INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Moraxella (Branhamella) catarrhalis (8, 14, 22) is recognized as the third-most-common pathogen causing otitis media and sinusitis in children, after Streptococcus pneumoniae and nontypeable Haemophilus influenzae (NTHi) (5, 23). This gram-negative diplococcus is also a cause of respiratory tract infections in adults (6, 43), especially those with chronic obstructive pulmonary diseases (39) or compromised immune systems (2, 22). The incidence of the diseases caused by M. catarrhalis appears to be increasing (37). Further, the percentage of clinical isolates producing beta-lactamase has increased over the last 20 years, with some geographical areas reporting an incidence rate as high as 90% (25). Currently, there is no vaccine for diseases caused by M. catarrhalis.

Vaccine development has been limited because there is little information about the protective antigens or an in vitro correlate of immunity against M. catarrhalis in humans. However, as for other bacterial pathogens (42), serum or bactericidal antibodies appear to be involved in immunity against M. catarrhalis diseases. For example, normal adults with natural immunity possibly resulting from colonization or infection have a lower carriage rate (1 to 6%) than children (50 to 78%) and elderly persons (>26%) and suffer fewer infections (20, 21, 23, 48). Children develop serum antibodies to M. catarrhalis gradually during the first 4 years of life (26). This seems to correlate with a decrease in the incidence of bacteremia and otitis media caused by M. catarrhalis (5, 13). Antibodies to whole cells, outer membrane proteins, and lipooligosaccharide (LOS) of M. catarrhalis have been detected in acute- and convalescent-phase sera of adult patients (12, 40, 41). The rise of antibodies to the above-mentioned antigens in paired sera during infection was variable, with about 50% of patients having an increase and only a few having a significant increase (4-fold). Most convalescent-phase sera (90%), however, demonstrated bactericidal activity against the corresponding M. catarrhalis isolate (9).

Efforts to date to study M. catarrhalis have focused primarily on surface antigens such as outer membrane proteins (4, 7, 31, 32, 38). Among them, a high-molecular-weight protein (UspA) and a major outer membrane protein (CD) have been studied since both are relatively conserved among different strains and are able to generate bactericidal antibodies (32, 38, 52). Passive immunization with monoclonal antibodies to UspA or immunization with UspA enhanced pulmonary clearance of M. catarrhalis strains in a murine challenge model (10, 32). Other surface antigens such as fimbriae have not been found in all clinical isolates (35), and it is still not clear whether a capsular polysaccharide (PS) exists (1).

LOS, a major surface component of M. catarrhalis, is a possible virulence factor in the pathogenesis of infections caused by this bacterium (15, 24). The LOS might be a potential vaccine candidate because serum antibodies to LOS developed in patients with M. catarrhalis infections (40), the convalescent-phase immunoglobulin G (IgG) anti-LOS from patients demonstrated bactericidal activity against M. catarrhalis strains (45), and the serological properties of LOS in humans suggest a less variable structure of LOS (40).

Three major antigenic types of M. catarrhalis LOSs can be distinguished (47), accounting for 95% of 302 strains (A, 61%; B, 29%; C, 5%). These LOSs contain an oligosaccharide (OS) linked to lipid A without an O-specific PS. Structural studies have shown that the OSs from all three serotypes are branched with a common inner core (17-19), and the lipid A portion is similar to that of other gram-negative bacteria (36a). In this study, we used strain 25238 LOS (type A) as a source of LOS (17, 36). LOS is toxic, and the detoxified LOS (hapten) is not immunogenic. Accordingly, LOS was detoxified and coupled to tetanus toxoid (TT) or high-molecular-weight proteins (HMP) to form conjugates designed to elicit serum LOS antibodies (41a). The detailed characterizations of the conjugates in vitro and in animal models are described.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains. Eleven wild-type strains of M. catarrhalis, ATCC 8176, 8193, 23246, 25238, 25239, 25240, 43617, 43618, 43627, 43628, and 49143, were purchased from the American Type Culture Collection (ATCC), Rockville, Md. Ten clinical isolates, M1 to M10, from patients with otitis media with effusions, were kindly provided by Goro Mogi, Oita Medical University, Oita, Japan.

LOS purification. Type A strain ATCC 25238 was grown on chocolate agar at 37°C in 5% CO₂ for 8 h and transferred to 250 ml of 3% tryptic soy broth (Difco Laboratories, Detroit, Mich.) in a 500-ml bottle. The bottle was incubated at 110 rpm in an incubator shaker (model G-25; New Brunswick Scientific Co., Edison, N.J.) at 37°C overnight. The culture was transferred to six 2.8-liter baffled Fernbach flasks, each of which contained 1.4 liters of tryptic soy broth. The flasks were shaken at 110 rpm and maintained at 37°C for 24 h. The culture was centrifuged at 15,000 × g and 4°C for 10 min to collect the cells.

Detoxification of LOS. Anhydrous hydrazine treatment of LOS removes esterified fatty acids from lipid A (29, 30). LOS (160 mg) was suspended in 16 ml of anhydrous hydrazine (Sigma Chemical Co., St. Louis, Mo.) and incubated at 37°C for 3 h with mixing. This suspension was cooled on ice and added dropwise with cold acetone until a precipitate formed. The mixture was centrifuged at 5,000 × g and 5°C for 30 min. The pellet was washed twice with cold acetone, dissolved in pyrogen-free water at a final concentration of 10 to 20 mg/ml, and then ultracentrifuged at 150,000 × g and 5°C for 3 h. The supernatant was passed through a column (1.6 by 90 cm) of Sephadex G-50 (Pharmacia LKB Biotechnology, Uppsala, Sweden) eluted with 25 mM ammonium acetate and monitored with a differential refractometer (R-400; Waters, Milford, Mass.). The eluate was assayed for carbohydrate by a phenol-sulfuric acid method (16). The carbohydrate-containing fractions were pooled, freeze-dried, and designated dLOS.

Derivatization of dLOS. Adipic acid dihydrazide (ADH; Aldrich Chemical Co., Milwaukee, Wis.) was bound to dLOS to form adipic hydrazide (AH)-dLOS derivatives, using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide HCl (EDC) and N-hydroxysulfosuccinimide (sulfo-NHS) (Pierce) (27). dLOS (70 mg) was dissolved in 7 ml of 345 mM ADH (molar ratio of ADH to LOS is ~100 to 1, based on an estimated M_r of 3,000 for dLOS) (17). Sulfo-NHS was added to a concentration of 8 mM, the pH was adjusted to 4.8, and EDC was added to a concentration of 0.1 M. The reaction mixture was stirred and maintained at pH 4.8 for 3 h. The reaction mixture was adjusted to pH 7.0 and passed through the G-50 column as described above. The eluate was assayed for carbohydrate and for AH (34). The peaks containing both carbohydrate and AH were pooled, freeze-dried, and designated AH-dLOS. AH-dLOS was measured for its composition, using dLOS and ADH as standards (16, 34).

Conjugation of AH-dLOS to proteins. TT was obtained from Connaught Laboratories Inc., Swiftwater, Pa., and HMP was purified from NTHi strain 12 (3). AH-dLOS was coupled to TT or HMP to form conjugates (27). Briefly, AH-dLOS (30 mg) was dissolved with 3 ml of water and mixed with 15 mg of TT (5.9 mg/ml) or with 12 mg of HMP (4 mg/ml). The molar ratio of AH-dLOS to both TT (M_r, 150,000) and HMP (M_r, 120,000) was ~100 to 1. The pH was adjusted to 5.4, and EDC was added to a concentration of 0.05 to 0.1 M. The reaction mixture was stirred, and the pH was maintained at 5.4 for 3 h. The reaction mixture was adjusted to pH 7.0, centrifuged, and passed through a column (1.6 by 90 cm) of Sephacryl S-300 in 0.9% NaCl. Peaks that contained both protein and carbohydrate were pooled and designated dLOS-TT or dLOS-HMP. Both conjugates were analyzed for their composition of carbohydrate and protein, using dLOS and bovine serum albumin (BSA) as standards (16, 44).

Preparation of hyperimmune sera. Two New Zealand White rabbits (female, 2 to 3 kg) were injected subcutaneously and intramuscularly (two injecting sites) twice with a 4-week interval with 10⁹ bacteria from strain 25238 plus incomplete Freund's adjuvant (1:1). Blood samples were collected before and 2 weeks after each injection.

Antigenicity. Antigenicity of the dLOS and conjugates was tested by double immunodiffusion and/or enzyme-linked immunosorbent assay (ELISA), using rabbit hyperimmune serum to 25238 whole cells. Double immunodiffusion was performed in 0.8% agarose in phosphate-buffered saline (PBS; pH 7.4).

Immunogenicity. Immunogenicity of the conjugates was examined in mice and rabbits. Five-week-old female mice (NIH Swiss), 10 to 20 per group, were injected subcutaneously with 5 µg of dLOS-TT or dLOS-HMP (carbohydrate), LOS, or dLOS plus TT or HMP (5 µg for protein) in 0.2 ml of 0.9% NaCl with or without Ribi-700 adjuvant (50 µg of monophosphoryl lipid A and 50 µg of synthetic trehalose dicorynomycolate; Ribi ImmunoChem Research, Inc., Hamilton, Mont.). The injections were given three times at 3-week intervals, and the mice were bled 14 days after the first injection and 7 days after the second and third injections.

Bactericidal assay. Pre- and postimmune sera (mouse sera after three injections and rabbit sera after two injections) were inactivated at 56°C for 30 min and tested for bactericidal activity against strains of M. catarrhalis. A complement-mediated bactericidal assay was performed as described previously (29) except that a guinea pig serum (1:1 dilution, 10 or 20 µl per well) was used as a source of complement (Sigma) and the reaction plate was incubated at 37°C for 30 min before plating onto agar plates. The highest serum dilution causing >50% killing was expressed as the reciprocal bactericidal titer.

LAL assay. The LOS and dLOS were tested by the Limulus amebocyte lysate (LAL) assay (33). All reagents were obtained from the Food and Drug Administration, Bethesda, Md. The sensitivity of the LAL assay is 0.2 endotoxin unit (EU)/ml.

SDS-PAGE and silver staining. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and silver staining were performed as described previously (46).

Statistical analysis. Antibody levels are expressed as the geometric mean ELISA units or titers (reciprocal) of n independent observations ± standard deviation or range (n, <4). Significance was tested with the two-tailed t test, and P values smaller than 0.05 were considered significant.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of dLOS. Silver staining of an SDS-polyacrylamide gel showed a single band of LOS (50 to 200 ng) (Fig. 1). In contract, as much as 20 µg of dLOS did not show a band at the location of LOS, which indicated that its residual LOS was less than 0.25%.

View larger version (64K): [in this window] [in a new window]   FIG. 1.   Silver-stained SDS-PAGE patterns of LOS and dLOS from M. catarrhalis 25238. Lanes 1 and 2 contain 200 ng each of Salmonella minnesota LPS Ra and Rc, respectively; lanes 3 through 6 contain 200, 100, 50, and 25 ng, respectively, of LOS; lane 7 contains 20 µg of dLOS from M. catarrhalis.

View larger version (109K): [in this window] [in a new window]   FIG. 2.   Double immunodiffusion. The central well contains a rabbit hyperimmune serum against strain 25238 whole cells. Other wells: 1, LOS, 2 mg/ml; 2, dLOS-TT, 103 µg/ml; 3, dLOS-HMP, 220 µg/ml; 4 and 5, dLOS, 1 mg/ml and 200 µg/ml, respectively; 6, HMP, 500 µg/ml.

Characterization of AH-dLOS and the conjugates. The molar ratio of AH to dLOS in AH-dLOS was 0.38, and the yield, on the basis of carbohydrate content, was 93%. The molar ratios of dLOS to TT and to HMP in two conjugate preparations were 19:1 and 31:1, and the yields were 8 and 19%, respectively, on the basis of carbohydrate content (Table 1). The two conjugates and the LOS formed identical precipitation lines by double immunodiffusion (Fig. 2, wells 1 to 3). The two conjugates showed similar levels of binding to a rabbit hyperimmune serum by ELISA (Table 1).

LOS antibodies in mice. A nonconjugated mixture of dLOS and TT or HMP did not elicit anti-LOS antibodies (Table 2). Both conjugates elicited low levels of anti-LOS IgG after the second but not the first injection, and there was a 50- to 100-fold rise of anti-LOS IgG after the third injection (P < 0.01). dLOS-TT and dLOS-HMP elicited similar levels of anti-LOS IgG after 3 injections. LOS alone and the conjugates elicited similar levels of anti-LOS IgG.

Protein antibodies in mice. For anti-TT antibodies, the mixture of dLOS and TT elicited a level of IgG higher than that elicited by dLOS-TT after each injection (Table 3). dLOS-TT elicited a low level of IgG after the first injection; that level rose significantly after the second and third injections (P < 0.01). The Ribi adjuvant enhanced the level of IgG in the dLOS-TT group. All immunogens elicited low levels of anti-TT IgM.

LOS antibodies in rabbits. A mixture of dLOS, TT, and HMP or LOS alone elicited low levels of anti-LOS IgG or IgM antibodies after three injections (Table 4). dLOS-TT elicited a significant rise of anti-LOS IgG after the first and second injections (37- and 700-fold above the preimmune serum levels). dLOS-HMP showed lower levels of IgG than dLOS-TT (6- and 347-fold above the preimmune serum levels). The Ribi adjuvant enhanced the levels of anti-LOS IgG in both conjugate groups after each injection (40- to 2,000-fold above the preimmune serum levels), and there was no significant difference between the two conjugates with the adjuvant after two injections.

Protein antibodies in rabbits. For anti-TT antibodies, the mixture of dLOS, TT, and HMP elicited a higher level of anti-TT IgG than that of dLOS-TT, especially after one injection (Table 5). dLOS-TT elicited a significant level of IgG after two injections (389-fold above the preimmune serum levels). The Ribi adjuvant enhanced the levels of IgG elicited by dLOS-TT fourfold after two injections. All immunogens elicited low levels of anti-TT IgM.

Bactericidal activity of mouse and rabbit antisera. In the mouse model, only 20% (4 of 20 mice) of conjugate-immunized sera or 45% (9 of 20 mice) of conjugate (with adjuvant)-immunized sera showed bactericidal activity against the homologous strain at a mean titer of 1:6 (1:2 to 1:32) after three injections of dLOS-TT or dLOS-HMP. There was a low correlation between LOS IgM ELISA levels and the bactericidal titers among 40 mice (r = 0.13, P = 0.02).

View larger version (50K): [in this window] [in a new window]   FIG. 3.   Bactericidal activity of conjugate-induced antisera against M. catarrhalis 25238. Rabbit antisera after two vaccinations of LOS, conjugates, and conjugates with Ribi adjuvant were tested, and each group contained two or three rabbits. The bactericidal titers were expressed as the fold increase above the value for preimmune sera based on the serum dilution causing >50% killing of the bacteria and expressed as the geometric mean and standard deviation for each group. The bactericidal titer for the hyperimmune sera elicited by whole cells was 1:1,600.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Two approaches for detoxification of lipopolysaccharide (LPS) or LOS have been applied to obtain clinically acceptable PS or OS from LPS or LOS. Mild-acid treatment of LOS cleaves the lipid A portion from the LOS molecule at the 3-deoxy-D-manno-octulosonic acid-glucosamine linkage (27); mild-alkali treatment of LOS removes ester-linked fatty acids while preserving amide-linked fatty acids of lipid A (29, 30). We used the latter method for the detoxification of M. catarrhalis LOS because the resulting dLOS gave a better yield and immunogenicity after conjugation to protein carriers (data not shown). Treatment of M. catarrhalis LOS with hydrazine resulted in a 20,000-fold reduction in the level of endotoxin, which makes it clinically acceptable (51).

After detoxification, dLOS retained antigenic determinants but was not immunogenic in vivo. dLOS was rendered immunogenic when covalently bound to protein carriers. The resulting conjugates induced IgG antibody responses to LOS in mice and rabbits. In mice, the conjugates elicited levels of anti-LOS IgG similar to those elicited by LOS. However, in rabbits, levels of anti-LOS IgG following immunization with the conjugates were significantly higher than those after LOS. In addition, the two conjugates elicited similar levels of anti-LOS IgG except when dLOS-HMP was formulated with Ribi adjuvant. This latter formulation elicited higher levels of anti-LOS IgG than dLOS-TT after the third injection in mice, indicating that the HMP is a good protein carrier for conjugate vaccines. Both conjugates were more immunogenic in rabbits than in mice, and the immunogenicity of the conjugates in both animals was enhanced with the adjuvant, consistent with our previous studies (27, 29).

The complement-mediated bactericidal activities of the conjugate-induced antisera were examined with homologous strain 25238 and heterologous strains of M. catarrhalis from ATCC and Japan, using guinea pig serum as a source of complement. Most mouse antisera showed no bactericidal activities against the homologous strain. Addition of Ribi adjuvant to the conjugates increased the percentage of mice with bactericidal activity from 20 to 45%. Other sources of complement (53) such as human AB serum or infant rabbit serum were tested, but none of them were applicable because they were bactericidal to M. catarrhalis alone. All conjugate-induced rabbit sera showed bactericidal activity against the homologous strain 25238. Some representative rabbit sera also showed cross-bactericidal activity against 9 of 10 strains from ATCC and Japan. There was a correlation between LOS antibody levels and the bactericidal titers. Immunizing with meningococcal LOS-derived OS-TT conjugates (27) and NTHi dLOS-protein conjugates (29) induced serum IgG anti-LOS in mice and rabbits, but only the latter had bactericidal activity against both homologous and heterologous strains, using infant rabbit serum as a source of complement.

Many studies have shown that PS (from bacterial capsule or LPS) or dLPS-protein conjugates are usually effective immunogens in mice and rabbits as well as in humans (30, 41a, 44a). Regarding the use of LOS as a component for conjugates, many factors may result in differences of antibody response generated in vivo. These include different bacterial strains and species, detoxification and conjugation methods, quality and composition of the conjugates, protein carriers, immunization routes, adjuvants, animal species, detecting methods, etc. In meningococcal LOS conjugate studies, Jennings et al. (33a) showed that dephosphorylated OS-TT conjugates, emulsified in Freund complete adjuvant, elicited bactericidal antibodies in rabbits to the homologous strain. Verheul et al. (48a) reported that OS-TT conjugates were immunogenic in rabbits, but none of the antisera was bactericidal for the homologous strain despite the use of different sources of complement. We synthesized an OS-TT conjugate which was immunogenic in both mice and rabbits, and the antiserum from the latter showed bactericidal activity against both homologous and heterologous strains (27).

Two available studies of NTHi LOS conjugates showed that an OS-CRM₁₉₇ conjugate was a poor immunogen in mice (26a); however, dLOS-TT and dLOS-HMP prepared by us were highly immunogenic in both mice and rabbits without adjuvants (29). In addition, the rabbit antisera but not mouse antisera showed bactericidal activity against the homologous and heterologous strains. These preliminary published and unpublished data from meningococcal, NTHi, and M. catarrhalis LOS-derived conjugates indicate that dLOS conjugates are usually more immunogenic than OS conjugates, both conjugates are more immunogenic in rabbits than in mice, and rabbit antisera are more likely to be bactericidal than mouse antisera.

It has been documented that serum bactericidal LPS or PS antibodies confer immunity to many pathogens in humans, including H. influenzae type b, Neisseria meningitidis, Vibrio cholerae, and Shigella sonnei (12, 42). Recently, we have shown that systemic immunization with dLOS from NTHi conjugated to proteins confers protection from experimental otitis media on chinchillas (26b). In the case of M. catarrhalis infections, the role of anti-LOS antibodies in conferring protective immunity has not been elucidated (40). The convalescent-phase anti-LOS IgG from patients demonstrated bactericidal activity against M. catarrhalis strains (45). We suggest that dLOS-TT or dLOS-HMP induced serum anti-LOS antibodies (IgG) with bactericidal activity can transude to mucosal surfaces (49) of nasopharynges and lyse the inocula of strains causing otitis media (26b) and respiratory diseases caused by M. catarrhalis. For further evaluation of the dLOS-protein conjugate as a potential vaccine for human use, animal protection studies are planned, realizing that M. catarrhalis is an exclusively human pathogen (37a).