|Home | About | Journals | Submit | Contact Us | Français|
Pseudomonas aeruginosa is a serious pathogen in hospitalized, immunocompromised, and cystic fibrosis (CF) patients. P. aeruginosa is motile via a single polar flagellum made of polymerized flagellin proteins differentiated into two major serotypes: a and b. Antibodies to flagella delay onset of infection in CF patients, but whether immunity to polymeric flagella and that to monomeric flagellin are comparable has not been addressed, nor has the question of whether such antibodies might negatively impact Toll-like receptor 5 (TLR5) activation, an important component of innate immunity to P. aeruginosa. We compared immunization with flagella and that with flagellin for in vitro effects on motility, opsonic killing, and protective efficacy using a mouse pneumonia model. Antibodies to flagella were superior to antibodies to flagellin at inhibiting motility, promoting opsonic killing, and mediating protection against P. aeruginosa pneumonia in mice. Protection against the flagellar type strains PAK and PA01 was maximal, but it was only marginal against motile clinical isolates from flagellum-immunized CF patients who nonetheless became colonized with P. aeruginosa. Purified flagellin was a more potent activator of TLR5 than were flagella and also elicited higher TLR5-neutralizing antibodies than did immunization with flagella. Antibody to type a but not type b flagella or flagellin inhibited TLR5 activation by whole bacterial cells. Overall, intact flagella appear to be superior for generating immunity to P. aeruginosa, and flagellin monomers might induce antibodies capable of neutralizing innate immunity due to TLR5 activation, but solid immunity to P. aeruginosa based on flagellar antigens may require additional components beyond type a and type b proteins from prototype strains.
Pseudomonas aeruginosa is an opportunistic pathogen responsible for a large proportion of ventilator-associated, hospital acquired pneumonia and is also a major cause of morbidity and mortality in cystic fibrosis (CF) patients. P. aeruginosa is motile via a single polar flagellum that has the added structural feature of being glycosylated (39). Flagellin is the primary protein component of the flagellar filament, and it can be classified into two serotypes, types a and b. Flagella carry out many functions, such as motility and attachment of bacteria to host cells, and can also elicit the activation of the host inflammatory response via Toll-like receptor 5 (TLR5) (6, 15, 29, 31). Importantly, promising results in terms of prevention of the acquisition of P. aeruginosa infection in CF patients immunized with a bivalent type a and b flagellum vaccine have been published (12).
Several animal studies have not only demonstrated the importance of flagella as a virulence factor in P. aeruginosa but also validated them, or their flagellin component, as target antigens for vaccination. In the burned-mouse model of infection, chemically mutagenized or genetically produced flagellum-negative strains were less virulent than flagellum-positive strains (5, 26). It has also been shown with this model that motility is necessary for dissemination from the site of infection, since an intact flagellum structure is essential for death due to sepsis (5). In a neonatal model of acute P. aeruginosa pulmonary infection, flagella were essential for full virulence (14), although this was not found to be the case for adult mice with pulmonary P. aeruginosa infection (6). In regard to protection mediated by flagella or flagellin, immunization with flagella provided protection against infection and decreased the spread to major organs in the burned-mouse model (19). In a rat model of P. aeruginosa-induced pneumonia, administration of human monoclonal antibodies (MAbs) to flagella provided protection against infection and decreased lung injury (24), and another set of human MAbs provided protection in a murine model of pneumonia during neutropenia (28). A DNA vaccine encoding recombinant type a or type b P. aeruginosa flagellin also induced protective immunity against lethal P. aeruginosa lung infection (33), although this study curiously found better heterologous protection than homologous protection when DNA encoding wild-type flagellin was incorporated into the vaccine. A fusion protein of outer membrane protein F (OprF) residues 311 to 341, mature OprI residues 21 to 83, and flagellins a and b (termed OprF311-341-OprI-flagellins) generated significant immune responses in mice and promoted enhanced clearance of strain PA01 in a pulmonary challenge model (42). None of these studies directly compared the vaccine potential of flagellin with that of flagella.
In addition to being highly immunogenic, the flagellin component of flagella serves as a pathogen-associated molecular pattern (PAMP), activating TLR5 and inducing innate immunity in the lung, stimulating a protective inflammatory response that contributes to the eradication of the pathogen (15, 32, 33, 35). Instillation of recombinant flagellin into the lungs of mice elicits a significant induction of innate immunity (20), and application of flagellin to the cornea of mice or intraperitoneal (i.p.) injection prior to corneal injury and local P. aeruginosa infection protects against pathological destruction of this tissue (22, 23). Finally, overexpression of flagellin monomers enhances virulence of P. aeruginosa (6).
Of great interest is that the TLR5-binding domain of flagellin is not exposed in the intact flagella (36), and thus, flagellin monomers must be released or extracted from the intact flagella to promote TLR5 activation. Therefore, the comparative TLR5 agonist activity of flagellin, flagella, and even intact P. aeruginosa bacteria has not been evaluated, nor is it clear if the TLR5 activation component of flagellin would be immunogenic when immunizing with the intact polymeric flagella.
Since P. aeruginosa serotype a and b flagella are conserved, contribute to virulence, stimulate innate immunity, and have induced protective efficacy in both animal (19, 24) and human (12) vaccine studies, it is clear that the flagellum or the flagellin monomer may be a useful target as a vaccine component, particularly as a carrier protein to link to protective carbohydrate antigens such as lipopolysaccharide (LPS) O-side chains or the alginate capsule (11, 30, 37). To our knowledge, no comparative analysis of the vaccine efficacy of flagellin versus that of flagella has been described for P. aeruginosa or other pathogens. Thus, it is not clear if it is flagellum or flagellin that is the best vaccine candidate, if either or both could be effectively utilized as a component of a conjugate vaccine, and if use of these vaccines could induce a state of enhanced susceptibility to infection by blocking flagellin-TLR5 interactions that promote effective innate immunity, as was found with antibodies induced by a DNA vaccine encoding P. aeruginosa flagellin (33). The purpose of this study was to compare whether immunity to P. aeruginosa flagella and that to flagellin are comparable or distinct and to evaluate if antibodies neutralizing TLR5 activation are induced and whether this impacted TLR5 activation by flagellin, flagella, or intact P. aeruginosa cells.
The P. aeruginosa strains used for these studies were as follows: PAK, a serogroup O6, type a flagellated strain; PA01, a serogroup O2/O5, type b flagellated strain; PAKΔfliC, a fliC deletion strain of P. aeruginosa strain PAK (9); and PA01ΔfliC, a gentamicin insertion mutant carrying the mutation in the fliC gene of P. aeruginosa strain PA01 (17); PA01ExoU+, a PA01 strain expressing the ExoU cytotoxin (1); and clinical isolates CF6, -12, -19, -20, -38, and -42, obtained from CF patients who participated in a flagellar vaccine clinical trial (12) and, despite having been immunized with flagella, nonetheless were infected with flagellum-positive strains. The flagellar mutant strains were provided by Reuben Ramphal (University of Florida).
C57BL/6 mice were obtained from Charles River Laboratories. New Zealand White rabbits were from Millbrook Breeding Labs, Amherst, MA. All animal studies were conducted in accordance with protocols approved by the Harvard Medical Area Institutional Animal Care and Use Committee.
Flagella were purified from P. aeruginosa strains as previously described, with some modifications (25, 40). Briefly, bacteria were grown statically overnight in tryptic soy broth (TSB), harvested by centrifugation, and resuspended in cold phosphate-buffered saline (PBS). Flagella were removed from the cells by shearing in a cold Waring blender for 35 s. The cells were separated from the flagella by centrifugation at 16,000 × g for 15 min. The supernatant thus obtained was then ultracentrifuged at 40,000 × g for 3 h. Flagella, obtained as a pellet, were then suspended in a minimum amount of PBS and filtered through a 0.45-μm-pore-size Millipore filter. Flagellar preparations were also passed through a polymyxin B column to remove lipopolysaccharide (LPS) contaminants and analyzed by a Limulus amebocyte lysate assay with Escherichia coli O113:H10 endotoxin as the standard (Cape Cod Associates) and with P. aeruginosa LPS used as a comparator. No endotoxin was detected at a level of >0.01% in the flagellar preparations. Flagella purified by this method, similar to that used in the vaccine trial of CF patients (12), also contain the FliD cap protein and components of the flagellar basal apparatus (4).
Recombinant flagellins were purified from E. coli BL21(DE3) carrying the pET15BVP vector with the His-tagged type a or b fliC gene as previously described (39).
Female New Zealand White rabbits were immunized subcutaneously (s.c.) with 100 μg of flagellin or 10 μg of flagella suspended in incomplete Freund's adjuvant (Sigma) on days 1 and 8. The rabbits were boosted intravenously with three 100- or 10-μg doses the following week. Further booster doses of 100 or 10 μg were given intravenously at intervals of 2 to 4 months.
Enzyme-linked immunosorbent assay (ELISAs) were performed by standard methods as described previously (30). In brief, microtiter plates were coated with flagellin or flagella (0.2 M carbonate buffer, pH 9.6) and kept overnight at 4°C. Between incubation steps, plates were washed three times with PBS containing 0.05% Tween 20 (PBS-Tw). Blocking was performed with 1% bovine serum albumin (BSA) in PBS overnight at 4°C. An alkaline phosphatase conjugate goat anti-rabbit IgG antibody diluted 1:1,000 was used as a secondary antibody, and p-nitrophenyl phosphate was used as a substrate (1 mg/ml in diethanolamine buffer, 0.5 mM MgCl2 [pH 9.8]; Sigma). After 60 min of incubation at 37°C, the absorbance was measured at 405 nm. ELISA titers were calculated by linear regression analysis of the average of duplicate measurements; the titer was the serum dilution giving a final optical density (OD) value of 0 at 405 nm as calculated from the linear regression curve.
An opsonophagocytic assay was used as previously described (2, 18) with the following modifications: to prepare bacteria for use in the assay, 3 ml instead of 10 ml of TSB was inoculated with bacteria from a tryptic soy agar (TSA) plate grown overnight at 37°C, with the inoculum placed in the TSB tubes, and bacteria were grown statically at 37°C until an OD at 650 nm (OD650) of 0.2 was obtained; RPMI medium was used with 10% heat-inactivated fetal bovine serum and 25 mM HEPES as the diluent; and infant rabbit serum adsorbed with the target strain for 1 h at 4°C was used as a complement source instead of human serum. The opsonic activity of immune sera was compared to that of sera obtained before vaccination. Negative controls included tubes from which polymorphonuclear leukocytes, complement, or serum was omitted. After the 90 min of incubation, a 50-μl portion was removed and diluted in TSB containing 0.05% Tween 20 as previously described (2). The opsonic activity of the serum was calculated as follows: [1 − (CFU immune serum at 90 min/CFU of preimmune serum at 90 min)] × 100 (37). To determine the amount of serum needed to kill 50% of the target bacteria (EC50), dilutions were log transformed and logistic regression analysis used to estimate the EC50 and 95% confidence interval (CI) using the PRISM 4 software package.
All studies involving human participants were approved by the Partners Health Care Institutional Review Board, and all participants gave written consent.
Motility assays were performed as previously described (7, 21), with some modifications. Bacteria were grown in TSB (with 100 μg gentamicin/ml for fliC mutants) statically at 37°C. Approximately 105 log-phase organisms were inoculated onto plates made with lysogeny broth (LB) and 0.3% agar in the presence or absence of various dilutions of antisera raised to flagellin or flagella and incubated at 30°C for 18 h. Results were photographically recorded after 18 h.
Flagellar typing of CF strains was accomplished by PCR amplification of the central region of the flagellin gene using primers specific for N-terminal (CW46 [5′-GGCCTGCAGATCNCCAA-3′]) and C-terminal (CW45 [5′-GGCAGCTGGTTNGCCTG-3′]) conserved regions as previously described (8, 43). The genomic DNA was isolated by using the Wizard genomic DNA purification kit (Promega) according to the manufacturer's instructions.
Female C57BL/6 mice, 6 to 8 weeks old, were given 200 μl of antibodies raised to either flagellin or flagella or given 200 μl of normal rabbit serum (NRS) i.p. 48, 24, and 4 h prior to infection.
Female C57BL/6 mice, 6 to 8 weeks old, were immunized intranasally (i.n.) as described previously (11). Briefly, prior to vaccination, mice were anesthetized by i.p. injection of 0.2 ml of xylazine (1.3 mg/ml) and ketamine (6.7 mg/ml) in sterile water. For active vaccination, mice were given 2 μg of either flagellum type a or b, 2 μg of BSA, or 0.2 μg of LPS i.n. in a 12.5-μl volume (6.25 μl per nostril) on days 1, 8, and 15. Mice were infected on day 40 ± 2.
For bacterial challenge, P. aeruginosa strains were grown on TSA plates overnight, bacteria from these plates were inoculated into TSB (with 400 μg of carbenicillin/ml for the PA01ExoU+ strain) at an OD650 of 0.1 and grown statically to an OD650 of 0.2 at 37°C. Bacteria were recovered by centrifugation, resuspended in PBS to the desired dose for infection, and washed three times in this buffer. Prior to administration to animals, bacterial cells were plated on MacConkey agar plates to determine the concentration. For infection, 25 μl of P. aeruginosa, prepared as described above, was slowly pipetted onto the nares of anesthetized mice, and animals were observed for survival twice a day for up to 5 days (1, 34).
Mouse sera were collected as described previously with some modifications (10). Briefly, blood samples were collected from the tail vein of each mouse after warming with a heat lamp and making a small nick with a sterile scalpel. Serum was separated from cells by centrifugation at 1,700 × g for 10 min, and the sera were stored at −20°C until use.
A TLR5-expressing A549 lung epithelial cell line stably transfected with a nuclear factor kappa B (NF-κB) luciferase reporter plasmid (NFκB-luc; Panomics, Freemont, CA) was used to detect cellular activation by P. aeruginosa flagellin, flagella, or intact bacterial cells. These cells are human lung adenocarcinoma cells expressing a luciferase-linked reporter regulated by multiple copies of the NF-kB response element. The cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum and selected by 100 μg hygromycin B/ml in T75 flasks. After ~90% of confluence was reached, A549/NFκB-luc cells were transferred to 96-well solid white plates (Costar) at a concentration of 5 × 104 cells/well. Antisera were diluted 1:50 and then added along with flagellin, flagella or P. aeruginosa live cells to the plates, which were then incubated for 5 h. After 5 h, a Steady-Glo luciferase reagent (Promega) was added, and the resultant luminescence was read in a luminometer after 10 min of incubation at room temperature. Lactate dehydrogenase (LDH) release and cell viability were analyzed using the LDH-based in vitro toxicology assay kit (Sigma-Aldrich).
Survival data for the different mouse groups were analyzed by using Kaplan-Meier survival curves and the log rank test.
After four rabbits were immunized with type a or b flagella or flagellin, antibody titers to those proteins were evaluated by ELISA. All four proteins were highly immunogenic (Table (Table1)1) and showed strong cross-reactivity between homologous flagella and flagellin proteins. The antisera also cross-reacted with the heterologous proteins but to a lower level than with the homologous proteins. Antibody to type a flagella or flagellin reacted better with type b proteins than did antisera raised to the type b proteins with the type a flagella or flagellin.
To determine the functional activity of the antisera to P. aeruginosa flagellin or flagella, we evaluated the ability of the antisera to inhibit motility of P. aeruginosa strains PAK and PA01. In these assays, NRS was used as a negative control. Antisera to flagellin at a dilution of 1:75 or less inhibited the motility of the corresponding type of P. aeruginosa. Antisera to flagella had a higher titer of motility-inhibiting antibodies, showing full inhibition of motility at a dilution of 1:200. The antiserum to flagellum type a showed some cross-reactivity to flagellum type b since it inhibited the motility of strain PA01 (Fig. (Fig.1),1), consistent with the titer determinations (Table (Table1).1). The six clinical isolates from CF patients were also tested in this assay, and the flagellar type identified by PCR was confirmed by showing inhibition of motility by the corresponding antiserum at a dilution of 1:50 (data not shown).
In an opsonophagocytic killing assay using antibodies raised to either flagellin or flagella, we determined the overall activity and estimated the serum dilution mediating killing of 50% of the bacterial cells (EC50) (Table (Table22 and Fig. Fig.2).2). Antisera to flagellin had little to no opsonic killing activity, with the serum raised to type a flagellin having low opsonic killing activity (EC50 = 10) using homologous strain PAK and no killing of strain PAO1 (Fig. (Fig.2A),2A), while an antiserum to type b flagellin had no killing activity (EC50 < 4) against homologous strain PA01 or heterologous strain PAK (Fig. (Fig.2B).2B). Antiserum to polymeric type a flagella was quite active, with an EC50 of 937 against homologous strain PAK (Fig. (Fig.2C),2C), while antisera to type b flagella had a modest EC50 of 19 against homologous strain PA01 (Fig. (Fig.2D).2D). After flagellar typing by PCR of P. aeruginosa CF isolates and ascertainment of functionality in motility assays (not shown), we randomly chose three strains of each type (for type a, CF 6, 38, and 42; for type b, CF 12, 19, and 20) to evaluate the opsonic activities of the antibodies raised to flagella. Antisera to type a flagella were highly active in mediating opsonic killing of the three type a CF clinical isolates, although the estimated EC50s were lower than those for activity against the homologous PAK strain (Table (Table2).2). Antisera to type a flagella also had modest killing activity against the three type b clinical isolates, comparable to that against the type b strain PAO1. The antiserum to type b flagella did not mediate killing of the type a clinical isolates (EC50 < 4) but had modest to high activity against the flagellar type b clinical isolates (Table (Table22).
Rabbit antibody raised to either type a or type b flagellin or flagella was used to passively immunize C57BL/6 mice i.p. After serum injections, mice were challenged i.n. with a P. aeruginosa flagellar type a or b strain at approximately two times the 50% lethal dose (LD50). Antibodies raised to either type of monomeric flagellin showed no protection against strain PAK or PA01ExoU+ (Fig. 3A and B) or against the clinical isolate CF6 (data not shown). In contrast, antibody to polymeric type a flagella achieved 87.5% survival following challenge with strain PAK (P < 0.0001 versus results for the NRS group) and 43.75% survival using strain CF6 (P < 0.0028 versus results for the NRS group). Antibody to the heterologous type b flagella did not protect against the type a strains (Fig. 3C and E), indicating specificity of this antiserum for the type b flagellum immunogen prepared from P. aeruginosa cells. Immunization with antibody to polymeric type b flagella resulted in 70.8% survival of strain PA01ExoU+ (P < 0.0001 versus results for the NRS group) and 12.5% survival of strain CF20 (P = 0.025 versus results for the NRS group). The antiserum to the type a flagella again exhibited some cross-reactivity with type b strains, providing 50% survival of mice challenged with strain PA01ExoU+ (P < 0.0005 versus results for the NRS group) and also resulted in an increase in the mean time to death to 36 h in of mice vaccinated with the anti-flagellum type a serum and infected with type b strain CF20. This increase was approximately 1.5 (95% CI, 1.12 to 1.88) times longer than that for mice vaccinated with NRS (P = 0.009) (Fig. 3D and F). However, antiserum to type a flagella did not improve the overall survival of mice infected with the type b clinical isolate CF20, again indicating that there was a high degree of specificity of the antibody for engendering survival against strains expressing the type a flagella isolated from P. aeruginosa cells.
Since passive immunization with antisera to flagellin was poorly opsonic and not protective against pneumonia, we focused the active vaccination studies on the protective activity of polymeric flagella. Mice were vaccinated i.n. with homologous or heterologous type a or type b flagella, BSA, or homologous LPS, which was used at a dose 1/10 that of the flagella to control for possible effects from potentially contaminating, albeit undetectable, LPS in the vaccines. After vaccination, mice were challenged i.n. 40 to 42 days later with strain PAK or PA01ExoU+ (Fig. 4A and B). Before infection, blood samples were collected from each mouse to evaluate the antibody response to the flagellar antigens. Both flagella were highly immunogenic by the i.n. route, as determined by finding high serum IgG antibody titers in an ELISA (data not shown). After infection with strain PAK, 83.3% of the mice immunized with homologous type a flagella survived, compared with 25% of the mice in the group immunized with BSA (P = 0.0016) and 8% survival in mice immunized with a low dose of LPS from strain PAK (P < 0.0001). There were no survivors in the group of mice immunized with heterologous type b flagella (P < 0.0001). Immunization with type b flagella was highly protective against infection with the homologous type b flagellum strain PA01ExoU+, with a survival rate of 83.3%, whereas in the group of mice immunized with BSA only, 8% of mice survived (P < 0.0001) and there were no survivors in mice immunized with LPS from strain PA01 or type a flagella (P < 0.0001, both comparisons). Immunization with type a flagella also protected against infection with type a strain CF6, with a survival rate of 41.67%, compared to 8.33% for the group immunized with BSA (P = 0.030), 16.67% for immunization with LPS (P = 0.09), and no survivors in the group of mice immunized with type b flagella (P < 0.0007) (Fig. (Fig.4C).4C). Immunization with type b flagella showed borderline protection against type b strain CF20, with a survival rate of 25%, while in the group of mice immunized with BSA, only 8.33% of mice survived (P = 0.09) and there were no survivors among the mice immunized with homologous LPS or type a flagella (P = 0.013) (Fig. (Fig.4D).4D). These results indicate high-level, flagellar-type-specific protection against pneumonia and death following i.n infection with strains PAK and PA01ExoU+ but clearly less protection when the challenge was with clinical isolates obtained from CF patients that had been immunized with a flagellar vaccine but nonetheless became colonized with P. aeruginosa. Since neither LPS nor immunization with the heterologous flagella provided protection, the protective immune responses of the mice were clearly specific to the immunizing antigen and not due to potential cross-reactive antibodies elicited by contaminating antigens.
We next evaluated whether antibody to flagella and/or flagellin inhibited TLR5 activation. In these assays, purified flagellin, as expected, was a more potent activator of TLR5 than flagella (Fig. (Fig.5),5), which are well known to release flagellin monomers during storage (6), making it difficult to determine if the activation was due to intact flagella or released flagellin. When comparing activation mediated by type a versus type b flagellin, there were no differences in the degree of luminescence, but polymeric type a flagella activated the cells to a higher level than polymeric type b flagella, particularly at concentrations lower than 1 μg/ml (Fig. (Fig.5).5). This comparative difference has been previously noted (42). When antibodies to flagellin or flagella were added to test for inhibition of TLR5 activation by flagellin or flagella, antibody to monomeric flagellin had a higher titer with greater neutralization of TLR5 activation than did antibodies raised to polymeric flagella (Fig. (Fig.6).6). In these assays, we noted the NRS control had some ability to activate the cells, which could be due to serum factors as has been described for TLR4 activation by LPS and CD14 and the LPS binding protein (LBP) (16, 38).
We also assessed the activation of TLR5 by live P. aeruginosa cells, the obvious entity which would trigger innate immunity during infection, and the inhibition of this activation by the corresponding antisera. In these assays, P. aeruginosa PAK and PA01 bacteria activated the NF-κB activity of the cells (Fig. (Fig.7A);7A); however, strain PAK was a better activator than strain PA01, which coincided with the results seen when testing the purified flagellar proteins from these strains. Analysis of LDH release as an indicator of cytotoxicity during incubation with the bacterial cells showed there was no more than a 10% difference in LDH release between cells incubated with bacteria and controls incubated without bacteria (data not shown). The ΔfliC mutant bacterial strains used as controls did not promote any NF-κB-dependent activation, which also confirmed the specificity of the cells for responding to flagellin (Fig. (Fig.7A).7A). Inhibition of the activation of TLR5 by antibody to flagella or flagellin showed that the two antisera had comparable activities at inhibiting TLR5 activation by strain PAK cells (Fig. (Fig.7B).7B). Notably, we did not find antibody to either flagella or flagellin capable of inhibiting the activation of TLR5 by P. aeruginosa strain PA01 cells (Fig. (Fig.7C)7C) compared to the activity of added NRS, which had a small inhibitory effect on its own. When we tested antiserum to type b flagellin or flagella along with two other type b strains in the TLR5 activation assay, we obtained an identical result (data not shown). Thus, while antibody to both type a flagella or flagellin inhibited TLR5 activation by strain PAK, no such inhibitor effect was seen with strain PA01 or other type b flagellum strains of P. aeruginosa.
The goals of this study were to compare the immunogenicity and functionality of flagellin and flagella as possible candidates for a vaccine against P. aeruginosa pneumonia and also to determine if such immunogens might give rise to antibodies that could interfere with the host's ability to detect the presence of flagellated pathogens via TLR5 activation (15, 32, 35). Prior work with animals (3, 13, 27) and CF patients lacking detectable P. aeruginosa colonization (12) has validated the vaccine potential of flagella. Vaccines based on monomeric flagellin have shown some efficacy in animals when a DNA vaccine encoding P. aeruginosa flagellin lacking TLR5 agonist activity is used (33), as has a multimeric construct of OprF(311-341)-OprI-flagellin fusion proteins (42) in models of burn wound and pulmonary infection. However, no direct comparison of the protective efficacy against infection by immunization with flagellin or flagella has been reported for P. aeruginosa or, as far as we can determine, for other microorganisms. In vitro studies suggest that the protective activity of antibody to flagella or flagellin correlates with inhibition of motility (24, 27, 28) and opsonic killing (42). Here we found that polymeric flagella, which by nature contain, in addition to the polymerized FliC protein, components such as the FliD cap protein and possible basal body components, were clearly superior to monomeric flagellin at inducing antibodies that inhibited motility, mediated opsonic killing, and provided protective immunity against acute lung infection. Also, the specificity of the protection for strains expressing only the homologous flagellar type indicated that potentially contaminating antigens in flagellar preparations isolated from P. aeruginosa cells were not contributing to the protective effects observed. Additionally, immunization with flagella induced lower titers of antibody that could interfere with the TLR5 agonist activity of flagellin and flagella, but only antibody to type a flagella and flagellin inhibited TLR5 activation by whole bacterial cells. These findings suggest a superiority of intact flagella over monomeric flagellin as a component of a P. aeruginosa vaccine.
In evaluating effects on P. aeruginosa motility and opsonic killing in vitro, we found flagellar-type-specific activity with antibodies to type b antigens, whereas the antibodies raised to type a flagella had some activity against type b strains. Since amino acid segments of the two flagella types are partially similar, it is possible that the type a flagellum vaccine was better able to induce antibodies to these shared components. Another potential basis for these differences may be glycosylation of flagellar proteins (5), wherein the type a flagellum is glycosylated by larger and more heterogeneous oligosaccharides than type b (40), possibly enhancing its immunogenicity. In addition, the lack of the glycan groups on the recombinant monomeric flagellins could also partly explain the poor ability of the antibodies raised to these proteins to promote opsonic killing. Nonetheless, for both motility inhibition and opsonic killing, the intact flagella were superior to monomeric flagellin at inducing antibodies mediating these in vitro correlates of protection.
In the mouse pneumonia model, we showed that passive immunization with antibodies to flagellin did not confer protection against infection with either type of P. aeruginosa. This outcome is different from that in the work of Saha et al. (33), who reported that immunization with a flagellin DNA vaccine protected against heterologous but not homologous bacterial challenge. This is curious in that almost all prior studies showed protection following flagellar vaccination was type specific (3, 13, 19, 27). The DNA vaccine construct may be superior at stimulating cross-protective humoral and/or cellular immune responses compared to intact flagellin protein, or active immunization with a DNA vaccine may provide results superior to those of passive immunization. Weimer et al. (42) obtained enhanced clearance of P. aeruginosa strain PA01 following immunization of mice with a construct of OprF311-341-OprI-flagellins, although the challenge dose was insufficient to cause mortality in the controls immunized with OprF311-341-OprI vaccine lacking flagellins. Similarly, immunization of young African green monkeys with this construct elicited serum antibodies that when passively administered to mice could promote clearance of P. aeruginosa from the lungs (41). However, since our goal was to compare flagellin-mediated immunity to that induced by polymeric flagella, we were able to validate a superior effect of the latter in passive transfer studies wherein injection of rabbit antibodies raised to flagella demonstrated high levels of protection in mice against lethal lung infection.
Active vaccination with intact flagella also showed high levels of type-specific, LPS-independent protection in a mouse pneumonia model and modest but less protection against motile clinical isolates from CF patients. These isolates were obtained from CF patients enrolled in the flagellar vaccine trial (12) who, in spite of vaccination, nonetheless became colonized with P. aeruginosa. We suspected these strains might be less susceptible to protection mediated by flagella derived from strains PAK and PA01, and the findings in our mouse studies bear this out. This could have important consequences for future vaccine trials incorporating flagella as vaccines and suggests that there may be additional flagellar components needed in a comprehensive vaccine, such as those expressing different subtype antigens on type a flagella (3). Since virtually all prior studies in this area have only evaluated protection against strain PAK or PA01, conclusions about the utility of flagellin or flagella as a vaccine have to be tempered with the lack of a comprehensive demonstration of efficacy against multiple strains, including clinical isolates, and, now that they are available, clinical isolates from flagellum-vaccinated CF patients unable to resist colonization by P. aeruginosa.
Another concern related to use of flagellin or flagella as vaccines is whether they will induce antibodies that interfere with TLR5-mediated innate immunity (15, 22, 32, 35, 36), which has been suggested by the studies of Saha et al. (33). However, the experimental design of that study is not informative as to the potential for antibody to flagella or flagellin to interfere with innate immunity during an actual infection with live P. aeruginosa cells. These investigators used purified, recombinant flagellin to enhance TLR5-mediated innate immunity by first incubating it with either antibody to wild-type flagellin or antibody raised to the R90A flagellin variant lacking the TLR5 binding domain. They subsequently inoculated BALB/c mice with this flagellin-antibody mixture 2 h prior to challenge with live P. aeruginosa PA01. The flagellin mixed with antibody to wild-type flagellin was less able to confer protection from lung infection than was flagellin mixed with antibody to the R90A variant, presumably due to inhibition of TLR5-mediated innate immune responses. However, this experiment did not indicate if antibody to flagellin inhibited TLR5 activation from whole bacteria during infection, thus increasing the animal's susceptibility to infection.
When testing the activation of TLR5 by P. aeruginosa bacterial cells, strain PAK was a more potent activator of the receptor than was strain PA01, consistent with the findings using purified flagella from these strains. However, for unclear reasons, while antibody to both flagella and flagellin could inhibit in a dose-dependent manner TLR5 activation mediated by strain PAK, neither of them could inhibit activation mediated by PA01 or by two other type b flagellum strains. Since there was no luciferase signal from cells infected with the ΔfliC strains, other P. aeruginosa PAMPs, such as LPS, were not active in this assay. The mechanisms that might explain the inability of antibody to flagella or flagellin to inhibit the activation of TLR5 by type b strains is not clear, but it may not necessarily be related to preventing TLR5 binding but rather perhaps to some other property of antibody to flagella or flagellin, such as inhibition of motility. In this regard, type a strains such as PAK may be less able to interact with the cells if motility is inhibited by the antibody to flagella, whereas the type b strains may either be less inhibited in their overall motility by antibody or use an alternative means of locomotion, such as pilus-mediated twitching motility, to interact with cells, or the antiserum to type b flagella or flagellin may be less potent at inhibiting motility than antibody to type a flagella, although this was not apparent in our in vitro motility inhibition assays.
Taken together, it seems that flagella, composed primarily of the FliC protein but also containing the type-specific FliD cap protein and basal body components, would be a better candidate for a vaccine against P. aeruginosa than flagellin, since immunization with flagella was demonstrated to be more proficient in generating flagellar-type-specific antibodies that inhibit motility, mediate opsonic killing, and protect against acute P. aeruginosa lung infection. In addition, although the antibodies to flagella inhibited activation of TLR5, they were less potent than antibodies to flagellin in this regard, minimizing the potential for interference with the induction of innate immunity mediated by activation of TLR5. Of note, however, use of flagella from strains PAK and PA01 as vaccines provided only modest protection against clinical isolates from flagellar-vaccine-immunized CF patients who nonetheless became colonized with flagellated P. aeruginosa. This suggests additional flagellar antigens may need to be incorporated into a comprehensive vaccine or strains other than the prototype PAK and PA01 may need to be utilized as a source for the flagellar antigens to find some that either are more immunogenic or give rise to more cross-reactive antibodies. Identifying the optimal formulation of the components of flagella needed for a maximally effective P. aeruginosa vaccine should enhance the utility of this approach for future evaluations.
This work was supported by NIH grants AI048917 and HL058398.
Editor: J. N. Weiser
Published ahead of print on 7 December 2009.