Potential of the adhesin complex protein of Neisseria meningitidis for next-generation meningococcal vaccines

Keywords::

• adhesin complex protein
• bactericidal antibody
• meningitis
• Neisseria meningitidis
• sepsis
• vaccine

Eradication of meningitis and septicemia caused by Neisseria meningitidis is approaching reality. Capsule polysaccharide (CPS)-conjugate vaccines are expected to virtually eliminate meningococcal disease caused by serogroups A, C, Y and W [1], but the development of serogroup B (MenB) vaccines has been more intractable. MenB CPS is poorly immunogenic and displays molecular mimicry to human fetal NCAM glycoprotein, directing research toward using subcapsular vaccine antigens [2]. MenB outer membrane vesicle (OMV) vaccines have been successful during clonal epidemics, but human immune responses to OMVs are very heterogeneous and short-lived. New platforms for identifying MenB vaccine antigens have been developed, for example, reverse and structural vaccinology (RV, SV), proteomics, immuno-proteomics and microarray transcriptomics [3–6]. These strategies use reductionist identification criteria of antigen surface exposure or secretion, greater conservation, immune recognition and the ability to induce bactericidal antibodies, which is the generally accepted correlate of protection [1]. Often, these antigens are involved in host–pathogen interactions.

The complete sequence of the first MenB genome (strain MC58) [7] provided the basis for the RV approach to develop the Novartis 4CMenB vaccine [8]. This vaccine contains three recombinant proteins, namely the Neisseria heparin binding antigen (NHBA, genome-derived Neisseria antigen (GNA) 2132), factor H binding protein (fHbp or lipoprotein (LP)2086, GNA1870) and Neisseria adhesin A (NadA, GNA1994). Antigens GNA1030 and GNA2091 are covalently fused to NHBA and fHbp, respectively, in order to enhance immunogenicity. These proteins are mixed with a MenB strain-specific OMV vaccine, MeNZB (NZ98/254, P1.7–2,4), which provides additional protection against ST 41/44 clonal complex strains and immunomodulatory functions. In contrast to RV, Pfizer have used a conventional approach based on differential proteome fractionation to develop a vaccine containing two subfamilies of recombinant OM LP, rLP2086 [9].

4CMenB has been used in trials involving almost 8000 infants, children, adolescents and adults worldwide and the vaccine received a European Commission marketing authorization as Bexsero® in January 2013. The clinical trials also afforded the opportunity to estimate the potential coverage of Bexsero, using a newly developed meningococcal antigen typing system (MATS). Because a broad range of NHBA, NadA and fHbp variants exist in meningococci circulating among the population and protein expression levels vary, incomplete strain coverage with Bexsero is a potential issue. In several European countries, MATS predicts that approximately 73–78% of all MenB strains would be killed by post-vaccination sera [10] and in Canada, the predicted strain coverage was estimated to be only 66% (95% CI: 46–78%) [11]. The Pfizer rLP2086 bivalent vaccine has reached Phase II, and estimating its breadth of strain coverage is complex, for example, in toddlers and adolescents-young adults, protective bactericidal titers ranged from 44 to 100% and 68 to 98%, respectively, against MenB strains expressing heterologous fHbp proteins [9]. Based on these estimates, there is still a need for basic discovery of new candidate vaccine antigens and the complexity of the meningococcal OM presents a reservoir of untried components.

Biological function & vaccine potential of ACP

The adhesin complex protein (ACP, 124 amino acids, Mr 13.3 kDa, and product of gene NMB2095) was identified by proteomic analyses as present in meningococcal OM/OMV preparations [12–14]. Microarray analyses showed that the acp gene was upregulated under iron-depleted conditions [15] and the protein was abundant in MC58ΔGNA33 [16] and NMB0573-deficient mutant strains [17]. However, the biological function of ACP and its potential as a vaccine candidate were unknown, so we decided to investigate these properties [18]. ACP is a surface-exposed adhesin molecule that plays a role in pathogenesis, as demonstrated from in vitro infection studies with deletion mutants (ΔACP) and the use of purified recombinant ligands in cell binding assays. ACP mediates the adhesion of capsulated meningococci with epithelial, endothelial and meningeal cells. Recombinant (r)ACP binds to the surface of different human cell types and meningococcal adhesion can be inhibited by anti-rACP antibodies. ACP can function as an invasin of epithelial cells in the absence of capsule, but invasion is not impressive, as the numbers of recovered bacteria per human cell are low. A BLAST database search for amino acid sequence similarity of Neisseria ACP revealed the presence of an ‘adhesin complex protein’ from the Gram-negative aerotolerant anaerobe Dichelobacter nodosus, as well as several putative adhesins and adhesin components from other bacteria, for example, from Cardiobacterium hominis, Eikenella corrodens, Simonsiella muelleri and Kingella denitrificans. Although there is ≤40% sequence similarity between these ‘adhesin’ proteins and Neisseria ACP, the ‘ACP’ from D. nodusus, the causative agent of ruminant footrot, was identified by RV as a potential vaccine candidate and a recombinant protein could be recognized by sera from infected sheep [19].

In addition to surface-exposure, ACP fulfils several other important criteria for a meningococcal vaccine antigen [18]:

• 1. ACP is highly conserved: from initial sequencing of the acp gene in meningococcal isolates from colonized individuals and patients and alignment of translated amino acid sequences, we identified three types of ACP proteins (I, II and III). These different proteins share a high degree of similarity (98–99%): compared with type I ACP, type II ACP protein has only one amino acid difference (Asp²⁵ → Asn²⁵) and type III ACP has two substitutions (Asp²⁵ → Asn²⁵ and Ala¹² → Ser¹²). Subsequent interrogation of the BIGS Database [101], which contains the complete genome sequence data for 205 Neisseria strains, of which 173 are meningococci, showed that the meningococci only possessed genes corresponding to the type I ACP (encoded by 15 strains represented by allele 1) or to type II ACP (encoded by 158 strains represented by allele 2). No additional sequences were found in the NCBI database following BLAST analysis [102]. Type II ACP protein was exclusively encoded by meningococci, while a variety of Neisseria strains, including N. sicca (Ns), N. polysaccharea (Np) and N. lactamica (Nl), contained the acp gene encoding type I ACP. Notably, genes encoding type I, II and III ACP proteins were not present in N. gonorrhoeae, but gonococci expressed two different ACP proteins that showed 94% similarity with meningococcal ACP. Examination of the recently available and expanded Meningitis Research Foundation Meningococcus Genome library of approximately 680 sequenced Neisseria strains [103], confirms that meningococci express predominantly type I or type II ACP.

• 2. ACP proteins are expressed at similar levels by meningococci, that is, no variable expression has been observed thus far in strains of different serogroup, serotype and serosubtype examined.

• 3. ACP is immunogenic: murine immunization with a recombinant type I ACP induced high levels of complement-mediated bactericidal activity against meningococci expressing homologous antigen. Bactericidal activity could be induced by ACP administered with several human-compatible adjuvants and in saline alone, the latter a consequence probably of its LP nature. Notably, type I ACP antisera showed cross-strain protection by killing heterologous strains expressing type II and III ACP, with similar bactericidal titers. FACS and western blot analyses confirmed that these bactericidal antibodies were specific and bound to surface-exposed ACP on meningococci.

Commentary & prospectus

Ostensibly developed for MenB, Bexsero is marketed as a ‘first-generation universal’ vaccine, as the recombinant antigen components are present in other serogroup meningococci and thus potentially capable of providing cross-serogroup protection. Only the introduction of Bexsero and rLP2086 into routine immunization programs will demonstrate the efficacy of these vaccines. However, if the estimates of strain coverage hold true, as a consequence of antigen variability, then vaccine refinement or new vaccines will be necessary to eliminate remaining disease and the potential for strain replacement. Other vaccine strategies using major OM proteins such as the immuno-dominant PorA, or PorB, Opa, Opc, FetA and other iron-regulated proteins, generate the same concerns for strain coverage due to antigen amino acid sequence variability and phase variation. In a significant development, the Joint Committee on Vaccination and Immunisation has recently issued an interim position statement in which it does not recommend the UK release of Bexsero® for routine immunization [104]. Thus, identification of antigens with more conservation and capable of inducing cross-protective antibody responses could provide the missing element(s) for a ‘second-generation’ MenB vaccine.

ACP may be one such potential component of new meningococcal vaccines and it is encouraging to note that rACP matches rPorA in its ability to induce murine bactericidal antibodies. However, further studies on ACP are required prior to human trial evaluation. We are addressing several key questions: is ACP expressed to similar levels in larger numbers of meningococcal strains and how does iron-restriction regulate its expression? What is the structure of ACP and the identity of the bactericidal epitope(s)? Secondary structure prediction models suggest that the majority of the ACP protein comprises β-strands linked by short loops (residues 20–126) and the N-terminal 20 residues are a single helix that exhibits low complexity. Can we include rACP in multi-component formulations and is the protein stable? Do sera from colonized individuals, convalescent patients and from individuals vaccinated with OMVs contain anti-ACP antibodies? What is the nature of the cell surface receptor for ACP and does ACP expression play a role during infection in vivo? ACP shows a remarkable similarity to NadA in biological function and vaccine potential [18], leading to the hypothesis that such antigens are co-operative and possibly form complexes during host cell interactions? Also, NadA can induce host cell responses [20]: is this true for ACP? By extension, the gonococcal ACP will be examined as a potential vaccine antigen and for its role in pathogenesis; indeed, do dissimilarities between the gonococcal and meningococcal ACP proteins mediate the tropism for colonization of different mucosal epithelia?

In summary, Neisseria ACP is the first member of a class of small Mr OM antigens for which a role in human host cell interactions has been demonstrated. Moreover, Neisseria ACP fulfils the selection criteria for a ‘universal ‘meningococcal vaccine antigen and in the future, ACP-like proteins identified in other bacterial pathogens may become components of both human and veterinary vaccines.

Financial & competing interests disclosure

M Christodoulides and JE Heckels are named inventors on meningococcal vaccine patents owned by the University of Southampton. M Christodoulides has received honoraria from Novartis for consultancy: these honoraria were paid into funds administered by the University of Southampton for research purposes only. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was used in the production of this manuscript.

Since the print version of this article has been published, it has been determined that there are errors in the disclosure, in addition to a ≤ sign missing and some incorrect formatting in the article.