Comparison of Phenotypic and Genotypic Approaches to Capsule Typing of Neisseria meningitidis by Use of Invasive and Carriage Isolate Collections

Neisseria meningitidis, a Gram-negative bacterium that causes both epidemic and endemic life-threatening disease, is also an obligate human commensal organism that colonizes the nasopharyngeal mucosa with no or minimal harm to the host, a phenomenon known as carriage (1). Asymptomatic pharyngeal colonization with N. meningitidis in young adults is relatively common, and these asymptomatic carriers represent a potential reservoir for the transmission of pathogenic isolates in the community (2, 3). The rates of asymptomatic carriage in the United States and Europe are highest among adolescents and young adults, peaking at the age of 19 years, and are estimated to be 10% to 35% (4–6). Rates of transmission and carriage are higher in closed and semiclosed populations, such as university students living in dormitories and military recruits housed in barracks (7). Higher rates of carriage are also found among people in close contact with patients with active meningococcal infections (8). For the majority of people, carriage is an immunizing process that results in systemic, serogroup-specific, protective antibody responses (9, 10). Invasive meningococcal disease (IMD) usually occurs shortly after the onset of colonization of a susceptible host, when the bacteria penetrate the mucosal membranes, overcome host defenses, and invade the bloodstream or meninges, leading to septicemia or meningitis, respectively (11).

IMD is cyclical and varies by age group, being more common among children, especially those <5 years of age. A second peak in disease occurs in adolescence, with rates declining during early adulthood and subsequently decreasing for older age groups (12, 13). Asymptomatic carriers represent a potential source of virulent isolates that cause disease; however, the factors that lead to a change in the carriage state of the organism, resulting in invasive disease, are still poorly understood. Approximately 450 cases of meningococcal disease occurred in the Unites States in 2012, yielding a rate of 0.15 cases/100,000 population (14). The overall incidence rates in Canada (2011 data) and Europe (2012 data) were 0.55 cases/100,000 population and 0.68 cases/100,000 population, respectively (15, 16). In outbreak and epidemic settings, the case rates are considerably higher, i.e., 100 cases/100,000 population to 1,000 cases/100,000 population (17, 18).

The polysaccharide capsule is an important meningococcal virulence factor that protects the bacterium from complement-mediated bactericidal activity of human serum, as well as opsonophagocytic killing in the bacteremic stages of meningococcal disease (19–21). N. meningitidis strains are classified into serogroups based on structural differences in their capsular polysaccharides (7, 22, 23). Twelve serogroups have been described, with six (i.e., N. meningitidis serogroup A [MnA], MnB, MnC, MnW, MnX, and MnY) being implicated in the majority of IMD (24–27).

Capsule antigens (polysaccharide and polysaccharide conjugate) are potent and efficacious in preventing IMD when formulated as vaccines (28, 29). However, the common sialic acid-containing epitope of the MnB capsule is also present on human neural tissue, which may explain the poor immunogenicity of the MnB capsule; therefore, out of an abundance of caution, the MnB polysaccharide has not been considered a viable vaccine candidate (30). As a result, protein antigens have been employed in the development of MnB vaccines (31). Two vaccines have recently been licensed for the prevention of MnB disease, namely, Trumenba (bivalent rLP2086) (32) and Bexsero (4CMenB) (33).

In areas in which MnC conjugate vaccines were introduced and had widespread implementation, including receipt of the vaccines by very large proportions of adolescents, disease incidence in nonvaccinated populations was dramatically affected due to indirect prevention of disease through reductions in carriage rates, e.g., herd protection/immunity (34–37). This effect was not observed with the use of nonconjugated polysaccharide vaccines. It will be important to study whether MnB vaccines have this additional benefit (38). A recent study demonstrated similar levels of herd protection for protein-based (4CMenB) and non-protein-based (MenACWY-CRM) meningococcal vaccines (39). Meningococcal carriage studies can be challenging, due to a combination of reduced capsule expression by carriage strains and a lack of standardized approaches to evaluate carriage isolates in order to identify their associated capsule polysaccharide groups.

Invasive disease isolates are traditionally serogrouped by slide agglutination serogrouping (SASG) assays (23, 40), which use commercial polyclonal antibody reagents raised against prototypical N. meningitidis isolates representing the important disease-associated serogroups. The reagents and laboratory methods used for SASG assays are not standardized across laboratories, especially because the endpoint of the assay is read subjectively with the unaided eye (23, 40). Other serogrouping approaches have employed capsule-specific monoclonal antibodies (MAbs) applied to different platforms, such as dot blotting (41), whole-cell enzyme-linked immunosorbent assays (ELISAs) (42), and flow cytometry (43). Large proportions (70 to 80%) of invasive isolates can be successfully serogrouped using expression-based phenotypic assays (40). In contrast, high percentages of isolates recovered from healthy carriers are reported as nonserogroupable in phenotypic assays; the general explanation for this finding has been the low levels or complete absence of expression of the capsule genes during carriage (44–47), which can be explained by several mechanisms, including phase variation of capsule expression genes (44), insertion of genetic mobile elements (48), or a complete absence of the genes required for capsule production, leading to a capsule-null (cnl) phenotype (49, 50). The cnl mutation was reported for ∼16% of nasopharyngeal meningococci collected from healthy children and young adults in two independent carriage studies (45, 50). A recent report by Loh et al. (51) described a complex regulatory mechanism linking growth temperature to capsule gene expression. Molecular mechanisms underlying capsule biosynthesis have been elucidated; polysaccharide synthesis and translocation genes are clustered at a single chromosomal locus called cps. The cps locus consists of six different regions, i.e., A to D, D′, and E (52). Of particular interest for the present study are genes in region A, which encode the enzymes responsible for polysaccharide synthesis (53, 54), and genes in region C, which encode proteins that facilitate polysaccharide transport across bacterial membranes (55, 56).

Capsular genogrouping of carriage isolates can be achieved using rapid sensitive DNA-based methods (40). A number of PCR approaches have been developed to detect targets within the ctrA (capsule transport) gene (52, 57), as well as genes required for serogroup-specific capsule biosynthesis (8, 40, 58–61). For serogroup prediction (capsular genogroup assignment), PCR using primers designed for serogroup-specific genes can be applied (62–66). Fluorescence-based real-time PCR (rt-PCR) assays were recently developed to confirm the capsule typing of meningococcal isolates (59, 67, 68). Whole-genome sequencing (WGS) data also are being used increasingly for epidemiological typing of bacterial pathogens (69–71). In this study, we compared different capsular serogrouping methodologies for the assignment of invasive and carriage isolates to capsular polysaccharide serogroups, in order to determine a methodology for the characterization of carriage isolates that is broadly applicable for different laboratory settings and has the capacity to be standardized.

The meningococcal capsule is a virulence factor, as exemplified by the low frequency of unencapsulated N. meningitidis isolates causing invasive disease (86, 87). The capsule protects the bacterium from bactericidal activity during invasion and dissemination (51, 88–90). In this study, we demonstrated that ≥97% of invasive disease isolates expressed capsule in vitro and could be readily characterized by both genotypic and phenotypic methodologies. Serogrouping of carriage isolates phenotypically, as well as by rt-PCR, was challenging; this was particularly true for MnB isolates. This is due to reduced or no capsule expression in the carriage state, which has been reported to be the result of phase variation and/or genetic deletions or insertions in the cps locus (44–47). A subset of nongroupable isolates has been described in the literature, with deletion of the complete cps locus yielding capsule-null (cnl) mutants (45, 50, 84). WGS confirmed the loss of the ctr (capsule transport) operon and capsule biosynthesis genes in 32 (34%) of the carriage isolates tested. In agreement with published work, the capsule-null isolates in our collection were restricted to the following clonal complexes: sequence type 53 (ST53) (38%), ST198 (41%), ST1117 (12%), and ST1136 (9%) (91). Additionally, 10% of the carriage isolates were negative by rt-PCR for serogroup-specific genes, although they did carry the ctrA gene. Although carriage isolates may be missing the capsule locus or harbor loci with insertions or deletions, this does not necessarily predispose such isolates to be nonvirulent, as N. meningitidis is naturally genetically competent and has the capacity to obtain the cps pathway genes through horizontal genetic exchange (92–94). In the carriage state in the human nasopharynx, cocolonization with other pathogenic and nonpathogenic N. meningitidis strains has the potential to lead to genetic exchange and capsule switching (95, 96). Furthermore, although IMD caused by cnl meningococci is rare, there have been reports of cnl isolates causing disease, mainly in immunocompromised patients (86, 87).

WGS identified 44 carriage isolates (47%) that contained a complete cps operon; of those isolates, less than one-half were phenotypically positive by the various phenotypic diagnostics tests (Table 5). Phenotypic assays were more effective in serogrouping invasive isolates than carriage isolates; ≥95% of invasive isolates were successfully grouped by FACS, dot blotting, and LCPA (Table 5). Although the capsule is an important virulence factor, cps expression is regulated at different stages of growth and infection (47). The recent study by Loh et al. (51) demonstrated that environmental factors (e.g., temperature) play roles in the expression of capsule and surface antigens through a thermosensor. Those authors identified an 8-bp repeat motif whose deletion resulted in dysregulation of the cssA promoter. In the present study, alignment of the intergenic region containing the ctrA and cssA promoters from carriage isolates that failed to express the capsule and capsule-positive control isolates did not reveal any clues, at the sequence level, to explain the lack of capsule expression by the carriage isolates (data not shown). Isolates in our collection were shown to have as many as three of the 8-bp repeat motifs described by Loh et al. (51); additional studies will be required to determine the role of these sequences in capsule expression in vivo.

WGS is rapidly becoming the gold standard for bacterial epidemiological studies (69–71). In this study, we have highlighted its potential usefulness for the characterization of meningococcal carriage isolates. Although the rt-PCR assay appears to be highly predictive of capsule genogrouping, especially for invasive N. meningitidis isolates (99%), this approach relies on the detection of a small fragment (<120 bp) of one gene in the cps operon for each capsular genogroup, and assay sensitivity can be affected by the high degree of genetic plasticity, which we confirmed especially when evaluating carriage isolates. Selection of the target sequences is critical and may be different in different geographical locations.

Several targets (porA, ctrA, and sodC) have been employed as N. meningitidis-specific diagnostic targets for rt-PCR assays (97, 98). In this study, we looked at all three targets by WGS in both of our isolate collections. As prior studies have found (50, 84), ctrA is absent in a large percentage of carriage isolates; in our collection, only 54/93 carriage isolates (58%) were positive for ctrA, whereas 100% of invasive isolates were positive for this gene. Both porA (99% and 97%, respectively) and sodC (97% and 100%, respectively) were well represented in the carriage and invasive isolate collections. Our data confirm that either of these target genes could be a reliable N. meningitidis-specific marker for an rt-PCR diagnostic assay.

The carriage and IMD isolate collections were not contemporary or from the same geographical locations. However, data collected over the past 5 years from diverse geographical locations, which showed that the levels of capsule expression in carriage isolates were the same (data not shown), suggested that geographical locations may not affect our final conclusions. Additionally, the IMD isolates were previously characterized by SASG assays; therefore, the isolates tested were not from an unbiased set, although all assays were conducted in a blinded fashion. As reported previously, IMD isolates are efficiently serogrouped by SASG assays; therefore, this was not thought to affect the overall conclusions of the study (40).

With the introduction of vaccines targeting MnB, carriage studies are essential for monitoring the ability of the new vaccines to prevent carriage acquisition and/or to shorten the duration of carriage episodes and for assessing the development of herd protection (12). Identification of the appropriate diagnostic tools to conduct surveillance following vaccination will be critical. The present study demonstrates the utility of using WGS to accurately type bacterial isolates recovered during clinical studies. Based on our field study, rt-PCR diagnostic assays seems unlikely to provide the reliability necessary for carriage surveillance, due to a large number of misclassified isolates with a mutated or deleted cps locus. Our findings suggest that the use of WGS to type carriage isolates accurately, in conjunction with the use of FACS to determine the capsule phenotype, may offer a superior algorithm to characterize carriage isolates for most purposes, compared to single methods. Remaining challenges for the development of WGS as an interpretive tool for disease surveillance, confirmatory diagnosis, and case management include the development of an appropriate infrastructure for data generation, analysis, and storage. We conclude that a combination of genotypic and phenotypic methods will improve N. meningitidis serogroup determination and the understanding of N. meningitidis carriage and epidemiology.

We acknowledge Alita Miller and Robert Donald for helpful conversations and Li Hao for assistance with bioinformatics analyses.

This study was sponsored by Pfizer Vaccine Research.

C.H.J., N.M., E.R., L.A., J.H., J.C.H., L.K.M., Q.J., L.P., J.E., K.U.J., and A.S.A. are current or former employees of Pfizer and may own stock in the company. R.G. and P.D.W. have received research grants and/or reimbursement of travel expenses from Pfizer. L.W.M. and X.W. have no conflicts of interest and do not receive financial support or have other interests in commercial companies.