The Type a and Type b Polysaccharide Capsules Predominate in an International Collection of Invasive Kingella kingae Isolates

Kingella kingae has emerged as a significant cause of septic arthritis, osteomyelitis, and bacteremia in young children. A recent study examining a diverse collection of K. kingae isolates from Israel revealed four different polysaccharide capsule types in this species, designated types a to d. To determine the global distribution of K. kingae capsule types, we assembled and capsule typed an international collection of K. kingae isolates. The findings reported here show that the type a and type b capsules represent >95% of the invasive isolates, similar to the Israeli isolate collection, suggesting that a polysaccharide-based vaccine targeting these two capsules could be an attractive approach to prevent K. kingae disease.

T he use of improved culture techniques and PCR-based diagnostics in recent years has revealed that the Gram-negative bacterium Kingella kingae is a significant etiology of osteoarticular infections in children 6 to 48 months of age in countries where these sensitive detection methods are routinely employed (1). K. kingae is a normal component of the upper respiratory tract flora in young children and is present in the posterior pharynx in approximately 10% of healthy children 6 to 48 months of age at any given point in time (2)(3)(4)(5). This organism is readily transmitted from person to person by close contact among young children, and longitudinal studies have estimated that children have an approximately 70% chance of being colonized with K. kingae during the first 2 years of life (3,6).
In most individuals, colonization with K. kingae persists for weeks to months and is then cleared (3,6). On occasion, the organism is able to breach the respiratory epithelial barrier, enter the bloodstream, and disseminate to distant sites, causing invasive disease. Analysis of the K. kingae population structure suggests that only some K. kingae strains are able to cause invasive disease (7). The primary clinical presentations of K. kingae disease include septic arthritis, osteomyelitis, spondylodiscitis, tenosynovitis, bacteremia without a focus, and endocarditis (8). The annual incidence of cultureproven disease among children younger than 5 years of age in Israel is 9.4 per 100,000. However, given the difficulty in cultivating K. kingae, this figure represents a minimal estimate (9). Recognizing that the use of species-specific nucleic acid amplification improves the detection of K. kingae by 500% compared to its detection by culture, the true incidence of K. kingae invasive disease in the Israeli population is likely comparable to the incidence of invasive Haemophilus influenzae type b disease before the introduction of H. influenzae type b conjugate vaccines (Hib conjugate vaccines) (i.e., Ͼ50 per 100,000) (10). The incidence of K. kingae invasive disease in other countries has not been defined but appears to be high in parts of Europe and a number of other countries around the world.
Recent studies have shown that isolates of K. kingae elaborate a polysaccharide capsule (11)(12)(13). Interestingly, elimination of encapsulation results in attenuated virulence in an infant rat model of invasive disease (14), suggesting that the capsule is an important virulence factor. Although the specific mechanism by which the capsule facilitates invasive disease has not been defined, the polysaccharide capsules of other pathogenic organisms prevent phagocytosis and block complement-mediated serum killing, promoting bacterial survival in the host. Given this information and the widespread success of polysaccharide conjugate vaccines in reducing morbidity and mortality due to a variety of encapsulated pathogenic bacteria, we recently defined the polysaccharide capsule repertoire in a diverse set of Ͼ400 Israeli K. kingae isolates (13). We found that four distinct polysaccharide capsule structures (capsule types a, b, c, and d) were present in this collection and that Ͼ95% of invasive disease isolates expressed the type a or type b capsule. Furthermore, we identified the csa, csb, csc, and csd capsule synthesis loci that are necessary for the expression of the type a, type b, type c, and type d capsules, respectively.
To gain a broader view of capsule type diversity in the global K. kingae population, in this study, we assembled an international collection of K. kingae isolates and determined the capsule type of each isolate using a multiplex PCR approach. We found that the same four capsule types identified in the Israeli isolate collection were present in the international collection and that no new capsule types were present. In addition, we established that over 95% of invasive isolates expressed the type a or type b capsule.
In a previous study, we developed a K. kingae capsule type PCR screening approach that relied on the use of five separate reactions: reactions specific to each of the four capsule types and a reaction that used primers flanking the capsule synthesis locus (13). To streamline the K. kingae capsule-typing process, in this work, we developed a single multiplex PCR approach for one-step identification of the capsule type. The reaction mixture contains four sets of primer pairs specific to each of the four capsule synthesis loci, designed to each produce a different size amplicon. As shown by the results in Fig. 1, the multiplex PCR produces an~2,000-bp amplicon for the csa locus (encoding the synthesis gene for the type a capsule), an~1,500-bp amplicon for the csb locus (encoding the synthesis genes for the type b capsule), an~1,000-bp amplicon for the csc locus (encoding the synthesis genes for the type c capsule), and an~500-bp amplicon for the csd locus (encoding the synthesis genes for the type d capsule).
To investigate global capsule type diversity in K. kingae, we assembled an international strain collection consisting of 150 isolates (Table 1). Genomic DNA was recovered from each isolate and was used as the template in the multiplex PCR assay. We hypothesized that any isolate that failed to produce a capsule locus-specific amplicon could potentially contain a novel capsule synthesis locus or could be a nonencapsu-lated strain. In total, 89 isolates were capsule type a (59.3%), 49 isolates were type b (32.7%), 8 isolates were type c (5.3%), 3 isolates were type d (2.0%), and 1 isolate (0.7%) yielded no PCR product. Further examination of the isolate that yielded no PCR product revealed that it was nonencapsulated, as assessed by alcian blue staining of surface extracts and inspection of colony morphology (Fig. 2). Using PCR primers that flank the capsule synthesis locus, a 2.5-kb product was amplified from this isolate. Sequencing of this product revealed a truncated csa locus (data not shown). Accordingly, for the purposes of this study, this isolate was considered capsule type a.
A detailed summary of the K. kingae capsule-typing results by country and syndrome (invasive isolate, carrier isolate, or unknown) at the time of isolation is shown in Table 2. Among the 126 isolates that were known to be recovered from systemic sites in patients with invasive K. kingae disease (osteoarticular infections, bacteremia, or endocarditis), 76 were type a (60.3%), 45 were type b (35.7%), 4 were type c (3.2%), and 1 was type d (0.8%), similar to the breakdown among invasive isolates in Israel (13). Of the 13 carrier isolates from oropharyngeal swabs, 7 were type a (53.8%), 3 were type b (23.1%), 1 was type c (7.7%), and 2 were type d (15.4%). While the small number of carrier isolates precludes any meaningful analysis of capsule type distribution, the relative representation of each type was similar to the distribution in the previously published Israeli carrier isolate collection (49.0% type a, 19.2% type b, 12.1% type c, and 19.7% type d) (13). With regard to the countries with at least 10 isolates, Ͼ80% of isolates were type a or type b, and in all countries except Iceland, the most common capsule type was type a. Interestingly, all 10 of the New Zealand isolates were type a, including the 4 carrier isolates.
The four previously identified K. kingae capsule types that were originally characterized in a diverse collection of Israeli isolates were also identified in the international collection of isolates presented here. No new capsule types were identified in this international collection, confirming the limited capsular repertoire of K. kingae compared to those of other pathogens that reside in the respiratory tract, such as Neisseria meningitidis (13 capsule types) and Streptococcus pneumoniae (over 90 capsule types). The study has the limitation that carrier isolates were poorly represented, reflecting the lack of K. kingae colonization studies using culture-based methods outside Israel. As a consequence, we cannot exclude the possibility that other capsule types or nonencapsulated strains exist in the countries examined in this study, keeping in mind the likelihood that some K. kingae strains are able to colonize the oropharynx but are unable to cause invasive disease. In addition, there is little information on the epidemiology of K. kingae disease in developing countries, which may harbor unidentified capsule types. Another limitation of this study is the lack of complete epidemiological   Given that the capsule-typing multiplex PCR approach is specific for the four known K. kingae capsule types, we anticipated that new capsule types in this international collection of isolates would be associated with the lack of a PCR amplicon. The only isolate that failed to yield an amplicon in the multiplex PCR was strain CAN8. Using PCR primers that flank the capsule synthesis locus, we were able to amplify a 2.5-kb amplicon from this strain. Nucleotide sequencing of this amplicon revealed a truncated csaA gene with a large, 1,085-bp deletion in the 3= region, resulting in a nonencapsulated phenotype. Thus, CAN8 is a type a strain that presumably experienced a deletion event and does not have a novel capsule type.