Intranasal immunization with heterologously expressed polysaccharide protects against multiple Pseudomonas aeruginosa infections

Abstract

Surface-expressed bacterial polysaccharides are often immunodominant, protective antigens. However, these antigens are chemically and serologically highly heterogeneous, and conjugation to protein carriers is often necessary to enhance their immunogenicity. Here we show the efficacy of intranasal immunization of mice with attenuated Salmonella enterica serovar Typhimurium expressing the O antigen portion of Pseudomonas aeruginosa lipopolysaccharide. P. aeruginosa is an ideal model system because it can cause a myriad of localized and systemic infections. In particular, this bacterium is a leading cause of hospital-acquired pneumonia and is responsible for infections after burns and after eye injury. In addition, there are mouse models of infection that mimic the clinical manifestations of P. aeruginosa infections. Immunized mice were highly protected against infection, with long-lasting immunity to acute P. aeruginosa pneumonia, whereas mice immunized with Salmonella containing only the cloning vector or PBS were not. Prophylactic and therapeutic administration of sera from vaccinated animals protected naive mice. Intranasal vaccination also provided complete protection from infections after burns and reduced pathology after corneal abrasions. These results indicate that intranasal delivery of heterologously expressed polysaccharide antigens provides protection at distinct sites of infection. This approach for the expression and delivery of polysaccharide antigens as recombinant immunogens could be easily adapted to develop vaccines for many infectious agents, without the need for complicated purification and conjugation procedures.

Bacterial surface polysaccharides make effective components for vaccines against serious infections because they are generally immunodominant. However, their utility is limited by significant chemical, and hence serologic, variability that requires a large number of components for comprehensive coverage against pathogens capable of expressing different polysaccharide serotypes. Also, purified polysaccharides are often poorly immunogenic, necessitating synthesis of protein–polysaccharide conjugate vaccines. These problems could potentially be overcome by use of recombinant DNA to induce synthesis of the polysaccharides in a heterologous antigen-delivery system. We evaluated this potential by expressing the Pseudomonas aeruginosa LPS O antigen in attenuated Salmonella enterica serovar Typhimurium SL3261 and used this construct to immunize mice intranasally (IN) to evaluate protective immunity against P. aeruginosa infection in the lung and eye, as well as systemic infection emanating from burn wounds.

The model polysaccharide that we have used for this study is the P. aeruginosa serogroup O11 O antigen. This serogroup is one of the most prevalent (1) and has been associated with acute infections (2), burns (3), and ulcerative keratitis (4). In addition, P. aeruginosa serogroup O11 strains are among the most pathogenic and lethal because of the expression of the cytotoxic phospholipase ExoU (2). When the P. aeruginosa serogroup O11 O antigen is expressed in Escherichia coli from the cosmid pLPS2, it is attached to the E. coli lipid A-core (5); when Salmonella contains pLPS2, P. aeruginosa serogroup O11 expression does not exclude expression of Salmonella O antigen (6).

We previously reported that oral or i.p. administration of a live, attenuated Salmonella vaccine strain expressing the P. aeruginosa serogroup O11 O antigen via the plasmid pLPS2 induced a robust P. aeruginosa O antigen-specific serum immune response after vaccination (7). Because P. aeruginosa is associated with respiratory tract infections, we evaluated vaccine efficacy by IN challenge of mice to produce pneumonia. Oral, but not i.p., immunization enhanced pulmonary bacterial clearance and increased the time to death with P. aeruginosa after IN challenge, but no protection against mortality was achieved. Lung lavage fluid from orally vaccinated animals had detectable levels of P. aeruginosa-specific IgA and IgG, with a much lower P. aeruginosa-specific IgG level detected in these fluids of i.p. vaccinated mice with no evidence of local IgA. These findings suggest that protection correlates with the presence of P. aeruginosa-specific O antigen antibodies in the lung (7).

We postulated that a potent local mucosal immune response would enhance protection if effective immunity could be induced via vaccination by the IN route. Here we report that IN immunization elicited robust P. aeruginosa-specific antibody production in serum and in the upper and lower respiratory tracts. IN vaccination was sufficient to provide complete protection to P. aeruginosa infection in an acute respiratory challenge model for up to 6 months after vaccination. In addition, this vaccination regimen elicited less pathology from a P. aeruginosa corneal infection and complete protection after infection of a burn wound indicating protection at distinct sites of infection. This strategy for heterologous expression of polysaccharide antigens in an attenuated Salmonella strain and delivery by the IN route could be applicable to the development of vaccines to other infectious agents.

Results

Active IN Immunization Provides High-Level Protection to Acute P. aeruginosa Pneumonia.

BALB/c mice were IN vaccinated with S. enterica serovar Typhimurium SL3261 expressing P. aeruginosa serogroup O11 antigen and challenged with the serogroup O11 strain 9882-80 at six or 12 times the LD₅₀ (Fig. 1 A and B). Complete protection against the 6-LD₅₀ challenge dose and 72% protection against the 12-LD₅₀ challenge dose was achieved and no signs of morbidity were observed in surviving mice that received the vaccine. All but one of the mice given PBS or immunized with the vector control succumbed to infection. These findings indicate that the IN vaccination provides increased protection to acute pneumonia caused by P. aeruginosa strain 9882-80 as compared with only an extended time to death that we previously observed with oral immunization (7).

Fig. 1.

IN vaccination promotes robust protection to P. aeruginosa acute pneumonia that is long-lasting. BALB/c mice were challenged with 9882-80 (≈10⁸ cfu) (log-rank: PBS vs. vaccine, P = 0.0003; vector vs. vaccine, P = 0.0008; PBS vs. vector, P = 0.6722; median survival: PBS, 42 h; vector, 42 h; vaccine, undefined) (A), with 9882-80 (2.4 × 10⁸ cfu) (log-rank: PBS vs. vaccine, P < 0.0001; vector vs. vaccine, P < 0.0001; PBS vs. vector, P = 0.2373; median survival: PBS, 28.5 h; vector, 26.5 h; vaccine, undefined) (B), with strain PA103 (≈3.4 × 10⁵ cfu) (log-rank: PBS vs. vaccine, P = 0.0003; vector vs. vaccine, P = 0.0004; PBS vs. vector, P = 0.1202; median survival: PBS, 30.5 h; vector, 40 h; vaccine, undefined) (C), and after 6 months with PA103 (≈1–2 × 10⁶) (log-rank: PBS vs. vaccine, P = 0.0005; vector vs. vaccine, P = 0.0004; PBS vs. vector, P = 0.3226; median survival: PBS, 24 h; vector, 24 h; vaccine, undefined) (D) and carefully monitored for survival. PBS- and vector-immunized mice served as the controls in all experiments. Results are represented in Kaplan–Meier survival curves and analyzed by the log-rank test. n refers to the number of animals.

We next examined whether the vaccine could provide protection to serogroup O11 strain PA103, which is ≈200 times more virulent than strain 9882-80 (data not shown). Vaccinated mice were completely protected from infection with strain PA103 at three times the LD₅₀ and showed no signs of morbidity throughout the experiment (Fig. 1C). Interestingly, immunogenicity and protection against lethal infection were not dependent on the mouse strain [see supporting information (SI) Table 1].

We next tested whether the vaccine could confer long-term protective immunity in immunized animals. Mice were infected with PA103 at >10 times the LD₅₀ 6 months after a second immunization. Vaccine-immunized mice were completely protected from infection and showed no signs of morbidity, indicating that the recombinant vaccine is capable of promoting long-lasting immunity against acute pneumonia caused by strain PA103 (Fig. 1D). Subsequent studies using a single vaccination of 10⁴ cfu also showed complete survival to infection with strain PA103 (data not shown).

Antibody Response to IN Vaccination.

The titer of antibodies in sera from PBS-, vector-, or vaccine-immunized animals was determined by ELISA on plates coated with whole cells of strain PA103 (Fig. 2). Sera isolated from BALB/c mice that received the vaccine exhibited robust P. aeruginosa-specific IgG and IgA antibody titers when compared with sera from the vector-immunized or PBS control mice (Fig. 2A). Moreover, pooled sera from IN vaccinated animals were reactive to four additional serogroup O11 strains, suggesting that IN delivery of the vaccine produced a broad serogroup O11 response (data not shown). No difference in reactivity to whole cells of the Salmonella vector was noted between vector- and vaccine-immunized animals (data not shown).

Fig. 2.

Anti-P. aeruginosa serogroup O11 serum and mucosal antibody response after IN vaccination of BALB/c mice. (A) PA103-specific total IgG and IgA reactivity from serum. Box and whisker plots represent endpoint titers in log₁₀ as determined by linear regression. The box is marked by the median and depicts the first and third quartiles, and the whiskers extend to the range (Mann–Whitney U test: P < 0.0001 for IgG and IgA for vaccine vs. vector and vaccine vs. PBS). (B) PA103-specific antibody reactivity from BAL of IN vaccinated mice. (C and D) PA103-specific antibody reactivity observed from NW before (C) and after (D) infection with PA103. Error bars represent the mean and SEM.

Identification of P. aeruginosa-Specific Antibodies in Mucosal Secretions.

We next examined the relative level of antibodies present in mucosal lavage fluid from vaccinated animals. Bronchoalveolar lavage (BAL) fluid from IN vaccine-immunized BALB/c mice contained high levels of P. aeruginosa-specific total IgG and IgA in the lower respiratory tract with the level of IgA reactivity exceeding IgG by nearly 10-fold (Fig. 2B). To identify the presence of P. aeruginosa-specific antibodies in the upper respiratory tract of IN vaccinated mice, nasal washes (NW) were performed. Interestingly, only Pseudomonas-specific IgA could be detected in this location before P. aeruginosa infection with no evidence of Pseudomonas-specific IgG reactivity above background levels (Fig. 2C). In addition, P. aeruginosa-specific IgG and IgA could be detected in the saliva of IN vaccinated animals (data not shown).

During an active infection, the ensuing inflammatory response is likely to cause the respiratory epithelium to become leaky, allowing transudation of serum components, such as antibodies onto the mucosal surface (8). We postulated that P. aeruginosa-specific Ig, besides IgA, might gain access to the nasal cavity during an active infection in vaccine-immunized mice that could aid in clearance. To assess this, vaccine-immunized animals were infected and NW fluid was obtained 6 h after PA103 infection and analyzed for antibody reactivity; at that time, antigen-specific IgG reactivity was detected in the NW fluid (Fig. 2D).

Opsonophagocytic Killing of P. aeruginosa Serogroup O11 Strains Is Mediated by Antisera and BAL Fluid from Immunized Mice.

The production of O-antigen-specific antibodies that can mediate bacterial uptake by phagocytes is correlated with the protective efficacy of P. aeruginosa LPS vaccines. By using antisera or BAL fluid from IN vaccinated vector and vaccine animals, opsonic killing assays were performed. Pooled antisera from vaccinated mice mediated killing at levels approaching 96% in a 1:100 dilution to all serogroup O11 strains tested (see SI Fig. 6). No killing was observed in the absence of complement (data not shown). Significant killing (>60%) was seen to each serogroup O11 strain at a 1:400 antiserum dilution, suggesting the vaccine induced a potent P. aeruginosa-serogroup O11-specific antibody response. Killing was not observed to a serogroup heterologous organism (strain 6294, serogroup O6) (see SI Fig. 6). In addition, killing was not observed to any P. aeruginosa organisms when using antisera obtained from animals immunized with the vector (<5% killing) (data not shown). Efficient killing of PA103 (56–86%) was also mediated by various dilutions of BAL fluid isolated from vaccine-immunized animals (see SI Fig. 6), but there was no killing with BAL fluid isolated from vector-immunized mice when tested at the same dilutions (data not shown).

Passive Antisera Transfer from Immunized Animals Provides Protection to Acute P. aeruginosa Pneumonia in Naive Mice.

In an initial experiment, antisera from PBS-, vector-, or vaccine-immunized mice were transferred IN to naive BALB/c mice at the same time that they were infected with strain PA103. Vaccine antisera completely protected mice from challenge with no signs of morbidity (Fig. 3A). Transfer of diluted antisera (1:10) produced similar results (Fig. 3B). Mice that received vector antisera were not protected from subsequent P. aeruginosa challenge.

Fig. 3.

Acute P. aeruginosa pneumonia survival studies in BALB/c mice using passively IN transferred antisera. (A) Passive transfer of whole antisera immediately followed by challenge with PA103 (≈3.45 × 10⁵ cfu) (log-rank: PBS vs. vaccine, P = 0.0001; vector vs. vaccine, P < 0.0001; PBS vs. vector, P = 0.0037). (B) Passive transfer of diluted (1:10) antisera immediately followed by PA103 challenge (≈3.4 × 10⁵ cfu) (log-rank: PBS vs. vaccine, P = 0.0031; vector vs. vaccine, P = 0.0016; PBS vs. vector, P = 0.2207). (C) Passive transfer of whole antisera 6 h after challenge with PA103 (≈3.0 × 10⁵ cfu) (log-rank: PBS vs. vaccine, P = 0.0047; vector vs. vaccine, P < 0.0001; PBS vs. vector, P = 0.2054). (D) Passive transfer of whole antisera immediately followed by 6294 challenge (7.0 × 10⁷ cfu) (log-rank: PBS vs. vaccine, P = 0.6627; vector vs. vaccine, P = 0.9305; PBS vs. vector, P = 0.6585). Results are represented in Kaplan–Meier survival curves and were analyzed by the log-rank test.

To identify whether mice could be therapeutically rescued from the consequences of pneumonia initiated by infection with strain PA103, undiluted antisera was transferred to mice 6 h after infection. Seventy-five percent of mice receiving vaccine antisera survived infection, whereas nearly all animals receiving either PBS or vector antisera succumbed to the challenge (Fig. 3C). We then determined whether passive transfer of vaccine antisera to mice could provide protection against infection with LPS heterologous strain 6294 (serogroup O6); no difference in survival was noted among the mice, indicating that protection correlates to the presence of serogroup O11-specific antibodies (Fig. 3D).

To determine further that the protective serum component was antigen-specific, diluted vaccine antisera (1:10) were adsorbed before transfer to naive animals with either strain PA103 or strain PA103galU (9), an O-antigen-deficient mutant of strain PA103. Complete protection was only seen in infected mice receiving antisera adsorbed with strain PA103galU (see SI Fig. 7). ELISA analysis for IgG antibodies in the vaccine antisera adsorbed with strain PA103 revealed a drastic reduction in P. aeruginosa-specific IgG reactivity when compared with antisera adsorbed with the PA103galU strain, suggesting that depletion of serogroup O11 O-antigen-specific antibodies were responsible for the lack of protection in this model (see SI Fig. 7).

Prevention of Dissemination and Pneumonia in Vaccine-Immunized Mice.

Bacterial dissemination and clearance were monitored in naive, vector-, and vaccine-immunized BALB/c mice by using the IVIS Imaging System (Xenogen, Alameda, CA) and the findings confirmed by determination of viable P. aeruginosa cfu in tissue homogenates. For these experiments, we constructed a luminescent strain of PA103 (PA103.lux) and determined that the detection limit of PA103.lux in the IVIS Imaging system was ≈10⁴ organisms (data not shown). Integration of the lux gene locus into the PA103 genome had no affect on the virulence of the strain (data not shown). PA103.lux infected BALB/c mice were initially imaged at 6 h, which did not reveal luminescence in any tissues. In the nasopharynx, the lack of lux signal may be due to the fact that the majority of bacteria were translocated to the lung within minutes after IN application (10); the lack of signal from the lungs at this time may be due to an initial decline in the number of bacteria after IN administration, as we have observed previously for some strains (unpublished results). However, strong signals could be detected in the upper respiratory tract and lungs of both naive and vector-immunized mice at 12 and 24 h after infection (Fig. 4 A and B). In addition, signals were detected from the gastrointestinal tract, which is likely because of partial swallowing of the luminescent bacteria (Fig. 4 A and B). Interestingly, we did not detect any evidence of luminescence in vaccine-immunized mice at either 12 or 24 h after infection, suggesting that these animals were able to keep the bacterial numbers below the level of detection of the photometric camera (Fig. 4C). NWs from naive and vector-immunized animals confirmed the high level of PA103.lux colonization of the upper respiratory tract at both 12 and 24 h, whereas viable cfu in NW from vaccine-immunized animals were minimal at 12 and absent at 24 h (Fig. 4D). Imaged animals were killed, and organs were harvested to confirm the level of PA103.lux infection. Consistent with the IVIS images, vaccine-inoculated animals had fewer detectable bacteria in their lungs at 12 h after infection with no signs of dissemination to the spleen or liver (see SI Fig. 8). By 24 h, no evidence of PA103.lux cells were detected in any of the tissues of vaccinated mice, indicating that IN vaccine immunization promoted sterile immunity (see SI Fig. 8). Similar results were obtained when using strain 9882-80 as the infecting organism (data not shown).

Fig. 4.

Real-time monitoring of PA103.lux in the acute pneumonia model. Representative BALB/c mice from naive (A), vector-immunized (B), and vaccine-immunized (C) animals were used to follow in vivo IN infection using the Xenogen IVIS CCD camera at 12 and 24 h after infection with PA103.lux (≈5.0 × 10⁵ cfu). The animals were imaged ventrally while under isoflourane anesthesia. (A–C) Luminescence is observed emanating from the nasopharynx, lungs, liver, and gastrointestinal area. The color bar indicates the intensity of the luminescent signal, with red and blue serving as the high and low signals, respectively. (D) NW fluid isolated from IVIS imaged animals at 12 and 24 h confirms the luminescent signal in the nasopharynx.

We next examined the histopathology of vaccinated animals infected with strain PA103 at 24 h after infection as an independent measure to determine whether vaccine-immunized then infected animals were protected from pneumonia. Low magnification images of lungs from PBS- and vector-immunized animals demonstrated pneumonic consolidation with filing of alveolar spaces with neutrophils that was more evident at higher magnification (see SI Fig. 9). In contrast, lungs from animals that received the vaccine show characteristic peribronchial and perivascular lymphoid infiltrates but a complete absence of alveolar cellular infiltration, indicative of bacterial clearance, prevention of pneumonia, and efficient protective immunity (see SI Fig. 9). The lungs of vaccinated mice after infection resembled those of vaccinated mice before infection (see SI Fig. 10).

IN Immunization Promotes Protection Against Corneal Infection with P. aeruginosa.

A P. aeruginosa corneal infection model has been previously established by using C3H/HeN mice (11). These mice were vaccinated IN with the Salmonella vector or serogroup O11-expressing strain, and the serum antibody response was determined by ELISA (see SI Table 1). After infection, mice receiving vaccine showed significantly less pathology than was observed in animals receiving the PBS and vector control (Fig. 5A). The number of total bacteria present in the cornea was then determined. Mice receiving the vaccine had fewer numbers of bacteria in their corneas compared with the controls (Fig. 5B). Invasion of corneal epithelial cells by P. aeruginosa strain PA103 was also determined: mice receiving the vaccine had fewer numbers of bacteria within corneal epithelial cells whereas higher bacterial internalization was observed in the PBS and vector controls (Fig. 5B).

Fig. 5.

Protection of vaccinated mice from P. aeruginosa infection after corneal damage or burns. (A) Corneal pathology scores were obtained from PBS-, vector-, and vaccine-immunized C3H/HeN mice 48 h after PA103 infection (10⁶ cfu) (PBS vs. vaccine, P = 0.0003; vector vs. vaccine, P = 0.0047). (B) The number of PA103 organisms able to adhere to or invade mouse corneas after infection of vaccinated animals (PBS vs. vaccine, P < 0.0001; vector vs. vaccine, P = 0.0005). Each point represents one eye from each animal and is marked by the median. Statistical analyses were performed by the Mann–Whitney U test. (C) Survival analysis from PBS-, vector-, and vaccine-immunized NSA-1 mice in a burn model after 6073 infection (4.4 × 10⁶ cfu) (log-rank: PBS vs. vaccine, P = 0.0024; vector vs. vaccine, P < 0.0001; vector vs. PBS, P = 0.0027). Results are represented in Kaplan–Meier survival curves and were analyzed by the log-rank test. n refers to the number of animals.

IN Immunization Induces Complete Protection in a P. aeruginosa Burn Infection Model.

A P. aeruginosa burn infection model has been established by using NSA-1 mice (12, 13). Therefore, we immunized these animals to test vaccine efficacy in this model. After an immune response was generated (see SI Table 1), vaccinated NSA-1 mice were burned, as previously described (12, 13), followed by a s.c. injection of the serogroup O11 strain 6073, which we had determined to have the lowest LD₅₀ among the serogroup O11 strains tested in naive mice using this model (data not shown). The animals were monitored for survival over 5 days. We observed complete protection in mice immunized with the vaccine (Fig. 5C).

Discussion

Although it is well established that optimal immunity to acute P. aeruginosa infection is mediated by antibodies to the O antigen of the LPS (14), it has been technically and conceptually challenging to produce an immunogenic and protective vaccine based on this principal. Reasons include the serologic diversity of these antigens, the poor immunogenicity of the purified O antigens, and the need to conjugate isolated O antigens to protein carriers for maximal effects. Another limitation previously encountered when using purified P. aeruginosa O antigen polysaccharides is that the protective immunity elicited by an antigen isolated from a single strain protected mice only against infection with the strain from which the polysaccharide was isolated and not against infection with other strains within the serogroup (15). Here we show that IN immunization with live, attenuated S. enterica serovar Typhimurium heterologously expressing the P. aeruginosa LPS serogroup O11 O antigen elicited opsonic antibody in serum and BAL that protected against infections with multiple serogroup O11 strains in the lung, cornea, and burn-wound infection models. This is a very promising preclinical result supporting the potential efficacy of this approach, especially because serogroup O11 is one of the 10 most clinically important P. aeruginosa O antigens (14). The further development of a multivalent vaccine composed of multiple recombinant attenuated S. enterica strains, each expressing the other prevalent P. aeruginosa LPS serogroup antigens, can likely be constructed to completely cover the most commonly encountered clinical isolates of P. aeruginosa. Furthermore, we have shown that this vaccine provides protective immunity from acute pneumonia using five different strains of mice (see SI Table 1), indicating that there is not a strong genetic component modulating development of protective immunity using this vaccination protocol.

A vaccine composed of eight P. aeruginosa O antigens conjugated to P. aeruginosa exotoxin A as a carrier has been injected intramuscularly into cystic fibrosis patients lacking detectable P. aeruginosa infection. This vaccine was found to elicit good serum immune responses (16, 17) and reduced the proportion of patients infected with P. aeruginosa, as well as delayed the time to infection (18). This same group is developing passive reagents (19); however, a clinical trial using i.v. hyperimmune globulin elicited with this same vaccine was stopped because of adverse effects and no evidence of efficacy (20). In mouse experiments, either intramuscular or IN administration of monovalent O antigen conjugates that were included in the octovalent vaccine was found to elicit an IgG response in the lung but only in the presence of adjuvant; only the IN immunization with adjuvant elicited an IgA response at that site. Mice immunized IN with adjuvant showed some protection after respiratory challenge or infection after burns (21). Here we show that delivery of live, attenuated Salmonella expressing P. aeruginosa O antigen IN induced both local and systemic immunity to P. aeruginosa infection, indicating its potential to provide optimal protective efficacy against the multiple manifestations of P. aeruginosa infection. This route of immunization with the attenuated Salmonella strain did not cause disease but did elicit an inflammatory response that was evident upon histopathological analysis. Importantly, however, IN immunization with the O antigen-expressing vaccine resulted in protection from acute pneumonia, which is in contrast to our previous findings wherein oral immunization with this same vaccine produced only an increased time to death (7).

Importantly, we also found that antisera obtained from IN immunized mice could be delivered therapeutically during an established infection and mediate significant protection. The ability to rescue mice 6 h after IN infection using the sera from vaccine-immunized mice, suggests that this approach could be used to induce both active immunity or obtain serum Ig for production of a passive therapeutic reagent.

Mechanistically, protective immunity in the lung was associated with the rapid appearance of local IgG in BAL fluid after infection, while protection against corneal and burn-wound infection was likely due to the systemic response elicited by IN vaccination. Antibodies to P. aeruginosa LPS have previously been shown to protect against pathology associated with corneal infection by reducing the numbers of organisms in the cornea (22). Although invasion of corneal cells by the bacteria is essential for pathogenesis (23), most of the bacterial cells are extracellular and susceptible to killing by phagocytes and antibodies. Thus, the ability of vaccine-induced antibody to protect the cornea against pathology is likely dependent on its ability to transudate onto the cornea along with PMN and complement to promote phagocytic killing of extracellular P. aeruginosa and reduce the overall bacterial numbers. In the burn-wound infection model, mice that are burned and challenged with P. aeruginosa die of systemic infection (12). However, at this time we cannot delineate whether mice immunized with our vaccine were protected at the site of infection or from the dissemination of bacteria. Further studies using local application or systemic administration of vaccine antisera are needed using both these infection models to identify the mechanisms of protection.

IN vaccination elicited distinct antibody profiles in the upper and lower respiratory tracts. P. aeruginosa-specific IgA was only detected in the upper respiratory tract whereas IgG and IgA were present in the lungs of IN vaccinated animals before infection. However, IgG and IgA were observed in the upper respiratory tract of vaccine-immunized animals at 6 h after infection. Likely local IgG was the major mediator of immunity, because no killing of P. aeruginosa in opsonophagocytosis assays using NW fluid containing only IgA was observed (data not shown), whereas antisera or BAL fluid containing high titers of IgG promoted efficient killing. The source of IgG in the respiratory tract after IN immunization has not yet been determined; however, we suspect that local lymphoid tissues may be a source for some of the protective antibodies, rather than transudation from the circulation alone. This interpretation is based on our previous results in which we noted that i.p. immunization with Salmonella expressing P. aeruginosa serogroup O11 promoted a potent serum antibody response but did not mediate protection from IN challenge with the P. aeruginosa serogroup O11 strain 9882-80 (7).

IN vaccination with the S. enterica serovar Typhimurium serogroup O11 construct also induced high-level protection against infection with P. aeruginosa serogroup O11 strains in a scratched cornea infection model and in burned mice. Therefore, the vaccine construct and IN delivery was highly effective at inducing immunity to multiple manifestations of distinct mucosal and systemic P. aeruginosa infections. We have already expressed the P. aeruginosa serogroup O11 in the approved vaccine strain, Salmonella typhi Ty21a (5), suggesting the easy translation of this technology to human infections. A multivalent vaccine to the other prevalent P. aeruginosa serogroups can likely be constructed to protect against all those commonly encountered in infection. Immunizing at-risk populations IN would be potentially a highly useful strategy for tackling the vexing problem of serious P. aeruginosa infection, which is more and more frequently being caused by organisms resistant to virtually all effective antibiotics. In addition, it is likely that this same Salmonella expression and delivery system could be applied to express other polysaccharides vaccine antigens to induce protective immunity to a variety of pathogens.

Materials and Methods

Mice.

Six-week-old BALB/c, C57BL/6, B6129SF2/J, and C3H/HeN mice and 10-week-old non-Swiss albino (NSA-1) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) or Harlan Sprague–Dawley Farms (Chicago, IL). All mice were housed under pathogen-free conditions and fed autoclaved rodent feed and acid-free water.

Preparation of Bacterial Strains for Vaccination.

Vector and vaccine strains (7) were inoculated in selective media (10 μg/ml tetracycline) for overnight growth, subcultured, and grown to an OD₆₅₀ of 0.5. Cells were harvested by centrifugation and washed twice in PBS. Before administration to animals bacterial cells were adjusted spectrophotometrically to obtain the desired inoculate and numerated to ensure viable cell counts.

IN Vaccination of Mice.

This procedure was approved by the University of Virginia Animal Care and Use Committee. Before vaccination, mice were anesthetized by i.p. injection of 0.2 ml of ketamine (6.7 mg/ml) and xylazine (1.3 mg/ml) in 0.9% NaCl. On day 0, 10 μl of PBS or bacterial inoculum was placed onto each nostril (total of 20 μl per mouse) followed by a booster at day 14. All mice were immunized twice unless otherwise noted.

Murine Acute Pneumonia Model.

This procedure was approved by the University of Virginia Animal Care and Use Committee. Guidelines for humane endpoints were strictly followed. P. aeruginosa strains 9882-80 (O11), PA103 (O11), and 6294 (O6) were used in the infection studies. For active and passive immunization survival analysis, strains were grown on tryptic soy agar for a maximum of 12 h at 37°C and suspended in PBS to an OD₆₅₀ of 0.5, which was previously determined to contain ≈10⁹ cfu/ml. Inoculums were adjusted spectrophotometrically to obtain the desired challenge dose in either 20 μl (active study) or 10 μl (passive study). Infections were performed on anesthetized animals by instilling the inoculums directly onto each nostril. For passive immunization studies, antisera from PBS-, vector-, or vaccine-immunized animals were delivered to anesthetized animals (5 μl per nostril) immediately before or 6 h after infection. All animals were carefully observed for up to 1 week. For most infections, mice were infected 2–3 weeks after a second immunization. For long-term survival studies, animals were infected 6 months after the second immunization.

Murine Corneal Infection Model.

This procedure was approved by the Harvard Medical School Animal Care and Use Committee. Immunized C3H/HeN animals were anesthetized followed by initiation of three 1-mm scratches on the cornea and superficial stroma of one eye of each mouse using a 27-gauge needle under a dissection microscope, as previously described (11). Mice were infected with PA103 (10⁶ cfu) by placing the bacteria directly on the eye (5 μl). Forty-eight hours after infection, corneal pathology scores were determined (11). At that time, mice were killed by CO₂ asphyxiation and the corneas were removed from the mice with a sterile scalpel and microdissecting scissors. Corneas were immediately placed in 1.0 ml of DMEM supplemented with 10% FBS and vortexed for 1 min to remove adherent bacteria followed by enumeration by serial dilution and plating on tryptic soy agar. Determination of intracellular bacteria was performed by exposure of corneas to 200 μg/ml gentamicin in DMEM plus 10% FBS for 1 h followed by extensive washing. Corneas were homogenized in 0.5% Triton X-100 to release intracellular bacteria followed by enumeration by serial dilution and plating on tryptic soy agar.

Murine Burn Infection Model.

This procedure was approved by the University of Cincinnati Animal Care and Use Committee. Hair from the backs of isoflourane-anesthetized immunized NSA-1 mice was clipped followed by the placement of a flame-resistant card against the shaved area. The exposed back was covered with 0.5 ml of ethanol, ignited, and allowed to burn for 10 seconds, as previously described (12, 13). This procedure produces a nonlethal 15% total-body surface area third-degree burn. Mice were immediately resuscitated with 0.5 ml of saline and infected s.c. beneath the burn with P. aeruginosa serogroup O11 strain 6073 (≈4.0 × 10⁷ cfu). Animals were carefully monitored for survival for 5 days.

Collection of Serum, BAL, and NW Fluid.

Serum and BAL fluids samples were collected from animals as described previously (7). NW fluid was collected from the nose after the cannula was placed into the nasopharyngeal duct followed by instillation of 0.5 ml of PBS-B.

Histological Analysis of Lungs After IN Vaccination and Infection.

Lungs of vaccinated BALB/c mice were isolated 28 days after a single vaccination before or 24 h after infection with strain PA103 (≈5 × 10⁵ cfu). Aseptically, the tracheas of each animal were exposed and profused with 0.3 ml of 4% paraformaldehyde and immediately placed in 1.0 ml of 4% paraformaldehyde. Paraffin-embedding, sectioning, and hematoxylin and eosin staining were performed by the University of Virginia Research Histology Core Facility.

ELISA and Opsonophagocytosis Assays.

Assays were performed as described previously (7). For ELISA, total IgG and IgA serum titers were adjusted based on OD₄₀₅ values of preimmunized samples and quantified by linear regression using the x-intercept as the endpoint titer. P. aeruginosa-specific total IgG and IgA from diluted pooled mucosal fluids were corrected by subtraction of the OD₄₀₅ values of vector-immunized samples to account for nonprotective antibody binding to vector antigens.

Construction of PA103.lux.

Primers LacZPfw 5′-(XhoI)GCGGCCGCCAGTCGACTGCAGCTCCAGCTTTTGTTCCCTTT-3′ and LacZRev 5′-(PstI)-TTCATTAATGCTCGAGGATCCACAGGTTTCCCGACTGGAAAG-3′ were used to amplify the lac promoter from pBluescript (Stratagene, Cedar Creek, TX), followed by gel purification and cloning into pCR-2.1-TOPO (Invitrogen, Carlsbad, CA). This 256-bp XhoI-PstI fragment, consisting of the lac promoter, was excised and cloned into the XhoI-PstI sites of the promoterless miniCTX-lux plasmid (24), generating a lac-lux fusion. The ensuing clone was transferred to PA103 by biparental mating, followed by selection of plasmid integrants on Pseudomonas isolation agar (Difco, Detroit, MI) containing tetracycline (100 μg/ml). PA103 plasmid integrants were then subjected to FLP-mediated excision by mobilization with plasmid pFLP2 (24). lac-lux insertion at the attB locus of PA103 was confirmed by colony PCR using lac primers in conjunction with primers (24) specific to the adjacent tRNA^Ser gene.

Tissue Colonization and in Vivo Imaging.

P. aeruginosa colonization in naive and vaccinated animals was analyzed after IN infection. Lung, spleen, and liver tissues were aseptically removed, homogenized in 1 ml of PBS-B, and plated for viable cfu on tryptic soy agar at various time points after infection as a function of tissue weight. To follow infection in vivo, animals were infected with PA103.lux and images were acquired with the In Vivo Luminescence and Fluorescence Imaging (IVIS) CCD camera system and analyzed with Living Image 2.11 software (Xenogen).

Statistical Analyses.

Statistics were performed by using Prism version 4 (GraphPad, San Diego, CA). Antibody titers and viable cfu from tissue counts were compared by using the Kruskal–Wallis U test or the Mann–Whitney U test. For survival studies, data were analyzed by the log-rank test.

Abbreviations

BAL

bronchoalveolar lavage

IN

intranasal

NW

nasal wash.