Vol. 46, No. 2INFECTION AND IMMUNITY, Nov. 1984, p. 408-4140019-9567/84/110408-07$02.00/0Copyright C 1984, American Society for Microbiology

Development of a Neisseria meningitidis Group B Serotype 2bProtein Vaccine and Evaluation in a Mouse Model

LI YA WANGt AND CARL E. FRASCH*

Office of Biologics, Center for Drugs and Biologics, Bethesda, Maryland 20205

Received 21 May 1984/Accepted 20 August 1984

Although serotype 2 remains the predominant cause of group B Neisseria meningitidis disease in many partsof the world, most cases of this disease are now due to serotype 2b rather than 2a. For this reason, we adaptedthe serotype 2a vaccine method of C. E. Frasch and M. S. Peppler (Infect. Immun. 37:271-280, 1982) to theproduction of a serotype 2b protein vaccine. A spontaneously occurring nonencapsulated mutant of the group Bserotype 2b strain 3006 was obtained by selection on group B antiserum agar. Serotype 2b outer membraneprotein vaccines were prepared with less than 1% lipopolysaccharide contamination. The immunogenicity ofthese vaccines was evaluated in mice in the presence and absence of meningococcal group B and group Ccapsular polysaccharides. The group B and group C polysaccharides equally potentiated the antibody responseto the serotype 2b protein. Addition of aluminum hydroxide or aluminum phosphate markedly improved theantibody response to the serotype 2b protein, but aluminum hydroxide-adjuvanted vaccines consistentlyelicited higher antibody levels. Aluminum hydroxide-adsorbed serotype 2a and 2b protein vaccines wereevaluated for induction of cross-protective bactericidal antibodies. The 2a vaccines were 2a specific, whereasthe 2b vaccines elicited antibodies strongly bactericidal for both 2a and 2b meningococcal strains and protectedagainst bacteremia in a mouse model. It may therefore be possible to provide protection against both 2a and 2bdisease by using an aluminum hydroxide-adsorbed protein vaccine containing a single serotype 2 proteincomponent.

Although the proportion of Neisseria meningitidis (menin-gococcal) disease due to the individual serogroups variesfrom country to country, over 95% of the disease is causedby strains of serogroups A, B, C, Y, and W135. Highlyeffective capsular polysaccharide vaccines for serogroups Aand C were developed over 10 years ago (17, 18). Recently,the polysaccharides for serogroups Y and W135 have beencombined with those of serogroups A and C into a tetrava-lent vaccine (19). In contrast, group B capsular polysaccha-ride vaccines have proven to be essentially nonimmunogenic(30, 36), and attempts to improve the immunogenicity of thepolysaccharide through conjugation to protein have beenunsuccessful (21). Serogroup B is presently the major causeof meningococcal disease in most temperate countries, andthis, combined with the lack of immunogenicity of the Bpolysaccharide, has necessitated development of alternativevaccines based upon the serotype proteins.Meningococcal groups B and C have been subdivided into

18 different serotypes based on the immunological specific-ities of their major outer membrane proteins (10, 31). Thecurrently used procedures for meningococcal serotyping areagar gel double diffusion (11), inhibition of solid-phaseradioimmunoassay (31), and coagglutination by using mono-clonal antibodies absorbed to protein A-Sepharose. Forreasons not yet understood, only a few of these serotypesare associated with most group B and C disease. Of thedisease isolates 60 to 80% are serotype 2 (2, 7), which hascaused most B and C meningococcal epidemics (8).

Serotype 2 has now been subdivided into serotypes 2a, 2b,and 2c based on specific immunodeterminants present on the

* Corresponding author.t Present address: National Vaccine and Serum Institute, Beijing,

People's Republic of China.

41,000-dalton major outer membrane protein (27). Serotype2a was the major serotype associated with disease withingroup B until ca. 1980 and is still the major serotype amonggroup C strains (1, 2, 27). Poolman et al. (27) has docu-mented the change from serotype 2a to 2b in The Nether-lands, and D. M. Jones and J. Eldridge observed thatserotype 2b has become more common in England (Abstr.Int. Conf. Cerebrospinal Meningitis, 5th, p. 9, Marseille,France, 1983). At present most group B serotype 2 disease isdue to serotype 2b (2).

Serotype 2a protein vaccines have been prepared fromouter membranes isolated in the form of membrane vesicles(13). A number of these vaccines have been examined fortoxicity and immunogenicity in animals (26) and evaluatedfor safety and immunogenicity in adult volunteers, in whomthey were found to be both safe and immunogenic (14, 36).The solubility characteristics of the protein vaccines werefound to be important, because early vaccines which werepredominantly protein became insoluble and failed to inducebactericidal and antibodies in humans (14), although theywere immunogenic in animals (15).

Soluble serotype 2a protein vaccines containing noncova-lently bound group B meningococcal polysaccharide havebeen shown to induce serotype 2-specific bactericidal anti-bodies in several clinical studies (9). When these vaccineswere evaluated in young children, they were found to be lessimmunogenic than in adults (12), indicating that the use ofadjuvants such as aluminum hydroxide or aluminum phos-phate may be required.The studies reported here deal with production and evalu-

ation of serotype 2b protein vaccines. We found that addi-tion of aluminum hydroxide at pH 7.0 to the combinedpolysaccharide-serotype 2b protein vaccine significantly in-creased the antibody responses of mice to both the serotype2a and 2b proteins.

408

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

GROUP B MENINGOCOCCAL VACCINE 409

MATERIALS AND METHODS

Vaccine strain. The meningococcal group B serotype 2b(B:2b) strain 3006 was obtained from Jan Poolman, Universi-ty of Amsterdam. The vaccine strain was obtained as a

spontaneously occurring nonencapsulated mutant, selectedon antiserum agar plates by multiple subcultures, and was

designated 3006M2. The serological and chemical character-istics of the mutant were compared with the parent strain,and no apparent differences were found in either the outer

membrane protein patterns or in the serological reactivity,except for the B polysaccharide.

Vaccine. The serotype 2b outer membrane protein vaccinepreparation procedure was based upon the method describedby Frasch and Peppler for serotype 2a (13), with minormodifications. To the concentrated culture supernatant ob-tained by ultrafiltration was added 3 volumes of 95% etha-nol. The resulting precipitate was dissolved in water, thenadjusted to 30 mM Tris-2 mM sodium EDTA (pH 8.5)-5%Brij 96 (polyoxyethylene oleyl ether; Sigma Chemical Co.,St. Louis, Mo.). The ammonium sulfate precipitation stepwas omitted due to poor solubility of the precipitated materi-al. After the outer membrane vesicle preparation was treatedwith Brij 96 to selectively solubilize the lipopolysaccharide(LPS) and ultracentrifuged, the pellet was redissolved inwater rather than 5% sodium deoxycholate for sterile filtra-tion. The protein was next ethanol precipitated and washedwith ethanol to remove residual detergent. The proteinconcentration was determined on a test sample, and basedupon this, an appropriate amount of 3% (wt/vol) lactose was

prepared in distilled water. High-molecular-weight group Bor C polysaccharide (provided by Connaught Laboratories,Inc., Swiftwater, Pa.) was added to the lactose solution to a

concentration of 250 to 1,200 p.g of polysaccharide per ml.The polysaccharide-lactose and lactose solutions were usedto prepare vaccine at 250 to 1,200 jig of protein per ml (seeTable 1), and 1.0 ml was added per vial for lyophilization.The dried vaccines were reconstituted with 5.0 ml per vial ofnormal saline. These vaccines were compared with thepreviously examined serotype 2a vaccines (13; 26). Eachvaccine was analyzed for protein by the Lowry method (23)and for ketodeoxyoctonate by the method of Osborn (25). Atrichloroacetic acid precipitation procedure was used to

separate the outer meembrane vesicle material from thepolysaccharide and lactose to determine protein and keto-deoxyoctonate on the final container material (3). Outermembrane protein patterns were determined by Tris-glycinesodium dodecyl sulfate-polyacrylamide gel electrophoresisas described (29).

Adjuvants. Aluminum phosphate and aluminunm hydroxidewere prepared in our laboratory with alunminum chloride as

the starting material and were sterilized by autoclaving. Theadjuvant preparations contained 6 mg of alutninum phos-phate or 1.32 mg of Al per ml and 8 mg of aluminumhydroxide or 2.77 mg of Al per ml, respectively. Theadjuvants were diluted with saline or phosphate-bufferedsaline (pH 7.4), the final pHs being 6.0 and 7.0, respectively,and the adjuvant concentration was adjusted to 1 mg/ml.Different concentrations and proportions of adjuvant toprotein were used to determine the optimal combination as

assessed by an enzyme-linked immunosorbent assay

(ELISA) and bactericidal antibodies.Mouse immunization and bacterenlia model. For most of

the antibody studies female NIH general-purpose miceweighing 12 to 14 g (3 weeks old) were immunized subcuta-neously (s.c.) or intraperitoneally (i.p.) with 1 ,ug of vaccine

protein. Control mice received 10 ,ug of meningococcalgroup B or group C polysaccharide. Blood was obtained 3 to

8 weeks after immunization by cutting the right subclavianartery. The mouse bacteremia model was used as described(6) to evaluate the level of protection induced by serotype 2bprotein group B polysaccharide vaccines. For mouse bacte-remia studies the mice received a single 10-,ug s.c. injectionof vaccine. Three to four weeks after immunization, animalswere challenged i.p. with 3 x 103 CFU of the serotype 2b3006 parent strain or the serotype 2a strain S946.ELISA. The ELISA was performed as described by

Peppler and Frasch (26). The outer membrane vesicle fromserotypes 2a (M986 NCV-1) and 2b (3006M2) were used at 2,ug/ml to coat 96-well polystyrene plates. The absorbanceobserved at 405 nm was extrapolated to 100 min, and theresults were expressed as ELISA units.

Bactericidal assay. The microbactericidal assay of Fraschand Robbins (15) employing serum from 4-week-old rabbitsdiluted 1:2 as a source of complement was used to compare

bactericidal antibody levels induced by different 2b vaccineswith and without adjuvant. The ehdpoint was defined as thehighest serum dilution causing greater than a 50% reductionin the viable count of strain B16B6 (B:2a) or the 3006 parent

strain (B:2b).Statistical analysis. The geometric mean ELISA values and

bactericidal titers were calculated with a 1-standard-devi-ation confidence interval given. The statistical significancewas determined by using the two-tailed Students t test on themean and standard deviation of log-transformed data.

RESULTS

Analysis of serotype 2b vaccines used for immunogenicityand production studies. The mutant 3006 M2 had an outer

membrane profile on sodium dodecyl sulfate-polyacrylamidegel electrophoresis indistinguishable from that of the parent

strain, which is very similar to that of the 2a pattern

previously reported (13). The vaccine lots used in thesestudies are shown in Table 1. The chemical composition was

determined on the final container material. The polysaccha-ride content shown is based upon the dry weight added to thelactose solution. All vaccines had low amounts of LPS. Theprotein pattern of the vaccines on sodium dodecyl sulfate-polyacrylamide gel electrophoresis was the same as that ofthe 2a vaccines reported previously (13).

Effect of group B and C meningococcal polysaccharides on

TABLE 1. Chemical characteristics of the serotype 2b proteinvaccines used in immunogenicity studies

ProteinPolsaccharideLPS" (~Lg/

Vaccine lot' Prteinaolysccg/vial) mg of(~~.~vial) (p.glvial) protein)

820622V 278 6.9820622VB 170 250 6.9820622VC 170 250 6.9821012VC 182 250 ND'830207VM 715 1,000 <6830207VBM 485 500 <6830207VCM 525 500 <6830207VBH 1,366 1,200 <6

a The vaccine containing only protein was identified by the suffix V andvaccines prepared with group B or C polysaccharide were given suffixes VBand VC, respectively. Vaccines designated VBM and VBH were prepared formouse and human studies, respectively, from the same vaccine bulk.

b The LPS content was calculated based upon a ketodeoxyoctonate contentof 5%.

' ND, Not determined.

VOL. 46, 1984

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

410 WANG AND FRASCH

TABLE 2. Effect of the meningococcal group B and Cpolysaccharides on the immunogenicity of the type 2b protein in

mice

Immunogenicity of the 2b protein by:Vaccine No. of

lot" mice ELISAb Bactericidaltiter"

820622V 6 0.07 (0.02-0.24) 40 (25-62)820622VB 6 0.63 (0.30-1.31) 50 (28-88)820622VC 6 0.55 (0.34-0.92) 56 (16-191)830207V 10 0.38 (0.13-1.03) 24 (9-56)830207VB 10 0.70 (0.25-1.90) 116 (62-221)830207VC 10 0.70 (0.21-2.32) 60 (30-135)

a The vaccines were administered s.c. at a dose of 1.0 ,ug of protein. Theanimals were bled 4 weeks later. Suffixes are as described in Table 1, footnote a.

b ELISA values are equal to the absorbance at 405 nm extrapolated to 100min for sera diluted 1/225. The geometric means and the 1-standard-deviationconfidence limits of the ELISA values are presented.

' The geometric means and the 1-standard-deviation confidence limits ofthe bactericidal titers were calculated from and expressed as the reciprocal ofthe highest serum dilution producing >50% killing.

immunogenicity of the protein. Previous results obtained withserotype 2a vaccines indicated that addition of group Bmeningococcal polysaccharide improved the solubility andimmunogenicity of the protein. The group B polysaccharidealone is nonimmunogenic in mice. The group C meningococ-cal polysaccharide, although poorly immunogenic alone, isquite immunogenic when noncovalently combined with me-ningococcal outer membrane protein (5). We therefore want-ed to determine whether the immunogenicity of the proteincould be further improved by the use of an immunogenicpolysaccharide carrier (Table 2). Although both polysaccha-rides improved the antiprotein response, no difference wasfound in the antibody levels to the serotype 2b protein invaccines containing either group B or group C polysaccha-ride. The group B polysaccharide was therefore used for thelater experiments.

Studies with adjuvants. Aluminum adjuvants were exam-ined for their ability to improve the immunogenicity andcross-protection of 2b vaccines, since aluminum hydroxideand aluminum phosphate are known to be highly effectiveadjuvants for protein antigens. Initial experiments comparedmixing aluminum hydroxide with fluid vaccine and using it toreconstitute lyophilized vaccine. Both procedures yieldedcomparable immunogenicity. Most experiments were there-fore performed by using aluminum hydroxide or aluminumphosphate as a diluent for reconstitution of lyophilizedvaccine. The adjuvant properties of aluminum hydroxide

and aluminum phosphate for serotype 2b protein were

compared (Table 3). Preliminary experiments indicated thatpeak antibody responses without alum occurred at 3 and 4weeks, respectively, following i.p. and s.c. immunization,but that the magnitude of the antibody response was thesame (data not shown). The s.c. route approximates theimmunization route to be used in clinical studies and was,

therefore, used for these studies. The adjuvants gave thebest results when the protein and alum were combined withthe protein at the higher pH (Table 3). At pH 7 bothadjuvants caused significantly higher bactericidal antibodyresponses compared to vaccine lot 830207VB with thevaccine without adjuvant. The aluminum hydroxide hadbetter adjuvant properties than aluminum phosphate for theserotype 2b protein and resulted in higher ELISA andbactericidal antibody levels that were cross-reactive with theserotype 2a protein. With adjuvant, peak antibody levelsoccurred 5 to 6 weeks after immunization.The antibody response to the group B polysaccharide in

mice given the 830207VB vaccine with aluminum adjuvantwas examined by the polylysine precoat ELISA procedure(22) (data not shown). It was hoped that the adjuvant wouldhelp induce an antibody response to the protein-complexedpolysaccharide, but with sera obtained 4 weeks after immu-nization the polysaccharide appeared to be essentially non-

immunogenic. Less than 1 mouse in 20 responded withmeasurable B, polysaccharide antibodies (immunoglobulinM).We examined the effect of different protein-to-aluminum

hydroxide ratios on the antibody response to the serotype 2aand 2b proteins (Table 4). Aluminum hydroxide at pH 7.0was combined with the protein-polysaccharide vaccine atratios of protein to adjuvant from 1:10 to 1:100. The 1:100ratio was superior. Without adjuvant the degree of cross-

reactivity between serotypes 2b and 2a was very low. The 2bantibody response to 1 pg of vaccine at the 1:100 ratio as

measured by ELISA and bactericidal titer did not appear tobe significantly different from the 20- or 50-[Lg doses.The effect of a booster immunization upon the serotype 2b

antibody response was examined (Table 5). After a singlebooster injection, with or without adjuvant, the ELISA andbactericidal antibody levels were elevated, but less thantwofold. The second injection had a greater relative effect onthe vaccine given without adjuvant.Comparison of serotype 2a and 2b vaccines. The 2a and 2b

protein-polysaccharide vaccines were given with and with-out aluminum hydroxide to compare the ability of the twovaccines to induce cross-reactive antibodies (Table 6). In-

TABLE 3. Comparison of the effects of different adjuvants on the antibody response of mice to serotype 2b protein vaccine

Vaccine d.b ELISA" Bactericidal titer'lota Adjuvant pH 2a 2b 2a 2b

830207V None NDd 0.03 (0.01-0.17) 0.34 (0.18-0.60) ND 24 (9-56)830207VB None ND 0.07 (0.04-0.14)e 0.59 (0.28-1.62/ ND 116 (62-221)9

A1(OH)3 6.0 0.07 (0.04-0.14) 0.61 (0.21-1.81) <20 42 (5-351)7.0 1.19 (0.58-2.46)e 0.98 (0.49-1.95/ 806 (285-2,280) 1,881 (1,339-2,643)9

A1(PO4) 6.0 0.27 (0.15-0.50) 0.35 (0.25-0.49) 123 (20-756) 348 (127-958)7.0 0.84 (0.24-2.87)e 0.93 (0.72-1.18/ 92 (36-229) 1,194 (377-3,782)9

a The vaccines were administered s.c. at a dose of 1.0 ,ug into groups of 10 mice. The mice were bled 4 weeks later. Suffixes are as described in Table 1, footnote a.b Aluminum hydroxide and aluminum phosphate were used at a protein-to-adjuvant ratio of 1:100.C For definition of the values, see Table 2, footnotes b and c.d ND, Not determined.e Both of the adjuvants at pH 7.0 caused significantly higher antibody responses than were caused without adjuvant. P < 0.001.f No adjuvant versus either adjuvant. P < 0.1.g No adjuvant versus either adjuvant. P < 0.001.

INFECT. IMMUN.

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

GROUP B MENINGOCOCCAL VACCINE 411

TABLE 4. Effects of different ratios of protein to aluminum hydroxide on the antibody response of mice injected with serotype 2bprotein group B polysaccharide vaccines

Protein/ Dose ELISABVaccine lot' adju- (jig of No. of Bactericidal titer'

vantb protein) mice 2a 2b (2b protein)

830207VBM 1:0 1.0 10 0.02 (0.01-0.15)d 0.32 (0.18-0.56)e 92 (39-215Y1:10 1.0 10 0.36 (0.14-0.91)d 1.13 (0.83-1.55)' 393 (139-1,110Y)1:20 1.0 10 0.30 (0.13-0.70) 1.11 (0.91-1.35) 242 (57-1,020)1.40 1.0 10 0.54 (0.35-1.21) 1.41 (1.04-1.91) 368 (106-1,270)1:100 1.0 10 0.98 (0.71-1.32)d 1.72 (1.48-1.99)e 787 (231-2,681)f1:10 20.0 5S 0.73 1.45 640

50.0 5 0.88 1.55 1,2801:20 20.0 5 1.24 1.52 1,280

50.0 5 1.43 1.55 1,280a The animals were immunized s.c. and bled 4 weeks later. The suffix is as described in Table 1, footnote a.b The amount of adjuvant present was determined as aluminum hydroxide.' For definition of the values, see Table 2, footnotes b and c.d The 2a immune response at the 1:100 ratio was significantly higher than at either the 1:0 or 1:10 ratios. P < 0.001.e No adjuvant versus that at a 1:10 ratio, P < 0.001; adjuvant at a 1:10 ratio versus that at a 1:100 ratio, P < 0.002.f Adjuvant at a 1:10 ratio was significantly better than no adjuvant (P < 0.001), but was not different from that at a 1:100 ratio (P < 0.2).9 The ELISA and bactericidal results were obtained on serum pools, with five mice per pool.

duction of bactericidal antibodies was assumed to be anindicator of cross-protection. The 2b vaccine lot 830207VBcombined with aluminum hydroxide at pH 7.0 stimulated thehighest bactericidal antibody levels against both serotype 2aand 2b strains. The 2a vaccines stimulated almost no bacteri-cidal antibodies against the serotype 2b test strain, with orwithout adjuvant.To estimate the degree of cross-immunity that could be

expected from the use of a 2b serotype protein vaccine, the2b vaccine was evaluated in the mouse bacteremia model forability to induce both homologous (2b) and heterologous (2)protection (Table 7). Groups of mice were immunized with10 ,ug of protein in adjuvant and challenged 3 to 4 weeks laterwith a B:2a or B:2b strain. The serotype 2b vaccine inducedprotection against both the 2a and 2b strains. Aluminumhydroxide was the better adjuvant. In similar experiments,immunization with the group B polysaccharide alone failedto protect the mice against group B challenge.

DISCUSSIONWork on development of an effective vaccine for preven-

tion of group B meningococcal disease has until now concen-trated upon serotype 2a (14, 35, 36). Recent studies byPoolman et al. (27) and Ashton et al. (1, 2) indicate,however, that serotype 2b has become the predominantcause of group B serotype 2 disease. Studies were thereforerequired to compare the potential of serotype 2a and 2bvaccines to induce cross-protective antibodies and to deter-mine whether a single serotype 2 protein could provideprotection against both 2a and 2b.

Serotype 2a vaccines have been evaluated in severalclinical studies (9, 14, 16, 36). The initial clinical studies wereconducted in adults with particulate protein vaccines thatinduced a relatively poor immune response (14, 34). Zol-linger et al. later tested a soluble group B polysaccharide-serotype 2a protein vaccine in adults and found betterimmunogenicity (36). Soluble protein vaccines preparedunder relatively nondenaturing conditions (13) were com-pared in mice with and without group B polysaccharide. Thepolysaccharide formed a noncovalent complex with theprotein and significantly improved the immunogenicity ofthe protein. Soluble serotype 2a protein vaccines containingeither protein alone or protein and group B polysaccharidehave been evaluated in adult volunteers (9, 16). The com-bined protein-polysaccharide vaccine was clearly the superi-or immunogen, inducing two- to fourfold higher antibodylevels as measured by both ELISA and a serotype 2-specificbactericidal assay.The serotype 2a protein group B polysaccharide vaccines

have been tested in progressively younger age groups (9, 14).The vaccines caused mild local reactions in most vaccinerecipients, but were associated with fewer systemic reac-tions (primarily fever) in the younger children (9). Thecombined group B polysaccharide serotype 2a protein vac-cines induced a much lower antibody response in children 6months to 5 years of age than in adolescents and adults (12).The purpose of the present studies was twofold: (i) adaptionof our published serotype 2a vaccine procedure (13) to theproduction of a serotype 2b vaccine and (ii) to furtherimprove the immunogenicity of serotype protein vaccines.

TABLE 5. Comparison of the homologous antibody response to serotype 2b vaccine lot 830207VB with and without a booster"Protein/ Bp pdadjuvant Boosterb ELISA' Pd Bactericidal titer' Pd1:0 No 0.21 (0.10-42) 116 (62-221)

Yes 0.41 (0.16-1.01) <0.1 240 (105-556) <0.051:100 No 0.76 (0.51-112) 1,194 (959-1,407)

Yes 1.12 (0.83-1.52) <0.05 1,470 (948-2,279) <0.1

a The vaccine was administered s.c. at a dose of 1.0 pLg of protein with and without aluminum hydroxide to groups of 10 mice. The nonboosted animals were bledat 4 weeks.

b The booster was given 4 weeks after the primary immunization, and animals were bled 2 weeks later." For definition of the values, see Table 2, footnotes b and c.d In each case, the t test was performed on the log-transferred values, comparing for each set with and without booster.

VOL. 46, 1984

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

412 WANG AND FRASCH

TABLE 6. Comparison of the antibody response of mice to serotype 2a and 2b protein vaccines with and without aluminum hydroxide

Vaccine Adjuvant ~~~~~~~~~~~ELIsAb 13actericidal titerbVaccie Sertype Adjuvant pH_________________lot' Serotype present pH 2a 2b 2a 2b

790626VB 2a No 0.66 (0.21-2.11)c NDd 60 (41-87) <10Yes 6.0 1.45 (0.96-2.18)c ND 80 (49-130) 17 (3-98)

3179VB 2a No 0.64 (0.20-2.07)c ND 30 (5-172) <10Yes 6.0 1.65 (0.95-2.87)C ND 69 (22-200) <10

821012VC 2b No ND 0.71 (0.33-1.54) ND 80 (16-396)Yes 6.0 ND 1.25 (0.90-1.74) 52 (36-77) 242 (69-854)

830207VB 2b No 0.02 (0.01-0.06) 0.24 (0.16-0.36)e ND 92 (54-154)Yes 7.0 0.89 (0.26-3.00) 1.13 (0.84-1.52)e 519 (95-2,843), 1,689 (453-6,291)f

a The animals received 1 FLg of vaccine protein s.c. and were bled 6 weeks later. There were 5 and 10 mice per group for the 2a and 2b vaccines, respectively.The suffixes are as described in Table 1, footnote a.

b For definition of the values, see Table 2, footnotes b and c.c The antibody respoises to 2a vaccines were not statistically different with and without adjuvant.d ND, Not determined.eVaccine with adjuvant versus that without adjuvant. P < 0.001.f The serotype 2a and 2b imnmune responses were not statistically different at P < 0.05.

The serotype 2b vaccine was prepared from a nonencapsu-lated variant selected by multiple passages on group Bantiserum agar. The LPS content of the serotype 2b vaccines(Table 1) was considerably lower than that of the serotype 2avaccines (13), even though both were prepared by similarprocedures. The LPS is apparently more firmly associatedwith the protein in the serotype 2a vaccine strain M986NCV-1 than in the 2b vaccine strain 3006M2.

Addition of group B polysaccharide to the detergent-treated serotype 2a outer membrane vesicle improved vac-cine solubility, as shown by electron microscopy (13). Thisimproved solubility may have contributed to the betterimmune response of mice to the combined vaccine (26). Thegroup B polysaccharide similarly improved the antibodyresponse to the serotype 2b protein (Table 2). The group Bpolysaccharide is poorly immunogenic (30), whereas thegroup C polysaccharide is immunogenic in mice when com-plexed with protein (5). However, the antibody response tothe serotype 2b protein when combined with either of thesetwo polysaccharides was the same.A relatively poor correlation was found between the

ELISA and bactericidal results, similar to an observationmnade by Zollinger et al. (33). The ELISA measures antibodyto all of the outer membrane surface antigens, and only some

TABLE 7. Cross-protection induced by serotype 2b proteingroup B polysaccharide vaccine lot 830207VB in a mouse

bacteremia model

Challenge Protein % protection' atstrain Adjuvant adjuvant' 3 h 6 h

3000(B:2b) A1(OH)3 1:10 40 1001:20 40 100

A1PO4 1:10 20 1001:20 20 100

S-946(B:2a) A1(OH)3 1:10 40 1001:20 40 100

A1PO4 1:10 40 801:20 40 60

Adjuvant 0 0control

a For all groups, excepting the adjuvant control, the mice received a singles.c. injection of 10 ,ug of protein with the protein-to-aluminum hydroxideratios shown."Groups of 10 mice were challenged i.p. with 3 x 104 CFU of strain 3006

(B:2b) or S-946 (B:2a) 3 to 4 weeks after s.c. immunization. The percentage ofmice without detectable bacteremia at 3 and 6 h after challenge was calculatedas percent protection.

of these antibodies are functionally bactericidal. Studies inour laboratory (unpublished data) have shown that individ-ual mice respond differently in relative amounts of antibodyto different outer membrane proteins, as detected byimmunoblotting. The observed poor correlation may, there-fore, be due to differences in the immune response tospecific membrane antigens. There was, however, generalagreement between the two assays regarding the effects ofadjuvants.Aluminum hydroxide and aluminum phosphate gels are

known to be effective adjuvants for protein antigens, proba-bly functioning in part by retaining the antigen at theinjection site longer. We compared these two adjuvants fortheir effects on the serotype 2b antibody response of mice.Addition of aluminum hydroxide to the protein at pH 7.0resulted in a higher antibody response that when aluminumhydroxide was combined with protein at pH 6.0. This may berelated to the fact that the aluminum hydroxide was anopalesent "solution" at pH 6, but was insoluble at pH 7. Thealuminum phosphate was insoluble at both pHs, the adjuvantproperties being approximately 3 times better at pH 7.0. Animportant function of the adjuvant was to stimulate forma-tion of cross-reactive bactericidal antibodies effectiveagainst both serotype 2a and 2b strains after immunizationwith 2b, and in this regard aluminum hydroxide was clearlybetter than aluminum phosphate. The cross-reactive bacteri-cidal antibodies were probably not directed against the Bpolysaccharide, because we were unable to detect B poly-saccharide antibodies by ELISA.The proportion of aluminum adjuvant to protein was found

to be an important variable in optimizing the antibodyresponse to the protein. Although initial studies comparingaluminum hydroxide and aluminum phosphate were donewith an arbitrary protein-to-adjuvant ratio of 1:100, this ratioproved to provide the best adjuvant properties. The antibodyresponse to 1 ,ug of serotype 2b protein at the 1:100 ratio wascomparable to that of 20 to 50 jxg of protein injected withsuboptimal amounts of aluminum hydroxide.A problem that has complicated administration of ad-

sorbed diphtheria and tetanus toxoids and pertussis vaccineis the formation of sterile abscesses or nodules at the site ofinjection (4). Studies indicated that these complicationsregarding administration of this vaccine were due either totoo large amounts of alum or to injection of the vaccine s.c.rather than intramuscularly (4, 28). The amount of aluminumhydroxide or aluminum phosphate that can be administered

INFECT. IMMUN.

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

GROUP B MENlNGOCOCCAL VACCINE 413

is limited by U.S. regulations to 0.85 mg of elementalaluminum per dose. Whereas the serotype 2a meningococcalprotein vaccines were used in clinical studies in doses of 25to 100 jig, the aluminum hydroxide-adsorbed serotype 2bvaccine dose for humans will be limited by the aluminumcontent (1:100 ratio) to ca. 25 jig of protein. The mouse dataindicate that in the presence of aluminum hydroxide, theeffective dose of serotype protein would probably be lowerthan it would be without adjuvant.

Since the adsorbed toxoid vaccines are given as a series ofinjections, we compared the effects of two injections ofserotype 2b protein given 1 month apart with and withoutaluminum hydroxide. After a single s.c. injection we foundthat peak antibody levels, as measured by ELISA, occurredca. 4 weeks after immunization without adjuvant but at 5 to 6weeks with adjuvant. A second injection resulted in in-creased antibody levels: 0.5-fold with adjuvant and 2-foldwithout adjuvant. After two injections, the geometric mean

bactericidal antibody levels elicited by the adsorbed proteinwere sixfold higher than when no adjuvant was used.

Serotype 2 is the major disease-causing serotype withinmeningococcal serogroups B and C (2, 7, 8). Poolman et al.determined that serotype 2 actually consisted of three close-ly related serotypes, 2a, 2b, and 2c (27); serotypes 2a and 2bwere associated with group B and group C disease, whereasserotype 2c was found only among group Y strains. Until themid to late 1970s most group B and C disease was caused byserotype 2a. After that time group B serotype 2 diseasebecame predominantly serotype 2b (1, 27). Since the group Cpolysaccharide does not appear to be an effective immuno-gen in children under 2 years of age and the B polysaccharideis ineffective, bactericidal antibodies reacting with bothserotype 2a and 2b strains may be required to protect againstgroup B and group C serotype 2 disease in young children.To avoid the necessity of using a combined 2a-2b protein

vaccine, serotype 2a and 2b protein vaccines were comparedfor their ability to stimulate cross-protective bactericidalantibodies. In contrast to the results with 2a vaccines, the 2bvaccines induced bactericidal antibodies effective againstboth serotypes 2a and 2b, when given adsorbed to aluminumhydroxide. These experimental findings are in agreementwith the immunochemical studies of Poolman et al., whichindicated that a serotype 2a vaccine would not be sufficientlycross-reactive with the serotype 2b outer membrane proteinto provide protection against serotype 2b (27). Serotype 2aand 2b strains share common 46,000-molecular-weight class1 proteins, and monoclonal antibodies to this protein killboth 2a and 2b strains (L. F. Mocca and C. E. Frasch,unpublished data). The 28,000-molecular-weight class 5 pro-

teins of the two strains are antigenically distinct but are

highly antigenic, eliciting strain-specific bactericidal anti-bodies (Mocca and Frasch, unpublished data). The 2a vac-

cine may have induced relatively less bactericidal antibodyagainst the class 1 protein.

In agreement with our bactericidal data, the 2b vaccineprevented bacteremia in mice challenged with either 2a or 2bstrains. Similar experiments with serotype 2a vaccines failedto protect mice against bacteremic infection with a 2b strain(unpublished data). Thus, aluminum hydroxide-adsorbedserotype 2b protein-group B polysaccharide vaccines may

provide protection against both serotype 2a and 2b disease.Clinical studies are planned to substantiate the mouse

immunogenicity data reported here. Adults and then youngchildren will be given serotype 2b protein-group B polysac-charide vaccine intramuscularly with and without aluminumhydroxide as an adjuvant. Since a significant proportion of

group B meningococcal disease in some countries is due toserotype 15 (20, 24), combined serotype 2-serotype 15 vac-cines are being prepared for clinical evaluation.

LITERATURE CITED1. Ashton, F. A., J. A. Ryan, C. Jones, B. R. Brodeur, and B. B.

Diena. 1983. Serotypes of Neisseria meningitidis associatedwith an increased incidence of meningitis cases in the Hamiltonarea, Ontario, during 1978 and 1979. Can. J. Microbiol. 29:129-136.

2. Ashton, F. E., A. Ryan, B. B. Diena, A. M. R. MacKenzie, and F.Chan. 1980. Serotypes among Neisseria meningitidis sero-groups B and C strains isolated in Canada. Can. J. Microbiol.26:1480-1488.

3. Bensadoun, A., and D. Weinstein 1976. Assay of proteins in thepresence of interfering materials. Anal. Biochem. 70:241-250.

4. Bernier, R. H., J. A. Frank, Jr., and T. F. Nolan, Jr. 1981.Abscesses complicating DTP vaccination. Am. J. Dis. Child.135:826-828.

5. Beuvery, E. C., F. Miedema, R. W. van Delft, J. Haverkamp,A. B. Leussink, B. J. te Pas, K. S. Teppema, and R. H. Tiesjema.1983. Preparation and physicochemical and immunologicalcharacterization of polysaccharide-outer membrane proteincomplexes of Neisseria meningitidis. Infect. Immun. 40:369-380.

6. Craven, D. E., and C. E. Frasch. 1979. Protection against groupB meningococcal disease: evaluation of serotype 2 proteinvaccines in a mouse bacteremia model. Infect. Immun. 26:110-117.

7. DeMaeyer, S., J.-M. Seba, and G. Reginster. 1981. Epidemiol-ogy of meningococcal meningitis in Belgium. J. Infect. 3(Suppl.1):63-70.

8. Frasch, C. E. 1979. Noncapsular surface antigens of Neisseriameningitidis, p. 304-337. In L. Weinstein and B. N. Fields (ed.),Seminars in infectious disease, vol. 2. Stratton IntercontinentalMedical Book Corp., New York.

9. Frasch, C. E. 1983. Immunization against Neisseria meningiti-dis, p. 115-144. In C. S. F. Easman and J. Jeljaszewicz (ed.),Medical microbiology, vol. 2. Academic Press, Inc., New York.

10. Frasch, C. E., and S. S. Chapman. 1972. Classification ofNeisseria meningitidis group B into distinct serotypes. I. Sero-logical typing by a microbactericidal method. Infect. Immun.5:98-102.

11. Frasch, C. E., and S. S. Chapman. 1972. Classification ofNeisseria meningitidis group B into distinct serotypes. II.Extraction of type-specific antigens for serotyping by precipitintechniques. Infect. Immun. 6:127-133.

12. Frasch, C. E., G. Coetzee, J. M. Zahradnik, H. A. Feldman, andH. J. Koornhof. 1983. Development and evaluation of group Bprotein vaccines: report of a group B field trial. Med. Trop.43:177-180.

13. Frasch, C. E., and M. S. Peppler. 1982. Protection against groupB Neisseria meningitidis disease: preparation of soluble proteinand protein-polysaccharide immunogens. Infect. Immun.37:271-280.

14. Frasch, C. E., M. S. Peppler, T. R. Cate, and J. M. Zahradnik.1982. Immunogenicity and clinical evaluation of group B Neis-seria meningitidis outer membrane protein vaccines, p. 263-267. In J. B. Robbins, J. C. Hill, and J. C. Sadoff(ed.), Seminarsin infectious disease, vol. 4. Thieme-Stratton, Inc., New York.

15. Frasch, C. E., and J. D. Robbins. 1978. Protection against groupB meningococcal disease. III. Immunogenicity of serotype 2vaccines and specificity of protection in a guinea pig model. J.Exp. Med. 147:629-644.

16. Froholm, L. O., B. P. Berdal, K. Bovre, P. Gaustad, A. Harboe,E. Holten, E. A. Heiby, A. Lystad, T. Omland, and C. E. Frasch.1983. Meningococcal group B vaccine trial in Norway. Prelimi-nary report of results available November 1983. NIPH Ann.6:133-138.

17. Gold, R., and M. S. Artenstein. 1971. Meningococcal infections.2. Field trial of group C meningococcal polysaccharide vaccinein 1969-1970. Bull. W.H.O. 45:279-282.

VOL. 46, 1984

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

414 WANG AND FRASCH

18. Gotschlich, E. C., T. Y. Liu, and M. S. Artenstein. 1969. Humanimmunity to the meningococcus. lII. Preparation and immuno-chemical properties of group A, group B and group C meningo-coccal polysaccharides. J. Exp. Med. 129:1349-1365.

19. Hankins, W. A., J. M. Gwaltney, J. R. Hendlery, J. 0. Farque-har, and J. S. Samuelson. 1982. Clinical and serological evalua-tion of a meningococcal polysaccharide vaccine groups A, C, Yand W135. Proc. Soc. Biol. Med. 169:54-57.

20. Holten, E. 1979. Serotypes of Neisseria meningitidis isolatedfrom patients in Norway during the first six months of 1978. J.Clin. Microbiol. 9:186-188.

21. Jennings, H. J., and C. Lugowski. 1981. Immunochemistry ofgroups A, B, and C meningococcal polysaccharide-tetanustoxoid conjugates. J. Immunol. 127:104-108.

22. Leinonen, M., and C. E. Frasch. 1982. Class-specific antibodyresponse to group B Neisseria meningitidis capsular polysac-charide: use of polylysine precoating in an enzyme-linkedimmunosorbent assay. Infect. Immun. 38:1203-1207.

23. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

24. Mocca, L. F., G. del Real, and C. E. Frasch. 1983. Serotypes andpolyacrylamide gel electrophoresis types among disease-associ-ated isolates of group B Neisseria meningitidis in Spain. 1976-1979. J. Infect. Dis. 148:249-253.

25. Osborn, M. J. 1963. Studies on the gram negative cell wall. I.

Evidence for the role of 2-keto-3-deoxy-octonate in the lipo-polysaccharide of Salmonella typhimumium. Proc. Natl. Acad.Sci. U.S.A. 50:499-506.

26. Peppler, M. S., and C. E. Frasch. 1982. Protection against group

B Neisseria meningitidis disease: effect of serogroup B polysac-charide and polymyxin B on immunogenicity of serotype pro-tein preparations. Infect. lmmun. 37:264-270.

27. Poolman, J. J., C. T. P. Hopman, and H. C. Zanen. 1980.Immunochemical characterization of Neisseria meningitidis se-rotype antigens by immunodiffusion and SDS-polyacrylamidegel electrophoresis immunoperoxidase techniques and the dis-

tribution of serotypes among cases and carriers. J. Gen. Micro-biol. 116:465-473.

28. Sako, W. 1947. Studies in pertussis immunization. J. Pediatr.30:24-40.

29. Tsai, C.-M., and C. E. Frasch. 1980. Chemical analysis of majorouter membrane proteins of Neisseria meningitidis: comparisonof serotypes 2 and 11. J. Bacteriol. 141:169-176.

30. Wyle, F. A., M. S. Artenstein, B. L. Brandt, E. C. Tramont,D. L. Kasper, P. L. Altieri, S. L. Berman, and J. P. Lowenthal.1972. Immunological response of man to group B meningococ-cal polysaccharide vaccines. J. Infect. Dis. 126:514-522.

31. Zollinger, W. D., and R. E. Mandrell. 1977. Outer membraneproteins and lipopolysaccharide serotyping of Neisseria menin-gitidis by inhibition of a solid-phase radioimmunoassay. Infect.Immun. 18:424-433.

32. Zollinger, W. D., and R. E. Mandrell. 1983. Importance ofcomplement source in bactericidal activity of human antibodyand murine monoclonal antibody to meningococcal group Bpolysaccharide. Infect. Immun. 40:257-264.

33. Zollinger, W. D., and R. E. Mandrell. 1983. Studies of thehuman antibody response to specific meningococcal outer mem-brane proteins of serotype 2 and 15. Med. Trop. 43:143-147.

34. Zollinger, W. D., R. E. Mandrell, P. Altieri, S. Berman, J.Lowenthal, and M. S. Artenstein. 1978. Safety and immunoge-nicity of a meningococcal type 2 protein vaccine in animals andhuman volunteers. J. Infect. Dis. 137:728-739.

35. Zollinger, W. D., R. E. Mandrell, and J. M. Griffiss. 1982.Enhancement of immunologic activity by noncovalent complex-ing of meningococcal group B polysaccharide and outer mem-brane proteins, p. 254-262. In J. B. Robbins, J. C. Hill, and J. C.Sadoff (ed.), Seminars in infectious disease, vol. 4. Thieme-Stratton, Inc., New York.

36. Zollinger, W. D., R. E. Mandrell, J. M. Griffiss, P. Altieri, andS. Berman. 1979. Complex of meningococcal group B polysac-charide and type 2 outer membrane protein immunogenic inman. J. Clin. Invest. 63:836-848.

INFECT. IMMUN.

on March 19, 2020 by guest

http://iai.asm.org/

Dow

nloaded from