Correlates of protection for meningococcal surface protein vaccines: lessons from the past

ABSTRACT

Introduction

Recombinant surface protein meningococcal serogroup B (MenB) vaccines are available but with different antigen compositions, leading to differences between vaccines in their immunogenicity and likely breadth of coverage. The serology and breadth of coverage assessment for MenB vaccines are multifaceted areas, and a comprehensive understanding of these complexities is required to appropriately compare licensed vaccines and those under development.

Areas covered

In the first of two companion papers that comprehensively review the serology and breadth of coverage assessment for MenB vaccines, the history of early meningococcal vaccines is considered in this narrative review to identify transferable lessons applicable to the currently licensed MenB vaccines and those under development, as well as their serology.

Expert opinion

Understanding correlates of protection and the breadth of coverage assessment for meningococcal surface protein vaccines is significantly more complex than that for capsular polysaccharide vaccines. Determination and understanding of the breadth of coverage of surface protein vaccines are clinically important and unique to each vaccine formulation. It is essential to estimate the proportion of MenB cases that are preventable by a specific vaccine to assess its overall potential impact and to compare the benefits and limitations of different vaccines in preventing invasive meningococcal disease.

KEYWORDS:

• Correlates of protection
• invasive meningococcal disease
• Neisseria meningitidis
• vaccine
• vaccine effectiveness

1. Introduction

Neisseria meningitidis is a Gram-negative bacterium that can cause invasive meningococcal disease (IMD), a life-threatening condition [1,2]. IMD most frequently presents as meningitis and/or septicemia and can cause debilitating, long-term sequelae among survivors. Although 12 meningococcal serogroups have been identified, five (A, B, C, W, and Y) have historically caused almost all IMD cases worldwide [3,4]. Meningococcal serogroup B (MenB) is currently a major cause of IMD in most Western regions [3]. Other less common disease-causing serogroups are also identified, such as serogroup X, which has caused outbreaks in sub-Saharan Africa [3,5].

Because of the significant mortality and morbidity of IMD even with optimal treatment, coupled with its sudden onset, rapid clinical course, and difficulty in diagnosing cases at an early stage, preventive vaccination is likely the most effective method for protecting against IMD [1]. Vaccines are now available for the prevention of disease caused by the five most prevalent meningococcal disease-causing serogroups either individually or in combination (i.e., serogroups A, B, C, W, and Y) [2,6]. Others, including vaccines targeting the serogroup X polysaccharide, are under development (e.g., NCT03295318).

Consideration of the meningococcal cell surface structure is critical to understanding the development and assessment of meningococcal vaccines. Capsular polysaccharides are major virulence factors and provide protection against host-mediated bactericidal activity; with very few exceptions, invasive strains express capsules, with strains distinguished by differences in the serogroup-specific structure of the capsular polysaccharide [1,7–9]. These differences in composition, therefore, have important implications for vaccine development.

Efficacious vaccines using capsular polysaccharides conjugated to a protein carrier have been developed for meningococcal serogroups A, C, W, and Y [2]. Unlike the serogroup A, C, W, and Y polysaccharides, the MenB capsular polysaccharide is poorly immunogenic because of homology to the self-antigen neural cell adhesion molecule and may induce auto-antibodies [10,11]; therefore, early MenB vaccines were based on outer membrane vesicles (OMVs) rather than capsular per Merriam-Webster [10,12,13]. More recently, recombinant surface protein MenB vaccines have become available, with initial licensure of 4CMenB (Bexsero®, MenB-4C; GlaxoSmithKline Vaccines, Srl, Siena, Italy) in the European Union in 2013 and MenB-FHbp (Trumenba®, bivalent rLP2086; Pfizer Inc, Philadelphia, PA, USA) in the United States in 2014 [14,15]. 4CMenB and MenB-FHbp have different antigen compositions, which has led to considerable differences between the vaccines in the immunogenicity and the breadth of coverage they are likely to provide.

Clinical efficacy trials are not feasible because of the low incidence of IMD and therefore are not used in meningococcal vaccine development and licensure [16]. Instead, serological correlates of protection have to be used for this purpose [17]. The serology and breadth of coverage assessment for MenB vaccines are multifaceted areas, and a comprehensive understanding of these complexities is required to appropriately compare licensed vaccines and those under development. This is the first of two companion papers that comprehensively review the serology and breadth of coverage assessment for MenB vaccines. Using a targeted literature search of PubMed databases without date limits, the focus herein is on the history of the earlier meningococcal vaccines to identify the transferable lessons applicable to the currently licensed and future MenB vaccines and their serology, which are the focus of the companion paper.

2. Meningococcal vaccines

The first efficacious meningococcal vaccines were developed in the 1960s against serogroups A and C and were based on the capsular polysaccharide; quadrivalent vaccines against serogroups A, C, W, and Y (MenACWY) followed [4]. Although these were safe and immunogenic in older children and adults, they were generally poorly immunogenic in the very young, did not induce immunologic memory, and were associated with hypo-responsiveness on repeated dosing [4,18–20].

Conjugation of the polysaccharide antigen to a protein carrier was subsequently shown to address the shortcomings of plain polysaccharide vaccines and, unlike polysaccharide vaccines, these vaccines were suitable for use in infants and interrupted transmission, thereby resulting in herd effects [4]. Monovalent serogroup C polysaccharide-protein conjugate vaccines were initially developed, and the first of these was licensed in 1999 [21]. In 2005, the first MenACWY conjugate vaccine was licensed and has since been followed by three other formulations [2,22–24].

Because of the aforementioned limitations of the MenB polysaccharide [4,25], researchers searched for the ‘holy grail’ antigen: one that was harbored and conserved across MenB disease-causing strains, was expressed on the bacterial surface and accessible to antibodies, and could induce a bactericidal immune response [26]. Unfortunately, no such antigen has yet been found. This necessitated the use of surface-expressed antigens that induce serum bactericidal antibody (SBA) activity, but are diverse and/or not harbored by all strains.

In lieu of identifying potential broadly protective antigens, the first efficacious MenB vaccines were based on OMVs produced using outbreak strains. These contained hundreds of outer membrane antigens but no serogroup B polysaccharide, which is removed during purification and detoxification steps [27–29]. Three monovalent OMV vaccines were developed and subsequently licensed in Cuba (used in Cuba and Brazil), Norway (used in Norway and France), and New Zealand [10,12,13,30–32]. Despite the success of these vaccines in preventing MenB disease associated with single clonal outbreaks [10,12,13,30,31], multiple doses are required [13,33], protection is of relatively short duration [33], responses are age dependent [10], and the vaccines do not provide breadth of coverage across diverse MenB strains [17]. The limited breadth of coverage afforded by these OMV vaccines is because the immunodominant component, the Porin A protein (PorA), is highly diverse (360 and 988 unique sequences identified for variable regions 1 and 2, respectively [34,35]), but there is limited cross-reactivity to other PorA variants [36].

To broaden coverage against diverse MenB strains, multivalent PorA/OMV vaccines have been developed, including bivalent, hexavalent, and nonavalent formulations [37–40]. None of these, however, have progressed beyond clinical testing [36]. Because of the limitations of OMVs and the immunodominance of the PorA antigen, researchers have continued the search for antigens eliciting the characteristics of the putative protective responses described above for the next generation of MenB vaccines.

The first surface protein vaccine to be successfully licensed with the potential to provide broad coverage was 4CMenB [41]. Potential antigens were identified from genomic sequencing of an MenB strain with factor H binding protein (FHbp), Neisserial Heparin Binding Antigen (NHBA), and Neisseria adhesin A (NadA) eventually selected as the most promising [42]. The FHbp and NHBA antigens are each combined with accessory proteins also identified from reverse vaccinology to create two fusion proteins to improve immunogenicity [43]. To increase immunogenicity and provide broader MenB strain coverage, the final formulation of 4CMenB combines these recombinant proteins with OMVs from the New Zealand outbreak strain NZ 98/254 [42].

The second surface protein vaccine licensed was MenB-FHbp [15]. Traditional vaccinology approaches independently and concurrently identified FHbp as a promising antigen. Pivotal early studies indicated that lipidation of FHbp increased immunogenicity and antibody responses cross-reactive to different FHbp variants [26,44,45]. Despite the benefit of lipidation, to provide coverage across a broad diversity of FHbp variants and therefore across MenB strains, it was necessary to include two FHbp antigenic variants in the final formulation, one from each of the two antigenically distinct FHbp subfamilies identified [46,47].

The two surface protein vaccines can potentially be considered as from the same vaccine class. However, these two vaccines are very different in their antigen compositions, and therefore, the immune response they induce and the breadth of coverage they provide likely differ. This is in contrast to previous meningococcal vaccines, in which vaccines in plain polysaccharide or conjugate classes were broadly similar to each other.

3. History of the meningococcal surrogate of protection: the serum bactericidal antibody assay

Initial reports of protection against IMD being mediated by SBA activity were nearly 100 years ago [48]; multiple studies have since confirmed these initial findings [49–51]. Consequently, the SBA assay, which measures complement- and antibody-mediated bacterial lysis of a test strain, has become the ‘gold standard’ assay and surrogate indicator of protection for vaccines across all meningococcal serogroups [9]. The importance of the SBA assay is further highlighted by the fact that meningococcal vaccines have been licensed on SBA data alone, as traditional efficacy trials are not feasible because of the low incidence of IMD, which would necessitate a prohibitively large sample size to demonstrate vaccine efficacy [16].

In 1969, seminal work by Goldschneider and colleagues showed the correlation between SBA in human sera and protection from meningococcal serogroup C (MenC) disease in military recruits [49]. They used the quantitative SBA assay using human complement (hSBA) in which the highest reciprocal serum dilution that killed ≥50% of the bacteria in the assay was reported as the SBA titer. By comparing MenC hSBA titers of recruits who subsequently went on to develop MenC disease with those from recruits who did not, a protective hSBA titer of ≥1:4 was proposed as the correlate of protection. Over the last half a century, the hSBA titer of ≥1:4 correlate has been validated for serogroup B, applied to all meningococcal groups, and used for vaccine licensure [33,52,53].

A major difficulty of the hSBA assay is procurement of exog-enous human complement sources from the few individuals who do not have intrinsic SBA against meningococci but who have normal complement hemolytic activity [54,55]. An additional challenge is the need to ensure standardization between donations from the same donor and between donors [56]. One possibility is the use of endogenous complement within each clinical sample, which circumnavigates the need for exogenous sources. However, this introduces extra variability into the assay and results in an inability to assay dilutions >1:4 as exogenous complement becomes insufficiently concentrated at these greater dilutions. A potential solution to the difficulties of obtaining exogenous human complement is development and validation of the SBA assay using rabbit complement (i.e., rSBA assay), which is readily available in large amounts and possesses an inherently low level of intrinsic bactericidal activity [50,55]. The rSBA assay was recommended by the World Health Organization in 1976 for the evaluation of meningococcal serogroup A (MenA) and MenC vaccines [57], and the assay has continued to be used for the evaluation of MenACWY vaccines.

For the MenB SBA assay specifically, only human complement is appropriate because rabbit complement causes higher SBA titers than those obtained with human complement alone [16,58,59]. These differences are thought to be partially attributed to the presence of anti-MenB polysaccha-ride immunoglobulin M (IgM) in clinical samples, which is functional with rabbit but not human complement [60]. Similarly, unlike human factor H (FH), rabbit FH is not bound by meningococcal FHbp and therefore does not downregulate complement activity; this, in turn, likely leads to increased rSBA titers for serogroup B and other meningococcal serogroup strains [51,61]. An additional factor to consider in the choice of test strains and complement source is that expression of FHbp and other FH-binding molecules can differ between meningococcal strains [62–64]. Of note, MenB meningococci differ in binding FH compared with other serogroups because sialic acid in the MenB polysaccharide renders complement protein 3 (C3) more accessible to the down-regulatory action of factors H and I and thus enhances inactivation of C3 [60,65]. These discrepancies explain why hSBA is the only acceptable assay for evaluation of MenB vaccines.

4. Interpretation of serum bactericidal antibody assay data and the correlate of protection

An hSBA titer of ≥1:4 remains the only internationally accepted correlate of protection against MenB IMD [9,54]. This cutoff, originally proposed in 1969, was subsequently validated for protection against MenB disease in a large Norwegian study and remains an integral component of assessment of meningococcal vaccine before licensure [33,49,52,53].

The SBA assay methodology does not incorporate a standard serum from which the SBA titer is calculated, therefore potentially resulting in differences between the titers determined in different laboratories [66]. This is exemplified by a study in which methodologic differences between four laboratories resulted in large discrepancies in the proportions of samples with MenB hSBA titers ≥1:4. Subsequent harmonization of assay conditions among the laboratories resulted in generally good agreement between hSBA assay results. Despite this knowledge, lack of harmonization and standardization of MenB SBA assay methodology across laboratories worldwide continues. Other developments have seen the use of a colorimetric method for bacterial enumeration to increase assay throughput [67–69]. The impact of this divergence from traditional methodology on SBA titer results compared with other laboratories continuing to use standard techniques remains to be fully elucidated.

The methodology used for SBA titer calculation also differs between laboratories and can impact SBA assay results [26,70,71]. For example, the 50% killing threshold for titer determination can be calculated from the number of colonies at the start (T₀) or end of the assay incubation (e.g., T₆₀) period [72]. Because it is usual for bacterial growth to continue over the incubation period, there is a tendency for higher SBA titers when calculated from T₀ compared with those calculated from T₆₀. This can lead to higher proportions of samples being reported as achieving hSBA titers ≥1:4 or any higher cutoff applied when using T₀ data.

Another way in which the SBA titer calculation differs is the use of interpolation to determine the exact value/titer at which the 50% killing threshold is achieved in the SBA [73,74], resulting in titers being assigned on a continuous scale (e.g., 5, 6, and 7); in contrast, the reciprocal dilution approach will always be more conservative and the doubling dilution will always be lower than the interpolated titer (e.g., 4 vs 5, 6, or 7). The use of interpolation has the associated impact of both higher estimates of protection [72] and an increased geometric mean titer (GMT).

The hSBA titer of ≥1:4 cutoff has traditionally been applied to meningococcal surface protein vaccine evaluation and described as ‘putatively protective’ or ‘sero-protective’ [10,12,75]. More recently, higher cutoff titers have been applied for varying reasons. An hSBA titer ≥1:5 cutoff, only feasible when using interpolation, has been used during the later development of 4CMenB as it was reported that the higher value represented 95% confidence that the individual had an hSBA titer ≥1:4 [73,74]. Higher cutoffs for some target strains have also been necessitated by hSBA assay validation, showing that the lower limit of quantification (LLOQ) is greater than the 1:4 correlate. This explains why 1:8 and 1:16 cutoffs have been used for MenB target strains for MenB-FHbp evaluations [76]. The application of these higher cutoffs can result in significant numbers of individ-uals being categorized as negative or unprotected who actually have hSBA titers ≥1:4 and, therefore, could be considered protected (Table 1) [77]. Furthermore, it must also be considered that the hSBA titer ≥1:4 correlate was already deemed to be a conservative measure [49], and hSBA titers <1:4 are not necessarily indicative of lack of protection [78].

Table 1. Numbers and percentages of sera showing hSBA titers ≥1:4 and ≥1:8 against the different meningococcal strains before and after hexavalent OMV vaccination [77]

Bleed	hSBA titer cutoff	hSBA target strain (PorA)
		M00 242,966 (P1.19,15–1)^a	M00 243,286 (P1.19–1,15–14)	M00 242,768 (P1.19–1,15–13)	M00 243,089 (P1.19–12,15–11)	M01 241,666 (P1.19–4,15)
Prevaccination	≥1:4	8 (28%)	2 (12%)	4 (14%)	0 (0%)	14 (49%)
	≥1:8	4 (14%)	1 (6%)	0 (0%)	0 (0%)	6 (21%)
Postvaccination	≥1:4	24 (92%)	12 (92%)	14 (70%)	12 (63%)	22 (88%)
	≥1:8	23 (88%)	12 (92%)	5 (25%)	7 (37%)	20 (80%)

hSBA = serum bactericidal antibody using human complement; MenB = meningococcal serogroup B; OMV = outer membrane vesicle; PorA = porin A.

^aMatches the vaccine.

The main disadvantage of applying hSBA titer cutoffs to the evaluation of meningococcal subcapsular vaccines is that individuals included in the study may have preexisting, naturally acquired SBA against the MenB target strain. If an individual has an hSBA titer greater than or equal to the cutoff prevaccination, an immune response cannot be determined by this measure because the individual is already considered ‘protected.’ One approach to this sensitivity issue is to select hSBA target strains with low ‘baseline positivity,’ but this may result in nonrepresentative strains and, therefore, nonrepresentative results; therefore, strain selection is an important consideration as discussed below [79].

Alternative approaches for analyzing SBA data have been used. Following on from the recommendation for meningococcal polysaccharide vaccines [52], the proportion of individuals with a ≥4-fold rise in hSBA titer from prevaccination to postvaccination has been routinely used for surface protein vaccines [41,80]. Unlike absolute SBA titers, ≥4-fold rises have been shown to be generally constant between laboratories [81]. However, ≥4-fold rises in hSBA titers can only be estimated using exogenous complement and may be calculated using different methods [32,54,55,70,71,82,83] and are dependent upon similar definitions of a 4-fold rise across analyses. Early OMV vaccine studies would accept any ≥4-fold rise regardless of the minimum postvaccination titer [70,82,83]; later studies were more conservative, requiring a minimum postvaccination titer predicated by the assay LLOQ (Table 2) [32,70,71].

Table 2. Effect of different calculation methods on the percentages of individuals (>18 years of age) achieving ≥4-fold rises in hSBA titer from baseline to postvaccination with MenBvac [70]

hSBA target strain	Percentages with ≥4-fold rise in hSBA titer
	Calculation method 1^a	Calculation method 2^b
44/76-SL	75.0	62.5
NZ 98/254	52.2	34.8
M01 240,013	34.8	34.8
M01 240,101	17.4	8.7
M01 240,149	26.1	17.4
M01 240,185	17.4	17.4
M01 240,355	34.8	34.8

hSBA = serum bactericidal antibody using human complement; MenBvac = Norwegian outer membrane vesicle vaccine.

^aIncludes all individuals with ≥4-fold rises in hSBA titer (including individuals with a baseline titer of <1:2 achieving a titer of ≥1:4 after the third dose of MenBvac).

^bOnly includes individuals with ≥4-fold rises in hSBA titer and a titer ≥1:8 after the third dose of MenBvac (i.e., for individuals with a baseline titer of <1:2, a titer of ≥1:8 had to be attained after the third dose of MenBvac).

Another way in which vaccine responses have been evaluated is by calculation of GMTs and comparison of prevaccination and postvaccination values [76,84]. Being a fully quantitative measure, GMTs allow full interpretation of a vaccine-induced immune response and are particularly useful for determination of the greater anamnestic responses compared with that of primary responses [33]. Other approaches used less frequently for measurement of vaccine response have included the assessment of fold rises in both individual titers and in GMTs [85,86].

5. Alternative immunoassays for measuring antibody response and determining protection

Among invasive bacterial infections provoked by meningococci, pneumococci, and Haemophilus influenzae, protection against IMD seems to be specifically linked to complement-mediated lysis; increased susceptibility to IMD is more commonly associated with the terminal complement pathway than for pneumococci and H influenzae [87]. This explains why the SBA assay is the only accepted surrogate of protection provided by vaccines against meningococcal disease. Nonfunctional assays measuring antibody responses have also been investigated as potential alternatives. These laboratory measurements may correlate with protection and thereby provide a surrogate of protection, but do not measure protective SBA titers [58]. For example, the enzyme-linked immunosorbent assay (ELISA) can be used to quantify a specific immunoglobulin G (IgG) or IgM, but cannot determine the proportion of IgG or IgM that is functional [9,56]. Furthermore, antibody paratopes and the target bacterial epitopes can also be denatured by ELISA reagents and inhibit antigen-antibody interactions [9].

Opsonophagocytosis is a type of bactericidal activity against meningococci, and complement-mediated antibody-dependent opsonophagocytic activity (OPA) is the correlate of protection against invasive pneumococcal disease [9,70]. However, the use of OPA as a correlate of protection against meningococcal disease has not been established. In contrast to the SBA assay, which requires specific antibody and activation of the entire complement cascade for lysis of the bacterium, only a portion of the complement cascade is utilized for OPA [9]. This potentially explains the weak correlation between MenB OPA and hSBA assay results, causing OPA assays to be restricted to research studies [70].

Unlike the SBA assay, the whole-blood assay (WBA) measures complement-mediated lysis along with bactericidal activity of blood, including phagocytosis [70,88]. The WBA has been proposed as being more sensitive than the SBA assay and uses the individual’s own blood. The main disadvantage of the WBA is that it requires significant volumes of fresh blood and is challenging to standardize and control, precluding its routine use as a correlate of protection against IMD [70,88,89].

6. The fundamental importance of breadth of coverage of meningococcal surface protein vaccines

In contrast to meningococcal surface protein vaccines, the breadth of coverage of capsular polysaccharide vaccines does not need any consideration because each capsular group’s polysaccharide is invariant except for potential minor differences in acetylation status, and capsule expression is a prerequisite for bacterial survival within the immunocompetent human host [17,90,91]. The breadth of coverage thereby provided by capsular polysaccharide vaccines is ubiquitous to all invasive strains with the corresponding capsule, irrespective of subcapsular diversity and genetic lineage. As a consequence, the serologic evaluation of these vaccines has also been relatively uncomplicated and primarily used a single SBA target strain for each capsular polysaccharide in each vaccine [51]. The MenC strain C11 exemplifies this uncomplicated SBA approach; this strain is used for evaluation of MenACWY vaccines and was previously used for evaluation of monovalent MenC and plain polysaccharide vaccines, having first been used by Goldschneider and colleagues >50 years earlier [49,92,93].

If the protein(s) used in meningococcal surface protein vaccines were harbored by, expressed sufficiently on, and invariant across all invasive MenB strains, the breadth of coverage assessment and serologic evaluation of MenB vaccines would be equally straightforward. However, no such protein has been found, and OMV, 4CMenB, and MenB-FHbp vaccines use proteins that, to various extents, do not meet these ideal criteria. Consequently, meningococcal surface protein vaccines are not expected to provide protection against all invasive MenB strains. The breadth of coverage afforded (i.e., the proportion of MenB strains that the vaccination is expected to protect against) varies depending on the vaccine formulation, may vary across regions because of geographic differences in circulating disease-causing strains, and varies by age group [94–97].

Monovalent OMV vaccines were primarily developed in response to clonal outbreaks of MenB disease and were, therefore, intended to provide protection against the specific outbreak clone [17]. Immunogenicity assessment of these vaccines was undertaken using an outbreak strain in the hSBA assay, which was usually the same strain used to produce the OMVs [10,30]. Results from these hSBA assessments provided information on the immunogenicity and protection afforded against the outbreak clone but no information on the breadth of coverage against other MenB strains. This was generally not a concern, as these vaccines were primarily intended to provide protection against the specific clonal outbreak of disease. Because both 4CMenB and MenB-FHbp were developed to provide broad protection [26,98], assessment and quantification of breadth of coverage are crucial to understanding their utility. This is especially important as MenB disease in most localities is caused by unrelated and genetically diverse strains [99].

The first consideration for assessing breadth of coverage of meningococcal surface protein vaccines is that not all MenB strains harbor the genes for the proteins used as vaccine antigens, meaning that no protection will be afforded by the corresponding vaccine component against such strains. The percentage of strains with the presence or absence of genes varies by protein. For example, although <0.1% of European MenB isolates in a strain collection from 2006–2008 did not harbor the fHbp gene, 77.7% were missing the NadA gene [95]. This directly influences the breadth of coverage assessment for this particular vaccine component; no matter how immunogenic an NadA antigen may be, at best, it will only be able to provide protection against (i.e., cover) 22.3% of this European MenB isolate collection.

The second consideration is that for the strains that do harbor the protein gene, these genes have a significant number of sequence variants (Table 3) [34,100–102]. These sequences encode many diverse allelic forms of each protein that, depending on their antigenicity, affects the ability of vaccine antigen-derived immune responses to recognize these related but different structures. For example, a single antigenic variant of FHbp, NadA, NHBA, or PorA is unable to provide clinical cross-protection to all other variants of the same protein, and the breadth of cross-reactivity is antigen specific [95,103–105].

Table 3. Known diversity of major surface-expressed meningococcal proteins, including those used in current surface protein vaccines [34,100–102]

Protein/antigen	Alleles, n^a
PorA VR2	988
FHbp	1728
NHBA (NEIS2109)	1706
NadA (NEIS1969)	314

FHbp = factor H binding protein; NadA = Neisserial adhesin A; NHBA = Neisseria heparin-binding antigen; PorA = porin A; VR = variable region.

^aAs of October 2020 (https://pubmlst.org/).

The third consideration is that the level of expression of proteins on the bacterial surface can vary between strains [106]. A vaccine can only provide protection against a strain if the protein is sufficiently expressed on the surface to enable SBA to bind and initiate bacterial lysis. Furthermore, in a small number of cases, these genes may be interrupted or contain premature stop codons, resulting in a lack of expression and an associated lack of coverage [95,107–109].

The fourth consideration is independent of the strain and associated with how cross-reactive the immune response is to different protein variants. Immunogenicity can be influenced by the protein variant used and can vary for the same protein variant in its presentation and its genotype [95,103]. For example, FHbp has been formulated to have a fused accessory protein to improve overall immunogenicity (i.e., in 4CMenB) or a lipid tail to enhance the breadth of response to other FHbp variants (i.e., in MenB-FHbp) [26,103,110]. Different antibody populations have also been shown to work in a synergistic manner and can affect breadth of coverage [111]. This is a factor that should be considered as all licensed meningococcal surface protein vaccines have included more than one antigen. To complicate matters further, the cross-reactivity of immune responses to protein antigens may be related to age. This was most clearly shown for 4CMenB in which the SBA response was broader for toddlers and older age groups compared with that in infants [75,112–114].

All of these considerations are unique to meningococcal surface protein vaccines and must be collectively considered when evaluating their breadth of coverage against diverse MenB strains. The serologic evaluation of these vaccines must, therefore, account for the diversity of MenB strains to provide the necessary information on breadth of coverage and a vaccine’s potential utility in a population. Basing a breadth of coverage assessment on only one or a nonrepresentative set of strains will undoubtedly result in bias, irrespective of the methodology used. It is unfeasible to test large numbers of MenB strains in the hSBA, partly because of clinical serum volume limitations, a factor exacerbated in pediatric studies [115]. A balance must therefore be achieved that provides appropriate age-related information on breadth of coverage without testing every MenB strain. To ensure that balanced results are achieved, it is also important to ensure that results from each individual strain are not biased by assay conditions or by the strain itself. For example, the use of strain expressing a nonrepresentative high amount of a protein would give a positive bias to results.

Determination of vaccine effectiveness after large-scale vaccine implementation has traditionally been undertaken for meningococcal vaccines [116,117]. For meningococcal surface protein vaccines, this does not provide as much information as initially perceived because such data will only confirm effectiveness against disease-causing MenB strains for the location and age group in question, which may not translate to effectiveness in other locations or age groups. It is, however, possible to determine gaps in the breadth of coverage in which breakthrough cases occur in immunocompetent individuals as has been recently shown with 4CMenB use in England [118]. The ability to select strains for the hSBA assay and the knowledge that hSBA correlates with effectiveness result in a situation in which the breadth of protection of meningococcal surface proteins can be fully elucidated within the laboratory.

7. Lessons learned from OMV vaccine immunogenicity

Over the past 30 years, a significant amount of experience has been gained by assessing the breadth of coverage for OMV vaccines. Many studies have investigated the immunogenicity of the different OMV formulations confirming the immunodominance of PorA and that responses raised against a PorA variant are generally not cross-reactive to others [36]. What is often overlooked, however, is that proteomic analyses have identified >100 different components within OMVs, of which many may induce an immune response [27]. For example, a clinical study of the New Zealand OMV vaccine, later included as a component of 4CMenB, showed the induction of IgG responses against many identified and unidentified components and significant increases against outer membrane protein 85 (Omp85), ferric enterobactin receptor A (FetA), PorA, porin B (PorB), 37-kDa ferric binding protein (FbpA), reduction-modifiable protein meningitidis (RmpM), and lipopolysaccharide (LPS) components of the vaccine strain [119].

The importance of these responses to minor OMV components is 2-fold. First, these responses contribute to overall immunogenicity and, therefore, the protection afforded by the OMV vaccine against the outbreak strain [115]. It is inappropriate to oversimplify the perceived immunogenicity of OMV vaccines to only being driven by PorA despite its overall immunodominance. This aspect generally does not need consideration if an OMV vaccine is being evaluated for use against a clonal outbreak because the vaccine, outbreak, and SBA strains are usually the same. In such cases, SBA responses against antigens other than PorA are important for protection against the outbreak strain [120]. An overestimation of breadth of coverage can arise when responses measured against a vaccine and SBA strain are considered representative for other nonclonal strains with the same PorA but differences in other OMV components of the vaccine. This was shown in multiple studies for the OMV component used in 4CMenB, whereby lower hSBA GMTs and smaller proportions of the tested population had hSBA titers ≥1:4 when measured against an unrelated PorA P1.7–2,4 strain rather than against the vaccine strain [112,120,121]. The use of the OMV vaccine strain in the hSBA assay to determine immunogenicity remains standard practice and is appropriate when considering protection against the outbreak strain; however, such data may not be transferable to other unrelated strains expressing the same PorA and may hinder the breadth of coverage assessment.

The second issue is that immune responses to these minor OMV components may provide protection against MenB strains containing a different PorA, which the OMV-derived PorA antibodies do not recognize [70,119]. A study of the Norwegian OMV vaccine confirmed this when it reported that 17.4% to 52.2% of individuals achieved ≥4-fold rises in hSBA titers against six diverse PorA-expressing strains (Table 4) [70]. Similar results were reported with a bivalent OMV vaccine; 27.9% to 46.3% of individuals achieved ≥4-fold rises in hSBA titers against three diverse PorA heterologous strains [37]. A study with the New Zealand OMV vaccine showed that it induced a 6-fold rise in median hSBA titer to the 44/76-SL strain, which had an unrelated PorA and genetic lineage [119]. The importance of these non-PorA responses to the breadth of coverage afforded by OMV vaccines will vary depending on the minor antigens in the OMV compared with those present in the circulating invasive strains.

Table 4. Individuals (>18 years of age) showing ≥4-fold rises in hSBA titer from before to after MenBvac administration against different serogroup B target strains [70]

hSBA target strain	Phenotype	Sequence type	Clonal complex	≥4-fold rise in hSBA titer, n/N (%)^a
Homologous vaccine strain
44/76-SL	B:15:P1.7,16	32	32	18/24 (75)
Heterologous strains
NZ 98/254	B:4:P1.7–2,4	42	41/44	12/23 (52.2)
M01 240,013	B:NT:P1.22,9	275	269	8/23 (34.8)
M01 240,101	B:NT:P1.19–1,15–11	1049	269	4/23 (17.4)
M01 240,149	B:4:P1.7–2,4	41	41/44	6/23 (26.1)
M01 240,185	B:2a:P1.5–1,10–8	11	11	4/23 (17.4)
M01 240,355	B:1:P1.22,14	213	213	8/23 (34.8)

hSBA = serum bactericidal antibody using human complement; MenBvac = Norwegian outer membrane vesicle vaccine.

^aIndividuals with ≥4-fold rises in hSBA titer from before to after third dose of MenBvac.

8. Lessons learned with OMV vaccine hSBA target strains

The choice of target strain has repeatedly been shown to be instrumental to the outcome of an hSBA assay. The lessons learned with OMV vaccines are important to consider currently and are applicable to all meningococcal surface protein vaccines. For reasons that are not fully understood, strains that are perceived to be identical or at least similar can produce significantly different results in the hSBA assay. The first report of this was with the hexavalent OMV vaccine in which immunogenicity was initially underestimated because the vaccine strain resulted in lower hSBA responses than that measured in subsequent studies using wild-type strains with an identical PorA [122]. In a study of the Norwegian OMV vaccine, the proportion of responders against a wild-type P1.7–2,4 strain was about half that determined against another P1.7–2,4 strain, NZ 98/254 (Table 4) [70]. However, PorA matching is still considered a central element in determining hSBA responses with OMV-based vaccines.

A comprehensive evaluation of this phenomenon was undertaken using P1.7–2,4 isolates and P1.1,15 MenB isolates, matching the two PorA antigens of the bivalent OMV vaccine used [123]. Strains within these groups gave significantly different hSBA results prevaccination and postvaccination with GMTs differing by ≥5.5-fold and absolute titers varying by up to 9 titers. Interestingly, the proportion with hSBA titers ≥1:4 was also different across the strains in each group, and despite high baselines, there was still ≤20% difference depending on the strain. Follow-up investigations failed to identify the underlying cause of these differences, although PorA expression and accessibility to antibody on the bacterial surface were shown not to be causative. Although the strains used in these comparisons had the same PorA and, in many cases, identical sequence types, they were not all from a clonal outbreak, and it is therefore possible that differences in the minor OMV components were responsible for variable results. A study with the New Zealand OMV vaccine, however, showed that up to 22% fewer individuals achieved ≥4-fold rises in hSBA against other indistinguishable outbreak strains (i.e., differing from the vaccine strain by capsular group, PorB type, PorA variable region specificities, or PorA expression) compared with the NZ 98/254 vaccine strain [120]. This confirmed that some strains are inherently more susceptible to lysis in the hSBA assay. The reasons for this are undetermined; however, it is noteworthy that, in the aforementioned study, the strain in question utilized a different exogenous complement source. This concept is supported by an interlaboratory study that concluded that the major factor for discrepant results related to how the hSBA target strain was sent to each laboratory and was handled on receipt [66]. Differences in hSBA responses were reported before and after vaccination. Perhaps the most striking result was that the percentage of individuals with hSBA titers ≥1:4 prevaccination varied between 9.5% (2/21 subjects) and 90.5% (19/21 subjects).

The use of different complement sources in the SBA assay could contribute to variable SBA results with different meningococcal strains, which can be theoretically supported by the need to use unique complement sources for different strains in some of the studies reported. Notably, the converse interpretation could also be derived, as the reason different complement sources are required is because the strains are unique and some are susceptible to intrinsic antibody within the complement source, while others are not. It is the authors’ opinion that the differences in results between strains reported in this review are not explained by different complement sources where used. This view is supported by studies which reported differences in strain results when the same complement source was used [123]. The validation of complement sources for the SBA assay is inherently important in the accuracy and reproducibility of results. As long as the complement validation process ensures an appropriate complement source that does not contain intrinsic activity is selected to accurately quantify antibody in a test sample, then differences in results gained with different complement sources will be negligible [54,55].

9. Conclusion

Determination and understanding of the breadth of coverage of surface protein vaccines are of paramount importance and unique to each vaccine formulation because of differences in antigenic immunogenicity of vaccine components and also the large diversity of MenB strains and their surface proteins. Although the MenB hSBA assay and hSBA titer ≥1:4 are the internationally accepted surrogate and correlate of protection, respectively, there is a lack of standardization of methodology and titer calculation between laboratories, and differences in methodology can directly impact assay results. Understanding the proportion of MenB cases that are preventable by a given vaccine is crucial to assess the overall utility of the vaccine in a given setting/situation and to compare benefits and limitations of different vaccines in preventing IMD.

10. Expert opinion

This review shows that understanding correlates of protection and the breadth of coverage assessment for meningococcal surface protein vaccines is considerably more complex than that for capsular polysaccharide vaccines. The MenB hSBA assay and use of hSBA titers ≥1:4 are the internationally accepted surrogates for and correlates of protection, respectively [9,54]. There remains, however, a lack of methodological and titer calculation standardization between laboratories. Importantly, it is clear that different methodologies, including the choice of target strain and laboratory handling, can directly affect assay results.

These issues must be collectively considered as they influence the overall clinical applicability of the SBA assay results, hindering a comparison of results between laboratories and may result in a ‘protective’ titer in one assay and not in another. Any bias will directly affect the proportions of individ-uals categorized as achieving an hSBA titer ≥1:4 or any other cutoff applied but may have a lesser effect on measures of response such as fold rises in titers that are self-referenced against prevaccination titers. The ability to quantify overall responses in an immunized population, including the magnitude of responses, is an important aspect of the standard hSBA assay that is not achievable by the endogenous complement assay. Because many aspects can influence the hSBA assay, it is important to consider the overall impact on results holistically and any bias that may result leading to inappropriate comparisons or inaccurate conclusions on the potential breadth of coverage of a vaccine formulation.

Determination and understanding of the breadth of coverage of surface protein vaccines are clinically important and unique to each vaccine formulation. Estimation of the proportion of MenB cases that are preventable by a specific vaccine is critical to assess their overall potential impact on and to compare the benefits and limitations of different vaccines in preventing IMD. In addition, this estimation is crucial to fuel cost-effectiveness models that are an important step in decision-making for implementation of vaccine programs [124–126]. In the companion manuscript, we review the lessons learned with 4CMenB and MenB-FHbp and combine them with those within the historical context provided in the current review to highlight the key principles for future assessment of meningococcal surface protein vaccines.

In the future, it is expected that additional breadth of coverage data will emerge. These are likely to include hSBA assay data from diverse MenB strains and vaccine effectiveness findings from different countries and programs. Given the specific association between complement and susceptibility to IMD, the SBA assays will most likely continue to be the most appropriate licensure end point for MenB vaccines. However, these data will require careful interpretation because of the aspects identified in this review, including that findings are not necessarily transferrable as further reviewed in the companion manuscript. In this context, the conduct of head-to-head studies and further investigation regarding the approach to select (i.e., those representing the main burden of disease and typical expression levels) and analyze representative strains are prudent.

Other issues to consider for the use of SBA assays to assess meningococcal vaccines may include the limited availability of sources for human complement. To address this, transgenic mice expressing different human complement components or purification of human complement may be developed and be accepted as novel approaches in addition to the traditional hSBA assays. In addition, the increasing number of IMD cases that are only confirmed by PCR will limit the availability of cultured representative isolates, available for use in hSBA assays, and potentially impact the evaluation of the breadth of coverage. However, the development of genetic approaches to predict breadth of coverage may offer an alternative in post-licensure surveillance.

Article highlights

• In the first of two companion papers, we focus on the history of early meningococcal serogroup B (MenB) vaccines to identify the transferable lessons applicable to the currently licensed MenB vaccines and their serology, both of which are the focus of the second companion paper.

• Understanding correlates of protection and the breadth of coverage assessment of meningococcal surface protein vaccines is significantly more complex than that for capsular polysaccharide vaccines.

• The MenB serum bactericidal antibody using human complement (hSBA) assay is the internationally accepted surrogate and the hSBA titer ≥1:4 is the correlate of protection. However, there is a lack of standardization of methodology and titer calculation between laboratories, and different methodology can directly impact assay results.

• Determination and understanding of the breadth of coverage of surface protein vaccines is of paramount importance and unique to each vaccine formulation because of differences in antigenic immunogenicity of vaccine components and because of the significant diversity of MenB strains and their surface proteins.

• Understanding the proportion of MenB cases that are preventable by a given vaccine is essential to assess its potential impact in preventing invasive meningococcal disease and to compare the benefits and limitations of current and future vaccines.

List of abbreviations

4CMenB	=	Bexsero®, MenB-4C
C3	=	Complement protein 3
ELISA	=	Enzyme-linked immunosorbent assay
FbpA	=	37-kDa ferric binding protein
FH	=	Factor H
FHbp	=	Factor H binding protein
GMT	=	Geometric mean titer
hSBA	=	Serum bactericidal antibody assay using human complement
IgG	=	Immunoglobulin G
IgM	=	Immunoglobulin M
IMD	=	Invasive meningococcal disease
LLOQ	=	Lower limit of quantification
LPS	=	Lipopolysaccharide
MenA	=	Meningococcal serogroup A
MenACWY	=	Meningococcal serogroups A, C, W, and Y
MenB	=	Meningococcal serogroup B
MenB-FHbp	=	Trumenba®, bivalent rLP2086
MenC	=	Meningococcal serogroup C
NadA	=	Neisseria adhesin A
NHBA	=	Neisserial Heparin Binding Antigen
Omp85	=	Outer membrane protein 85
OMV	=	Outer membrane vesicle
OPA	=	Opsonophagocytic activity
PorA	=	Porin A
PorB	=	Porin B
RmpM	=	Reduction-modifiable protein meningitidis
rSBA	=	Serum bactericidal antibody assay using rabbit complement
SBA	=	Serum bactericidal antibody assay
T0	=	Number of colonies at the start of the assay incubation
T60	=	Number of colonies at the end of the assay incubation
WBA	=	Whole-blood assay

Declaration of interest

J Findlow, C Burman, and P Balmer are employees of Pfizer Inc and may hold stock or stock options. J Lucidarme performs contract work for Public Health England on behalf of GSK, PATH, Sanofi Pasteur, and Pfizer Inc. M-K Taha reports grants from GSK, grants from Pfizer Inc, and grants from Sanofi Pasteur, outside the submitted work and also has a patent NZ630133A with GSK ‘Vaccines for serogroup X meningococcus’ issued.

Reviewer disclosures

Peer reviewers for this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

Editorial/medical writing support was provided by Tricia Newell, PhD, and Sheena Hunt, PhD, of ICON plc (North Wales, PA, USA) and was funded by Pfizer Inc.

Additional information

Funding

This work was supported by Pfizer Inc.