|Home | About | Journals | Submit | Contact Us | Français|
Streptococcus pneumoniae is a leading cause of morbidity and mortality among children worldwide and particularly in developing countries. In this study, we evaluated PsaA, a conserved antigen important for S. pneumoniae adhesion to and invasion into nasopharynx epithelia, for its ability to induce protective immunity against S. pneumoniae challenge when delivered by recombinant attenuated Salmonella vaccine (RASVs) strains. RASVs were engineered to synthesize PsaA peptides of various lengths. Vaccination with an RASV synthesizing full-length PsaA induced high titers of anti-PsaA antibodies in both systemic (IgG in serum) and mucosal (IgA in vaginal washes, nasal washes, and lung homogenates) sites. BALB/c (haplotype H2d) or C57BL/6 (haplotype H2b) mice vaccinated either orally or intranasally exhibited a significant reduction in colonization of nasopharyngeal tissues after intranasal challenge with S. pneumoniae strains compared to controls, although protection was not observed with all challenge strains. None of the vaccine constructs provided protection against intraperitoneal challenge with S. pneumoniae strain WU2 (serotype 3). Immunization with RASVs synthesizing truncated PsaA generated lower titers of IgA and IgG and did not provide significant protection. Our results showed that RASVs synthesizing full-length PsaA can provide protection against nasal colonization by some S. pneumoniae strains. PsaA may be a useful addition to a multivalent vaccine, providing protection against pneumonia, otitis media, and other diseases caused by S. pneumoniae.
Streptococcus pneumoniae is responsible for a number of serious diseases in humans, including pneumonia, meningitis, bacteremia, otitis media, and sinusitis (31). It is a major cause of childhood mortality, 90% of which occurs in developing countries. The current vaccines against pneumococcal infections include a 23-valent capsular polysaccharide vaccine for adults and a 7-valent conjugate vaccine licensed for children (75, 77). However, some nonvaccine serotypes have become prevalent in the face of continued use of polysaccharide vaccines (63, 79). Also, certain high-risk groups have poor immunological responses to some of the polysaccharides in the vaccine formulations (28). There are also several concerns about the conjugate vaccines related to the cost and complexity of manufacture due to the different prevalent serotypes in different geographical areas. A meta-analysis showed that vaccination appears efficacious in reducing pneumococcal pneumonia in low-risk adults but not in high-risk groups (24). A more recent meta-analysis of 22 trials involving 101,507 participants found that the current 23-valent polysaccharide vaccine does not appear to be effective in preventing pneumonia, even in populations for which the vaccine is currently recommended (33, 52). There is a need to develop an improved and effective vaccine based on conserved antigens across all capsular serotypes to induce more effective and durable immune responses that could potentially protect against all clinically relevant pneumococcal capsular types and cover some high-risk groups who may not respond well to the current vaccine, while still keeping the cost low enough to be used in developing countries.
Studies of S. pneumoniae protective antigens have identified several candidate proteins that may be useful as vaccine components and drug targets, including PsaA, PspA, PspC, autolysin, pneumolysin, several neuraminidase enzymes, PcsB, and SktP (25, 80, 81, 88).
PsaA is a metal-binding lipoprotein with specificity for Mn2+ and Zn2+ (21, 41). psaA expression is upregulated during adherence to human lung epithelial cells and in blood or cerebrospinal fluid (20, 32, 61), and the protein plays a significant role in pneumococcal adherence and colonization. E-cadherin has been identified as the receptor for PsaA (1). These results indicate that PsaA is a critical factor in the first step for pneumococcal nasopharyngeal colonization and carriage. Mutations in psaA result in pleiotropic effects on a number of virulence functions in addition to adherence, including hypersensitivity to oxidative stress, a deficiency in Mn2+ transport and virulence (6, 14, 49, 57, 85). PsaA is a conserved antigen. It was present in all examined strains representing the 90 S. pneumoniae serogroups known at the time of the study, as well as other viridans streptococcal species (34, 54, 71). In addition, PsaA is immunogenic (9, 36), making it a desirable candidate for inclusion in a vaccine.
The primary translation product of the psaA gene is a 309-amino-acid (aa) polypeptide that includes a 20-aa N-terminal leader sequence containing the prolipoprotein recognition sequence LXXC recognized by signal peptidase II, two (β/α)4 domains, and an α-helical linker. Signal sequence cleavage results in a 290-aa mature protein anchored to the bacterial membrane via the resultant N-terminal Cys-linked lipid tail. The remainder of the protein is composed of the two (β/α)4 domains linked by an α-helix, forming two lobes with a cleft where the metal-binding site is located (41, 62).
Immunization with PsaA induced significant protection against S. pneumoniae colonization but only modest protection against invasive infection (8, 64, 81). Because PsaA and PspA have different functions in virulence, protection induced by these proteins may be additive. Indeed, promising results have been found for the combination of PsaA and PspA in the prevention of colonization and otitis media in animal models (3, 9, 58). Nasal immunization with six doses of lactic acid bacteria expressing psaA has been shown to induce anti-PsaA antibodies and to decrease colonization of the nasopharynx after intranasal challenge (59), although protection against intraperitoneal challenge was modest and not statistically significant. While these studies are promising, use of a more invasive vector may provide better stimulation of the immune system with fewer doses. Recombinant attenuated Salmonella vaccines (RASVs) can effectively colonize deep lymphoid tissues to induce long-lasting immune responses to delivered recombinant antigens as well as to vector antigens. In this work, we evaluated the utility of using a live attenuated Salmonella strain to deliver PsaA.
The bacterial strains and plasmids used in this study are listed in Table Table1.1. Salmonella enterica serovar Typhimurium vaccine strains were derived from the highly virulent parent strain, χ3761 (UK-1) (18). Bacteriophage P22HTint was used for generalized transduction (72). Serovar Typhimurium cultures were grown at 37°C in LB broth (7) or on LB agar with or without 0.05% arabinose. Diaminopimelic acid (DAP) was added (50 μg/ml) for the growth of Δasd strains (55). LB agar without NaCl and containing 5% sucrose was used for sacB gene-based counterselection in allelic exchange experiments. S. pneumoniae strains were cultured on brain heart infusion agar containing 5% sheep blood or in Todd-Hewitt broth plus 0.5% yeast extract (10). Growth on MOPS (morpholinepropanesulfonic acid) minimal medium with and without 10 μg/ml p-aminobenzoic acid was used to confirm the phenotype of pabA pabB mutants. The ΔasdA16 mutation was confirmed by the inability to grow in medium without DAP. The ΔaraBAD23 mutation was verified by PCR and its white colony phenotype on MacConkey agar with 1% arabinose.
Two complementary oligonucleotides P1 and P2 (Table (Table2)2) flanked by NcoI and SalI sites, respectively, containing the signal peptide of the lpp gene from Salmonella serovar Typhimurium LT-2 were annealed and cloned adjacent to the Ptrc promoter into pYA3342 (38) digested with NcoI and SalI to generate pYA3627. Two complementary oligonucleotides P3 and P4 flanked by NcoI and SalI sites, respectively, containing DNA sequences that code for the psaA signal peptide (aa 1 to 24) from Yersinia pestis KIM6+ were annealed and cloned adjacent to the Ptrc promoter in pYA3342 digested with NcoI and SalI to obtain pYA3638.
Using S. pneumoniae Tigr4 genomic DNA as the template, PsaA aa 21 to 210 were amplified by primers P5 and P9, cut with BamHI/HindIII and BamHI/SalI, respectively, and cloned into pYA3493 and pYA3620 to generate pYA3752 and pYA3753, respectively. Using the same procedures, PsaA aa 20 to 210, PsaA aa 17 to 19 and 21 to 210, and PsaA aa 17 to 210 were amplified with primer pairs P6/P9, P7/P9, and P8/P9 into pYA3493 (38) to generate pYA3756, pYA3760, and pYA3764, respectively, and into pYA3620 (19) to generate pYA3757, pYA3761, and pYA3765, respectively. The constructions were confirmed by DNA sequencing. The fragment encoding PsaA aa 21 to 210, PsaA aa 20 to 210, PsaA aa 17 to 19 and 21 to 210, and PsaA aa 17 to 210 were also cloned into pBAD-HisC to generate pYA3751, pYA3755, pYA3759, and pYA3763, respectively. There were three codons (G23, GGA to GGT; L33, CTA to CTG; and G206, GGA to GGT) changed to commonly used codons in Salmonella to optimize expression.
A 580-bp fragment of PsaA (aa 21 to 210) was amplified using plasmid pYA3751 as a template with primers P10 and P11, digested with SalI/HindIII, and cloned into expression plasmids pYA3627 and pYA3638 to generate pYA4092 and pYA4093, respectively.
Primers P11 and P12 were used to extend the N terminus of the truncated psaA gene carried by plasmid pYA3764 to aa 1 of the native amino acid sequence. The resulting full-length gene was cloned into pYA3342 to generate pYA4359.
The codon-optimized, truncated psaA gene (aa 1 to 210) carried by plasmid pYA4359 was extended to aa 309 (full length). Primers P14 and P15 were used to generate a fragment as a source of aa 211 to 309 from the S. pneumoniae Tigr4 genome, and primers P12 and P13 were used to generate a PCR fragment containing the psaA gene in plasmid pYA4359. These two fragments were annealed and amplified using primers P12 and P15 to extend the psaA product's C terminus to full length to encode aa 309 and cloned into pYA3342 to generate plasmid pYA4729. During construction, we introduced an additional codon change at G306 from GGA to GGT to codon optimize the new Tigr4 sequence for better gene expression in Salmonella. Primers P16 and P17 were used to amplify full-length psaA from pYA4729 and cloned into pET28a by using NdeI/XhoI to generate plasmid pYA4730.
Plasmid pYA3700 carries a tightly regulated araC PBAD TT cassette. To construct pYA3700, two oligonucleotides, P18 and its complement P19, corresponding to the T4 ipIII transcription terminator (53) and additional enzyme sites were annealed, cut with KpnI-PstI, and cloned into pGEM3Z cut with the same enzymes to create plasmid pYA3698 (Table (Table1).1). The araC PBAD cassette was amplified using plasmid pYA3624 (17) as a template with the primer pair P20 and P21. The resulting PCR fragment was cut with KpnI-XbaI and cloned into plasmid pGEM3Z to generate plasmid pYA3699 and into pYA3698 to generate the plasmid pYA3700.
The lacI gene with the natural GTG start codon (GTG-lacI) was amplified from the chromosome of Escherichia coli strain χ289 by using the primer pair P22 and P23 and cloned into pCR-Blunt II-TOPO. ATG-lacI was amplified using primer pair P22 and P24. The codon optimization of ATG-lacI was done by PCR. Briefly, 22 pairs of primers were used to modify 15 rare codons in lacI by PCR. The PCR products were used as templates and amplified again using primer pair P22 and P24 to yield the codon-optimized ATG-lacI. The 15 codons modified were codons 35 (CGG to CGT), 49 (CCC to CCG), 101 (CGA to CGT), 155 (CCC to CCG), 168 (CGA to CGT), 213 (ATA to ATC), 216 (CGG to CGT), 239 (CCC to CCG), 272 (GGA to GGT), 320 (CCC to CCG), 326 (AGA to CGT), 332 (CCC to CCG), 339 (CCC to CCG), 351 (CGA to CGT) and 355 (CGA to CGT). Using Salmonella serovar Typhimurium strain χ3761 chromosomal DNA as a template, DNA upstream (ygcA) of the relA gene was amplified using primer pair P25 and P26 and DNA downstream (STM2955) of the relA gene using the primer pair P27 and P28. lacI, relA-N(ygcA), relA-C(STM2955), and the T4 ipIII transcription terminator in pYA3698 and the araC PBAD fragment in pYA3700 were used to generate three STM2955::araC PBAD lacI TT::ycgA cassettes which harbor different lacI genes. The cassettes were used to generate suicide plasmids pYA3784 (GTG-lacI, ΔrelA196::araC PBAD lacI TT), pYA3789 (ATG-lacI, ΔrelA197::araC PBAD lacI TT) and pYA4064 (codon-optimized lacI, ΔrelA198::araC PBAD lacI TT). The ΔrelA196 deletion was introduced into χ8914 and χ8916 to generate χ9017 and χ9018. ΔrelA197 was introduced into χ8914 to generate χ9099. ΔaraBAD23 was introduced into χ8914 and χ9099 to generate χ9097 and χ9101, respectively. ΔrelA198 was introduced into χ9097 to generate χ9241.
Samples of recombinant PsaA and whole-cell lysates of RASV strains and S. pneumoniae strains were separated by 12% SDS-PAGE gels and then transferred to nitrocellulose membranes. The membranes were blocked with 3% skim milk in phosphate-buffered saline (PBS) with 0.05% Tween 20 (pH 7.4), incubated with rabbit polyclonal antibody raised against full-length PsaA (kindly provided by D. Briles, University of Alabama at Birmingham) or GroEL (Sigma, St. Louis, MO) and then with an alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma). Immunoreactive bands were detected by the addition of BCIP (5-bromo-4-chloro-3-indolylphosphate)-NBT (nitroblue tetrazolium) solution (Sigma). The reaction was stopped after 2 min by washing with large volumes of deionized water several times.
The interaction of anti-PsaA antibody with the surface of intact S. pneumoniae was measured by flow cytometry according to the method of Gor et al. (29). Briefly, frozen stocks of five pneumococcal strains were streaked individually onto blood agar plates and incubated overnight at 37°C. Bacteria were harvested from the plates, washed in PBS, and resuspended in strain buffer (PBS with 1% bovine serum albumin). Approximately 1 × 107 CFU of bacteria were incubated with 20% serum from mice inoculated with RASV strains carrying a psaA expression plasmid or an empty vector plasmid. After incubation, bacteria were washed with PBS and incubated with goat anti-mouse IgG conjugate with fluorescein isothiocyanate (FITC; SouthernBiotech, Birmingham, AL). Bacteria were then washed with PBS and subjected to flow cytometry by using a Cytomics FC500 flow cytometer. The data were collected and analyzed by using CXP software (Beckman Coulter Inc., Fullerton, CA).
Female BALB/c mice and C57BL/6J mice, 6 to 8 weeks old, were obtained from the Charles River Laboratories and Jackson Laboratory, respectively. All animal procedures were approved by the Arizona State University Animal Care and Use Committees. Mice were acclimated for 7 days after arrival before the experiments were started.
RASV strains were grown statically overnight in LB broth with 0.05% arabinose at 37°C and then subcultured 1:100 into fresh prewarmed LB broth with 0.05% arabinose with aeration at 37°C to an optical density at 600 nm of 0.8 to 0.9. Cells were harvested by centrifugation at room temperature (6,000 × g for 15 min), and the pellet was resuspended in buffered saline with gelatin (BSG). Serial dilutions of the RASV strains were plated onto MacConkey agar supplemented with 1% lactose to determine titers. Mice were inoculated intranasally with 10 μl or orally with 20 μl of BSG containing 1 × 109 CFU of the RASV or control strain. In some experiments, the mice were boosted at week 6 with the same dose by using the same route as that used for primary immunization. Blood samples were obtained by mandibular vein puncture at biweekly intervals. Following centrifugation, the serum was removed from the whole-blood samples, pooled, and stored at −20°C. Vaginal wash samples were collected at biweekly intervals, pooled, and stored at −20°C.
Serovar Typhimurium lipopolysaccharide (LPS) was obtained from Sigma. Recombinant PsaA (rPsaA) protein was purified with His-Select resin (Sigma) according to protocols provided by the manufacturer. The rPsaA clones used were pYA3763 (aa 17 to 210) and pYA4730 (full-length PsaA).
An enzyme-linked immunosorbent assay (ELISA) was used to assay antibodies in serum to serovar Typhimurium LPS and rPsaA and in vaginal washes, nasal washes, and lung homogenates to rPsaA. Samples from nasal washes and lung homogenates were collected 5 to 6 days after challenge and filtered for ELISA. Briefly, 96-well Nunc-Immuno MaxiSop plates (Nalgene Nunc International, Denmark) were coated overnight with 100 ng/well of LPS or purified rPsaA at 4°C. After blocking with a buffer containing PBS (pH 7.4), 0.1% Tween 20, and 10% Sea Block blocking buffer (Pierce, Rockford, IL), 100 μl of a serially diluted sample was added to individual wells in triplicate and incubated for 1 h at 37°C. Plates were then treated with biotinylated goat anti-mouse IgG or IgA (SouthernBiotech, Birmingham, AL). Wells were developed with a streptavidin-alkaline phosphatase conjugate (SouthernBiotech, Birmingham, AL), followed by p-nitrophenylphosphate substrate (Sigma) in glycine buffer (pH 9.8; Sigma). Color development (absorbance) was recorded at 405 nm using an automated ELISA plate reader (SpectraMax M5; Molecular Devices, Sunnyvale, CA). Absorbance readings that were 0.1 higher than PBS control values were considered positive.
At week 10, mice were challenged either by intraperitoneal injection with 2 × 104 CFU of S. pneumoniae WU2 (equal to 100× the 50% lethal dose [LD50]) (38, 56) or intranasally with 20 μl containing 5 × 106 CFU S. pneumoniae strain L82016 or E134 or 1 × 107 CFU of strain A66.1 or D39 (9, 89, 90). Mice challenged intraperitoneally were monitored daily for 30 days. For intranasally challenged mice, nasal washes were performed using 1 ml of saline after 5 to 6 days. Mouse lungs were collected and homogenized in 1 ml PBS. Serial dilutions of the samples were plated onto blood agar containing 4 mg/ml gentamicin. Alpha-hemolytic colonies were counted after incubation of the plates for 24 h at 37°C, and the number of CFU/ml was calculated based on the recovered volume. The detection limit was 10 or 20 CFU/ml, depending on the volumes plated. For representation in graphic and statistical analysis, log10 was applied to the values, and recovery of 0 CFU was considered the detection limit of 10 or 20 CFU.
All statistics were carried out using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). Antibody titers were expressed as means ± standard errors. The means were evaluated with two-way analysis of variance and Bonferroni's test for multiple comparisons among groups. The Mann-Whitney U test was used for the analysis of S. pneumoniae colonization. Differences were considered significant at a P value of <0.05.
Initially, we constructed plasmids fusing the β-lactamase signal sequence, the Y. pestis psaA signal sequence or the Salmonella Lpp secretion signal sequence to α-helical hydrophilic segments of PsaA (Fig. (Fig.1).1). Although mice immunized with Salmonella carrying these plasmids generated anti-PsaA serum antibodies (see Fig. S1 in the supplemental material), none of these constructs induced protective immunity against intraperitoneal challenge with the S. pneumoniae WU2 strain (see Table S1 in the supplemental material). Thus, we utilized the native S. pneumoniae psaA signal sequence and constructed plasmid pYA4359, encoding aa 1 to 210 of PsaA. In addition, we constructed plasmid pYA4729 containing DNA that codes for the entire psaA gene product (aa 1 to 309), including a functional epitope at aa 245 to 272 for adherence to nasopharyngeal cells (36, 65, 69) not encoded by pYA4359. Both of these two plasmids were moved into attenuated Salmonella strain χ9241, carrying a regulated delayed expression cassette in the chromosome (ΔrelA198::araC PBAD lacI TT) which provides for arabinose-regulated lacI expression. Note that in plasmids pYA4359 and pYA4729, psaA transcription is driven by the LacI-repressible Ptrc promoter. Thus, when cells are grown in the presence of arabinose in vitro, LacI is produced, repressing the plasmid-encoded Ptrc promoter, repressing antigen expression. The LacI synthesis stops in vivo due to the lack of arabinose availability in host tissues. Under these conditions, psaA expression increases as LacI decreases as a consequence of cell division. This system will be described in detail elsewhere (S. Wang, Y. Li, G. Scarpellini, W. Kong, H.-Y. Shi, C.-H. Baek, B. Gunn, S.-Y. Wanda, K. Roland, X. Zhang, P. Senechal-Willis, and R. Curtiss III, unpublished data).
Whole-cell lysates from strains χ9241(pYA3342) (empty vector), χ9241(pYA4359) (aa 1 to 210), and χ9241(pYA4729) (aa 1 to 309) grown in LB medium were evaluated for PsaA synthesis. Both strains synthesized proteins of the expected sizes that reacted with anti-PsaA antisera, about 23.4 kDa for strain χ9241(pYA4359) and 34.6 kDa for χ9241(pYA4729) (Fig. (Fig.2).2). Strain χ9241(pYA4729) appeared to synthesize more PsaA protein than did χ9241(pYA4359). However, it is possible that because the shorter protein synthesized from pYA4359 contains fewer epitopes, it may not react as well with the antisera as would the full-length PsaA synthesized in the strain carrying pYA4729.
Strain χ9241(pYA4359) (aa 1 to 210), χ9241(pYA4729) (aa 1 to 309), or χ9241(pYA3342) (control) was used to intranasally or orally inoculate C57BL/6J mice. The levels of IgG against PsaA and Salmonella LPS in the sera and anti-PsaA IgA in vaginal washes from immunized mice were measured (Fig. (Fig.3).3). Both immunization routes generated high titers of anti-PsaA IgG in sera (Fig. (Fig.3A),3A), with strain χ9241(pYA4729) inducing significantly higher titers than strain χ9241(pYA4359) (P < 0.001). Similarly, the anti-PsaA IgA responses in vaginal washes were significantly greater in mice immunized with χ9241(pYA4729) than in mice immunized with χ9241(pYA4359) (Fig. (Fig.3B)3B) (P < 0.05). Intranasal immunization with χ9241(pYA4359) elicited higher anti-PsaA titers than did oral immunization at 6 and 8 weeks for IgG (P < 0.05) and 8 weeks for IgA (P < 0.001) (Fig. 3A and B). For mice immunized by χ9241(pYA4729), intranasal immunization generated higher titers of anti-PsaA IgG than oral immunization at 2 weeks for IgG (P < 0.001) and 4 weeks for IgA (P < 0.001), but less for IgA at 2 and 8 weeks (P < 0.001 and P < 0.05, respectively). The peak IgG and IgA antibody titers occurred at 4 weeks and did not increase, even at 8 weeks after boosting (Fig. 3A and B). No anti-PsaA IgG was detected in mice immunized with χ9241(pYA3342). The anti-LPS responses were similar for strains with blank vector for both immunization routes, while for strains harboring an expression plasmid, titers were slightly higher in mice immunized by the intranasal route than in those immunized by the oral route at 2, 4 and 6 weeks, and modest boosting was observed at week 8 for all strains expressing psaA (Fig. (Fig.3C3C).
Immunized mice were challenged intraperitoneally with 100 LD50s of S. pneumoniae strain WU2. Despite the high titers of anti-PsaA antibodies (see Fig. S1 in the supplemental material), none of the mice survived challenge, indicating that, in the context of our system, PsaA is not a protective antigen against systemic infection (see Table S1 in the supplemental material).
PsaA is a colonization factor; therefore, we evaluated the effect of immunization on colonization. Immunized C57BL/6 mice were challenged with S. pneumoniae strain L82016, and bacteria were recovered in nasal washes after 5 days. There were no significant differences observed in nasal colonization for mice immunized by either route with χ9241(pYA4359) (aa 1 to 210) compared to that for mice immunized with the control strain χ9241(pYA3342) (P > 0.05) (Fig. (Fig.44 A). In contrast, immunization by either route with the strain expressing full-length psaA, χ9241(pYA4729) (aa 1 to 309), led to a significant reduction in colonization (P = 0.01 for orally immunized mice and P < 0.05 for intranasally immunized mice) compared to that for mice inoculated with the control strain, χ9241(pYA3342).
We then determined the anti-PsaA IgA titers present in the same nasal washes used to determine colonization. Consistent with the colonization data, no anti-PsaA IgA antibody was detected in nasal washes from mice immunized with χ9241(pYA4359) (Fig. (Fig.4B).4B). The inability of the truncated PsaA to generate high anti-PsaA serum IgG and mucosal IgA titers probably accounts for its lack of protective efficacy. Anti-PsaA IgA was detected in mice immunized with χ9241(pYA4729) after challenge with L82016 (Fig. (Fig.4B).4B). Although the anti-PsaA IgA titers in orally immunized mice were significantly lower than those in intranasally immunized mice, the results suggest that the titers were sufficient to reduce L82016 colonization. Taken together, these results indicate that full-length PsaA, but not truncated PsaA, was necessary to induce protective immunity against nasal colonization by S. pneumoniae strain L82016.
Innate resistance to S. pneumoniae infection in mice has been associated with its major histocompatibility complex (MHC) haplotype (26). BALB/c mice (MHC haplotype H2d) are significantly more resistant to intranasal challenge with S. pneumoniae strain D39 than are C57BL/6 mice (MHC haplotype H2b) (26). To investigate whether this might affect protective immunity, we compared the immunogenicities and protective efficacies of χ9241(pYA4729) in C57BL/6 mice and BALB/c mice. Mice were immunized either intranasally or orally using the same regimen as that used in the previous experiment. Anti-PsaA serum IgG titers were significantly lower in BALB/c and C57BL/6 mice immunized orally than in those immunized intranasally at all weeks (Fig. (Fig.55 A) (P < 0.001). At 2 and 4 weeks postimmunization, the BALB/c mice generated lower antibody titers than did C57BL/6 mice in response to either intranasal or oral immunization. By 6 weeks, both groups of mice immunized with χ9241(pYA4729) had developed similar titers, while at 8 weeks, intranasally immunized BALB/c mice generated higher antibody titers than did intranasally immunized C57BL/6 mice. The biggest difference in immune responses between mouse strains was seen in the mucosal anti-IgA titers (Fig. (Fig.5B).5B). Overall, BALB/c mice developed higher titers than did C57BL/6 mice. Intranasally immunized BALB/c mice developed significantly higher titers than did any other group at all time points (P < 0.001). The anti-LPS responses followed a similar trend, with intranasally immunized mice developing significantly higher titers than orally immunized mice (P < 0.001) (Fig. (Fig.5C5C).
All mice were challenged intranasally with strain L82016. There was significant reduction in S. pneumoniae nasal colonization in the BALB/c mice immunized with χ9241(pYA4729) by both the intranasal (Fig. (Fig.66 A) and oral (Fig. (Fig.6B)6B) routes compared to that in the animals that received the control strain χ9241(pYA3342) (P < 0.02 for intranasal immunization and P < 0.05 for oral immunization). Similar results were obtained in C57BL/6 mice (P < 0.05 for both intranasal and oral immunizations).
Next, we evaluated protection against other S. pneumoniae serotypes elicited by immunization with strain χ9241(pYA4729). BALB/c mice were immunized intranasally or orally as in the previous experiments and challenged with serotype 23 S. pneumoniae strain E134. Mice immunized with χ9241(pYA4729) by either route showed a significant reduction in colonization by the challenge strain compared to that shown by mice inoculated with χ9241(pYA3342) (P < 0.02) (Fig. (Fig.6C6C).
We tested the efficacy of χ9241(pYA4729) immunization against nasal colonization by S. pneumoniae strains A66.1 (serotype 3) and D39 (serotype 2). However, after challenge, we did not detect any colonies in nasal washes from any group, including the controls (data not shown). We were, however, able to detect these strains in lung homogenates. Using this model, oral immunization of BALB/c mice with χ9241(pYA4729) did not provide protection against lung colonization by A66.1 and D39 (P > 0.05) (Fig. (Fig.6D6D).
Anti-PsaA mucosal IgA titers in nasal washes and lung homogenates were measured in mice after challenge with the different S. pneumoniae strains. In BALB/c mice challenged with either E134 or L82016, the intranasal immunization route elicited higher antibody titers than did oral immunization (Fig. (Fig.6E).6E). However, unlike the experiment shown in Fig. Fig.5,5, in this experiment, the two routes were not significantly different in C57BL/6 mice. This result is consistent with the vaginal-wash IgA titers (Fig. (Fig.5B).5B). The anti-PsaA IgA titers were significantly lower in the nasal washes than in the lung homogenates for mice challenged with strains A66.1 and D39 (Fig. (Fig.6F).6F). However, despite the relatively high anti-PsaA titers in the lung, no protection against these strains was observed (Fig. (Fig.6D6D).
Western blots were used to evaluate PsaA synthesis in the S. pneumoniae strains used in this study (see Fig. S2 in the supplemental material). Serum samples from C57BL/6 mice immunized intranasally with χ9241(pYA4359) or either intranasally or orally with χ9241(pYA4729) were diluted 1:1,000 for these experiments. Purified full-length rPsaA was used as the positive control. There are clear bands of ~35 kDa from SDS-PAGE, showing PsaA synthesis in all five S. pneumoniae strains, A66.1, D39, E134, L82016, and WU2 (see Fig. S2A in the supplemental material). Serum samples from immunized mice reacted with a single band of 35 kDa in all five lysates as well as the rPsaA positive control (see Fig. S2B in the supplemental material). The band density is weaker when serum samples from mice immunized with χ9241(pYA4359) were used as the primary antibody than when serum samples from mice immunized with χ9241(pYA4729) were used. The results are consistent with ours (Fig. (Fig.2)2) and those of Gor et al. (29).
Previous data have shown that PsaA on S. pneumoniae isolates is not readily accessible to antibody (29, 30, 37). We evaluated the accessibility of antibodies generated in mice immunized intranasally with χ9241(pYA4729) to surface PsaA on intact S. pneumoniae. Serum from mice inoculated with χ9241(pYA3342) did not bind to S. pneumoniae strains A66.1, D39, and WU2 but did bind to a very small fraction of E134 and L82016 (generally less than 0.1%) (see Fig. S3 in the supplemental material). Serum from mice inoculated with χ9241(pYA4729) did bind to strains D39, E134, and L82016, although at very low levels (<2% of cells bound antibody) (see Fig. S3), but did not bind to A66.1 and WU2. The relative inaccessibility of PsaA on S. pneumoniae cells that we observed correlates with previously published reports (29, 30).
PsaA has been studied extensively and investigated for its potential as a vaccine antigen (see references 35 and 64). PsaA protein, sometimes combined with other proteins, was used in most of the experiments, while delivery by live attenuated bacteria or viral vectors was seldom used (3, 47, 59, 60). In this work, we tested the ability of different PsaA constructs delivered by Salmonella vaccine strains to induce protective immunity. Previous work established that PsaA is an effective antigen to reduce nasal colonization by S. pneumoniae; however, few studies have shown that it can induce protection against intraperitoneal challenge, and one reported protection by intravenous challenge (29, 30, 82). We evaluated protection from intraperitoneal challenge with the virulent WU2 strain in mice immunized with our initial truncated PsaA constructions (see Table S1 in the supplemental material). These constructs failed to induce protective immunity, which is similar to the findings of Ogunniyi et al. (58) and Gor et al. (30). Compared to previously reported results using PspA as the antigen (42, 76), our intraperitoneal challenge results are disappointing, even when we immunized and boosted mice intranasally with a strain synthesizing full-length PsaA (see Table S1 in the supplemental material). One reason for these results may be the masking of PsaA by the cell capsule. Anti-PsaA antibodies cannot bind unless the capsule is removed (29, 37). S. pneumoniae has phase variations at a rate of about 10−3 to 10−6 between opaque, intermediate, and transparent phenotypes (87). Opaque cells produce as much as five times more capsular polysaccharide than transparent cells (15, 16, 39, 87), while transparent cells have greater adherence to cytokine-activated pneumocytes and vascular endothelial cells than do opaque cells (15, 16, 39, 87). Anti-PsaA antibody can bind to transparent cells but not to opaque cells. We found that in our hands, only ~1% of S. pneumoniae cells, at best, can directly bind anti-PsaA antibody (see Fig. S3 in the supplemental material) in spite of the fact that PsaA is abundantly synthesized by all S. pneumoniae strains tested (see Fig. S2A in the supplemental material), suggesting that the strains we used in the binding assay were highly encapsulated. The physiological state of the cell can also affect capsule synthesis. Bacteria obtained from log-phase cultures are typically highly encapsulated, and thus the surface-localized PsaA is not accessible to anti-PsaA antibodies, while bacteria obtained from stationary-phase culture are much less encapsulated and can be accessed by anti-PsaA antiserum (30). Thus, it is possible that changing the growth conditions and harvest time for our binding assay may have resulted in a greater number of cells bound by the anti-PsaA antibodies, at least for some strains. The character of the S. pneumoniae strain also affects the neutralization effect of antibody. WU2 is a highly encapsulated strain (11). All these factors may have influenced the binding of anti-PsaA antibody to the pneumococcal cells.
Another reason for the lack of protection against WU2 challenge is that the antibody titer against PsaA was not high enough to be effective (66). The highest reciprocal IgG antibody titer that we obtained after immunization with our first set of constructs was 210 (see Fig. S1 in the supplemental material). Although the titer is comparable to or higher than titers reported by others, it is lower than anti-PspA titers when PspA is delivered by Salmonella vectors, which are generally 212 to 215 (42, 76). Thus, there may not be enough antibodies to block the binding of S. pneumoniae to the nasopharyngeal cell surface. The situation was changed after we adopted the natural PsaA signal sequence, when we obtained anti-PsaA titers that were as high as 217 (Fig. (Fig.5).5). The antibody titer against PsaA may be increased further by including the sopB mutation in our Salmonella vector strain. Introduction of a ΔsopB mutation into attenuated Salmonella has been shown to enhance the immune response against vectored antigens (42, 46).
The structure of PsaA is likely to be critical for generating antibodies against conformational epitopes. Romero-Steiner et al. reported that the functional epitopes critical for anti-PsaA antibody binding are not continuous (69, 70). This possibility is consistent with the results reported by Giefing et al. (25). In that study, they used a genomic display library consisting of 15- to 150-aa peptides from S. pneumoniae to pan for conserved antigens that react with antibodies from volunteers previously exposed to S. pneumoniae. PsaA was not among the antigens identified in that screen, suggesting that the peptides used in the library were too short to contain conformational epitopes, although low anti-PsaA titers in the volunteers may have also played a role. The PsaA protein used in many previous immunogenicity studies was synthesized in E. coli as a fusion protein (2, 9, 74). These constructions may lead to proteins with conformations different from native PsaA that do not induce antibodies against conformational epitopes. We used the entire psaA gene in an effort to generate PsaA closer to the native structure to promote the formation of antibodies against conformational epitopes.
Compared with antiserum against PspA or capsule, the neutralization effects of anti-PsaA serum are greatly reduced against S. pneumoniae (29, 45). Since intraperitoneal challenge causes severe sepsis and death within a short time, normally around 3 days, the poor neutralizing ability of anti-PsaA could explain why PsaA has very limited protective function in this challenge model. Taken together, these results indicate that PsaA is not a good antigen to elicit protection against systemic infection due to masking by the pneumococcal capsule when the organism migrates outside the nasopharynx. In addition, it is possible that PsaA may not be important for virulence during sepsis.
Many researchers have shown that immunization with PsaA in mice can reduce nasal colonization by S. pneumoniae (27, 50, 51, 67, 78). When we challenged mice intranasally, our strains colonized to a level between 103 and 104 CFU/ml in nasal washes in unimmunized mice, similar to previously reported results (12, 91). Immunization with Salmonella synthesizing full-length PsaA resulted in significant protection against colonization with S. pneumoniae strains L82016 and E134. One reason for our success in reducing nasal colonization by strain L82016 may be related to the fact that 90% of this strain in nasal washes and the nasopharynx are in the transparent phase (12). In the transparent phase, bacteria have thick cell walls and sparse capsular polysaccharide, leading to more proficiency at initiating attachment, presumably due at least in part to the action of PsaA. With PsaA not masked with capsule, it is more accessible to blocking by specific antibodies. In contrast, our failure to protect against lung colonization by strains A66.1 and D39 in spite of fairly high anti-PsaA IgA titers may be due to the fact that transparent cells have no advantage over opaque cells for colonizing the lung (16). Salmonella induces an IL-17A response in infected C57BL/6 mice (73). Recent reports have indicated that induction of IL-17A plays a critical role in suppressing nasal carriage of S. pneumoniae, particularly in mice immunized intranasally (48). Therefore, it is likely that IL-17A played a role in the protection from nasal carriage that we observed with our vaccine. However, we note that it is also likely that PsaA-specific antibody was also required, since Salmonella alone or Salmonella expressing truncated antigens were not protective (see Fig. S1 and Table S1 in the supplemental material). In addition, there appears to be a direct correlation between protection and an antibody response (Fig. (Fig.3A3A and and4A).4A). Antigen-specific T cell is also important to clear bacteria (84). Clarifying and understanding the relationship between antibody, cytokines, and T-cell responses will undoubtedly lead to a more effective vaccine design.
Most clinical isolates of S. pneumoniae are a mixed population of opaque and transparent variants (44). Increased surface adherence correlated with the selection of transparent variants during carriage in an infant rat model of colonization (87). The transparent phenotype is predominant during natural carriage in humans (86). Recent results suggested that transparent-phase cells are highly adapted to various middle ear environments and that the transparent phenotype is the predominant phenotype responsible for the pathogenesis of pneumococcal otitis media (43, 44). Thus, anti-PsaA antibody could benefit in the prevention of otitis media caused by S. pneumoniae. Anti-PsaA antibody can reduce nasopharyngeal carriage of some serotypes and probably reduce the bacterial load in lungs, where S. pneumoniae infection can be further prevented by immunization with other antigens, such as PspA, PspC, or pneumolysin. In this work, we demonstrated that PsaA can be delivered by a Salmonella vaccine vector to elicit protective immunity. In future work, this antigen will be combined with other antigens to develop an effective multiantigen system delivered by Salmonella to prevent infection by S. pneumoniae.
The work was supported by NIH R01 AI24533, AI057885, and AI056289 and the Bill and Melinda Gates Foundation grant 37863.
We thank David Briles and Janice King (University of Alabama at Birmingham) for providing S. pneumoniae strains, anti-PsaA antiserum, and animal protocols and Erika Arch (Arizona State University) for her assistance with the manuscript. We also acknowledge the anonymous reviewers for their helpful suggestions that improved the quality of our manuscript.
Editor: J. N. Weiser
Published ahead of print on 17 May 2010.
†Supplemental material for this article may be found at http://iai.asm.org/.