purchased from BioXcell (West Lebanon, NH). Alum (Allhydrogel 2%) was obtained from Brenntag Biosector (Frederikssund, Denmark). A stimulatory 30mer CpG-containing oligodeoxynucleotide (CpG-ODN) was synthesized as previously described (34).

Preparation of GBS whole protein extract

Fifty ml of live unencapsulated GBS-III (strain COH1-13) bacteria were centrifuged at 3000 rpm for 20 min and the supernatant fluid was discarded. About 5 ml of B-PER bacterial protein extraction reagent from Pierce (Rockford, IL), was added to the bacterial pellet (1:10 ratio), mixed well by shaking and incubated at RT for 15 min. The treated bacterial culture was again centrifuged at 3000 rpm for 10 min. The supernatant fluid containing the soluble proteins was obtained and the concentration was determined by the BCA assay.

Immunizations

Mice (n=7/group) were immunized i.p. with 1 × 10⁹ colony forming units (CFU) of heat-killed or UV-inactivated GBS-III in saline (2 × 10⁸ CFU) or 10 μg of purified IIIPS in saline, or 1 μg of IIIPS-rAlp3 adsorbed on 13 μg of alum mixed with 25 μg of CpG-ODN. Serum samples for measurement of anti-protein and anti-PS Ig isotype titers by ELISA were prepared from blood obtained through the tail vein.

ELISA

For measurement of serum titers of IIIPS-specific Ig, Immulon 4 ELISA plates were coated with IIIPS-HSA (1 μg/ml, 100 μl/well) in PBS. To measure the serum titers of whole GBS-protein-specific Ig, the plates were coated with GBS protein extract (30 μg/ml, 50 μl/well). Plates werethen washed 3x with PBS + 0.1% Tween 20 and were blocked with PBS + 1% BSA for 2 h at 37°C. Three-fold dilutions of serum samples, starting at a 1/50 serum dilution, in PBS + 1% BSA were incubated overnight at 4°C and plates were then washed 3x with PBS + 0.1% Tween 20. Alkaline phosphatase-conjugated polyclonal goat anti-mouse IgM, IgG, IgG3, IgG1, IgG2b, or IgG2a Abs (200 ng/ml final concentration) in PBS + 1% BSA were then added, and plates were incubated at 37°C for 1 h. Plates were washed 3x with PBS + 0.1% Tween 20. Substrate (p-nitrophenyl phosphate, disodium; Sigma) at 1 mg/ml in TM buffer (1 M Tris 0.3 mM MgCl₂ [pH 9.8]) was then added for color development. Color was read at an absorbance of 405 nm on a Multiskan Ascent ELISA reader (Labsystems, Finland). Serum titers of PPS14-specific Ig were measured as described previously (13).

Statistics

Serum Ig isotype titers were expressed as geometric means ± SEM of the individual serum Ig isotype titers. Significance was determined by the Student’s t test. P-values of ≤ 0.05 were considered statistically significant. All experiments were performed at least twice.

Results

Intact GBS-III elicits a boosted secondary IIIPS-specific IgG response similar to a IIIPS conjugate vaccine

We previously demonstrated that the IgG PPS14-specific IgG response to intact Pn14 is distinct from that of a PPS14 conjugate vaccine, in that Pn14 elicits more rapid primary kinetics, and fails to boost serum titers of PPS14-specific IgG following secondary immunization, despite the primary response being CD4+ T-cell dependent (12, 35). As discussed above, we wished to compare the PS-specific Ig response to Pn14 with that elicited by another intact Gram-positive bacterium, GBS-III. We began by comparing the in vivo IIIPS-specific Ig response to intact GBS-III relative to that induced by a IIIPS-rAlp3 conjugate vaccine (33). Alps (alpha-like proteins) are a family of related surface-anchored proteins with large numbers of repeat units that are expressed by most GBS strains, with Alp3 expressed on most isolates of GBS-V and GBS-VIII (36).

BALB/c mice were immunized i.p. with either U.V.-inactivated GBS-III (strain M781) in saline or IIIPS-rAlp3 conjugate (1 μg/mouse) in alum + CpG-ODN, and boosted 28 days later in a similar manner. The primary IIIPS-specific IgG response to intact M781 exhibits relatively rapid kinetics of induction with peak titers observed by day 7 (Fig. 1A), similar to that elicited by IIIPS-rAlp3 (Fig. 1B). Although the degree of induction of serum titers of rAlp3-specific IgG in response to conjugate was similar to that of IIIPS-specific IgG by day 7, rAlp3-specific IgG titers continued to rise significantly with peak titers observed by day 21 (Fig. 1B). Further, M781, in addition to IIIPS-rAlp3, elicits a secondary IIIPS-specific IgG response that is >10-fold higher than the primary, similar to the level of boosting of the secondary rAlp3-specific IgG response to the conjugate. The primary IIIPS-specific IgM response to M781 also peaked early (day 7) with modest, but transient, secondary boosting. Thus, intact GBS-III, like Pn14 generates a relatively rapid primary PS-specific IgG response, but in contrast to that previously reported for Pn14, elicits a highly boosted secondary PS-specific IgG response, similar to that observed for a conjugate vaccine.

Figure 1. Intact GBS-III elicits a boosted secondary IIIPS-specific IgG response similar to a IIIPS conjugate vaccine.

BALB/c mice (7 per group) were immunized i.p. with 2 × 10⁸ CFU/ml of U.V.-inactivated intact GBS-III (strain M781) in saline (A) or 1 μg of IIIPS-rAlp3 in alum + CpG-ODN (B), and boosted i.p. with the same dose on day 28. Serum titers of Ag-specific IgM and IgG were determined by ELISA. Significance (*) p≤0.05 between secondary titers relative to peak primary titers.

The IIIPS-specific IgG responses to intact GBS-III versus isolated IIIPS are distinct

Isolated non-zwitterionic PS are unable to recruit CD4+ T cell help and thus typically induce primarily IgM, with relatively minor IgG responses following primary immunization. These PS also fail to induce IgG memory upon secondary challenge (7), although IgM memory responses to these antigens have been observed (37). In this regard, we directly compare the IIIPS-specific Ig response to isolated IIIPS versus intact GBS-III to determine the impact of the bacterial sub-capsular domain on the associated PS-specific Ig response. Mice were immunized with either isolated IIIPS or heat-inactivated M781 in saline and boosted on day 14 in a similar fashion. As illustrated in Fig. 2, the primary IgM responses to both isolated IIIPS and M781 were similar in titer and peaked by day 7. Whereas secondary immunization with isolated IIIPS produced no further increases in serum titers of IIIPS-specific IgM, M781 induced a boosted secondary response. No detectable primary or secondary IIIPS-specific IgG or IgG isotype response to isolated IIIPS was observed. In contrast, M781 induced a significant primary IIIPS-specific IgG response, peaking by day 7, with ~30 fold increase in titers following secondary immunization. Secondary IIIPS-specific IgG in response to M781 comprised boosted serum titers of IgG3, IgG1 and IgG2a, with little if any IgG2b. Immunization of mice with 1, 3, 10, or 30 μg of isolated IIIPS in saline generated essentially similar results (data not shown). Thus, the IIIPS-specific IgG response to intact GBS-III is distinct from that elicited by purified IIIPS, demonstrating the significant impact of the bacterial sub-capsular domain on the associated PS-specific IgG response, in agreement with our previous studies utilizing intact Pn14 and MenC (18).

Figure 2. The IIIPS-specific IgG responses to intact GBS-III versus isolated IIIPS are distinct.

BALB/c mice (7 per group) were immunized i.p. with 1 × 10⁹ CFU/ml of heat-inactivated intact GBS-III (strain M781) or purified IIIPS in saline, and boosted i.p. with the same dose on day 14. Serum titers of IIIPS-specific IgM, IgG, and IgG subclasses (IgG3, IgG1, IgG2b and IgG2a) were determined by ELISA. Significance (*) p≤0.05 between M781 and IIIPS.

The IIIPS-specific IgG elicited in response to intact GBS-III cross-reacts with PPS14

It has been reported that individuals immunized with either unconjugated or conjugated IIIPS vaccines produce antibodies that cross-react with PPS14 (29). In the present experiment, we wished to determine this potential cross-reactivity of the IIIPS-specific Ig response using three different strains of GBS-III: 1) COH1 that expresses the native type III capsule, 2) COH1-11, an isogenic mutant of COH1 that expresses the desialylated type IIIPS which is identical to the PPS14 capsule of Pn14, and 3) COH1-13, an isogenic mutant of COH1 that lacks the capsule. Mice were immunized with heat-killed COH1, COH1-11 or COH1-13 and boosted on day 21 in a similar fashion. As illustrated in Fig. 3, the primary IIIPS-specific IgG response to intact COH1 and COH1-11 peaked by day 7 with >10-fold enhancements following secondary immunization, similar to that observed in Figs. 1A and 2. In addition, coating of ELISA plates with purified PPS14, instead of IIIPS demonstrates a primary and secondary PPS14-specific IgG response to both intact COH1 and COH1-11 that is similar to that observed for IIIPS-specific IgG. Indeed, pre-incubating COH1- and COH1-11-induced immune sera with PPS14 largely abrogates the detection of IIIPS-specific IgG using the IIIPS-specific ELISA (data not shown). Thus, these data indicate that the majority of the IIIPS-specific IgG are directed towards the type III core antigen that is identical to PPS14. The COH1-13 bacterial strain that lacks capsular PS fails to induce any detectable serum titers of IIIPS-specific or PPS14-specific IgG, thus establishing the specificity of the ELISA assay for the capsular PS.

Figure 3. The IIIPS-specific IgG elicited in response to intact GBS-III cross-reacts with PPS14.

BALB/c mice (7 per group) were immunized i.p. with 1 × 10⁹ CFU/ml of heat-inactivated intact GBS-III (strains COH1, COH1-11 or COH1-13) in saline and boosted i.p. with the same dose on day 14. Serum titers of IIIPS-specific and PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between secondary titers relative to peak primary titers.

The primary and secondary IIIPS- and PPS14-specific IgG responses to GBS-III are dependent upon CD4+ T cell help and CD40L-dependent co-stimulation

While the PS-specific Ig responses to most isolated PS are TI, studies using intact Pn14 demonstrated dependence on CD4+ T cells, as well as CD40L, for the primary PPS14-specific IgG response (12, 35). Further, studies using intact MenC have shown that the secondary, but not primary, MCPS-specific IgG response is also dependent on CD4+ T-cells and CD40L (17). In this regard, we wished to determine the role of CD4+ T-cells and CD40L in the induction of the IIIPS-specific and PPS14-specific IgG responses to intact GBS-III. To determine this, BALB/c mice were injected with an anti-CD4+ T cell-depleting mAb (clone GK1.5) or control polyclonal rat IgG 24 h before immunization with heat-inactivated GBS-III (strain M781). Flow cytometric analysis confirmed that 95% of the CD4+ T cells were specifically depleted after 24 h, in the anti-CD4 mAb-injected group. Mice were boosted on day 14 in the absence of anti-CD4 mAb or control polyclonal rat IgG. In addition, athymic nude mice (BALB/c background), which are markedly deficient in T cells, were also immunized at the same time with M781 and boosted similarly on day 14. As illustrated in Fig. 4A, the primary IIIPS-specific and PPS14-specific IgG responses to M781 in mice injected with control rat IgG peaked by day 7 with a >10-fold boost following secondary immunization. In contrast, both the anti-CD4 mAb-injected mice and athymic nude mice, showed a nearly complete abrogation of the primary and secondary IIIPS-and PPS14-specific IgG responses compared to the control group, establishing their critical dependence on CD4+ T cells.

Figure 4. The primary and secondary IIIPS- and PPS14-specific IgG responses to GBS-III are dependent upon CD4+ T cell help and CD40L-dependent co-stimulation.

(A) BALB/c mice (7 per group) were injected i.p. with either depleting anti-CD4 mAb (clone GK1.5) or control polyclonal rat IgG (0.5 mg/mouse) 24 h before immunization. Both athymic nude mice and Ab-injected mice were immunized i.p. with 1 × 10⁹ CFU of GBS-III (strain COH1) and boosted on day 14 in the absence of depleting antibodies. Serum titers of IIIPS- and PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between control versus anti-CD4 mAb-injected mice or athymic nude mice. (B) BALB/c mice (7 per group) were injected i.p. with either blocking anti-CD40L mAb (clone MR1) or control polyclonal hamster IgG (0.5 mg/mouse) 24 h before i.p. immunization with 1 × 10⁹ CFU of GBS-III (strain COH1). Mice were boosted similarly on day 14 in the absence of antibodies. Serum titers of IIIPS- and PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between control versus anti-CD40L mAb-injected mice.

The induction of CD40L on activated CD4+ T cells is critical in mediating TD humoral immune responses, including induction of isotype switched antibodies, germinal center formation and the generation of memory, through delivery of co-stimulatory signals via CD40 on APC (38). Preventing CD40 CD40L interactions in vivo through the use of blocking antibodies largely abolishes TD humoral immunity (39). To determine the role for CD40L co-stimulation on the IIIPS- and PPS14-specific IgG responses to intact GBS-III, BALB/c mice were injected with a blocking anti-CD40L mAb (clone MR1) or polyclonal hamster IgG as a control, 24 h before immunization with intact heat-killed GBS-III (strain M781). Mice were boosted in a similar fashion on day 14 in the absence of blocking anti-CD40L mAb or control IgG. As shown in Fig. 4B, both the primary and secondary IIIPS-specific and PPS14-specific IgG responses were significantly inhibited compared to the control (primary: 4.6–7.0-fold reduction; secondary: (3.8–7.9-fold reduction). Of interest, blocking of CD40L did not prevent boosting of the IgG responses to either IIIPS or PPS14 following secondary immunization, although the serum titers were significantly reduced compared to the control. Collectively these data demonstrate that the primary and boosted secondary PS-specific IgG responses to intact GBS-III require CD4+ T cell help as well as CD40L-dependent co-stimulation.

B7- and ICOS-dependent co-stimulation is required for optimal induction of PS-specific IgG responses to intact GBS-III

CD28 constitutively expressed on CD4+ T cells binds to CD80/CD86 (B7-1/B7-2) expressed on APCs and this interaction is critical for the initiation of CD4+ T cell activation (40, 41). We previously demonstrated that the primary PPS14-specific IgG response to intact Pn14, and the secondary, but not primary, MCPS-specific IgG response to MenC, is dependent on CD28 (13). In this regard, we wished to determine the role of CD28-mediated signaling in the PS-specific Ig response to intact GBS-III. Therefore, mice were injected with blocking anti-CD80 (clone 16-10A1) and anti-CD86 (clone GL-1) mAbs or polyclonal hamster and rat IgG as a control, 24 h before immunization with intact GBS-III (strain M781), with secondary immunization on day 14. As illustrated in Fig. 5A, the primary and secondary IIIPS- and PPS14-specific IgG responses to M781 were markedly inhibited in the presence of blocking mAbs compared to the control (primary: 8.2–17.3-fold reduction; secondary: 14.9–133-fold reduction).

Figure 5. B7- and ICOS-dependent co-stimulation is required for optimal induction of PS-specific IgG responses to intact GBS-III.

(A) BALB/c mice (7 per group) were injected i.p. with either blocking hamster anti-CD80 mAb (B7-1) (clone 16-10A1) and rat anti-CD86 mAb (B7-2) (clone GL-1) or control polyclonal hamster and rat IgG (0.5 mg/mouse) 24 h before i.p. immunization with 1 × 10⁹ CFU of GBS-III (strain COH1). Mice were boosted similarly on day 14 in the absence of antibodies. Serum titers of IIIPS- and PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between control versus anti-CD80- and anti-CD86 mAb-injected mice. (B) BALB/c mice (7 per group) were injected i.p. with either blocking anti-ICOSL mAb (clone HK5.3) or control rat IgG2a mAb (clone 2A-3) (0.5 mg/mouse) 24 h before i.p. immunization with 1 × 10⁹ CFU of GBS-III (strain M781). Mice were boosted similarly on day 14 in the absence of antibodies. Serum titers of Ag-specific IgG were determined by ELISA. Significance (*) p≤0.05 between control versus anti-ICOSL mAb-injected mice.

Inducible costimulator (ICOS), a member of the CD28 family, induced on CD4+ T cells upon TCR crosslinking and CD28-mediated signaling, binds to ICOSL expressed on APC (40, 42). ICOS is the key regulator of germinal center formation and immunological memory (43, 44). Our earlier studies using Pn14 demonstrated that the PPS14-specific IgG response is ICOS-independent, while the secondary, but not primary, MCPS-specific IgG response to MenC required ICOS-co-stimulation. In light of the boosted secondary PS-specific IgG responses to GBS-III, we wished to determine whether or not it was ICOS-dependent. Therefore, we injected BALB/c mice with anti-ICOSL mAb (clone HK5.3) or control polyclonal rat IgG, 24 h before immunization with intact GBS-III (strain M781). Mice were boosted on day 14 in the absence of anti-ICOSL mAb or control Ab. As shown in Fig. 5B, the secondary, but not primary, IIIPS-specific IgG and PPS14-specific IgG responses to M781 were significantly, although not completely, inhibited in anti-ICOSL mAb-injected mice (3.2–3.3-fold reduction), whereas a near-complete abrogation of the secondary protein-specific IgG response was observed. Thus, these data demonstrate that B7-dependent, T cell co-stimulation is critical for induction of primary and boosted secondary IIIPS-specific and PPS14-specific IgG responses to intact GBS-III. Further, the importance of ICOSL-dependent co-stimulation for the boosted PS-specific IgG response to GBS-III supports the notion that GBS-III induces PS-specific IgG memory, likely dependent on GC formation.

A boosted secondary PPS14-specific IgG response is elicited by GBS-III in Pn14-primed mice

We demonstrated previously that intact Pn14 fails to generate a boosted PPS14-specific IgG response upon secondary immunization, despite the dependence of the primary response on CD4+ T cells. Although these data suggested that Pn14 failed to induce memory for PPS14-specific IgG, an alternative possibility was that Pn14 induced a state of memory, but was unable to elicit a boosted memory response upon secondary immunization, in contrast to GBS-III. Therefore, we sought to determine whether priming of mice with Pn14 (strain R614) or GBS-III (strain COH1-11) followed alternatively by secondary immunization on day 14 with COH1-11 or R614, respectively, would elicit a boosted PPS14-specific IgG response. We utilized strain COH1-11 since it expresses the identical PPS14 capsule as R614. As illustrated in Fig. 6A, primary and secondary immunization with R614, in contrast to COH1-11, failed to elicit a boosted PPS14-specific IgG secondary response, as previously observed. Of note, primary immunization with R614 followed by secondary immunization with COH1-11 resulted in a boosted PPS14-specific IgG response similar to that observed using COH1-11 for both immunizations (p<.001). These data thus indicate that R614 can induce memory for PPS14-specific IgG, but fails to elicit a boosted secondary response in R614-primed mice. Of interest, mice primed with COH1-11 followed by secondary immunization with R614 showed a more modest boost in the secondary PPS14-specific IgG response in two identical experiments, relative to COH1-11/COH1-11 and R614/COH1-11 immunized mice.

Figure 6. A boosted secondary PPS14-specific IgG response is elicited by GBS-III in Pn14-primed mice.

(A) BALB/c mice (7 per group) were immunized i.p. on d0 and d14 with 1 × 10⁹ CFU of either Pn14 (strain R614) or GBS-III (strain COH1-11) in various combinations as indicated. Serum titers of PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between secondary titers relative to peak primary titers. (B) BALB/c mice (7 per group) were primed with 5 × 10⁸ CFU of heat inactivated Pn14 (strain R614) or 5 × 10⁸ CFU of heat inactivated GBS-III (strain COH1-11) or R614+ GBS-III (strain COH1-13) (5 × 10⁸ CFU/mouse each) or COH1-11 + Pn2 (strain R36A) (5 × 10⁸ CFU each). Mice were boosted in a similar fashion on d14. Serum titers of PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between secondary titers of R614- and R614+COH1-13-immunized groups and secondary titers of COH1-11− and COH1-11+R36A-immunized groups.

We previously demonstrated that primary, followed by secondary, co-immunization of mice with R614 and MenC results in secondary boosting of the PPS14-specific IgG response, suggesting that provision of sufficient MenC-induced innate stimulation during the secondary response might compensate for inadequate signaling or an inhibitory effect of Pn14 (17). Therefore, we wished to determine whether GBS-III could provide a similar compensatory effect. To accomplish this, mice were co-immunized (primary and secondary) with R614 and the unencapsulated GBS-III strain COH1-13, to preclude a PPS14-specific contribution of GBS-III to the PPS14-specific IgG response. As shown in Fig. 6B, co-immunization with R614 and COH1-13 resulted in a significant boost in the secondary PPS14-specific IgG response, although significantly lower than in mice immunized only with the encapsulated COH1-11 strain. Mice immunized with COH1-11 in the presence of unencapsulated Pn14 (strain R36A) also elicited a significantly boosted secondary PPS14-specific IgG response similar to mice co-immunized with R614 + COH1-13, but significantly lower than mice immunized with COH1-11 alone. Collectively, these data suggest that the absence of a boosted secondary PPS14-specific IgG response to Pn14 may reflect an inhibitory effect mediated by this bacterium during secondary immunization of Pn14-primed mice, and that this might be overcome by stimulation by innate immune mechanisms.

The boosted secondary PPS14-specific IgG response to GBS-III requires GBS-III-associated PPS14 and CD4+ T cells during secondary immunization

The ability of COH1-11 to boost the secondary PPS14-specific IgG response in COH1-11- or R614-primed mice left unresolved whether boosting during secondary immunization simply required BCR crosslinking of PPS14-specific B cells in the absence of a putative inhibitory influence of R614, appropriate non-specific innate stimulation, or both, and whether this effect required CD4+ T cells during secondary immunization. To test this we immunized mice with heat-inactivated COH1-11 or R614 followed by secondary immunization with isolated PPS14 on day 14. As a control group, mice were immunized with isolated PPS14 for both the primary and secondary. In addition, to determine the BCR specificity of the secondary PPS14-specific IgG response, mice were primed with COH1-11 or R614 and boosted with their unencapsulated variants, COH1-13 and R36A respectively.

Neither isolated PPS14 nor non-specific stimulation by unencapsulated bacteria were able to elicit a boosted PPS14-specific IgG response in either COH1-11- or R614-primed mice (Fig. 7A). Additionally, as illustrated in Fig. 7B, mice primed with COH1-11 alone followed by secondary immunization with COH1-11 in the presence of depleting anti-CD4 mAb, failed to elicit a boosted PPS14-specific IgG response relative to mice receiving control antibody. Collectively, these data suggest that boosting of the secondary PPS14-specific IgG response in GBS-III-immunized mice requires both PPS14-specific BCR engagement and CD4+ T cells, the latter likely triggered via recognition of bacterial-derived peptides in association with MHC-II. These data also suggest that APC processing of intact Pn14 and GBS-III might generate cross-reactive peptides for the generation of CD4+ T cell help.

Figure 7. The boosted secondary PPS14-specific IgG response to GBS-III requires GBS-III-associated PPS14 and CD4+ T cells during secondary immunization.

(A) BALB/c mice (7 per group) were immunized and boosted i.p. on d0 and d14 with either 1 μg of PPS14 or 1 × 10⁹ CFU of either Pn 14 (strain R614) or GBS-III (strain COH1-11) as indicated. Serum titers of PPS14-specific IgG were determined by ELISA. (B) BALB/c mice (7 per group) were immunized i.p. with 1 × 10⁹ CFU of heat inactivated GBS-III (strain COH1) on d0 and d14. 0.5 mg of depleting anti-CD4 mAb (clone GK1.5) or control rat IgG was injected on d13. Serum titers of PPS14-specific IgG were determined by ELISA. Significance (*) p≤0.05 between secondary titers of control group versus anti-CD4-mAb-injected group.

Discussion

To determine whether the sub-capsular bacterial domain of distinct PS-encapsulated extracellular bacteria can differentially affect the nature of an in vivo PS-specific response, it was necessary to study two distinct bacteria with biochemically identical PS having similar bacterial attachments. In this regard, we made use of inactivated intact type III GBS, a Gram-positive extracellular bacterium that expresses a capsule containing a core PS (i.e. GBS-III capsular PS without the terminal sialic acid) identical to that of the capsular PS of S. pneumoniae type 14. We also employed an isogenic mutant of GBS-III that expresses the core PS alone. In vivo IIIPS-and PPS14-specific IgM and IgG responses to intact inactivated GBS-III were compared to those elicited by isolated IIIPS, as well as to intact inactivated Pn14, another Gram-positive extracellular bacterium, using the i.p. route of immunization. The use of inactivated, as opposed to viable, intact bacteria was to focus selectively on the properties of these pathogens as complex particulate immunogens, with the understanding that bacterial metabolism and alternate routes of immunization will add additional layers of complexity to the conclusions being drawn.

These data, combined with our previous observations using intact Pn14 and MenC (18) lend further support to the proposal that the bacterial sub-capsular domain converts a normally TI PS-specific Ig response to one that is dependent on CD4+ T cells, including a requirement for B7-, and potentially ICOS-dependent, co-stimulation, and CD40/CD40L interactions. The stronger reduction of the PPS14- and IIIPS-specific IgG responses to GBS that occurred in the absence of T cells or in the presence of blocking anti-CD80/86 mAbs, in contrast to that resulting from blockade of CD40L or ICOSL, may reflect either incomplete blockade by the mAbs used in the latter responses, or additional helper activities of CD4+ T cells independent of CD40 or ICOS costimulation. However, a more absolute requirement may exist for B7-dependent costimulation for the early activation of T cells. The CD4+ T cell help during the primary response to the Gram-positive bacteria, GBS-III and Pn14, induces a relatively rapid and robust PS-specific IgG response comprising multiple IgG isotypes (45) (Fig. 2). In contrast, the primary PS-specific IgG response to the Gram-negative bacterium, MenC, develops more gradually and is TI (17). Whether this represents a more general dichotomy between Gram-positive and Gram-negative bacteria remains to be determined. However, induction of PS-specific IgG memory, a process dependent on CD4+ T cells, can occur following primary immunization with all three bacteria although only MenC and GBS-III elicit a boosted PS-specific IgG response following secondary immunization (17) (Fig. 6). Thus, recruitment of CD4+ T cells may potentially enhance PS-specific adaptive immunity to these pathogens, although as discussed below, the sub-capsular domain of Pn also appears to contain a structure that is immunosuppressive.

The immunologic properties of PS expressed by intact bacteria are thus distinct from those observed using classical TI antigens (7), which in isolation are divorced from the more complex context likely observed during natural infections. In particular, the inability of isolated IIIPS to induce a detectable IIIPS-specific IgG response in vivo (Fig. 2), is consistent with our earlier in vitro studies demonstrating a requirement for a second signal in addition to multivalent BCR cross-linking, for induction of Ig secretion and class switching in response to PS (9). The requisite second signal may comprise a TLR ligand, various inflammatory cytokines, and/or BAFF/TACI, collectively recruited into the PS-specific Ig response by the sub-capsular bacterial domain in the presence of innate immune cells (9, 46, 47). The degree to which natural purified bacterial PS induce IgG response in vivo may depend on the degree of contamination of the PS preparation with activating bacterial cell wall structures. In this report isolated PPS14 in contrast to IIIPS, induced a detectable, though relatively modest, PPS14-specific IgG response. However, we previously demonstrated that this PPS14-specific IgG response was critically dependent upon a contaminating TLR2 ligand, likely derived from the Pn cell wall (48).

The dependence of the boosted secondary IIIPS- and PPS14-specific IgG responses to GBS-III on CD4+ T cells, CD40/CD40L interaction, and B7- and ICOS-dependent co-stimulation (Figs. 4 and 5) suggests that PS-specific memory B cells and/or bacterial peptide-specific memory CD4+ T cells may be generated within a germinal center reaction in response to intact GBS-III. In this context, the ability of GBS-III to trigger a boosted PPS14-specific IgG response in Pn14-primed mice, that is dependent on CD4+ T cells (Figs. 6 and 7), may indicate that these two different pathogens share cross-reactive CD4+ T cell epitopes. This observation is not unprecedented in nature as memory T cells that are specific for one virus can become activated during infection with an unrelated heterologous virus (49). Of interest, we observe no evidence for cross-reactive protein-specific B cell epitopes between GBS-III and Pn14, in that detectable serum titers of protein-specific IgG, using whole protein extract from the corresponding unencapsulated bacteria for ELISA, are only observed for the homologous bacteria (data not shown). Alternatively, it is possible that CD4+ T cells that mediate help for PS-specific IgG responses to intact bacteria are largely specific for PS, as recently illustrated using a conjugate vaccine of IIIPS and carrier protein (50). Nevertheless, the attachment of PS to MHC-II was via carrier-derived peptide, covalently linked to the PS. Thus, PS-specific IgG memory responses to conjugate vaccines appear to require the covalent attachment of PS and protein (50–53).

However, in bacteria, it does not appear that capsular PS is directly linked to protein via a covalent bond (19, 20). In this regard, in a previous study using intact vaccinia virus, CD4+ T cell help for an in vivo IgG response to a specific viral protein was induced only by CD4+ T cells with the same protein specificity as the B cell (54). This requirement for intra-molecular help would thus be satisfied by a soluble PS-protein conjugate vaccine, but not by an intact bacterium. In addition, BCR-mediated, Ag uptake by B cells involves endocytic vesicles that are only 50–150 nm in diameter, and thus would be predicted to exclude intact bacteria (54). However, a single study demonstrated BCR-mediated uptake of intact Salmonella typhimurium by human B cells with subsequent display of MHC-II/peptide complexes and primary CD4+ T cell activation (55). Alternatively, PS-specific B cells could potentially acquire bacterial protein subsequent to processing by macrophages and/or DC (56, 57). Given these considerations, it is possible that intact bacteria elicit a TD response that produces a stable expansion of un-mutated PS-specific B cells, lacking properties of bona-fide memory B cells. The requirement for CD4+ T cell help could be cognate at the level of initial DC priming, but non-cognate for B cells, or perhaps entirely non-cognate in nature. In any event, our data indicate that BCR cross-linking is necessary, but not sufficient for triggering B cells from GBS-III-primed mice for a boosted secondary PS-specific IgG response, as neither secondary immunization with isolated PPS14 or unencapsulated GBS-III can mediate this effect (Fig. 7). In addition, CD4+ T cells are also required during secondary immunization to affect the PS-specific IgG booster response (Fig. 7).

Although the majority of PPS14-specific IgG elicited in response to either Pn14 or GBS-III shares the same dominant idiotype (J. Colino et al, submitted), Pn14 fails to elicit a boosted secondary PPS14-specific IgG response in Pn14-primed mice, and induces a blunted booster response in GBS-III-primed mice (Fig. 6). Further, co-immunization with GBS-III and unencapsulated Pn also leads to a lower boosted PPS14-specific IgG response, relative to immunization with GBS-III alone (Fig. 6). In light of our previous report that Pn inhibits TD IgG responses to co-immunized soluble antigens independent of capsule expression (58), these data suggest that Pn expresses an immunosuppressive structure in its sub-capsular domain, that is not present in the GBS-III strains used in this study. This does not appear to be secondary to the generation of CD25+ regulatory T cells (59). One possible candidate is phosphorylcholine (PC), a haptenic moiety covalently attached to the Pn cell wall teichoic acid and membrane lipoteichoic acid (4). Thus, previous studies demonstrated the ability of a secreted filarial protein (ES-62) to mediate immunosuppression through a mechanism that is dependent on ES-62 expression of PC and the triggering of IL-10 secretion (60). Extracellular bacteria, in addition to Pn, such as Pseudomonas aeruginosa, N. meningitidis, N. gonorrhoeae, and Haemophilus influenzae can also express PC on their LPS or pili in a regulated manner (61–63), but the immunologic consequences of this remain largely unexplored. In light of the above, the ability of R614 to partially boost the PPS14-specific IgG response in COH1-11-, but not R614-primed mice may in some way be related to the high levels of serum PC-specific IgM and IgG, that can bind to R614, at the time of secondary immunization, observed in mice primed with R614, but not COH1-11. The ability of unencapsulated GBS-III to partially restore the boosted PPS14-specific IgG response when co-immunized with Pn14 (Fig. 6) suggests that innate immune activation mediated by the GBS-III sub-capsular domain can overcome this potential Pn-mediated suppressive effect. In this regard, inactivated GBS is a significantly more potent activator of macrophages and monocytes than inactivated Pn, inducing higher levels of TNF-α (64, 65).

Abbreviations

GBS-III

Group B Streptococcus (Streptococcus agalactiae) type III

PS

polysaccharide

IIIPS

capsular PS of GBS-III

M781 and COH1

wild-type strains of GBS-III

COH1-11

isogenic mutant of COH1 expressing desialylated IIIPS capsule

COH1-13

isogenic mutant of COH1 lacking IIIPS capsule

Pn14

intact inactivated Streptococcus pneumoniae, capsular type 14

R614

wild-type strain of Pn14

R36A

isogenic mutant of Pn2 (strain D39) lacking capsular PS

PPS14

pneumococcal PS type 14

MenC

Neisseria meningitidis, serogroup C

MCPS

Meningococcal type C PS