|Home | About | Journals | Submit | Contact Us | Français|
Nonserotypeable pneumococci (NSP) are commonly carried by Australian Indigenous children in remote communities. The purpose of this study was to characterize carriage isolates of NSP from Indigenous children vaccinated with the seven-valent pneumococcal conjugate vaccine (PCV7) and to use these data to guide decisions on reporting of NSP. A total of 182 NSP were characterized by BOX typing, antibiogram analysis, and multilocus sequence typing (MLST) of common BOX types. NSP positive for the wzg capsule gene were analyzed by a multiplex PCR-based reverse line blot hybridization assay (mPCR/RLB-H) targeting capsule genes to determine the serotype. Among 182 NSP, 49 BOX types were identified. MLST of 10 representative isolates found 7 STs, including ST448 (which accounted for 11% of NSP). Non-penicillin susceptibility was evident in 51% of the isolates. Pneumococcal wzg sequences were detected in only 23 (13%) NSP, including 10 that contained an ~1.2-kb insert in the region. mPCR/RLB-H identified serotype 14 wzy sequences in all 10 NSP, and 1 also contained a serotype 3-specific wze sequence. Among the remaining 13 wzg-positive NSP, few belonged to the serotypes represented in PCV7. It appears that most NSP identified in Australian Indigenous children are from a true nonencapsulated lineage. Few NSP represented serotypes in PCV7 that suppress capsular expression. High rates of carriage and penicillin resistance and the occasional presence of capsule genes suggest a role for NSP in the maintenance and survival of capsulated pneumococci. To avoid the inflation of pneumococcal carriage and antibiotic resistance rates, in clinical trials, we recommend separate reporting of rates of capsular strains and NSP and the exclusion of data for NSP from primary analyses.
Australian Indigenous children in remote communities of the Northern Territory experience dense, persistent nasopharyngeal colonization with Streptococcus pneumoniae (pneumococcus) from early infancy (13, 25). Cross-sectional studies report pneumococcal carriage rates of over 80% for these children (15). Pneumococcal serotype diversity contributed to swift serotype replacement following introduction of the seven-valent pneumococcal conjugate vaccine (PCV7) in 2001. The vaccine had no significant impact on the incidence of overall pneumococcal carriage or otitis media (14).
To date, 91 pneumococcal serotypes have been described (21). In addition, a population of nonserotypeable pneumococcal (NSP) isolates does not react with the capsular polysaccharide typing sera. Molecular typing has identified three categories of NSP: (i) pneumococci that lack capsule genes, (ii) pneumococci that possess capsule genes but that are phenotypically nonencapsulated, and (iii) atypical pneumococci that are phenotypically NSP but that are genetically divergent from pneumococci (8,18). Pneumococci that lack capsule genes make up a highly diverse population that includes strains that have spread globally (8). These strains have been associated with a variety of mucosal (and, occasionally, invasive) diseases (for a review, see reference 24). Phenotypically nonencapsulated pneumococci that possess capsule genes may represent pneumococci that have lost the ability to express the capsule or strains that have temporarily ceased or downregulated capsule expression. Understanding the distinction between these populations is particularly important in the context of widespread immunization with anticapsular vaccines.
NSP frequently colonize Indigenous Australian children. In our studies with young Indigenous children, NSP were detected in up to 18% of nasopharyngeal swab specimens. In a recent cross-sectional carriage study, NSP were the 3rd most common pneumococcal serotype (unpublished data from reference 16). It is also likely that we underestimate NSP carriage rates because of their morphological differences from their capsular counterparts; NSP tend to be smaller and dryer than capsular pneumococci, and the dimple is less conspicuous. The purpose of this study was to characterize NSP carriage isolates collected from Indigenous children after the introduction of PCV7. Our aim was to understand the potential importance of this population, particularly with regard to the presence of the capsule genes and antibiotic resistance. Importantly, we require evidence to guide reporting of NSP carriage and resistance in pneumococcal surveillance studies and as microbiological outcomes in clinical trials for otitis media.
Ethical approval for this study was granted by the Human Research Ethics Committee of the Northern Territory Department of Health and Families and the Menzies School of Health Research. Isolates were selected only if the parents or guardians consented at enrollment to the samples being used for future research.
NSP were grown on horse blood agar plates (Oxoid, Australia) at 37°C in 5% CO2 and identified as optochin sensitive, alpha-hemolytic colonies with a negative reaction to Quellung omniserum (1) and a positive autolysin (lytA) PCR result (19).
A total of 182 nasopharyngeal NSP isolates were randomly selected (Research Randomizer; http://www.randomizer.org) from four studies of otitis media (Table (Table1)1) conducted in the Northern Territory after the introduction of PCV7. The isolates were stored at −80°C in skim milk-tryptone-glucose-glycerol broth (20). When the specimens originated from longitudinal studies, sample selection was limited to the first NSP isolate recovered from each child.
Screening for penicillin and macrolide susceptibility was determined by a calibrated dichotomous susceptibility (CDS) disk diffusion assay (oxacillin and erythromycin, respectively) (3). Etests (AB Biodisk, Solna, Sweden) were used to determine the penicillin and erythromycin MICs of all nonsusceptible isolates (CDS radius < 6 mm). PCR for the detection of macrolide resistance genes (mefA/E and ermB was performed as described previously (18, 26).
NSP DNA for BOX typing and the detection of macrolide resistance genes was extracted by using the Instagene matrix (Bio-Rad), according to the manufacturer's instructions. NSP DNA for multilocus sequence typing (MLST) was extracted with a QIAamp DNA minikit (Qiagen), according to the manufacturer's instructions. The DNA concentration was determined with a Picogreen double-stranded DNA quantitation assay kit (Molecular Probes). The extraction of DNA for multiplex PCR/reverse line blot hybridization (mPCR/RLB-H) was done as described previously (12).
The capsule gene was detected by the method of Hanage et al. (8), with the exception that primers wzg-up (5′-ATCCTTGTCAGCTCTGTGTC) and wzg-down (5′-TCACTTGCAACTACATGAAC) amplified a 1,797-bp product.
BOX typing was done as described previously with 1 ng of DNA in each 50-μl reaction mixture (27). The BOX types were defined according to the banding patterns between 750 bp and 3,000 bp. MLST was undertaken essentially as described previously (6), except for modification of the spi forward primer (primer spiF3, 5′-CAGATTTTGCAAGCCTATGG), the recP reverse primer (primer recPR2, 5′-GCCGTGTACAGCATTAGTTC), and the ddl forward primer (primer ddlF2, 5′-GATGGCTCTGTTCAAGGATT). Molecular serotyping of wzg-positive isolates of NSP was by done mPCR/RLB-H, as described previously (11, 28), for 90 pneumococcal serotypes. This method targets serotype-specific regions of wzy for all serotypes with the exception of serotype 3. In the case of serotype 3, wze is targeted, as wzy is not present in serotype 3 strains.
Presumptive NSP in this study were identified on the basis of colonial morphology, optochin sensitivity in 5% CO2 for distinction from S. pseudopneumoniae (2), and a negative Quellung reaction. In longitudinal studies 1 and 2 (Table (Table1),1), presumptive NSP were reported in 18% and 8% of the nasopharyngeal swab specimens, respectively. Cross-sectional studies 3 and 4 reported presumptive NSP carriage prevalence rates of 8% and 5%, respectively. These numbers are likely to underestimate the true carriage rate because NSP were not systematically sought. Variations in the rates reported between studies may also be due to subjectivity in the selection of colonies.
Among the 182 randomly selected isolates of NSP (which were confirmed to be lytA positive), 49 BOX types were identified. The three most common BOX types (types 1, 14, and 3) represented 44% of the isolates. BOX types 1 and 14 were closely related, distinguished by a single band size difference of less than 50 bp.
Intermediate resistance to penicillin (MICs, 0.12 to 1 μg/ml; breakpoint for oral therapy of nonmeningeal strains) was evident in 49% (90/182) of the NSP isolates, penicillin resistance (MICs, ≥2 μg/ml) was detected in 2% (4/182) of the isolates, and erythromycin resistance (MICs, ≥4 μg/ml) was detected in 10% (19/182) of the isolates. Erythromycin resistance was mediated by the mefE macrolide resistance gene in all except one of the isolates; the latter isolate was negative for the mefA/E and ermB genes. Reduced susceptibility to these commonly used antimicrobials was detected in isolates from 32 BOX types.
Representative isolates of the three most common BOX types, including all antibiograms, were analyzed by MLST (Table (Table2).2). BOX type 1 and 14 isolates were represented by six sequence types (STs), all of which were new to the MLST database. Two STs were common to these BOX types. ST4485 was a double-locus variant of a single ST corresponding to a serotype 4 isolate. ST4487 and ST4488 were both double-locus variants of two STs corresponding to serotype 14 and 19A isolates. The BOX type 3 isolate was ST448, an NSP clone reported internationally. ST448 and STs 4486, 4484, and 4565 were not closely related (less than five identical MLST loci) to STs corresponding to capsular pneumococci that are currently in the MLST database. All isolates were confirmed to be pneumococci by using the “is it a pneumococcus?” analysis (www.mlst.net).
As described previously (8), we used PCR for the conserved wzg (cpsA) gene to test for the presence of a cps locus (Tables (Tables33 and and4).4). wzg was detected in 23 of 182 (13%) NSP isolates and 13 BOX types. However, 10 isolates provided a product of approximately 3 kb rather than the expected 1.8-kb product. Partial sequencing of the product demonstrated a sequence homologous to that of the pneumococcal wzg gene and an unidentified insert which was not related to streptococcal ali genes, as reported previously (9). The 10 isolates represented four different BOX types, suggesting that this insertion was not a recent event.
mPCR/RLB-H analysis of the isolates with the expected 1.8-kb wzg gene determined that five isolates corresponded to PCV7 serotypes 9V (or 9A) and 14, whereas eight corresponded to non-PCV7 serotypes 7A/F, 16F, 11A/D, 22A/F, and 13 (Table (Table3).3). Cocolonization with a capsular strain of the corresponding serotype was evident for 8 of 13 NSP with the 1.8-kb wzg product (Table (Table3).3). Subsequent testing of all strains with specific type antisera found that none reacted.
mPCR/RLB-H analysis determined that the 10 isolates with the 3-kb wzg PCR product corresponded to serotype 14. Two BOX type 14 isolates positive for the insertion were ST4488 (double-locus variants of STs corresponding to serotype 14 and 19A isolates).
One isolate consistently provided a serotype 14 result (based on the result for wzy) and a serotype 3 result (based on the result for wze), suggesting mosaicism of the strain and that both serotype 3 and 14 capsular genes may exist in this strain.
Reduced susceptibility to antimicrobials was common in all groups (Table (Table44).
Australian Indigenous children living in remote communities experience capsular pneumococcal carriage rates of approximately 80% and NSP carriage rates of 5 to 20%. These NSP carriage rates are likely to be underestimates because NSP are morphologically distinct from capsular pneumococci and NSP were not systematically sought in the studies that determined those carriage rates. NSP detection rates in a population appear to be dependent on the study protocol and subjectivity in colony selection.
Cocolonization with capsular and noncapsular pneumococci occurred frequently, providing an opportunity for an interaction between these populations. Previous studies of NSP report a highly diverse population with a pervasive nature, an ability to cause disease, and a likely important role in pneumococcal ecology (8, 23, 24). Although nonencapsulated forms are less likely to cause disease, their ability for gene transfer, including the transfer of antibiotic resistance and capsular/virulence genes, is predictably high due to their capacity for an increased rate of transformation (22).
Two distinct populations were detected in this study: NSP with and without the wzg (cpsA) gene. The majority (87%) of NSP in this study were wzg negative by PCR and were thus likely to be of the true nonencapsulated lineage. However, given that 10 of 23 wzg sequences were atypical due to an insertion, we cannot rule out the possibility that other isolates contained modified wzg sequences not detectable by our methods.
Our grant did not provide funding for MLST of all NSP isolates; however, MLST of representatives of the three most common BOX types identified authentic pneumococci, including six new STs and ST448 (which accounted for 11% of the NSP in this study). ST448 is a particularly interesting clone, in that it is internationally disseminated and has been associated with invasive disease (9).
Identification of the majority of NSP in this study was based on colonial morphology, optochin susceptibility, and a positive lytA PCR result. Therefore, it is likely that these strains were S. pneumoniae rather than S. pseudopneumoniae (which are generally optochin resistant in the presence of increased amounts of CO2) (2) or other closely related Streptococcus spp. However, as lytA is not ideal as a sole genetic criterion for pneumococcal identification, in due course MLST analysis may identify some strains as atypical pneumococci (divergent at the seven MLST loci by >5%) (7).
Capsular genes (wzg) were detected in a minority of the NSP isolates. Ten of these isolates had an insertion in wzg which may have interrupted capsule expression, while 13 contained capsule genes related to both PCV7 and non-PCV7 serotypes. Additionally, among 182 isolates, only 15 were potentially downregulated PCV7 types. These data do not support the hypothesis that PCV7 types are suppressing capsule expression to avoid vaccine-induced immune clearance.
Transformation is a common mechanism for the development of resistance to penicillin and macrolides in pneumococci (4, 5). As NSP have transformability rates up to 1,075-fold higher than those of capsulated strains (22), NSP may act as a vector for the transmission of resistance genes between viridans group streptococci and capsulated pneumococci. Indeed, a serotype 19F clone became increasingly non-penicillin susceptible to by the uptake of pbp2x gene fragments shared with NSP (10). Despite being present in low numbers, nonencapsulated forms of PCV7 serotypes are also likely to harbor the antibiotic resistance genes of their parent capsular vaccine serotypes and thus have the potential to transfer antibiotic resistance to newly emerging nonvaccine serotypes. Our 2005 surveillance of pneumococcal carriage by Indigenous children found that 14% of the isolates were non-penicillin susceptible (MIC ≥ 0.12 μg/ml) and that 6% of the isolates were azithromycin resistant (MIC ≥ 2 μg/ml) (15). The penicillin resistance data are in stark contrast to the 55.5% rate of non-penicillin susceptibility that we found in wzg-negative NSP and 26% in potentially revertible wzg-positive NSP. Thus, there is potential for the selection of existing capsular strains exhibiting antibiotic resistance and for the uptake of resistance genes by susceptible capsular pneumococci during exposure to NSP.
High carriage rates, antibiotic resistance, and the presence of capsule genes suggest a role for NSP in the maintenance and survival of capsular pneumococci. Until this role is better understood, we support continued reporting of NSP carriage rates. Since MLST is necessary to explicitly confirm the identity of nonserotypeable pneumococcal isolates, it will be necessary to analyze data obtained by using less rigorous methods of identification with caution.
In populations with high NSP carriage rates, combining the data for NSP and capsular pneumococci would inflate the rates of pneumococcal carriage and antibiotic resistance. Therefore, we recommend separate reporting of rates of carriage of NSP and capsular pneumococci in surveillance studies. When the microbiological outcomes of clinical trials are reported, we recommend the inclusion of only the capsular pneumococci in primary analyses.
We thank the Channel 7 Children's Research Foundation for providing financial support for this study.
We also thank the families who participated in our studies and agreed to storage and further analysis of their nasopharyngeal swab specimens. Without the support of these families, this study would not have been possible.
Published ahead of print on 30 December 2009.