Distribution of epidemic antibiotic-resistant pneumococcal clones in Scottish pneumococcal isolates analysed by multilocus sequence typing

Since the earliest report of penicillin resistance in pneumococci from Boston, USA, in the 1960s (Kislak et al., 1965), antibiotic resistance in this organism has continued to emerge. Levels of antibiotic resistance have expanded to include a wider range of antimicrobials, and an increased number of affected geographical locations (Klugman, 1990; Schreiber & Jacobs, 1995). DNA-sequencing techniques have led to the identification of pneumococcal antimicrobial-resistant clones. This phenomenon is due to the acquisition of resistance genes within specific clonal lineages (Munoz et al., 1992). In particular, clones originating from Spain, South Africa, Hungary and the USA have achieved a global distribution (Coffey et al., 1996; McDougal et al., 1995; Smith & Klugman, 1997). The understanding of the epidemiology of the emergence and spread of resistance within the pneumococcal population has been facilitated by the Pneumococcal Molecular Epidemiology Network (PMEN) (McGee et al., 2001). The PMEN system takes advantage of multilocus sequence typing (MLST), a DNA-sequence-based typing method, which uses data from a number of housekeeping genes to provide a sequence type (ST). MLST data are highly discriminatory and electronically portable between laboratories via the internet, making the technique ideal for investigating the clonal spread of antibiotic-resistant pneumococci (Enright & Spratt, 1998a). Criteria for inclusion of a clone in the PMEN database include widespread geographical distribution, persistence over a number of years, and resistance to one or more antibiotics (McGee et al., 2001).

Previous work (Jefferies et al., 2003) using MLST and serotyping has constructed a phylogenetic framework for a collection of pneumococci isolated in Scotland. We describe an analysis of these isolates in terms of their antimicrobial susceptibility, ST and relationship to the globally significant PMEN clones.

Isolates.
Isolates were selected from a collection referred to the Scottish Meningococcus and Pneumococcus Reference Laboratory (SMPRL) over the period 19962003. Isolates for this study (n=252) comprised 190 blood isolates, 8 cerebrospinal fluid isolates, 22 eye isolates, 9 ear isolates, 7 sputum isolates, and 16 isolates from miscellaneous sources. Isolates were selected initially on the basis of serogroup to include at least five members where available, and to represent a proportional number from the wide range of serotypes found in Scotland (Denham & Clarke, 2005); they included rare serotypes, and isolates well distributed with respect to geographical location within Scotland, type of disease caused, and age of patient. Isolates were not preselected according to antimicrobial resistance.

Serotyping.
Isolates were serotyped by co-agglutination, as described by Smart (1986).

MLST.
MLST (Enright & Spratt, 1998a) was performed on Streptococcus pneumoniae strains using a semi-automated protocol (Jefferies et al., 2003, 2004). Single-locus variants (SLVs) and double-locus variants (DLVs) are thought to be sufficiently related to be considered members of the same clonal group of the parental ST clone [based upon related sequence types (BURST), ].

Antimicrobial susceptibility testing.
MICs for Scottish and PMEN isolates were determined using Etest strips (AB Biodisk), in accordance with the manufacturer's instructions. Breakpoints for susceptibility were those given by the British Society for Antimicrobial Chemotherapy (; penicillin resistance 2 mg l¹, intermediate resistance 0·121·0 mg l¹; erythromycin resistance >1 mg l¹; clindamycin resistance >1 mg l¹; ceftriaxone resistance 2 mg l¹).

Demographic data
The 252 isolates in our collection are from 39 different serotypes, and they include eight non-typable isolates (Jefferies et al., 2004). The isolates were assigned to 109 STs. The mean age of patients was 45 years (range <1 month to 103 years). The range of penicillin susceptibility was 0·0042 µg ml¹, with the majority of isolates penicillin susceptible (MIC <0·12 µg ml¹, n=214). One isolate, an SLV of the Taiwan^19F-14 clone, was resistant to ceftriaxone, and the collection contained just two isolates with a ceftriaxone MIC of 2 µg ml¹. There were 26 erythromycin-resistant isolates, and these were found in STs 9, 90, 143, 156, 180, 814, 860, 271, 285, 320, 377 and 1000.

Isolates with identical STs to PMEN clones
There were 16 isolates with STs identical to three PMEN clones (Spain^9V-3, England¹⁴-9 and Utah^35B-24) (Table 1). Although members of the Spain^9V-3 clone are penicillin resistant, and erythromycin sensitive, all of the seven isolates in our collection with an identical ST to the Spain^9V-3 PMEN clone had lower penicillin MICs than the corresponding type strain. One of these isolates was an erythromycin-resistant, penicillin-intermediate variant of the Spain^9V-3 clone. Three of the isolates related to the Spain^9V-3 clone had different capsular types (serotypes 8 and 14, and a non-typable isolate). In contrast, the eight isolates with an ST identical to the erythomycin-resistant England¹⁴-9 clone were erythromycin resistant, and had an identical serotype to the type strain. The remaining isolate, which had an identical ST and serotype to the penicillin-resistant Utah^35B-24 clone, was an erythromycin resistant, but penicillin intermediate, variant of this clone.

Table 1. Selected Scottish pneumococci sharing an identical MLST with a representative strain from the PMEN PEN, penicillin; ERY, erythromycin; CLI, clindamycin; CRO, ceftriaxone. The strain identification number indicates the year of isolation, e.g. 01-3648 was isolated in 2001. Type strains are shown in bold.

SLVs and DLVs related to PMEN clones
Analysis of our 252 isolates for SLVs and DLVs of the PMEN clones (Table 2) revealed seven SLV groups related to seven PMEN clones, and four DLV groups related to four PMEN clones. For 14 of the 15 SLVs of the PMEN clones, the penicillin MIC was less than the corresponding PMEN type strain (intermediate resistance).

Table 2. SLVs and DLVs of the PMEN clones in Scotland

Analysis of the PMEN SLVs for erythromycin susceptibility revealed two variants that were erythromycin resistant, and penicillin intermediate or sensitive, one belonging to the Spain^23F-1 clone and one to the Spain^9V-3 clone. When comparing erythromycin-resistant PMEN clones, only one SLV (related to the Taiwan^19F-14 clone) demonstrated high-level erythromycin resistance, although the type strain demonstrates low-level resistance. Many of the SLVs of the PMEN clones demonstrated capsule switching (Table 2). Of interest was the observation that the SLVs occurred at the same loci for a particular clonal group; for example, the ddl locus for SLVs of Utah^35B-24 and Spain^9V-3, and gki for SLVs of Spain^6B-2.

DLVs were found for four clonal groups (Spain^9V-3, Tennessee^23F-4, Poland^23F-16 and Columbia^23F-26). One isolate (ST 176 03-1562) was a DLV of the Poland^23F-16 and Columbia^23F-26 clones. The DLV isolates related to both the Poland^23F-16 clone and the Columbia^23F-26 clone, and a DLV related to the Tennessee^23F-4 clone demonstrated capsule switching. The two DLV clonal groups related to pencillin-resistant PMEN clones (ST 163 and 176) had lower penicillin MICs than the corresponding PMEN type strain; both of these isolates had different ddl loci compared with the corresponding PMEN type strain. For the erythromycin-resistant PMEN clone (Tennessee^23F-4), the DLV isolates in our collection were erythromycin-sensitive variants, and all varied from the type strain at the gki and recP loci.

STs of antibiotic-resistant isolates not related to epidemic clones
Analysis of our collection of 252 pneumococci for the presence of resistance to penicillin or erythromycin revealed six isolates not found in the PMEN database (Table 3). Two of these isolates were resistant to penicillin, erythromycin and clindamycin, and belong to a multi-drug-resistant clone isolated in Australia, China and Hong Kong (ST 320). There were four isolates that were erythromycin-resistant variants of previously described clones. All these clones are not assigned to the PMEN database. In addition, we noted a serotype 6 isolate with high-level erythromycin resistance belonging to ST 814; to our knowledge, this is the first description of this ST.

Table 3. Scottish antibiotic-resistant isolates unrelated to PMEN clones

This study provides useful information on the spread of international clones within the Scottish pneumococcal population. It is necessary to understand whether antibiotic resistance in the pneumococcal population in Scotland is driven by expansion of the PMEN clones, or the horizontal donation of locally acquired resistance genes. Studies that have focussed on resistant pneumococcal isolates have represented a biased view of the pneumococcal population, and tended to support a clonal relationship amongst resistant isolates. The present study includes a population consisting of both sensitive and intermediate resistant isolates, which is more representative of the pneumococcal population (although it excludes the carried population), and it suggests that the PMEN clones are not widely distributed amongst disease-causing isolates in Scotland, with many failing to display a fully resistant phenotype. However, it is difficult to obtain a precise picture of the molecular epidemiology of resistance in this cohort of pneumoccal isolates without data on the penicillin-resistance genes.

The incidence of invasive pneumococcal disease in Scotland is similar to that reported in most studies from the USA, Sweden, and other parts of the UK, with serotypes/groups 14, 23, 9, 6 and 19 predominating among pneumococci that cause invasive disease (Kyaw et al., 2000, 2002; Clarke et al., 2003). Of interest is the relatively low prevalence of high-level penicillin resistance among pneumococci in Scotland; over the period 19921999, only two isolates were reported with MICs 2 mg l¹ (0·2 % of isolates) (Kyaw et al., 2000). More recent data for the year 2000 found no isolate (n=360) with a penicillin MIC >2 mg l¹ (Clarke et al., 2003). Levels of erythromycin resistance were higher, at 10 % over the period 19941999, rising to 13 % in 2000 (Kyaw et al., 2000; Clarke et al., 2003).

Our collection contained a small number of isolates identical to the Spain^9V-3 clone, but some of these displayed different capsular types (non-typable, 8 and 14), and had lower penicillin MICs than the corresponding PMEN type strain; the strains were tested using Etest methodology. Of interest was the sequence changes in the D-alanine, D-alanine ligase (ddl) locus of SLV and DLV variants of the Spain^9V-3, Poland^23F-16, Utah^35B-24 and Columbia^23F-26 PMEN clones. The ddl gene is known to be closely linked to the penicillin-binding protein pbp2b gene, particular alleles of which confer resistance to β-lactam antibiotics. These changes may account for the differing susceptibilities of these isolates to β-lactam agents (Beall et al., 2002). The pbp2b genes may have been acquired as part of a hitchhiking effect, along with the DNA fragments acquired during alterations in the ddl locus (Enright & Spratt, 1998b).

Most of the Scottish erythromycin-resistant isolates related to the PMEN clones were related to the England¹⁴-9 clone, which displays the most frequently identified serotype in Scotland (Kyaw et al., 2000). This clone displays the drug efflux erythromycin-resistance pattern (MIC 212·0 mg l¹), which is thought to be due to carriage of the mefA gene (Amezaga et al., 2002). Of the erythromycin-resistant isolates not related to the PMEN clones, all demonstrated the high-level erythromycin-resistance phenotype (MLS_B).

A unique feature of the global spread of drug-resistant pneumococci during the last four decades is the emergence of relatively few genetic lineages that have achieved unusually wide geographic spread. The mechanisms of dissemination and successful establishment of these clones are unknown. At present, the incidence of penicillin- and multi-drug-resistant pneumococci is relatively low in Scotland. Our collection of pneumococci from Scotland, which was not selected on the basis of antimicrobial susceptibility, has shown that isolates with an identical genotype to PMEN clones are circulating in the Scottish population; however, many of these isolates do not display the same phenotypic traits of high-level antimicrobial resistance or serotype as the parental clone. Further work is required to elucidate whether future changes in the antibiotic resistance pattern of Scottish pneumococci are driven by the clonal expansion of PMEN clones, and/or horizontal recombination of locally induced resistance genes.

This study was funded by the Scottish Executive Health Department, Chief Scientist Office. We acknowledge the use of the pneumococcal MLST database, which is located at Imperial College London, and is funded by the Wellcome Trust.

Footnotes

†Present address: Portsmouth City PCT, Finchdean House, Milton Road, Portsmouth, UK.