INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Group B Streptococcus (GBS) is one of the major pathogens that can be transferred to neonates from the mother through the vaginal tract and causes neonatal bacteremia, sepsis, and meningitis (13, 36, 37). Protective immunity to this organism in neonates can be achieved by maternal antibodies to the capsular polysaccharide (CPS) of GBS, which is transferred through the placenta (31). Like other bacterial CPSs, the conjugation of GBS CPS to an appropriate carrier protein, such as tetanus toxoid, may result in an increase of the immune response to CPS, probably due to both the recruitment of carrier-specific T helper cells and other as-yet-undefined mechanisms (1, 24, 30, 31, 44). The most common GBS serotypes associated with invasive disease are types III, Ia, and V. In several animal studies, these GBS serotype CPS-tetanus toxoid conjugates have been shown to be effectively protective for the offspring after systemic immunization of the mother (1, 24, 30, 44). Since colonization of the genital and lower intestinal tracts is important in transmission of GBS, effective immunity at genital and rectal mucosal sites may be necessary to diminish or eliminate the colonization of this organism, thus preventing it from spreading. In recent years, studies by several groups have shown that intranasal (i.n.) vaccination with Streptococcus pneumoniae CPS conjugate vaccine can not only protect mice against invasive S. pneumoniae infection but also effectively reduce colonization in the lungs (21, 22, 38). In previous studies, we showed that GBS CPS type III (CPS III) conjugated with the effectively mucosa-binding nontoxic B subunit of cholera toxin (CTB) using the reductive amination (RA) method could induce both strong systemic and local mucosal immune responses and also that the levels of serum antibodies correlated with the opsonizing activity (40, 41).

The efficacy of CPS-carrier protein conjugates may be influenced by several factors, such as (i) the conjugation methods used, (ii) the extent of cross-linking between the CPS and the carrier protein, (iii) the molecular weight of the conjugate, and (iv) the content of free polysaccharide in the conjugate, which has been shown to inhibit the immune responses elicited by the conjugated CPS (33, 34). For conjugation to CTB, an especially sensitive and important aspect is the preservation of the binding activity of the coupled CTB to its mucosal receptor, the GM1 ganglioside (17). However, the influence of these characteristics of CPS conjugates on their immunogenicities has not been adequately examined after mucosal vaccination.

For practical reasons, the ability to combine different conjugate vaccines in formulations that can be administered simultaneously is important to permit stimulation of protection against multiple serotypes of GBS infection within a simple immunization schedule. Thus, possible interactions between conjugates must be considered. It has been reported that mono- and multivalent GBS CPS conjugate vaccines can be formulated which are efficacious in inducing protective immunity in animal models by systemic immunization (32). The possibility of negative interactions in mucosal immunization with GBS CPS conjugates needs to be addressed.

In this study, we synthesized GBS CPS III-CTB conjugates with different linkage types with or without a spacer. The CPS III-CTB conjugates were fractionated into large- and small-molecular-weight batches. In addition, based on the results with the CPS III conjugates, GBS CPS Ia was also conjugated with CTB by the RA method. The anti-CPS responses were investigated after i.n. immunization with those conjugates in a mouse model to address (i) the effects of different conjugation methods of GBS CPS III with CTB, the molecule size of the conjugate, and the amount of free polysaccharide in conjugates on the anti-CPS specific immune responses; and (ii) the immunogenicity of the CPS Ia-CTB conjugate and the effect of combined immunization with CPS Ia-CTB and CPS III-CTB conjugates on the different types of anti-CPS specific immune responses.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. The following reagents were used: adipic acid dihydrazide (ADH) (Fluka Chimie AG, Buchs, Switzerland); avidin, cyanobromide (CNBr), 2(N-morpholino)ethanesulfonic acid (MES), and cystamine (all from Sigma Chemical Co., St. Louis, Mo.); dithiothreitol (DTT) (Calbiochem, La Jolla, Calif.); 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide-HCl (EDAC), N-hydroxysuccinimidobiotin, o-phenylenediamine, and sodium m-periodate (Sigma Chemical Co.); sodium cyanoborohydride (Aldrich Chemie, Steinheim, Germany); and N-succinimidyl-3-(2-pyridyldithio)propionate (SPDP) (Pharmacia Fine Chemicals, Uppsala, Sweden).

CPS antigens and cholera proteins. GBS CPS III was purified from the culture medium of Streptococcus agalactiae strain M732 as described previously (40). GBS CPS Ia was purified from S. agalactiae strain SS615 by the same methods used for the purification of CPS III. The purified CPS III was composed of 18 to 20% (wt/wt) sialic acid and contained <1% protein. Purified CPS Ia had a larger molecular weight than purified CPS III. It contained 13% (wt/wt) sialic acid and <0.5% (wt/wt) protein. Recombinant CTB (rCTB) was purified from culture medium of Vibrio cholerae strain 358 as described previously (29). Purified CT was obtained from List Biological Laboratories Inc. (Campbell, Calif.).

Preparation of CPS III-rCTB conjugate with cystamine and SPDP. To perform thiolation of CPS III using cystamine, a solution of CPS III (5 mg/ml) was activated by CNBr at pH 10.5 for 6 min at 4°C. The weight ratio of CNBr to CPS was 1.5:1. The reaction mixture was brought to pH 8.5 by 0.5 M NaHCO₃, and cystamine was added to a final concentration of 0.5 M. After being tumbled for 18 h at 4°C, the mixture was dialyzed against distilled water and lyophilized. The presence of NH₂ on the cystamine-derivatized CPS was verified by a 2,3,6-trinitrobenzenesulfonate (TNBS) assay (15). Cystamine-modified GBS CPS III (15 mg in 1.5 ml of 0.1 M sodium phosphate buffer containing 10 mM EDTA, pH 7.5) was reduced by 50 mM DTT. After incubation at room temperature for 4 h, the thiolated CPS III was separated from low-molecular-weight reagents by passage through a Sephadex G-25 column (Pharmacia). The SH content of the CPS III was measured by means of the Ellman test (16).

1

1

Preparation of the CPS III-rCTB conjugate by EDAC with ADH as a spacer. To modify GBS CPS III with ADH as described previously (40), briefly, a solution of CPS III (5 mg/ml) was activated by CNBr at pH 10.5 for 6 min at 4°C. The weight ratio of CNBr to CPS was 1.5:1. The reaction mixture was brought to pH 8.5 by addition of 0.5 M NaHCO₃, and ADH was added to a final concentration of 0.25 M. After being tumbled for 18 h at 4°C, the mixture was passed through a Sephadex G-25 column with distilled water as the eluant. The fractions containing CPS were pooled, dialyzed against distilled water, and lyophilized. The presence of NH₂ on the ADH-derivatized CPS was verified by the TNBS assay.

Preparation of CPS Ia- and III-rCTB conjugates by RA. Conjugation was performed as described previously (41). Briefly, GBS CPS Ia or III (20 mg) in 2 ml of phosphate-buffered saline (PBS), pH 7.0, was incubated in darkness at room temperature for 2 h with 2.75 mM sodium m-periodate for CPS Ia and separately for 1.5 h with 4 mM sodium m-periodate for CPS III. Glycerol was then added to consume any residual periodates. The mixture was passed through a Sephadex G-25 column with distilled water as the eluant and lyophilized. The oxidized CPS Ia or III (15 mg) was dissolved in 2 ml of 0.1 M sodium bicarbonate, pH 9.0, and mixed with rCTB (15 mg). Sodium cyanborohydride was added to a final concentration of 20 mg/ml, and the mixture was incubated at 37°C for 5 to 6 days. The progress of conjugation was monitored by analyzing aliquots of the mixture at various times with fast-pressure liquid chromatography (FPLC) on a Superose 6 HR 10/30 column (Pharmacia Fine Chemicals) with PBS as the eluant at a flow rate of 0.5 ml/min. Conjugation was indicated by a progressive increase in a broad high-molecular-weight protein peak monitored by measurement of UV absorbance at 280 nm. After the conjugation was completed, sodium borohydride (10 mg/ml) was added to the reaction mixture to reduce the remaining free aldehyde groups, and the conjugate was purified by gel filtration.

Purification and fractionation of conjugates. All of the conjugates were purified by gel filtration on a Sephacryl S-300 HR 16/60 column (Pharmacia Fine Chemicals) eluted with PBS to separate the conjugate from unbound rCTB. For the different CPS III-rCTB conjugates, fractions of the conjugate peak were divided into two pools according to elution volumes. The large-molecular-weight fraction pool (L) contained the elution volumes from 40 to 52 ml, and the small molecular-weight fraction pool (S) contained those from 53 to 62 ml. For the CPS Ia-rCTB conjugate, fractions corresponding to the high-molecular-weight material were pooled (elution volume, from 38 to 49 ml). The conjugates or pooled fractions were then concentrated by ultrafiltration on a Millipore membrane with a 10-kDa molecular mass cutoff.

Analyses of conjugates. To determine the relative distribution coefficient (K_av) of the two sizes of CPS III-rCTB conjugates, these conjugate preparations were tested by FPLC on a Superose 6HR 10/30 column equilibrated with PBS buffer.

Immunization of mice. C57BL/6 female mice, 8 to 10 weeks old, were obtained from B&K Universal (Stockholm, Sweden, and Bomholtsgård, Denmark). For the studies of the immunogenicities of the different CPS III-rCTB conjugates, the conjugates were tested after division into L and small S molecular weight preparations. A control group was immunized with a mixture of 30 µg of free CPS III and rCTB. To evaluate the immunogenicity of CPS Ia conjugate and the effect of the combination with CPS III-rCTB, four groups of mice were immunized with (i) purified CPS Ia alone, (ii) CPS Ia-rCTB conjugate, (iii) CPS III-rCTB conjugate, or (iv) CPS Ia-rCTB with CPS III-rCTB conjugates. Four or five mice per group were immunized i.n. with 30 µg of CPS of the conjugates plus 1.5 µg of CT per dose at 0 to 2, 14 to 16, and 28 to 30 days. Each dose was divided and given on three consecutive days with a micropipettor with a volume of 20 to 30 µl, since our previous work has shown that this gives rise to an immune response stronger than that achieved with a single bolus administration (41). The animals were lightly anesthetized with methoxyflurane (Schering-Plough Animal Health Corp., Union, N.J.) for all immunizations. Blood samples were taken from the tail vein before immunization. The animals were euthanized 7 to 10 days after the last immunization, and blood was drawn from the subclavian vein. Perfusion-extraction with saponin was used to obtain lungs and vaginal specimens for antibody detection as previously described (40, 41).

Serologic studies. Antibodies to GBS CPS Ia and III were estimated by ELISA using biotinylated CPS Ia and III as antigens (42). Plates (Greiner, Frickenhausen, Germany) were coated with avidin (5 µg/ml) overnight and then incubated with 2 µg of biotinylated GBS CPS Ia (North American Biologicals Inc., Rockville, Md.) or III/ml for 4 h at room temperature. The tested samples were added in threefold serial dilutions and incubated overnight. A pool of positive serum from mice immunized with CPS Ia-rCTB or CPS III-rCTB conjugate was used as a positive control. Horseradish peroxidase-labeled goat antibodies to mouse IgG (Jackson ImmunoResearch Laboratories, Inc.), IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA (Southern Biotechnology Associates, Inc., Birmingham, Ala.) were used, and the ELISA was developed with o-phenylenediamine and H₂O₂. The antibody titers were expressed as the reciprocal dilutions of specimens giving an absorbance of 0.4 above background.

Statistics. The geometric mean (GM), standard deviation (SD), and standard error of the mean (SEM) were calculated. Student's t test with Bonferroni correction was used to compare mean values in different groups of mice. Statistical significance was defined as a P of <0.05.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of the CPS III-rCTB conjugates. GBS CPS III was coupled to the rCTB by three different conjugation methods. The conjugate preparations obtained with cystamine and SPDP, or ADH and EDAC, or reductive amination in the presence of sodium cyanoborohydride were designated III-rCTB_SPDP, III-rCTB_ADH, and III-rCTB_RA, respectively (Table 1). The conjugates were gel filtered on an S-300 HR Sephacryl column after synthesis, and their elution profiles were found to be similar (Fig. 1). Each conjugate was divided into L and S material. Analyses by FPLC on a Superose 6HR 10/30 column showed that the K_avs were 0.20 to 0.21 for the L and 0.35 to 0.37 for S preparations, and the estimated molecular masses by comparison with globulin proteins ranged between 500 and 5,000 kDa for L and 400 and 500 kDa for S (Fig. 2). The different L conjugates had similar CPS/protein ratios (0.9 to 1 to 0.95 to 1), whereas CPS III-rCTB_SPDP (S) had a lower CPS/protein ratio (1.09 to 1). These and various other characteristics of the different conjugates are described in Table 1.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Gel filtration of GBS type III CPS-rCTB reaction mixture on the Sephacryl S-300 HR 16/60 column after conjugation. (a) III-rCTBSPDP; (b) III-rCTBADH; (c) III-rCTBRA conjugates. The flow rate was 19.8 ml/h, and the fraction size was 1 ml. The CPS was monitored by the phenol sulfuric assay (optical density [O.D.] at 490 nm (), and the protein was monitored by the optical density at 280 nm (). The peaks containing the III-rCTB conjugate were divided into L (elution volume, 40 to 52 ml) and S (elution volume, 53 to 62 ml) Mr pools.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Elution profile (optical density [O.D.] at 280 nm) of GBS CPS III-rCTB conjugates on Superose 6HR 10/30 gel filtration column. The larger-Mr conjugates, III-rCTBSPDP, III-rCTBADH, and III-rCTBRA (a), had peak Kav values of 0.2, 0.21, and 0.2, respectively; the smaller-Mr conjugates, III-rCTBSPDP, III-rCTBADH, and III-rCTBRA (b), had peak Kav values of 0.35, 0.36, and 0.37, respectively. The void and bed volumes of the column were 8.8 and 25 ml, respectively.

Immunogenicity of different CPS III-rCTB conjugates and L and S fractions. (i) CPS III antibodies in serum. The postimmunization sera from the group of mice immunized with a mixture of free CPS III and rCTB contained levels of anti-CPS III specific IgM, IgG, and IgA antibodies that were three- to fivefold higher than the preimmunization level (Table 2). Surprisingly, immunization with both the III-rCTB_SPDP (L) and (S) conjugates elicited only levels of IgM-, IgG-, and IgA-specific antibodies similar to those obtained with the CPS III and rCTB mixture (Table 2). In contrast, both the III-rCTB_AHD and III-rCTB_RA conjugates induced significantly higher levels of serum IgG and IgA antibodies than the III-rCTB_SPDP conjugate [III-rCTB_RA (L) versus III-rCTB_SPDP (L); P < 0.001; III-rCTB_ADH (L) versus III-rCTB_SPDP (L), P < 0.01] (Table 2). Immunization with the smaller-M_r III-rCTB_ADH (S) and III-rCTB_RA (S) conjugates induced levels of serum IgG antibodies similar to those of the large-M_r conjugates prepared by the same methods, but the levels of serum IgA antibodies were significantly lower than those obtained with the L conjugates (P < 0.05 and P < 0.01, respectively).

(ii) IgG subclasses. The levels of serum anti-CPS III IgG subclasses induced by the III-rCTB_RA (L) and (S) conjugates were tested and compared with those induced by CPS mixed with rCTB. The III-rCTB_RA (L) conjugate induced significantly higher levels of anti-CPS serum IgG1, IgG2a, IgG2b, and IgG3 antibodies than the CPS III and rCTB mixture (Fig. 3). There were no significant differences in the levels of IgG1, IgG2b, and IgG3 antibodies induced by the S conjugate and the L conjugate, whereas the level of IgG2a after immunization with III-rCTB_RA (S) was much lower than that obtained with the L conjugate (P = 0.004).

View larger version (23K): [in this window] [in a new window]   FIG. 3.   Serum anti-CPS specific IgG subclass titers after i.n. immunization with large- and small-Mr III-rCTBRA conjugates. Antibody titers are given as log10 of the GM titer + SEM. The horizontal line indicates maximal titers seen in the preimmunization samples (0.7 log10 for IgG1 and IgG2a, 1.4 log10 for IgG2b, and 2.0 log10 for IgG3). The open bars represent the titers of large-Mr conjugates, the solid bars represent titers of small-Mr conjugates, and the shaded bars represent titers of a CPS III and rCTB mixture. Statistically significant differences are shown as follows: ***, P < 0.001, and **, P < 0.01 (compared with the CPS III and rCTB mixture group).

CPS III antibodies in lungs and vagina. The anti-CPS III IgG and IgA titers in lungs and vaginal tissues elicited by the L conjugates prepared with different coupling reagents were evaluated. The perfusion-extraction technique used for preparing the tissue extract specimens analyzed has been shown to remove >97% of serum antibodies from the lungs and >95% from the vaginal tissues (23). In the lungs, the conjugate III-rCTB_SPDP (L) induced only low levels of IgG and IgA anti-CPS antibodies, similar to the levels obtained with the CPS III and rCTB mixture. In contrast, the III-rCTB_AHD (L) conjugate and, even more pronounced, the III-rCTB_RA (L) conjugate induced a significantly higher anti-CPS III IgG and IgA response than the CPS III and rCTB mixture did (Fig. 4a). Among three types of conjugate, III-rCTB_RA (L) induced the highest levels of anti-CPS III antibodies, both in IgG (III-rCTB_RA versus III-rCTB_SPDP and III-rCTB_AHD; P < 0.001) and in IgA (III-rCTB_RA versus III-rCTB_SPDP and III-rCTB_AHD; P < 0.05).

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Anti-CPS specific IgG and IgA in lungs (a) and vagina (b) after i.n. immunization with large- and small-Mr CPS III-rCTB conjugates. Antibody titers are given as log10 of the GM titer + SEM. The horizontal line indicates maximal titers seen in the PBS control group (0.95 log10 in lungs and 1.4 log10 in vagina). The open bars represent the titers of larger-Mr conjugates, the solid bars represent titers of small-Mr conjugates, and the shaded bars represent titers of a CPS III and rCTB mixture. Statistically significant differences are shown as follows: ***, P < 0.001; **, P < 0.01; *, P < 0.05 (compared with the CPS III and rCTB mixture group).

Preparation and testing of GBS CPS Ia-rCTB conjugate alone and in combination with CPS III-rCTB conjugate. As described above, the results with CPS III indicated strongly that conjugation to rCTB by RA and isolating the L fraction yielded the most immunogenic vaccine for i.n. immunization. We wished to examine whether the RA method would also work well for conjugating other GBS CPSs to rCTB and tested this with CPS Ia.

(i) Characterization of Ia-rCTB_RA conjugate. A small amount of the oxidized CPS Ia was found to be insoluble, and after the conjugation reaction was completed, there was also a small amount of insoluble material in the Ia-rCTB_RA conjugate preparation. The gel filtration profile of the Ia-rCTB_RA conjugate showed that a major portion of the rCTB was conjugated to CPS Ia (Fig. 5). The fractions containing the largest-M_r material (the void volume) were collected in order to avoid unconjugated CPS and rCTB. The ratio (wt/wt) of CPS Ia and rCTB in the resulting final conjugate was found to be 0.83 to 1, and the yield of total CPS recovered in the conjugate was 16%. The conjugated CPS Ia, in concentrations down to 10 ng/ml, reacted with anti-GBS type Ia hyperimmune rabbit serum, and the conjugated rCTB, down to a concentration of 8 ng/ml, reacted with the anti-CTB mouse monoclonal antibody (not shown). The results indicate that the conjugate had the expected in vitro immunologic and receptor-binding properties.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Gel filtration of GBS CPS Ia-rCTB reaction mixture on the Sephacryl S-300 HR 16/60 column after conjugation. The flow rate was 19.8 ml/h, and the fraction size was 1 ml. The CPS was monitored by the phenol sulfuric assay (optical density [O.D.] at 490 nm []), and the protein was monitored by the optical density at 280 nm (). The peak containing a large-Mr CPS Ia-rCTB conjugate was pooled (elution volume, 38 to 49 ml).

(ii) Immunogenicity of CPS Ia-rCTB conjugate and effect of combination immunization. Similar to the case for the CPS III-rCTB conjugate, the CPS Ia-rCTB conjugate induced higher levels of serum IgG and IgA antibodies than did unconjugated CPS Ia (P = 0.005 and P < 0.05, respectively). After i.n. immunization with the CPS Ia-rCTB conjugate, a 15-fold-higher level of specific IgG (P < 0.01) and a 6-fold increase in specific IgA were recorded in the lungs, compared with purified CPS Ia alone (Table 3). Mucosal immunization with a combination of CPS Ia-rCTB and CPS III-rCTB conjugates did not affect the immune responses to the CPS Ia and III antigens compared to immunization with each conjugate alone (P > 0.10 for all titers measured). Thus, the antibody levels achieved in both serum and lungs with the combination were fully comparable to those achieved with immunization with the monovalent CPS Ia-rCTB or CPS III-rCTB conjugate separately (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ideal vaccine against GBS infections should stimulate both local mucosal and systemic immunities. In vaccinated women, the role of mucosal antibodies would be to prevent colonization of the female genital tract and possibly also to defend the respiratory tract of the newborn, while the systemic humoral immunity via transplacental transfer of IgG antibodies could protect the neonate when bacteria reach the bloodstream.

This study was undertaken with the long-term aim of developing a multivalent vaccine against GBS based on a cocktail of selected serotypes of CPS conjugated to rCTB and to be administered i.n. in order to stimulate both mucosal and systemic immunity. This approach has been taken based on several considerations. Secretory antibodies against CPS have been found to inhibit colonization by a capsulated pathogen, Haemophilus influenzae type b, in an infant rat model (25). It was also shown, both in mouse models and in humans, that immunization or colonization with GBS in the rectum or uterine cervix or even the nasal cavity could induce cervicovaginal IgA antibodies to GBS (18, 19). I.n. immunization has been found to be superior to other routes for inducing local immunity in the respiratory and genital tracts together with a strong serum antibody response (3-6, 20, 23). Finally, CTB has been proved to be an effective mucosal carrier protein for several conjugated polysaccharide antigens; including dextran (3), H. influenzae type b CPS (4), and recently also GBS type III CPS (40, 41).

It has been reported that GBS CPS III conjugated with tetanus toxoid by means of RA could induce high levels of anti-CPS antibody in serum after systemic immunization in both mice and human trials (24, 30). However, parenteral immunization with the GBS CPS III conjugate was unable to stimulate any significant mucosal anti-CPS response (40). In contrast, our previous studies showed that GBS CPS III coupled to rCTB by the RA method and given by i.n. immunization could effectively elicit strong T-cell-dependent immune responses both in serum and at mucosal sites (40, 41). In this study, we show by comparison with two other conjugation methods that GBS CPS-rCTB conjugate prepared by RA is the most immunogenic and is especially superior when the isolated large-molecular-weigh fraction (L) is used as a vaccine.

The three large-M_r conjugates with identical CPS and carrier protein components and similar molecule weights and ratios of CPS to protein exhibited significant disparity in immunogenicity. It has been shown that when bacterial CPS was conjugated to a carrier protein via an ADH linkage, immunogenicity could be greatly enhanced (7, 11, 12, 26-28, 40). Consistent with this, in the present study, the GBS III-rCTB_ADH (L) conjugate could induce substantial levels of both IgG and IgA antibodies in serum, lungs, and vagina. However, the magnitude of the response induced by this conjugate was smaller than that induced by III-rCTB_RA (L), both in serum and organs. This difference may be due to the more precise or effective coupling of CPS to rCTB by RA. Compared with the more random coupling of GBS CPS to rCTB via ADH, the conjugation using RA permits the CPS III-terminal ends of sialic acid residues to be exposed to the immune system effectively by specifically coupling the side chain terminal of CPS to rCTB. In addition, it has been reported that conjugation with ADH as a spacer was difficult to standardize, often resulting in conjugates with highly variable immunogencities (39).

In this study, even the large-M_r CPS III-rCTB conjugate prepared with SPDP failed to evoke more than marginal IgG or IgA antibody responses, even though this conjugate had a relatively small amount of free CPS and a molecular mass similar to that of the strongly immunogenic ADH (L) and RA (L) conjugates. The exact reason for this discrepancy is unclear. It has been reported that the SPDP disulfide bond of the conjugate can be unstable and have a short life in vivo (11). It is possible that degradation of the III-rCTB_SPDP conjugate by respiratory tract enzymes might have caused the low levels of specific responses after i.n. immunization. On the other hand, SPDP-prepared conjugates between protein antigens and rCTB have been stably immunogenic when given i.n. (23), and also other polysaccharide-rCTB conjugates using SPDP coupling have worked well (3).

It is known that glycoconjugate vaccines generally induce stronger anti-polysaccharide antibody responses with a broader isotype range, which mainly consist of IgM and IgG1 antibodies in mice (8), than those obtained with polysaccharide alone. In this study, i.n. immunization with the III-rCTB_RA (L) conjugate induced not only high levels of anti-CPS specific IgG1, IgG2b, and IgG3, but also a substantial level of IgG2a antibodies in serum. It has been shown that IgG2a and IgG3 are able to fix complement and promote opsonophagocytosis effectively in mice (10, 14, 35). Thus, the wider range of anti-CPS antibody isotypes after i.n. immunization with III-rCTB_RA (L) conjugates may have a significant functional impact on the extent of protection against encapsulated GBS infection.

Generally, large-M_r conjugates are more immunogenic than smaller ones in inducing anti-CPS serum antibody responses following parenteral immunization (45). The presence of free CPS in conjugate formulations is known to suppress the T-helper cell-dependent anti-CPS response after systemic immunization (2, 43). In the present study, the small-M_r conjugates prepared with either ADH or RA contained 50 and 66% free CPS, respectively. These small-M_r conjugates could induce a level of IgG antibodies in serum, lungs, and vagina similar to that induced by the large-M_r conjugates, which contained 15 to 20% free CPS. Thus, there was no significant suppression by the larger amount of free CPS in the smaller-M_r conjugates. However, these small-M_r conjugates generally elicited a lower level of specific IgA antibodies in serum and vagina than the large-M_r conjugates. In addition, a disparate pattern of serum IgG subclasses was found after i.n. immunization with the two conjugates of different molecular sizes, with a significantly lower level of specific IgG2a being induced by III-rCTB_RA (S) than by a large-M_r conjugate. This finding suggests that the molecular size of a conjugate and the amount of free CPS in a conjugate may influence the qualitative characteristics of the antibody response achieved.

Since the GBS CPS III-rCTB conjugate prepared with RA could most effectively improve the immunogenicity of CPS III, we coupled GBS CPS Ia to rCTB by the same method. Our result demonstrated improved immunogenicity of rCTB-conjugated CPS Ia compared with that of free CPS Ia after i.n. immunization. However, the immune responses to the conjugated CPS Ia did not achieve the same high antibody levels noted for the immune responses to the conjugated CPS III, either in serum or lungs. This might be due to the fact that CPS Ia is a larger-M_r polysaccharide than CPS III and also, or alternatively, the fact that there is a higher percentage of oxidation of CPS Ia, and therefore more extensively cross-linked conjugates (1, 45) may form insoluble material. Still, the Ia-rCTB_RA conjugate was impressively immunogenic compared with the unconjugated CPS Ia, and most importantly, combining i.n. immunization with CPS Ia-rCTB and CPS III-rCTB had no negative effect on either the type Ia or the type III antibody responses in serum and lungs.

To summarize, GBS CPS III-rCTB conjugates obtained by different coupling methods have different effects on evoking anti-CPS specific systemic and mucosal responses after i.n. immunization. The molecular sizes of the conjugates and the presence of free CPS do not greatly affect the specific IgG isotype immune response, but these parameters may affect the serum IgG subclass distribution and have a negative effect on the mucosal IgA antibody levels achieved. The GBS CPS Ia-rCTB and CPS III-rCTB conjugates prepared by the RA method show promise as a bivalent, or component of a multivalent, mucosal conjugate vaccine against GBS infection.