Bordetella pertussis Filamentous Hemagglutinin Enhances the Immunogenicity of Liposome-Delivered Antigen Administered Intranasally

Abstract

In an attempt to increase the immunogenicity of mucosally delivered antigens, we incorporated the Bordetella pertussis filamentous hemagglutinin (FHA) adhesin into liposomes containing the glutathione S-transferase of Schistosoma mansoni (Sm28GST) as a model antigen. Outbred mice immunized twice intranasally with liposomes containing a constant suboptimal dose of Sm28GST and increasing doses of FHA produced anti-Sm28GST antibodies in a FHA dose-dependent manner. The addition of 3 μg of FHA to the liposomes induced more than 10-fold-higher anti-Sm28GST antibody titers, compared to those induced by liposomes without FHA. The presence of FHA did not alter the nature of the humoral immune response, and the sera contained anti-Sm28GST immunoglobulin G1 (IgG1), IgG2a, and IgG2b. However, anti-Sm28GST IgA was only detected when at least 3 μg of FHA was added to the preparation. These results show a promising potential for FHA to enhance the immunogenicity of mucosally administered antigens incorporated into liposomes.

One of the main objectives of current vaccine research is the development of vectors capable of inducing strong immune responses against protective antigens when delivered by a mucosal route. Compared to standard systemic routes, mucosal routes offer several advantages, including the ease of administration in a noninvasive fashion and diminished risk of contamination, which may be caused by injection. Different formulations have been developed in recent years to increase the immunogenicity of mucosally delivered antigens. One of these formulations is based on multilamellar liposomes containing dimyristoylphosphatidylcholine (DPPC) and dipalmytoylphosphatidyl glycerol (DMPG) (17). We have recently demonstrated that the association of an antigen with such liposomes was able to induce a protective immune response when given by the oral route (11). Although incorporation of the antigen into liposomes certainly enhanced its immunogenicity when administered by the oral route, large amounts of antigen were still needed.

We reasoned that immunogenicity might be further enhanced if the liposomes were targeted to mucosal sites by the addition of specific adherence molecules. It has recently been reported that coating liposomes with immunoglobulin A (IgA) enhances their uptake into Peyer’s patches and thereby increases both the mucosal and systemic immune responses after rectal administration together with cholera toxin (27). Furthermore, the B subunit of cholera toxin was also shown to target microparticles to the M cells of Peyer’s patches, resulting in an increase in immune responses (6). As an alternative to oral or rectal delivery of antigens, we explored the intranasal delivery of liposome formulations containing a schistosome model antigen and the Bordetella pertussis filamentous hemagglutinin (FHA) as an adhesin specific for the respiratory tract tissues (for a review, see reference 13). FHA expresses several adherence activities, including binding to carbohydrates on respiratory cilia (24); binding to sulfated carbohydrates (9), which is involved in the attachment of B. pertussis to epithelial cells and the extracellular matrix; and binding to macrophage integrins via an RGD sequence (10). Moreover, FHA is a strong mucosal immunogen, as evidenced by the high levels of anti-FHA antibodies produced in humans infected with B. pertussis (26).

In this study, therefore, we incorporated FHA into liposomes together with the Schistosoma mansoni glutathione S-transferase (Sm28GST) as a model antigen. Here, we show that the enhancement of the immune response to Sm28GST was dependent on the FHA dose, without resulting in a change in the isotypic profile.

Specific immune response obtained after intranasal administration of Sm28GST liposomes.

Recombinant Sm28GST (rSm28GST) produced in Saccharomyces cerevisiae and provided by Transgène S.A. (Strasbourg, France) was affinity purified as described previously (22). Liposomes were prepared as previously described (11) by using a mixture of two lipid components in a 9:1 (DPPC to DMPG) (Genzyme, Cambridge, Mass.) molar ratio. In order to determine the minimal immunizing dose of rSm28GST when incorporated into liposomes, we prepared liposomes with three different concentrations of rSm28GST (0.2, 1, and 5 mg/ml). However, the proportion of protein incorporated into liposomes at these concentrations was not strictly linear and corresponded to 75, 70, and 60%, respectively, of protein incorporation. The liposomes were washed three times in phosphate-buffered saline (PBS) and centrifuged at 10,000 × g for 30 min. The pellet was resuspended in PBS and adjusted to 400 μl (2 μmol of phospholipids per 40 μl). Six-week-old female OF1 mice (Iffa Credo, L’Arbesle, France) were anesthetized with 200 μl of 5% sodium pentobarbital (Sanofi, Libourne, France) per 10 g of body weight given intraperitoneally and immunized with 40 μl of the liposome suspension or PBS deposited in the nostrils. Thus, the dose administered per mouse at each instillation corresponded to 15, 70, or 300 μg of rSm28GST, depending on the concentration. The liposome preparations were given twice intranasally with a 2-week interval. The specific immune responses in the sera were analyzed 2 weeks after the second administration (i.e., on day 27). As shown in Table 1, anti-Sm28GST IgG1, IgG2a, and IgG2b were detected in significant amounts on day 27, and antibody levels increased in proportion to the dose of rSm28GST administered. A weak serum anti-Sm28GST IgA response was observed with the highest dose of rSm28GST. Individual analyses indicated that one mouse in five did not produce a detectable anti-Sm28GST immune response, even with the largest dose of antigen.

TABLE 1.

Dose-dependent immune response in OF1 mice after intranasal administration of Sm28GST liposomes

Dose (μg) of Sm28GST/ instillation	Antibody levela
	IgG1	IgG2a	IgG2b	IgG3	IgA
0	−b	−	−	−	−
15	80	89	−	−	−
70	191	158	182	−	−
300	228	1,062	1,090	−	49

^aResults are given as titers of pooled sera of five animals. Titers are the reciprocals of the dilutions giving an optical density three times that of the conjugate control for the pooled serum samples. 

^b−, titer < 40. 

Characterization of rSm28GST-FHA liposomes.

FHA was purified by heparin-Sepharose chromatography as previously described (15) with B. pertussis BPRA, a strain lacking the pertussis toxin gene (2), as the source.

In order to evaluate the possible effect of FHA associated with rSm28GST in liposomes on the anti-Sm28GST antibody response, we used an intermediate dose of 40 μg of rSm28GST, administered at each instillation. This dose was obtained by using a concentration of 0.6 mg of rSm28GST/ml in the rehydration solution containing various amounts of FHA (0.5, 5, or 50 μg/ml). The different liposome preparations were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis with 12% polyacrylamide gels and by immunoblotting with rat polyclonal anti-FHA or anti-Sm28GST antibodies. As shown in Fig. 1, the different liposome preparations contained similar amounts of Sm28GST, indicating that the addition of FHA, at least up to 50 μg/ml, did not interfere with the rate of incorporation of rSm28GST into the liposomes. In addition to the major 28-kDa protein, we detected larger-molecular-mass forms which may correspond to the polymeric forms of Sm28GST and their association with phospholipids. The number of liposomes loaded was adjusted in order to obtain comparable Sm28GST bands; for this reason FHA was only detected in liposomes prepared with the highest concentration of the adhesin.

FIG. 1.

Immunoblot analysis of Sm28GST-FHA liposomes. Five microliters of each liposome suspension used for in vivo experiments (corresponding to 0.25 μmol of phospholipids) was analyzed by immunoblotting with rat hyperimmune polyclonal anti-FHA (left panel) and rat hyperimmune polyclonal anti-Sm28GST (right panel). The nitrocellulose sheets were then incubated with anti-rat antibodies coupled to alkaline phosphatase (Sigma) (1/4,000 in PBS containing 1% fat dry milk) and developed with 5-bromo-4-chloro-3-indolylphosphate–nitroblue tetrazolium (Promega, Madison, Wis.). Lanes 1 contained 3 μg of free rSm28GST, and lanes 2 contained 1.5 μg of free FHA. Lanes 3, 4, and 5 contained the three different Sm28GST-FHA liposome preparations (i.e., high, medium, and low concentrations of FHA, respectively). Lanes 6 contained the Sm28GST liposome preparation. The molecular mass markers, expressed in kilodaltons (kD), are given in the left margin.

Since the sequence of FHA contains several hydrophobic stretches (5), it seemed likely that some parts of FHA are exposed at the surfaces of the liposomes. The surface exposure of FHA was assessed by flow cytometry. The liposomes were labeled in suspension with anti-FHA polyclonal antibodies diluted 1/100 in PBS containing 1‰ Tween 20 (PBS-T). After being washed, the liposomes were incubated with an anti-rat antibody conjugated to fluorescein isothiocyanate (Sigma, St. Louis, Mo.) at 1/100 in PBS-T. After two additional washes with PBS-T the liposomes were subjected to fluorescence-activated cell sorter (FACS) analysis (Coulter Epics Elite, Miami, Fla.). Specific surface labeling of the rSm28GST-FHA liposomes was demonstrated, indicating that at least some FHA epitopes are accessible to antibodies and therefore exposed at the surface of the liposomes (Fig. 2).

FIG. 2.

FACS analysis of Sm28GST-FHA liposomes. Sm28GST liposomes and Sm28GST-FHA liposomes were labeled with a rat anti-FHA polyclonal antibody followed by fluorescein isothiocyanate-labeled anti-rat antibody and then subjected to FACS analysis. The numbers (nb) of liposomes were plotted against fluorescence intensity as indicated.

Comparison of immune responses obtained after intranasal administration of rSm28GST-FHA liposomes containing increasing doses of FHA.

By considering the level of FHA incorporation to be comparable to that of Sm28GST, the actual doses of FHA at each instillation could be estimated at 3, 0.3, and 0.03 μg for the three separate liposome preparations. No FHA was added for the control liposomes. OF1 mice were immunized twice intranasally with a 2-week interval with either PBS, control rSm28GST liposomes (40 μg per administration), or one of the three rSm28GST-FHA liposomes described above. Two weeks after the second instillation, anti-Sm28GST and anti-FHA antibody responses were analyzed in pooled sera of the immunized mice (Table 2). The sera were analyzed by enzyme-linked immunosorbent assays as previously described (11). For specific detection, microtiter plates were coated as a first step with 50 μl (per well) of a solution containing either 5 μg of FHA per ml or 10 μg of Sm28GST per ml. Anti-Sm28GST antibodies were detected with either liposome preparation. However, their titers increased proportionally with the amount of FHA present in the liposomes. The addition of 0.03 μg of FHA resulted in a threefold increase in anti-Sm28GST antibody titers, and the addition of 0.3 μg of FHA increased the anti-Sm28GST titers approximately 10-fold. The level of specific antibodies to the parasite antigen obtained with 40 μg of Sm28GST in combination with 3 μg of FHA was comparable to that obtained with 300 μg of Sm28GST without FHA (Table 1), demonstrating the potential of FHA to increase the immune response.

TABLE 2.

FHA dose-dependent immune response in OF1 mice after intranasal administration of Sm28GST-FHA liposomes

Liposome or control	Dose (μg)	Antibody levela										No. positiveb/total		P valuec
		IgG1 anti-:		IgG2a anti-:		IgG2b anti-:		IgG3 anti-:		IgA anti-:		AntiSm28GST	Anti-FHA
		Sm28GST	FHA	Sm28GST	FHA	Sm28GST	FHA	Sm28GST	FHA	Sm28GST	FHA
Sm28GST	40	78	−d	51	−	−	−	−	−	−	−	5/15	0/15
Sm28GST-FHA	40/0.03e	228	−	142	−	130	−	−	−	−	−	5/15	0/15	0.1754 (NS)
	40/0.3	726	−	715	−	347	−	−	−	46	−	10/16	1/16	0.0228
	40/3	1,554	919	1,308	889	1,002	1,133	−	−	93	−	16/16	16/16	<0.0001
PBS		−	−	−	−	−	−	−	−	−	−	0/15	0/15

^aResults are given as titers of pooled sera of 15 to 16 animals. Titers are the reciprocals of the dilutions giving an optical density three times that of the conjugate control for the pooled serum samples. 

^bSerum response was arbitrarily considered to be positive when individual serum titer of Ig(H+L) anti-Sm28GST or anti-FHA was higher than 40. 

^cStudent’s t test analysis was performed on Ig(H+L) anti-Sm28GST values for individual mice between the group immunized with Sm28GST liposomes and the three groups immunized with Sm28GST-FHA liposomes. The probability value is not significant (NS) above 0.05. 

^d−, titer < 40. 

^eSm28GST dose/FHA dose. 

When compared with the response obtained with liposomes containing 300 μg of Sm28GST alone, the addition of FHA did not appear to strongly affect the isotype profile of the anti-Sm28GST antibodies, since IgG1, IgG2a, and IgG2b production all increased in proportion to increasing amounts of FHA in the liposome preparations. Anti-Sm28GST IgA became detectable only when mice were immunized with liposomes containing the two highest doses of FHA. Anti-Sm28GST IgG3 remained undetectable in the sera regardless of the presence of FHA in the liposomes. A similar antibody response in terms of profile and titers has been obtained in mice after intranasal infection with a recombinant B. pertussis strain producing Sm28GST fused to FHA (18).

Anti-FHA antibodies were only detected when 3 μg of FHA was incorporated into the liposomes. This immune response was also mixed, with IgG1, IgG2a, and IgG2b antibodies present. However, no anti-FHA IgG3 or IgA was detected in the sera.

Since outbred mice were used in this study, it was of interest to determine whether the presence of FHA in the liposomes stimulated the anti-Sm28GST immune response in the whole population, or only in a proportion of animals. Therefore, individual sera were analyzed for the presence of anti-Sm28GST and anti-FHA antibodies (Table 2). Arbitrarily, a serum titer was considered positive when it was higher than 40. In the group of mice receiving liposomes containing the highest concentration of FHA, all of the animals produced anti-Sm28GST antibodies. The number of seropositive animals decreased with decreasing doses of FHA. Similarly, anti-FHA antibody titers higher than 40 were detected in all mice that had received liposomes with the largest amount of FHA, while only one mouse was found seropositive for FHA in the group immunized with liposomes containing 0.3 μg of FHA. These results indicated that all outbred mice immunized with Sm28GST-FHA liposomes containing the highest dose of FHA produced antibodies to both Sm28GST and FHA, showing that an immune response was generated independently of the genetic background of the animals. However, the enhancement of this immune response was heterogeneous in this population, and the titers varied from 50 to 2,000. In all cases, the antibody response was maximum 2 weeks after the last administration. The immune response decreased slowly, but it was still strong after 7 weeks, since at that time the titers were at 50% of the maximum response. The decrease was independent of the presence of FHA on the liposomes.

Since FHA was prepared from culture supernatants of the gram-negative B. pertussis, and since lipopolysaccharide (LPS) can potentiate the immune response against a coadministered antigen, it was important to investigate whether the enhancement of the anti-Sm28GST antibody responses observed with Sm28GST-FHA liposomes was due to an LPS contamination. Therefore, a new series of liposomes was prepared with FHA previously subjected to polymyxin-Sepharose 4B chromatography to remove residual LPS (16). When mice were immunized with these preparations (4 IU of LPS/mouse/administration; Limulus amoebocyte lysate analysis), no significant difference in anti-Sm28GST antibody production was detected compared to that of the mice immunized with liposomes containing FHA not treated with polymyxin-Sepharose 4B (71 IU of LPS/mouse/administration), indicating that the increase in anti-Sm28GST antibody titers observed with rSm28GST-FHA liposomes was not due to the presence of endotoxins in our preparations but to that of FHA (data not shown).

It is likely that the enhancement of the immune response against Sm28GST by the addition of FHA is related to the properties of adherence of the protein to the mucosal tissues. In order to test the adhesive properties of Sm28GST-FHA liposomes, we performed an in vitro assay with intraperitoneal macrophages. Sm28GST or Sm28GST-FHA liposomes were fluorescently labeled with carbocyanin dye (DiI; 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate) (Molecular Probes, Eugene, Oreg.) (3). Mouse intraperitoneal macrophages were fixed and then incubated with fluorescent liposomes. FACS analysis demonstrated that macrophages incubated with Sm28GST-FHA liposomes were twice as fluorescent as macrophages incubated with Sm28GST liposomes (data not shown). In vivo, adherence might protect the liposomes from degradation and enhance their rapid interaction with specialized immune cells. FHA expresses several binding activities (13, 14). Some of the sites have been mapped on the protein and may be involved in the enhancement of the immune response against heterologous antigens when simultaneously incorporated into liposomes. This may ultimately lead to the identification of the target cell type involved in the immune enhancement. Alternatively, it is possible that the enhancement of the immune response after intranasal administration of liposomes containing FHA is the result of the multiple binding activities of the adhesin, which allow the liposomes to be presented to several different sites in the respiratory tract, including macrophages and epithelial cells, rather than to a single cell type.

In addition to the presence of the bacterial adhesin, the lipid composition of the liposomes may also contribute to the modulation of the immune response against the vectorized antigen. Our liposomes are composed of synthetic phospholipids, i.e., DPPC and DMPG. However, the DPPC is a natural lipid component of the pulmonary surfactant, which also contains proteins such as surfactant protein-A (SP-A) (25). It has been shown that SP-A specifically binds DPPC (12) and is able to promote its uptake by alveolar type II pneumocytes (23) or alveolar macrophages (19). This uptake is followed by intracellular processing involving different intracellular organelles (20) and promoting recycling and reutilization of alveolar surfactant. Therefore, the lipid composition of the liposomes may potentially modulate antigen presentation. Furthermore, uptake of negatively charged liposomes by alveolar macrophages appears to block their immunosuppressive activity, thereby facilitating antibody responses (4). The influence of the lipid composition of our preparations on immunogenicity remains to be characterized.

Although so far only tested for a single model antigen, it is likely that the association of FHA with liposomes may also enhance immune responses against antigens from other pathogens. The use of FHA as an adhesin on liposomes is expected to be safe for human use, as this molecule is already present in the new acellular vaccines against whooping cough (1, 7, 8, 21). Since the addition of FHA to liposomes allows the amount of antigen required for the induction of a desired immune response to be decreased approximately 10-fold, it may be particularly attractive for the optimization of synthetic vectors bearing several different antigens to be used as vaccine formulations for a multiantigenic presentation to the mucosa-associated lymphoid tissues.