cgMLST characterisation of invasive Neisseria meningitidis serogroup C and W strains associated with increasing disease incidence in the Republic of Ireland

Introduction and aims Since 2013 MenC and MenW disease incidence and associated mortality rates have increased in the Republic of Ireland. From 2002/2003 to 2012/2013, the average annual MenC incidence was 0.08/100,000, which increased to 0.34/100,000 during 2013/2014 to 2017/18, peaking in 2016/17 (0.72/100,000) with an associated case fatality rate (CFR) of 14.7%. MenW disease incidence has increased each year from 0.02/100,000 in 2013/2014, to 0.29/100,000 in 2017/18, with an associated CFR of 28.6%. We aimed to characterise and relate recent MenC isolates to the previously prevalent MenC:cc11 ET-15 clones, and also characterise and relate recent MenW isolates to the novel ‘Hajj’ clones. Methods Using WGS we characterised invasive (n = 74, 1997/98 to 2016/17) and carried (n = 16, 2016/17) MenC isolates, and invasive (n = 18, 2010/11 to 2016/17) and carried (n = 15, 2016/17) MenW isolates. Genomes were assembled using VelvethOptimiser and stored on the PubMLST Neisseria Bacterial Isolate Genome Sequence Database. Isolates were compared using the cgMLST approach. Results Most MenC and MenW isolates identified were cc11. A single MenC:cc11 sub-lineage contained the majority (68%, n = 19/28) of recent MenC:cc11 disease isolates and all carried MenC:cc11 isolates, which were interspersed and distinct from the historically significant ET-15 clones. MenW:cc11 study isolates clustered among international examples of both the original UK 2009 MenW:cc11, and novel 2013 MenW:cc11clones. Conclusions We have shown that the majority of recent MenC disease incidence was caused by strain types distinct from the MenC:cc11 ET-15 clone of the late 1990s, which still circulate but have caused only sporadic disease in recent years. We have identified that the same aggressive MenW clone now established in several other European countries, is endemic in the RoI and responsible for the recent MenW incidence increases. This data informed the National immunisation Advisory Committee, who are currently deliberating a vaccine policy change to protect teenagers.


Introduction
The Gram-negative diplococcus Neisseria meningitidis is normally a harmless coloniser of the upper respiratory tract. On rare occasions the meningococcus can cross the nasopharyngeal epithelium and cause invasive meningococcal disease (IMD), the onset and progression of which is often rapid. Associated mortality ranges between 10%-40%, and many survivors experience long term serious sequelae, including hearing impairment and limb loss [1].
The capsular polysaccharide defines the serogroup, and is considered the most important virulence determinant [2]. In Europe, most disease incidence is caused by meningococcal strains expressing serogroups B, C, W, or Y, to which vaccines are available [3].
Multi locus enzyme electrophoresis (MLEE), and later multi locus sequence typing (MLST), revealed meningococcal populations to exist as groups of genetically related clones termed electrophoretic types and clonal complexes (cc) respectively [4,5]. Whole genome sequencing (WGS) has been increasingly used over the past 10 years to resolve meningococcal populations into lineages with epidemiological consequence using a subset of meningococcal core genes [6]. Specific genetic lineages of the meningococcus can be associated with either asymptomatic carriage, which is highest in young adults, or with invasive disease, where the highest rates occur in those under 5 years and young adults [7]. Just a few lineages, the hyperinvasive lineages, are consistently isolated from invasive disease cases, and are responsible for the vast majority of IMD globally [4].
Of particular importance is the hyper-invasive cc11 meningococcal lineage (lineage 11) which are associated with high levels of morbidity and mortality, and can express serogroups C, W, B or Y [8][9][10]. The epidemiology of cc11 meningococci ranges from sporadic disease cases and clusters, to regional outbreaks (which can be protracted) and international epidemics [11][12][13].
The first glycoconjugate vaccines were used to combat an international epidemic associated with MenC:cc11 meningococci, which spread through Europe during the mid to late 1990s [13]. Since the early 2000's several small clusters of MenC:cc11 meningococci have been reported in men who have sex with men (MSM), and recently a protracted regional MenC: cc11 outbreak was reported in Italy [14][15][16][17].
A MenW:cc11 strain responsible for an outbreak associated with the Hajj pilgrimage at the turn of the century, became endemic in Europe where it persisted for several years [18]. The outbreak strain also disseminated into Africa, and then later to several South American countries [19,20]. In 2009, this strain began causing disease in the UK (the original 2009 UK strain). This clone is indistinguishable from the original Anglo-French Hajj associated strains by conventional MLST and antigen fine typing methods, and is recognised as MenW:cc11:P1.5,2:F1-1 [5,21]. A novel variant of the original 2009 strain (novel UK 2013 strain) was responsible for an outbreak among scouting enthusiasts following their return from the Japanese 2015 scout jamboree [22,23]. This clone is now endemic in some European countries including the UK, France, Sweden, and the Netherlands [24][25][26][27].
The highly clonal nature of the cc11 lineage makes the recognition of clonal variants challenging. MLST recognises virtually all invasive cc11 strains as sequence type 11 (https:// pubmlst.org/neisseria/), the majority of which are homogeneous for the PorA outer membrane protein (OMP), and type as P1.5,2 using mono clonal antisera. Antigen fine-typing too is unsatisfactory, distinguishing among some MenC:cc11 clones, but lacking the necessary discriminatory power to establish the true phylogenetic relationships among contemporary circulating clones, or to historically significant strain types.
The clone associated with the international MenC epidemic of the 1990s and early 2000s is recognised by MLEE as electrophoretic type 15 (ET-15), and is characterised by a point mutation in the fumerase gene, or by the presence of insertion sequence element IS1301 [28,29].
In response to increasing MenC incidence, the Republic of Ireland (RoI) introduced the meningococcal C conjugate (MCC) vaccine into the routine infant schedule (October 2000), and offered the vaccine on a one-off basis to all those under 23 years. Subsequently, MenC incidence decreased from an incidence rate (IR) = 4.37/100,000 population (164 cases To understand if recent increases in MenC incidence are due to the emergence of a novel clone, or the re-emergence of clones prevalent at the time of the introduction of MenC vaccine, we characterised IMD-associated MenC:cc11 isolated during the 1997/98 to 2016/17 epidemiological years by WGS. We also compared these genomes to international invasive MenC: cc11 genomes. To understand the nature of the currently circulating MenW strains, all IMDassociated MenW isolates collected between 2010/11 and 2016/17 were compared to international examples of the original Hajj clone, and the more recently emerged original 2009 UK, and novel 2013 UK lineages. MenC and MenW strains isolated during a recent national carriage study among university students were also included.

Materials and methods
In the Republic of Ireland (RoI) all invasive cases of meningococcal disease are reported to the Health Protection Surveillance Centre (HPSC) and the corresponding isolates and/or disease associated specimens are sent to the Irish Meningitis and Sepsis Reference Laboratory (IMSRL). N. meningitidis confirmation is by (PCR and real-time PCR) [31], and strain characterisation methods include serogroup [32], PorA and FetA variable region sequencing [33,34], multi-locus sequence typing (MLST) [5], and multi locus restriction typing (MLRT) [35].

Study isolates
From a national collection of previously characterised IMD-associated isolates, 74 invasive MenC strains isolated during the 1997/98 to 2016/17 epidemiological years were included in this study.
All invasive meningococci isolated since 2010/11 have been characterised by WGS and were included in the study (n = 31). Prior to 2010/11, we retrospectively identified MenC:cc11 isolates, and characterised these isolates by whole genome sequencing. Clonal complex was inferred from MLRT characterisation, which recognises cc11 meningococci as restriction type 1 (RT-1). All invasive MenC isolates with RT-1 identified between 2001/02 to 2010/11 were included (n = 23). During the 1997/98 to 2000/2001 epidemiological years, a period of high incidence, 110 MenC-RT-1 isolates were identified, and 20 were included in the study. For these isolates PorA sequence type data was used as a proxy for genotype diversity, and isolates were chosen to broadly represent genotype diversity. Also included were asymptomatically carried MenC strains (n = 16) isolated from university students between November 2016 and March 2017. In all 90 MenC isolates were analysed. The national carriage study was ethically approved by the Children's University Hospital ethics committee, and conducted in a fully anonymised manner where no participant identifiable information was collected.
MenW:cc11 isolates (n = 26) from the RoI were compared to an international collection of MenW:cc11 isolates, chosen to represent examples of the original Hajj strain (n = 170), and the more recent variants, the original UK 2009 and novel 2013 UK strains (n = 890). The isolate distribution by country was as follows: the UK (n = 878), the Netherlands (n = 121), France (n = 5), Sweden (n = 5) and Greece (n = 2).

Genome sequencing
Genomic DNA was sequenced using the Illumina HiSeq platform (Oxford Genomics Centre, Wellcome Trust Centre for Human Genetics, University of Oxford, United Kingdom). Short read sequences were assembled de novo using the VelvetOptimiser algorithm [36], and then added to the Bacterial Isolate Genome Sequence Database (BIGSdb) on the Neisseria PubMLST website (https://pubmlst.org/neisseria/) [37]. The genomes were automatically scanned and annotated using alleles previously defined in the sequence definition database. The number of contigs assembled per genome ranged from 107 to 322, with a median of 142. The corresponding contig total lengths were between 2.11 and 2.24 Mb, with a median of 2.14 mega bases.

Phylogenetic analysis
Genome comparisons were made using a gene-by-gene approach, where the 'Genome Comparator' tool identifies and assigns a numerical designation to each unique allele encountered [37]. The core genome MLST (cgMLST) scheme, which utilises a set of 1605 loci, was chosen as the basis of all genome comparisons [6]. The resultant distance matrix was visualised as a Neighbour-Net network in SplitsTree (version 4.14.4) [38,39]. Phylogenetic clusters were labelled consistently with a large international study of cc11 meningococci, which identified two main branches termed 11.1 and 11.2 [40]. Statistical tests were carried out in R (version 3.5.1, www.r-project.org).

Antimicrobial resistance
For isolated collected since the 2012/13 epidemiological, minimum inhibitory concentrations were determined for penicillin, rifampicin, and ciprofloxacin by the E-Test method, and using Muller-Hinton agar supplemented with 5% sheep's blood [41]. Phenotypes were assigned in accordance with EUCAST guidelines [41].

Incidence of MenC and MenW disease
The predominant serogroups associated with meningococcal disease in the RoI are B, W, C, and Y (Fig 1). Serogroup B disease incidence has declined since the turn of the century. Disease incidence associated with serogroups W and C have increased since 2013. The number of associated deaths by serogroup has also increased (Fig 1).
RoI MenC:cc11 isolates in the context of international MenC:cc11 strains. Irish MenC isolates were compared to a set of 779 contemporarily circulating strains in Europe (S2 Fig). The greater sample size of the international collection reveals the cc11 population structure in greater detail. Despite the disparity in sample size between the two isolate collections, the same basic MenC:cc11 population structure is observed, which indicates that the Irish isolates chosen from this period are representative of diversity circulating in Europe during the studied period and the robustness of the sub-lineage differentiation. The temporal distribution of the European isolates is consistent with the RoI isolate pattern, which suggests that MenC:cc11 strains have spread frequently between these countries. The comparison also shows that the dominant MenC clone circulating at present in the RoI is not unique to the RoI. Vaccine and antimicrobial resistance antigens. Among cc11 MenC and cc11 MenW isolated since 2014/15 epidemiological year (including carriage cc11 isolates) two Bexsero Antigen Sequence Types (BAST) dominated; BAST 8 (PorA 5,2; fHbp 22; NHBA 29; NadA 29) among MenC isolates, and BAST 2 (PorA 5,2; fHbp 22, NHBA 29, NadA 6) among MenW isolates. Antimicrobial resistance genes for ciprofloxacin, penicillin, and rifampicin (gyrA, penA, and rpoB respectively) were homogeneous among these isolates, and were consistent with isolates sensitive phenotypes (S1 Table).

Discussion
In the RoI, disease incidence and associated case fatality rates due to MenC and MenW meningococci have increased in recent years. We have used WGS to investigate the isolates associated with these increases to resolve the uncertainty arising from conventional molecular typing method's limited capacity to discriminate among cc11 sub-lineages. This study comprehensively characterised invasive MenC and MenW isolates by WGS, and compared genomes using cgMLST (n = 1605 loci).
Comparisons showed the invasive and carried MenC and MenW isolates studied were predominantly cc11. The majority of MenC incidence over the last five epidemiological years was due to a clone genetically distinct from the ET-15 MenC:cc11 epidemic strains. These isolates comprised two temporally structured clusters of highly similar clones. The emergence and expansion of this clone coincides with increasing MenC disease incidence observed since 2013. The interspersion of the carriage and disease isolates indicate that disease associated with this clone is largely a consequence of transmission. These isolates lacked the ET-15 fumC mutation, and were highly homologous at important antigenic loci such porA and other 4CMenB antigen targets, but would not be anticipated as susceptible to Bexsero elicited antibodies.
A minority (21%, 6/28) of invasive isolates from recent epidemiological years were found in sub-lineage 11.2 (red and orange). Given their variation, these strains appear more similar at a genomic level to the pre-MenC vaccine strains. So far, sub-lineage 11.2 group-E genotypes have been limited to causing sporadic disease only. While these strains must still circulate, no examples were isolated in the 2016/17 carriage study, indicating that they are circulating at a lower frequency than sub lineage 11.1 group-A strains; or less likely, they colonise the nasopharynx to a density undetectable by the sampling technique employed in the carriage study.
International comparisons of Irish and European MenC:cc11 isolates showed that while recent disease in the RoI is caused by a distinct MenC:cc11 clone, it is not unique to the RoI. This comparison further confirms the capacity of hyper invasive MenC:cc11 strains to disseminate and spread internationally, an inherent characteristic of MenC:cc11 strains.
Taken together, the significant increase in MenC disease incidence from a novel disease genotype, the higher associated case fatality rates, the increase in median incidence age, and the potential for capsule switching events to occur in this lineage, are all reasons to warrant concern, and highlight a changing epidemiological pattern that will require continuous and careful monitoring, if not immediate action [10,40]. The success of meningococcal conjugate C (MCC) vaccination programs is attributed to the vaccines capacity to elicit mucosal epithelial secretions containing anti-MenC antibodies, which can impact the carriage state and slow transmission [45]. Meningococcal carriage is age dependant, where the highest rates of are in teenagers and young adults [46]. Immunisation of this cohort disrupts MenC transmission and protects other non-vaccinated cohorts indirectly (herd immunity) [45,46]. Therefore, maintaining antibody titres among this cohort are essential to sustain the herd effect.
The longevity of MCC vaccination protection is unknown, although UK data had predicted a reduction of MenC disease for >15 years following vaccination [47]. A recent UK study reported that just 19% of 10 year olds immunised at 14 months with a single dose of MenC had protective antibody titres 9 years later [48]. While no similar seroprevalence data is available in the RoI, it is probable that MenC increases are due to waning herd immunity in the Irish population. Further, MCC herd immunity has likely reduced naturally acquired immunity to MenC strains as a consequence of reducing MenC transmission [48,49]. When MCC herd effects fail, transmission is likely to increase, and an immunologically naïve population (compared to a pre vaccine levels), might be at higher risk to MenC disease.
Prior to 2011 in the RoI, no examples of the characteristic Hajj genotype (MenW: cc11:1.5,1.2) have been identified among either disease isolates or associated diagnostic extracts. Therefore, disease incidence associated with the original Hajj clone was either rare or absent before 2011. The data presented here shows that since then, cc11 meningococci were responsible for the majority of invasive MenW disease. These isolates, and asymptomatically carried strains, were identified as the original 2009 UK and novel 2013 UK strain clones, and show that these strain types are now endemic in the RoI. The original Hajj clone however, was not observed in either carriage or disease. Why the original Hajj strain did not become endemic and the current strains have is unclear.
The year by year incidence increases, and the high CFR rates associated with this hyper virulent MenW strain reported here are consistent with reports from other countries [24][25][26]50,51,52]. In addition, reports of atypical clinical presentations have been associated with this strain, such as vomiting, diarrhoea, and necrotising fasciitis and septic arthritis and pericarditis [24,26,53]. We have clearly established that the majority of Irish MenW isolates are indistinguishable from those strains driving MenW disease incidence in the UK, Netherlands, Sweden, Italy and France. Further, we also report isolating this clone from asymptomatic carriage in University students, confirming that this clone is now endemic in the RoI.

Summary and conclusions
Unambiguous characterisation of bacterial isolates is essential to establishing relationships among bacterial isolates nationally and internationally, and relating contemporarily circulating strains to historically significant strain types. This analysis is invaluable to informing public health decisions to adapt vaccination policy to counter changes in epidemiological patterns. We have shown that the highly clonal nature of cc11 meningococcal strains can be comprehensively resolved by WGS characterisation and publicly accessible tools. This study also highlights the importance of regular carriage studies; relating disease associated genotypes to circulating carriage genotypes can help establish the degree to which novel or emerging disease associated genotypes have become endemic.
The national immunisation advisory committee (NIAC) are currently (April 2019) considering a change in vaccine policy in response to changing MenC and MenW epidemiological patterns the RoI. Data from this study clarifying the nature of recent meningococcal epidemiological changes was used to inform this pending decision.