Review of Meningococcal Group B Vaccines

Abstract

No broadly effective vaccines are available for prevention of group B meningococcal disease, which account for > 50% of all cases. The group B capsule is an autoantigen and is not a suitable vaccine target. Outer-membrane vesicle (OMV) vaccines appear to be safe and effective but serum bactericidal (SBA) responses of infants are specific for a porin protein (PorA), which is antigenically variable. To broaden protection, OMV vaccines have been prepared from more than 1 strain; from mutants with more than 1 PorA; or mutants with genetically detoxified endotoxin and overexpressed desirable antigens such as factor H-binding protein (fHbp). Also, recombinant protein vaccines such as fHbp, given alone or combined with other antigens, are in late-stage clinical development and may be effective against the majority of group B strains. Thus, the prospects have never been better for developing vaccines for prevention of meningococcal disease, including group B strains.

Introduction

Nearly half of all cases of meningococcal disease in the United States are caused by capsular group B strains for which there is no broadly effective vaccine [1]. In many European countries, the proportion is even higher (90%) [2, 3], in part, because of routine infant and/or toddler meningococcal group C polysaccharide-protein conjugate vaccination [4]. Group B strains cause a disproportionate number of cases in infants < 1 year, the age group with the highest incidence of disease [5–7]. These strains also cause prolonged epidemics, such as occurred in Cuba and Norway during the 1980s and, more recently, in New Zealand [8]. A quadrivalent group A, C, W-135, and Y polysaccharide-protein conjugate vaccine was introduced in the US and recommended for routine use beginning at age 11 years [9]. A more immunogenic quadrivalent conjugate vaccine [10, 11], and a Haemophilus influenezae type b-meningococcal group C and Y conjugate vaccine [12], both suitable for infants, are in late-stage clinical development. Control of meningococcal disease, however, will not be achieved until a broadly effective vaccine is available against group B strains, which is the subject of this review.

Protection Against Meningococcal Disease

Considerable evidence indicates that complement-mediated serum bactericidal antibody (SBA) confers protection against meningococcal disease (reviewed in [13, 14]). An SBA titer of 1:4 or greater when measured with human complement is generally accepted as a surrogate of protection [13]. Recent seroepidemiologic and experimental evidence also indicates that protection may be conferred by bactericidal activity present at serum dilutions < 1:4 and/or by opsonophagocytosis [15]. Nevertheless, because of high specificity, vaccine efficacy can be inferred from SBA titers ≥ 1:4, and the results can be used by national regulatory authorities for the licensure of new meningococcal vaccines.

Challenges for Group B Vaccine Development

When polysaccharides are conjugated to carrier proteins, the polysaccharide antigens become immunogenic in infants and prime for memory anticapsular antibody responses (reviewed in [16, 17]). The meningococcal group B polysaccharide, however, is a homolinear polymer of α(2→8) N-acetyl neuraminic acid (polysialic acid), and is an autoantigen [18]. The polysaccharide is expressed by a number of host tissues [19] and is a poor immunogen, even when conjugated to a protein carrier [20]. To increase immunogenicity, N-propionyl-derivatized group B polysaccharide-conjugate vaccines were prepared, which elicited SBA responses in mice [21] but not in humans [22]. Efforts to develop a group B vaccine, therefore, have focused on noncapsular antigens such as proteins or lipopolysaccharide (in meningococcus, referred to as lipooligosaccharide [LOS], because of the presence of repeating short saccharides instead of long-chain saccharides). The principal challenge has been to identify surface-exposed noncapsular antigens that are safe and antigenically conserved and that elicit broad SBA responses. Promising noncapsular group B vaccine approaches are discussed below and include outer-membrane vesicles (OMVs), recombinant proteins, and a combination of an OMV and recombinant proteins (Table 1).

Table 1.

Group B vaccines investigated in clinical trials

Formulation	Vaccine	Clinical Status	Immunogenicity Results (humans)
Polysaccharide-protein conjugate	N-propionylated group B polysaccharide derivative [22]	Phase 1; completed	Did not elicit SBA
Detergent-treated OMVs	OMV from 1 strain* [28–33]	Phase 3; completed	Elicited SBA and opsonic activity
	Mixture of OMV from 2 strains [40, 41]	Phase 1; completed	Elicited SBA
	Mixture of OMV from 2 mutants, each with 3 PorA proteins [45–47, 127, 128]	Phase 2; completed	Elicited SBA
	Mixture of OMV from 3 mutants, each with 3 PorA proteins [44]	Phase 1	Not yet reported
	OMV from Neisseria lactamica [58]	Phase 1	Elicited minimal SBA responses
Native OMV (not treated with detergents)	Mutant with attenuated endotoxin, 2 PorA proteins, overexpressed fHbp, and other mutations [75]	Phase 1	Not yet reported
Recombinant proteins	TbpB [92]	Phase 1; completed	SBA responses not reported
	NspA [89]	Phase 1; completed	Did not elicit SBA
	fHbp (2 antigenic variants) [120–122]	Phase 2	Elicited SBA
	2 fusion proteins, GNA 2091-fHbp variant 1 and GNA2132-GNA1030, and Neisseria adhesin A (NadA) [119] (Figure 6)	Phase 1	Elicited SBA and opsonic activity
Recombinant proteins + detergent-treated OMV	2 fusion proteins, GNA 2091-fHbp variant 1, and GNA2132-GNA1030, + NadA (Figure 6) + OMV [123, 124]	Phase 2/3	Elicited SBA

^*The vaccine from Cuba was combined with group C polysaccharide [30, 31].

Abbreviations used: OMV, outer membrane vesicles; PorA, one of two porin proteins, designated A; fHbp, factor H binding protein; TbpB, transferring-binding protein B; NspA, Neisserial surface protein A; GNA, Genome-derived Neisserial Antigens; NadA, Neisserial adhesin A

Strategies for Group B Vaccine Development

Detergent-extracted OMV vaccines

OMVs can be separated from meningococcal bacteria (figure 1, panel A) [23], or isolated as membrane blebs, which are released into media during bacterial growth. The OMVs are treated with detergents to extract LOS and decrease endotoxin activity [24]. By sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), the detergent-treated vesicles contain 4 or 5 major outer-membrane proteins (figure 1, panel B). Through more sensitive proteomic methods, the vesicles were shown to contain many other periplasmic and cytoplasmic proteins [25, 26]. The role of these proteins in safety or immunogenicity is unknown. Ample evidence indicates that OMV vaccines are safe [27] and effective in preventing group B meningococcal disease. OMV vaccines prepared from various wildtype strains have been tested in humans [28–33]. In 5 published studies, efficacy of 2 doses given to children 4 years or older, or to young adults, ranged from 57% to 83% against disease caused largely by homologous strains (reviewed in [15]). Recently, administration of 3 doses of an OMV vaccine to the New Zealand population 2 months to 20 years of age controlled a longstanding group B epidemic. Overall, vaccine efficacy was estimated at 73% [34]; and was 80% in the age group 6 months to 5 years [35].

Figure 1.

Detergent-extracted OMV vaccines. Panel A. Electron micrograph of outer membrane vesicles of N. meningitides. The scale bar is 100 nm and the vesicle diameter is about 50–200 nm (80nm on average). Panel B. Major outer membrane proteins (PorA, PorB, reduction modifiable protein (RmpM) and opacity protein A (OpcA)) as visualized by Coommassie-stained SDS PAGE. Lane 1, molecular mass standards; lane 2, strain NZ98/254; lane 3, strain H44/76. After immunization, the SBA–responses of infants and children are directly predominantly against PorA. Adapted from published data [23]. Reprinted from Vaccine, vol 27, Supplement 2, 2009, Holst J, Martin D, Arnold R, et al. Properties and clinical performance of vaccines containing outer membrane vesicles from Neisseria meningitidis, with permission from Elsevier.

One limitation of conventional detergent-treated OMV vaccines is that SBA responses of children are largely directed against surface-accessible loops on a porin protein, called PorA [36], which is antigenically variable [37]. The utility of OMV vaccines, therefore, is best suited for control of epidemics caused by a predominant strain [38, 39]. To broaden protection, OMV vaccines have been prepared from more than 1 strain [40, 41], or from mutant strains engineered to express more than 1 PorA molecule [42–47]. An OMV vaccine from 2 mutants (3 PorA variable region [VR) types per mutant) elicited PorA-specific SBA responses in human infants and primed for booster antibody responses [46, 48]. Certain PorA VR types, however, were poorly immunogenic, and point mutations in the PorA gene can result in small antigenic changes and resistance to SBA [49]. A vaccine consisting of a mixture of OMVs from 3 mutants, each expressing 3 PorA molecules (total of 9 PorA VR types), is in early-stage clinical development [44]. In the US, however, meningococcal disease is caused by strains with considerable PorA antigenic diversity (> 20 PorA VR types) [37, 50, 51]. Therefore, OMV vaccines that predominantly target PorA are unlikely to confer broad protection in infants and young children. These vaccines may be more useful in older age groups since SBA responses of OMV-immunized adults had broader SBA than those of immunized infants [36]. One possible reason is that most adults are naturally primed by exposure to Neisserial organisms and even small quantities of residual non-PorA antigens in the detergent-extracted OMV vaccines may be sufficient to boost memory SBA responses with broad activity. The quantity of these antigens, however, may be insufficient for immunogenicity in unprimed infants.

The incidence of meningococcal disease declines rapidly beginning in the second year of life coincident with the acquisition of colonization by N. lactamica, which is a common commensal of the nasopharynx of young children [52–54]. Exposure to crossreacting N. lactamica antigens and, later, to colonization by N. meningitidis, has been thought to contribute to naturally acquired meningococcal immunity although the specific antigenic targets and protective mechanisms are poorly understood [15]. To mimic naturally acquired immunity, and to circumvent possible immunodominance of PorA, which has been hypothesized to impair serum antibody responses to non-PorA antigens [55], an OMV vaccine was prepared from N. lactamica [52, 54, 56, 57], which does not express an ortholog of meningococcal PorA. In a phase 1 study, adults immunized with the N. lactamica OMV vaccine developed minimal SBA responses, even though nearly all of the subjects were naturally primed based on the presence of bactericidal activity in preimmunization sera [58]. The lack of SBA responses to the lactamica OMV vaccine is not surprising. Below age 10 years, when N. lactamica colonization is common, the prevalence of SBA is low [7, 16]. Also, although mice immunized with an OMV vaccine prepared from N. lactamica were protected against a lethal N. meningitides challenge [57], N. lactamica vaccines did not elicit SBA responses [57, 59]. The mechanism responsible for the mouse protection has not been defined.

Weynants et al. prepared detergent-treated OMV vaccines from mutant N. meningitidis strains in which the PorA gene had been inactivated to circumvent immunodominance [55]. The mutants also were engineered to overproduce several “minor” outer-membrane proteins that are normally expressed in low copy number (Transferrin-binding protein A, Neisserial surface protein A, and Outer Membrane Protein 85). The authors hypothesized that in the absence of PorA immunodominance, and with overexpression of these minor antigens, the breadth of the SBA responses to the mutant OMV vaccine would be increased. In immunized mice, only antibodies elicited by the OMV vaccine from the mutant in which all 3 minor antigens were overexpressed had SBA. The authors concluded that it was necessary to elicit antibodies directed against multiple “minor” antigens to achieve sufficient density of IgG on the surface of the bacteria to engage C1q and activate complement-mediated SBA. As described below, there may be exceptions to this model when the antibodies target a sparsely expressed antigen such as factor H-binding protein (fHbp), which regulates complement pathways [60].

Native (non detergent-treated) OMV vaccines

Detergent treatment, which is used to lower endotoxin activity of OMV vaccines, also extracts desirable antigens such as the lipoproteins, fHbp and Genome-derived Neisserial Antigen 2132 (GNA 2132), which are 2 recently discovered vaccine targets (see below) [61, 62]. To avoid the detergent step and preserve desirable detergent-soluble antigens, it may be possible to prepare native OMV vaccines from strains selected to have naturally low endotoxin activity [63], or to use genetic approaches to attenuate endotoxin activity [64–66].

The lipid A portion of the LOS molecule is responsible for its endotoxin activity. One promising mutant with attenuated endotoxin activity contains a deletion in the LpxL1 gene (also referred to as the msbB gene) [67]. This mutation results in penta-acylated lipid A, which is poorly recognized by human toll-like receptor 4 (TLR4) [68], instead of the more toxic hexa-acylated lipid A, which is present in most wildtype strains [63, 69, 70]. When incubated with human PBMCs, a native OMV vaccine from a penta-acylated mutant had more than 100-fold less endotoxin activity based on lower stimulation of multiple proinflammatory cytokines than a control native OMV from the wildtype strain [65, 66]. The native mutant OMV vaccine showed similar stimulatory activity as control, detergent-treated wildtype OMV vaccines that had been administered safely to humans. Representative TNF-∝ responses to the different WT and mutant OMV vaccines are shown in figure 2.

Figure 2.

Release of TNF-α after incubation of human PBMCs for 4 hours with OMV vaccines. The OMV concentrations that resulted in a 10-fold increased release of TNF-α concentrations above background are shown on the X intercepts. White circles, native OMV from wildtype strain; gray squares, OMV vaccine prepared from LpxL1 KO mutants; black triangles, detergent-extracted OMV vaccines from corresponding wildtype strains. Adapted from published data [65] with permission from American Society for Microbiology.

Many of the newly discovered vaccine targets identified by genome mining (an approach referred to as “reverse vaccinology” [71]), are naturally expressed in relatively low copy number by N. meningitidis strains (which is one reason why these antigens remained unrecognized before genome research). With a few exceptions, humans recovering from meningococcal disease showed low antibody responses to these antigens [72], and the antigens were poorly immunogenic in mice immunized with native OMV vaccines prepared from wildtype strains [66] [73]. To enhance immunogenicity, native (i.e. non–detergent-extracted) OMV vaccines have been prepared from mutants engineered to overexpress desirable antigens such as fHbp [65, 66, 74] (described in detail, below). In mice, a native OMV vaccine with genetically detoxified endotoxin and overexpressed fHbp elicited high titers of serum anti-fHbp antibodies with broad SBA (representative data, figure 3). By adsorption studies, the majority of the bactericidal antibodies were directed at fHbp [65, 66]. The native OMV vaccine, which expressed PorA, also elicited strain-specific bactericidal anti-PorA antibodies [65]. The safety and immunogenicity of a prototype native OMV vaccine from a mutant N. meningitidis strain with genetically attenuated endotoxin activity, overexpressed fHbp, more than one PorA VR type, and other mutations, is being investigated in a phase 1 clinical trial in adults [75].

Figure 3.

Serum bactericidal activity elicited in mice by a native OMV vaccine prepared from a mutant with attenuated endotoxin activity and overexpressed fHbp. N. meningitidis strain designations are shown below the X axis (all expressed fHbp in the variant 1 group). Vaccine groups: recombinant (open bars), multicomponent vaccine containing 2 fusion proteins, GNA 2091-fHbp variant 1 and GNA 2132-GNA 1030, and NadA (See Figure 6); Detergent-OMV, WT (hatched bars), a clinical lot of detergent-treated OMV vaccine from Norway (strain H44/76); native OMV, mutant (gray bars), a native (not treated with detergent) OMV from a mutant strain of H44/76 with attenuated endotoxin (LpxL1KO) and overexpressed fHbp in the variant 1 group. Al(OH)3 (black bars), adjuvant alone (also used for the 3 vaccines). Copyright © 2008 by the Infectious Diseases Society of America. All rights reserved.

LOS has become another potential meningococcal vaccine target since anti-LOS antibodies were reported to have SBA and/or opsonic activity [76–78]. However, the lacto-N-neotetraose (Gal-GlcNAc-Gal-Glc tetrasaccharide) on meningococcal LOS is shared by antigens on human red blood cells, which raises safety concerns. Also, the typical concentrations of detergent used to extract LOS from OMV vaccines to decrease endotoxin activity decreased LOS immunogenicity [66, 67].

To develop an immunogenic LOS-enriched OMV vaccine, Weynants et al. prepared an LpxL1 KO mutant with attenuated endotoxin activity [67]. The mutant also was engineered to express a truncated LOS lacking the terminal galactose of lacto-N-neotetraose to eliminate antigenic crossreactivity with red cell antigens. PorA was deleted to avoid the hypothetical possibility of PorA immunodominance and suppression of anti-LOS antibody responses. In mice, the mutant OMV vaccine, which contained approximately 15% LOS after mild detergent treatment, elicited high serum anti-LOS antibody titers with broad SBA activity. The SBA responses, however, were measured with rabbit complement, which gives much higher titers than human complement [79, 80]. One reason is that rabbit complement factor H (fH) binds poorly to N. meningitidis [81]. In the absence of fH bound to the bacterial surface, complement activation is not down-regulated (Figure 4A and B). The result is that there are higher bactericidal titers with rabbit complement than with human complement since human fH is bound by the bacteria [81]. In other studies, mice immunized with native meningococcal OMV vaccines also developed high titers of anti-LOS antibodies by ELISA, but the antibodies appeared to be of low avidity [66] and had minimal SBA when assayed with human complement [65, 66, 73]. Thus, the vaccine potential of vaccines that target LOS antigens is unknown.

Figure 4.

A, Activation of classical complement pathway. Binding of 2 optimally spaced IgG molecules to the bacterial surface engages C1q and activates the classical complement pathway, which results in increased deposition of C3b. Bound C3b can serve as an opsonin and can also lead to bacteriolyis by cleavage of C5 and assembly of the C5b-9 membrane attack complex. Not shown are the components of the alternative pathway, which can be activated by the classical pathway and serve as an amplification loop. B, Regulation of complement activation by binding of human factor H (fH) to the bacterial surface. Human fH binds to surface-exposed fHbp. fH accelerates the decay of alternative pathway C3/C5 convertases, which downregulates the positive feedback amplification loop of the alternative pathway. Binding of fH also leads to degradation of C3b by factor I (not shown), which decreases classical pathway activation and amplification by the alternative pathway. C, Binding of antibodies to fHbp activates classical complement pathway bacteriolysis and also inhibits binding of fH to the bacterial surface. With decreased amounts of fH bound to the bacterial surface, there is less downregulation of complement activation and the organism becomes more susceptible to complement-mediated bacteriolysis.

Recombinant protein vaccines

New vaccine discovery approaches including genome mining [62, 71, 82–84], proteomics [85, 86], and immunological approaches [87] have identified a large number of novel vaccine targets for prevention of group B meningococcal disease. These include Neisserial surface protein A (NspA) [87–89], Transferrin-binding proteins (Tbp) [90–92], Opacity proteins (Opc) [93–95], Genome-derived Neisserial Antigen (GNA) 2132 [96–99], fHbp (previously referred to as GNA 1870 or LP2086 [61, 100]), FetA [101], Neisserial adhesin A (NadA, also referred to as GNA 1994) [102–105], and others [62, 106]. To date, the vaccine potential of nearly all of these candidates has been limited by either antigenic variability (e.g., FetA and Opc), lack of the gene in strains from some hypervirulent lineages (e.g. NadA), phase-variability (Opc), or low constitutive expression of the antigen by some strains (fHbp, GNA 2132, and NspA). There also may be poor expression of important conformational epitopes by the recombinant protein (NspA) [89, 107]. Thus, a vaccine containing only 1 recombinant antigen is unlikely to be sufficient for broad protective meningococcal immunity.

One of the most promising of the new protein antigens is fHbp, which is a surface-exposed lipoprotein present in all N. meningitidis strains [61, 100, 108]. An important function of fHbp is to bind the human complement protein fH [60]. Binding of fH to the bacterial surface accelerates decay of C3/C5 convertase, which decreases alternative pathway activation (Figure 4B), and contributes to the ability of the organism to avoid complement-mediated killing by non-immune human serum or blood [60, 109–111]. In the presence of anti-meningococcal antibodies, binding of fH to the bacteria also decreases complement-mediated SBA titers [81] by enhancing factor I-mediated degradation of C3b, which decreases classical complement pathway activation, and by decreasing alternative complement pathway amplification (Figure 4B). Additional indirect evidence that fH contributes to the pathogenesis of meningococcal disease comes from a report that persons homozygous for a single nucleotide polymorphism in the promoter region, C496T, have increased serum fH protein levels, and also have an increased risk of developing meningococcal disease [112]. With increased fH concentrations, the bacteria became more resistant to SBA.

As described above, binding of fH to N. meningitidis was reported to be specific for human fH [81]. This human specificity adds another mechanism to explain why N. meningitidis is strictly a human pathogen. A crystal structure of a portion of fH in complex with fHbp has been reported [113], which provided a structural basis for specificity of binding human fH. As a vaccine antigen, fHbp is unique in that antibodies to fHbp both activate classical complement pathway bacteriolysis directly and block binding of fH to the bacterial surface [60] [114]. Inhibition of fH binding would be expected to enhance susceptibility of the organism to classical and alternative pathway bacteriolysis (Figure 4C).

Meningococcal fHbp can be subclassified into 3 antigenic variant groups based on antigenic crossreactivity and sequence similarity of the entire protein [61] (figure 5A). In general, antibodies prepared against fHbp in the variant 1 group (also referred to as subfamily B by Fletcher et al. [100]) were bactericidal against strains expressing fHbp from the variant 1 group but not against strains expressing fHbp in the variant 2 or 3 groups (together, referred to as subfamily A), and vice versa [61, 100, 103]. There also are subvariants of fHbp within each of the variant groups (proteins that differ by < 10% of amino acids from those of the canonical protein of the respective antigenic variant group) [103, 115]. Recently, the molecular architecture of fHbp has been shown to be “modular” (each fHbp variant contains different combinations of five variable segments, each of which is derived from one of two genetic lineages [116], Figure 5B). The breadth of protection conferred by anti-fHbp antibodies against strains expressing subvariants of fHbp, or fHbps from different “modular” groups, remains to be determined.

Figure 5.

A. Phylogram of 70 unique fHbp amino acid sequences showing division of the proteins into 2 subfamilies, designated as A and B by Fletcher et al. [100]. Subfamily B contains the proteins in the variant 1 group described by Masignani et a. [61]. Subfamily A is subdivided into 2 branches, designated by Masignani et al. as variants 2 and 3, respectively. Each branch represents a distinctive protein sequence. The scale bar represents 5 amino acid differences per 100 amino acids. B. Modular structure. The architecture of fHbp consists of different combinations of five variable segments, designated V_A to V_D [116]. Each segment is derived from one of two genetic lineages, designated alpha (gray segments) or beta (white segments). All of the distinctive fHbp protein amino acid sequences referred to in Panel A could be assigned to one of six “modular groups”, designated I–VI. Panel B is reprinted from Microbiology 2009;155:2873–83. Copyright ©2009 by the Society for General Microbiology.

Factor H-binding proteins are part of 2 promising meningococcal vaccines being investigated in humans. One vaccine, referred to as LP2086, contains 2 recombinant lipidated proteins from the fHbp subfamilies A and B. The second vaccine contains fHbp in the variant 1 group (subfamily B) as part of a multicomponent vaccine [96] (figure 6). Two of the components are fusion proteins (GNA 2091 fused with fHbp and GNA 2132 fused with GNA 1030), and the third component is recombinant NadA. Of these 5 antigens, fHbp, GNA 2132, and NadA were reported to be responsible for most of the SBA responses in mice [96].

Figure 6.

Schematic showing three recombinant proteins (five antigens) contained in a multicomponent meningococcal vaccine [96, 119]. Two of the components are fusion proteins (GNA 2132 fused with GNA 1030, and GNA 2091 fused with fHbp. The third component is recombinant NadA. N- and C- refer to the amino- and carboxy terminal portions, respectively, of the proteins. The scale bar represents 100 amino acids (AA). Of the five antigens, fHbp, GNA 2132 and NadA, shown in gray, were reported to be responsible for most of the SBA responses in mice [96]. In the vaccine formulation being investigated in phase 3 clinical trials, the three recombinant proteins are combined with a detergent-extracted OMV vaccine from group B strain NZ98/254.

GNA 2132 antigen is a surface-accessible lipoprotein of unknown function. The gene was detected in all N. meningitidis strains tested to date, and also was present in strains of N. lactamica and N. gonorrhoeae [62]. Based on sequence alignments, portions of GNA 2132 were highly conserved. In mice, recombinant GNA 2132 elicited SBA responses, although only a subset of strains was susceptible with human complement [99]. Human adults immunized with an OMV vaccine combined with recombinant GNA 2132 had higher SBA responses measured with human complement than those who were given a control OMV vaccine without the recombinant protein [97].

NadA is an adhesin/invasin that binds to epithelial cells in vitro [102]. In mice, recombinant NadA elicited SBA responses [105]. The antigen is conserved (> 96% amino acid identity) among group B strains but the gene is absent from strains from certain genetic lineages, which are responsible for approximately 40% of group B disease [103, 105, 117, 118].

In a phase 1 study [119], adults were immunized with the multicomponent recombinant protein vaccine described in figure 6. Their sera were assayed for bactericidal activity against three representative test strains, H44/76, NZ98254 and S3032, each from different genetic clonal complexes: Two of the strains were responsible for large group B epidemics in Norway (H44/76) and New Zealand (NZ98254), and S3032 was from a patient in the United States. High SBA responses were observed against strain H44/76, which expressed a homologous fHbp to the recombinant protein antigen in the vaccine (Serum GMT < 1:4 before immunization, which increased to 1:64 in sera obtained 4 weeks after a third dose of vaccine); 97% of subjects developed protective titers of ≥ 1:4 when measured with human complement (figure 7). Against the 2 other test strains, which were selected because neither had the gene for NadA and both expressed fHbp antigens mismatched for the vaccine antigen, SBA responses were much lower (figure 7). The majority of the postimmunization sera, however, conferred passive protection against these stains in an ex vivo human bacteremia model (figure 7) [119]. The greater sensitivity of the passive protection ex vivo bacteremia assay may reflect higher complement concentrations in the whole human blood assay (90%) than in the SBA assay (20%–25%), and/or the presence of phagocytic cells in the blood assay.

Figure 7.

Summary of the proportion of immunized subjects with titers ≥ 1:4 in SBA (3 strains), or passive protection (PP) in a human blood bacteremia model with group B strains NZ98/254 and S3032 (strain H44/76 was not tested). The bars represent the proportion of pre- (open bars) or postimmunization sera (closed bars) that were positive when tested at a 1:4 dilution in each assay and the respective 95% confidence intervals. Postimmunization passive protective activity was 2.6-to 2.8-fold more frequent then an SBA titer ≥ 1:4. Adapted with permission. Copyright 2009© American Society for Microbiology, Clin Vaccine Immunol, 16:785–91, 2009.

In the multicomponent vaccine formulation that has advanced to phase 3 testing, the 3 recombinant proteins (5 antigens) were combined with a detergent-treated OMV vaccine that had been used to control a group B epidemic in New Zealand [34]. As of Fall 2009, there are no published reports of the results of clinical trials with this vaccine or a bivalent fHbp vaccine. Early clinical data from testing these vaccines have been reported at conferences [120–124]. Both vaccines were well tolerated and elicited SBA responses in children and adults. Although the data were promising, detailed information is needed for critical assessment of the safety and efficacy of these vaccines.

Challenges in Assessing Vaccine Efficacy

Because of the low incidence of meningococcal group B disease, it is not feasible to perform prospective, randomized, controlled clinical trials to assess the efficacy of new group B vaccines. For vaccine licensure, the breadth of protection will be estimated from immunogenicity studies conducted in different age groups and in different geographic locales, and by testing SBA against genetically diverse strains. This approach presents several challenges. First, the quantity of sera available from the infant trials will be limited, which precludes performing assays against a large number of genetically diverse test strains. Given variability in antigen sequence and expression, it will be necessary to select a representative strain panel that insures that the resulting SBA data can be extrapolated to estimate protection against disease-causing strains in the population. Second, as discussed above, by relying only on SBA titers ≥ 1:4 as a measure of protection, the immunogenicity data likely will underestimate protection induced by vaccination [15, 97, 119]. A more sensitive assay is needed. Determining whether protection is conferred by antibodies that only have opsonic activity, or have SBA at titers < 1:4, would permit use of expanded serologic assays and advance vaccine development, as well as help formulate optimal public health strategies for introduction of the new vaccines. Third, up to a third of the effect of group C conjugate vaccines on control of meningococcal disease in a population has been attributed to decreased colonization and transmission of group C organisms (herd immunity) [125]. No information is available on whether protein-based meningococcal vaccines affect transmission of the organism, and no clinical information is available on protection against non-group B strains. The protein antigens used for group B vaccines are shared across strains with other capsular groups [126], and mice immunized with OMV vaccines from group B mutants developed broad SBA responses against epidemic group A, W-135, and X isolates from Africa [126]. Referring to the new protein vaccines as “group B” vaccines is, therefore, a misnomer since the vaccines likely also will protect against capsular group A, C, W-135, and Y strains, which will be an added bonus to vaccination.