Local and Systemic Antibody Responses in Mice Immunized Intranasally with Native and Detergent-Extracted Outer Membrane Vesicles from Neisseria meningitidis

Abstract

The mouse humoral immune response toward native or detergent-extracted outer membrane vesicles (NOMVs and DOMVs, respectively) from Neisseria meningitidis was determined after intranasal immunization. Both preparations elicited high frequencies of NOMV-specific antibody-forming cells (AFCs) locally in the nasal associated lymphoid tissue (NALT) after three or four weekly doses. The diffuse NALT (D-NALT) contained ca. 10-fold more NOMV-specific AFCs than those observed in the mediastinal lymph node, spleen, and bone marrow. AFCs observed in the D-NALT were primarily immunoglobulin A positive (IgA⁺) and were maintained for at least 1 month. In contrast, the organized NALT (O-NALT) contained low numbers of AFCs, and the response was relatively short-lived. In other lymphoid tissues, AFCs producing various IgG subclasses and IgM were present with IgG2b-producing AFCs being dominant or codominant with IgA or IgG2a. In serum and in all of the tissues examined, with the exception of the NALT, NOMVs clearly induced a stronger antibody response and a broader range of antibody isotypes than DOMVs. The development of NOMV-specific AFCs in spleen and bone marrow after intranasal immunization was slow compared to intravenous immunization but, once established, the intranasally elicited responses increased steadily for at least 75 days. NOMV-specific antibodies induced via several routes of immunization had high bactericidal activities in serum. Our results indicated that intranasally administered OMVs induced strong local and systemic antibody responses in mice that were relatively long-lived.

The human nasopharynx is the only natural niche for the mucosal commensal Neisseria meningitidis. Approximately 10% of the normal population are healthy carriers of N. meningitidis (23). Infrequently, meningococci penetrate the mucosal barrier and cause disseminated meningococcal disease, which remains a serious health problem worldwide. The clinical symptoms range in severity from a mild sore throat to acute meningococcemia, which if left untreated can rapidly lead to circulatory collapse, multiple organ dysfunction, and eventually death. The most common presentation, however, is acute purulent meningitis. Meningococcal disease mainly affects infants and teenagers. The disease rate is normally very low among individuals above 25 years of age. Natural immunity toward meningococcal diseases is thought to be acquired after asymptomatic colonization of the nasopharyngeal mucosa by meningococci. The precise mechanisms involved in the induction of immunity to meningococci are as yet undefined, but protective immunity correlate strongly with the induction of serum antibodies with bactericidal and/or opsonophagocytic activity (24).

A universal vaccine for meningococcal diseases caused by serogroup B N. meningitidis is currently unavailable due to the poor immunogenicity of its polysaccharide capsule and the antigenic variability of noncapsular surface components of meningococci (41). Serogroup B vaccines based on detergent extracts of meningococcal outer membrane vesicles (DOMVs) have been used in several countries, but the efficacy of intramuscularly administered DOMV vaccines was variable, and DOMV-induced bactericidal antibodies were strain specific (10, 11, 42, 46). Since the nasopharynx is the only natural habitat of meningococci, intranasal (i.n.) immunization with meningococcal antigens has been suggested to be an effective way of inducing both mucosal and systemic immunity. Recent studies of i.n. administered OMVs in mice and humans have provided support for this strategy. Some studies have shown that i.n. immunizations with DOMV vaccines induce long-lasting elevated levels of serum bactericidal antibodies (SBA) in humans (27; M. Fischer, M., J. Holst, I. S. Aaberge, I. L. Haugen, J. L. Burns, B. A. Perkins, and B. Haneberg, 12th Int. Pathogenic Neisseria Conf., abstr. 113, 2000), albeit the proportion of vaccinees with a ≥4-fold increase in bactericidal titers was only between 18 and 40% (Fischer et al., 12th Int. Pathogenic Neisseria Conf.). The safety and immunogenicity of native OMVs (NOMVs) administered i.n. in humans have also been demonstrated (19, 34). NOMVs are outer membrane material shed from meningococci during growth that contain relatively large quantities of lipopolysaccharide (LPS; 25 to 50% by weight relative to protein) compared to DOMVs (5 to 8% LPS). Despite the high level of LPS in NOMVs, these preparations have been well tolerated by humans immunized i.n. (19, 34). However, studies in humans have also shown that DOMVs induced significantly lower SBA levels when administered i.n. than via the intramuscular route (27). Therefore, the effectiveness of mucosal OMV-based vaccines needs to be improved. Determining where and how local and systemic immune responses develop after i.n. immunizations would aid in the assessment and design of mucosal meningococcal vaccines.

The nasal associated lymphoid tissues (NALT) play an important role in local immune responses in the upper respiratory tract. In mice, and other rodents, the NALT is divided into the organized and diffuse NALTs (O-NALT and D-NALT, respectively) (6, 7, 35). O-NALT, which has been described as the equivalent of Waldeyer's ring in humans, is the only well-organized mucosal associated lymphoid tissue in the upper respiratory tract. It consists of paired lymphoid cell aggregates located between the columnar epithelium and the palate. O-NALT is an inductive site with similarities to the Peyer's patches, although these two tissues differentially express certain addressins on their high endothelial venules (15). In contrast, D-NALT is composed of lymphoid tissue lining the nasal passages. D-NALT has been described as an effector site equivalent to the lamina propria of the gut. Responses in tissues other than the NALT, including lymph nodes, spleen and bone marrow, are also of importance upon i.n. immunization as they may contribute to the systemic response. Analysis of immune responses after i.n. immunizations or infections in the upper respiratory tract generally show that the magnitude, kinetics, localization, and longevity of the responses vary with the antigens or microorganisms, immunization protocols, and mouse strains used (28, 34, 53, 54). In the present study we examine the characteristics of humoral responses in various sites, including the NALT, lymph nodes, spleen, and bone marrow after immunization with DOMVs and NOMVs by various routes. We describe for the first time the salient features of the NALT humoral responses toward meningococcal OMVs after i.n. immunization, and we discuss factors that can influence the local and systemic antibody responses induced by OMVs.

(This study was presented in part at the 12th International Pathogenic Neisseria Conference in Galveston, Tex., in September 2000.)

MATERIALS AND METHODS

Bacterial strains and growth conditions.

N. meningitidis mutant-4 (Mu-4), a variant of wild-type strain 44/76 that expresses LPS with a complete inner core of sugars residues only, has been previously described (3, 12, 31). Meningococci were grown on brain heart infusion (BHI) agar plates, supplemented with 1% horse serum, in a 5% CO₂ atmosphere at 37°C. When meningococci were grown in liquid culture, 800 ml of BHI broth with 1% horse serum in a 2-liter conical flask was inoculated with a meningococcal suspension harvested from two BHI agar plates. The bacteria were then incubated at 37°C overnight with constant shaking at 150 rpm. Standard gram-staining procedures were performed to check cultures for possible contamination.

Preparation of NOMVs and DOMVs.

NOMVs were prepared from liquid cultures of Mu-4 as described above. Bacterial cells were collected by centrifugation, and the supernatant was concentrated 10-fold by ultrafiltration (filtration units with a 500-kDa cutoff; A/G Technology Corp., Needham, Mass.). Cell debris was removed from the concentrate by centrifugation (10,000 × g for 30 min). NOMVs were isolated by ultracentrifugation (twice at 100,000 × g for 4 h) in order to remove medium components. The NOMV pellet was resuspended in 0.2-μm (pore-size)-filtered deionized water, divided into aliquots, and stored at −80°C. All preparations were tested for sterility prior to use. DOMVs were prepared by detergent extraction as previously described (22).

Mice.

Female C57BL/6 and BALB/c (6 to 8 weeks old) mice were obtained from Charles Rivers Laboratories. Mice were maintained under specific-pathogen-free conditions in the animal care facilities of the Institute of Animal Health, Compton, Berkshire, United Kingdom. Experiments on animals were carried out according to the guidelines of the United Kingdom Home Office and the regulations of the Animals (Scientific Procedures) Act of 1986.

Immunizations and collection of samples.

Mice were immunized i.n. and intravenously (i.v.) with NOMVs or DOMVs as indicated in the text. Mice immunized i.n. were first anesthetized with a ketamine-xylazine mixture (100 μl) by the intraperitoneal (i.p.) route. Blood and tissue samples were withdrawn on days 11 or 12 and days 30 or 32 unless otherwise indicated in the text.

Preparation of lymphocytes.

Cells from the O- and D-NALTs were extracted as previously described by Asanuma et al. (6, 7). Cell suspensions were prepared from the O-NALT, mediastinal lymph nodes (MLN), and spleen by gently pressing the tissues between frosted glass slides. Bone marrow cells were prepared by flushing marrow of bare bones from hind limbs with Iscove modified Dulbecco medium (IMDM) with a 1-ml syringe with a 25-gauge needle. The cells were dispersed in IMDM or IMDM supplemented with 5% (vol/vol) fetal bovine serum (FBS), 100 U of penicillin/ml, and 0.1 mg of streptomycin/ml (cIMDM). Cell suspensions were then filtered through 35-μm-pore-size nylon gauze (BDH, Leicester, United Kingdom) and washed twice with cIMDM (400 × g for 10 min at room temperature). Red blood cells were lysed by incubating cell suspensions for 5 min at room temperature in the presence of red blood cell lysis buffer (Sigma). Cells were pelleted by centrifugation into a 1-ml cushion of FBS, washed twice in cIMDM as described above, and kept on ice until plated out for enzyme-linked immunospot (ELISPOT) assay analysis as described below.

Lung fragment cultures.

Antigen-specific antibodies in lung tissues were detected by a modification of a method used to assay antibody secreted from small intestine segments (14). Briefly, after removal of the lungs under sterile conditions, 1-mm duplicate slices of equivalent size were cut from the middle of one lobe and placed in individual wells of a 24-well plate with 1 ml of cIMDM containing 1% Amphoteracin (Gibco, Paisley, United Kingdom). Samples were incubated for 5 days at 37°C in an atmosphere of 95% O₂ and 5% CO₂. Supernatants were then removed from each well and assayed for the presence of specific antibodies by enzyme-linked immunosorbent assay (ELISA).

ELISA.

Microtiter plates (i.e., flat-bottom 96-well Maxisorp plates; Nunc, Roskilde, Denmark) were coated overnight at 4°C with 100 μl of NOMVs per well (4 μg of protein/ml) dissolved in 50 mM sodium carbonate-bicarbonate buffer (pH 9.6). Unoccupied protein-binding sites were blocked with 150 μl/well of 0.5% (wt/vol) bovine serum albumin in phosphate-buffered saline (PBS) for 1 h. Primary antibodies were incubated overnight at 4°C and alkaline phosphatase-conjugated goat anti-mouse immunoglobulin isotype-specific or subclass-specific antibodies (used at 1:1000; Southern Biotechnology Associates, Inc., Birmingham, Ala.) were incubated for 2 h at room temperature. Between each incubation step, the plates were washed five times with PBS containing 0.02% Tween 20. The substrate, p-nitrophenylphosphate (Sigma, St. Louis, Mo.), dissolved at 1 mg/ml in a 0.1 M Tris-HCl buffer (pH 9.8) containing 10% (vol/vol) diethanolamine, 0.1 M NaCl and 5 mM MgCl₂] was added at 100 μl per well. The enzymatic reaction was terminated after 30 min by the addition of 50 μl/well of 1 M NaOH. The optical density (OD) was measured at 405 nm (Spectra MAX 340 ELISA plate reader; Molecular Devices Corp., Sunnyvale, Calif.). The results are expressed as log₂ reciprocal endpoint titers. A cutoff for positive OD values was calculated as the average (arithmetic mean) plus three standard deviations of all dilutions from nonimmunized control mice (n = 3). The endpoint titer was defined as the reciprocal value of the first dilution below the cut off.

ELISPOT assay.

Antibody-forming cells (AFCs) were detected by ELISPOT assay as previously described (16, 43). Microtiter plates (96-well nitrocellulose-based filtration plates; Millipore, Watford, United Kingdom) were coated with NOMVs (100 μl per well of 5 μg of protein/ml of 0.2-μm [pore-size]-filtered PBS) and were incubated at 4°C overnight in a humidified chamber. Plates were washed three times with 200 μl of PBS before unoccupied protein binding sites were blocked with 2% (vol/vol) FBS in PBS for 1 h at room temperature. Plates were then washed as previous, and then 100-μl volumes of lymphocyte suspensions (10⁴ to 10⁵ cells/ml) were added to the plates, and these were incubated overnight at 37°C in a humid 5% CO₂ atmosphere. Plates were washed successively once with 200 μl of PBS, once with 200 μl of PBS plus 0.1% Tween 20, and three times with 200 μl of PBS. Goat anti-mouse immunoglobulin isotype-specific or subclass-specific antibodies conjugated with alkaline phosphatase (diluted 1:500) in 5% (wt/vol) bovine serum albumin-PBS were used to detect plaques. The secondary antibodies were incubated for 3 h at room temperature. Plates were then washed as described above before incubating them with the substrate BCIP (5-bromo-4-chloro-3-indolylphosphate; Sigma) dissolved at 1 mg/ml in 0.1 M Tris-HCl (pH 9.5), containing 10% (vol/vol) diethanolamine, 0.1 M NaCl, and 5 mM MgCl₂. The enzymatic reaction was stopped, after 30 min incubation at room temperature, by washing the plates four times with PBS. Plaques, which are indicative of AFCs, were enumerated by using a direct light microscope in conjunction with an overhead light source.

Serum bactericidal assays.

Assays were performed as previously described (42). Briefly, a bacterial inoculum of 70 to 120 CFU was incubated in the presence of 25% human complement in microtiter plates. Twofold dilution series (starting at 1:8) for nonimmunized or immunized sera were tested for SBA against wild-type strain 44/76. The results are presented as the SBA titer, which is defined as reciprocal of the highest dilution in serum causing more than 50% killing of the target strain.

Statistical methods.

Statistical significance between groups was determined by the Student t test if the data passed the normality test. The Mann-Whitney test was used for data that failed the normality test.

RESULTS

Large numbers of NOMV-specific AFCs were found in D-NALT.

In the present study, OMVs prepared from the LPS mutant Mu-4, and not from the parent strain 44/76, were used as the immunogen because we also wanted to study the generation of anti-LPS inner-core-specific antibodies (5). The quantity and quality of antibodies to major outer membrane proteins, except class 4 outer membrane proteins, induced by OMVs prepared from strain 44/76 and Mu-4 have previously been found to be very similar (4). Mice were immunized i.n. with 15 μg (protein content) of NOMVs or DOMVs in a 50-μl volume. Mice were immunized either once a week for 4 weeks with NOMVs or DOMVs (group 1) or given two doses of NOMVs 1 week apart and a third dose 28 days later (group 2). Both NOMVs and DOMVs induced NOMV-specific AFCs in all tissues examined, except in the O-NALT, as quantified by ELISPOT assay (top panels in Fig. 1A and B, respectively). The anti-NOMV response was high both at the early (day 11 or 12 for groups 1 and 2, respectively) and the late (day 30 or 33 for groups 1 and 2, respectively) time points. In both groups of mice, the frequency of NOMV-specific AFCs in the D-NALT (450 to 700 AFCs per 5 × 10⁵ cells) was up to 10-fold higher than that of the MLN, spleen, and bone marrow.

FIG. 1.

Frequencies and isotypes of NOMV-specific AFCs in O-NALT, D-NALT, MLN, spleen, and bone marrow after i.n. immunization with NOMVs or DOMVs. In group 1, C57BL/6 mice (n = 3 per group) were immunized with either NOMVs or DOMVs (15 μg of protein/dose in 50 μl of PBS) on days 0, 7, 14, and 21. In group 2, C57BL/6 mice (n = 3) were immunized on days 0, 7, and 32 with NOMVs (15 μg of protein/dose in 50 μl of PBS). Single cell suspensions from mice in each group were pooled and plated out on ELISPOT plates coated with NOMVs.

In group 1 mice, the frequency of AFCs in the D-NALT at the early time point after NOMV immunization was higher than that found at the late time point, whereas the opposite was evident in group 2 mice (Fig. 1, top panels). This indicates that the frequency of NOMV-specific AFCs detected in the D-NALT fluctuates significantly. Consequently, it is difficult to estimate or establish when the response peaked, or whether the NOMVs or the DOMVs were more potent in the induction of anti-NOMV responses in the D-NALT. However, the results show clearly that i.n. immunization, with either NOMVs or DOMVs, was able to induce very high frequencies of AFCs that were maintained in the D-NALT for at least 30 days. In contrast, the antibody response in the O-NALT was weak and short-lived. The frequencies of NOMV-specific AFCs in the spleen and bone marrow indicate that NOMVs induced a stronger and more prolonged antibody production than DOMVs under the current immunization protocol.

The number of immunoglobulin A (IgA)-producing AFCs in the D-NALT was ∼10-fold higher than the number of AFCs producing the two other isotypes, IgG2a and IgG2b (Fig. 1, middle and bottom panels). In contrast, the response in the MLN, spleen, and bone marrow showed a significantly higher proportion of AFCs producing the various IgG subclasses and IgM. In these three tissues, at both the early and late time point postimmunization, the IgG2b-producing AFCs dominated over IgG2a-specific AFCs in majority of the mice, which in turn dominated over IgG1- and IgG3-specific AFCs.

NOMVs induced a stronger antibody response and a broader range of immunoglobulin isotypes than DOMVs in serum and lung tissue.

Serum pools and individual lung fragment culture supernatants (LFCs) from mice in groups 1 and 2 were analyzed by ELISA with NOMVs as the capturing antigen. Serum pools from mice immunized i.n. with NOMVs had higher levels of NOMV-specific antibodies than those from mice immunized i.n. with DOMVs. Serum pools from the latter group contained only significant titers of IgG2b. In contrast, serum pools from NOMV-immunized mice contained relatively high titers of IgG2b in addition to intermediate titers of IgG1 and IgG2a. In these serum pools the IgG1 titers dominated over IgG2a (Fig. 2). Sera from nonimmunized mice did not contain detectable levels of antibodies to NOMVs (data not shown).

FIG. 2.

Isotype and subclass distribution of NOMV-specific antibodies in serum after i.n. immunization with NOMVs or DOMVs. Mice (C57BL/6, n = 3 per group) were immunized as described in Fig. 1. Sera from each group were withdrawn on day 11 or 12 and day 30 or 33, pooled, and assayed for NOMV-specific antibodies by ELISA. The results are expressed as the OD at 405 nm.

Isotypes of NOMV-specific antibodies found in LFCs were similar to those in sera at days 11 and 30 for group 1 (Fig. 3a and b) and at days 12 and 33 for group 2 mice (Fig. 3c and d). The results are given as individual log₂ endpoint titers from three i.n. immunized mice per group. Mice immunized with NOMVs had significantly (P = 0.006) higher antibody (total immunoglobulin) levels in the lungs than mice immunized with DOMVs at day 30. The dominating NOMV-specific immunoglobulin isotype in LFCs appeared to be IgG2b, followed closely by IgA. LFCs from mice immunized with DOMVs contained only low or undetectable levels of other IgG subclasses. In contrast, LFCs from group 1 NOMV-immunized mice also contained significant quantities of IgG2a, followed by IgG1. IgM was also detected in LFCs from this group (Fig. 3a and b).

FIG. 3.

NOMV-specific antibodies in lung fragment culture supernatants after i.n. immunization with NOMVs or DOMVs. Mice (C57BL/6, n = 3 per group) were immunized as described in Fig. 1. Lung fragments from individual mice (numbered 1 to 3) were collected on day 11 or 12 and day 30 or 33 after initial immunization with NOMVs or DOMVs. Endpoint titers were defined as the first dilution to give an OD value below the cutoff value. The cutoff value was defined as the mean OD of all of the dilutions of a negative control (three nonimmunized mice) plus three standard deviations. n.d., Not determined.

The kinetics of NOMV-specific AFCs in spleen and bone marrow after i.n. and i.v. immunization with NOMVs differed markedly.

In order to study the kinetics of the development of NOMV-specific AFCs in primary and secondary lymphoid tissues, spleen, and bone marrow were analyzed after i.n. immunization of BALB/c mice with 15 μg of NOMVs in 30 μl of PBS once a week for 4 weeks. For comparison, control groups were immunized i.v. on days 0 and 21 with 5 μg of NOMVs in 100 μl of PBS, and the antibody response was monitored in parallel (Fig. 4). Sera from two other groups of BALB/c mice, subjected to the same immunization protocols, were analyzed for anti-NOMV antibodies on day 28, 7 days after the last dose (Fig. 5). After two i.n. doses of NOMVs, the mice had low but detectable levels of NOMV-specific AFCs (5 per 5 × 10⁵ cells to 35 per 5 × 10⁵ cells) in both spleen and bone marrow as early as day 14 (Fig. 4a and b, respectively). The frequencies of AFCs producing antibodies of IgG subclasses were relatively constant throughout the study despite two additional immunizations on day 14 and day 21. In contrast, IgA-producing AFCs were not detected until day 23, but from then on the level increased significantly in the bone marrow throughout the 75-day observation period (Fig. 4a and b).

FIG. 4.

Development of AFCs producing NOMV-specific antibodies in spleen and bone marrow after i.n. or i.v. immunization with NOMVs. Mice (BALB/c) were immunized i.n. with 15 μg of NOMVs in 30 μl of PBS under anesthetic on days 0, 7, 14, and 21 (a and b) or i.v. with 5 μg of NOMVs in 100 μl on days 0 and 21 (c and d). Single cell suspensions of spleens (a and c) and bone marrow (b and d) from mice (n = 3 per time point assayed) were collected and analyzed individually by ELISPOT assay with NOMVs as the capturing antigen. The results are expressed as the mean number of AFCs with the standard deviations.

FIG. 5.

Isotype and subclass distribution of anti-NOMV serum antibodies on day 28 after i.n. (a) or i.v. (b) immunization. Mice (BALB/c, n = 3 per group) were immunized as described in Fig. 4. Endpoint titers and cutoff values were defined as described for Fig. 3.

A single i.v. dose of NOMVs (5 μg) elicited comparable levels of AFCs in the spleen at day 10 to that of two weekly i.n. doses (15 μg) at day 14 (Fig. 4a and c). After the i.v. booster dose given on day 21, the frequency of AFCs in the spleen increased dramatically (up to 250 per 5 × 10⁵ cells) 2 days later (i.e., day 23), followed by a fivefold decrease on day 28 (Fig. 4c). After the booster immunization the frequencies of AFCs producing IgG1, IgG2a, and IgG2b were slightly higher than that of IgG3-producing AFCs. A similar IgG subclass distribution was also found in the bone marrow from i.v. immunized mice, even though detectable levels of NOMV-specific AFCs were only apparent from day 35 and onward (Fig. 4d). Furthermore, significant levels of IgA-producing AFCs were not detected in spleen or bone marrow after i.v. immunization with NOMVs (Fig. 4c and d).

Serum samples from the i.n. immunized group collected on day 28, 7 days after the last dose, showed that the subclasses IgG2a and IgG2b dominated over IgG1 and IgG3 (8-16-fold higher endpoint titer). A relatively lower endpoint titer for IgA and an insignificant titer for IgM were also detected. In contrast, sera from the i.v. immunized group contained equal levels of IgG1, IgG2a, and IgG2b; slightly lower levels of IgG3; and undetectable levels of IgA and IgM (Fig. 5).

Immunization with meningococcal OMVs induced SBA.

Protection against meningococcal diseases correlates with the presence of SBA. Figure 6 shows the bactericidal activities of sera from mice immunized with 0 to 3 doses of NOMVs via the i.n. or i.v. route. We found that i.v. immunization with a single dose of NOMVs generated SBA titers of 1:1,024, and this was further increased to more than 1:8,192 with a second dose given 3 weeks later. The SBA titers for the i.p. immunized mice were almost identical to those obtained for the i.v. immunized mice (data not shown). In contrast, i.n. immunized mice required at least three weekly doses of NOMVs to generate SBA titers of 1:1,024. Mice immunized i.n. with one or two doses did not produce SBA titers more than the control mice whose SBA titers were 1:8 or less. Although sufficient amounts of sera for both ELISA and serum bactericidal assay were not obtained from every immunized group of mice used in the present study, it was clear that NOMV-specific antibodies generated via all three routes were highly bactericidal. The highest levels were found in the group immunized i.v.

FIG. 6.

Bactericidal activity of sera from NOMV-immunized mice. The effect of the immunization route was examined with mice immunized either i.n. (▪) or i.v. (□) with zero to three doses of NOMVs. Mice were immunized i.n. once a week for a maximum of 3 weeks, and sera were collected 7 or 14 days after the last dose for serum bactericidal assays. Mice immunized i.v. with one dose were bled 3 weeks later, and those immunized with two doses 3 weeks apart were bled 7 days after the second dose for serum bactericidal assays.

DISCUSSION

Antimeningococcal antibodies play a central role in the protection against diseases caused by N. meningitidis (24). Although intramuscularly injected vaccines based on capsular polysaccharides are effective in protecting meningococcal diseases due to serogroups A, C, W135, and Y, the only currently available vaccines for serogroup B meningococci are based on DOMVs. i.n. immunization is a practical alternative route to parenteral immunization for inducing immunity against this mucosal pathogen, and OMVs have documented acceptable immunogenicity and safety profiles in humans (19, 27, 34). However, local and systemic antimeningococcal antibody responses in humans could not be readily studied in detail. The main aim of our study is to define the characteristics of the local (i.e., the NALT) and systemic antimeningococcal antibody responses in mice after immunization with NOMVs and DOMVs by various routes. Our results revealed that the magnitude, immunoglobulin isotype, and kinetics of humoral responses toward meningococcal antigens are dependent on the route of immunization and on the nature of the OMVs. We found that i.n. immunization with either meningococcal DOMVs or NOMVs, which are particulate antigens, elicited NOMV-specific AFCs in various lymphoid tissues. The highest frequency of AFCs was found in the D-NALT, and more modest responses were observed in other tissues such as the spleen, MLN, and bone marrow. In contrast, the antibody response in the O-NALT was short-lived, and the frequencies of NOMV-specific AFCs were low or undetectable. This observation is in accordance with other reports indicating that the O-NALT is the inductive site and the D-NALT is the effector site of the nasal tissues (51, 52, 54). Whether or not IgA-producing AFCs can originate from lymphoid cells associated with mucosal tissues other than the O-NALT was not determined in the present study.

IgA responses play an important role in mucosal defense against pathogens via a number of potential mechanisms (36). The role of IgA in meningococcal infection and disease is unclear. At the nasopharynx, IgA may mediate antibody-dependent cellular cytotoxicity (38) against meningococci in addition to its general role in providing an immune barrier against microbes at mucosal surfaces (36). Antimeningococcal specific IgA, a non-complement-fixing isotype, also has the potential to block complement-mediated lysis of meningococci by IgG or IgM in serum by competing for the same antigens (26). The present study shows that significant levels of IgA-producing AFCs were induced in various tissues with NOMVs or DOMVs after i.n., but not i.v., immunization. The insignificant or very low levels of IgA detected in serum upon i.n. or i.v. immunization are possibly due to the fact that rodents have very efficient polymeric immunoglobulin receptors that rapidly transport IgA from the circulation to mucosal surfaces (44). The IgA/IgG ratio observed in the D-NALT was much higher than in the O-NALT and other lymphoid tissues studied. Mucosal B cells seem to have a preference for immunoglobulin class switch recombination from μ to α, probably through stimulation with transforming growth factor β, interleukin-10, and CD40, as indicated by a number of in vitro studies (13, 18, 33, 40, 50). The mechanism for preferential class switching to IgA is not fully understood, but the O-NALT was recently proposed to possess a unique machinery that provides an enrichment of high-affinity IgA- but not IgG-specific cells in the memory compartment (45). The same study also showed that the frequency of IgG2b-expressing cells in the O-NALT which peaked at day 7 and persisted up to day 11, was about three to four times the number of IgA-producing cells (45). In contrast, our study showed that either only IgA-producing AFCs or equal numbers of IgA- and IgG2b-producing AFCs, were found in the O-NALT on day 11 or 12, and no specific AFCs were detected on day 30 or 33. Our results are consistent with a recent report on the isotype profile and frequency of IgA- and IgG2b-producing cells in the NALT after influenza virus infection (37).

The fate of NOMV-specific IgA-producing AFCs was not determined in the present study, but it has been shown that IgA-producing cells activated in the O-NALT were subsequently found in the posterior cervical lymph nodes and in other nonmucosal lymphoid tissues (45). NOMV-specific IgA-producing AFCs were found in significant numbers in the spleen and bone marrow after i.n, but not i.v., immunization. Intriguingly, these cells increased in numbers in the bone marrow but not in the spleen and persisted in both tissues for at least 75 days. These AFCs may contribute to the long-term mucosal and systemic antibody response against meningococcal antigens, since virus-specific long-lived plasma cells found in spleen and bone marrow have been shown to contribute significantly to the long-term systemic antibody response after influenza virus infection in mice (32, 48). Investigation into the source(s) and fate of NOMV-specific IgA AFCs in the NALT would provide further insights into the generation and maintenance of local antibody responses after i.n. immunization with NOMVs.

Both NOMVs and DOMVs elicited very high, and almost equal, frequencies of AFCs in the D-NALT after i.n. immunization, but striking differences in AFC frequencies between these OMVs were observed among the other lymphoid tissues investigated. NOMVs induced a stronger antibody response than DOMVs in the MLN, spleen, bone marrow, lung, and serum. Also, NOMVs induced a broader range of the IgG subclasses, suggesting that they were more potent inducers of B-cell class-switching events than their DOMV counterparts. It is noteworthy that these differences were only observed when the i.n. administered volume was no less than 50 μl (unpublished results). A previous study on the effect of i.n. administered volume on the distribution of microspheres has indicated that the use of 50 μl but not 10 μl given i.n. resulted in the deposition of microspheres at sites other than the nasal cavity (20). Overflow of OMV antigens in the nasal cavity is a possible means by which cells from other immunological sites, such as the lung, were exposed and responded to i.n. administered NOMVs or DOMVs. Our results suggest that the generation of a strong systemic NOMV-specific antibody response by NOMVs via the i.n. route requires the involvement of lymphoid tissues other than the NALT. Similarly, Katial et al. have found that humans immunized both i.n. and oropharyngeally developed higher levels of both IgG and IgA in the nasal fluid, even though the SBA titers from these vaccinees were not different from those who were only immunized i.n. (35). The use of large volumes to enhance systemic antibody responses elicited by i.n. administered OMV vaccines must proceed with caution since excessive inflammation in the lower respiratory tract has been reported for other antigens and adjuvants (29, 47).

The main difference between NOMVs and DOMVs is that the latter contain ∼5-fold less LPS than the former due to the detergent extraction process used (49). The less-pronounced difference in the D-NALT responses between mice immunized with NOMVs and DOMVs could be explained by the fact that cells associated with the mucosal surface are less susceptible to the biological activity of LPS. Human cervicovaginal and intestinal epithelial cells have recently been found to be hyporesponsive to LPS due to the absence or low-level expression of MD-2 and Toll-like receptor 4 (TLR-4), two pivotal components of LPS-mediated cell signaling (1, 21). Although the levels of MD-2 and TLR-4 on cells in the respiratory tract have not been determined, LPS hyporesponsiveness could be one of the main reasons for the lack of overt or harmful side effects in human volunteers after i.n. immunization with NOMVs (34). Both NOMVs and DOMVs have components with immunostimulatory effects on local and systemic immune responses since strong antibody responses were induced by OMVs in the absence of any added adjuvant (17). Neisserial porins, which are major components in both preparations, have been found to stimulate B cells and upregulate the surface expression of costimulatory B7-2 (CD86) molecules through a TLR-2- and MyD88-dependent signaling mechanism (39). Interestingly, in contrast to TLR-4, TLR-2 seems to be expressed by cells from various mucosal sites (1, 8, 21). Thus, activation of TLR-2-mediated signaling pathways by non-LPS meningococcal components may be required for the induction of a vigorous antibody response in the NALT. In contrast to its apparent lack of endotoxic effect on the NALT, LPS can act as an adjuvant upon parenteral administration and induce cytokines that may influence the immunoglobulin isotypes and subclasses produced (2). It is not known whether non-LPS meningococcal components have similar influence on antibody production.

NOMV-specific antibodies generated by i.n. immunization with OMVs are bactericidal, with high titers of ca. 1:1,024. This represents a 128-fold increase over nonimmune mouse serum controls. However, it is difficult to compare SBA titers from OMV-immunized mice and humans because SBA titers of 1:4 or greater have been considered to be protective in humans (24, 25, 30). Human vaccine trials have also shown that both NOMVs and the DOMV vaccines induced only modest increases in systemic antibody responses and two- to fourfold increases in SBA titers upon i.n. immunization (19, 27, 34). Nevertheless, the mouse immune responses to i.n. administered OMVs suggest similarity to a phenomenon observed in OMV-immunized humans. In mice primed i.n. with four weekly doses of the Norwegian DOMV vaccine, SBA titers did not increase after a new round of i.n. immunizations (9). We have previously made similar observations when mice were immunized with NOMVs (unpublished results). Human adults with moderately high basal levels of SBA titers of 1:8, due most likely to asymptomatic carriage of meningococci in the nasopharynx, do not respond to i.n. OMV vaccines by increasing SBA titers significantly. Interestingly, a significant increase in SBA titer was found in mice primed i.n. and boosted subcutaneously with one dose of DOMV vaccine (9). Whether humans would benefit from this immunization schedule for OMV vaccines remains to be determined.

The present study has identified the D-NALT as the major site of antimeningococcal AFC production after i.n. immunization with OMVs and showed that the kinetics, magnitude, and immunoglobulin isotype of the systemic antibody response were dependent on the form of OMVs being used and the immunization route. With the exception of the NALT where IgA dominated the isotype of OMV-induced antibodies, NOMVs were better than DOMVs at inducing higher antibody levels and class switching to complement-fixing isotypes. The i.n. route required several doses of NOMVs to generate antimeningococcal systemic antibody responses equivalent to one dose of NOMVs via the i.v. or i.p. route. Although it was more difficult to initiate local and systemic anti-NOMV antibody responses via the i.n. route, the established antibody responses were relatively long-lived, and presumably these antibodies are protective as indicated by their SBA titers and immunoglobulin isotypes. In addition to OMV formulation and immunization route, it is important to determine other factors that influence the initiation of antibody responses against meningococcal antigens in the NALT. It will also be important to identify the immunological tissues other than the NALT that are required to be activated in order to achieve a strong systemic immune response in humans. A better understanding of the underlying mechanisms would thus be critically important not only in the development of improved OMV-based vaccines but also in the design of novel mucosal vaccines for meningococcal diseases.