ABSTRACT
In serogroup C Neisseria meningitidis, the cssA (siaA) gene codes for an UDP-N-acetylglucosamine 2-epimerase that catalyzes the conversion of UDP-N-acetyl-α-d-glucosamine into N-acetyl-d-mannosamine and UDP in the first step in sialic acid biosynthesis. This enzyme is required for the biosynthesis of the (α2→9)-linked polysialic acid capsule and for lipooligosaccharide (LOS) sialylation. In this study, we have used a reference serogroup C meningococcal strain and an isogenic cssA knockout mutant to investigate the pathogenetic role of surface-exposed sialic acids in a model of meningitis based on intracisternal inoculation of BALB/c mice. Results confirmed the key role of surface-exposed sialic acids in meningococcal pathogenesis. The 50% lethal dose (LD50) of the wild-type strain 93/4286 was about four orders of magnitude lower than that of the cssA mutant. Compared to the wild-type strain, the ability of this mutant to replicate in brain and spread systemically was severely impaired. Evaluation of brain damage evidenced a significant reduction in cerebral hemorrhages in mice infected with the mutant in comparison with the levels in those challenged with the wild-type strain. Histological analysis showed the typical features of bacterial meningitis, including inflammatory cells in the subarachnoid, perivascular, and ventricular spaces especially in animals infected with the wild type. Noticeably, 80% of mice infected with the wild-type strain presented with massive bacterial localization and accompanying inflammatory infiltrate in the corpus callosum, indicating high tropism of meningococci exposing sialic acids toward this brain structure and a specific involvement of the corpus callosum in the mouse model of meningococcal meningitis.
INTRODUCTION
Neisseria meningitidis is a leading cause of sepsis and meningitis worldwide in humans. Invasive disease is preceded by asymptomatic nasopharyngeal colonization occurring in up to 18% of the normal population. In some individuals this common transitory colonizer is able to breach the mucosal barrier, get into bloodstream and multiply uncontrollably, and finally cross the blood-brain barrier (BBB) to cause meningitis. Both host and bacterial factors seem to be involved in this switch from harmless transitory colonization to devastating disease (1).
N. meningitidis infects only humans because of the high specificity of both meningococcal surface structures and iron uptake systems for human receptors and transport proteins (2–4). The lack of valuable animal models of disease due to the narrow host range, along with the meningococcal high degree of genetic (phase and antigenic) variation of surface structures, has greatly hindered progress in understanding the pathogenesis of meningococcal disease and developing effective vaccines. Much of our knowledge about cellular and molecular biology of this human pathogen and its virulence determinants, including capsular polysaccharide, lipooligosaccharide (LOS), and a number of surface-adhesive and secreted proteins, comes from cell and organ culture systems or animal models that, however, fail to reproduce the complexity of the infectious cycle in the human host (5).
Among the virulence factors described so far, surface-exposed sialic acids occupy a prominent position (6, 7). Thirteen N. meningitidis serogroups have been described on the basis of serologic differences of the capsular polysaccharides; of these, five (A, B, C, Y, and W-135) cause the majority of invasive disease. According to recent WHO data, meningococcal serogroup C is still one of the most widespread serogroups in the world (www.who.int/emergencies/diseases/meningitis/serogroup-distribution-2018.pdf). Recently, a new meningococcal meningitis clone of serogroup C has been expanding in sub-Saharan Africa, associated with a huge risk of a major epidemic in the next 2 years (WHO, 2018 [https://www.who.int/emergencies/diseases/meningitis/meningitis-c-epidemic-risk/en/]). In addition, since January 2015, in Tuscany, Italy, there has been an unexpected increase in cases of invasive meningococcal disease (a total of 43 cases, of which 10 were fatal) due to infection with serogroup C N. meningitidis. Thirty-five out of the samples analyzed in this study were confirmed as C:P1.5-1,10-8:F3-6:ST-11 (clonal complex [CC] 11) (8).
Among the serogroups responsible for epidemics, four (B, C, Y and W-135) carry sialic acids in their capsular polysaccharides (5). Sialic acid is also found as a modification (in place of the terminal galactose residue) of the meningococcal LOS in serogroups with a sialic acid-containing capsule (9). The large abundance of surface-exposed sialic acids is associated with virulence and serum resistance to both phagocytosis and complement-mediated killing via alternative pathway activation (10–16), resulting in enhanced survival in the bloodstream and central nervous system (CNS) (17). There is also evidence that the meningococcal polysialic acid capsule is important for bacterial survival within human cells (18), that it mediates the interaction of bacteria with host cell microtubules during cell infection (19), and that it protects the bacteria against cationic antimicrobial peptides (CAMP), including human cathelicidin LL-37 (18, 20). On the other hand, expression of the polysialic acid capsule hinders colonization and invasion of the nasopharyngeal barrier by masking adhesins/invasins (21–23). For these reasons, capsular polysaccharide expression is subjected to frequent phase variation via slipped-strand mispairing affecting cssD (siaD) or cssA (siaA) (21, 24) or reversible insertion of IS1301 mobile elements in cssA (22) and is tightly regulated at the transcriptional level (25). Loss or downregulation of polysialic acid capsule expression facilitates meningococcal attachment (18, 22, 23, 26) and biofilm formation and correlates with the nasopharyngeal carriage in humans (6).
In the past, our research group has developed a model of meningococcal meningitis (MM) based on intracisternal (i.cist.) infection of adult mice with mouse-passaged bacteria (27). Survival and clinical parameters of infected mice and microbiological and histological analyses of the brain demonstrated the establishment of meningitis with features comparable to those of the disease in humans. Meningococci were also found in the blood, spleen, and liver of infected mice, and bacterial loads in different organs were dependent on the infectious dose. The model was used to assess the virulence of a mutant strain deficient in the l-glutamate transporter GltT (27). The aim of the present study was to evaluate the role of surface-exposed sialic acids in the establishment of meningitis and meningoencephalitis in mice when the bacteria are directly injected i.cist using the MM model in mouse. To this purpose, we have used the reference serogroup C meningococcal strain 93/4286 and an isogenic cssA knockout mutant defective in UDP-N-acetylglucosamine 2-epimerase that catalyzes the first step of sialic acid (N-acetylneuraminic acid) biosynthesis, i.e., the conversion of UDP-N-acetyl-α-d-glucosamine into N-acetyl-d-mannosamine and UDP (28). The 50% lethal dose (LD50) of these strains were determined as well as their abilities to replicate in the brain and other organs. To investigate the infectious dynamics and histopathological correlates of the disease in the MM mouse model, histological evaluation, cerebral bleeding analysis, and localization of bacteria in brain structures were carried out.
RESULTS
Construction of a serogroup C cssA-defective isogenic mutant.The isogenic mutant 93/4286ΩcssA of the serogroup C reference strain 93/4286 was obtained by insertional inactivation of the cssA gene (locus tag NMC0054 or NMC_RS00310), coding for the UDP-N-acetylglucosamine 2-epimerase (EC 3.2.1.183). In the genome of serogroup C strains, cssA is the first gene of the region A capsule synthesis locus, which comprises the conserved cssABC (formally denominated siaABC) genes for CMP-N-acetylneuraminic acid biosynthesis followed by the serogroup C-specific loci, csc (siaDC) coding for α2→9 polysialyltransferase and cssE (oatC) encoding the O-acetyltransferase (29) (Fig. 1A). Southern blot analysis confirmed disruption of the ccsA gene. By using a cssA-specific probe, two NdeI DNA fragments of the expected sizes (3,180 bp and 2,183 bp) were detected in the 93/4286ΩcssA strain, whereas only a single 791-bp NdeI fragment was observed in the wild-type strain 93/4286 (Fig. 1B). The absence of capsular polysaccharide expression in the mutant was confirmed by a latex slide agglutination test using antibodies against serogroup C meningococci capsular polysaccharide (see Fig. S1 in the supplemental material).
Knockout of the cssA gene in the 93/4286 strain. (A) Experimental design for cssA disruption by single crossover. The genetic map of the cps locus region A (capsule synthesis) of N. meningitidis serogroup C was constructed on the basis of the available nucleotide sequences of FAM18 (ET-37) in the NCBI data bank (accession number NC_008767) with arrows depicting gene orientations. The genetic determinants of plasmid pDEΔcssA are indicated: a DNA fragment containing a DNA uptake sequence (DUS), required for efficient DNA uptake during transformation; ermC, the erythromycin resistance gene used as a selective marker for transformation; ΔcssA, a 644-bp XbaI DNA fragment spanning the central part of cssA; and a plasmid polylinker region (indicated by a black box). The physical map of the pDEΔcssA plasmid is also indicated. (B) Southern blot analysis demonstrating inactivation of cssA. Chromosomal DNA was extracted from the parental strain 93/4286 (lane 1) and from an isogenic mutant, 93/4286ΩcssA (lane 2), obtained by transformation with pDEΔcssA and selection with erythromycin. Chromosomal DNA was analyzed by Southern blotting using a cssA-specific probe. Bars on the right indicate cssA-specific fragments whose sizes were deduced on the basis of the relative migration pattern of DNA ladders (bars on the left).
Characterization of 93/4286ΩcssA mutant under in vitro conditions.In order to exclude any differences during bacterial replication, the growth rates of wild-type strain 93/4286 and its derivative cssA mutant were preliminarily analyzed in gonococcus (GC) broth at 37°C. The cssA mutant exhibited growth curves comparable to those of the wild-type strain, with a growth rate (μ = 0.97 ± 0.08) comparable to that of the reference strain (μ = 0.85 ± 0.06) without any statistically significant difference (Fig. S2 and Table S1). In GC medium, the in vitro competition index of the cssA-defective strain was also determined. The growth rate of the mutant during the logarithmic phase of growth compared to that of the wild-type strain provided a relative fitness value of the mutant of 108% compared to that of the wild-type strain (10.77 ± 0.85 generations for the mutant and 9.93 ± 0.68 generations for the wild type). In a competition experiment using equal numbers of CFU, the difference in fitness (D0–1.0 OD) was determined as 0.057 ± 0.01, and the cost per generation was −0.058 ± 0.01 for the mutant versus the wild type. Moreover, the cssA mutant had a growth curve and colony morphology similar to those of the wild-type strain even in Dulbecco’s modified Eagle’s medium (DMEM) (Fig. S3 and data not shown). In cell culture medium, the growth rate of the cssA-defective mutant (μ = 0.25 ± 0.004) was comparable to that of wild-type strain (μ = 0.30 ± 0.04) (Table S2).
In addition, the cssA mutant, grown in GC broth, exhibited a slight upregulation in the expression of virulence-associated surface adhesins such as pilin (pilE; 1.64 ± 0.63-fold change) and nonfimbrial adhesins such as opacity protein (opa; 1.66 ± 0.65-fold change), neisserial heparin-binding antigen (nhbA; 2.08 ± 0.38-fold change), neisserial adhesin A (nadA; 2.07 ± 0.36-fold change), adhesin/invasin (hrpA; 1.98 ± 0.38-fold change), and factor H binding protein (fHbp; 2.08 ± 0.23-fold change) (Fig. S4). The difference in the expression levels of all analyzed genes was not, however, statistically significant.
Survival of mice infected with the 93/4286ΩcssA mutant is significantly increased.The virulence of the cssA-defective strain was assessed in the MM model by analyzing animal survival at different bacterial doses. In order to determine the dose lethal for 50% of animals (LD50), three groups of mice were infected by i.cist. injection of 104, 105, and 106 CFU of the wild-type strain 93/4286 or 107, 108, and 109 CFU of the mutant strain 93/4286ΩcssA. Preliminary data showed higher survival rates in animals infected with the cssA-defective mutant than in those infected with the wild-type strain at the same dose (data not shown). In accordance with previous results (25), mouse death, weight loss, and temperature drop generally occurred within the first 72 h after meningococcal inoculation. Results with the wild-type strain 93/4286 indicated that 50% and 16.6% of rodents survived meningococcal challenge with 104 and 105 CFU, respectively, while all mice died at the dose of 106 CFU (Fig. 2A). A significant difference was observed between the three groups (log rank test, P < 0.05). In contrast, at the lowest dose of 107 CFU, there was 83.3% survival in the group infected with the mutant strain 93/4286ΩcssA, while 50% survival was recorded in mice inoculated with 108 CFU (Fig. 2B), indicating a 10,000-fold-increased LD50 of the cssA-defective mutant.
Survival of mice infected with a wild-type or cssA-defective N. meningitidis strain. Three groups of BALB/c mice (n = 6/dose) were infected i.cist. with 104, 105, and 106 CFU per mouse of the wild-type strain 93/4286 (A) and with 107, 108, and 109 CFU per mouse of the cssA-defective mutant (B). Mice were monitored for a week, and survival was recorded. Results are expressed as percent survival at different doses over time; the log rank P value was <0.05 for mice infected with the wild-type strain.
The 93/4286ΩcssA mutant is severely impaired at replicating in the mouse brain.To determine the number of meningococci in the brain at different stages of disease, animals were injected i.cist. with 5 × 105 CFU of the 93/4286 or 93/4286ΩcssA strain and sacrificed at different time points after challenge (Fig. 3A). A rapid increase in CFU counts was observed for wild-type bacteria that reached the highest numbers 24 h after inoculation (8.519 ± 0.072 log CFU). In contrast, bacterial loads in the brain of mice challenged with the cssA-defective mutant progressively dropped over time, reaching 2.026 ± 1.774 log CFU at 72 h postinfection (Fig. 3A). Bacterial clearance from the infection site occurred in 33.3% of subjects challenged with the mutant, whereas infection was never eradicated from the brain of mice that had received the wild-type strain.
Bacterial loads over time in mice inoculated with the 93/4286 or 93/4286ΩcssA strain. (A) Time course of bacterial loads in the brain following i.cist. infection. Two groups of BALB/c mice (n = 20/group) were infected by the i.cist. route with 5 × 105 CFU of either the wild-type strain 93/4286 or the cssA-defective mutant. Animals were sacrificed at 4, 24, 48, and 72 h after infection (3/time point). Brains were collected and homogenized in GC medium, and viable counts were determined. Results are expressed as means ± SD log CFU numbers per organ at different time points after inoculation. Asterisks indicate statistical significance (**, P < 0.01). (B) Bacterial loads over time in spleen and liver. Two groups of BALB/c mice (n = 5/group) were infected i.cist. with 2 × 106 CFU of either the wild-type strain 93/4286 or the cssA-defective mutant. Animals were sacrificed 48 h after infection. Spleens and livers were collected and homogenized, and viable counts were determined. Results are expressed as log CFU numbers per organ. Horizontal bars indicate mean values of bacterial titers. Each symbol represents a single animal. Asterisks indicate statistical significance (***, P < 0.001).
The cssA-defective mutant is cleared systemically from mice with MM.To evaluate clearance of bacteria from infected mice, two groups of animals were inoculated with 5 × 105 CFU of either the 93/4286 or 93/4286ΩcssA strain, and bacterial viable counts in the spleen and liver were determined (Fig. 3B). Systemic meningococcal infection caused by the cssA-defective mutant was entirely cleared within 48 h from i.cist. challenge, whereas none of the animals inoculated with the wild type had eliminated bacteria from spleen and liver. Two days after inoculation, mean CFU counts of the wild-type strain in the spleen and liver were still 3.212 ± 3.354 log CFU and 6.949 ± 1.37 log CFU, respectively. Differences in bacterial loads in the liver between the two animal groups were statistically significant (P < 0.001).
Serogroup C wild-type meningococci induced severe MM in mice with preferential localization in the corpus callosum.To compare the disease induced by the wild-type 93/4286 and cssA-defective mutant, histological analysis and bacterial immunostaining were performed on brain slices from infected mice 48 h after infection.
MM was considerably more severe in animals infected with the wild type (Fig. 4A) than in those challenged with the cssA-defective strain (Fig. 4B). Histological analysis showed the typical features of bacterial meningitis, including the presence of inflammatory cells in the subarachnoid (Fig. 4C, black arrowheads) and perivascular and ventricular spaces (Fig. 4D, white arrowheads). Vasculitis (Fig. 4C, white arrowheads) and hemorrhages (Fig. 4A, black arrows) were observed mainly in animals infected with the wild-type strain. Interestingly, inflammatory infiltrates were detected in the corpus callosum (Fig. 4E). Indeed, 80% of mice infected with the wild type presented with severe inflammation in the corpus callosum (Fig. 4E, white arrowheads). In contrast, except for one mouse, no massive evident inflammatory infiltrates, but only few immune cells, could be observed in the corpus callosum of animals infected with the mutant strain (Fig. 4F, white arrowheads). The presence and localization of bacteria were further investigated by immunofluorescence. In animals infected with the wild-type 93/4286 strain, immunoreactivity with a meningococcal antiserum was mostly detected in the corpus callosum (Fig. 5A and B), in association with neutrophils in the ventricles (Fig. 5C), or on the meninges (data not shown). A positive signal associated with the cells lining the ventricle or possibly cells from the choroid plexus was also detected (data not shown). In contrast, immunostaining of meningococci revealed no signal in the corpus callosum of animals infected with the cssA-defective mutant (Fig. 5D and E). A weak immunoreaction was detected in association with cells in the ventricles (Fig. 5F) or on the meninges (data not shown).
Cresyl violet stained sections of brains from animals infected with the wild-type or cssA-defective N. meningitidis strain. BALB/c mice were challenged by the i.cist. route with either the wild-type 93/4286 or the mutant 93/4286ΩcssA strain. At 48 h, brains were harvested and treated for histological analysis. Coronal sections (45 μm) were stained with cresyl violet. Images provide an overview (mosaic reconstruction from individual pictures at a magnification of ×10) of the hippocampal region of a representative animal infected with 93/4286 (A) or 93/4286ΩcssA (B). (C) Overview (20× objective) of the meninges (black arrowheads) and an inflamed penetrating vessel (white arrowheads) in an animal infected with 93/4286. (D) Close-up view (40× objective) of the ventricular space of the animal infected with 93/4286 showing inflammatory cells (white arrowheads) and possible intraventricular hemorrhage (black star). Close-up views (40× objective) are shown of the corpus callosum of animals infected with 93/4286 (E) or 93/4286ΩcssA (F) with the presence of infiltrated inflammatory cells (white arrowheads).
Immunofluorescence analysis of brain sections of mice infected with the wild-type 93/4286 or mutant 93/4286ΩcssA strain. Mice were infected and sacrificed as described in the legend of Fig. 4. Brain sections (10 μm) were treated with a rabbit meningococcal antiserum and then a goat anti-rabbit Cy3 serum. Slides were counterstained with DAPI and observed using a Zeiss fluorescence microscope. An overview of the hippocampal region of two representative animals infected with either the wild-type 93/4286 (A to C) or the mutant 93/4286ΩcssA (D to F) strains is shown. In the insets (20× magnification), corpus callosum (B and E) and ventricles (C and F) from the brain of mice challenged with 93/4286 or 93/4286ΩcssA are shown. Large quantities of bacteria are detected in samples from mice infected with the wild-type strain. Red, N. meningitidis bacteria immunostained with meningococcal antiserum; blue, DAPI.
Mice infected with the cssA-defective mutant showed reduced intracerebral hemorrhages.In a previous study, cerebral bleeding was identified as a consistent readout in the brain of mice with MM (30). To perform a quantitative analysis of brain bleeding, the number and area of cerebral bleedings were determined in mice infected by the wild type or the cssA-defective strain. In accordance with histological data, results showed a significant reduction in macroscopical assessment of cerebral hemorrhages (Fig. 6A), in the number of bleeding spots (Fig. 6B; P = 0.01), and in the hemorrhagic area (Fig. 6C; P = 0.048) in mice challenged with cssA-defective bacteria.
Cerebral bleeding in mice infected with the 93/4286 or 93/4286ΩcssA strain. BALB/c mice were infected with either the wild-type strain 93/4286 (n = 8) or the mutant strain 93/4286ΩcssA (n = 8) and sacrificed at 48 h. Brains were collected and immediately frozen in dry ice. Hemispheres were cut in 30-μm cryosections and photographed to determine the number of hemorrhagic spots and the areas of bleeding. (A) Macroscopical assessment of cerebral hemorrhages in animals challenged with the wild-type or the mutant strain. (B and C) Enumeration of bleeding spots and measurement of hemorrhagic areas were carried out on five comparable brain sections/mouse. Data are represented as means ± SD. Differences were assessed by a Mann-Whitney U test (*, P < 0.05).
DISCUSSION
Sialic acids are a family of nine-carbon carboxylated sugars, which include more than 50 different members classified based on various substituents on carbons 4, 5, 7, 8, and 9. The substituent on carbon 5 defines the four most common types of sialic acids: neuraminic acid (Neu), N-acetylneuraminic acid (Neu5Ac), N-glycolylneuraminic acid (Neu5Gc), and 2-keto-3-deoxy-nonulosonic acid (Kdn). They can be found as terminal sugars of glycoconjugates such as glycoproteins and glycolipids on cell surfaces of vertebrates and higher invertebrates (31, 32). By modulating contact-dependent mechanisms, sialic acids and their metabolism play key roles in many physiological and pathological processes, including nervous system embryogenesis, regulation of immune system, cancer metastasis, and bacterial and viral infection (31, 33–37). Sialic acids are also important constituents of LOS and capsular polysaccharides of some bacterial pathogens (38, 39); moreover, other bacterial pathogens not producing sialic acid, like some viruses (40), are equipped with sialidases or neuraminidases, which have been shown to be key virulence factors (41–44).
The crucial importance of sialic acids in the host-pathogen interplay is also well exemplified by our knowledge of the function and genetic regulation of polysialic acid capsule and LOS decoration with terminal sialic acid (Neu5Ac) in N. meningitidis. Most of the attention has been focused on the role of surface-exposed sialic acids in mediating resistance to both phagocytosis and complement-mediated killing via alternative pathway activation (10–16), resulting in enhanced meningococcal survival in the intracellular environment (18–20), in the bloodstream, and in the CNS (17). In the present study, we first aimed at validating the MM model by using a reference serogroup C strain and its attenuated isogenic cssA knockout mutant unable to produce sialic acids. Then, comparison of the virulence of the two strains was also instrumental to further explore the pathogenesis of MM and subsequent cerebral damage by analyzing possible interactions between meningococcal surface-exposed sialic acids and brain structures.
In this study, the inbred BALB/c mouse strain was used instead of the outbred CD-1 strain that was originally employed to establish the MM model (27). Outbred mice present with larger genetic variability that may be more suitable to uncover universal effects in a more diverse cohort and may provide results more applicable to the human population (45, 46). However, such variability requires larger sample sizes to reach sufficient statistical power and may hamper standardization procedures and targeted studies. In our case, the LD50 of strain 93/4286 (without passage in mice) in inbred BALB/c mice was 104 CFU, while the LD50 in outbred CD-1 animals was approximately 107 CFU of mouse-passaged bacteria (27). In this regard, it is noteworthy that BALB/c mice carry the susceptibility (s) mutation in the solute carrier family 11a member 1-encoding gene (Slc11a1), which truncates the encoded protein (also known as natural resistance-associated macrophage protein 1, Nramp1) and increases susceptibility to infection with Mycobacteria spp. and Salmonella spp. (47–51).
The key role of surface-exposed sialic acids in meningococcal pathogenesis was confirmed in the present experimental MM model. The LD50 of the wild-type strain 93/4286 was about four orders of magnitude lower than that of the 93/4286ΩcssA mutant (Fig. 2). Compared to the wild-type strain, the ability of the mutant to replicate in the brain (Fig. 3A) and spread systemically (Fig. 3B) was severely impaired. Histological analysis and bacterial immunostaining on brain slices confirmed higher disease severity with more pronounced inflammation, vasculitis, and hemorrhages in mice infected with the wild-type strain than in those challenged with the cssA-defective mutant (Fig. 4 and 6). The histopathological finding is reminiscent of cerebral infarction that in humans represents a complication in about 25% of patients suffering from bacterial meningitis and in 9% of MM cases (52). Interestingly, 80% of mice infected with the wild-type strain 93/4286 presented with severe inflammation in the corpus callosum (Fig. 4), and most of the immuno-positive signal was localized in this brain structure by immunofluorescence with a meningococcal antiserum (Fig. 5). As expected, meningococci were also detected on the meninges, in the ventricles, and in the choroid plexus. Massive presence of bacteria in the vessels as well as in the epithelium of the choroid plexus and ventricular system is a very common finding in histopathological examination of patients with MM (53, 54). Indeed, the choroid plexus is considered an important gateway for meningococcal traversal from the bloodstream into the CNS during meningitis in humans (55). It is very likely that the bacteria utilize this highly vascularized site to spread systemically from the CNS in the intracerebral (i.c.) mouse model of MM by using a reverse route. In contrast, the remarkable localization of meningococci in the corpus callosum is unexpected, suggesting a certain tropism of N. meningitidis for this brain structure. From a theoretical point of view, in order to accumulate within the corpus callosum, in the i.c. mouse model of MM the bacteria have to leave the cerebrospinal fluid (CSF) space (since the CSF is generated within the plexus choroideus and flows toward the subarachnoid space), survive in the bloodstream, and reenter the brain. Our data seem to suggest the corpus callosum as a major site of bacterial reentry in the i.c. mouse model. This could be due to a high concentration of adhesion molecules relevant to meningococcal-host cell interactions at the level of the cerebral vessels or other structures in the corpus callosum.
Of note, there is evidence in a murine model that heparan sulfate receptors (heparan sulfate proteoglycan [HSPGs]), which are targeted by meningococcal Opa, Opc, and NhhA proteins (5) and mediate the interaction with both epithelial and endothelial cells, are highly expressed in the corpus callosum (56). In addition, it was also reported that the carcinoembryonic antigen-related cell adhesion molecule-1 (CEACAM-1), which serves as a receptor for several meningococcal Opa adhesins/invasins (5), is highly expressed by oligodendrocytes, which are abundant in the corpus callosum (57). Furthermore, the CEACAM-1 pathway activates matrix metalloproteinases that may be involved in blood-brain barrier breakdown (58). Noteworthy, oligodendrocytes specifically express the myelin-associated glycoprotein (MAG), which is a member of the Siglec family of proteins (sialic acid-binding, immunoglobulin-like lectins) capable of binding sialic acid (59). Thus, it is possible, although speculative, that these molecular interactions could recruit the wild-type meningococci and guide their reentry through the corpus callosum. In contrast, the absence of the cssA-defective mutant in this brain structure might be due to both/either its inability to survive in the bloodstream, as demonstrated by its complete clearance at 48 h postinfection in the peripheral organs (Fig. 3B), and/or to the absence of surface-exposed sialic acid.
Although corpus callosum involvement as a complication of MM or invasive meningococcal disease is reported to be a rare occurrence, a case of involvement of the corpus callosum with cerebral ischemia and consequent callosal disconnection syndrome has been recently documented by magnetic resonance imaging and diffusion tensor tractography (60). More recently, a case of a reversible splenial lesion of the corpus callosum associated with MM has also been reported (61). Whether the involvement of this brain structure in meningococcal meningitis/meningoencephalitis, as revealed by advanced imaging technologies, may have actually been underestimated in the past is not clear yet. In fact, the histological evidence of the localization of meningococci in the corpus callosum of patients who died of meningococcal disease does not yet exist. This limits our findings to the analyzed meningococcal serogroup and strain and to the i.c. mouse model of MM used in this study.
The results of our study are consistent with the data reported by Vogel et al. (7) with a bacteremia model in infant rats infected with serogroup B N. meningitidis strain B1940 and a set of isogenic mutants defective in either capsule synthesis or LOS sialylation. Infection of infant rats with the wild-type strain caused severe bacteremia, while an isogenic mutant strain defective in capsule synthesis (but expressing a sialylated LOS) caused bacteremia only when a 106- CFU-higher bacterial dose was used. In addition, when infant rats were infected with encapsulated meningococci that were unable to sialylate the LOS, bacteremia could never be induced, even with an infective dose as high as 108 CFU, suggesting that both forms of sialic acid on the bacterial cell surface are indispensable for systemic meningococcal survival in the infant rat model (7). Our study further expands these data to CNS infection dynamics, having in mind, however, all the limitations of an i.c. mouse model that exploits a nonnatural infection route. Histological analysis and bacterial immunostaining indicate surface-exposed sialic acid as a main determinant for meningococcal intracellular growth/survival as reported before (18, 19) and also as a possible mediator in the interaction between meningococci and neuronal cells in the pathogenesis of invasive meningococcal disease. Noteworthy, sialic acid-dependent interactions play a major role in metastatic invasion of the corpus callosum by tumor cells. In humans, the frequencies of polysialic acid-positive cells and polysialyltransferase expression were higher in diffuse and recurrent astrocytoma (associated with the invasion of the corpus callosum) than in astrocytoma with lower spreading potential (62). This study suggests that tumor cells expressing polysialic acid on neural cell adhesion molecules (NCAMs) may interact with adhesive receptors in the corpus callosum, thus allowing tumor cell migration and localization in this brain structure (62). Similar molecular interactions may explain the massive localization of wild-type meningococci in the corpus callosum in our MM model, proposing a new role of microbial surface-exposed sialic acids in the interplay between N. meningitidis and the host in the pathogenesis of meningococcal disease that, however, should be further explored.
MATERIALS AND METHODS
Bacterial strains and growth conditions.The meningococcal strains used in this study are the serogroup C strain 93/4286 and the sialic acid-deficient isogenic mutant 93/4286ΩcssA. The 93/4286 strain belonging to the ET-37 hypervirulent lineage (CC ST-11) was kindly provided by Novartis Vaccine and Diagnostics, Siena, Italy. Meningococci were cultured on gonococcus (GC) agar/broth (Oxoid S.p.A., Milan, Italy) supplemented with 1% (vol/vol) Polyvitox (Oxoid) at 37°C with 5% CO2. When needed, erythromycin (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) was added to a final concentration of 7 μg ml−1. Meningococci were also cultured in Dulbecco’s modified Eagle medium (DMEM) (Microgem, Naples, Italy) with 10% fetal bovine serum, heat inactivated (Microgem), and 2 mM l-glutamine (Microgem). To evaluate the fitness of each strain, at every stage of growth, serial dilutions were plated on GC agar in the presence or absence of erythromycin and incubated at 37°C with 5% CO2 for 24 h. After growth, viable cell counts were determined by the CFU method. The growth rate μ (h−1) of the cssA mutant and the wild-type strain was calculated as described by Hall and coworkers (63), and the number of generations (G) and time (t) per generation were calculated as described by Billington and coworkers (64).
All experiments were performed in triplicate with three independent cultures; the results obtained were analyzed and graphically reported by using GraphPad Prism (version 4) software, and statistical significance was examined by the Student's t test. Pairwise competition experiments were used to estimate the in vitro fitness of the cssA-defective mutant relative to that of the wild-type strain. Equal numbers of CFU of the isogenic mutant and wild-type strains were mixed together (1:1), and the bacteria were allowed to grow together competitively in antibiotic-free GC broth at 37°C. Experiments were conducted as previously described (65).
Inocula for mouse challenge were prepared by cultivating bacteria in GC broth until mid-logarithmic phase. Viable cell counts were determined, and bacteria were frozen at −80°C with 10% glycerol until use. Escherichia coli strain DH5α was used in cloning procedures. This strain was grown in Luria-Bertani (LB) (Oxoid) medium. To allow plasmid selection, LB medium was supplemented with ampicillin (50 μg ml−1) (Sigma-Aldrich/Merck KGaA).
DNA procedures, plasmids, and transformation of N. meningitidis.High-molecular-weight genomic DNA from N. meningitidis strains was prepared as previously reported (66). DNA fragments were isolated by using acrylamide slab gels and recovered by electroelution as described previously (67). Oligonucleotide synthesis and DNA sequencing were performed by Ceinge-Advanced Biotechnologies, Naples, Italy. DNA sequence analysis was carried out by using the GeneJockey Sequence Processor software (Biosoft) and the multiple-sequence alignment tool Clustal W (http://www.ebi.ac.uk/Tools/msa/clustalw2/).
To construct the pDEΔcssA vector, a genomic fragment of cssA (also known as siaA, synA, or neuA before proposal of a unified nomenclature for capsule loci [29]) (644 bp) was amplified from genomic DNA of 93/4286 strain using the primers CssAXbaF (5′-ATTGAACCTCTAGAGGTCATGATTCACGGCGACCG-3′) and CssAXbaR (5′-TGGCGTTCTAGAACATCAATTGAAGGGACACCG-3′). Amplification reaction programs were as follows: 45 s of denaturation at 94°C, 45 s of annealing at 65°C, and 60 s of extension at 72°C for a total of 30 cycles. Reactions were carried out in a MyCycler thermal cycler (Bio-Rad, Laboratories S.r.l., Segrate, Milan, Italy). The amplicon was cloned into the XbaI site of Neisseria-E. coli shuttle plasmid pDEX (66, 68). Plasmid pDEΔcssA was then used to genetically inactivate by single crossover the cssA gene (locus tag NMC0054 or NMC_RS00310) of strain 93/4286 coding for the UDP-N-acetylglucosamine 2-epimerase. Transformation experiments were performed using 0.1 to 1 μg of plasmid DNA as previously described (69). Transformants were selected on GC agar medium supplemented with erythromycin. Gene disruption was demonstrated by Southern blot hybridization using a 644-bp-long cssA-specific 32P-labeled probe. Southern blot hybridizations were carried out according to standard protocols (67). 32P-labeling of DNA fragments was performed by random priming using the Klenow fragment of E. coli DNA polymerase I and [α-32P]dGTP (3,000 Ci mmol−1) (67). To detect capsular polysaccharide expression, a rabbit polyclonal antibody-coated latex suspension against groups C and W-135 N. meningitidis strains was used in a latex slide agglutination test (BD Directigen Meningitis Combo Test; BD Italia, Milan, Italy).
Real-time RT-PCR experiments.Semiquantitative analysis of the pilE, fHbp, hrpA, nadA, opa54, and nhbA transcripts normalized to the level of expression of the 16S rRNA gene (65) was performed by real-time reverse transcriptase PCR (RT-PCR). Wild-type strain 93/4286 and the cssA-defective mutant were grown to late logarithmic phase (optical density at 600 nm [OD600], 1.0) in GC broth. Total bacterial RNAs were then extracted by use of an RNeasy minikit (Qiagen, Venlo, the Netherlands) according to the manufacturer’s instructions. Before extraction, samples were treated with 2 volumes of RNA Protect bacteria reagent (Qiagen). DNA contamination was avoided by on-column treatment with an RNase-free DNase set (Qiagen) according to the manufacturer’s instructions. This procedure was performed in triplicate for each strain. The concentration and integrity of the RNA samples were assessed by measurement of the A260/A280 and A260/A230 ratios and verified using a NanoDrop Lite spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA). Then, for each sample, total RNAs (2 μg) were reverse transcribed into cDNA as previously described (65). For semiquantitative analysis, about 64 to 128 ng of each reverse transcription reaction mixture was used to run a real-time PCR on an Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) with KiCqStart SYBR Green quantitative PCR (qPCR) Ready Mix, Low ROX(Sigma-Aldrich), and with a specific primer pair as reported in Table S3 in the supplemental material.
PCR was conducted according to the manufacturer’s guidelines as follows: initial holding and activation at 95°C for 10 min, followed by 40 cycles at 95°C for 20 s, 60°C for 30 s, and 72°C for 30 s. The post-PCR melt curve was performed between temperatures of 60°C and 95°C with 1% temperature increments. Previously, standard curves were analyzed to determine the efficiency of amplification and the 2−ΔΔCT method (where CT is threshold cycle) was used for the analysis of the level (n-fold) of change. Samples were run in the real-time PCR in triplicate, and statistical significance was examined by a Mann-Whitney U test. To quantify gene expression, we expressed the data as fold change obtained using the 2−ΔΔCT method.
Mice, MM model, and experimental design.Eight-week-old female inbred BALB/c mice weighing 19 to 20 g were purchased from Charles River Italia (Lecco, Italy). The mice were fed with laboratory food pellets and tap water ad libitum and were housed under specific-pathogen-free conditions. All efforts were made to minimize animal suffering and reduce the number of mice in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC). The study was approved by the Ethical Animal Care and Use Committee (protocol number 2, 14 December 2012) and the Italian Ministry of Health (protocol number 0000094-A-03/01/2013). For brain histology and cerebral bleeding analysis, animal experiments were performed at the Azienda Ospedaliera Universitaria Senese and authorized by the Local Ethics Committee (Comitato Etico Locale, Azienda Ospedaliera Universitaria Senese, 21 May 2012) and the Italian Ministry of Health (document no. 131/2013, 30.05.2013).
Mice were infected by the i.cist. route as previously described (27, 30, 70). Bacteria for mouse challenge were prepared as previously reported (27), thawed, centrifuged for 15 min at 1,500 × g, and suspended in GC broth with iron dextran (5 mg/kg; Sigma-Aldrich/Merck KGaA). Approximately 2 h before infection, animals were injected intraperitoneally (i.p.) with iron dextran (250 mg kg−1). Animals were lightly anesthetized (50 mg/kg ketamine and 3 mg/kg xylazine or Zoletil 20 [30 mg/kg; VirbacSrl] and Xilor [8 mg/kg; Bio 98 Srl]), and bacteria (suspended in a total volume of 10 μl) were inoculated by hand puncturing the cisterna magna of mice using a 30-gauge needle (BD Italia, Milan, Italy). Mice were monitored for possible seizures due to inoculation. Clinical signs were monitored according to a previously described coma scale (71), and mice with a score of 2 were euthanized and recorded as dead for statistical analysis.
For brain histology and cerebral bleeding analysis, brains were removed and dissected into the two hemispheres and cerebellum. One hemisphere was fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) (wt/vol) for histological analysis, while the other one was frozen in dry ice for assessment of intracerebral bleeding. Samples were not collected from animals found dead or humanely sacrificed before 48 h.
Animal survival and CFU counts.Different bacterial doses ranging from 104 to 106 CFU per mouse for the 93/4286 wild-type strain and from 107 to 109 for the 93/4286ΩcssA isogenic mutant were used to inoculate animals (n = 6/dose). Control mice were inoculated with GC broth. Every day throughout the whole experiment, animals were monitored for clinical signs, and body weight and temperature were measured as previously described (27). Survival was recorded for a week. To determine the number of meningococci in the brain over time, animals were infected with 5 × 105 CFU/mouse and sacrificed at different time points (4, 24, 48, and 72 h) (n = 3/time point) after infection. To compare the virulence of the wild-type strain versus that of the cssA-defective mutant, two groups of mice (n = 5/group) were infected with 5 × 105 CFU/mouse and sacrificed 48 h after challenge for organ collection. Brain, spleen, and liver were excised and homogenized in 1 ml of GC medium. Viable cell counts were performed by plating 10-fold dilutions onto GC agar plates with (mutant) or without (wild type) erythromycin.
Brain histology.Experiments were performed with a total number of 20 mice, of which 8 were infected with the wild-type strain, 8 were challenged with the cssA-defective mutant, and 4 served as controls. Infection dose was 6 × 105 CFU/mouse. Brain hemispheres from mice infected for 48 h were prepared for cryopreservation by incubation in 18% (wt/vol) sucrose in PBS at 4°C overnight. Hemispheres were mounted in 22-oxacalcitriol (OCT) compound and cut coronally using a Leica 3050S cryostat (Leica Biosystems, Wetzlar, Germany). Coronal sections (45 μm thick) were sampled at a frequency of every 15th slice. Additional 10-μm sections were obtained for immunofluorescence analysis. Histopathological evaluations were made on sections stained with cresyl violet for Nissl substance.
Immunofluorescence.Slices were incubated with a primary rabbit polyclonal antibody against whole-cell preparation of serogroup A, B, and C N. meningitidis (6122; ViroStat, Portland, ME, USA) at a dilution of 1:1,000. This antibody was reactive against both encapsulated and unencaspulated meningococci (18, 19). Sections were washed three times with PBS and incubated with the secondary antibody (goat anti-rabbit Cy3; Jackson, West Grove, PA, USA) for 45 min at room temperature in the dark. Primary and secondary antibodies were diluted in Tris-buffered saline (TBS; Sigma-Aldrich/Merck KGaA) containing 0.5% bovine serum albumin. After a washing step, slides were counterstained with 4′,6′-diamidino-2-phenylindole (DAPI) for 1 min, washed, and mounted with Mowiol (Merck) containing 2.5% diazabicyclooctane (DABCO; Sigma-Aldrich/Merck KGaA). Pictures were obtained using a Zeiss fluorescence microscope (Axio Imager M1; Zeiss, West Germany) equipped with a digital camera (AxioCamHRc). Overview pictures were created by combining photos obtained with a 10× objective and mosaic reconstruction using the AxioVision, version 4.8, software (Zeiss, Oberkochen, Germany).
Analysis of cerebral bleeding.Cerebral hemorrhages were assessed as previously described (72). Briefly, brain hemispheres were cut in a frontal plane into 30-μm-thick sections, and serial sections were photographed with a digital camera at 0.3-mm intervals. For each animal, five comparable brain sections were analyzed. The bleeding spots were counted, and the relative areas of bleeding were measured by using the UTHSCSA Image Tool (Texas, USA). Cumulative bleeding areas were divided by the whole slice area and computed into a total bleeding area/whole slice area ×1,000.
Statistical analysis.Bacterial counts in different organs and time points were represented as means ± standard deviations (SD) of CFU numbers isolated from single mice. Differences in growth rates in in vitro experiments and differences in bacterial loads between mice infected with the wild-type strain or the mutant were examined by Student's t test. Mouse survival was estimated by Kaplan-Meier survival analysis, and differences were compared using a log rank test (P < 0.05). Differences in expression levels of surface adhesins determined by reverse transcriptase real-time PCR and differences in cerebral bleeding were evaluated using a Mann-Whitney U test (P < 0.05).
ACKNOWLEDGMENTS
This research was supported in part by PRIN 2012 (grant number 2012WJSX8K), “Host-microbe interaction models in mucosal infections: development of novel therapeutic strategies.”
We thank Robert Lukesch for excellent technical assistance and Tiziana Braccini for excellent technical assistance in in vivo experiments.
FOOTNOTES
- Received 12 September 2018.
- Returned for modification 4 October 2018.
- Accepted 29 January 2019.
- Accepted manuscript posted online 4 February 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00688-18.
- Copyright © 2019 American Society for Microbiology.
