Unraveling Neisseria meningitidis pathogenesis: from functional genomics to experimental models

Neisseria meningitidis is a harmless commensal bacterium finely adapted to humans. Unfortunately, under “privileged” conditions, it adopts a “devious” lifestyle leading to uncontrolled behavior characterized by the unleashing of molecular weapons causing potentially lethal disease such as sepsis and acute meningitis. Indeed, despite the lack of a classic repertoire of virulence genes in N. meningitidis separating commensal from invasive strains, molecular epidemiology and functional genomics studies suggest that carriage and invasive strains belong to genetically distinct populations characterized by an exclusive pathogenic potential. In the last few years, “omics” technologies have helped scientists to unwrap the framework drawn by N. meningitidis during different stages of colonization and disease. However, this scenario is still incomplete and would benefit from the implementation of physiological tissue models for the reproduction of mucosal and systemic interactions in vitro. These emerging technologies supported by recent advances in the world of stem cell biology hold the promise for a further understanding of N. meningitidis pathogenesis.


Introduction
Neisseria meningitidis is a versatile organism capable of adapting to the different environments it encounters during colonization and invasive disease. Like many other bacterial pathogens, it finds it beneficial to keep the host alive to allow transmission. However, it is a fact that in crowded settings such as military camps, universities, and schools, N. meningitidis tends to become more virulent 1 . Whether this is related to the chance to encounter more appropriate environmental conditions (for example, weakened immunity, affordable nutrients, and reduced niche competition) or to the fact that, under low population density, selection pressure would keep the host alive until transmission is possible is still indefinite. Nevertheless, household contacts of patients with meningococcal disease have been shown to be at increased risk of meningococcal carriage and disease. From a genomic perspective, N. meningitidis is a highly diverse species, undergoing frequent recombination characterized by horizontal gene transfer 2 . However, phylogenetic and genealogical analyses have revealed the presence of a limited number of clonal complexes associated with invasive disease (often referred to as "hyper-invasive lineages") 2 . These lineages show a recurrent antigenic and disease phenotype and have been an important paradigm for designing intervention strategies. The advent of "next-generation" sequencing has revolutionized the molecular epidemiology field by offering the opportunity of a complete picture of N. meningitidis genotypes and improving our understanding of meningococcal pathogenesis (for an in depth review on recent advances in population genomics, see 3). In this context, initiatives such as the Meningitis Research Foundation meningococcus genome library (http://www.meningitis.org/ research/genome) are expected to facilitate not only population genomics approaches but also functional genomics by guiding the selection of the most appropriate isolates and reduce the use of often irrelevant laboratory strains. An interesting application of this tool has been in the vaccine field, where this library has been instrumental in establishing that a recent rise in serogroup W cases since 2009 belongs to ST-11, a particularly virulent sequence type with a high case fatality rate 4 .
For years, the specificity of N. meningitidis for humans has been the main bottleneck in unravelling the mechanisms beyond its invasive behavior. In particular, the lack of appropriate animal models resembling the clinical presentations of the human disease has affected the capacity to develop efficacious preventive interventions. In the last decade, molecular and structural evidence has highlighted a number of surface molecules with a strong specificity for human serum factors. In particular, factor H-binding protein (fHbp) has been at the center of great interest not only for its role in N. meningitidis pathogenesis 5 but also for its capacity to generate strong bactericidal antibodies after immunization in humans 6 . fHbp is currently one of the components of the recently approved vaccines against type B meningococcus and likely to contribute to the extraordinary data on the efficacy of serogroup B meningococcal vaccine in the UK 7 . Serogroup B is now the most common cause of outbreak-associated disease, and the fact that the novel, multicomponent, protein-based Bexsero™ vaccine turned out to be 82.9% effective after two doses in preventing serogroup B N. meningitidis disease in British infants younger than 12 months of age 7 turns a promise into reality. However, the success of the strategy, as for that of all vaccines, will depend on the breadth of implementation and the promptness of the pathogen to epidemiologically adapt to the evolutionary pressure introduced by vaccination campaigns. Therefore, whatever would be the most optimistic scenario, it is important to continue to monitor, investigate, and consider all of the subtle strategies beyond the peculiar habit of this "smart" microorganism disguised as a commensal but with the license to kill. Several scientists refer to these events as an "accidental lethality" or "pathogenic commensalism". In this commentary, we will go through the salient steps of N. meningitidis pathogenesis that, thanks to the support of "omics" technologies and advanced infection models, have been fully unraveled in the last decade.
Disclosure of Neisseria meningitides pathogenesis by "omics" and experimental models N. meningitidis usually resides in the human nasopharynx where it spends most of its life as a commensal microorganism by exploiting nutrients present on the mucosae 8,9 . Notably,Veyrier et al. 10 recently postulated that cell shape evolution of N. meningitidis (from bacillus to coccus) has allowed an increased adaptation to the nasopharynx by reducing the cell surface sensible to immune attacks through the modification of the peptidoglycan and by redistributing surface determinants such as pili 10 . The initial steps of colonization and pathogenesis are graphically summarized in Figure 1, in which the emphasis is on the factors that have been identified so far as essential for N. meningitidis "sojourn" in the host.
Crossing of the epithelial cell layer of the nasopharynx is a rare event but, when it occurs, leads to invasion of the bloodstream, where bacteria are capable of eluding the immune system and of reaching the meninges. The ability of N. meningitidis to bind to ligands present on the surface of host cells allows the bacterium to easily enter in contact with the endothelial cell layer of the brain vessels and to form microcolonies 11 . This interaction, mediated mainly by the type IV pili, modulates the endothelial cytoskeleton leading to the formation of docking structures similar to the ones elicited by leukocytes during extravasation and the consequent opening of the intercellular junctions 12 . The sterility of the cerebrospinal fluid of the subarachnoidal space and its low serum protein content but richness in nutrients (including glucose, sodium chloride, and urea) greatly favor the replication of N. meningitidis and its dissemination throughout the meninges 13 . Another phenomenon linked to N. meningitidis invasive disease is a generalized sepsis, in which bacteria associated with microvessels induce extensive thrombosis, coagulation, congestion, and vascular leak, leading to an extensive necrosis of the skin and surrounding tissues 14 .
Many of the characters playing a pivotal role in this drama have recently been disclosed by the application of "omics" technologies to a number of experimental models mimicking different steps of N. meningitidis pathogenesis. Functional genomics, by linking genotype to phenotype, have allowed study of the correlation between gene transcript abundance or deficiency and the capacity of N. meningitidis to behave under various physiological conditions of the host. The first example of functional genomics in N. meningitidis was reported by Tang's group almost 20 years ago 15 . By genome-wide signature-tagged mutagenesis (STM), 73 genes essential to bacteremia were identified in an infant rat model 15 . A few years later, with the advent of the microarray technology, new studies focused on the transcriptional events occurring during the interaction of N. meningitidis with host cells [16][17][18][19] . Then comparative genomics 20-23 , in vitro 24,25 and ex vivo 26,27 transcriptomics, proteomics 28,29 , and further STM 30,31 completed the picture.
The scenario derived by these studies (intuitively represented in Figure 1) offers a number of considerations. As expected, adhesion molecules (such as type IV pili) and serum resistance factors (like the lipooligosaccharide and genes involved in the synthesis of the polysialic acid capsule) turned out to be essential to preserve the fitness of the bacterium under stress conditions or only to maintain its "colonizer" status. These molecules, by sensing the external milieu, need to rapidly respond to changes, whether this means the proximity to host cell ligands, the interaction with serum factors, or the availability of nutrients. This substantial surface remodeling has been exploited to identify putative vaccine candidates, as the augmented expression of surface antigens under physiological conditions has been considered a discriminating factor for selection 16,17,19 . However, the dynamics ongoing during the adaptation of N. meningitidis to the host are far more complex and pivotal to maintain bacterial fitness. Indeed, the most intriguing results generated from functional genomics studies were relative to the modulation of genes involved in regulatory functions and metabolism. (For in-depth reviews, see 32 and 33, respectively.) It is not by chance that genomic regions coding for metabolic functions exhibit high rates of recombination 22,23 , a feature shared with genes contributing to pathogenicity. On the other hand, 35 of the 73 genes reported by Sun et al. as "essential" to in vivo bacteremia encode for enzymes involved in metabolism and transport of nutrients 15 . This trend was further corroborated by ex vivo transcriptomic data showing that N. meningitidis grown in human blood differentially expresses several genes involved in nutrient transport and central metabolism 26,27 . Overall, transcriptomic studies have highlighted that differential expressions of genes involved in metabolism of lactate, oxidative stress response, glutathione metabolism, and denitrification pathways are among the most frequent examples of adaptive response during pathogenesis. In particular, the capacity of N. meningitidis to promptly catabolize lactate has been considered fundamental to bacterial survival. Lactate is broadly present in the human body at considerable concentrations (approximately 0.3 to 1.3 mM). Being a substrate for the synthesis of N-acetyl-neuraminic acids via the N-acetyl-neuraminic acid synthase (NeuB) synthetic enzyme, lactate contributes to enhanced serum resistance 34-36 and nasopharyngeal colonization 37,38 . The evidence that, in human blood, lactate permease was significantly upregulated 26 further confirmed the importance of this sugar in immune where it adheres to the mucosae of the mucociliary epithelium and resides as a commensal microorganism until environmental conditions are suitable for dissemination. (C) Crossing of the mucosal epithelial barrier occurs by intra-or inter-cellular routes allowing entry into the bloodstream, where it quickly proliferates. This event causes sepsis and eventually (after translocation of a further physiological barrier such as the blood-brain barrier) meningitis. The right bottom panel is a list of factors/pathways defined by functional genomics studies to be determinant during various steps of N. meningitidis pathogenesis such as colonization of the nasopharynx and survival in blood. App, adhesion and penetration protein; fHbp, factor H-binding protein; FNR, fumarate and nitrate reductase regulator protein; Fur, ferric uptake regulation protein; Hfq, cofactor RNA-binding protein; LPS, lipopolysaccharide; MIP, macrophage infectivity potentiator MspA, meningococcal serine protease A; NadA, Neisseria adhesin A; NalP, Neisseria autotransporter lipoprotein; NhhA, Neisseria hia homologue A; NspA, Neisserial surface protein A; Opa, opacity protein; Opc, opacity protein C; sRNA, small non-coding RNA.
evasion. However, whether the reported phenotypic behavior may act as a paradigm for N. meningitidis increased colonization of the nasopharynx is not clear. Indeed, an increased synthesis of sialic acid by enhancing capsule levels and lipopolysaccharide sialylation may result in an impaired ability to bind to mucosal surface. Therefore, the balance between carriage and invasive attitude is quite arguable, and more data are needed to understand the contribution of metabolic and virulence factors to N. meningitidis pathogenesis. The evolutionary success of N. meningitidis relies on an efficient replication within the bloodstream because of not only an efficacious uptake of nutrients but also the concomitant ability to evade the innate and acquired immune defenses by exploiting the benefits of an appropriate sugar decoration 39 . Iron metabolism is also central to the fitness and ability of N. meningitidis to out-compete neighborhood bacteria and host defenses. Although iron is pivotal for DNA replication, electron transfer in the respiratory chain, and oxidative metabolism, free iron is scarcely available in the host and meningococci possess several iron uptake systems 40 . Acquisition of iron from host complexes is mediated by surface-located receptors: two hemoglobin receptors (HmbR and the heterodimeric HpuAB complex) and TbpBA and LbpBA reported to bind iron-loaded transferrin and lactoferrin, respectively. However, although iron uptake is essential to N. meningitidis immune evasion, HmbR was recently suggested not to be required during the early stages of disease, calling into question the importance of hemoglobin in meningococcal pathogenesis 41 . Microarray analysis of the effect of iron addition to N. meningitidis culture revealed a large modulation of genes involved in energy metabolism, protein synthesis, and cell envelope assembly 42 . These events appear to be largely under the control of the ferric uptake regulation protein (Fur) regulator that, in response to iron, affects the expression of target genes 42-44 . For example, since the mucosal surface is rich in lactoferrin and the bloodstream contains high amounts of hemoglobin, these proteins were suggested to serve as niche indicators for N. meningitidis, leading to specific changes in gene expression 42,45 .
Environmental oxygen levels represent another important stress event encountered by N. meningitidis during pathogenesis. An in-depth analysis of the importance of FNR (fumarate and nitrate reductase regulator protein) in sensing oxygen concentrations was reported by Bartolini et al. 46 , who elucidated a number of metabolic pathways modulated under limited oxygen conditions, as faced in the brain microcirculation.
As mentioned previously, regulatory functions are currently the hot topic in functional genomics, especially after the discovery of small non-coding RNAs (sRNAs) 47,48 . In N. meningitidis, a great deal of importance has been given to Hfq, an RNA binding protein contributing to base pairing between sRNA and mRNA 49,50 , found to be modulated in blood 26 and essential for serum resistance 15 . A number of transcriptomic and proteomic studies confirmed the relevance of Hfq in Neisseria response to stress conditions 49,51-53 and its capacity to modulate sRNAs 54 . Of importance, Capel et al., by exploiting a Tn-seq strategy coupled to high-throughput DNA sequencing technologies, reported a comprehensive analysis of sRNAs essential to colonize epithelial cells and primary brain endothelial cells, providing a new tool to further investigate meningococcal pathogenesis in different environments 55 .
Unfortunately, our understanding of the pathways activated by N. meninigitidis in response to environmental changes is limited by the relevant number of functionally unknown open reading frames that have often been reported among the most modulated targets. To this end, Exley et al. found that six out of eight mutants attenuated for their capacity to adhere to nasopharyngeal explants had transposon insertions in genes of unknown function 30 . Currently, this major gap still keeps the whole picture incomplete. It is important to notice that much of the reported evidence on the contribution of meningococcal "armaments" to adaptation and virulence was obtained by employing laboratory isolates often belonging to rare genotypes or not relevant to N. meningitidis pathogenesis. This is to highlight that we may still underestimate the impact of "hidden" pathways relevant to hyper-virulent lineages associated with outbreaks.
In vitro transcriptomic and mutagenesis studies were mainly carried out by incubating bacteria in the presence of immortalized human cell lines derived from epithelial and endothelial tissues. Although they have been a remarkable pioneering attempt to resemble human physiology of the upper respiratory tract and microcirculation, these in vitro studies were limited by the specificity of the events triggered by N. meningitidis in vivo. Recent studies on mucosal pathogens have revealed the fundamental contribution of mucosae components in triggering signals to host tissues. Nevertheless, the human specificity of this bacterium makes studying the pathogenesis of Neisseria infections in vivo very difficult. Seminal in vitro studies were characterized by the use of cell lines derived from organs relevant to the meningococcal disease, such as the respiratory epithelium and the brain endothelium 56-59 . Although the results of these studies have been pivotal to the understanding of N. meningitidis pathogenesis, they were limited by the lack of environmental attributes that contribute to the in vivo response of the host to pathogens. Experimental models of fulminant meningococcemia in human skin-grafted immune-compromised mice have recently been engineered 60,61 . Under these conditions, N. meningitidis adheres to implanted human vessels, triggering extensive vascular damage, similar to that observed in patients 62 . We expect that this kind of model, together with the increased accessibility to organoids and three-dimensional (3D) bioprinted organs, will be extensively exploited not only to confirm the current knowledge on N. meningitidis pathogenesis but to disclose the hidden pathways that are essential to bacterial fitness and that could be unraveled only by recreating a physiological environment. In this context, Deosarkar et al. reported the first dynamic in vitro neonatal blood-brain barrier on a chip closely mimicking the in vivo microenvironment 63 . On the other hand, models for skin, bronchi, blood vessels, and microcirculation are widely engineered for all sorts of different applications from basic research to drug discovery (nicely reviewed in 64). We therefore foresee the adaptation of 3D cellular models in novel multi-organ systems to study N. meningitidis pathogenesis, as has extensively been done for intestinal and gastric organoids to study enteric and Helicobacter pylori infections, respectively. In this context, Marrazzo et al. 65 recently established an in vitro 3D system which recapitulates the human tracheo-bronchial mucosa comprehensive of the pseudostratified epithelium and the underlying stromal tissue. This model has been exploited to study initial colonization events triggered by non-typeable Haemophilus influenzae but could easily be adapted to any other microorganism colonizing the nasopharynx. Therefore, only by stemming from the field of regenerative medicine, we could find the right approaches to unravel unknown signaling pathways occurring during N. meningitidis pathogenesis. Researchers working in the field of cancer progression or environmental damage to respiratory organs are generating sophisticated examples of human airways that scientists working in the infectious disease world should start considering. Table 1 is a list of 3D tissue models that have mainly been developed to study organ physiology but that could be customized to carry out studies on the strategies used by N. meningitidis to adapt, colonize, and induce disease in humans.

Closing remarks
This commentary started with the hope of a world without meningococcal meningitis thanks to the implementation of current vaccines. However, a lot is still needed to fully understand the pathophysiology of such a disease. The progress obtained so far in disclosing N. meningitidis pathogenesis reveals that the ample evidence for "culprits" is not sufficient to completely unravel the "murder scene". Population and functional genomics have had a great role in defining many of the key pathways activated by N. meningitidis to successfully colonize our organism, but biotechnologies like in vitro 3D human experimental models are emerging as the new frontier to establish the appropriate environment to study bacterial pathogenesis. To this end, a multi-disciplinary approach would be vital to ensure the required progress for fighting human infections. In the last decade, the input of engineers, physics, mathematicians, and statisticians has been crucial to several biology and medicine areas (particularly in "omics" disciplines) and is expected to have even more relevance in the future. In the area of infectious diseases, they are becoming the principal interlocutors of molecular and cellular microbiologists by playing a pivotal role in designing, fabricating, miniaturizing, and validating in vitro tissue models to be exploited in host-pathogen interaction studies. Technology centers in Europe and the US (for example, the Francis Crick Institute, London, UK, and Wake Forest Institute, Winston-Salem, NC, USA) and international biotech companies (for example, Organovo, San Diego, CA, USA, and 3D Bioprinting Solutions, Moscow, Russia) are already investing in this direction by holding the promise of a future with curable diseases and a better quality of life.

Competing interests
The author declares that he has no competing interests.

Grant information
The author(s) declared that no grants were involved in supporting this work.