Antibodies to the Iron Uptake ABC Transporter Lipoproteins PiaA and PiuA Promote Opsonophagocytosis of Streptococcus pneumoniae

Abstract

PiaA and PiuA are the lipoprotein components of the Pia and Piu Streptococcus pneumoniae iron uptake ABC transporters and are required for full virulence in mouse models of infection. Active or passive vaccination with recombinant PiuA and PiaA protects mice against invasive S. pneumoniae disease. In this study we have analyzed the antibody responses and mechanism of protection induced by PiuA and PiaA in more detail. For both proteins, two booster vaccinations induced stronger antibody responses in mice than a single or no booster vaccinations, and 5 μg of protein induced similar levels of antibody responses as 20 μg. Immunoglobulin G (IgG) subclass-specific enzyme-linked immunosorbent assays demonstrated that the antibody response to PiuA and PiaA was predominantly IgG1, with induction of only low levels of IgG2a. Anti-PiaA and anti-PiuA polyclonal rabbit antibodies bound to the surface of live S. pneumoniae when assessed by flow cytometry but did not inhibit growth of S. pneumoniae in cation-depleted medium or bacterial susceptibility to the iron-dependent antibiotic streptonigrin. However, anti-PiaA and anti-PiuA did increase complement-independent and -dependent opsonophagocytosis of different serotypes of S. pneumoniae by the human neutrophil cell line HL60. Hence, vaccination with PiaA and PiuA protects against S. pneumoniae infection by inducing antibodies that promote bacterial opsonophagocytosis rather than inhibiting iron transport.

In both developing and industrialized countries, the gram-positive pathogen Streptococcus pneumoniae is responsible for a significant burden of potentially preventable disease (2, 16). The currently available S. pneumoniae vaccines are based on capsular polysaccharide antigens, but as S. pneumoniae is capable of synthesizing at least 90 antigenically different capsular polysaccharides, a polyvalent vaccine containing many capsular antigens is needed for protection against the common serotypes. The 23-valent polysaccharide vaccine protects against 80% of clinically significant serotypes but is ineffective in children under the age of 2 and has reduced efficacy in the elderly, the two main risk groups for severe disease (15). Conjugated vaccines composed of 7 to 11 selected polysaccharides bound to nonpneumococcal protein carriers have been developed and protect children aged less than 2 years from invasive S. pneumoniae infections (3, 14). However, the conjugated vaccines protect against only a limited number of capsular serotypes, and their use results in nonvaccine serotypes becoming more prevalent (11). In addition, conjugated vaccines are expensive, limiting their availability in the developing world. Because of these drawbacks with capsular antigen vaccines, there has been much interest in developing a protein-based S. pneumoniae vaccine, using surface-expressed proteins which are well conserved across all serotypes (1, 8, 21, 22, 25, 28).

PiaA and PiuA are recently identified lipoprotein components of two separate S. pneumoniae iron uptake ABC transporters that are required for full virulence in mouse models of infection (4). Both PiuA and PiaA are likely to be expressed on the bacterial cell surface (24, 26), and active or passive vaccination of mice with recombinant PiuA and PiaA protects against invasive S. pneumoniae disease, especially when mice are vaccinated with both antigens (8). Mouse antibodies to recombinant PiuA and PiaA derived from a capsular serotype 2 S. pneumoniae strain react with identical proteins from nine other capsular serotypes (8), and human sera recovered from patients with recent episodes of S. pneumoniae septicemia due to different capsular serotypes contain anti-PiuA and anti-PiaA which cross-react with recombinant PiuA and PiaA from a single capsular serotype (27). Hence, PiuA and PiaA are potential candidates for an S. pneumoniae vaccine that may induce protection against many different S. pneumoniae capsular serotypes and overcome some of the limitations of the existing capsular antigen vaccines. However, ABC transporter lipoproteins are probably attached to the outer surface of the cell membrane (24, 26) and therefore separated by both the capsule and cell wall from the external environment, and recent data have shown that the surface accessibility to antibodies of another lipoprotein vaccine candidate, PsaA, was poor (12). It is therefore unclear whether antibodies to PiuA and PiaA can bind to the lipoprotein in live S. pneumoniae cells and assist opsonophagocytosis of the bacteria by host phagocytes. Potentially, antibodies to PiuA and PiaA could aid immunity by other mechanisms, such as inhibiting the function of these proteins and therefore reducing iron uptake by S. pneumoniae. As iron uptake is required for in vivo growth (4-6), this would result in delayed replication of S. pneumoniae and protection of the host against infection.

In the present study, we have characterized the antibody responses to different schedules of vaccination with these antigens in more detail. In addition, we have investigated the mechanism of protection provided by anti-PiuA and anti-PiaA, using growth in cation-depleted medium and streptonigrin sensitivity assays to assess their effect on iron transport and a flow cytometry assay to assess their effect on opsonophagocytosis of S. pneumoniae.

MATERIALS AND METHODS

Bacterial strains and media.

The capsular serotype 2 strain D39 and its derivatives were used for all experiments. D39 ΔpiuB and ΔpiaA strains were constructed by standard transformation techniques using DNA isolated from the previously described strain 0100993 capsular serotype 3 ΔpiuB and ΔpiaA strains and selection for mutant strains with appropriate antibiotics (4). Representative strains of common strain types (according to multilocus sequence typing) for the capsular serotypes 6A, 6B, 9V, 14, and 23F were kind gifts from Brian Spratt. S. pneumoniae was cultured in an atmosphere of 5% CO₂ and 95% air at 37°C on Columbia agar containing 5% horse blood and supplemented when necessary with erythromycin at 0.2 μg ml⁻¹ or chloramphenicol at 4 μg ml⁻¹ or in Todd-Hewitt broth supplemented with 0.5% yeast extract; when necessary, cations were depleted using Chelex 100 as previously described (4). Single-use aliquots of D39 for use in the experiments were prepared from cultures grown in Todd-Hewitt-yeast (THY) broth (optical density at 540 nm [OD₅₄₀], 0.2 to 0.4) and stored at −70°C.

Cloning, expression, and purification of recombinant PiuA and PiaA.

Recombinant PiuA and PiaA proteins (excluding the N-terminal lipoprotein peptidase signal sequence) were expressed and purified using the QIAexpress (QIAGEN) expression system. Fragments of piuA and piaA were amplified using high-fidelity PCR and S. pneumoniae D39 chromosomal DNA as the template with the oligonucleotide primer pairs PiaA forward (5′-CGGGATCCTTCTAATTCTGTTAAAAATGAA-3′)-PiaA reverse (5′-CGAGCTCGGCAGATTATAAAAATTATGC-3′) and PiuA forward (5′-CGGGATCCAACTCAAGCACTAGTCAGACA-3′) and PiuA reverse (5′-GAACTGCAGGAAGTGGCTTATTTCAAAG-3′). The restriction sites incorporated into the primers (underlined) were BamHI for the forward primers for both genes, SstI for the PiaA reverse primer, and PstI for the PiuA reverse primer. The products were ligated into the QIAexpress vector pQE31, transformed into Escherichia coli strain DH5α, and then transferred into the E. coli K-12 expression strain M15 (carrying a kanamycin resistance repressor plasmid pRep) by using standard protocols (8). In-frame fusion of the sequences of the piuA and piaA fragments with the N-terminal His₆ tag-encoding region of pQE31 was confirmed by automated dye-terminator sequencing (5′-GGCGTATCACGAGGCCCTTTCG-3′). Addition of isopropyl-d-thiogalactoside induced high-level expression of PiaA and PiuA, which were then purified using standard methods (8). The purified PiaA and PiuA proteins were resuspended in 50 mM sodium phosphate (pH 7.0), glycerol was added to a final concentration of 50%, and the proteins were stored at −20°C.

Immunization experiments.

Purified PiuA and PiaA proteins were used as antigens for the immunization experiments. The concentrations of these proteins were calculated using the Bradford reagent (Bio-Rad), and their purity was ascertained to be greater than 95% by sodium dodecyl sulfate-polyacrylamide gel electrophoresis after staining with Coomassie brilliant blue R250. Proteins were prepared for immunization at 50- or 100-μg ml¹⁻¹ concentrations in 10% alum adjuvant (Inject Alum no. 77161; Pierce, Chicago, Ill.). Groups of 5 to 10 female BALB/c mice (6 to 8 weeks old) were immunized by intraperitoneal (i.p.) inoculation of 100 to 200 μl of each protein preparation in alum adjuvant on days 0, 7, and 14 (8, 21). Sera were collected from five mice from each group by tail bleeding (5 μl per mouse) on days 0, 7, 14, 21, and 28 just prior to repeat vaccinations and were pooled. For the passive protection experiments, groups of 20 mice were immunized by i.p. injection of 100 μl of rabbit anti-PiaA and/or anti-PiuA sera at 0 h, followed by i.p. challenge at 12 h with 10⁶ CFU of D39 and a boost of 100 μl of rabbit anti-PiaA and/or anti-PiuA at 24 h. Anti-PiuA serum was diluted in phosphate-buffered saline (PBS) to have an antibody titer similar to that of anti-PiaA serum. The development of disease in mice was monitored as previously described (4, 8), and mice were sacrificed when they exhibited the following signs of severe disease: hunched posture, poor mobility, and piloerection.

ELISAs.

Specific antibody titers in pooled sera from five mice for each group were measured by enzyme-linked immunosorbent assays (ELISA) using 96-well polystyrene Maxisorb plates (Nunc) coated with purified antigens as described previously (21). Bound antibodies were detected by using alkaline phosphatase-conjugated goat anti-mouse immunoglobulin G (IgG), IgG1, or IgG2a (Bio-Rad) and disodium p-nitrophenyl phosphate as the substrate. The OD₄₀₅ was measured, and antibody titers were calculated as the lowest dilution giving an OD₄₀₅ equal to or greater than 0.30.

IgG binding assays.

Flow cytometry assays of IgG deposition on the surface of S. pneumoniae were performed using a previously described protocol (7). Briefly, bacterial stocks were thawed, washed by centrifugation at 20,000 × g for 6 min in PBS, incubated in 10 μl of mouse serum for 20 min, washed twice with 500 μl of PBS-0.1% Tween 20, and resuspended in 50 μl of PBS-0.1% Tween 20 containing a 1:100 dilution of phycoerythrin-conjugated goat anti-rabbit IgG (Sigma). After incubation on ice for 30 min, the bacteria were washed with 500 μl of PBS-0.1% Tween 20 and resuspended in 400 μl of PBS for flow cytometry analysis.

Opsonophagocytosis assays.

To assess the effect of anti-PiuA and anti-PiaA antibodies on the interaction of S. pneumoniae with phagocytes, we measured the proportion of a neutrophil cell line associated with fluorescent bacteria by using a previously described flow cytometry opsonophagocytosis assay (10, 17, 23). The complement source used was commercially available baby rabbit serum (Sigma S7764; rabbit HLA-ABC). S. pneumoniae strains were fluorescently labeled by incubation with 5,6-carboxyfluorescein-succinimidyl ester (FAM-SE; Molecular Probes, Eugene, Oreg.) solution (10 mg/ml in dimethyl sulfoxide; Sigma) in 0.1 M sodium bicarbonate buffer for 1 h at 37°C and then washed six times with Hanks balanced salt solution (HBSS) in 0.2% bovine serum albumin and stored in aliquots at −70°C in 10% glycerol (approximately 10⁹ CFU/ml). The human cell line HL-60 (promyelocytic leukemia cells; CCL240; American Type Culture Collection, Manassas, Va.) was used to provide the effector cells after differentiation into granulocytes by using previously described protocols (17). Differentiated HL60 cells were harvested by centrifugation (160 × g, 8 min, 4°C) and washed twice with HBSS and once with HBSS in the presence of Ca²⁺ and Mg²⁺. FAM-SE-labeled bacteria (10⁶ CFU) were opsonized with 1/100, 1/40, and 1/10 dilutions of serum in a 96-well plate for 20 min at 37°C with horizontal shaking (170 rpm). Negative controls were included, using the same volume of HBSS. HL60 cells (10⁵) were added to the opsonized bacteria in the microplate plate and incubated for 30 min at 37°C with shaking, after which the bacteria and cells were fixed using 3% paraformaldehyde and analyzed using flow cytometry. A minimum of 6,000 cells per sample were analyzed.

Rabbit polyclonal antibodies to PiaA or PiuA.

Rabbit polyclonal anti-PiaA and anti-PiuA were prepared by CovalAb (United Kingdom) by immunization of 6- to 8-week-old rabbits three times at three weekly intervals i.p. with either PiaA or PiuA and collecting sera by ear bleeds. A proportion of the antisera was used for affinity purification of anti-PiaA and anti-PiuA by using a protein A bead column, providing approximately 2 ml of affinity-purified antibody at 4 mg ml⁻¹.

Streptonigrin sensitivity assays.

Streptonigrin susceptibility assays were performed using bacteria grown in cation-depleted medium in the presence or absence of rabbit affinity-purified anti-PiuA or anti-PiaA (4, 5). One hundred microliters of the culture was removed, to which was added 2.5 μg ml¹⁻¹ of streptonigrin (Sigma), followed by incubation at 37°C for up to 80 min. Aliquots of the reaction cultures were diluted and plated at different time points and the CFU milliliter⁻¹ for each time point calculated. Each experiment used triplicate samples and was repeated three times. Results were expressed as a proportion of the starting CFU milliliter⁻¹ to allow comparison between experimental groups.

Statistical analysis.

Data presented are representative of data from experiments repeated two or three times. Data from the survival experiments were analyzed using log rank tests, and the results for all other comparisons were analyzed using two-tailed t tests.

RESULTS

Effects of booster immunization with PiuA and PiaA and IgG subclass responses.

To investigate the optimum immunization regimen, the effects of one or two booster vaccinations on the strength of mouse antibody responses to PiaA and PiuA were assessed. Groups of mice were given one, two, or three 10-μg doses of PiuA or PiaA by i.p. injection on days 0, 7, and 14, and the titers of total IgG to the antigens in sera obtained on days 7, 14, 21, and 28 immediately before booster vaccinations were measured using ELISAs. For both PiuA and PiaA, antibody titers after one dose of antigen were relatively weak, but they were substantially increased when mice were given one or two booster vaccinations (Table 1). Antibody titers were higher after two than after one booster dose. As previously reported, PiuA appears to be more immunogenic than PiaA, with maximum antibody titers of 1/12,800 compared to 1/3,200 (8).

TABLE 1.

Specific IgG titers in mouse sera over time in response to one, two, or three i.p. vaccinations with PiuA and PiaA

Antigen	No. of vaccinations	Titer after day since first vaccinationa:
		7	14	21	28
PiuA	1	1/400	1/800	1/400	NDb
	2	1/200	1/1,600	1/3,200	1/1,600
	3	1/200	1/1,600	1/12,800	1/12,800
PiaA	1	—c	—	1/100	ND
	2	—	1/200	1/400	1/400
	3	—	1/200	1/3,200	1/3,200

^aMice were vaccinated on day 0 and after sera were obtained on days 7 and 14.

^bND, not done.

^c—, no detectable response.

The combination of immunization with PiuA and PiaA improves protection against systemic S. pneumoniae infection (8); therefore, additional experiments investigated the effect of vaccination with varying and combined doses of PiuA and PiaA on the antibody titers. Different groups of animals were vaccinated with three doses of 10 or 20 μg of individual antigen or 5 or 10 μg each of PiuA and PiaA in combination, and the antibody responses were measured 14 days after the last vaccination. There were only small differences in the antibody titers to PiuA and PiaA in mice given 5, 10, or 20 μg of each antigen singularly or in combination (Fig. 1A and B), suggesting that vaccination with 5 μg of each antigen should be as effective as higher doses. The predominant IgG subclass responses after vaccination with PiuA and PiaA were assessed using ELISAs specific for mouse IgG1 and IgG2a. Both PiuA and PiaA elicited strong IgG1 subclass responses, with only weak IgG2a responses (Fig. 1C and D).

FIG. 1.

(A and B) Antibody titers to PiuA (A) and PiaA (B) in mouse sera after i.p. vaccination with 10 μg (diagonally hatched bars) or 20 μg (vertically hatched bars) of single antigen or with 5 μg plus 5 μg (horizontally hatched bars) or 10 μg plus 10 μg (open bars) of PiuA and PiaA in combination. (C and D) Serum IgG1 (C) and IgG2a (D) subclass antibody titers to PiuA and PiaA on day 7 (open bars, 7 days after the initial i.p. vaccination), 14 (diagonally hatched bars, 7 days after first booster vaccination), or 21 (vertically hatched bars, 7 days after the second booster i.p. vaccination). Solid bars represent results for sera from mice vaccinated with alum alone, and error bars represent standard deviations. Titers are presented as log₁₀ values of the reciprocal dilutions giving an OD₄₀₅ equal to or greater than 0.30.

Rabbit anti-PiuA and anti-PiaA protect mice against systemic S. pneumoniae infection.

In order to analyze the mechanism of protection induced by anti-PiaA and anti-PiuA antibodies, high-titer polyclonal rabbit anti-PiaA and anti-PiuA were obtained (titers of 1/8,000 and 1/32,000, respectively). Before assessment of the mechanisms by which anti-PiuA and anti-PiaA can prevent S. pneumoniae infection, whether polyclonal rabbit anti-PiaA and anti-PiuA can protect mice against S. pneumoniae infection was confirmed using passive immunization experiments. Groups of mice were given 100 μl of rabbit anti-PiuA (diluted in PBS to a titer of 1/8,000) and anti-PiaA sera by i.p. injection, followed 12 h later by i.p. challenge with 10⁶ CFU of S. pneumoniae D39 and a second i.p. vaccination with rabbit antisera at 24 h. In keeping with previous data using mouse antisera, passive immunization with rabbit anti-PiuA and anti-PiaA in combination significantly enhanced survival of mice (55%, compared to 10% for mice given rabbit preimmune sera; P < 0.01) (Fig. 2) (8). Passive immunization with the individual antisera resulted in smaller, nonsignificant improvements in survival (30% for PiuA and 20% for PiaA). These results demonstrate that rabbit anti-PiuA and anti-PiaA in combination can protect against S. pneumoniae infection.

FIG. 2.

Survival curves for groups of 20 mice vaccinated passively with rabbit polyclonal anti-PiuA and anti-PiaA sera at a titer of 1/8,000. Circles, preimmune serum; diamonds, anti-PiaA serum; squares, anti-PiuA serum; triangles, combination of anti-PiuA and anti-PiaA serum. For passive vaccination with anti-PiuA and anti-PiaA in combination versus preimmune serum, P < 0.01. The differences in survival between mice given either anti-PiuA or anti-PiaA and mice given preimmune serum or the combination of anti-PiuA and anti-PiaA were not significant.

IgG binding to S. pneumoniae.

Whether polyclonal rabbit anti-PiuA and anti-PiaA bound to the surface of live S. pneumoniae was measured using viable D39 cells and a flow cytometry assay. IgG was deposited on a high proportion of S. pneumoniae cells after incubation in either undiluted anti-PiuA or anti-PiaA serum, and to a lesser degree in diluted anti-PiuA or anti-PiaA serum (Fig. 3A). The level of IgG binding to PiaA was similar to that to PiuA despite the higher titer of antibody to PiuA, consistent with reverse transcription-PCR data which show that the piaA operon is more highly expressed than the piuA operon in laboratory medium and during infection in mice (5; J. S. Brown, unpublished data). To confirm that IgG binding to S. pneumoniae in rabbit anti-PiuA and anti-PiaA sera was antigen specific, IgG binding experiments were repeated using the ΔpiaA and ΔpiuB strains, in which, respectively, the expression of PiaA and PiuA is disrupted (4). Although there was some nonspecific IgG binding, the proportion of bacteria positive for IgG increased substantially when the ΔpiuB strain was incubated with anti-PiaA antibodies and when the ΔpiaA strain was incubated with anti-PiuA antibodies (Fig. 3B). These experiments demonstrate that anti-PiuA and anti-PiaA bind to the surface of live S. pneumoniae and that this binding is mainly antigen specific.

FIG. 3.

Anti-PiuA and anti-PiaA binding to S. pneumoniae D39 assessed by identifying bacteria coated with antibody by using phycoerythrin-conjugated goat anti-rabbit IgG and flow cytometry. (A) Proportion of wild-type bacteria positive for IgG after incubation in rabbit polyclonal anti-PiuA and anti-PiaA. Solid bars, PBS negative control; open bars, 1/20 dilution of antisera; diagonally hatched bars, 1/5 dilution of antisera; vertically hatched bars, undiluted antisera. For all antiserum dilutions versus PBS, P < 0.015. (B) Proportions of ΔpiuB, ΔpiaA, and ΔpiuB ΔpiaA bacteria positive for IgG after incubation in a 1/10 dilution of preimmune sera (solid bars), rabbit polyclonal anti-PiuA (open bars), and anti-PiaA (diagonally hatched bars). For the ΔpiaA strain, results for anti-PiuA compared to the results for anti-PiaA, P < 0.0001. For the ΔpiuB strain, results for anti-PiaA compared to the results for anti-PiuA, P < 0.0001. Error bars represent standard deviations.

Effects of anti-PiuA and anti-PiaA on iron transport.

Whether anti-PiuA and anti-PiaA can inhibit iron uptake by the Piu and Pia ABC transporters was investigated using growth of S. pneumoniae in cation-depleted medium and bacterial susceptibility to streptonigrin. Growth of the wild-type D39 S. pneumoniae strain in cation-depleted medium with or without addition of anti-PiuA and anti-PiaA was assessed by measuring the ODs of broth cultures over time. To avoid contamination with iron-containing compounds present in serum, affinity-purified anti-PiuA and anti-PiaA were used for these experiments. As previously demonstrated for a capsular serotype 3 S. pneumoniae strain, growth of a ΔpiuB ΔpiaA double mutant strain derived from D39 was substantially impaired in cation-depleted medium compared to that of the wild-type parental strain (Fig. 4). However, addition of the combination of anti-PiaA and anti-PiuA antibodies did not delay growth of the wild-type strain, indicating that these antibodies do not substantially inhibit the function of the Pia and Piu iron transporters. In addition, the presence of anti-PiaA and anti-PiuA did not inhibit growth of the ΔpiuB and ΔpiaA strains, respectively, in cation-depleted medium (data not shown).

FIG. 4.

Growth curves as measured by OD₅₄₀ of the S. pneumoniae D39 wild-type (diamonds) and ΔpiuB ΔpiaA (circles) strains in Chelex-THY in the absence (open symbols) or presence (closed symbols) of rabbit affinity-purified anti-PiaA and anti-PiuA (1/20 final concentration). For the differences in OD₅₄₀ between the D39 and ΔpiuB ΔpiaA strains at 10 h, P < 0.01.

To further investigate whether anti-PiaA and anti-PiuA may affect iron transport, the susceptibility of D39 strains to streptonigrin in the presence or absence of the antibodies was assessed. Streptonigrin is an antibiotic whose efficacy is dependent upon intracellular iron content, and it is a sensitive indirect indicator of low iron levels in S. pneumoniae (4). However, addition of anti-PiuA in combination with anti-PiaA antibody to D39, of anti-PiuA to the ΔpiaA strain, or of anti-PiaA to the ΔpiuB strain grown in cation-depleted medium did not affect S. pneumoniae susceptibility to streptonigrin (data not shown). Hence, data for both growth in cation-depleted medium and susceptibility to streptonigrin suggest that anti-PiuA and anti-PiaA do not inhibit Piu and Pia function, and these antibodies probably protect against S. pneumoniae infection by other mechanisms.

Anti-PiuA and anti-PiaA increase opsonophagocytosis of S. pneumoniae.

Anti-PiuA and anti-PiaA may assist immunity to S. pneumoniae by opsonizing bacteria and improving phagocytosis. To investigate this possibility, we used an in vitro flow cytometry assay to measure the association of S. pneumoniae with the human neutrophil cell line HL60 after incubation with 1/100, 1/40, and 1/10 dilutions of affinity-purified anti-PiuA and anti-PiaA. The role of complement was investigated by repeating the experiments with and without exogenous supplementation with rabbit complement. In the absence of complement there was a small increase in the association of D39 with HL60 cells in the presence of either anti-PiuA or anti-PiaA (Fig. 5A). Overall, the level of association of D39 with HL60 cells was considerably higher when rabbit complement was added to the assay, and this increase was more marked in the presence of anti-PiuA and anti-Pia (Fig. 5B). Further opsonophagocytosis assays with 1/10 dilutions of affinity-purified anti-PiuA and anti-PiaA were performed using representative strains of the common S. pneumoniae serotypes 6A, 6B, 9V, 14, and 23F. For all the serotypes investigated, in the presence of complement both anti-PiuA and anti-PiaA increased the association of bacteria with HL60 cells compared to HBSS alone (Fig. 5C). These results demonstrate that anti-PiuA and anti-PiaA assist opsonophagocytosis of D39 and a range of other S. pneumoniae capsular serotypes.

FIG. 5.

(A and B) Opsonophagocytosis measured by association of S. pneumoniae D39 with HL60 cells in different dilutions of affinity-purified rabbit anti-PiuA or anti-PiaA in the absence (A) or presence (B) of purified rabbit complement. Solid bars, HBSS; open bars, 1/100 dilution of serum; diagonally hatched bars, 1/40 dilution of serum; vertically hatched bars, 1/10 dilution of serum. For the 1/10 dilution of anti-PiuA compared to HBSS, P < 0.01 in either the presence or absence of complement; for the 1/10 dilution of anti-PiaA compared to HBSS, P < 0.005 in either the presence or absence of complement. For the 1/20 dilutions, P < 0.03 for both antisera in the presence or absence of complement. Error bars represent standard deviations. (C) Opsonophagocytosis measured by association with HL60 cells of different serotypes of S. pneumoniae in the presence of complement after incubation in HBSS (open bars), a 1/10 dilution of anti-PiuA (diagonally hatched bars), or a 1/10 dilution of anti-PiaA (vertically hatched bars). Anti-PiuA and anti-PiaA results that are significantly different from the results for HBSS for that strain are marked by asterisks.

DISCUSSION

Several surface-expressed S. pneumoniae proteins are under current investigation as alternative vaccine candidates that may overcome the limitations of vaccines based on capsular polysaccharide (25). The work presented in this paper further characterizes the antibody responses to two of these candidate proteins, the iron uptake ABC transporter lipoproteins PiuA and PiaA. Previously we have shown that active vaccination with either PiuA or PiaA improves protection against systemic S. pneumoniae infection and that vaccination with both antigens is highly protective (8). In addition, passive vaccination with polyclonal mouse anti-PiuA and anti-PiaA was protective against S. pneumoniae (8), and the results presented here confirm that this was also true for passive vaccination with rabbit anti-PiaA and anti-PiuA. To help optimize vaccination schedules for future studies, we have investigated the effects of one or two booster vaccination doses and different quantities of antigen on the strength of mouse antibody responses to PiaA and PiuA. The results indicate that three doses of vaccine induce the strongest response and that using 5 μg of each protein should be adequate to stimulate good antibody titers. As has previously been shown for other S. pneumoniae protein vaccine candidates (29), i.p. vaccination of BALB/c mice with PiaA and PiuA induces mainly IgG1 subclass antibodies, with only a small rise in IgG2a. The predominance of IgG1 suggests that in this model of vaccination there is mainly a Th2 immune response, contrasting with the Th1 response seen after DNA vaccination (19, 20).

The mechanism of protection afforded by many vaccine candidates remains ill-defined. Passive vaccination with antibodies to various S. pneumoniae protein antigens prevents infection, indicating that the protection provided by active vaccination is likely to be at least partially dependent on antibodies (8, 18, 21, 28). Antibodies to PiaA and PiuA are detectable in convalescent-phase sera from patients recovering from S. pneumoniae infections (27), suggesting that these antigens are immunogenic during natural infection. However, these antibodies could be generated by bacterial breakdown products rather than by responses to cell membrane antigens in live bacteria, and whether S. pneumoniae lipoprotein vaccine candidates such as PiuA and PiaA are accessible to antibodies in extracellular fluid or are shielded by the capsule and the cell wall was unclear. Using a flow cytometry assay, we have now demonstrated that live bacteria incubated in either anti-PiuA or anti-PiaA are coated in IgG. Moreover, using bacterial strains deficient in either antigen, we have shown that the IgG binding to S. pneumoniae was dependent on the presence of specific antigen, providing good evidence that these antigens are accessible to antibodies even in live bacteria. There was also a degree of nonspecific IgG binding, especially when the ΔpiaA strain was incubated in anti-PiaA serum. This may reflect nonspecific binding of IgG to the bacterial surface or, as previously has been demonstrated, cross-reactivity between anti-PiaA and PiuA (8). Higher levels of anti-PiaA and anti-PiuA binding to ΔpiuA and ΔpiaA strains than in wild-type bacteria could be due to increased expression of the remaining iron uptake systems in the mutant strains.

These results contrast with those of a recently published study in which anti-PsaA failed to detect significant levels of expression of the lipoprotein vaccine candidate PsaA on the surface of S. pneumoniae (12). Why there are differences between the accessibilities of PiuA, PiaA, and PsaA to antibody binding is unclear, but they could reflect differences in the capsular structure and relative expression levels of each protein. In addition, although lipoproteins are considered to be bound mainly to membrane, potentially a proportion of some lipoproteins may also be found within the cell wall or capsule, where there would be more likely to bind to antibody and aid immunity.

Iron is an essential cofactor for many pathogenic bacteria, but within the host, iron availability is greatly restricted. In common with many pathogens, S. pneumoniae has specialized mechanisms for iron uptake that are necessary for full virulence, including the Pia and Piu ABC transporters (4, 6). Binding of specific antibody to PiuA and PiaA could protect against infection by inhibiting Piu and Pia function, and this would readily explain the synergistic effect of dual vaccination with these antigens. However, using growth in cation-depleted medium and the highly sensitive streptonigrin susceptibility assays, we have been unable to demonstrate any inhibitory effect of anti-PiuA or anti-PiaA on iron uptake, even in mutant strains already deficient in either PiuA or PiaA. It is therefore unlikely that anti-PiuA or anti-PiaA protects against S. pneumoniae infection by preventing iron uptake by Piu or Pia.

Alternatively anti-PiuA and anti-PiaA may prevent infection by assisting host immunity. IgG serotype-specific antibodies to capsular polysaccharides are known to assist opsonophagocytosis of S. pneumoniae, as do antibodies to the cell surface-associated proteins PspA and PmpA (1, 17, 20, 23). We have adapted a flow cytometry assay of opsonophagocytosis widely used for assessing the functional effect of anticapsular antibody in human sera to investigate the effect of rabbit anti-PiuA and anti-PiaA on S. pneumoniae opsonophagocytosis (17). Both affinity-purified rabbit anti-PiuA and anti-PiaA increased opsonophagocytosis of the S. pneumoniae serotype 2 strain in either the presence or absence of complement. In addition, anti-PiaA and anti-PiuA improved oposonophagoctyosis of representative strains of the common capsular serotypes 6A, 6B, 9V, 14, and 23, demonstrating that these antibodies are likely to provide protection against many of the clinically important S. pneumoniae serotypes. In general, anti-PiaA stimulated opsonophagocytosis to a greater degree than anti-PiuA, possibly because PiaA is expressed to a higher degree than PiuA (5). Hence, antibodies to membrane-anchored proteins such as PiuA and PiaA and potentially other vaccine candidates located in the same cellular compartment (9, 19, 21) can be effective opsonins and promote humoral immunity, and this is likely to be one mechanism by which anti-PiuA and anti-PiaA protect against S. pneumoniae infection. Recent data suggest that cellular immune responses are also important for immunity to S. pneumoniae (13). Whether immunization with PiuA and PiaA, in addition to stimulating humoral immunity, can also improve cellular immune responses to S. pneumoniae infection requires further investigation.

This study provides further data on the antibody response induced by vaccination with PiuA and PiaA and the mechanism by which these proteins induce protection against i.p. infection with S. pneumoniae. Future research is required to demonstrate whether vaccination with PiaA and PiuA can also protect against S. pneumoniae after intranasal inoculation, an infection model that leads to pneumonia and represents the natural route of infection in humans. In addition, combinations of PiaA and/or PiuA and other existing protein vaccine candidates need to be investigated to identify the most effective combination vaccine for preventing S. pneumoniae infections (21).