INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Neisseria meningitidis is a human pathogen for which no fully effective vaccine is available. Vaccines based on the capsular polysaccharides from meningococci of serogroups A, C, Y, and W-135 have been developed. However, the capsular polysaccharide from serogroup B meningococci, the main causative agent of meningococcal disease in many countries, is poorly immunogenic and is cross-reactive with components found in human neural tissue, raising the possibility of autoimmunity (7). Therefore, vaccines based on outer membrane vesicles (OMVs) are currently under investigation for their efficacy in protecting against group B meningococcal disease (4, 36). These vaccines consist primarily of outer membrane proteins (OMPs) of N. meningitidis, in particular the PorA protein, which was shown to be an important target for bactericidal antibodies (31). Lipopolysaccharide (LPS), another major component of OMVs, could play a potential role as immunogen and adjuvant in such a vaccine (38). Unfortunately, LPS must be partly removed from the current vaccines, because the lipid A part of LPS is responsible for reactogenicity, while the similarity of the oligosaccharide chain to host antigens also makes the use of native LPS in vaccine formulations undesirable (20, 40). Therefore, we have investigated the genetics of LPS biosynthesis, with the aim of isolating mutants with modified LPS in which similarity to host antigens no longer exists, endotoxicity is reduced, and adjuvant activity is retained.

Isolation and mutagenesis of the genes involved in the biosynthesis of meningococcal LPS has resulted in a series of LPS mutants with stepwise truncations of the oligosaccharide chain (8, 12-14, 27, 30, 34, 43). These mutants no longer exhibit structures identical to those of the host, making them more suitable for use in vaccine development. The endotoxin activity of LPS can be reduced by chemical or enzymatic modification of the fatty acyl composition of the lipid A part (21, 44). The isolation of the meningococcal lpxD-fabZ-lpxA locus involved in lipid A biosynthesis has opened new possibilities to modify the fatty acyl composition at the level of biosynthesis (29). Our attempts to alter the fatty acyl specificity of the UDP-GlcNAc acyltransferase encoded by the lpxA gene led to the unexpected discovery that it is possible to inactivate this gene while maintaining cell viability, a phenomenon not seen before in gram-negative bacteria (3, 5). In this way, we obtained a completely LPS-deficient meningococcal mutant which had lost all endotoxin activity but still expressed the immunodominant outer membrane proteins in normal amounts (28).

The experiments in this study were designed to assess the immunogenicity of heat-inactivated bacteria or outer membrane complexes (OMCs) of such an endotoxin-free mutant derived from meningococcal strain H44/76. In addition, immunogens derived from this mutant were used in combination with N. meningitidis H44/76 LPS and a series of meningococcal LPS derivatives to elucidate the role of LPS in adjuvant activity. Furthermore, the availability of outer membrane preparations without any LPS allowed us to investigate the enhancement of the immune response by other less toxic adjuvants.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, growth conditions, and inactivation of bacteria. N. meningitidis H44/76 and LPS H44/76 were grown overnight at 37°C on GC medium base (Difco) supplemented with IsoVitaleX in a humid atmosphere containing 5% CO₂ or in liquid medium as previously described (35). Bacteria were heat inactivated in a 56°C water bath for 1 h.

Isolation of OMCs. Meningococci grown in liquid medium were used for the isolation of OMCs by sarcosyl extraction as described by van der Ley et al. (33). The protein content was determined by using the bicinchoninic acid protein assay reagent (Pierce Chemical Co.), with bovine serum albumin as a standard. The protein composition was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (35), and proteins were stained with Coomassie brilliant blue.

LPS and adjuvants. Meningococcal H44/76 wild-type and mutant LPS were isolated by the hot-phenol extraction method described by Westphal and Jann (42). The icsB LPS consisting of lipid A-(KDO)₂-(Hep)₂ · phosphoethanolamine-GlcNAc was previously described by van der Ley et al. (34), and meningococcal rfaC LPS consisting only of lipidA-(KDO)₂ was described by Stojiljkovic et al. (30). In addition, heptose deficiency of rfaC LPS was confirmed in a heptose assay in which heptose was determined by the cysteine-H₂SO₄ reaction (25). Meningococcal htrB1 LPS was analyzed by tandem mass spectrometry to determine its lipid A acyloxyacyl composition (37). Detoxified H44/76 LPS (H44/76 dLPS) was obtained by subjecting wild-type H44/76 LPS to alkaline hydrolysis as described by Myers et al. (22), resulting in the loss of the O-linked 3-OH fatty acids, which was confirmed by gas chromatography-mass spectrometry.

Immunization of mice. (i) Procedure A. BALB/c mice, 5 weeks old (five animals in each group), were immunized on day 0 subcutaneously with 5 × 10⁷ CFU of heat-inactivated bacteria dissolved in 0.5 ml of phosphate-buffered saline (PBS). For the dose-response experiment, mice were immunized on day 0 subcutaneously with 10, 25, or 50 µg of OMCs dissolved in 0.5 ml of PBS. To study the role of LPS, one group of five mice was immunized on day 0 with 10 µg of LPS-deficient OMCs combined with 2.5 µg of LPS dissolved in 0.5 ml of PBS. Immunization was repeated on days 14 and 28 with a double dose of each preparation. The mice were bled on day 42, and sera were collected and stored at 4°C.

(ii) Procedure B. BALB/c mice, 6 to 8 weeks old (five animals in each group), were immunized on days 0, 14, and 28 subcutaneously with 20 µg of LPS H44/76 OMCs supplemented with adjuvant (see Table 3) dissolved in 0.5 ml of PBS. The mice were bled on day 42, and sera were collected and stored at 4°C.

(iii) Procedure C. BALB/c mice, 6 to 8 weeks old (five animals in each group), were immunized on days 0, 14, and 28 subcutaneously with 20 µg of LPS H44/76 OMCs supplemented with 20 µg of Quil A total or 20 µg of Quil A subfraction dissolved in 0.5 ml of PBS. MF59 was mixed 1:1 with 400 µg of LPS H44/76 OMCs per ml dissolved in PBS. A 0.1-ml volume of this antigen-adjuvant emulsion was administered subcutaneously on day 0, and immunization was repeated on days 14 and 28. The mice were bled on day 42, and sera were collected and stored at 4°C.

ELISA. Antibody titers against H44/76 whole cells, purified PorA protein, and LPS were determined by enzyme-linked immunosorbent assay (ELISA) as described elsewhere (1). Flat-bottom 96-well microtiter plates were coated overnight at 37°C with 100 µl of a 2-µg/ml PorA solution in PBS or a 10-µg/ml LPS solution in PBS. Antibody titers were measured for each individual serum. A four-parameter curve fit was made for optical densities values of serial dilutions, and the antibody level was calculated in reciprocal dilutions that gave 50% of the maximum absorbance.

Serum bactericidal assay. The serum bactericidal activity was performed as previously described (9). N. meningitidis H44/76 (B:15P1.7,16:L3,7,9) and H44/76-derived mutant strain HI5, lacking PorA, were used. Sera from mice were heat inactivated for 30 min at 56°C prior to use. Serum samples and bacteria were allowed to incubate for 10 to 15 min at room temperature before the addition of complement. A final concentration of 20% baby rabbit complement was used. The serum bactericidal titer was measured as the reciprocal serum dilution showing more than 90% killing of the number of bacteria used.

Statistical methods. Before statistical analysis, antibody and bactericidal titers were log₁₀ converted, which normalized their distributions. Results are expressed as the mean of log₁₀ titers of five independent observations. Analysis of variance was used for evaluation of statistical data. The significance of the differences between the mean values was determined by the least-significant-difference (LSD) test at a confidence level of 95%.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Immunogenicity of the LPS-deficient H44/76 mutant. Both wild-type and mutant heat-inactivated bacteria and OMCs were used for immunization of BALB/c mice by to procedure A as described in Materials and Methods. Wild-type and mutant OMCs used for immunization showed equal amounts of the PorA, PorB, and RmpM OMPs when analyzed by SDS-PAGE, while the LPS-deficient mutant expressed increased amounts of an Opa protein (Fig. 1). However, the absence of this Opa protein from the wild-type H44/76 test strain used in the ELISA and serum bactericidal assay means that it plays no role in the observed immune response. Sera were analyzed for anti-H44/76 antibodies in a whole-cell ELISA (Table 1). LPS H44/76 heat-inactivated bacteria induced a 10-fold-lower immunoglobulin G (IgG) ELISA titer than did wild-type H44/76 heat-inactivated bacteria. The reduced immunogenicity of the LPS-deficient mutant was even more apparent when OMCs were used for immunization. Sera evoked with 20 µg of LPS H44/76 OMC protein showed an 80-fold-lower ELISA titer than those elicited with the same amount of H44/76 OMC protein.

View larger version (48K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE of OMC proteins from H44/76 (lane 3) and LPS H44/76 (lane 2). Lane 1 contains molecular mass markers of 94, 67, 43, 30, 20.1, and 14.4 kDa.

View this table: [in this window] [in a new window]   TABLE 2.   PorA-specific antibodies of antisera evoked with H44/76 or LPS H44/76 immunogens

Influence of adjuvants on the immune response to LPS-deficient OMCs. BALB/c mice (five mice in each group) were immunized by procedure B (see Materials and Methods) with LPS H44/76 OMC (20 µg of protein) supplemented with adjuvant as listed in Table 3. The sera were analyzed by ELISA with H44/76 whole cells as the coating antigen, and the functional activity of the antibodies was determined in a bactericidal assay against strain H44/76 (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ability of bacterial LPS to function as an adjuvant and enhance the immune response to antigens has been recognized since 1956 (15). The results reported in this paper are consistent with these earlier findings, since the poor immunogenicity of OMCs derived from the recently isolated LPS-deficient meningococcal mutant could be enhanced by the addition of external LPS. However, LPS was not able to enhance the immunogenicity of LPS-deficient heat-inactivated bacteria (data not shown). A possible explanation is that a direct physical interaction of LPS with the antigen is required to establish its function as an adjuvant. Such an interaction is probably easier to accomplish with OMPs in an OMC formulation than in whole cells. It is possible that LPS is able to interact with OMCs but not with whole cells to form a kind of liposomal structure which enhances the immunostimulatory activity of LPS (26).

E. coli LPS, meningococcal H44/76 LPS, and oligosaccharide-truncated LPS derived from the meningococcal icsB and rfaC mutants were able to restore the immunogenicity of LPS-deficient OMCs. Andersen et al. (2) postulated that epitopes on the PorA protein, present in meningococcal OMV vaccines, require the presence of LPS with a certain carbohydrate chain length to induce bactericidal antibodies. In contrast, our data demonstrate that the composition and length of the carbohydrate chain of LPS do not influence the induction of bactericidal antibodies against the immunodominant PorA protein, since the wild-type meningococcal LPS can be replaced with both E. coli LPS and truncated meningococcal rfaC or icsB LPS without reduction of immunogenicity. Moreover, QuilA, subfractions of QuilA, and MF59 were also able to restore the immunogenicity of LPS-deficient OMCs, indicating that LPS is not required per se to elicit bactericidal antibodies to PorA. In agreement with this, Verheul et al. (39) demonstrated that the immunogenicity of purified PorA could be increased with QuilA or the nonionic block polymer L121. Ward et al. (41) and Idänpään-Heikkilä et al. (10) showed that it is possible to reconstitute denatured PorA by incorporating it into liposomes, without using LPS. In addition, PorA was reconstituted in the presence of Zwittergent or Triton X-100 (11). Both methods resulted in the expression of native PorA epitopes able to induce functional antibodies in the absence of LPS. Since the PorA in LPS-deficient immunogens did not elicit high proportions of antibodies against denatured epitopes, we conclude that it still at least resembles its native conformation and that the reduction in immunogenicity is exclusively related to the absence of LPS adjuvant activity.

The adjuvant activity and the toxicity of LPS have been established to reside in its lipid A moiety. The amount of esterified and amidated fatty acids in lipid A, as well as its phosphorylation pattern, determine activity (16, 32). This is clearly reflected in the reduced adjuvant activity found with R. sphaeroides LPS, MPL, and alkali-hydrolyzed meningococcal LPS, which all have incomplete lipid A structures and as a result are less toxic than meningococcal wild-type LPS. Interestingly, no reduction in adjuvant activity compared to H44/76 LPS was seen when meningococcal htrB1 LPS, missing one of the acyloxyacyl groups in its lipid A, was used. The biological activity of this mutant LPS is still under investigation, but htrB mutants of Salmonella typhimurium and Haemophilus influenzae have already been shown to express reduced endotoxic activity (17, 23). Retained adjuvant activity combined with reduced toxicity would obviously make the htrB1 mutant LPS an interesting vaccine component.

The moderate adjuvant activity found with R. sphaeroides LPS, MPL, and detoxified meningococcal LPS was also apparent when AlPO₄ was used. The subclass distribution of the antibodies elicited with these less active adjuvants showed a relatively large amount of IgG1 antibodies, whereas predominantly IgG2a antibodies were found with E. coli LPS, meningococcal H44/76 and truncated LPS, Quil A, and subfractions of Quil A as adjuvants. The mouse IgG1 antibodies are thought to be less active in complement activation than are antibodies of the IgG2a isotype, which might explain the lower bactericidal activity of the sera with a relatively high IgG1 content (6). Still, a high proportion of the non-complement-fixing IgG1 antibodies does not always lead to reduced bactericidal activity, since sera evoked with LPS-deficient OMCs and MF59, containing relatively high levels of IgG1 antibodies, also showed high bactericidal activity. Apparently, the bactericidal activity of an antiserum is not exclusively related to the amount of non-complement-fixing antibodies but probably depends on the complete reportoire of quantity, affinity, isotype, and epitope specificity of the antibodies induced (18).

In conclusion, our results demonstrate that the poor immunogenicity of meningococcal LPS-deficient OMCs can be restored with adjuvants less toxic than wild-type LPS. In this respect, QA3, which has a low acute systemic toxicity and a hemolytic activity of about 10% compared to total Quil A, and MF59, which is known to be generally well tolerated in humans (24), are very interesting candidates. Furthermore, meningococcal htrB1 LPS might be of great interest for vaccine development if it does turn out to have an improved ratio of adjuvant activity to toxicity. Of course, it remains to be established whether the results reported here for mice can be fully extrapolated to humans. In any case, the availability of an LPS-deficient mutant allows the easy replacement of wild-type LPS with other, improved adjuvants more suitable for use in humans.