A Chimeric Influenza Virus Expressing an Epitope of Outer Membrane Protein F of Pseudomonas aeruginosa Affords Protection against Challenge with P. aeruginosa in a Murine Model of Chronic Pulmonary Infection

Abstract

The ability of a chimeric influenza virus containing, within the antigenic B site of its hemagglutinin, an 11-amino-acid (AEGRAINRRVE) insert from the peptide 10 epitope of outer membrane (OM) protein F of Pseudomonas aeruginosa to serve as a protective vaccine against P. aeruginosa was tested by using the murine chronic pulmonary infection model. Mice immunized with the chimeric virus developed antibodies that reacted in an enzyme-linked immunosorbent assay with peptide 10, with purified protein F, and with whole cells of various immunotype strains of P. aeruginosa but failed to react with a protein F-deficient strain of P. aeruginosa. The chimeric-virus antisera reacted specifically with protein F alone when immunoblotted against proteins extracted from cell envelopes of each of the seven Fisher-Devlin immunotype strains and had significantly greater in vitro opsonic activity for P. aeruginosa than did antisera from wild-type influenza virus-immunized mice. Subsequent to intratracheal challenge with agar-encased cells of P. aeruginosa, chimeric-virus-immunized mice developed significantly fewer severe lung lesions than did control mice immunized with the wild-type influenza virus. Furthermore, the chimeric influenza virus-immunized group had a significantly smaller percentage of mice with >5 × 10³ CFU of P. aeruginosa in their lungs upon bacterial quantitation than did the control group. These data indicate that chimeric influenza viruses expressing epitopes of OM protein F warrant continued development as vaccines to prevent pulmonary infections caused by P. aeruginosa.

Pseudomonas aeruginosa is an important opportunistic pathogen that causes severe infections in compromised humans, including those with cystic fibrosis (CF). In CF patients, P. aeruginosa remains the major cause of morbidity and mortality (2, 4, 17, 24, 33) due to chronic colonization of the CF lung. No means are currently available to prevent the colonization of the CF lung by P. aeruginosa or the concomitant pulmonary problems that follow.

Development of a vaccine that could successfully prevent the colonization of CF children with P. aeruginosa is a much sought-after goal. Among the more promising vaccine candidates for use in this clinical situation is outer membrane (OM) protein F of P. aeruginosa (11, 12, 34, 35). Protein F is a major OM protein (36) that is surface exposed in wild-type cells (12, 18, 27). Furthermore, it is present and immunologically cross-reactive in all strains of P. aeruginosa (3, 12, 28). Antibodies elicited by immunization with protein F are opsonic for P. aeruginosa (1, 7) but do not cross-react significantly with cells of other genera of gram-negative bacteria (3, 12, 28). Purified protein F from P. aeruginosa and recombinant protein F have been shown to provide significant protection in immunized animals against subsequent infection by P. aeruginosa in various animal models (7, 8, 11, 22, 23). Two linear B-cell epitopes within protein F have been identified through the use of synthetic peptides (10, 16) and have been shown to provide protection against both chronic (12) and acute (15) pulmonary infections with P. aeruginosa in animals immunized with each of the peptides conjugated to keyhole limpet hemocyanin as carrier. These two peptides (peptide 9, TDAYNQKLSERRAN, amino acid residues 261 to 274 of mature protein F, and peptide 10, NATAEGRAINRRVE, residues 305 to 318) appear to have potential for development as a vaccine for use in humans. Inducing effective systemic and local mucosal immune responses against epitopes of OM protein F might enhance protection against P. aeruginosa. Toward this end, we have produced chimeric influenza A viruses containing various lengths of the peptide 10 epitope incorporated into the antigenic B site of the viral hemagglutinin (HA) (9). In this study, we examined the ability of the chimeric influenza A virus expressing a peptide 10 epitope to serve as a protective vaccine in a mouse chronic pulmonary infection model.

The construction of chimeric influenza virus HG10-11 by ribonucleoprotein transfection was reported previously (9, 29, 31). The HG10-11 virus contains the P. aeruginosa OM protein F peptide 10 sequence AEGRAINRRVE inserted into site B of the HA of the influenza A/WSN/33 (WSN) virus between amino acids 158 and 159 (HA1 numbering). Two immunization protocols were used. (i) Initially, mice (5-week-old, female, specific-pathogen-free ICR mice from Harlan-Sprague Dawley, Indianapolis, Ind.) were immunized with either the WSN wild-type influenza virus (control) or the HG10-11 chimeric virus in accordance with the following protocol. Five immunizing doses were administered, all given at 2-week intervals and with no adjuvant. The first three doses consisted of 10³ PFU of virus in 50 μl of phosphate-buffered saline, pH 7.3, administered intranasally (i.n.) to anesthetized mice. The last two doses consisted of 10³ and 10⁵ PFU of virus, respectively, in 200 μl of saline administered intramuscularly (i.m.) into alternate hips of the mice. Two weeks after administration of the fifth and final immunizing dose, the mice were either bled for antisera or challenged with P. aeruginosa. (ii) In a revised immunization protocol, mice were immunized with five immunizing doses, all given at 2-week intervals with no adjuvant. The viral dose and route of administration were as follows: 10³ PFU of virus given i.n., 10⁵ PFU given i.m., 10⁵ PFU given i.n., 10⁵ PFU given i.m., and 10⁵ PFU given i.m. Two weeks after the fifth and final immunizing dose, the mice were either bled for antisera or challenged with P. aeruginosa.

Mice were immunized with various PFU doses of wild-type and chimeric viruses i.n., and their lungs were examined over a period of weeks to ensure that no lung lesions were found due to viral pathology at 2 weeks after immunization, i.e., that any lung lesions seen in mice following challenge with P. aeruginosa were not caused by the viral vaccine itself.

For in vitro analyses of antiserum activities, antisera from two or three mice were pooled following collection (2 weeks after administration of the final immunizing dose) from mice immunized in accordance with the revised immunization protocol (described above) with the WSN wild-type virus or with the chimeric HG10-11 virus. These antisera were tested for titers of immunoglobulin G (IgG) antibodies against various enzyme-linked immunosorbent assay (ELISA) antigens, including peptide 10, purified OM protein F, whole cells of various strains of P. aeruginosa (PAO of Fisher-Devlin [FD] immunotype 7, FD immunotypes 1 to 6 [8, 22]), and KG1077, a protein F-deficient mutant of the PAO strain [13] obtained from N. Gotoh, Kyoto, Japan, and the two (WSN and HG10-11) influenza viruses. The procedures for performing these ELISAs have been published previously (16, 22). The ELISA was performed a minimum of three times with each of the antisera.

The pooled antisera were also used for Western immunoblotting, performed as described previously (22), against purified OM protein F and against proteins extracted from cell envelopes of each of the FD immunotype strains and the KG1077 protein F-deficient strain of P. aeruginosa. In opsonophagocytic assays, the ability of the antisera from HG10-11-immunized mice to mediate the uptake of two heterologous-immunotype (FD immunotype 2 and 4) strains of P. aeruginosa by human polymorphonuclear leukocytes (PMNs) was compared with the ability of antisera from the WSN-immunized mice to do likewise. The assay was performed as described previously (7), and duplicate assays were run for each antiserum-immunotype strain mixture. Briefly, bacterial cells were mixed with heat inactivated (56°C for 30 min) sera and incubated with gentle shaking at 37°C for 30 min. Human whole blood was added to the mixture and incubated for another 30 min at 37°C. After incubation of the blood with the bacteria and antisera, slides of each mixture were prepared and stained with Giemsa stain. Each slide was examined microscopically, and the number of bacterial cells contained within the first 75 isolated, intact PMNs encountered was determined for each reaction mixture. The mean number (± the standard deviation [SD]) of bacterial cells per PMN was calculated. The statistical significance of differences noted between groups was evaluated by using the unpaired Student t test, and P ≤ 0.05 was considered statistically significant.

Mice were challenged by using a model of chronic pulmonary infection with P. aeruginosa (8, 32). Two weeks after the final immunization, the mice were challenged with agar beads containing P. aeruginosa cells of the FD immunotype 4 strain. The mice were first anesthetized with an intraperitoneal injection of sodium pentobarbital and then inoculated via a tracheal incision with 50 μl of an agar bead slurry encasing approximately 5 × 10³ CFU of P. aeruginosa. A beaded-tip 22-gauge needle was gently guided to favor inoculation of the left lung. The incision was closed with sterile wound clips. Eight days after the challenge, the mice were sacrificed by administering an overdose of halothane (Ayerst Laboratories, Inc., New York, N.Y.). Protection afforded to immunized mice by the chimeric virus was assessed by two methods. First, the lungs were examined macroscopically for the presence of lesions (5–8, 12). Lesions were scored as 0 to 4+ based on the scale detailed in Table 1, footnote b. Scoring of the pulmonary lesions was performed by an investigator well experienced in macroscopic lung lesion scoring. Second, bacterial quantitation of the number of P. aeruginosa CFU present in the lungs was performed as described previously (7, 8).

TABLE 1.

Scoring of macroscopic lung lesions in immunized^a mice following challenge with FD immunotype 4 P. aeruginosa in a chronic pulmonary infection model

Immunization group	No. of mice having lesions scoredb as:					No. of mice having lesions scored as 3+–4+/total no. of mice (%)	P valuec
	0	1+	2+	3+	4+
WSN parent virus	1	1	1	14	3	17/20 (85)
HG10-11 virus	3	3	6	5	4	9/21 (42.9)	0.006

^aInitial immunization protocol: 10³ PFU of virus administered i.n. for three immunizations, followed by 10³ PFU administered i.m. and then 10⁵ PFU administered i.m. 

^bLesions were scored as follows: 0, absence of any macroscopic lesion; 1⁺, presence of one or two small lesions not exceeding 1 mm in diameter; 2⁺, presence of three or more small lesions not exceeding 1 mm in diameter; 3⁺, presence of a medium lesion 2 to 5 mm in diameter; 4⁺, presence of a large lesion exceeding 5 mm in diameter. 

^cP value was determined by Fisher’s exact test. 

Statistical analyses of the differences between control WSN virus-immunized mice and HG10-11 virus-immunized mice upon scoring of lung lesions and quantitation of the bacteria present in the lungs were performed with the IBM EpiStat Basic Statistics Program, and P values were calculated by the Fisher exact test. P values of ≤0.05 were considered to be significant. All animals used in this study were handled in accordance with the guidelines of the Louisiana State University Medical Center-Shreveport Animal Care and Use Committee.

We investigated the protective efficacy of a chimeric influenza virus expressing one of the previously identified (10, 16) B-cell epitopes within OM protein F of P. aeruginosa. Chimeric viruses containing inserts of various lengths (5 to 11 amino acids) of the peptide 10 epitope (10, 16) of OM protein F were constructed previously (9). Oligonucleotides representing the subsets of peptide 10 were synthesized, and PCR products were ligated into the antigenic B site of the cloned HA molecule. Transcripts from the plasmids containing the peptide 10 hybrid P. aeruginosa HA were transfected as influenza ribonucleoprotein complexes into influenza virus-infected cells. This permitted the rescue of chimeric viruses containing the bacterial peptide insert. From the four chimeric viruses thus recovered (9), the HG10-11 chimeric influenza virus containing the longest epitope (AEGRAINRRVE) of peptide 10 within the antigenic B site was selected for further study.

This chimeric influenza virus, containing the 11-mer insert of peptide 10, afforded significant protection against chronic pulmonary infection upon subsequent challenge of virus-immunized mice. The initial immunization protocol tested involved administering three doses of 10³ PFU of the chimeric virus i.n. at 2-week intervals, followed at 2-week intervals by i.m. administration of 10³ and 10⁵ PFU of virus. This initial immunization protocol elicited a 160 titer of IgG antibodies directed toward FD immunotype 4 whole cells as determined by whole-cell ELISA of pooled sera collected from mice 2 weeks after administration of the fifth and final immunizing dose of the HG10-11 chimeric virus. Upon challenge with the FD immunotype 4 strain of P. aeruginosa in the chronic pulmonary infection model, the mice immunized with the HG10-11 chimeric virus by this initial protocol were afforded statistically significant (P = 0.006) protection from development of the more severe (≥3+, i.e., medium or large) lung lesions (Table 1) compared to control WSN virus-immunized mice. However, when all severe lung lesions were considered (i.e., lesions graded ≥2+), the HG10-11 virus-immunized mice were not significantly different from the WSN virus-immunized mice (15 [71.4%] of 21 versus 18 [90%] of 20, respectively; P = 0.134).

To elicit a higher titer of P. aeruginosa-specific antibodies, the immunization protocol was revised to consist of doses (all given at 2-week intervals) administered as follows: 10³ PFU of virus i.n., 10⁵ PFU i.m., 10⁵ PFU i.n., 10⁵ PFU i.m., and 10⁵ PFU i.m. This revised protocol succeeded in eliciting higher titers of IgG antibodies directed against whole cells of P. aeruginosa. Two weeks following administration of the final immunizing dose, two or three mice were bled and the resultant pooled antisera were checked by ELISA to determine titers of IgG antibodies reactive with various antigens. Due to viral antigens, antisera from WSN and HG10-11 virus-immunized mice each reacted at titers of 10,240 to the WSN virus and to the chimeric HG10-11 virus. The WSN virus antisera were nonreactive with all of the P. aeruginosa antigens tested, whereas the HG10-11 virus antisera reacted positively with whole cells of strains of all seven of the FD immunotypes (Table 2). The HG10-11 virus antisera gave an approximately equivalent reaction, ranging in mean titer from 266 to 960 for immunotypes 1 to 4 and 6 and 7, but had a mean titer of 33 for the FD immunotype 5 strain. This FD immunotype 5 strain had a greatly reduced protein F band upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis, which agrees with the reduced ELISA titer in comparison with those of strains of the remaining six immunotypes. This indicates that the reaction seen with the various strains depended upon the presence of protein F. This is further supported by the fact that the HG10-11 antisera had a titer of 0 against the KG1077 strain, which is a protein F-deficient mutant derived from the PAO strain. Western immunoblotting revealed that the HG10-11 antisera reacted specifically with purified protein F and with the protein F band in strains of all seven FD immunotypes but failed to react with the KG1077 strain (data not shown). The HG10-11 antisera reacted in an ELISA with protein F purified from the PAO strain at a titer of 640 and with peptide 10 at a titer of 160, indicating that the antibodies elicited were reactive with the peptide 10 epitope of protein F.

TABLE 2.

Mean titers of IgG antibodies in antisera as determined by ELISA with whole cells of various strains of P. aeruginosa as antigens

Antisera	Mean titera determined by ELISA with whole cells of strain:
	FD1b	FD2c	FD3	FD4d	FD5	FD6	FD7 (PAO)	KG1077e
Antisera elicited by:
WSN virus (wild type)	0 ± 0	1.7 ± 2.9	0 ± 0	3.3 ± 5.8	0 ± 0	0 ± 0	3.3 ± 5.8	0 ± 0
HG10-11 virus (chimeric)	960 ± 453	560 ± 160	266 ± 92	773 ± 896	33 ± 12	266 ± 323	480 ± 277	0 ± 0

^aMean ± SD of three determinations, unless noted otherwise. 

^bFD, Fisher-Devlin immunotype. 

^cMean ± SD from four determinations. 

^dMean ± SD from six determinations. 

^eKG1077 is a protein F-deficient mutant of strain PAO. 

We confirmed that the antibodies elicited by the HG10-11 chimeric influenza virus had potentially protective functional activity as opsonins and that the opsonic activity was comparable to that of antisera elicited by purified protein F itself. The HG10-11 virus-immunized mice produced antisera that were significantly more opsonic for strains of both of the FD immunotypes (2 and 4) used in this assay than were the antisera from WSN virus-immunized mice (Table 3). The duplicate assays in all four cases likewise showed statistically significant opsonic enhancement by the test serum (protein F- or HG10-11-immunized serum) over the control serum (normal mouse serum [NMS] or WSN-immunized serum), respectively (data not shown). This suggests that the peptide 10 epitope has importance in providing immunological protection against infection by P. aeruginosa.

TABLE 3.

Quantitation of phagocytic uptake by PMNs of P. aeruginosa cells following exposure to various sera

Test sera	FD immunotype of bacteria	Mean no. of bacteria associated/ PMN ± SD	Ratioa	P valueb
NMSc	2	5.5 ± 3.4
Protein F immunizedd	2	10.2 ± 8.6	1.9	<0.0001
WSN immunized	2	3.1 ± 2.9
HG10-11 immunized	2	6.5 ± 6.9	2.1	0.0001
NMSc	4	12.6 ± 7.7
Protein F immunizedd	4	19.4 ± 11.2	1.5	<0.0001
WSN immunized	4	12.4 ± 7.4
HG10-11 immunized	4	16.9 ± 9.4	1.4	0.0002

^aEach ratio represents the mean number of bacteria associated per PMN with test sera divided by the mean number of bacteria associated per PMN with the appropriate control sera. 

^bP values were determined by the unpaired Student t test. 

^cCollected from 28 unimmunized mice. 

^dProtein F-immunized sera were pooled from 30 mice following immunization with purified protein F (10 μg with aluminum hydroxide adjuvant for two immunizations, followed by 10 μg with no adjuvant for two immunizations, 2-week interval between immunizations, all injections given i.m. in hips). 

Mice immunized by the revised immunization protocol were afforded significant protection against challenge with the FD immunotype 4 strain of P. aeruginosa in the murine chronic pulmonary infection model, as determined by two different methods. Mice vaccinated with the HG10-11 chimeric virus were protected against the development of both severe (≥2+) and more severe (≥3+) lung lesions (Table 4). Following challenge, 91% of control WSN virus-immunized mice had lung lesions scored as ≥3+, whereas only 37% of HG10-11 virus-immunized mice developed lung lesions scored as ≥3+. When lesions scored as ≥2+ were considered, 56% of the HG10-11 virus-immunized mice exhibited such lesions, which represented a significant (P = 0.005) reduction from the proportion (91%) of control WSN virus-immunized mice having ≥2+ lesions. A second indicator of protection afforded by the chimeric virus vaccine was the decrease in bacteria present in the lungs of immunized mice 8 days after challenge with the FD immunotype 4 P. aeruginosa strain (Table 5). Although the proportion of mice whose lungs yielded no bacterial growth was increased from 21.7% of control WSN virus-immunized mice to 40.7% of HG10-11 virus-immunized mice, this increase was not statistically significant at the 0.05 level. However, if one considers the percentage of mice which had <5 × 10³ CFU in their lungs upon bacterial quantitation, the HG10-11 virus-immunized mice had a significant (P = 0.011) increase (66.7% of the mice had fewer than 5 × 10³ CFU versus only 30.4% of WSN virus-immunized control mice). The 5 × 10³-CFU cutoff was selected because it represented a 99% (or 2-log) decline from the mean number of bacteria found in the lungs of challenged, WSN virus-immunized control mice in an initial experiment. Fourteen control mice were challenged, and then the number of bacteria present in their lungs at day 8 after challenge was determined. Three (21.4%) of the 14 yielded no growth. The mean number of bacteria present in the remaining 11 mice was 7.94 × 10⁵ ± 11.3 × 10⁵ (SD), and only 3 (27.3%) of these 11 mice had fewer than 5 × 10³ CFU in their lungs. Hence, the 5 × 10³-CFU cutoff was predicted to represent a greater-than-99% reduction from the mean number of bacteria present in control lungs following challenge. The lungs of 18 control WSN virus-immunized mice (Table 5) yielded a mean bacterial growth of 1.72 × 10⁵ ± 3.1 × 10⁵ (SD) CFU. In this experiment, the 5 × 10³-CFU cutoff actually represented a greater-than-97% reduction from the mean number of P. aeruginosa CFU found in control lungs. In addition, approximately 5 × 10³ CFU were inoculated into the lungs at the time of challenge. Thus, this 5 × 10³-CFU cutoff also indicates whether P. aeruginosa was capable of establishing an infection in which it substantially increased in number within the lung or remained contained. Significant protection was demonstrated, since 8 days after a challenge with P. aeruginosa, 70% of the control WSN virus-immunized mice had >5 × 10³ CFU of P. aeruginosa in their lungs, whereas only 33% of the HG10-11 virus-vaccinated mice had >5 × 10³ CFU of P. aeruginosa in their lungs (P = 0.01).

TABLE 4.

Scoring of macroscopic lung lesions in mice immunized by the revised protocol^a following challenge with FD immunotype 4 P. aeruginosa in a chronic pulmonary infection model

Immunization group	No. of mice having lesions scoredb as:					No. of mice having lesions scored as 2+–4+/total no. of mice (%)	P valuec	No. of mice having lesions scored as 3+–4+/total no. of mice (%)	P valuec
	0	1+	2+	3+	4+
WSN parent virus	1	1	0	9	12	21/23 (91.3)		21/23 (91.3)
HG10-11 virus	7	5	5	5	5	15/27 (55.6)	0.005	10/27 (37.0)	<0.0001

^aRevised immunization protocol: 10³ PFU of virus administered i.n., 10⁵ PFU administered i.m., 10⁵ PFU administered i.n., followed by two final 10⁵ PFU doses administered i.m. 

^bLesions were scored as detailed in Table 1, footnote b. 

^cP values were determined by Fisher’s exact test. 

TABLE 5.

Quantitation of P. aeruginosa present in the lungs of mice immunized by the revised protocol^a following challenge in a chronic pulmonary infection model

Immunization group	No. of mice with lungs yielding no growth/total no. of mice (%)	P valueb	No. of mice with lungs yielding <5 × 103 CFUc/total no. of mice (%)	P valueb
WSN parent	5/23 (21.7)		7/23 (30.4)
HG10-11	11/27 (40.7)	0.129	18/27 (66.7)	0.011

^aRevised immunization protocol as detailed in Table 4, footnote a. 

^bP values were determined by Fisher’s exact test. 

^c<5 × 10³ CFU represents a decrease of >97% from the mean number (1.72 × 10⁵) of bacteria present in the lungs of control mice at day 8 following challenge.  

Chimeric influenza viruses seem to be well suited to serve as vaccines for immunization against chronic lung infections caused by P. aeruginosa for a number of reasons (9, 29). Chimeric viruses containing P. aeruginosa epitopes can be delivered i.n. to stimulate a local mucosal immune response to P. aeruginosa in the lung upon replication of the viruses in the upper, and possibly the lower, respiratory tract. Many serotypes of influenza virus are available should it be necessary to change the serotype of the virus to prevent any unwanted memory response against the virus in booster immunizations. If the attenuated chimeric viruses that replicate in the upper (and lower) regions of the lung prove to be deleterious to the lung, then either attenuated or cold-adapted virus vaccines could be used. Cold-adapted influenza virus strains, which would replicate only in the nares and upper respiratory tract, have previously been shown to be immunogenic and safe for use in humans with CF and their family members (14). Hence, the use of influenza virus as a vaccine vector for delivery of selected epitopes of P. aeruginosa into the lungs of CF children appears to hold much promise.

The reverse genetic system for incorporating heterologous epitopes into the antigenic B site of the influenza virus HA molecule has also been used successfully to incorporate epitopes from Plasmodium yoelii (20, 30) and viruses (19, 21, 25, 26), and such chimeric viruses induce a humoral immune response at the mucosal level (26, 29). The present findings also indicate that the influenza virus vector system can successfully serve as a vaccine carrier for epitopes from bacteria. Finally, we believe that the HG10-11 chimeric influenza virus containing an 11-amino-acid insert from the peptide 10 epitope of OM protein F of P. aeruginosa incorporated within the antigenic B site of the viral HA has potential for continued development as a vaccine capable of affording protection against infection by P. aeruginosa.