Genomic analysis of Neisseria meningitidis carriage isolates during an outbreak of serogroup C clonal complex 11, Tuscany, Italy

Background In 2015–2016, a cross-sectional carriage survey was performed in Tuscany Region, Italy, during an outbreak of invasive meningococcal disease due to Neisseria meningitidis serogroup C clonal complex 11 (MenC:cc11). This study aims to evaluate the genomic profile of meningococcal carriage isolates collected during the survey. Methods Whole-genome sequencing (WGS) was performed using Illumina MiSeq on 85 cultivated meningococcal carriage isolates received at the Dept. of Infectious Disease, National Institute of Health (Istituto Superiore di Sanità, ISS), as National Reference Laboratory (NRL) for Invasive Meningococcal Disease (IMD). De novo assembled genomes were scanned by the BIGSdb platform to assign: the genotypic profiles, the cgMLST, the vaccine antigen variants and allele types of antimicrobial resistance associated genes, together with denitrification pathway loci. Results Capsule null and non-groupable meningococci accounted for 52.9% and 10.6%, respectively. Among the remaining carriage isolates, serogroup B was the predominant (71.0%). Serogroup C meningococci were culture negative and unavailable for WGS. Overall, 64 genotypic profiles were identified and, based on cgMLST, isolates clustered according to clonal complexes. Eight isolates (9.4%) harbored at least one gene encoding a 4CMenB vaccine antigen. Mutated penA alleles were found in more than 82%. Finally, complete aniA and norB coding sequences were detected among 71.8% of carriage isolates. Conclusions Meningococcal carriage isolates collected during the MenC:cc11 outbreak were characterized by an extensive genetic diversity. The lack of outbreak-related isolates among carriage might be attributable to the high transmissibility with low duration of colonization of MenC:cc11 meningococci.


Methods
Whole-genome sequencing (WGS) was performed using Illumina MiSeq on 85 cultivated meningococcal carriage isolates received at the Dept. of Infectious Disease, National Institute of Health (Istituto Superiore di Sanità, ISS), as National Reference Laboratory (NRL) for Invasive Meningococcal Disease (IMD). De novo assembled genomes were scanned by the BIGSdb platform to assign: the genotypic profiles, the cgMLST, the vaccine antigen variants and allele types of antimicrobial resistance associated genes, together with denitrification pathway loci.

Results
Capsule null and non-groupable meningococci accounted for 52.9% and 10.6%, respectively. Among the remaining carriage isolates, serogroup B was the predominant (71.0%). Serogroup C meningococci were culture negative and unavailable for WGS. Overall, 64 genotypic profiles were identified and, based on cgMLST, isolates clustered according to clonal complexes. Eight isolates (9.4%) harbored at least one gene encoding a 4CMenB vaccine antigen. Mutated penA alleles were found in more than 82%. Finally, complete aniA and norB coding sequences were detected among 71.8% of carriage isolates. PLOS

Introduction
Neisseria meningitidis (also known as meningococcus) can be considered a common commensal bacterium of the human pharynx, which represents its natural reservoir [1]. Pharyngeal carriage prevalence is age related, increasing through adolescence [2]. Occasionally, meningococci can invade the bloodstream and other normally sterile sites, leading to invasive meningococcal disease (IMD), whose most frequent clinical presentations are sepsis and meningitis. In Europe, infants and young children are the most affected, followed by adolescents and young adults [3]. Only a minority of strains, referred to as hyper-invasive, are responsible for the majority of IMD cases worldwide [4]. It is reasonable to assume that such hyper-invasive strains are transmitted through respiratory droplets from asymptomatic carriers more frequently than from patients with IMD [1]. However, the relationship between carriage and development of IMD is not fully understood.
Between 2015-2016, Tuscany Region reported an unexpected increase of IMD cases due to the meningococcal C:P1.5-1,10-8:F3-6:ST-11 (cc11) strain [15][16][17]. As already described [18], several initiatives and studies were implemented in the Region, of which: 1) a reactive vaccination campaign with a single dose of the tetravalent ACYW conjugate vaccine or monovalent MCC vaccine, targeting the population aged between 11 and 45 years and, from December 2016, discos and lesbian, gay, bisexual, and transgender (LGBT) associations; 2) a cross-sectional carriage survey with the aim to characterize meningococcal carriage prevalence and related risk factors in the outbreak context [19].
As already published [19], 110 carriage samples resulted positives for N. meningitidis and 85 of them were cultivated in vitro. Hereby, all the 85 culture positive carriage meningococci were further characterized to evaluate the genomic profile.

Ethics considerations
Ethical approval was obtained by the Regional Ethic Committee of Tuscany (registration no. 965) [19]. A written informed consent was signed by all the participants and by the parents/ legal guardians of subjects aged less than 18 years.

Whole-genome sequencing and typing
Eighty-five cultivated carriage isolates [19] were sent to the Italian National Reference Laboratory (NRL) for Invasive Meningococcal Diseases (IMD) at the National Institute of Health (Istituto Superiore di Sanità, ISS) for genomic investigation.
Whole-genome sequencing (WGS) was performed as previously described [15]. De novo assembled genomes have been uploaded onto BIGSdb platform, hosted at PubMLST.org (http://pubmlst.org/neisseria/), and analyzed through the gene-by-gene annotation approach [20]. In case of new alleles or incomplete loci, single gene sequencing and manual curation were performed. According to designation tools included in the Neisseria pubMLST website, isolates were characterized by capsular genogrouping, finetyping of the outer membrane proteins PorA and FetA and multilocus sequence typing (MLST). Such information defines the genotypic profile as follows: genogroup: PorA (P1). VR1,VR2: FetA (F)VR: ST (cc). Phylogenetic analysis was performed by core genome MLST (cgMLST) v1.0 [21] on 26 carriage isolates of this study plus 37 invasive meningococci, sharing the same clonal complex, isolated in 2016 in Italy (Table A in S1 Appendix). Incomplete loci were automatically removed from the distance matrix, and the results were visualised as neighbour-net networks, generated by Split-sTree4 (version 4.13.1) [22]. Genes encoding MenB vaccines antigens (fHbp, NHBA, NadA and PorA VR2) were profiled and the Bexsero antigen sequence types (BASTs) were assigned. Moreover, penA gene, which encodes penicillin-binding protein 2, and genes involved in denitrification pathway, aniA (reference number NEIS1549) and norB (reference number NEIS1548), were typed.
As reported in Table 1 A comparison between Tuscany's carriage (n = 26) and invasive isolates collected in Italy in the same time period (n = 37), belonging to cc23, cc32, cc41/44, cc162, cc213 and cc865 (Table A in S1 Appendix), was performed by cgMLST. Meningococci belonging to the same clonal complex grouped together (Fig 1), and a high similarity between carriage and invasive isolates was found in the main cc23 subcluster (average distance among components of 77 loci).
Isolates were further analyzed for the MenB vaccine antigens and for the Bexsero antigen sequence types (Fig 2 and Table B in S1 Appendix).
An entire coding sequence for NadA was found in 3 isolates: MenB:cc32, encoding the subvariant NadA-1.1; MenZ:UNK and NG:cc175, encoding the subvariant NadA-2/3.8 ( Fig 2C  and Table B in S1 Appendix), included in the 4CMenB formulation. Two MenB:cc213 isolates (2.3%; 2/85) were negative for NadA subvariant due to the presence of a frameshift mutation in the nadA sequence that would result in phase-off gene expression, while the remaining  (Fig 2D). The PorA VR2 P1.4, component of the 4CMenB vaccine, was not detected.

Discussion
Since 2012, a significant increase in the proportion of MenC cases has been observed in Italy, making it one of the most frequent serogroups causing IMD in the country [23]. Due to a MenC:cc11 outbreak occurred in Tuscany Region [15][16][17] a cross-sectional carriage survey was conducted in order to evaluate how the N. meningitidis carriage played a role in the strain spread during the outbreak [18,19].
Here, the genomic analysis was carried out on cultivated carriage isolates to analyze their main molecular traits. Unfortunately, the unique 4 MenC:cc11 carriers, previously described by Miglietta et al. [19], resulted positive only by molecular methods and the isolates were not available for WGS analysis.
Overall, the majority of carriage isolates subjected to WGS were cnl (53%). Moreover, in agreement with previous carriage studies [24,25], MenB was the most frequently identified serogroup among groupable isolates (71%), followed by MenY (23%). Of note, 2 MenB and 2 MenY identified by Miglietta et al. [19], resulted non-groupable by WGS due to the lack of cps biosynthetic genes. Despite the high heterogeneity, the carriage isolates clustered by capsular serogroup and clonal complex. The three prevalent cc s , the cc1136, the cc198 and the cc53, were already associated with cnl meningococcal carriage [26]. Interestingly, strains belonging to cc41/44, cc32 and cc23, three of the main cc s associated with IMD in Italy and in Europe [27][28][29], were also identified among the carriage isolates collected during the Tuscan epidemic.
A high variability with respect to the MenB vaccine antigen-encoding genes was also found, as suggested by the BASTs results. Only eight carriage isolates presented at least one 4CMenB (Bexsero) matching gene and two isolates showed the gene encoding the A subvariant included in the bivalent MenB vaccine (Trumenba). Due to the lack of MATS-ELISA assay [36], the results do not permit to estimate the vaccine coverage among the examined meningococci. However, recently, a new genetic Meningococcal Antigen Typing System (gMATS) was set up to correlate antigen genotypes and coverage estimates by MATS [37]. Some gMATS predictors of coverage identified by Muzzi et al. [37] were also found in this work: fHbp peptides 4 and 510; NHBA peptides 2, 10, 20 and 21. Consequently, 29% of the carriage isolates were estimated to be covered by 4CMenB (Bexsero) vaccine.
Of note, more than 82% of the Tuscany's carriage isolates harbored mutated penA alleles encoding a modified penicillin binding protein 2 that is involved in reducing susceptibility to penicillin G [38]. In the last two decades, this phenotype has been increasingly reported in several countries, including Italy, where it represents the majority of meningococci causing invasive disease [14,38,39].
The ability of meningococci to adapt to oxygen deficiency, even though not essential for meningococcal survival in the pharynx [40], may facilitate new niche of adaptation [41] through the expression of the aniA and norB genes [10,40,41]. More than 71% of the carriage isolates showed an intact coding sequence for both of the genes, suggesting that the transmission chain may include low-oxic environments, as the urethra [41]. However, the results need to be further investigated in order to define a precise role of the denitrification pathway in the survival of carriage isolates associated to specific genomic characteristics.
To conclude, this is the first genomic analysis of meningococcal carriage isolates collected during an outbreak in Italy. To this regard, it should be underlined that WGS offers a greater degree of accuracy to define the main genomic traits of the strains, including capsular group determination, compared to phenotypic and/or polymerase chain reaction assay, which frequently misclassify carriage isolates [42,43].
Overall, this study provided evidence of an extensive diversity among meningococcal carriage isolates during the MenC:cc11 outbreak in Tuscany. Even though an outbreak is a multifactorial event resulting from changes in host-pathogen interactions, the results suggest a quite low recovery degree of the MenC:cc11 in the pharynx, as already described in previous carriage surveys [19,44], and are consistent with a high transmission rate of MenC:cc11strain [45].
Supporting information S1 Appendix. This appendix contains Tables A and B.  (DOC)