A Multi-Component Meningococcal Serogroup B Vaccine (4CMenB): The Clinical Development Program

The introduction of safe and immunogenic glycoconjugate vaccines against invasive meningococcal disease (IMD) in Europe, Canada, Australia, and Brazil (Neisseria meningitidis serogroups ACWY and C) and in the meningitis belt in sub-Saharan Africa (serogroup A) has increased awareness of the need for disease prevention of the remaining prominent cause of IMD: serogroup B (MenB) [1]. Laboratory-confirmed cases of MenB disease resulting in substantial mortality have been documented throughout the world, most notably in countries where surveillance systems are robust and clinical outcomes are well documented [2‐4]. Serogroup B is the predominant cause of IMD in most of Europe and Australia, especially where serogroup C vaccination is part of routine recommendations and in New Zealand due to a serogroup B epidemic. Similarly, in Canada, serogroups B, C, and Y represented the main causes of IMD after introduction of serogroup C conjugated vaccines by different provinces from 2001–2009; in 2010, serogroup B was responsible for 64 % of IMD, compared with 19 and 7 % for serogroups Y and C, respectively [4]. In the USA, where vaccination of infants is not practiced but where vaccination of adolescents against serogroups A, C, W, and Y is routinely recommended, serogroups B, C, and Y represent the main causes of IMD in approximately equal proportions [3]. In Latin America, overall, serogroup B has predominated over C, W, and Y, although the situation is highly variable—recent increases in serogroup C have seen it replace serogroup B as the predominant form in Brazil [5], and serogroup W has recently emerged as an important group in Argentina (34 % of all IMD in 2011) and Chile (55 % of IMD in 2012) [6]. The epidemiologic picture in Brazil may continue to change if, as expected, the recently implemented infant MenC immunization campaign reduces the incidence of serogroup C disease. Serogroup dynamics will most probably continue to change during the future years worldwide.

Prevention of serogroup B-related IMD presents unique challenges. Glycoconjugate vaccines against serogroups A, C, W, and Y exploit the antigenicity of the capsular polysaccharides that characterize the serogroups, but poor immunogenicity of the MenB capsular polysaccharide [7] and antigenic diversity of the MenB surface proteins have significantly complicated vaccine development [8]. To provide broad coverage against the diversity of MenB strains, vaccine candidates require well conserved antigen variants and mixtures of multiple surface proteins, with the capacity to induce bactericidal antibodies against a majority of circulating strains [7].

Strong evidence that a protein-based meningococcal vaccine would provide protective efficacy originated from clinical experience with vaccines derived from meningococcal outer membrane vesicles (OMVs) to control regional outbreaks of specific MenB strains [9‐11]. OMVs contain several different molecular moieties, but the porin protein, PorA, is the principle antigenic source of bactericidal antibodies. The limitations of these vaccines are well recognized; effectiveness tends to be limited to strains containing the same PorA protein (serosubtype-specific), especially in young children, limiting use to strain-specific outbreaks [11]. Worldwide, the vast majority of MenB disease is not related to one or even a limited set of disease-causing strains. Thus, advancements toward MenB IMD protection using OMV have been limited, as their effective coverage would not be sufficiently broad to justify development and implementation [1].

To confer broad protection against MenB, a vaccine has to account for high levels of antigenic diversity, variability in types and expression of antigens, and also anticipate the risk of escape mutants [12, 13]. To do so, constituent antigens should be highly conserved, important for virulence and/or survival of N. meningitidis, surface-expressed, and able to induce bactericidal antibodies.

In the late 1990s, reverse vaccinology applied a complete analysis of the genetic sequence of a MenB strain to systematically identify numerous surface-exposed proteins that were screened for their ability to induce bactericidal activity against MenB in animal models [14]. The most promising of these candidates were included in investigational vaccine formulations aimed at protecting against a broad spectrum of MenB strains circulating worldwide [13]. Further refinements, including fusion of certain vaccine antigens to stabilize protein conformation, were made in the final formulation [15].

This review focuses on the clinical development program of the Novartis Vaccines and Diagnostics meningococcal vaccine, 4CMenB, now licensed in Europe and Australia as Bexsero^® (Fig. 1). The goal was to develop a MenB meningococcal vaccine proven to be safe and immunogenic in all age groups, particularly in infants, who represent the major group at risk of IMD, and to induce long-term immunity across heterologous subtypes to be effective against a majority of circulating strains [16]. 4CMenB contains the four main components, two of which are presented as fusion proteins with two other antigenic proteins identified in the screening process, GNA2091 and GNA1030: Neisseria adhesin A (NadA), factor H-binding protein (fHbp) with GNA2091, Neisseria heparin-binding antigen (NHBA) with GNA1030, and OMV from the New Zealand outbreak strain NZ98/254 (NZ OMV), which contains the PorA 1.4 protein. The rationale for choosing these components for inclusion in 4CMenB, details of the vaccine development, and the distribution of the vaccine antigens in meningococcal strains have been described in detail elsewhere [15].

We provide updated and consolidated information regarding the 4CMenB clinical program, reviewing results from the early-phase studies, summarizing details of the immunogenicity and safety results from pivotal studies completed in adolescents and infants, and potential coverage of this vaccine based on available global surveillance data. We also address the potential public health impact of this novel vaccine.

For relatively rare diseases such as IMD, with annual disease rates ranging from 0.5 to 5 per 100,000 individuals, efficacy studies are impractical because of the large sample sizes required. Instead, an immunologic correlate of protection using the serum bactericidal assay (SBA) has been accepted as an approach for vaccine licensure [17]. Correlates of protection using SBA with exogenous human complement (hSBA) were obtained approximately 40 years ago in studies performed in US Army recruits that provided data supporting a relationship between susceptibility to serogroup A, B, and C meningococcal disease and absence of bactericidal antibodies [17‐19]. These data, combined with the observation of peak disease rates in age groups with the lowest detected hSBA antibody levels (6–24 months of age) support the concept that bactericidal antibodies are sufficient to confer protection against IMD; the level established as surrogate protective correlate is a serum dilution of 1:4 (or titer of 4) in the hSBA assay [17]. These observations have been demonstrated to be accurate in practice and are currently accepted to provide a path to licensure for all disease-causing serogroups, including serogroup B [20].

In the early phase I and II clinical studies of 4CMenB, functional antibody responses were assessed against 7–15 reference strains to estimate the potential efficacy of the protein-based vaccines [21‐23]. Strains used in these studies were chosen based on genetic diversity, geographic and temporal distribution, and representation of key pathogenic clonal complexes (Table 1). In later phase III studies, indicator strains were selected, each of which is sensitive to antibodies against only one of the four vaccine antigens on its surface, in order to assess bactericidal killing directed against one particular vaccine antigen [21‐23]. As an indicator strain representing bactericidal response directed primarily at the NHBA antigen was only available at the latter stages of the clinical program (strain M10713) [24], serum antibody responses to NHBA were measured by enzyme-linked immunosorbent assay (ELISA) against NHBA as a surrogate for hSBA in the earlier studies [21‐23, 25].

SBA titer of 4 was initially used [17]. Subsequent phase III studies used the more conservative limit of ≥5, which ensures that within assay variability with a >95 % confidence that the result was at least 4.

The first study of 4CMenB, reported by Toneatto et al. [23], was a randomized, single-center study conducted in Switzerland beginning in February 2006 in healthy adults aged 18–40 years (V72P5) to assess the immunogenicity and tolerability of the three selected recombinant protein antigens in the absence (rMenB) or presence of OMV from the New Zealand strain (NZ OMV, equivalent to the final 4CMenB formulation) or the Norwegian (NW OMV) outbreak strain. Participants were randomized to receive three doses of the respective formulations on a 0-, 1-, and 2-month schedule. Bactericidal responses were observed for participants in all three study arms against a diverse panel of reference strains. At least 75 % of subjects achieved hSBA titers ≥4 against 13 of 15, 10 of 15, and 7 of 15 strains after two doses of 4CMenB, rMenB + NW OMV, and rMenB, respectively. Both OMV-containing vaccines elicited immune responses against more meningococcal serogroup B strains than rMenB alone. The 4CMenB formulation also elicited a response to both OMVs tested, including the heterologous NW OMV, and was thus the formulation, along with rMenB, on which further development was focused.

A phase II study beginning in July 2007, reported by Kimura et al. [24], studied the immunogenicity and tolerability of three doses of 4CMenB and a single dose of the quadrivalent glycoconjugate against serogroups A, C, W, and Y (MenACWY-CRM; Menveo^®, Novartis Vaccines and Diagnostics) in laboratory workers in Germany and Italy highly exposed to meningococcal isolates. Participants received three doses of 4CMenB on a 0-, 2-, and 6-month schedule followed by a single dose of MenACWY-CRM 1 month after the last 4CMenB (V72P4; NCT00560313). Sera were tested against the fHbp, NadA, and PorA 1.4 indicator strains with 91–100 % achieving hSBA titers ≥4 after the second 4CMenB dose, with similar immunogenicity (range 92–100 %) after the third 4CMenB dose. The response to the subsequent MenACWY-CRM was as expected, with 83–100 % achieving hSBA titers ≥8 against serogroups A, C, W, and Y. Both vaccines were well tolerated, although reactogenicity was higher with 4CMenB than with MenACWY-CRM, but no serious adverse events associated with either vaccine were observed. This study indicated that a two-dose schedule of 4CMenB is sufficient in adults to provide robust immune responses against the vaccine antigens, although this observation needed to be confirmed in subjects whose occupation does not offer the possibility of priming through pre-exposure to a range of meningococcal strains.

Two phase II studies were conducted in parallel in UK infants to evaluate inclusion of OMV, in different vaccination schedules, and concomitant administration with routine infant vaccines. Both studies included groups who received 4CMenB or rMenB.

The first study, reported by Findlow et al. [21], was an open-label, randomized, controlled trial beginning in September 2006 in 147 healthy 2-month-old infants (V72P6; NCT00381615). Two groups of infants received either rMenB or 4CMenB at 2, 4, 6, and 12 months of age, while a third group of MenB vaccine-naïve toddlers received one dose of rMenB or 4CMenB at 12 months. In the cohort who received 4CMenB, 74–100 % achieved hSBA titers ≥4 against the fHbp, NadA, and PorA 1.4 indicator strains after the second dose and 85–95 % after the third dose. Responses in the rMenB cohort, with the exception of NadA, were lower both in terms of proportions achieving ≥4 and as geometric mean titers (GMT). Sera were also tested against a broader panel of MenB heterologous reference strains, which reinforced the immunologic benefit of including the OMV component. Percentages achieving hSBA titers ≥4 against these heterologous strains were lower than the indicator strains, but killing was evident across a broad panel of strains as early as 5 months of age. At 13 months of age, following a booster dose at 12 months, a majority (57–100 %) of infants receiving four doses of 4CMenB achieved hSBA titers ≥4 against six of seven strains, while the rMenB-only group elicited responses in over 20 % of subjects against three of seven strains (44/76-SL, 100 %; 5/99, 97 %; M01 240364, 70 %). Comparisons of titers after third and fourth doses, and in toddlers given their first dose of 4CMenB at 12 months of age, indicated that the increases in titers after the fourth dose of 4CMenB were anamnestic responses, suggesting this was a true booster response. At 5 months of age, following the second dose at 4 months, although percentages achieving hSBA titers ≥4 against reference strains were lower, killing was already evident across a broad panel of strains.

The persistence of these immune responses was examined in a small follow-up study (NCT01027351) at 40 months of age following vaccination of infants and toddlers at 2, 4, 6, and 12 months [25]. At 40 months of age, 41–76 % of the 17 children tested still maintained protective bactericidal antibody titers (titer ≥4) to each of the four vaccine components following 4CMenB (fHbp, 65 %; NadA, 76 %; NZ OMV, 41 %; NHBA, 67 %). When a booster dose of 4CMenB was administered, there was an anamnestic response to achieve protective titers against the four reference strains in 89–100 % of subjects 1 month later, showing that memory had been induced. Similarly high responses were observed against a panel of other strains, with 94–100 % achieving titers ≥4 against three strains, and 50 % against a fourth.

The second study, described by Snape et al. [22], was performed in 60 older infants who were vaccinated with two doses of either 4CMenB or rMenB beginning at 6–8 months of age with a third dose at 12 months of age (V72P9, NCT00433914), and their responses assessed against the same panel of seven strains as the V72P6 trial. Reflecting the results in younger infants, after two doses of 4CMenB most infants (67–100 %) achieved an hSBA ≥4 against six of the seven test strains, but this was only true for three of seven strains after rMenB (95–100 %). Persistence of the immune responses in this study were also investigated at 40 months of age [26], and were consistent with the previous study [25]; 36 % had titers ≥4 against fHbp, 100 % against NadA, 14 % against NZ OMV, and 79 % against NHBA after 4CMenB vaccination 2 years earlier. Persistence was lower in those vaccinated only with the experimental rMenB formulation. All subjects responded anamnestically to a booster dose of 4CMenB, and also demonstrated responses against four other non-reference strains as in the previous study.

In both of these infant phase II studies, the two vaccine formulations were generally well tolerated, with infrequent severe local or systemic reactogenicity, although with a modest increase observed in reactogenicity in the 4CMenB group [21, 22]. When present, fever was of short duration and no medically attended fever events occurred.

This phase I study in adults (V72P5) and early phase II studies in infants and toddlers (V72P6, V72P9) were the first indication that 4CMenB had the potential to confer a more robust, and/or broader range of protection against MenB strains compared with OMV-only or rMenB-only vaccines; including OMV enhanced the immune response over rMenB alone both in magnitude of the hSBA response and with a broader coverage of different strains, with protection evident across a wide range of ages, including infants. These data resulted in the 4CMenB formulation advancing into larger phase IIb/III trials.

For several traditional bacterial vaccines that target fixed and highly conserved antigenic epitopes, antibody titers correlate with protection against the majority of circulating strains. For MenB, antigenic variability and level of surface expression of outer membrane proteins presents a challenge for determining vaccine coverage against the myriad of circulating strains. The use of hSBA (as an immunologic correlate) enables detection of functional antibodies in serum, and, by extension, clinical protection. However, a limitation of hSBA is that it does not define how killing is mediated (i.e., which specific components the bactericidal activity is directed against). Nor can hSBA detect whether antibody-mediated killing against different antigens expressed on the bacterial surface can exert synergistic effects, which may influence vaccine efficacy [33]. Effectiveness of a protein-based vaccine against MenB will therefore depend on the presence and quantity of target antigens on the circulating strain. Any method predicting vaccine effectiveness must account for the genetic diversity of MenB strains, which, through gene regulation, phase variation, and sequence diversity, can affect the levels of expression of these surface antigens [33].

The meningococcal antigen typing system (MATS) uses polyclonal antibodies against fHbp, NadA, and NHBA in an ELISA to quantify the expression and amounts of these individual antigens, which cross-react with the corresponding antigen used in the vaccine. The PorA component is determined using PCR, as this is universally expressed and is only dependent upon the presence of the PorA 1.4 gene [33]. 4CMenB is expected to provide protection against strains that meet a minimum threshold of reactivity in the MATS ELISA and/or contain the PorA 1.4 antigen. MATS coverage estimates were developed using pooled sera from infants who received a three-dose series of 4CMenB with an additional dose at 12 months. MATS provides an estimate of vaccine coverage against a panel of strain isolates from a given region. In a large panel of invasive MenB isolates from across Europe, MATS determined that 78 % of MenB strains would be covered by vaccination with 4CMenB [34]. Predicted strain coverage rates in Europe using MATS are shown by country in Fig. 6, with additional data for non-EU countries.

Coverage estimates provided by MATS may be an underestimation since MATS does not account for the activity of antibodies from non-PorA OMV antigens, or for synergistic effects among the multiple vaccine components [33]. Furthermore, in adults and adolescents, MATS is considered to provide a conservative estimate of strain coverage because the breadth of bactericidal activity is expected to be higher than predicted by MATS, due to greater levels of cross-reactivity observed at these older ages [33]. Some infant studies have shown hSBA activity against non-reference strains [25, 26], but the small amounts of sera available from infants precludes analyses of infant coverage against panels of strains, except through pooling of sera as used in the MATS analyses. MATS-based predictions were recently validated in an analysis of a representative MenB strain panel from England and Wales, which showed that whereas MATS predicted a coverage of 70 % (95 % CI 55–85), the actual coverage based on killing by pooled infant and adolescent immune sera in hSBA was observed to be 88 % (95 % CI 72–95) [35]. However, the method is dependent upon pooled sera, and thus misses an assessment of individual seroresponse levels. Indications of individual or proportional protection of a population comes from the observed responses in clinical trials of individuals, which show that almost 100 % of individual infants and adolescents produced hSBA titers ≥4 or 5 against three or more of the vaccine antigens, noting that the use of hSBA titers ≥4 or 5 as indicators of protective responses is itself a surrogate for protection [17‐19]. However, data from the UK following the success of the polysaccharide-protein conjugate vaccines against meningococcal serogroup C, and the use of OMV serogroup B vaccines against outbreaks in Cuba, Brazil, and Norway, appear to confirm the utility of this surrogate.

MATS is a powerful tool that will allow the assessment of the likely effectiveness against individual outbreak strains or, where adequate surveillance is in place, strain coverage predictions for MenB endemic disease in a given country or geographic region. Information collected using MATS will be useful for ongoing global surveillance of pathogenic bacterial isolates through established reference laboratories via a public–private partnership [36]. MATS can provide ongoing data on temporal and geographic changes in MenB epidemiology, and potential development of escape mutant strains as a result of selective pressure from vaccination. Additional assessments are ongoing but it may be that MATS may also be applied to assess 4CMenB coverage rates against other (non-B) meningococcal serogroups. Lastly, MATS provides data that could assist in cost-effectiveness or cost/benefit analyses of 4CMenB if implemented in national immunization programs.

The 4CMenB clinical development program has established an acceptable safety and tolerability profile for the vaccine across all age groups. The post-vaccination increase in fever will require education of parents/caretakers on appropriate management. Nevertheless, as with all newly approved vaccines, additional long-term studies are needed to demonstrate a more comprehensive picture of the vaccine’s safety profile. Post-licensure studies of 4CMenB will monitor long-term safety events across a large cohort of vaccine recipients. The full spectrum of clinical protection provided by 4CMenB against meningitis, invasive disease, and fulminant septicemia will be determined by well designed surveillance programs in place in countries before, during, and after introduction in national vaccine programs. Moreover, as new vaccines are developed, particularly for infants and young children, the safety and efficacy of 4CMenB used concomitantly in multi-vaccine regimens will require evaluation. Future efficacy considerations will involve monitoring the emergence of regional strain variants, since the selective pressure of 4CMenB could potentially drive the emergence of escape mutants and may reduce vaccine efficacy. MATS is expected to be an efficient tool for monitoring MenB strain variations. Currently, strain panels specific for defined geographical regions are being developed and are expected to hasten the prediction of strain coverage in these regions [35].

Herd immunity has been demonstrated to be an important factor in the success of the polysaccharide-protein conjugate vaccines against meningococcal serogroup C, as well as other bacterial causes of infant meningitis and septicemia (Haemophilus influenzae type b, Streptococcus pneumoniae). This will only become apparent when active surveillance programs in place (once the vaccine is widely used in specific populations) provide effectiveness data. A preliminary report of a recent study on the impact of 4CMenB and a quadrivalent MenACWY vaccine on nasopharyngeal carriage in university students in England (clinicaltrials.gov NCT01214850) demonstrated proof of principle of an effect on serogroup B carriage [37], but the results have yet to be fully reported and further evaluated.

Other protein-based MenB vaccines are in development using different approaches compared with 4CMenB [38]. The most advanced of these is the bivalent rLP2806 vaccine being developed by Pfizer, which has been through phase II and is currently being evaluated in phase III trials [39]. This vaccine is based on lipidated forms of two immunologically distinct sub-families (A and B) of the fHbp protein, which is expressed in the vast majority of MenB strains. For comparison, 4CMenB includes an fHbp fusion protein from the B sub-family. Phase I and phase II data indicate that the rLP2806 vaccine elicits a strong immune responses in adolescents, 75–100 % achieving hSBA titers ≥4 against a panel of clinical isolates [39]. As of this writing, the trials in which the vaccine is being studied include subjects down to the age of 10 years, with no indication yet of age de-escalation to infants and toddlers, so it remains to be seen whether the vaccine will be appropriate for these important age groups.

In summary, 4CMenB is a multi-component vaccine containing three broadly conserved surface-expressed recombinant antigens and a specific OMV, thereby providing broad protection against circulating heterologous strains of MenB. The 4CMenB formulation has proven to be immunogenic in adults, adolescents, and young infants with a well characterized tolerability profile in clinical trials performed in the most susceptible age groups. Incremental increases in reactogenicity (including fever) compared with routine childhood vaccinations when 4CMenB is added to the routine schedule have been observed. Concomitant administration with routine vaccines (DTaP-HBV-IPV/Hib, PCV7, MMRV, and rotavirus vaccines) does not appear to interfere with the immunogenicity of any of the vaccines, but the use with more recently introduced vaccines, e.g., PCV13, or other routine vaccines such as MenC conjugates will need to be investigated. The pyrogenic profile observed after vaccination with 4CMenB in infants and toddlers tends to be of short duration, does not increase with successive doses, and few vaccinees experienced temperatures ≥40 °C. As with all new vaccines, post-marketing surveillance will be conducted to monitor for any other safety signals.

Estimating strain coverage of 4CMenB is crucial for public healthcare policy; MATS will provide valuable information for policy makers and healthcare workers and inform decisions relating to vaccination programs against MenB at a population level, which has important implications in shaping public health policy regarding IMD caused by MenB. For the first time, a vaccine has the potential to limit the devastating effects of IMD caused by heterologous serogroup B strains.