Neisseria meningitidis (the meningococcus) is a Gram-negative diplococcus that transiently colonizes the nasopharynx of healthy subjects and occasionally causes invasive meningococcal disease (IMD) (Hill et al., 2010). Despite the availability of modern and effective vaccines and the optimization of therapeutic protocols, IMD, with its most frequent manifestations, sepsis and meningitis, continues to claim victims and to represent a major public health problem. On a global scale, from 5 to 15% of the population is estimated to be composed of asymptomatic nasopharyngeal carriers of N. meningitidis, and less than 1% of colonized individuals develop IMD (World Health Organization, 2025). Incidence rates of meningococcal carriage and IMD are, however, strongly affected by age, population, geography, and time-related variations. The carriage rate is generally low in infants and children, while it increases and reaches its peak in adolescents. Differences in host genetics, demographic factors, social behaviors, circulating meningococcal strains, public health infrastructures, surveillance, and prophylactic interventions may explain differences in meningococcal epidemiology between different countries (Bogaert et al., 2005; Weckx et al., 2017; Kizil et al., 2021). For instance, in the USA, the carriage rate is around 24% (Peterson et al., 2019; Santos-Neto et al., 2019), in the African meningitis belt, it is between 10 and 20% but reaches 80% during outbreaks (Tefera et al., 2020), while in Asian countries it is generally low (1.5-9.1%) (Serra et al., 2020). However, despite the low carriage rate, in some Asian countries, the mortality rate of IMD is high due to weak surveillance and/or the lack of routine vaccination (Aye et al., 2020). IMD is characterized by high morbidity and mortality in children and adults. Lethality rates vary between 7 and 15% when the disease is treated but can exceed 50% when IMD is not treated (Wang et al., 2019). Rough estimations point to around 500,000 IMD cases worldwide, causing around 50,000 deaths each year (World Health Organization, 2025). Moreover, between 11 and 19% of patients who survive have severe sequelae, such as hearing loss, limb amputation, and neurological complications (Olbrich et al., 2018; Walter et al., 2021).

Key questions about meningococcal infection, which still await clear and exhaustive answers but are essential for identifying increasingly effective prophylactic and therapeutic strategies, are: i. How does the transition from asymptomatic colonization to disease occur? ii. Why is IMD so devastating? iii. What are the bacteria-, host-, and environment-related factors associated with increased incidence, severity, and mortality of meningococcal infection? iv. How did this bacterium, now considered an “accidental” pathogen, evolve? v. Is there a possibility of new pathogens emerging able to cause systemic disease in humans within the Neisseria genus? To address these questions, it is mandatory to understand the infection cycle of this bacterium.

To cause sepsis and meningitis, N. meningitidis has to cross cellular barriers. In fact, this common transitory colonizer of the human nasopharynx is able to cross the nasopharyngeal epithelial barrier (NEB), entering and replicating in the bloodstream, resulting in septicemia and/or septicemic shock. It is also able to cross the blood-brain barrier (BBB) to reach the subarachnoid space of the leptomeninges, causing meningitis with or without septicemia (van Deuren et al., 2000). The crossing of these cellular barriers is essential for the development of the IMD, and although it has been intensively studied, it remains not fully understood. Different mechanisms have been proposed in different cellular, organ, and animal models, including transcellular and paracellular routes. An intracellular stage of the meningococcal infection has also been identified, favored by specific bacterial factors differentiating it from other Neisseria spp. such as the polysaccharidic capsule (Nikulin et al., 2006; Spinosa et al., 2007), as well as other virulence determinants that will be reviewed here. Thus, N. meningitidis adds to the growing list of pathogenic bacteria traditionally classified as extracellular and now considered to have a dual intracellular/extracellular lifestyle (Silva, 2012; Casadevall and Fang, 2020). Indeed, cellular infection by meningococci and the complex interactions between bacteria and the host cell in the intracellular microenvironment are essential not only for the progression of the meningococcal infectious cycle but also for determining the signs and symptoms of the IMD. In this article, we will review what is actually known about these interactions, which involve signal transduction, membrane trafficking, the cytoskeleton, metabolic cross-talk, and programmed cell death.

Genomic and phylogenetic analysis indicate that all commensals and pathogenic Neisseria spp. evolved from a common rod-shaped ancestor (Chen et al., 2021; Nyongesa et al., 2022). This was adapted to mucosal colonization and likely unencapsulated (Clemence et al., 2018; Priniski and Seifert, 2018). The two human pathogens belonging to the genus Neisseria, N. meningitidis and N. gonorrhoeae, have evolved from a common more recent ancestor and share a high similarity with genetic identity of 80-90% (Bennett et al., 2014; Maiden and Harrison, 2016; Vigué and Eyre-Walker, 2019). However, the N. meningitidis population is much more diverse than the N. gonorrhoeae (Vigué and Eyre-Walker, 2019). Potential explanations are: i. lower effective population size of N. gonorrhoeae because it evolved from N. meningitidis and went through a bottleneck after speciation as a consequence of ecological isolation in the human genital tract (Vázquez et al., 1993) or because of the specificity of its biology; ii. lower mutation rates in N. gonorrhoeae population than N. meningitidis population, which is characterized by the frequent occurrence of mutator clones in disease-associated lineages (Bucci et al., 1999; Richardson and Stojiljkovic, 2001; Richardson et al., 2002; Colicchio et al., 2006a; Hall and Henderson-Begg, 2006; Omer et al., 2011); iii. less diversity acquired by DNA recombination in N. gonorrhoeae than in N. meningitidis (Vigué and Eyre-Walker, 2019).

Intriguingly, although N. meningitidis is a naturally transformable species, its population is structured into rather stable clonal complexes (ccs), which cluster genetically associated sequence type (ST) strains defined by multi-locus sequence typing (MLST) of seven conserved housekeeping genes (Maiden et al., 2013). Meningococci belonging to different ccs are associated differently with carriage or disease status. For example, meningococcal strains belonging to cc11 are rarely found as colonizers and are overrepresented in IMD, with an estimated disease/carriage ratio of 12.59, whereas cc53 generally behaves as a commensal with an estimated disease/carriage ratio of <0.1 (Mullally et al., 2021).

The ccs associated with IMD are termed hypervirulent lineages (Maiden, 2008). There is no core pathogenome to distinguish carriage from invasive N. meningitidis strains. In fact, even factors such as the capsule, which is a crucial virulence factor finely regulated, are associated with multiple ccs and can be present in carriage isolates (Oldfield et al., 2018; Whaley et al., 2022). However, it is worth noting that capsule null strains or strains with mutations impeding capsule expression are more frequently clustered in carriage-associated ccs, such as cc198, cc1136, and cc53 (Moreno et al., 2015; Neri et al., 2019; Olof et al., 2023).

Recently, other genes have emerged to be potentially enriched in invasive meningococcal isolates compared to carriage ones. Mullally and coworkers found genomic islands (GIs) associated with hyperinvasive lineages (absent in cc53), which encode functions that facilitate meningococcal access to different cell types, leading to an increased risk for IMD (Mullally et al., 2021). These GIs are involved in meningococcal adhesion and subsequent host cell invasion, iron uptake (hpuAB), meningococcal survival in the intracellular environment, modulation of the host cell cycle, innate immune escape and evasion of phagocytic killing, epigenetic control of gene expression (modB), and competition within the meningococcal population in the human nasopharynx. Among these GIs, Mullally and coworkers found the glutathione peroxidase-encoding gene gpxA, the capsular biosynthetic genes, and pglI coding for an enzyme involved in the O-acetylation of the pilin glycan (Mullally et al., 2021). This latter modification is known to affect the chain length of the pilin glycan on the surface of the pilus structure, which in turn modulates the interaction between pili and the immune system, but also affects the meningococcal adhesion and subsequent invasion of the host cell (Power et al., 2003; Mubaiwa et al., 2017).

Mullally and coworkers also noted that the autotransporter NadA has been acquired by a branch of the hyperinvasive common ancestor (Mullally et al., 2021). NadA is a surface protein and is involved in adhesion to epithelial cells (Capecchi et al., 2005), brain microvascular endothelial cells (Kulkarni et al., 2020) and binds with high affinity to sialic acid-binding immunoglobulin-type lectins (Siglec)-5 and Siglec-14, promoting bacterial invasion (Benucci et al., 2024). In another study, the gene encoding phage transposase NEIS1048 and the associated single-nucleotide polymorphisms (SNP) glmU S373C encoding the enzyme N-acetylglucosamine 1 phosphate (GlcNAc 1 P) uridyltransferase have been associated with invasive lineages. The latter is involved in the synthesis of UDP-N-acetylglucosamine pyrophosphorylase (UDP-GlcNAc), which is a substrate for the synthesis of lipooligosaccharide (LOS), capsule, and CMP-NANA, the substrate for sialic acid (Eriksson et al., 2023). Meningococcal carriage and invasive isolates have also displayed differences in subsequences, k-mers, of fkbp, glmU, pilC, and pilE, suggesting that variation in these genes could play a role in infection capability (Eriksson et al., 2023). PilE is, in fact, the core structural protein of Type IV pili, while PilC is the adhesin located at the top of the pilus, which is the most crucial factor for initial adhesion of meningococci to host cells (Virji et al., 1992; Forest and Tainer, 1997; Merz and So, 2000). In contrast, commensal colonizer lineages, such as cc53, have not acquired these GIs and are characterized by thirteen unique loss-of-function loci (Mullally et al., 2021).

Intriguing differences were also reported in the structure of a two-partner secretion system (TPS)-encoding locus, tspA/tpsB (also named hrpA/hrpB), between hyperinvasive ccs and cc53 (Mullally et al., 2021). TPS is a secretion pathway that appears to play diverse roles in several Gram-negative bacteria and includes a large secreted protein (generally referred to as TpsA) and a channel-forming β barrel outer membrane activator/transport protein (TpsB) (Henderson et al., 2004; Schielke et al., 2010; Jacob-Dubuisson et al., 2013; Willett et al., 2015). Meningococci contain up to three different TPS systems. System 1 (TpsA/TpsB, also known as HrpA/HrpB) appears to be meningococcal-specific, while systems 2 and 3 are overrepresented in disease isolates compared to carriage isolates, but are also present in N. lactamica, which also has a distinct system 4 that is absent in meningococci (van Ulsen et al., 2008). In N. meningitidis, the HrpA/HrpB TPS was implicated in diverse functions, including competition between meningococci as a toxin-antitoxin fratricide system, biofilm formation, adherence to epithelial cells, intracellular survival, vacuolar escape, interaction with dynein, and modulation of apoptosis/pyroptosis (Schmitter et al., 2007; Talà et al., 2008, Talà et al., 2022; van Ulsen et al., 2008; Neil and Apicella, 2009; Arenas et al., 2013).

The tspA/tpsB locus evolved in the hyperinvasive lineages with the appearance of repeated N-terminal-truncated tpsA genes (tpsC cassettes) and Immunity Open Reading Frames (IORFs), which are absent in cc53 (Mullally et al., 2021). It has been shown that low-frequency recombination with silent tpsC cassettes, which share sequence similarity with the central region of tpsA but show an entirely different 3’-terminal sequence, may introduce different toxic modules at the variable C-terminus of meningococcal TpsA (Arenas et al., 2013). Thus, the presence of the tpsC cassettes and the cognate IORFs could confer increased plasticity to a genomic region that appears to be involved in the evolution of virulence in meningococcus.

More importantly, as will be discussed below, because the genome content tends to be highly conserved between bona fide carriage and disease isolates, subtle differences in the ability to modulate gene expression in host microenvironments, to engage in productive cross-talk with the cells also at metabolic level, to elude the host immune defenses, and to evolve over a short period of time have been proposed as major determinants of the hyperinvasive phenotype. The extraordinary ability of N. meningitidis to evolve rapidly (microevolution) leads to considerable diversity among circulating meningococcal strains and a different ability of these strains to colonize the human host and cause IMD.

As a human-adapted pathogen, N. meningitidis has developed multiple structures to interact with the host cell and evade the immune system. These include structures and proteins specialized in entering the host cell and surviving in the intracellular environment, differentially expressed between strains more closely associated with the disease and strains associated with carrier status.

The first step in the infectious cycle of the meningococcus is the adhesion to the host cell. Initially, this is mediated by type IV pili, filamentous structures composed of the major pilin PilE and three minor pilins PilV, PilX, and ComP (Forest and Tainer, 1997; Maier et al., 2004). It has been shown that, following type IV pili-mediated contact, the pilE gene encoding the major pilin and the genes for capsule synthesis are downregulated, and this is thought to allow for more intimate adhesion through other outer membrane adhesins. The best studied of these adhesins are the opacity proteins Opa and Opc (Bhat et al., 1991; Hauck, 2003). The former is present in both the meningococcus and the gonococcus, while Opc is only present in the meningococcus. Additionally, a variety of other meningococcal outer membrane proteins fulfill the role of minor adhesins, such as NadA, NhhA, App, MspA, and HrpA of the meningococcal TPS system 1 HrpA/HrpB (Hadi et al., 2001; Scarselli et al., 2006; Schmitt et al., 2007; Nägele et al., 2011; Khairalla et al., 2015). Notably, all these minor adhesins appear to be meningococcal-specific and are not found in gonococci (Comanducci et al., 2002, Comanducci et al., 2004; Turner et al., 2006; Schielke et al., 2010).

These multiple interactions mediating the intimate adhesion of meningococci to the host cell led to the activation of host signaling pathways for the rearrangement of the host cytoskeleton to form protrusions that engulf the bacteria. The internalization frequency, however, is not fixed, but it depends on both the receptors expressed by the host cell and meningococcal factors. For instance, antigenic variability of the PorB porin affects its binding to TLR2 and consequently bacterial internalization (Toussi et al., 2016). At the same time, LOS sialylation, regulated by the availability of exogenous sialic acid, the regulation of genes for endogenous sialic acid production, and the expression of Lst sialyltransferase, enhance internalization in immune cells expressing Siglec-5 and sialoadhesin (Jones et al., 2003; Chang and Nizet, 2014) but inhibits Opc-mediated interaction with the host cell (de Vries et al., 1998).

Once internalized, it is mandatory for the meningococcus to avoid lysosomal killing; thus, it modifies the internalization vacuole via Lysosomal Associated Membrane Protein 1 (LAMP1) cleavage through IgA protease (Lin et al., 1997) and ultimately mediates the rupture of the vacuole. HrpA has been involved in this process because of its Mn²⁺-dependent hemolytic activity residing in its C-terminal domain (Talà et al., 2022). LAMP1 is one of the most abundant proteins of the lysosomal membrane. It is highly glycosylated and provides a protective barrier from the action of lysosomal hydrolases, thus maintaining lysosomal integrity (Eskelinen, 2006; Li and Pfeffer, 2016), in addition to playing a role in maintaining the lysosomal acidic pH (Zhang et al., 2023). Meningococci interaction through type IV pili has been shown to induce an increase in Ca²⁺ levels sufficient to redistribute LAMP1 to the plasma membrane, making it accessible to the IgA protease (Ayala et al., 2001). Meningococcal type IV pili interact with CD147 (Basigin/EMMPRIN) (Bernard et al., 2014; Maïssa et al., 2017) and with Platelet Activating Factor Receptor (PAFr) (Jen et al., 2013), as demonstrated in endothelial cells and upper airway epithelial cells, respectively. The interaction with these receptors is probably the reason for the increase in Ca²⁺ levels. In fact, platelet-activating factor binding to PAFr leads to the production of inositol trisphosphate (IP3) through the phosphatidylinositol cycle, which in turn binds to inositol 1,4,5-triphosphate receptor type 1 (IP3R1) to induce the release of Ca²⁺ stored in the endoplasmic reticulum (Kroegel et al., 1989; Mazer et al., 1991; Bito et al., 1992). CD147, on the other hand, activates the FAK-Scr pathway, leading to tyrosine phosphorylation of IP3R1, which enhances the affinity of the receptor for IP3 (Jayaraman et al., 1996; Tang et al., 2015).

In the late endosomes and lysosomes, accumulation of Mn²⁺ may be obtained through the cation transporter ATP13A2 (PARK9), which has been identified as fundamental to prevent manganese toxicity, and proposed to shuttle manganese and other cations from the cytosol into the lysosomal lumen (Schmidt et al., 2009; Santoro et al., 2011; Nyarko-Danquah et al., 2020). Thus, meningococci-mediated depletion of LAMP1 potentially affects the internalization vacuole pH and its integrity. At the same time, the HrpA hemolytic activity is favored by the accumulation of Mn²⁺ in this compartment through PARK9. Overall, this can ultimately lead to the lysis of the vacuole, releasing meningococci into the cytosol. In agreement, meningococci deleted for IgA protease (Lin et al., 1997) or for HrpA (Talà et al., 2008) showed a dramatic reduction in survival/growth within different cell types. The proposed mechanism of meningococcal evasion from the internalization vacuole is illustrated in Figure 1.

Prevented from undergoing lysosomal degradation, the meningococcus must face the host defense mechanisms present in the intracellular environment. The capsule here plays a crucial role, probably protecting the bacterium from oxidative stress and antimicrobial peptides (Zaragoza et al., 2008; Brissac et al., 2021). Unencapsulated meningococci are, in fact, not able to survive inside the host cell, and capsule biosynthesis is upregulated in the intracellular environment (Spinosa et al., 2007). Spermidine has been shown to induce meningococcal capsule upregulation, increase meningococcal invasion by 5-fold, and enhance survival within macrophages (Kanojiya et al., 2022a). Spermidine is a polyamine whose concentration can reach the millimolar range in the intracellular environment, while extracellularly it is present in the low micromolar range (Heller et al., 1978; Lumkwana et al., 2022). Spermine, instead, is more easily found in the extracellular environment, and it has been shown to increase meningococcal adherence to epithelial cells and the expression of pilE. (Kanojiya et al., 2022b). Thus, meningococci may sense the intracellular environment through spermidine and increase capsule expression to survive in this microenvironment.

In the intracellular milieu, the meningococcus has been found to interact with Dynein Light Chain Tctex-Type 1 (DYNLT1) component of the motor protein dynein through a middle region of HrpA (Talà et al., 2022). This interaction enables the meningococcus to move along microtubules and come in contact with other organelles on this route, such as the mitochondrion (Talà et al., 2022). N. meningitidis can inhibit the intrinsic apoptosis pathway, translocating its porin PorB to the voltage-dependent anion channel (VDAC) on the mitochondrial outer membrane to prevent cytochrome c release (Massari et al., 2000, Massari et al., 2003). PorB-dependent apoptosis inhibition relies on HrpA-mediated meningococcal movement along microtubules, since infection in DYNLT1-silenced cells triggers apoptosis (Talà et al., 2022). Besides PorB, different meningococcal factors manipulate host cell death pathways. Among these, the autotransporters App and MspA can reach the nucleus, where they bind and cleave Histone H3 through their serine endopeptidase activity. This cleavage ultimately leads to a caspase-dependent cell death, which can be prevented by pan-caspase inhibitor Z-VAD-FMK (Khairalla et al., 2015). It has also been found that meningococcal IgA protease can translocate to the nucleus of the host cell because of the presence of a nuclear localization signal. Here, it cleaves the p65/RelA portion of NF-κB, inactivating it. Thus, the meningococcus can regulate NF-κB signaling, triggering its activation early in infection and then interrupting it. The signal, however, is not present in IgA protease of all meningococci but seems to be more represented in hyperinvasive strains (Besbes et al., 2015).

Meningococcal infections are characterized by a high inflammatory state (van Deuren et al., 1994; Møller et al., 2005; Zughaier, 2011; Ibrahim et al., 2024). This relies on the activation of signaling pathways upon the recognition of Pathogen-Associated Molecular Patterns (PAMPs) on bacteria or their Outer Membrane Vesicles (OMVs), mainly Toll-Like Receptors-4 (TLR4) and TLR2 signaling mediated by LOS (Christodoulides, 2021) and PorB recognition (Toussi et al., 2016), respectively. Moreover, PorB activation of TLR2 has been demonstrated to enhance bacterial internalization, suggesting that the expression of different PorB variants may be another factor affecting meningococcal internalization (Toussi et al., 2016).

OMVs are abundantly released from meningococci during infection and have been isolated in the blood and cerebrospinal fluid (CSF) of IMD patients (Stephens et al., 1982; Brandtzaeg et al., 1992). These vesicles are internalized through clathrin-mediated endocytosis, bringing multiple PAMPs into the host cell for the activation of inflammatory pathways. Key pathways activated include the canonical and non-canonical inflammasome. In the canonical pathway, NF-κB-dependent transcription of inflammasome components and their assembly is triggered by alterations in intracellular homeostasis. Consequently, pro-caspase-1 self-cleavage produces the activated caspase-1, which in turn cleaves gasdermin-D (GSDMD), and the N-terminal GSDMD fragments finally form pores into the plasma membrane through which inflammatory mediators are released, among which IL-1β and IL-18 (Franchi et al., 2009; Miao et al., 2010). In the non-canonical pathway, LOS on OMVs or intracellular bacteria is first recognized by guanylate-binding proteins (GBPs) that recruit caspase-4 (murine caspase-11) (Vanaja et al., 2016; Wandel et al., 2020). Activation of caspase-4 subsequently activates GSDMD, and the Damage-Associated Molecular Pattern (DAMPs) released can activate the inflammasome (Kayagaki et al., 2015; Liu et al., 2016). Excessive activation of these pathways leads to strong inflammation and death of the cell mediated by Ninjurin1 (NINJ1)-dependent cell lysis (Kayagaki et al., 2021), termed pyroptosis (Bergsbaken et al., 2009).

N. meningitidis activates both canonical and non-canonical pathways in the host, with predominance of the alternative caspase-3 mediated-GSDME activation in vitro in different cell lines (Talà et al., 2022) and of the GSDMD-dependent pathway in vivo in an intracisternal infected mouse model of meningitis (Pagliuca et al., 2024). Canonical pathway activation seems to be dependent on the intracellular localization of the bacterium. In fact, the infection with meningococci defective for HrpA/HrpB TPS is strongly impaired in the activation of caspase-1 and IL-1β and IL-18 release (Talà et al., 2022; Pagliuca et al., 2024), but if the integrity of the internalization vacuole is destabilized through DYNLT1-silencing (Driskell et al., 2007; Flores-Rodriguez et al., 2011; Yap and Winckler, 2022; Yap et al., 2022), caspase-1 activation occurs as in cells infected with wild-type meningococci (Talà et al., 2022).

Finally, the Cas9 protein of the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-Cas9 system has been implicated in meningococcal ability to adhere to and invade the host cell, and especially to survive in the intracellular environment (Sampson et al., 2013). The CRISPR-Cas9 system is normally used by bacteria as a defense mechanism against invading exogenous nucleic acid, but evidence suggests that this system has a more extensive function in bacterial physiology. For instance, in Francisella novicida, Cas9 has been found to repress expression of an endogenous mRNA (Sampson et al., 2013). Thus, meningococci may use the CRISPR-Cas9 system to regulate the expression of crucial factors for intracellular survival.

The final fate of intracellular meningococci is largely unknown, but it has been found that GPB1, mediating the initial assembly of GBP platform on cytosolic Gram-negative intracellular bacteria, and GBP3, a component of the GBPs recruited to the platform for caspase-4 activation, selectively lyse F. novicida and N. meningitidis (Feng et al., 2022). Therefore, the intracellular cytosolic cycle of meningococci may end with GBPs-mediated killing or with the release from the cell if pyroptosis occurs.

Overall, scientific efforts over the past twenty years have shed light on the behavior of meningococcus within the host cell and have supported a model in which, after internalization by the host cell, the meningococcus is able to avoid lysosomal killing by cleaving LAMP1 through IgA protease, and to escape from the internalization vacuole through the C-terminal hemolytic domain of HrpA, reaching the cell cytosol. Here, the meningococcus is able to sense the cytosolic environment, modulating the expression of the capsule and other virulence determinants. It exploits the microtubule cytoskeleton and dynein to move within the cell and interact with cellular structures, including mitochondria and the inflammasome. Using these mechanisms, which involve capsule, LOS, type IV pili, IgA protease, HrpA, PorB and other virulence determinants, the meningococcus is able to survive/multiply within infected cells and modulates the balance between apoptosis and pyroptosis.

The brain is normally protected from exogenous insults, such as pathogenic bacteria and toxic substances, through the meninges, the blood-brain barrier (BBB), and the blood-cerebrospinal fluid barrier (BCSFB). The meninges envelop the Central Nervous System (CNS) and are composed of three membranes: from the outermost, the dura mater, the arachnoid mater, and the pia mater (Kadry et al., 2020). Between the arachnoid membrane and the pia mater, there is the subarachnoid space, which contains cerebrospinal fluid (CSF) produced by the choroid plexuses (Menéndez González, 2023). The BBB regulates the passage of ions, nutrients, and oxygen from the blood to the brain and consists of endothelial cells supported by astrocytes and pericytes. The barrier function is established by the lack of fenestrae in the vessels, the tight junctions connecting the endothelial cells, which prevent the paracellular route for external agents, and the low level of transcytotic vesicles (Rodrigues and Granger, 2015). The stability of tight junctions between endothelial cells is guaranteed by interactions with pericytes (Jo et al., 2013), and by Wnt proteins, vascular Endothelial Growth Factor (VEGF), and Transforming Growth Factor-β (TGF-β), released by astrocytes (Guérit et al., 2021; Pivoriūnas and Verkhratsky, 2021; Sobral et al., 2025). Pericytes are also crucial for the suppression of transcytosis. This is achieved through the induction of endothelial expression of Major Facilitator Superfamily Domain-containing 2a (Mfsd2a), which is a transporter of docosahexaenoic acid (DHA) (Ben-Zvi et al., 2014; Nguyen et al., 2014). The incorporation of DHA creates a unique lipid composition of brain endothelial membranes, suppressing the formation of caveolae (Nguyen et al., 2014). A perivascular space, located between endothelial cells and astrocytic endfeet, can be present. This, often absent in capillaries, is called the Virchow-Robin perivascular space, and it is filled with interstitial fluid that has lymphatic functions (Esiri and Gay, 1990).

The BCSFB barrier is composed of an ependymal epithelium that produces the CSF, characterized by apical tight junctions located at the choroid plexuses. Choroid plexuses are ventricular structures that surround a stroma rich in fenestrated capillaries (Damkier and Praetorius, 2020). Postmortem examination of patients with purpura fulminans and IMD revealed meningococcal microcolonies associated with the endothelium of different vessels of the two barriers, such as the capillaries of the choroid plexuses, meningeal vessels, and brain parenchyma vessels (Mairey et al., 2006; Dando et al., 2014; Ridpath et al., 2014). This supports the hypothesis that meningococcal translocation across brain barriers occurs primarily through the endothelium, without the active participation of immune cells. Notably, no bacteria were observed in contact with the ependymal epithelium of the choroid plexuses, suggesting that meningococcus does not cross this barrier (Dando et al., 2014).

A deeper understanding of the dynamics of meningococcal colonization of different vascular beds was possible thanks to the use of a mouse model grafted with human skin (Manriquez et al., 2021). Similar to what was observed in human patients, the meningococcus colonized capillaries, venules, and arterioles with the progressive accumulation of recruited neutrophils in the humanized mouse model. Neutrophil recruitment was found to be subordinate to the Type IV pili-dependent adhesion to the endothelium, and it was crucial in controlling the number of adherent meningococci (Manriquez et al., 2021). In fact, neutrophils were able to detach meningococci bound to the endothelium and phagocytize them. However, N. meningitidis rapidly colonized capillaries and arterioles before the activation of an efficient inflammatory response. Neutrophils fail to translocate into capillaries and arterioles because of the absence of adhesion molecules such as E-selectin on these vessels (Manriquez et al., 2021). Thus, these can be used by the meningococcus to evade detection by neutrophils, allowing the infection to proceed.

Meningococcal adhesion and invasion of endothelial cells have proven to be a prerequisite for further dissemination. It has been demonstrated that at least 10 bacteria per microcolony on endothelial cells are needed to induce the recruitment of ezrin and the consequent cytoskeletal remodeling leading to the internalization of the bacteria (Soyer et al., 2014). Meningococcal internalization is an early event that is subsequently inhibited when the microcolonies develop into a more diffuse layer on endothelial cells (Eugene et al., 2002). The formation of the internalization vacuole for endothelial cell invasion has been extensively investigated. Type IV pili binding to CD147/Basigin triggers tyrosine phosphorylation through c-Src and focal adhesion kinase, and EGFR is activated (Slanina et al., 2010, Slanina et al., 2012, Slanina et al., 2014), and acid sphingomyelinase increases ceramide levels on the plasma membrane, forming platforms for the recruitment of host receptors (Simonis et al., 2014; Fohmann et al., 2023). Ezrin, moesin, CD44, ICAM-1, and the tyrosine kinase receptor ErbB2 are recruited. Rho and Cdc42 GTPases orchestrate actin polymerization, and microvilli-like structures are formed (Hoffmann et al., 2001; Eugene et al., 2002). Opc-positive meningococcal lineages have been demonstrated to be more efficiently internalized. This is related to the Opc binding to fibronectin and vitronectin in the host serum, which act as a bridge for the connection with integrin α5β1 and αvβ3, respectively. Opc-mediated internalization is likely host-specific; in fact, in vitro, only in the presence of human serum, and not the routinely used bovine-derived serum, were meningococci internalized into endothelial cells (Unkmeir et al., 2002). Additionally, Opc binds preferentially to sulphated tyrosine residues of activated vitronectin (Sa E Cunha et al., 2010), the positions of which are different between vitronectins from different species. Coherently, Opc was able to bind in vitro activated human vitronectin but not bovine or murine vitronectins (Sa E Cunha et al., 2010). Opc-mediated binding to integrins led to the formation of focal adhesions with the recruitment of vinculin, talin, paxillin, and cortactin. The last binds and activates the Arp2/3 complex, a central organizer of actin filament structure crucial for Opc-mediated internalization. In fact, a cortactin mutation interrupting the binding to the Arp2/3 complex has been demonstrated to reduce the number of internalized bacteria into endothelial cells (Slanina et al., 2012).

It is worth noting that the internalization vacuole formed during gonococcal infection is quite different from that formed during meningococcal infection. N. meningitidis triggers the formation of microvilli-like protrusions (Eugene et al., 2002) while N. gonorrhoeae induces membrane ruffles and lamellipodia (Edwards et al., 2000). This discrepancy may contribute to the differences in the intracellular fate of the two pathogens, likely due to the distinct target cells and virulence factors possessed by the bacteria. For instance, Opc is possessed only by the meningococcus, and Opa proteins have different affinities for CEACAM members, with meningococcal Opa primarily targeting CEACAM1 and gonococcal Opa proteins binding to CEACAM1, CEACAM5, CEACAM6, and CEACAM3 on epithelia, neutrophils, and lymphocytes (Sadarangani et al., 2011; Martin et al., 2016).

N. meningitidis binds to the hetero-oligomeric complexes formed by CD147 and the G-protein-coupled receptor (GPCR) β2-adrenergic receptor on endothelial cells through Type IV pili (Coureuil et al., 2010). This led to a biased activation of β-arrestin, which regulates GPCRs (Coureuil et al., 2010), promoting endocytosis. Disruption or delocalization of ZO-1, occludin, and claudin-5 has been observed in vitro in endothelial cells infected with N. meningitidis (Martins Gomes et al., 2019). The β-arrestin signaling has been associated with junctional protein depletion; in fact, GPCRs bind to PDZ-domain containing proteins such as ZO-1, affecting the localization and stability of junctional proteins (Dunn and Ferguson, 2015; González-Mariscal et al., 2018). Moreover, GPCR signaling has been shown to transactivate different metalloproteinases (MMPs) (Lin, 2025), and MMP-8 has been found critical in occludin cleavage upon meningococcal infection in endothelial cells (Schubert-Unkmeir et al., 2010). Mislocalization and degradation of these proteins have been associated with a leakage of the endothelial barrier sufficient for the meningococcus to cross it via the paracellular route. However, meningococci enter the brain with a minimal increase in BBB permeability. Moreover, meningococcal meningitis is not associated with the formation of thrombotic lesions (Join-Lambert et al., 2013), making it unlikely that the barrier will be crossed following cell death. Most observations suggestive of barrier loss were made in vitro, and although a paracellular route is possible and likely to occur, the internalization of meningococci in endothelial cells has been clearly demonstrated (Eugene et al., 2002; Doulet et al., 2006; Nikulin et al., 2006; Dupin et al., 2012; Martins Gomes et al., 2019). The N. meningitidis MC58 strain was shown to be internalized into the endothelial cells and to cross the barrier within 24 hours without disruption of junctional proteins or alteration in TEER, indicating a transcellular crossing (Endres et al., 2022). Alterations in barrier functions were instead observed by the authors later in the infection (Endres et al., 2022). In agreement, genes involved in barrier permeability, as well as genes involved in cell surface rearrangements and endocytosis, were found to be differentially regulated in endothelial cells infected with the meningococcus (Káňová et al., 2019).

Type IV pili interaction with CD147 on endothelial cells has also been demonstrated to induce an increase in sphingosine 1-phosphate receptor 2 (S1PR2) and the activation of the S1P-S1PR2-EGFR axis, leading to bacterial uptake (Fohmann et al., 2023). S1P is a signaling sphingolipid produced by the phosphorylation of sphingosine by two sphingosine kinases, SphK1 and SphK2 (Maceyka et al., 2012). Meningococcal infection induces sphingosine phosphorylation through SphK1 in a mechanism strictly dependent on pili-mediated interaction with the cell (Fohmann et al., 2023). Consequently, a continuous release of S1P is observed (Fohmann et al., 2023). The latter binds to the receptors S1PR1-5. S1PR1, S1PR2, and S1PR3 are ubiquitous, while S1PR4 and S1PR5 expression is restricted to hematopoietic cells (Sun et al., 2024). Notably, while S1PR2 is upregulated and induces the internalization of meningococci, the activation of S1PR1 or S1PR3 reduces the number of intracellular bacteria (Fohmann et al., 2023). S1P is a regulator of BBB integrity. The interaction of S1P with S1PR1 improves barrier function, while interaction with S1PR2 causes barrier leakage. The balance between S1PR1–3 and S1PR2 signaling is critical for both paracellular and transcellular integrity of the BBB (Garcia et al., 2001; Prager et al., 2015; Wiltshire et al., 2016). In particular, S1PR1 signaling promotes the assembly and maintenance of adherent junctions (Anwar and Mehta, 2020) while S1PR2 signaling disrupts adherent junctions via Rho/ROCK, which can also induce the phosphorylation of occludin and claudin-5, affecting tight junctions (Hirase et al., 2001; Sanchez et al., 2007).

Additionally, internalized bacteria can amplify the inflammatory response, again affecting junctional proteins. Intracellular bacteria have been shown to induce activation of BAD, BAX, caspase-3, and AMPK (Schubert-Unkmeir et al., 2007). Alteration of ion homeostasis is also observed, together with the induction of HIF1 signaling and oxidative stress (Martins Gomes et al., 2019). Caspase-3 activation can have a strong impact on barrier integrity. In fact, it has been demonstrated that caspase-3 leads to ZO-1 and claudin-5 disruption in endothelial cells even in the absence of apoptotic death, indicating a non-apoptotic role in barrier regulation (Zehendner et al., 2011; Yu et al., 2016).

Thus, considering also the natural inhibition of endocytosis in the BBB barrier, it is possible that early in infection, only a few meningococci can cross the barrier through the transcellular route. Later in the infection, the combination of signals and pathways induced for bacterial internalization may promote a localized loss of barrier function, allowing the paracellular route to occur, making bacterial internalization a requisite step for further invasion. Although the barrier function is maintained after the bacteria cross it, this step is essential in meningococcal pathogenesis. In fact, this allows reaching the CNS before an intense inflammatory response is elicited. Once in this microenvironment, the replication of the meningococcus in the CFS promotes the intense inflammation leading to fulminant meningitis. Hence, BBB crossing prevention, blocking N. meningitidis dissemination to the CNS, could have a major impact on disease outcomes and its sequelae. The proposed mechanisms for the BBB crossing by the meningococcus are illustrated in Figure 2.

Since, as mentioned above, most virulence determinants are uniformly distributed in the N. meningitidis population with no obvious difference between the clonal complexes mostly associated with IMD and those rarely found in IMD cases, differential gene expression has been proposed as an important determinant of the hyperinvasive phenotype (Joseph et al., 2011; Schoen et al., 2014; Tan et al., 2016; Ampattu et al., 2017; Ren et al., 2017). This indication arises from the results of some fairly recent comparative transcriptome studies, but also from some previous studies on the expression of single genes.

By comparatively analyzing the overall transcriptional profile of the MC58 strain, belonging to the hyperinvasive cc32 (ST-32), with that of the α710 strain, a carrier isolate belonging to the cc136 (ST-136), during adhesion to human nasopharyngeal cells, Joseph et al. (2010) found notable differences in the expression of specific genes mainly involved in meningococcal metabolism. In particular, the 55 most highly expressed genes in the invasive MC58 strain included genes encoding proteins involved in amino acid transport and metabolism (argH, aroA, aroB, ilvC, and gdhA), genes for ATP synthase subunits (atpA, atpD, atpG), genes involved in sialic acid capsule biosynthesis (siaB, siaC), and an operon encoding Na⁺-translocating NADH-quinone reductase subunits (nqrB, nqrC, nqrD). In contrast, the 81 genes that were higher expressed in the carriage strain α710 comprised, among the others, genes involved in inorganic ion transport and metabolism and genes encoding two sigma factors (rpoD and rpoE), the tuf genes for the elongation factor Tu, clpX (encoding the Clp protease), a tpsA gene, and 14 genes encoding ribosomal proteins.

Furthermore, it is interesting to note that, with the exception of abcZ, which encodes an ABC transporter, the other six housekeeping genes used to analyze the structure of the meningococcal population by MLST to identify strains belonging to certain ccs, more or less associated with meningococcal disease or carrier status, encode enzymes involved in the main metabolic pathways (Yazdankhah et al., 2004; Maiden, 2006). Specifically, adk encodes adenylate kinase, aroE encodes shikimate dehydrogenase required for chorismate biosynthesis, fumC encodes fumarate hydratase of the tricarboxylic acid cycle (TCA), gdh encodes the glucose-6-phosphate 1-dehydrogenase, pdhC (aceE) encodes the pyruvate dehydrogenase subunit E1, and pgm encodes phosphoglucomutase. Although genetic variations at these loci have long been considered neutral, the availability of a significant amount of epidemiological data and the use of mathematical models have allowed us to discover that some combinations of alleles at these loci may be subject to selection and that some combinations in some hypervirulent lines are associated with small gains in fitness for meningococcal transmission (Buckee et al., 2008). These data support the link between metabolic efficiency, increased transmission fitness, and increased incidence of IMD in several clonal complexes as theorized (Stollenwerk et al., 2004; Moxon et al., 2006).

This conclusion was supported by comparative genomic, transcriptomic and phenotypic profiles of meningococcal isolates from disease patients and their household contacts, which led to identification of potentially important metabolic differences between carriage and disease isolates including the sulfate assimilation pathway, in addition to the observation that several carriage isolates had lost their type IV pili and that this loss correlated with reduced induction of tumor necrosis factor alpha (TNF-α) expression when cultured with epithelial cells (Ren et al., 2017).

An interesting study compared the gene expression of the invasive MC58 strain and the α522 carrier strain under ex vivo conditions simulating the commensal and virulence compartments to assess the strain-specific impact of gene regulation on meningococcal virulence (Ampattu et al., 2017). The authors found that these strains differed in the expression of over 500 genes under conditions mimicking infection, in spite of indistinguishable ex vivo phenotypes. These differences specifically included metabolic genes, including those involved in the incomplete denitrification pathway, information processing genes, as well as genes known to be involved in host damage, and numerous LOS biosynthesis genes. Furthermore, a model-based analysis of transcriptomic differences in human blood suggested differences in metabolic flux in the energy, glutamine, and cysteine metabolic pathways, along with differences in stringent response activation (Ampattu et al., 2017). This computational result was supported by experimental analyses that revealed differences in conditional cysteine and glutamine auxotrophy, as well as a strain- and condition-dependent essentiality of the (p)ppGpp synthase gene relA associated with a short non-coding AT-rich repeat element in its promoter region (Ampattu et al., 2017). The conclusion of this study is that some differences between different meningococcal strains in the expression of certain metabolic traits may be cryptic due to “transcriptional buffering”, and can be revealed only under particular stress conditions. This study also highlights the importance of the stringent response in meningococcal virulence and the differences in its activation in different meningococcal strains. This study also shows that the gene differences between the MC58 pathogenic strain and the α522 carrier, in particular in the blood, may be caused by the activation of different sets of regulatory genes, in addition to RelA, including those encoding: i. MisR, a PhoP-family response regulator of the two-component signal transduction system MisR/MisS that regulates various cellular processes in pathogenic Neisseria, including LOS structure and modification, iron assimilation, and resistance to host defenses such as serum complement and oxidative stress; ii. Fur, the major transcriptional regulator of iron homeostasis; iii. FNR, the master regulator involved in the adaptation to oxygen-limited conditions; iv. the alternative sigma factor E (σE) that in pathogenic Neisseria species is activated in response to oxidative stress and negatively controls transcription of aniA, encoding the nitrite reductase (Ampattu et al., 2017).

The analysis of the genes involved in glutamate metabolism may provide a concrete example of how differences in gene regulation between meningococcal strains are directly associated with their pathogenic potential. Northern blot analysis showed differential expression of ghdA in different N. meningitidis strains, and the highest levels of gdhA mRNA were observed in strains belonging to the hypervirulent clonal complexes ST-32 (ET-5, serogroup B) and ST-4 (subgroup IV-1, serogroup A) (Pagliarulo et al., 2004). The underlying mechanisms were investigated, leading to the discovery that in strains expressing high levels of gdhA mRNA, two promoters, gdhA P1 and gdhA P2, initiate transcription of gdhA. In contrast, in strains expressing low levels of mRNA, gdhA P2 was not active due to the weak expression of gdhR, an associated regulatory gene, as a result of the insertion of the miniature neisserial element (nemis also known as Correia element) downstream of gdhR that destabilizes its mRNA. The study also showed that: i. transcription from gdhA P2 was maximal in complex medium during late logarithmic growth phase and in chemically defined medium when glucose instead of lactate was used as a carbon source in the presence of glutamate; ii. 2-oxoglutarate, a product of the catabolic reaction of the NADP-GDH and an intermediate of the TCA cycle, inhibits the binding of GdhR to gdhA P2 (Pagliarulo et al., 2004).

N. meningitidis is known to possess a relatively low number of transcriptional regulators compared to other bacterial species. Thirty-five are annotated in the MC58 strain genome, and of these 35, only a few have been well characterized: i. CrgA, a LysR-type regulator that is upregulated upon contact with human epithelial cells and represses its own transcription and that of the type IV pili subunits pilE and pilC1; ii. AsnC, a global transcriptional regulator that controls the response to poor nutritional conditions, which are sensed by binding of this regulator to leucine and methionine; iii. Fur, the above-mentioned global regulator of iron homeostasis, which exerts its function also indirectly controlling the expression of small regulatory RNAs; iv. the fumarate and nitrite reductase regulator FNR, which in case of oxygen limitation allows the meningococci to survive by changing the metabolism toward sugar fermentation and denitrification; v. NsrR, a repressor of a compact regulon of genes in the absence of nitric oxide; vi. HexR, an RpiR-like transcriptional regulator, HexR, that is responsible for part of the glucose-responsive regulation and affects the fitness of the meningococcus in vivo (Claus et al., 2007; Schielke et al., 2010; Antunes et al., 2015).

In addition to transcriptional regulators, N. meningitidis utilizes small regulatory RNAs that are important for meningococcal adaptation to specific environmental changes during colonization and invasive infection. Three of these transcoding noncoding RNAs have been characterized: NrrF, AniS, and NmsR. NrrF is essential for iron homeostasis, is expressed under iron-limiting conditions under the control of Fur and is known to inhibit the translation of sdhA and sdhC mRNAs, which encode subunits of succinate dehydrogenase complex, an iron-containing enzyme. AniS is part of the FNR regulon and may be responsible for the downregulation of FNR-repressed genes. NmsR downregulates mRNAs targeting prpB and prpC, encoding proteins involved in the methylcitrate cycle (Eichner et al., 2022).

Despite this relatively small number of transcriptional regulators and small regulatory RNAs, meningococcus has a broad capacity to regulate gene expression at the population level, thanks to the presence of sophisticated ON/OFF phase variation mechanisms. Specifically, the N. meningitidis genome is characterized by an abundance of homopolymeric DNA repeats (simple sequence DNA repeats, SSRs) that undergo stochastic and reversible mutations (insertion/deletion) at a high frequency by slipped-strand mispairing during DNA synthesis associated with DNA replication or recombination (Saunders et al., 2000; Snyder et al., 2001). The length of the SSRs, when located within a coding region, can change translation by introducing a frameshift in the reading frame, or, when located in the proximity of a promoter, can modulate transcription either by altering the promoter sequence or by switching between alternative promoter sites. This sophisticated mechanism of genetic variation of “contingency” genes is responsible for a broad phenotypic diversity at the population level and is thought to facilitate the adaptation of meningococci to dynamic environmental changes (Alexander et al., 2004; Davidsen and Tønjum, 2006; Moxon et al., 2006; Klughammer et al., 2017; Green et al., 2020). It is noteworthy that the rate of phase variation is modulated in the meningococcal population due to the presence of mutator phenotypes (see below). A population-scale comparative genomic analysis identified 277 genes and classified them into 52 strong, 60 moderate, and 165 weak candidates for phase variation. Deep-coverage DNA sequencing of single colonies grown overnight under non-selective conditions confirmed the presence of high-frequency, stochastic variation in 115 of them (Siena et al., 2016). Functional characterization of these phase-variable 115 genes demonstrated an enrichment for those encoding already known surface determinants (capsule, LOS, adhesins, capsule, nutrient-scavenging proteins) or DNA metabolism. However, among phase-variable genes, genes involved in a broad spectrum of other metabolic functions were found. In addition, most of the variable SSRs were predicted to induce phenotypic changes by modulating gene expression at a transcriptional level or by producing different protein isoforms rather than mediating on/off translational switching through frameshifts as previously known (Siena et al., 2016).

Among the phase-variable genes, there are genes encoding DNA methyltransferases (Mod) belonging to the type III restriction-modification systems. This is of particular interest because Mod proteins mediate epigenetic control (Srikhanta et al., 2005) Random, reversible mutation of simple sequence DNA repeats within the open reading frame of mod genes leads to frameshift mutations and ON/OFF Mod expression. This results in a different methylation pattern of the genome and modified expression of specific sets of genes under the control of specific Mod proteins. These regulons are termed “pasevarions” (phase variable regulons) and are implicated in strain differences in virulence traits (Srikhanta et al., 2005; Tan et al., 2016).

In conclusion, gene regulation is increasingly recognized as a key determinant of meningococcal virulence. In fact, virulence determinants are uniformly distributed in the N. meningitidis population, with no obvious differences between clonal complexes more or less associated with IMD. Therefore, their differential regulation makes the difference. Differential gene expression between hypervirulent and carrier strains affects numerous genes, many of which are involved in meningococcal metabolism, highlighting its importance in meningococcal disease, while other genes encode virulence determinants or proteins involved in genome stability and gene regulation (Table 2). However, there is no uniform pattern that can dichotomously distinguish hypervirulent strains from carrier strains, which indicates that meningococcal pathogenicity is complex and multifactorial.

Several studies on N. meningitidis, a microorganism characterized by high genomic plasticity, have highlighted the relationship between microevolution, antimicrobial resistance, vaccine escape, and hyperinvasive phenotype (Kugelberg et al., 2008; Brehony et al., 2016; Mikucki and Kahler, 2023). The high plasticity of the genome arises from several factors, among which the most relevant appear to be: i. the natural competence of N. meningitidis for natural transformation; ii. The presence of sophisticated mechanisms of phase variation of “contingency” genes; iii. The presence of specific mechanisms of antigenic (and functional) variation of key determinants involved in the interaction with the host; iv. The high frequency in the meningococcal population of allelic variants coding for proteins with reduced functionality involved in DNA mismatch repair encoding proteins (MutS, MutL) and homologous recombination (RecB, RecC, UvrD) (Feil et al., 1999; Holmes et al., 1999; Richardson and Stojiljkovic, 2001; Richardson et al., 2002; Colicchio et al., 2006a).

The high propensity for genetic transformation of meningococcus, a bacterium with minimal genetic barrier to recombination and a unique, non-regulated competence system that allows it to capture DNA throughout the entire growth cycle, is such that this species presents non-clonal, but panmictic characteristics (Maiden, 1993, Maiden, 2006; Spratt and Maiden, 1999; Bentley et al., 2007). Furthermore, the mutator mutS and mutL alleles contribute to reducing the genetic barrier to recombination (Alexander et al., 2004; Davidsen and Tønjum, 2006), as it is known that MutS and MutL mismatch binding activity edit (prevent) homologous and homoeologous recombination, limiting the stability of heteroduplex intermediates, with the contribution of the RecBCD nuclease (Štambuk and Radman, 1998).

The high incidence of mutS and mutL mutator alleles in the N. meningitidis population is thought to be a primary factor in the adaptive evolution of meningococcus, increasing overall spontaneous mutation rates and the rate of phase variation of “contingency” genes (Saunders et al., 2000; Siena et al., 2016; Klughammer et al., 2017; Green et al., 2020). Evidence is provided that asymptomatic meningococcal carriage on the nasopharyngeal mucosal surface is facilitated by localized hypermutation in “contingency” genes, which, together with horizontal gene transfer, affects the expression of a plethora of surface determinants involved in meningococcal adhesion, metabolism, intracellular survival, intra-species and interspecies competition, and escape of immune response (Green et al., 2020).

It has been proposed that meningococcal virulence results from the accidental emergence of invasive variants during carriage (Klughammer et al., 2017). It is worth mentioning, in this regard, a study that, analyzing a laboratory accident with a mutS-defective mutator strain of Neisseria meningitidis, which was responsible for a case of IMD that fortunately evolved with antibiotic treatment until complete recovery, describes the genotypic and phenotypic modifications of the bacterium after accidental passage into humans (Omer et al., 2011). The in vivo passage was responsible for the modification in key phase-variable genes. In particular, compared to the parental strain, the meningococcal isolate from the patient’s blood utilized a different hemoglobin-bound iron receptor (HpuA/B) than the parental strains (HmbR), expressed different pilin variants with different adhesion properties, and showed a different LOS immunotype. The study of this episode demonstrates the dangerousness of mutant strains of meningococcus.

Based on phase variation frequencies in hpuAB and hmbR, encoding two distinct hemoglobin receptors, Richardson and Stojiljkovic (2001) identified mutL alleles encoding MutL variants with multiple amino acid substitutions in MutL, associated with three distinct switching phenotypes, slow, medium, and fast, together with a mutS allelic variant associated with fast switch and high spontaneous mutation frequency to rifampicin-resistance. In a reference strain belonging to the hypervirulent ET-37 electrophoretic type (clonal complex ST-11), 93/4286, mutS was found to be inactivated by IS1106, while mutL allelic variants associated with high spontaneous mutation frequency were detected among isolates belonging to ET-24 electrophoretic type (lineage 3, clonal complex ST-41/44) (Colicchio et al., 2006a), supporting the idea that the activity of the DNA mismatch repair system, in particular MutL activity, is subject to a sophisticated mechanism of modulation within the species N. meningitidis. Furthermore, the mutator phenotype associated with defective MutL activity was suppressed when a non-functional recB allele, derived from ET-37 meningococcal strains, replaced the functional recB allele, suggesting that in MutL-deficient strains, hypermutation mostly arises during post-replicative DNA synthesis associated with the activity of the RecBCD recombination pathway (Colicchio et al., 2006a). The high number of allelic variants in genes involved in DNA mismatch repair and RecBCD recombination, which could correspond to functional variations in the encoded proteins, could, at least in part, explain why the N. meningitidis population is much more diverse than that of N. gonorrhoeae (Vigué and Eyre-Walker, 2019), as well as the rapid evolution of hypervirulent clones.

N. meningitidis is a paradigmatic microorganism that exploits the high plasticity of its genome to continuously adapt to its human host by genome microevolution. The high genome plasticity arises from well-defined and interconnected factors: i. natural competence for transformation; ii. specific mechanisms of phase variation of “contingency” genes; iii. specific mechanisms of antigenic (and functional) variation; iv. high frequency of allelic variants coding for proteins with reduced functionality involved in DNA mismatch repair and RecBCD recombination. The association between microevolutionary capacity and propensity to cause IMD in several hypervirulent lineages is particularly intriguing, as it has been proposed that meningococcal virulence results from the accidental emergence of invasive variants during carriage.

Within a genus such as Neisseria, characterized by notable genomic plasticity and consequent genetic heterogeneity, a key point of concern is the emergence of species potentially pathogenic for humans and other animals. In this regard, the evolutionary history of the gonococcus, discovered by Albert Neisser in 1879, closely related to the meningococcus, can exemplify this concept.

Although description of gonorrhea-like symptoms dates back to ancient civilizations like the Romans, Jews, and Arabs, genomic analysis suggests that the modern gonococcal population originated much more recently, possibly in the 16^th century, and was subsequently spread globally (Golparian et al., 2020). The genomic analysis provides evidence that the modern gonococcus is not as old as previously thought and shows that strains from the postmodern era (1980-twenty-first century) have evolved from strains belonging to the pre-antibiotic era (pre-1950s) and mainly form a separate clade in the tree. Furthermore, the gonococcal genome has become increasingly conserved over time, with isolates becoming increasingly clonal and the level of recombination not substantially different in different isolates (Golparian et al., 2020). This is consistent with the finding that N. meningitidis has acquired more of its diversity by recombination than N. gonorrhoeae (Vigué and Eyre-Walker, 2019), consistent with the phylogenetic reconstruction of Neisseria species performed with housekeeping genes transmitted with high verticality in prokaryotes (Gogarten et al., 2002; Nagies et al., 2020), showing a considerably higher radiation of N. meningitidis sequences compared to N. gonorrhoeae sequences available in the NCBI database (Figure 4).

Despite the high genetic similarity between N. meningitidis and N. gonorrhoeae, these two bacteria occupy distinct niches in the human host, i.e, the nasopharynx in the case of N. meningitidis and the urogenital tract in the case of N. gonorrhoeae, are responsible for distinct diseases, and are characterized by a very different infectious cycle. This is the result of a continuous adaptation of the gonococcus to its diverse ecological niche, an adaptation influenced by the use/misuse of antibiotics in more recent times, and which has mainly involved: i. modifications of some metabolic pathways (Potter and Criss, 2024); ii. modifications affecting global and pathway-specific regulatory systems (Schielke et al., 2010); iii. modifications of some determinants of the bacterial surface, involved in the interaction with the host (Schielke et al., 2010; Song et al., 2020). For example, the sulfate assimilation pathway for cysteine ​​biosynthesis and the methylcitrate pathways are in decay in the gonococcus (Potter and Criss, 2024; Talà et al., 2025), and in the 1980s, a gonococcus clade with an arginine, hypoxanthine, and uracil auxotrophy was associated with disseminated gonococcal infection (Noble et al., 1984). Furthermore, some gonococcal strains have been shown to harbor the truncated lactoferrin receptor LbpAB, despite the fitness advantage conferred by possessing a functional LbpAB in experimental infection of the human male urethra (Anderson et al., 2003). It is also interesting to note that N. meningitidis and N. gonorrhoeae share a large number of transcriptional regulators, but some, such as FadR and GdhR, have differential functions in the two bacteria (Pagliarulo et al., 2004; Schielke et al., 2010; Rouquette-Loughlin et al., 2017; Ayala et al., 2022; Mortimer, 2022). Regarding gonococcal surface structures, including type IV pili and the CEACAM-binding protein Opa, these have evolved primarily to establish and maintain colonization while inhibiting gonococcal penetration, reducing their likelihood of encountering subepithelial immune cells, explaining why most female infections are localized and asymptomatic (Song et al., 2020). This is possible due to the extraordinary ability of the gonococcus to modulate the interaction with different types of cervical epithelial cells in the female tract, using distinct mechanisms, including the modulation and control of distinct epithelial cell-cell adhesion complexes through the manipulation of host cell signaling and the ability to survive intracellularly within a gonococcal-containing vacuole (Edwards and Butler, 2011; Song et al., 2020).

A notable difference between N. meningitidis and N. gonorrhoeae is that N. gonorrhoeae lacks a capsule because it lacks the corresponding genes for capsule biosynthesis (Virji, 2009). This difference between the two closely related bacteria is crucial because the capsule is essential for N. meningitidis to cause invasive disease. Another important difference is the lack in the gonococcus of the HrpA/HrpB TPS (Talà et al., 2008; Schielke et al., 2010), which, as mentioned above, was implicated in diverse functions including adherence to epithelial cells, intracellular survival, vacuolar escape, interaction with dynein, and modulation of apoptosis/pyroptosis (Schmitt et al., 2007; Talà et al., 2008, Talà et al., 2022; van Ulsen et al., 2008; Neil and Apicella, 2009; Arenas et al., 2013), and was required for the establishment of IMD in a mouse model of meningitis (Pagliuca et al., 2024). The lack of HrpA/HrpB TPS in the gonococcus would explain why the gonococcus remains predominantly confined within a vacuole during cellular infection, unlike the meningococcus, which is able to escape, and why the meningococcus, after reaching the cytosol, induces pyroptotic pathways during cellular infection, a phenomenon not described for the gonococcus. The presence of the capsule and the HrpA/HrpB TPS represents distinctive characteristics of the meningococcus compared to the gonococcus and partly explains the different pathogenic behaviors of the two bacteria.

Two cases of human infection caused by Neisseria brasiliensis, a new Neisseria species phylogenetically distant from N. meningitidis, have been recently described, one of which involved bacteriemia (Mustapha et al., 2020). The first patient was a 64-year-old man from Rio Grande do Sul, Brazil, who developed congestive heart failure with bilateral pulmonary infiltrates and pleural effusion on chest X-ray in June 2016. The second patient was a 74-year-old woman with leprosy from Paraná, Brazil, who developed a polymicrobial infected ulcer on her left lower extremity in February 2016. The two cases were separated in time and by a distance of more than 400 km and had no known epidemiological link (Mustapha et al., 2020). Nevertheless, N. brasiliensis, a species not generally associated with human disease, was able to cause systemic disease in both cases. Despite the comorbidities of the patients, it is unusual for this bacterium to cause disease, suggesting that it may have acquired some virulence factors.

Interestingly, the two isolates (N.95–16 isolates from patient 1 and No.177–16 from patient 2) tested positive by slide agglutination for N. meningitidis capsular groups (ABCEWXYZ), and real-time PCR identified isolate 1 as N. meningitidis capsular group X and isolate 2 as capsular group B. Genomic analysis of the two isolates confirmed the presence of capsular biosynthetic genes in these two isolates (Mustapha et al., 2020). In particular, isolate N.95–16 contained csxABC genes that shared 98% amino acid identity with the meningococcal serogroup X reference strain α388, while isolate N.177–16 contained cssABC-csb genes that shared 99% amino acid identity with the meningococcal serogroup B reference strain H44/76 (Mustapha et al., 2020). The presence of serogroup B and X capsular biosynthetic genes is of particular interest because these genes are normally associated with IMD, and detection of these genes in N. brasiliensis could provide us with clues about the transfer of the capsular locus even among Neisseria that are phylogenetically and ecologically unrelated to N. meningitidis. In addition, the available genome sequences of N. brasiliensis N.95–16 and No.177-16 (Mustapha et al., 2020) revealed the presence of genes (MRN37431.1; MRN37432.1; QGL24436.1) encoding proteins with considerable protein sequence identities with N. meningitidis HrpA and HrpB.

These findings offer us the opportunity to continuously and directly analyze the evolution of Neisseria pathogenicity, and on the other hand, suggest the need to activate continuous surveillance also for Neisseria not normally associated with human colonization, in order to better define their pathogenic and evolutionary potential.

Despite numerous research studies and resources invested, the question of what determines the transition from pharyngeal colonization to IMD, which has been raised since Weichselbaum’s meningococcus, the main agent of epidemic cerebrospinal meningitis, was identified (Weichselbaum, 1887), remains without a definitive answer. Answering this question has become essential to further improve prevention and treatment strategies for meningitis and meningococcal sepsis. In fact, despite the availability of modern vaccines and effective therapeutic protocols, IMD continues to claim victims worldwide. WHO estimates that 500,000 cases of IMD occur each year, approximately 50,000 of which are fatal (Findlow et al., 2025; Poulikakos et al., 2025; Shah et al., 2025). Understanding the intracellular phase of the N. meningitidis infectious cycle could provide useful information to address the molecular mechanisms underlying the transition from asymptomatic (or oligosymptomatic) nasopharyngeal colonization to IMD. What is increasingly emerging from the literature is that, although it is traditionally considered an extracellular pathogen, the meningococcus engages in complex interactions with host cells in the intracellular microenvironment. These interactions involve signal transduction, membrane trafficking, cytoskeleton, metabolic cross-talk, and control of programmed host cell death. The intracellular phase of meningococcal infection may, in fact, be relevant for IMD pathogenesis. Emerging evidence indicates that this phase can be a key step for both NEB and BBB crossing, enabling bacterial dissemination into the host. The intracellular environment can be used to hide from the complement system and from innate immune mechanisms in the early phase of the infection. Moreover, intracellular meningococci induce a cellular response, such as caspase-3 activation, that participates both in the inflammatory state and in the disruption of cell-to-cell junctions. This indicates that a further role of the intracellular phase may be to favor the paracellular crossing of host barriers by the bacteria. Another important point is the considerable genetic variability within the meningococcal population, which is reflected in marked differences between the various strains, even those belonging to hypervirulent lineages, in their mechanisms of interaction with the host cell and microevolution. For instance, different meningococcal strains differ in their capability to evoke inflammation and exploit different routes to cross the NEB and BBB. These differences could explain, at least in part, some inconsistencies found among various studies, attributable to the different strains used. Caution is therefore advised when drawing general conclusions from studies conducted on a single strain or a single hypervirulent lineage.

Although there is now a fairly extensive literature confirming that the intracellular phase is crucial in meningococcal pathogenesis, there are important limitations that this article seeks to highlight. First, as mentioned above, numerous studies have been conducted with phylogenetically distant meningococcal strains, which exhibit different behaviors even in comparable experimental systems. This is due to the considerable genetic and phenotypic variability of the meningococcus. Furthermore, the results of some studies, even when using similar strains, are difficult to compare due to the very different experimental settings. A more systematic use of control reference strains and a better standardization of experimental settings could help us obtain more consistent results and also to better highlight the differences existing between the various strains and their specificities.

Second, it should be noted that much of what we know about the intracellular lifestyle of N. meningitidis comes from studies conducted on single or co-cultured cell lines that only remotely mimic real in vivo conditions. On the other hand, the use of animal models to study invasive meningococcal disease, in addition to presenting the obvious difficulties in investigating aspects of the extracellular/intracellular phase of meningococcal infection, is hampered by the fact that meningococcus has a very narrow host specificity, essentially restricted to humans. Furthermore, the use of animal models is limited for ethical reasons by increasingly stringent regulations. The use of alternative models to study meningococcal pathogenicity outside of a living organism, including organoids, tissue cultures, and ex vivo organ culture systems, could help us confirm observations made in cell line infection studies using traditional systems.

Third, the experimental settings used to study the intracellular/extracellular phase of meningococcal infection ignore crucial factors such as host genetics, age, environment, and lifestyle, and only very limited information is available on the interaction between N. meningitidis and other microorganisms and viruses in the nasopharynx. This information is limited to some observational studies and a few experimental studies involving co-culture or co-infection experiments. Further information on these aspects is essential to understand the dynamics of meningococcal infection and the main factors leading from asymptomatic colonization to disease onset, as well as the evolution of the pathogenic phenotype in Neisseria, in line with the idea that microbial pathogenicity is a multifactorial and complex trait, as indicated by René Dubos in the mid-1990s, who also challenged the then-dominant notion of bacterial fixity and urged bacteriologists to take note of the plasticity of bacteria in order to be ready to deal with them (Dubos, 1955, Dubos, 1959).