Insights from the comparison of diverse Neisseria species
In addition to being medically important, the genus Neisseria is an attractive system for developing and applying population genomic approaches to linking bacterial phenotype with genotype. This is because the genus comprises a number of distinct species with different phenotypes in terms of their association with humans and animals. Three closely related members of the genus have been of particular interest from a medical perspective:
- Neisseria meningitidis, the meningococcus, is a major cause of meningitis and septicaemia worldwide;
- Neisseria gonorrhoeae, the gonococcus, is responsible for one of the most widespread sexually transmitted diseases; and
- Neisseria lactamica as a common harmless commensal member of the microbiota of children.
We have used genomic approaches, including ribosomal multilocus sequence typing (rMLST) and core-genome multilocus sequence typing (cgMLST), to resolve relationships within the genus at high resolution. This has demonstrated that Neisseria polysaccharea and Neisseria bergeri, a novel species that we have defined, are the closest relatives to the meningococcus and the gonococcus. All of these bacteria only colonise mucosal surfaces, with the meningococcus and N. polysaccharea, N. bergeri, and N. lactamica living in the nasopharynx and the gonococcus in the urogenital tract. This tissue tropism is an important differential phenotype. Further, the meningococcus has the ‘accidental pathogen’ phenotype: most infections are asymptomatic and invasion of the blood and cerebrospinal fluid is not a mechanism for host-to-host transmission. The gonococcus, on the other hand is frequently a pathogen, causing a range of sexually transmitted diseases. The association of some of the other Neisseria species with different hosts presents the prospect of unravelling host association. This is of particular relevance as there are no good animal models of human infection.
The assembly of a comprehensive collection of Neisseria isolates and their whole genome sequences (WGSs) has enabled us to define the species groups associated with the different phenotypes precisely. This is permitting us to explore the genetic bases of these phenotypes by genomic comparisons. A particularly intriguing insight is that, with the exception of the meningococcal capsule, which is limited to one species, many other cell components putatively involved in virulence are distributed among ‘non-pathogenic’ members of the genus.
Carriage and disease: investigating the meningococcal hyperinvasive phenotype
Although the meningococci have a commensal life cycle, with virtually all infections resulting in asymptomatic colonisation, in a minority of cases they invade to cause the severe conditions of meningitis and septicaemia, ether individually or in combination. Meningococci are therefore best referred to as ‘accidental’ pathogens, as the pathogenic process does not lead to transmission and indeed is inimical to it. The propensity to invade is not evenly distributed among meningococci and depends on many factors, a principal one being the possession of a polysaccharide capsule. Meningococci can have one of twelve capsules, or have no capsule at all, but only six capsules, corresponding to the meningococcal serogroups A, B, C, W, X and Y, are regularly associated with disease. Further, notwithstanding high rates of horizontal gene transfer (HGT), meningococcal populations contain stable genotypes (lineages, recognised as clonal complexes or ccs in MLST analyses). These are also differentially associated with disease. The consequence of this is that there are large differences in collections of meningococci isolated from disease and asymptomatic carriage.
A major part of our work assembles structured samples of disease and carriage isolates, which are comparable as far as is possible. The comparison of these isolates enables the investigation of the disease phenotype and the meningococcal determinants of that phenotype. Although the possession of a capsule is necessary in the great majority of cases of disease, it is not sufficient and there are other factors that make meningococci more likely to invade, or ‘hyperinvasive’. Together with colleagues, we have defined some of these, for example various proteins involved in iron metabolism and the meningococcal disease-associated island, but it is becoming increasingly clear that each of the major hyperinvasive lineages (there are at least a dozen of them) have different factors promoting virulence. Even within lineages, closely related organisms can have different, definable disease phenotypes. With whole genome sequencing we can examine these isolates to identify those genes or allelic variants of genes that are responsible for given phenotypes, leading to an improved understanding of the molecular mechanisms and potentially identifying targets for novel vaccines.
Understanding meningococcal epidemics
Meningococcal disease levels are highly variable: in many countries the disease is endemic at low levels, but in others hyper endemics – elevated levels of disease over longer periods of time – occur. Some locales experience disease outbreaks, for example in schools, universities, or military recruit camps and large-scale epidemics can occur, especially in the African meningitis belt. Global pandemics of meningococcal disease also occur. We know that elevated levels of disease are generally associated with particular hyperinvasive lineages, but it is not known why the fluctuation of disease rates occur and at the present time these fluctuations are unpredictable. For example, for most of the past 100 years, meningococcal disease rates have been low in England and Wales with increases in incidence at times of social disruption (the two world wars and the depression) with short-term peaks in incidence at times of global outbreaks. From the early 1980s until the early 2000s there was a sustained and dramatic increase in disease in England and Wales. We now known that this was due to simultaneous epidemics of at least four hyperinvasive lineages, but we don’t know why this occurred.
We are embarking on an ambitious study to dissect this period of high incidence. Over the whole epidemic period the Public health England Meningococcal Research Unit (PHE MRU) and Scottish Meningococcal, Pneumococcal, Heamophilus and Listeria Reference Laboratory collected essentially all of the UK disease isolates, forming an enormously valuable resource. In addition, the United Kingdom Meningococcal Carriage Study (UKMenCar) collected over 9000 carriage isolates at the peak of disease incidence in 1999-2001, from more than 50,000 UK teenagers. These isolates were collected to assess the impact of the meningococcal C conjugate vaccines on carriage. Together with colleagues around the UK, we are conducting a follow-up study (UKMenCar4) in 2014-15. This will assemble about 3000 carriage isolates from the current period of low meningococcal disease incidence. Our combined expertise in meningococcal epidemiology, genomic analysis, and population structure will enable us to compare the genes and genotypes of meningococci from disease and carriage at high and low incidence periods. With a sample of this size we shall be able to look at all of the major hyperinvasive lineages in detail as well as exploring within lineage variation. This will exhaustively identify the genetic differences among these meningococci, refining our understanding of the meningococcal invasive phenotype. Precise definition of the meningococcal carriage and disease phenotypes will enable us to assess the likely role of genes and allelic variants in these phenotypes.
Implications for vaccine development and implementation
The most successful vaccines against the meningococcus are those based on capsular polysaccharides. The first such vaccines, which comprised only the polysaccharide, were poorly immunogenic, providing short-term protection and only in adults. A major step forward was the conjugate polysaccharide vaccines that, by virtue of chemically linking the polysaccharide to a T-cell epitope typically a protein toxoid, elicited an anamnestic affinity matured response. These vaccines have been highly successful in large part due to their effects on carriage, which generates a very strong herd immunity effect. We played a major part in establishing the effects of such vaccines on carriage both in the UKMenCar Project, which we led during the introduction of meningococcal C conjugate vaccines (MCC), and by collaborating in the MenAfriCar consortium, which conducted a similar large-scale trial in the African Meningitis Belt over the period of introduction of MenAfriVac, a serogroup A meningococcal conjugate vaccines.
Unfortunately, due to safety concerns arising from the similarity of the B polysaccharide to human antigens, and its consequent poor immunogenicity, conjugate approaches have not ben exploited for serogroup B meningococci, which remains a major threat world wide. Efforts have concentrated on the development of ‘B substitute’ vaccines based on protein antigens. We have worked on the variation of these diverse antigens since 1988 in addition to developing our own vaccine candidate based on the PorA and FetA antigens (MenPF) which along with our colleagues we recently developed to a phase I trial. In addition we have collaborated with most of the major vaccine companies helping them to evaluate the variation of the components in their own particular vaccines.