Protection against Streptococcus suis Serotype 2 Infection Using a Capsular Polysaccharide Glycoconjugate Vaccine

Abstract

Streptococcus suis serotype 2 is an encapsulated bacterium and one of the most important bacterial pathogens in the porcine industry. Despite decades of research for an efficient vaccine, none is currently available. Based on the success achieved with other encapsulated pathogens, a glycoconjugate vaccine strategy was selected to elicit opsonizing anti-capsular polysaccharide (anti-CPS) IgG antibodies. In this work, glycoconjugate prototypes were prepared by coupling S. suis type 2 CPS to tetanus toxoid, and the immunological features of the postconjugation preparations were evaluated in vivo. In mice, experiments evaluating three different adjuvants showed that CpG oligodeoxyribonucleotide (ODN) induces very low levels of anti-CPS IgM antibodies, while the emulsifying adjuvants Stimune and TiterMax Gold both induced high levels of IgGs and IgM. Dose-response trials comparing free CPS with the conjugate vaccine showed that free CPS is nonimmunogenic independently of the dose used, while 25 μg of the conjugate preparation was optimal in inducing high levels of anti-CPS IgGs postboost. With an opsonophagocytosis assay using murine whole blood, sera from immunized mice showed functional activity. Finally, the conjugate vaccine showed immunogenicity and induced protection in a swine challenge model. When conjugated and administered with emulsifying adjuvants, S. suis type 2 CPS is able to induce potent IgM and isotype-switched IgGs in mice and pigs, yielding functional activity in vitro and protection against a lethal challenge in vivo, all features of a T cell-dependent response. This study represents a proof of concept for the potential of glycoconjugate vaccines in veterinary medicine applications against invasive bacterial infections.

INTRODUCTION

Streptococcus suis is a Gram-positive encapsulated bacterium and one of the most important bacterial pathogens in the porcine industry, resulting in important economic losses (1). To date, 35 S. suis serotypes have been described based on the capsular polysaccharide (CPS) antigenic diversity. S. suis serotype 2 is considered the most virulent; it is the serotype most frequently isolated from clinical samples and associated with disease in pigs in most countries (2). S. suis, mainly serotype 2, is also an important emerging zoonotic agent for humans in close contact with pigs or pig-derived products (2). The natural habitat of S. suis is the upper respiratory tract of pigs, more particularly the tonsils and nasal cavities, as well as the genital and digestive tracts (3). Of the various manifestations of the disease, septicemia and meningitis are by far the most striking, but other clinical outcomes can also be observed (4). Although the incidence of disease in swine varies over time and is generally less than 5%, mortality rates can reach 20% in the absence of treatment (5). Affected animals are generally between 5 and 10 weeks of age, but infections have also been reported in newborn piglets to 32-week-old pigs (3).

The thick surface-associated S. suis CPS confers on the bacteria protection against the immune system, notably by resisting phagocytosis (6). As with most extracellular encapsulated bacteria, protection against S. suis is likely mediated by opsonizing antibodies which induce bacterial clearance by opsonophagocytosis. Anti-CPS antibodies have previously been demonstrated to be protective against S. suis serotype 2 infection following passive immunizations (7–9).

Research has been ongoing for years in the hope of developing an efficient commercial vaccine to protect postweaning pigs against S. suis disease. Yet, to our knowledge, no such vaccine is available. Commercial or autogenous bacterins, which are suspensions of heat-killed or formalin-killed bacteria, are used in the field with limited success (1, 10–13). Other strategies have been experimentally tested, such as live strains and subunit vaccines. The use of live avirulent strains gave inconsistent results and may present some safety concerns (zoonosis) (11, 14, 15). After several years of research, there is still no proven and commercially available protein-based subunit vaccine using well-characterized virulence factors and/or protective antigens (15). While some of these candidates failed, others are promising and still in early stages of clinical investigations. Because this pathogen has a multifactorial virulence mechanism and presents a relatively high phenotypic heterogeneity, vaccine development is a real challenge (2, 16).

As the CPS is the most external bacterial layer in contact with the host, antibodies against it are highly opsonizing and protective, as demonstrated for several encapsulated pathogens (17–20). Paradoxically, due to their carbohydrate nature, CPSs are in general considered poorly immunogenic and do not generate long-lasting adaptive immune responses. Indeed, polysaccharides, unlike proteins and peptides, are generally recognized as T cell-independent (TI) antigens, which illustrate their innate inability to stimulate helper T (Th) cells via major histocompatibility complex (MHC) class II signaling, resulting in low immune cell proliferation, no antibody class switching or affinity/specificity maturation, and, more importantly, lack of immunological memory (21). The scientific advancement allowing for the widespread use of carbohydrate-based vaccines was the discovery that when properly conjugated to protein carriers serving as T cell-dependent (TD) epitopes, polysaccharides become potent vaccine antigens (21). These vaccines were named “glycoconjugate vaccines.” Since it has been reported that anti-CPS antibodies, if produced, have a high protective potential against infection caused by S. suis (22, 23), an interesting strategy for the development of an S. suis type 2 vaccine would be the use of a glycoconjugate vaccine made of CPS coupled to an immunogenic carrier protein, such as tetanus toxoid (TT), diphtheria toxoid, cross-reactive material 197 (CRM₁₉₇), or many more (24). As a matter of fact, glycoconjugate vaccines are very successful in the fight against encapsulated human pathogens such as Haemophilus influenzae (Hiberix), Neisseria meningitidis (MenACWY), and Streptococcus pneumoniae (PCV13) (20). Despite the popular use of glycoconjugate vaccines in human medicine, this strategy has been poorly developed for veterinary practice. A CPS conjugate vaccine has been suggested for veterinary use against Actinobacillus pleuropneumoniae, the etiological agent of porcine pleuropneumonia (25–27).

In 2010, Van Calsteren et al. reported the exact structure of the repeating unit for the serotype 2 CPS (28). The CPS repeating unit is composed of a unique arrangement of 1 rhamnose, 1 glucose, 3 galactoses, 1 N-acetylglucosamine, and 1 sialic acid (also named N-acetylneuraminic acid [Neu5Ac]). The sialic acid (Neu5Ac) is found to be terminal on a branch with an α2,6 linkage to a galactose. The precise knowledge of the S. suis serotype 2 CPS structure provides the chemical bases for the construction of a glycoconjugate. Thus, the main objective of this study was to explore the feasibility/potential, immunogenicity, and protection in mice and in pigs of a CPS-TT conjugate vaccine against S. suis serotype 2. The effects of different adjuvants on the immunological features of the antibody response against the CPS were also studied.

MATERIALS AND METHODS

All chemical reagents used were of ACS grade or higher and were from either Sigma-Aldrich (Oakville, Ontario, Canada) or Acros Organics (Fisher Scientific, Ottawa, Ontario, Canada).

Bacterial strains and growth conditions.

S. suis serotype 2 reference strain S735 (ATCC 43765) was used as the source of type 2 CPS (28), as the target strain for in vitro opsonophagocytosis assays (OPA), and to prepare the heat-killed bacteria used to hyperimmunize mice. Isolated colonies on sheep blood agar plates were inoculated in 5 ml of Todd-Hewitt broth (THB) (Oxoid, Nepean, Ontario, Canada) and incubated for 8 h in a water bath at 37°C with agitation at 120 rpm. Working cultures were prepared by transferring 10 μl of 8-h cultures diluted 1:1,000 with phosphate-buffered saline (PBS) into 30 ml of THB, which was incubated for 16 h. Bacteria were washed once and resuspended in PBS to obtain 5 × 10⁸ CFU/ml. Heat-killed bacterial cultures were obtained as previously described (29). Briefly, overnight cultures were washed once with PBS and then resuspended in 30 ml of fresh THB. A sample was taken to perform bacterial counts on Todd-Hewitt agar (THA). Bacteria were immediately killed by incubating at 60°C for 45 min and then were cooled on ice. Bacterial killing was confirmed by absence of growth on blood agar for 48 h. Strains used for the swine challenge model are described below.

Isolation and purification of type 2 S. suis CPS.

Bacterial cultures were performed as described by Calzas et al. (30). CPS extraction and purification, followed by quality control procedures comprising protein determination with the modified Lowry protein assay kit (Pierce, Rockford, IL), nucleic acid quantification using an ND-1000 spectrometer (NanoDrop, Wilmington, DE), and one-dimensional (1D)/2D ¹H nuclear magnetic resonance (NMR) analysis to ensure purity and identity were performed as described by Van Calsteren et al. (28).

Depolymerization of type 2 CPS.

Twenty milliliters of a 2-mg/ml solution of CPS in 50 mM NH₄HCO₃ was transferred to a 50-ml conical polypropylene tube in an ice bath. A titanium 1/8-in. microtip probe mounted on a Virsonic 600 sonicator (VirTis, Gardiner, NY) was immersed in the CPS solution, and sonication was performed at 20 kHz and 24 W for 60 min. After sonication, a sample of CPS was taken to determine the weight-average molar mass (M_w) by size exclusion chromatography coupled with multiangle light scattering (SEC-MALS) as previously described (31), with some modifications. Briefly, the chromatographic separation was performed with two 8- by 300-mm Shodex OHpak gel filtration columns connected in series (SB-806 and SB-804), preceded by a SB-807G guard column (Showa Denko, Tokyo, Japan). Elution was done at 0.5 ml/min using 0.1 M NaNO₃ as the mobile phase. Molar masses were determined using a Dawn EOS MALS detector (Wyatt, Santa Barbara, CA), and calculations were performed with the ASTRA software version 6.1.1.17 (Wyatt) using 11 detectors from angle 34.8° to 132.2° (detectors 5 to 15) for the depolymerized samples. The remaining solution of depolymerized CPS was dialyzed against water (Spectra/Por; molecular weight cutoff [MWCO], 3,500 [Spectrum Laboratories, Rancho Dominguez, CA]) and lyophilized. The optimal conditions for sonication were determined in pretests using different time points (see Fig. S1 in the supplemental material).

Mild periodate oxidation of depolymerized CPS.

Depolymerized S. suis serotype 2 CPS (8.8 mg, 6.7 μmol) was incubated with 620 μM sodium periodate in 1.1 ml of water in the dark with a stirring magnet at room temperature for 1 h. An excess of two equivalents of triethylene glycol per periodate was added and left for 1 h to consume any residual periodate. The mixture was dialyzed against water (Spectra/Por; MWCO, 1,000 [Spectrum Laboratories]) and lyophilized. The optimal conditions for oxidation were determined in preliminary tests (data not shown).

The degree of oxidation of the sialic acid (Neu5Ac) residues was assessed by gas chromatography (GC) analysis of the peracetylated methyl glycosides adapted from a previously described method (32). Briefly, oxidized CPS (0.4 mg) was reduced by treatment with 100 μl of NaBH₄ (10 mg/ml) in water for 1 h at room temperature. The reaction was quenched with 5% acetic acid solution in methanol and evaporated to dryness using a stream of N₂. Evaporations were repeated 3 times by the addition of 250 μl of methanol each time. The composition of the residue was determined by methanolysis. To this end, methanol (465 μl) and acetyl chloride (35 μl), which generate HCl, were added to the residue. The solution was heated for 17 h at 75°C and evaporated to dryness, followed by addition of 500 μl of tert-butanol and evaporation to dryness again. The methyl glycosides were acetylated with 150 μl of pyridine and 150 μl of acetic anhydride at 100°C for 20 min. The cooled solution was partitioned with 5 ml of water and 1 ml of CH₂Cl₂. The organic layer containing the peracetylated methyl glycosides was analyzed by GC using flame ionization detection (GC-FID). GC-FID analysis was done on a Hewlett-Packard model 7890 gas chromatograph equipped with a 30-m-by-0.32-mm (0.25-μm particle size) HP-5 capillary column (Agilent Technologies, Santa Clara, CA) using the following temperature program: 50°C for 2 min, an increase of 30°C/min to 150°C, then an increase of 3°C/min to 230°C, and a hold for 5 min. The temperatures of the injector and the flame ionization detector were 225°C and 250°C, respectively.

Purification of TT monomer.

Tetanus toxoid (TT) monomer was obtained by gel filtration chromatography before conjugation. One milliliter of a liquid preparation containing 4.5 mg/ml protein (as determined by the modified Lowry protein assay) was loaded onto a XK16-100 column filled with Superdex 200 Prep Grade (GE Healthcare Life Sciences, Uppsala, Sweden) equilibrated in PBS (20 mM NaHPO₄ [pH 7.2], 150 mM NaCl) and eluted with the same buffer. The protein eluted from the column in two peaks: the earlier-eluting peak contained oligomerized toxoid, and the later-eluting peak, corresponding to a M_r of 150,000, contained TT monomer. Fractions corresponding to the later (monomer) peak were pooled, desalted against deionized water, concentrated using a Centricon Plus-70 centrifugal filter device (30K Ultracel PL membrane; Millipore, Billerica, MA), and then lyophilized.

Conjugation of type 2 CPS to TT by reductive amination.

Periodate-treated type 2 CPS (3.6 mg, 40 nmol) and purified TT monomer (3.0 mg, 20 nmol) were dissolved in 2.2 ml of 0.1 M sodium bicarbonate (pH 8.1) for the 2:1 conjugate ratio. Sodium cyanoborohydride (7.5 mg, 120 μmol) was added, and the mixture was incubated at 37°C with orbital agitation for 2 days. For the 1:1 conjugate ratio, conjugation was performed in the same manner as described above except with 1.8 mg (20 nmol) of oxidized CPS. Sodium borohydride (4.7 mg, 124 μmol) was then added to the reaction mixture to reduce any remaining free aldehyde groups. The produced 2:1 and 1:1 mixed conjugates were extensively dialyzed against water (Spectra/Por; MWCO, 3,500 [Spectrum Laboratories]) and lyophilized. Conjugation was controlled by gel shift on SDS-PAGE, immunoblotting, and high-performance liquid chromatography (HPLC) as described below. The conditions for conjugation by reductive amination were determined in pretests using different CPS-to-TT ratios, different percentages of CPS oxidation, and different incubation times (data not shown).

SDS-PAGE and immunoblotting.

Samples were diluted in twice-concentrated SDS-PAGE loading buffer containing 2-mercaptoethanol, denatured by heating at 100°C for 5 min, and separated by SDS-PAGE on 7.5% acrylamide gels. The gels either were stained with Coomassie G-250 stain or a silver stain kit (Bio-Rad Hercules, CA) or were transferred to nitrocellulose Western blot membrane (Bio-Rad). The membrane was then blocked with a solution of Tris-buffered saline (TBS) (10 mM Tris-HCl [pH 7.4], 150 mM NaCl) containing 2% skim milk for 1 h. The membrane was washed 3 times with TBS and incubated for 2 h with either a mouse monoclonal antibody (MAb) against serotype 2 CPS of S. suis (clone Z3) (23) diluted 1:50 or a mouse MAb against TT (clone HYB 278-01; Abcam, Toronto, Ontario, Canada) diluted 1:500 in blocking buffer. The membrane was then washed 3 times with TBS and incubated for 1 h with a goat anti-mouse IgG-IgM(H+L) horseradish peroxidase (HRP)-conjugated antibody (Jackson ImmunoResearch, West Grove, PA) either diluted 1:3,000 for CPS or diluted 1:1,000 for TT in blocking buffer. The membrane was washed 3 times with TBS and developed with a 4-chloro-1-naphthol solution (Sigma-Aldrich).

HPLC analysis of the conjugates.

HPLC analysis of the glycoconjugate preparations was done by size exclusion chromatography. The chromatographic separation was performed with three 8- by 300-mm Shodex OHpak gel filtration columns connected in series (two SB-804 and one SB-803) preceded by a SB-807G guard column (Showa Denko). The glycoconjugate vaccine was eluted with 0.1 M NaNO₃ at a flow rate of 0.4 ml/min using a Knauer Smartline system equipped with a differential refractometer (RI) detector model 2300 and a UV detector model 2600 at a wavelength of 280 nm. The conjugate preparation (8-mg/ml solution in the mobile phase) was injected using a 50-μl injection loop. In selected experiments (see below), the fractions eluting at the void volume, which correspond to the conjugate fractions, were pooled, dialyzed against water (Spectra/Por; MWCO, 12,000 to 14,000 [Spectrum Laboratories]), and lyophilized. This corresponds to the 2:1 fractionated conjugate.

Mouse immunizations.

All experiments involving mice were conducted in accordance with the guidelines and policies of the Canadian Council on Animal Care and the principles set forth in the Guide for the Care and Use of Laboratory Animals by the Animal Welfare Committee of the University of Montreal. Five- to 6-week-old C57BL/6 female mice (Charles River, Wilmington, MA), largely used for studies on S. suis pathogenesis (17, 33–35), were immunized subcutaneously with different doses of the S. suis CPS conjugate preparations (see Results) in 0.1 ml PBS on day 0 and boosted on day 21. In a first set of experiments aimed to compare different adjuvants, 3 groups (n = 10) received 25 μg of the 2:1 mixed conjugate dissolved in PBS adjuvanted with either 20 μg of CpG oligodeoxyribonucleotide (ODN) 1826 (InvivoGen, San Diego, CA), Stimune (Prionics, La Vista, NE), or TiterMax Gold (CytRx Corporation, Norcross, GA) following the manufacturers' recommendations. Three placebo groups (n = 5) received only PBS adjuvanted as described above. In a second set of experiments, a dose-response study was performed using groups of mice (n = 8) immunized with either 1, 2.5, 5, or 25 μg of the 2:1 mixed conjugate emulsified 1:1 (vol/vol) with TiterMax Gold. Mice (n = 8) immunized with similar doses of free (unconjugated) depolymerized CPS emulsified with TiterMax Gold were included for comparison purposes. A placebo group (n = 5) was also included. In a third set of experiments, to compare the efficiencies of different conjugates, groups of mice (n = 10) received 25 μg of either the 1:1 mixed conjugate, the 2:1 fractionated conjugate, or a free (unconjugated) mixture of 2 CPS to 1 TT. All preparations were emulsified with TiterMax Gold, and a placebo group was also included.

In all experiments, to follow antibody responses, mice were bled (10 μl) weekly on days −1, 7, 14, 21, 28, 35, and 41 postimmunization by the tail vein. Diluted blood was directly used in the enzyme-linked immunosorbent assay (ELISA) as described below. At day 42 postimmunization, mice were humanely euthanized and sera collected and frozen at −80°C for ELISA Ig titration and isotyping and for OPA analyses (see below).

Immunization and challenge of pigs.

Animals were treated in accordance with the ethical guidelines of the Institutional Animal Care and Use Committee of Boehringer Ingelheim Vetmedica, Inc. A total of 44 3-week-old piglets (±5 days) were obtained from a commercial herd in Nebraska, USA, for this study. Tonsil swabs from the study piglets were negative by S. suis serotype 2 PCR prior to study initiation (2). The piglets were blocked by litter and then randomly assigned to one of four groups by a biostatistician using SAS version 9.3: group 1, n = 14; group 2, n = 10, group 3, n = 15; and group 4, n = 5. Groups 1 to 4 were commingled until group 4 (strict control) was removed at study day 35. Blood samples were collected on study days 0, 21, and 34 for determination of serum antibody levels.

The vaccines and the placebo were adjuvanted with Stimune according to the manufacturer's instructions. The piglets were injected intramuscularly twice at a 3-week interval (study days 0 and 21) with 2 ml of the vaccine or placebo: group 1 was vaccinated with formalin-inactivated, adjuvanted S. suis type 2 culture containing 2 × 10¹⁰ bacteria (isolate 3977B, a virulent field isolate of S. suis serotype 2 previously used in a bacterin efficacy study [A. Neubauer, unpublished]), group 2 was injected with the adjuvanted 2:1 mixed conjugate vaccine containing either 85 μg (first vaccination) or 56 μg (second vaccination) of the antigen, and group 3 was given 2 ml of adjuvanted PBS.

On day 36, groups 1 to 3 were challenged intraperitoneally with 2 ml (3 × 10⁹ CFU/dose) of a late-exponential-phase S. suis type 2 culture grown in THB with 5% fetal bovine serum. The animals were challenged with isolate ATCC 700794, which is the established challenge strain in the model used. Following challenge, pigs were monitored daily over a period of 7 days for the presence of clinical signs. The individuals observing the animals were blinded to the treatments. Assessed were behaviors, including those indicating functional alteration of the central nervous system (CNS), and locomotion. The observations for general behavior were numerically scored as follows: 0, physiological; 1, depression; and 2, apathy. Observations for locomotion were scored as follows: 0, physiological; 1, slightly to moderately lame; 2, severely lame/reluctant to stand; and 3, animal partially/completely down, i.e., animals can rise but lie down again within 10 s. CNS signs were scored as follows: 0, absent; and 1, present. In accordance with the humane endpoints defined in the animal use protocol as well as 9CFR 117.4 (36), animals unresponsive to stimuli, animals exhibiting CNS signs, and animals that were assigned a lameness score of 3 were humanely euthanized. At study day 43, all remaining animals were humanely euthanized.

All animals found dead or that had been euthanized were necropsied. Meninges, pericardium, and joint swabs were collected and streaked out on Columbia blood agar plates. Alpha-hemolytic colonies present on the plates following overnight incubation were tested with type 2 antisera and, as deemed necessary, by S. suis type 2 PCR. Gross pathology observations were recorded for the thoracic cavity (presence of fibrin, fluid, congestion, and pericarditis), and for the joints (presence of fibrin and fluid). The individuals conducting the necropsies were blinded to the treatments.

Control mouse antiserum.

Hyperimmune mice (n = 6) were obtained by repeated immunization of 5-week-old female C57BL/6 mice with 7.5 × 10⁸ CFU/ml heat-killed S. suis serotype 2 strain S735 in THB by intraperitoneal injection on days 0, 7, 21, and 28. On day 42, serum was collected, pooled, aliquoted, and stored at −80°C.

Measurement of antibodies against type 2 S. suis CPS and TT.

To measure specific antibodies, 200 ng of either native S. suis serotype 2 CPS or TT in 0.1 M NaCO₃ (pH 9.6) were added to wells of an ELISA plate (Nunc-Immuno Polysorp; Canadawide Scientific, Toronto, Ontario, Canada). After overnight coating at 4°C, plates were washed with PBS containing 0.05% (vol/vol) Tween 20 (PBST) and blocked by treatment with PBS containing 1% (wt/vol) bovine serum albumin (BSA) (HyClone, Logan, UT) for 1 h. After washing, mouse blood or mouse/porcine serum samples diluted in PBST were added to the wells and left for 1 h. After washing, the plates were incubated for 1 h with an HRP-conjugated isotype specific antibody diluted in PBST as described below. The enzyme reaction was developed by addition of 3,3′,5,5′-tetramethylbenzidine (TMB) (Invitrogen, Burlington, Ontario, Canada) and stopped by addition of 0.5 M H₂SO₄, and the absorbance was read at 450 nm with an ELISA plate reader.

To follow the kinetics of total (IgG plus IgM) antibody responses to CPS and TT, mouse blood collected from the tail vein was diluted 1:100 or 1:20,000, respectively. Dilution optimization had previously been conducted (data not shown). HRP-conjugated goat anti-mouse IgG-IgM(H+L) at a dilution of 1:2,500 (Jackson ImmunoResearch) was used as a detection antibody.

To perform the titration of mouse Ig isotypes, day 42 serum was serially diluted (2-fold) in PBST, and antibodies were detected using either HRP-conjugated goat anti-mouse IgG plus IgM as mentioned above, goat anti-IgM diluted 1:1000, goat anti-IgG1, goat anti-IgG2b, goat anti-IgG2c, or goat anti-IgG3 diluted 1:400 (Southern Biotech). For porcine serum, 2-fold serial dilutions were performed in PBST and antibodies were detected using HRP-conjugated goat anti-swine total Ig (IgG plus IgM) diluted 1:4,000 (Jackson ImmunoResearch). To detect porcine IgG subclasses, unconjugated mouse anti-swine IgG1 or mouse anti-swine IgG2 (AbD Serotec, Raleigh, NC) diluted 1:250 was added, followed by incubation with HRP-conjugated goat anti-mouse secondary antibody. For both mouse and pig serum titration, the reciprocal of the last serum dilution that resulted in an optical density at 450 nm (OD₄₅₀) equal to or lower than 0.2 (as a preestablished cutoff for comparison purposes) was considered the titer of that serum. For representation purposes, negative titers (less than or equal to the cutoff) were given an arbitrary titer value of 10.

To control interplate variations, an internal reference positive control was added to each plate. For titration of mouse antibodies, this control was a pool of sera from hyper-immunized mice (produced as described above). For titration of pig antibodies, this control was a serum of a pig hyper-immunized with 10⁸ CFU of a killed suspension of S. suis serotype 2. Reaction in TMB was stopped when an OD_{450 nm} of 1 was obtained for the positive internal control. Optimal dilutions of the coating antigen (CPS or TT), the positive internal control sera and the HRP-conjugated anti-mouse or anti-pig antibodies were determined during preliminary standardizations.

OPA.

Blood was collected by intracardiac puncture from naive C57BL/6 mice, treated with sodium heparin, and then diluted to obtain 6.25 × 10⁶ leukocytes/ml in RPMI 1640 supplemented with 5% heat-inactivated fetal bovine serum, 10 mM HEPES, 2 mM l-glutamine, and 50 μM 2-mercaptoethanol. All reagents were from Gibco (Invitrogen, Burlington, Ontario, Canada). All blood preparations were kept at room temperature. Using washed bacterial cultures grown as described above, final bacterial suspensions were prepared in complete cell culture medium to obtain a concentration of 1.25 × 10⁶ CFU/ml. The number of CFU/ml in the final suspension was determined by plating samples onto THA using an Autoplate 4000 automated spiral plater (Spiral Biotech, Norwood, MA). All bacterial preparations were kept on ice. Diluted whole blood at 5 × 10⁵ leukocytes was mixed with 5 × 10⁴ CFU of S. suis (multiplicity of infection [MOI] of 0.1) and 40% (vol/vol) of serum from naive or vaccinated mice in a microtube to a final volume of 0.2 ml. The tube tops were pierced using a sterile 25-gauge needle, and then the microtubes were incubated for 2 h at 37°C with 5% CO₂, with gentle manual agitation every 20 min. After incubation, viable bacterial counts were performed on THA using an Autoplate 4000 automated spiral plater. Tubes with addition of naive rabbit serum or rabbit anti-S. suis type 2 strain S735 serum (37) were used as negative and positive controls, respectively. The bacterial killing percentage was determined using the following formula: percentage of bacteria killed = [1 − (bacteria recovered from sample tubes/bacteria recovered from negative-control tubes with naive mouse sera)] × 100. Final OPA conditions were selected based on several pretrials using different incubation times and MOIs (38).

Statistical analyses.

All data are expressed as means ± standard errors of the means (SEM). Data were analyzed for significance using analysis of variance (ANOVA) from SigmaPlot version 11.0, except for the survival curves analysis, which was performed using the log rank test from GraphPad version 5.01. Significance is denoted in the figures as follows: *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Summaries of and data analyses for the pig study were conducted by a biostatistician using SAS version 9.3. Clinical observations (death, lameness, CNS signs, and behavioral changes) were summarized as frequencies by day and treatment. The incidence of normal versus not normal for each characteristic was analyzed, where appropriate, using the GLIMMIX procedure of SAS with binomial error and logit link. The model included the fixed effect of treatment and the random effects of litter and residual. In addition, the proportion of the observations for each animal that were not normal was analyzed. Prior to analysis, the proportion was transformed using the arcsine square root transformation. The mixed model included the fixed effect of treatment and the random effects of litter and residual. Comparisons of interest include the following and were evaluated using a two-sided test with alpha = 0.05: groups 1 (bacterin) and 2 (conjugate) versus 3 (protection provided against challenge with isolate ATCC 700794).

RESULTS

Preparation of the conjugate vaccines.

Using highly purified CPS from S. suis type 2 containing less than 1% (wt/wt) of proteins or nucleic acids (as previously described by Calzas et al. [30]), we investigated whether conjugation to a carrier protein, such as TT, would circumvent the TI antigenicity of the CPS and instead induce a TD protective humoral response in animals.

Due to its high M_w, found to be between 410,000 and 480,000 Da by SEC-MALS (28, 30), the native polysaccharide of S. suis type 2 must first be depolymerized into smaller fragments. This reduces polydispersity and yields higher ratios of polysaccharide to protein following coupling. To perform this depolymerization, we opted for ultrasonic irradiation as described by Szu et al. (39). By monitoring the CPS M_w of samples by SEC-MALS during pretests, it was shown that depolymerization plateaued after 45 min of sonication (see Fig. S1 in the supplemental material). Based on these observations, we selected a depolymerization time of 60 min of sonication, at which two different lots were produced with reproducible results, giving an average M_w of 115,000 Da (113,000 to 118,000 Da) (see Table S1 in the supplemental material). ¹H NMR investigations of these two lots found no structural alteration of the polysaccharide other than the depolymerization itself (data not shown). These depolymerized CPSs were used in the subsequent preparation of the conjugate vaccine formulations.

The presence of sialic acid (Neu5Ac) as a constituent in the repeating-unit sequence of S. suis serotype 2 CPS allowed the use of mild conditions in order to achieve an oxidative cleavage between C-8 and C-9 of the glycerol side chain, thus leaving free terminal aldehydes as reacting groups for subsequent conjugation to TT by reductive amination (40). To preserve CPS immunogenicity, a 10% level of oxidation was targeted, leaving 90% of all Neu5Ac untouched. For an ∼115,000-Da CPS, this resulted in an average of 9 oxidized Neu5Ac per chain. Following pretests, 0.1 equivalent of sodium periodate per Neu5Ac was selected, and the two different CPS lots were oxidized. Reproducible oxidation levels at C-8 of 9.2 to 9.4% were obtained as determined by GC-FID analysis of the peracetylated methyl glycosides. No oxidation at C-7 was observed under these conditions. ¹H NMR investigations found no other structural modification of the polysaccharide (data not shown).

The depolymerized-oxidized CPS and purified TT monomer were conjugated at a molar ratio of 2 chains of CPS to 1 TT or at a molar ratio of 1:1 by reductive amination (41). The optimal incubation time for conjugation was found to be 2 days during pretests (data not shown). After incubation, remaining aldehyde groups were reduced by the addition of sodium borohydride. Reagents were then eliminated from the conjugate mixes by extensive dialysis against water.

Gel shift, Western blotting, and HPLC analysis confirm successful conjugation of CPS to TT.

The presence of conjugates in the different preparations was verified by gel shift and Western blot experiments (Fig. 1) and by HPLC analysis (Fig. 2). For the gel shift experiments, both Coomassie blue (Fig. 1A) and silver (Fig. 1B) staining showed a considerable shift from the purified TT monomer at 150 kDa (lanes 2) to a thick band of over 250 kDa in the conjugates (lanes 3 and 4). This shift resulted from the covalent addition of a random number of 115-kDa CPS chains to the protein. Interestingly, the selected silver stain kit included one step of dichromate oxidation, favoring a higher-intensity signal from glycoproteins (42). Accordingly, a strong reaction was observed with the conjugate preparations (Fig. 1B, lanes 3 and 4). Neither Coomassie blue (Fig. 1A), silver staining (Fig. 1B), nor Western blot using an anti-CPS MAb (Fig. 1C) revealed any band for the depolymerized CPS included as a control in all gels (lanes 5), illustrating its weak capacity to migrate through the gels under these assay conditions. As such, the positive signal observed for the bands of >250 kDa when revealed using an anti-CPS MAb definitely proved the covalent nature of the linkage between CPS and TT in the conjugates (Fig. 1C, lanes 3 and 4). Control staining using an anti-TT MAb (Fig. 1D) shows that the epitope to which the MAb binds was preserved in the conjugates. Preservation of TT antigenicity is essential, since it is the key mechanism allowing for the production of a T cell-dependent anti-CPS humoral response. It should be noted that differences in signal intensities between the 2:1 and 1:1 mixed conjugates (Fig. 1, lanes 3 and 4) are likely related to the total amounts of protein content (4.5 μg versus 6.3 μg, respectively) within the 10 μg of sample loaded per lane.

FIG 1.

SDS-PAGE and Western blot characterization of the glycoconjugates. Separating gels containing 7.5% acrylamide either were stained with Coomassie blue stain (A) or silver stain (B) or were transferred onto a nitrocellulose membrane and revealed by immunoblotting with either an anti-S. suis serotype 2 capsular polysaccharide (CPS) monoclonal antibody (MAb) (C) or an anti-tetanus toxoid (TT) MAb (D). Lanes 1, 10 μl of SeeBlue Plus 2 prestained protein standard (Ladder; Invitrogen); lanes 2, 10 μg of TT purified monomer; lanes 3, 10 μg of the 2:1 mixed conjugate; lanes 4, 10 μg of the 1:1 mixed conjugate; lanes 5, 10 μg of free depolymerized S. suis serotype 2 CPS.

FIG 2.

HPLC elution profile of the 2:1 mixed conjugate. Differential refractometer (RI) (solid line) and UV detection (dotted line) profiles are shown. Similar profiles were observed for the 1:1 mixed conjugate.

HPLC analysis (Fig. 2) showed the elution of the conjugate (>250 kDa), of free CPS (100 kDa), and of free TT (150 kDa). By integrating the UV signal at 280 nm from the chromatograms, it was estimated that 48% ± 6% (mean ± standard deviation [SD]) of the protein content from the mixture is indeed found in the conjugate fraction. Taken together, the gel shift and Western blot experiments combined with HPLC analysis of the two conjugate samples revealed the presence of conjugates in the 2 CPS:1 TT and 1:1 preparations.

Emulsifying adjuvants show higher immunomodulatory properties than CpG ODN for a polysaccharide antigen.

Using the 2:1 mixed conjugate, optimization of the immunization protocol was performed in a murine model using inbred C57BL/6 mice. The performances of three different adjuvants were compared. CpG ODN is a synthetic version of a bacterial oligonucleotide with unmethylated CpG motifs and acts as a Toll-like receptor 9 (TLR9) ligand with immunostimulatory properties toward a Th1 response (43). Stimune (Specol) is a water-in-oil adjuvant composed of purified and defined mineral oil (Markol 52) with Span 85 and Tween 85 as emulsifiers (44). It has been used as a good alternative to Freund's adjuvant for weak immunogens in animals, such as mice and pigs (45). TiterMax Gold is also a water-in-oil adjuvant consisting of squalene as a metabolizable oil, sorbitan monooleate 80 as an emulsifier, and CRL8300 (a patented block copolymer) and microparticulate silica as stabilizers (44). TiterMax Gold has been suggested as a superior alternative to Freund's adjuvant, providing comparable titers with fewer injections and less undesired reactivity in mice (46).

Based on the literature (47), a dose of 20 μg of CpG was chosen to be added for adjuvanting. In parallel, conjugates were emulsified with recommended ratios of 4 parts aqueous antigen per 5 parts adjuvant for Stimune or 1:1 for TiterMax Gold. Doses of 25 μg of the 2:1 mixed conjugate vaccine for each adjuvant were administered to mice on days 0 and 21. The kinetics of total IgG-plus-IgM antibody responses against CPS or TT were followed weekly in tail vein blood samples (Fig. 3). Overall, CpG ODN 1826 (Fig. 3A) gave the lowest anti-CPS and anti-TT responses. Also, anti-CPS Ig isotyping showed a strict IgM isotype response (data not shown). In contrast, Stimune (Fig. 3B) and TiterMax Gold (Fig. 3C) gave comparably strong anti-CPS and anti-TT total IgG-plus-IgM responses. Furthermore, anti-CPS Ig isotype switching was observed with both emulsifying adjuvants (see below). Although a memory antibody response against TT was observed with all three adjuvants, Stimune and TiterMax Gold induced faster and higher anti-CPS antibody levels after boosting, suggesting that generation of immunological memory against the CPS antigen is favored by these emulsifying adjuvants. Finally, it should be noted that no placebo mice, injected with only PBS and adjuvant, produced any nonspecific antibody response (data not shown).

FIG 3.

Kinetics of total antibody responses of mice immunized with 25 μg of the 2:1 mixed conjugate adjuvanted with either CpG (A), Stimune (B), or TiterMax Gold (C). Mice (n = 10) were immunized on day 0 and boosted on day 21. ELISA plates were coated with either native capsular polysaccharide (CPS) or tetanus toxoid (TT) and incubated with blood samples diluted 1:100 or 1:20,000 to measure anti-CPS and anti-TT antibodies, respectively. Total (IgG plus IgM) antibody levels are shown for individual mice, with horizontal bars representing means (±SEM) of values for optical density at 450 nm. The arrow at day 21 indicates the boost. To simplify the graph, kinetics for the respective placebo groups (which were all negative) are not shown.

As TiterMax Gold is recognized as one of the best adjuvants for mice (48, 49), it was selected for further immunizations with this species. Because its response is comparable to that of TiterMax Gold and it has been previously evaluated in large animals (49, 50), Stimune was selected as the adjuvant for immunization of pigs.

While the free polysaccharide is nonimmunogenic, a dose-response effect on antibody levels is observed with the conjugate vaccine.

Using TiterMax Gold as the adjuvant, mice were immunized on days 0 and 21 with doses of 1, 2.5, 5, or 25 μg of the 2:1 mixed conjugate to evaluate the dose-response effect on antibody production. Groups of mice were also immunized with different doses of S. suis serotype 2 free (unconjugated) CPS to assess if it could be immunogenic by itself when adjuvanted with TiterMax Gold.

Even at a high dose (25 μg) of free CPS, no significant total IgG-plus-IgM primary or memory antibody responses were observed throughout the immunization period (Fig. 4). In contrast, a dose-response effect was observed when mice were immunized with the 2:1 mixed conjugate, with the 25-μg dose yielding the highest total IgG-plus-IgM anti-CPS antibody response as measured in blood samples collected weekly.

FIG 4.

Dose-response effect on total antibody levels of mice immunized with either free depolymerized capsular polysaccharide (CPS) or 2:1 mixed conjugate at 1, 2.5, 5, or 25 μg adjuvanted with TiterMax Gold. Mouse groups (n = 8) were injected on day 0 and boosted on day 21. ELISA plates were coated with native CPS and incubated with blood samples diluted 1:100. Total (IgG plus IgM) anti-CPS antibody levels are shown for individual mice, with horizontal bars representing means (±SEM) of values for optical density at 450 nm. The arrow at day 21 indicates the boost.

Conjugation of S. suis type 2 CPS to TT induces antibody isotype switching in mice.

Not only a stronger response following boosting (as illustrated in Fig. 3 and 4) but also antibody isotype switching is a good indicator of conjugate immunogenicity and ability to induce a T cell-dependent response. Therefore, titers of the different anti-CPS antibody isotypes were determined in mice immunized with 25 μg of the 2:1 mixed conjugate vaccine adjuvanted with TiterMax Gold. As shown in Fig. 5, not only strong IgM titers but also high levels of all IgG subclasses were observed, including IgG1, IgG2b, IgG2c, and IgG3 specific for the CPS antigen. To evaluate if isotype switching was dependent on the adjuvant, serum samples from mice immunized with 25 μg of the 2:1 mixed conjugate adjuvanted with Stimune were analyzed. Stimune also induced isotype switching in mice; however, levels were lower and profiles differed from those observed with TiterMax Gold, with no production of the IgG2c and IgG3 subclasses (Fig. 6A).

FIG 5.

Titers of different anti-CPS antibody isotypes in mice immunized with conjugate vaccines adjuvanted in TiterMax Gold. Mouse groups were as follows: placebo (n = 5), 2:1 mixed conjugate (n = 18, animals from the 2 previous immunizations with 25 μg), 1:1 mixed conjugate (n = 10), 2:1 fractionated conjugate (n = 10), and 2 CPS:1 TT unconjugated control mixture (n = 10). All mice were immunized with 25 μg of antigen in TiterMax Gold on day 0 and boosted on day 21, and sera were collected at day 42. A pool of hyperimmune mouse sera from 6 mice was also included for comparative purposes. For the titration, ELISA plates were coated with native CPS and incubated with 2-fold serial dilutions of sera. Isotypes were detected using specific HRP-conjugated anti-mouse IgG plus IgM, IgM, IgG1, IgG2b, IgG2c, or IgG3 antibodies. Titers for individual mice are shown, with horizontal bars representing means ± SEM. #, titers significantly different from those of the placebo group (P < 0.05). Differences between other groups are denoted as follows: **, P < 0.01; ***, P < 0.001. Abbreviations: CPS, capsular polysaccharide from S. suis serotype 2; TT, tetanus toxoid.

FIG 6.

Isotyping and functional studies of antibodies induced in mice immunized with 2:1 mixed conjugate vaccine in Stimune. Mice (n = 10) were immunized on day 0 and boosted on day 21 with 25 μg of the 2:1 mixed conjugate adjuvanted with Stimune. Placebo mice (n = 5) were similarly injected with PBS adjuvanted with Stimune. Sera were collected on day 42. (A) For titration of anti-CPS antibody isotypes, ELISA plates were coated with native CPS and incubated with 2-fold serial dilutions of sera, and isotypes were detected using specific HRP-conjugated anti-mouse IgG plus IgM, IgM, IgG1, IgG2b, IgG2c, or IgG3 antibodies. (B) Opsonophagocytosis killing of S. suis type 2 strain S735 by day 42 sera from mice immunized with 25 μg of the 2:1 mixed conjugate adjuvanted with Stimune. A 40% (vol/vol) sample serum and a bacterial MOI of 0.1 were added to fresh whole blood from naive mice to perform the assay. Viable bacterial counts were determined after 2 h of incubation. To determine bacterial killing, viable bacterial counts from tubes incubated with sample sera were compared to those incubated with control naive mouse sera. Results are expressed as percent bacterial killing for individual mice, with horizontal bars representing means ± SEM. Abbreviations: CPS, capsular polysaccharide from S. suis serotype 2.

In order to determine the effect of the CPS-to-TT ratio on the conjugate immunogenicity, another mixed conjugate, this time using a ratio of 1 CPS to 1 TT, was prepared (Fig. 1) and emulsified in TiterMax Gold. Immunized mice showed similar IgM titers, reduced (but not significantly different) IgG1 and IgG3 titers, and significantly lower IgG2b titers (P < 0.01) than those induced by the 2:1 mixed conjugate in TiterMax Gold (Fig. 5). Interestingly, the 1:1 mixed conjugate failed to induce significant titers of the IgG2c subclass.

In order to prove that the observed immunogenicity was in fact due to the conjugate present in the vaccine formulation and not only due to remaining free CPS and TT, two additional controls were included in the study. A first control was the HPLC-isolated specific fraction corresponding to the conjugate from the 2:1 mixed conjugate (2:1 fractionated conjugate). The second control was a mixture of free CPS and free TT in the same ratio of 2:1 as before conjugation. In general, no major differences were observed between the 2:1 mixed conjugate and the specific 2:1 fractionated conjugate, yet higher titers of IgG1, IgG2b, and IgG3 were observed with the latter preparation (Fig. 5) (P < 0.01). In contrast, the mixture of unconjugated CPS and TT gave a strong IgM titer but very low titers of IgG1 compared to those with the 2:1 mixed conjugate (P < 0.01). In addition, no production of IgG2c subclass was observed in mice immunized with this control unconjugated preparation.

Finally, the control hyperimmune sera (from mice repeatedly injected with heat-killed bacteria) resulted in a high titer of IgM, production of IgG2b and IgG2c, but absence of IgG1 and IgG3 subclasses against the CPS antigen (Fig. 5). Similar results were obtained when mice were hyperimmunized with heat-killed bacteria adjuvanted in TiterMax Gold (data not shown).

Functional activity of antibodies.

The protective capacity of sera from immunized mice was evaluated using an OPA test, a recognized correlate of immunity for encapsulated Gram-positive bacteria, such as S. pneumoniae (51). Instead of using a cell line or a single cell type, the OPA was standardized using whole blood from naive mice. This model takes into account all blood leukocytes and thus represents a more realistic model of the complex interactions between all immune cells and the bacteria during a systemic infection, as is the case for S. suis. As shown in Fig. 7, sera from mice immunized with the 2:1 mixed conjugate adjuvanted with TiterMax Gold induced high bacterial killing levels ranging from 64 to 77%. Sera from mice immunized with the 2:1 fractionated conjugate gave higher, but not significantly different, bacterial killing values ranging from 74 to 98% (Fig. 7). The effect of the adjuvant was also evaluated in the OPA test; sera from mice immunized with the 2:1 mixed conjugate adjuvanted with Stimune induced bacterial killing levels ranging from 39 to 74% (Fig. 6B), which were not significantly different from those induced by TiterMax Gold (P > 0.05). When the OPA was performed using sera from mice immunized with the unconjugated CPS and TT mixture, significantly lower values (between 0 and 47%) of bacterial killing were observed compared to those with conjugates (Fig. 7) (P < 0.001). Finally, pooled sera from mice hyperimmunized with killed whole-cell preparations gave bacterial killing values highly similar to those with the unconjugated mixture (Fig. 7) (P > 0.05).

FIG 7.

Opsonophagocytosis killing of S. suis type 2 strain S735 by day 42 sera from mice immunized with different CPS conjugate vaccines adjuvanted with TiterMax Gold. Mouse groups were as follows: placebo (n = 5), 2:1 mixed conjugate (n = 10), 2:1 fractionated conjugate (n = 10), and 2 CPS:1 TT unconjugated control mixture (n = 10). All mice were immunized with 25 μg of antigen in TiterMax Gold on day 0 and boosted on day 21, and sera were collected at day 42. A pool of hyperimmune mouse sera from 6 mice was also included for comparative purposes (n = 8 experimental replicates). A 40% (vol/vol) sample serum and a bacterial MOI of 0.1 were added to fresh whole blood from naive mice to perform the assay. Viable bacterial counts were performed after 2 h of incubation. To determine bacterial killing, viable bacterial counts from tubes incubated with sample sera were compared to those incubated with control naive mouse sera. Results are expressed as percent bacterial killing for individual mice, with horizontal bars representing means ± SEM. #, values significantly different from those of the placebo group (P < 0.01). Differences between other groups: ***, P < 0.001. Abbreviations: CPS, capsular polysaccharide from S. suis serotype 2; TT, tetanus toxoid.

Immunogenicity and protection in pigs.

Based on previous results, the 2:1 mixed conjugate was selected to evaluate the immunogenicity and protection in the S. suis natural host, the pig. The adjuvant Stimune was chosen, as it had been previously included in S. suis bacterin-based vaccines (12, 52). The performance of the conjugate was compared to that of an S. suis type 2 bacterin adjuvanted with Stimune. To this end, pigs were injected intramuscularly twice at a 3-week interval, and serum samples were collected on days 0, 21, and 34 for titration and isotyping of anti-CPS antibodies (Fig. 8A). On day 21 postimmunization, total IgG-plus-IgM anti-CPS titers induced by the 2:1 conjugate vaccine were significantly higher (P < 0.01) than those with the placebo and bacterin. After boosting, on day 34 postimmunization, an increase in anti-CPS titers was observed for both vaccinated groups compared to those on day 21. However, only the titers from pigs vaccinated with the 2:1 mixed conjugate were significantly higher than those for the placebo control group (P < 0.001). Titers of the different swine IgG subclasses, namely, IgG1 and IgG2, were also assayed post-boost injection (day 34). Forty percent of pigs immunized with the 2:1 mixed conjugate showed significant levels of anti-CPS IgG1 subclass (Fig. 8A). No switch to the IgG2 subclass was observed (data not shown). In contrast, vaccination of pigs with the bacterin failed to induce an anti-CPS Ig class switch (Fig. 8A and data not shown).

FIG 8.

Immunogenicity and protection studies in pigs. Animals were blocked by litter and then randomly assigned to one of four groups: group 1, n = 14; group 2, n = 10, group 3, n = 15; and group 4, n = 5. Groups 1 to 4 were commingled until group 4 (strict control) was removed at study day 35. Blood samples were collected on study days 0, 21, and 34 for determination of serum antibody levels. The piglets were injected intramuscularly twice at a 3-week interval (study days 0 and 21) with 2 ml of the respective vaccine or placebo adjuvanted with Stimune: group 1 was vaccinated with adjuvanted S. suis type 2 bacterin, group 2 was injected with the adjuvanted 2:1 mixed conjugate, and group 3 was given 2 ml of adjuvanted PBS. (A) Kinetics of serum antibody response of immunized pigs. ELISA plates were coated with native capsular polysaccharide and incubated for 1 h with 2-fold serial dilutions of sera, and isotypes were detected using specific HRP-conjugated anti-pig IgG plus IgM or IgG1 antibodies. Antibody titers for individual pigs are shown, with horizontal bars representing means ± SEM. The arrow at day 21 indicates the boost. **, P < 0.01; ***, P < 0.001 (as determined by one-way ANOVA). (B) Protection study. On day 36, groups 1 to 3 were challenged intraperitoneally with 3 × 10⁹ CFU/dose of S. suis type 2 isolate ATCC 700794. Following challenge, pigs were monitored daily over a period of 7 days for the presence of clinical signs. Note that on day 21, one animal from the bacterin group was euthanized due to complications following serum collection, which leaves n = 14 at day 34 and for the challenge. **, P < 0.01 for both bacterin- and 2:1 mixed conjugate-vaccinated groups compared to placebo (challenge control) group.

On study day 36, pigs were challenged intraperitoneally with a dose of 3 ×10⁹ CFU of ATCC 700794, a virulent S. suis serotype 2 strain. Most pigs in the placebo group died during the systemic phase of S. suis infection, reaching a mortality of 86.7%. In contrast, pigs immunized with the bacterin or the 2:1 mixed conjugate showed mortalities of 28.6% and of 30.0%, respectively. Analysis of survival curves (Fig. 8B) showed a significant difference as soon as day 3 between both immunized groups and the placebo group (P = 0.009). Protection induced by the 2:1 mixed conjugate was similar to that for the control type 2 bacterin during the systemic phase of an S. suis challenge infection in pigs.

Pigs were also monitored for clinical signs (behavior, locomotion problems, or CNS signs) for seven consecutive days after challenge. In 31.6% of all observations for the bacterin-vaccinated group and in 28.1% of all observations for the conjugate-vaccinated group, abnormal behavior was observed. This was significantly lower than the findings in the placebo group, in which 90.7% of the observations revealed abnormal behavior (Table 1) (adjusted P value, <0.05). Lameness was observed in 26.3% of all observations for the bacterin-vaccinated group and in 33.5% of all observations for the conjugate-vaccinated group, compared to 89.3% for the placebo group (Table 1) (adjusted P value, <0.05). These differences were also observed when the distribution of clinical scores was analyzed daily for each group (data not shown). CNS signs were observed in only a few pigs, and no statistically significant differences were observed between the three challenged groups (Table 1). This can be explained by the fact that animals were observed for only a 7 day-period postchallenge, as the study design focused mainly on the systemic phase of the disease. All pigs found dead as well as all euthanized pigs were necropsied. The frequency of gross lesions in the thoracic cavity (i.e., fibrin, excess fluid, or pericarditis) or in the joints was overall reduced in vaccinated animals compared to the placebo group, though the observed differences did not reach statistical significance (see Table S2 in the supplemental material). The conjugate vaccine significantly reduced recovery of the challenge strain from joint swabs (P < 0.01). The S. suis challenge strain was also less frequently isolated from the meningeal and pericardial swabs than from the placebo group, yet the differences were not statistically different (see Table S3 in the supplemental material).

TABLE 1.

Clinical evaluation of immunized pigs after experimental challenge with S. suis serotype 2^a

Group	Abnormal behavior		Abnormal locomotion		CNS clinical signs
	%b	P valuec	%	P value	%	P value
Type 2 bacterin	31.6	0.0209	26.3	0.0064	1.2	NSd
2:1 mixed conjugate	28.1	0.0308	33.5	0.0335	0	NS
Placebo (challenge control)	90.7		89.3		2.6

^aAssessed were behavior, including any behavior indicating an effect of challenge on the central nervous system (CNS), and locomotion. The observations for behavior were numerically scored as follows: 0, physiological; 1, depression; and 2, apathy. Observations for locomotion were scored as follows: 0, physiological; 1, slightly to moderately lame; 2, severely lame/reluctant to stand; and 3, animal partially/completely down (animal can rise but lies down again within 10 s). CNS signs were scored as follows: 0, absent; and 1, present.

^bPercentage of evaluations where behavior, locomotion, or CNS signs gave a value of >0 across days. Assessment was for the cumulative observation period. Data are expressed as least-squares means (back-transformed).

^cAdjusted P value (Scheffé's test). All values are compared to the challenge control group; the strict control group was not included in the assessment.

^dNS, not significant (P > 0.05).

DISCUSSION

While glycoconjugate vaccines (made from CPS conjugated to an immunogenic protein carrier) are very effective in the fight against encapsulated bacteria in human medicine, so far no such vaccine is available for swine veterinary practice. Although some conjugates have previously been suggested for veterinary medicine use (25–27, 53), so far none have been marketed. For the first time, we report the feasibility of a glycoconjugate vaccine to protect pigs from S. suis type 2 infections. This glycoconjugate, made with type 2 CPS coupled to TT, is immunogenic in mice and pigs by inducing production of IgG antibodies, which are functional in vitro and protective in vivo.

Capsular polysaccharides are considered TI antigens, as they are in general unable to stimulate MHC class II-dependent Th cell help (30), which results in poor immunogenicity. In contrast to TD antigens, TI antigens do not induce the classical antibody isotype switch and high-affinity memory B cells (54). Yet, purified CPSs from S. pneumoniae (Pneumovax, 23 valent) and from group B Streptococcus (GBS) serotype III can induce not only IgM but also IgG antibody responses in mice and in adults without adjuvant (19, 55–58). In contrast, S. suis serotype 2 CPS alone is unable to induce any significant antibody response, even when adjuvanted with TiterMax Gold or Stimune or when combined with the TLR ligand CpG in mice (unpublished results). Furthermore, previous studies using live S. suis serotype 2 infection showed modest IgM and no isotype-switched IgG anti-CPS antibody titers in pigs and in mice even after an experimental reinfection (17). Thus, S. suis serotype 2 CPS seems to be particularly nonimmunogenic, making it an ideal candidate for evaluation of the effect of glycoconjugation on immunogenicity. A precedent exists in the literature, where serotype 2 CPS was conjugated to bovine serum albumin with the aim to obtain anti-CPS control sera for in vitro studies (59). However, neither the biochemical characteristics nor the immunogenicity and functional activity of that conjugate were investigated.

To produce the glycoconjugate, it was first decided to depolymerize the CPS to a smaller size in order to improve the efficacy of the conjugation and the residual exposure of the T cell peptide epitopes of the protein carrier. Ultrasonic irradiation (sonication) was chosen over fragmentation by chemical (60–62) or enzymatic (63–66) methods to avoid chemical alterations (67). In addition, it is easy to use, is reliable for labile epitopes, and does not require elimination of excess reagents. Another great advantage of ultrasonic irradiation is the reduction in sample polydispersity (39), facilitating biochemical characterization, particularly within the glycoconjugate.

In order to conjugate the CPS to a protein carrier by reductive amination, aldehydes must be introduced, ideally regioselectively, in the CPS by oxidation with sodium periodate. The presence of sialic acid offers a unique opportunity to use mild oxidative conditions, leaving the remainder of the polysaccharide unmodified, as already described for conjugates against different serotypes of GBS (41, 68–73). The glycerol side chain of the sialic acid residues is particularly reactive toward mild periodate oxidation. The desired percentage of oxidation is also a critical parameter: too few reactive groups will yield a poor conjugate, while too many will leave few intact epitopes of the native polysaccharide (40). For S. suis type 2, we found that a low oxidation level (10%) yielded immunogenic conjugates and was thus preferred over very high oxidation levels. In accordance, some immunogenic conjugates for GBS types Ia, Ib, and II have been described with an oxidation level of 10% (68–70). In contrast, conjugates of other GBS types, including type III, routinely use higher oxidation levels of 25% (41, 70, 71, 73) or between 40 and 50% (72) and could go up to 66% without loss of immunogenicity for serotype III (74), even though the type III GBS CPS conformational epitope is controlled by the sialic acid residue (75, 76).

Two conjugate vaccine formulations with different CPS to TT ratios were obtained. The 2:1 mixed conjugate was found to be the most immunogenic, resulting in significantly higher titers of IgG2b and IgG2c anti-CPS isotypes. This difference in immunogenicity may arise from the higher percentage of total CPS in the 2:1 than in the 1:1 mixed conjugate, which might influence the capacity of the conjugate to modulate the immune cells, including antigen-presenting cells (APCs), presumably through its higher molecular mass/size, which might ease uptake and internalization. In this regard, it has been shown that the polysaccharide size used for conjugation, the obtained conjugate size, and the degree of polysaccharide-protein cross-linking influence the immunogenicity and protective efficacy of a GBS type III-TT conjugate vaccine (74, 77, 78). Further studies evaluating how the aforementioned parameters affect the immunogenicity of an S. suis type 2 CPS conjugate, which can contribute to improvements of the design of the conjugate vaccine, are under way.

During pretrials, it was observed that conjugation alone was not enough to induce a robust immunological response against S. suis type 2 CPS (data not shown). In this regard, subunit vaccines are known to induce more potent and durable antigen-specific immunity if combined with an adjuvant (79). It has been shown that adjuvants not only can improve the immunogenicity of conjugate vaccines but also can direct the antipolysaccharide antibody isotype switch toward the desired IgG subclasses (80, 81). CpG ODN 1826 (82) was shown to enhance the isotype switching from IgM to IgG2a and IgG3 subclasses for pneumococcal conjugates in a serotype- and mouse age-dependent manner (43, 80, 83). In our study, CpG ODN 1826 failed to significantly improve the immunogenicity of the S. suis type 2 CPS mixed conjugate or to induce isotype switching. Conditions differing from those used in this study, such as the vaccine and adjuvant doses, might explain the observed differences. The biochemical properties of different CPSs might also influence the adjuvant capacity of CpG (83). As such, the choice of adjuvant is vital to maximize isotype switching. In the present study, high levels of IgM as well as of both Th1 (IgG2b, IgG2c, and IgG3) and Th2 (IgG1) IgG subclasses were observed in mice immunized with conjugates in TiterMax Gold. Similarly, Stimune induced high levels of IgM and a mixed Th1/Th2 IgG response. However, the Th1 antibody response was restricted to the IgG2b subclass when using Stimune. Our data are in agreement with previous studies suggesting that water-in-oil adjuvants, such as TiterMax Gold and Stimune, induce strong humoral responses by means of the depot effect and coimmunostimulatory properties, which are associated with mixed Th1 and Th2 antibody responses (46, 84–86). While Stimune is composed principally of purified and defined mineral oil, TiterMax Gold includes metabolizable squalene as the oil phase and also includes block copolymers in the formulation. This special formulation might contribute to the increased isotype switching activity observed with this adjuvant in mice. These block copolymers have been shown to influence the localization and retention of antigen in lymphoid tissues and to recruit and activate APCs and lymphocytes (48, 87).

However, even when using TiterMax Gold as adjuvant, not all mice developed anti-CPS IgG subclass-switched antibodies, although they all exhibited strong anti-CPS IgM titers in addition to high antibody levels to TT. Isotype switching from IgM to IgG depends on the signals (cytokines and costimulatory molecules) provided by APCs to T cells and then by T cells to B cells. These signals are induced by TT and the adjuvant, since the CPS is not immunostimulatory per se (24, 48). The intensity of these signals might vary among individual mice. As such, some of them will not be able to switch or will present isotype profiles different from those of other individuals from the same group.

Compared to the conjugate vaccine, a TiterMax Gold-adjuvanted control mixture of free CPS and TT induced a reduced TD anti-CPS antibody response. This also confirms that the immunogenicity in the 2:1 mixed conjugate is in fact due to the conjugate present in the vaccine formulation. Free polysaccharides in conjugate vaccines have been shown to have adverse effects on the immune response (by presenting mixed TI and TD forms of the CPS antigen) (88), as demonstrated with pneumococcal CPS types 3, 4, 6B, and 14 (89–91). In contrast, based on our results, free S. suis CPS does not seem to influence the immunogenicity of the conjugate vaccine. This is suggested by the facts that no suppressive effect was observed when using high doses of vaccine preparations and that the 2:1 fractionated conjugate behaves similarly to the mixed vaccine conjugate. The 2:1 mixed conjugate vaccine was also sufficient to induce protection in pigs, rendering the purification of the postconjugation preparation by chromatography unnecessary for potential use in the field.

While ELISA titers are great tools to evaluate the immunogenicity of vaccine prototypes, they are not generally considered reliable correlates of immunity. A strong antibody response does not necessarily reflect upon the protection of an individual (38). In this regard, functional assays are preferred, such as the OPA, a recognized correlate of protective immunity against extracellular encapsulated Gram-positive bacteria (18, 92). During the OPA, opsonizing antibodies from the immunized serum will opsonize the target bacteria, which results in bacterial phagocytosis and bactericidal activity (38, 93–95). Specific cell type activation depends on the Ig isotypes/subclasses present in the immune serum, since each isotype/subclass possesses different preferences for binding to Fc receptors, which differently influences the cell response (38, 94). Besides IgM, the predominant subclass of protective antibodies to TI antigens in mice is IgG3 (54, 96–98). A study using mouse monoclonal antibodies proposed that the type 1 subclasses (IgG3 ≫ IgG2b ≥ IgG2a) are superior in both opsonophagocytosis activity and complement activation to the type 2 IgG1subclass. Yet, these functional properties of mouse IgG subclasses seem to depend on the target antigen (protein versus carbohydrate), antigen distribution, and susceptibility of the bacteria to antibody/complement attack (99, 100). In our OPA experiments, the two groups which obtained the highest bacterial killing values were the 2:1 mixed or fractionated conjugate adjuvanted with TiterMax Gold, both containing the highest titers of all type 1 IgG subclasses. They were closely followed by the 2:1 mixed conjugate adjuvanted with Stimune, although this group lacks production of IgG3 and IgG2c. It has been shown that mice lacking the dominant IgG3 subclass made to bacterial CPS are more susceptible to fatal S. pneumoniae sepsis than wild-type mice, yet these mice can be rescued by induction of IgG1 using an S. pneumoniae glycoconjugate (100). In the IgG3^−/− mouse model, it was proposed that low titers of IgG1 anti-CPS antibodies in combination with complement deposition mediated by IgM anti-CPS antibodies may have been adequate to opsonize pneumococci for uptake by macrophages (100). Thus, in the absence of high levels of the most opsonizing antibodies (such as IgG3), other Ig isotypes/subclasses might compensate. As such, protection depends on the right balance between functionality, affinity, and quantity of different anti-CPS Ig isotypes/subclasses induced by the vaccine. In this regard, control mouse groups immunized with the mixture of free CPS and free TT or mice hyperimmunized with killed bacteria failed to adequately perform in the OPA test, probably due to the combined absence or low levels of several IgG subclasses, including IgG1. Together with the absence of protection in placebo groups, the OPA data confirm that anti-CPS antibodies generated by the glycoconjugate play a crucial role in opsonophagocytosis of S. suis and thus in the protective effect observed with the vaccine.

For an ultimate proof of concept, evaluation of the conjugate was conducted in the natural host of S. suis, the pig. In the field, pigs usually present clinical signs between 5 and 10 weeks of age (3), which would correspond to approximately 7 to 14 days postboost. Pigs immunized with the 2:1 mixed conjugate showed significant protection against a challenge systemic infection. As mentioned above, the combined action of IgM and IgG1 might be sufficient to confer protection in the swine model, yet the functional activities of different swine isotypes/subclasses remain to be elucidated. During swine immunization trials, a bacterin adjuvanted with Stimune was used as control. Although the bacterin induced levels of protection similar to those for the conjugate vaccine, this protection was not related to anti-CPS antibodies but probably was related to antiprotein antibodies. Yet, and in contrast to CPS, which is a universal antigen for S. suis type 2, protein antigens vary depending on the strain origin or sequence type (ST) within the same serotype (16, 101–105). Strains belonging to the same ST (ST1) were used as bacterin and challenge strains, and thus the capacity of such a vaccine preparation to protect against a heterologous challenge (ST25 and ST28 strains) remains to be elucidated (2). It can, however, be expected that protection conferred by a bacterin might be strain/ST dependent, while that provided by a CPS conjugate vaccine is serotype specific and thus strain/ST independent (106).

In conclusion, the conception and laboratory-scale production of a glycoconjugate vaccine against S. suis serotype 2 are reported for the first time. In its unpurified form as a desalted postconjugation formulation (mixed conjugate), the vaccine was shown to be efficient when emulsified with water-in-oil adjuvants. In mice, anti-CPS IgM and isotype-switched IgG antibodies were observed and found to be functional in an in vitro opsonophagocytosis assay. In pigs, anti-CPS IgM and IgG1 antibodies were detected and found to be significantly protective in an in vivo lethal-dose challenge with virulent S. suis serotype 2. At this stage, these results represent a proof of concept for the potential of glycoconjugate vaccines in veterinary medicine applications against invasive bacterial infections.

Funding Statement

This work was mainly supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) through a grant to MS (# 342150) and by Boehringer Ingelheim Vetmedica, Inc. Research Agreement No. 16579. Boehringer Ingelheim Vetmedica, Inc. provided support for the conduct of the research and the preparation of the article. The sponsor contributed to the in vivo study design; in the collection, analysis and interpretation of swine data; in the writing of the report; and in the decision to submit the article for publication. Partial contribution was also provided by Canadian Institutes of Health Research grant to MS, RR and MG (# 203549), by a NSERC Canadian Research Chair to RR (# UBR 303231), and by a Fonds de recherche du Québec - Nature et technologies (FRQ-NT) team grant to RR, MS, and MG.