|Home | About | Journals | Submit | Contact Us | Français|
Increasing frequencies of community-acquired methicillin-resistant Staphylcoccus aureus (MRSA) strain isolation have been reported from many countries. The overall prevalence of MRSA in Norway is still very low. MRSA isolates (n = 67) detected between 1995 and 2003 in northern Norway were analyzed by pulsed-field gel electrophoresis, multilocus sequence typing, and staphylococcal cassette chromosome mec (SCCmec) typing. Sixty-seven isolates were associated with 13 different sequence types. Two successful MRSA clones predominated. Sequence type 8 (ST8) (40%) and ST80 (19%) containing SCCmec type IV were detected in hospitals and communities in different geographic regions during a 7-year period. In general, there was a low level of antimicrobial resistance. Only 26% of the isolates were multiresistant. International epidemic clones were detected. The frequent findings of SCCmec type IV (91%) along with heterogeneous genetic backgrounds suggest a horizontal spread of SCCmec type IV among staphylococcal strains in parallel with the clonal spread of successful MRSA strains.
Methicillin resistance in staphylococci is mediated by acquisition of staphylococcal cassette chromosome mec (SCCmec), which integrates in a site-specific manner into the chromosome (23). Methicillin-resistant Staphylococcus aureus (MRSA) has become a worldwide problem. The prevalence of MRSA varies widely between countries. Its prevalence is consistently higher in the United States, Japan, and Southern Europe than in other countries; more than 30% of individuals in these countries are infected, compared with less than 2% in Scandinavia, The Netherlands, and Switzerland (10, 34). Recent reports suggest that community-acquired MRSA (CA-MRSA) infections in healthy persons without the known risk factors for MRSA infection are increasing in frequency (7, 20, 33). In Europe, the extracellular product Panton-Valentine leukocidin (PVL) has been associated with severe necrotizing pneumonia and community-associated staphylococcal skin infections, such as furunculosis (11, 16).
Infections with MRSA have been notifiable in Norway since 1995, but colonization without infection is not notifiable. The overall prevalence of MRSA is still very low in Norway. In 2003 there was an increase in the number of reported cases: 216 cases compared to 142, 122, and 67 in the three previous years. Only 0.3% of S. aureus isolates in blood cultures were identified as MRSA in 2002 according to the Norwegian surveillance system NORM (28). In Norway, about 50% of the MRSA isolates are imported from various countries. The remainders are termed “Norwegian” strains and are presumed to be of domestic origin, with no apparent connection with recent travel abroad (4, 28, 38). Reports from Norway documenting the clonality of MRSA isolates are scarce, and we do not know much about the distribution of CA-MRSA and hospital-acquired MRSA.
The aim of this study was to define the MRSA clonal types and their molecular epidemiology over time in northern Norway by different molecular typing techniques, including SCCmec typing, pulsed-field gel electrophoresis (PFGE), and multilocus sequence typing (MLST).
Northern Norway consists of three counties (Nordland, Troms, and Finnmark), with 11 hospitals and a total population of 460,000 inhabitants. From these counties, a total of 143 MRSA isolates were reported to and identified at the University Hospital of North Norway, Tromsø, Norway, in the period from 1995 to 2003. Sixty-seven out of the 143 MRSA strains were selected for this study based on the following criteria: only one MRSA isolate from each patient was included, and only one isolate from each verified and known outbreak was included; i.e., strains from the same outbreak, but with one band difference in PFGE banding pattern, were included. Samples from nonresidents of the region and patients with a recent history of hospitalization abroad (i.e., within 1 year, based on clinical information given by the physician in charge) were excluded from the study.
Fifty out of 67 strains were from outpatients without any known link to health care facilities, and 17 strains were from hospitalized patients at five different hospitals. Seven out of 67 isolates were from nursing home residents, and they were grouped as nonhospital/outpatient samples. Both clinical and screening samples were included. The study included only two invasive isolates, while the remaining 65 isolates were recovered from wound secretions (n = 31), abscesses (n = 8), nasal swab screenings (n = 14), and other sites (n = 12). The 14 nasal swab samples were taken in association with the screening of hospital patients (n = 4), nursing home residents (n = 5), and hospital and nursing home staff (n = 5). Seven of the 14 nasal swab samples were from nonhospital individuals/outpatients (staff, n = 2; patients, n = 5), and seven of the 14 nasal swab samples were from the hospital (hospital staff, n = 3; hospital patients, n = 4).
For reference, the mecA-negative strain S. aureus NCTC 8325 and the SCCmec-positive strains S. aureus NCTC 10442, S. aureus N315, and S. aureus 85/2082 (21) were included in all analyzes. Five international MRSA clones (Iberian clone, Brazilian clone, pediatric clone, New York/Japan clone, and Hungarian clone) were included in the PFGE analyzes. The collection and analyzes of data were approved by the regional committee for medical research ethics and by the Norwegian social science data services.
All isolates were identified by positive Gram stain, the presence of catalase, and the presence of clumping factor (bound coagulase) by Staphaurex Plus* (Murex Biotech, Dartford, England). All strains were tested for the presence of the mecA gene and nuclease (nuc) gene by a multiplex mec/nuc PCR (6, 31).
Antimicrobial susceptibility testing was performed using disk diffusion and breakpoints according to the Norwegian Working Group on Antibiotics (5) for ciprofloxacin (10 μg), doxycycline (30 μg), gentamicin (30 μg), trimethoprim-sulfamethoxazole (8 μg), fusidic acid (50 μg), rifampin (5 μg), streptomycin (30 μg), chloramphenicol (30 μg), and kanamycin (30 μg). The medium used for disk diffusion was the paper disk methods medium (PDM; AB Biodisk, Stockholm, Sweden). The NCCLS uses wider ranges when defining their breakpoint values, resulting in fewer isolates being classified as resistant. The Norwegian Working Group on Antibiotics set its breakpoints close to that of the native (often susceptible) bacterial population (24). Susceptibility to arbekacin (30 μg; Eiken Chemical Ltd., Japan) was determined by disk diffusion on Mueller-Hinton agar with an inoculum at a McFarland standard of 0.5. No NCCLS breakpoint for arbekacin exists, but the same zone diameters as for gentamicin were used for interpretation: for resistance, ≤ 13 mm; for intermediate susceptibility, 14 to 17 mm; and for susceptibility, ≥ 18 mm. MICs of oxacillin, linezolid, and mupirocin were determined by Etest (AB Biodisk, Solna, Sweden) according to NCCLS interpretive standards (27). Susceptibility to vancomycin was tested by a vancomycin agar screen (36). Detection of inducible clindamycin resistance was performed according to the standard NCCLS disk diffusion test using unsupplemented Mueller-Hinton agar and 15-μg erythromycin disks and 2-μg clindamycin disks as described by Fiebelkorn et al. (15).
Production of beta-lactamase was tested with nitrocefin disks (Oxoid, Basingstoke, United Kingdom) according to the instructions of the manufacturer.
Chromosomal DNA was prepared and SmaI digestion and PFGE were performed as described previously (19). The PFGE gels were analyzed both visually according to the method of Tenover et al. (35) and by GelCompar II software version 2.5 (Applied Maths, Kortrijk, Belgium). The PFGE types were defined on the basis of the DNA banding patterns in accordance with the criteria of Tenover et al. (35) for bacterial strain typing.
The SCCmec types were determined by PCR typing of the mec and ccr gene complexes as described previously (19, 21). All strains were tested for the presence of IS1272 and type IV SCCmec with the primers and PCR described in the work of Okuma et al. (29).
Southern blot transfer of PFGE SmaI-digested genomic DNA to a positively charged nitrocellulose membrane (Roche, Mannheim Germany), ccrAB probe labeling, and Southern blot hybridization were carried out as described previously (19).
PCR products were purified with the EXO/SAP (shrimp alkaline phosphatase and exonuclease I) PCR product presequencing kit (USB Corporation, Ohio). Heterogeneity in the ccrA and -B genes was identified by bidirectional DNA sequencing as described by Hanssen et al. (19).
The ccrAB nucleotide sequences and deduced amino acid sequences were edited by using the Chromas software (version 2.21) and aligned using the BioEdit sequence alignment editor (v5.0.9) (18). Nucleotide sequences were compared to sequences in GenBank, and protein sequences were compared to nonredundant GenBank coding sequence translations, by using the BLASTN, BLASTP, and BLASTX local alignment search tools (3).
Chromosomal DNA was extracted using the GenoM-48 robotic workstation (GenoVision, Oslo, Norway) and the protocol MagAttract DNA tissue. MLST was performed according to the protocol of Enright et al. (12) with a few modifications of the PCR master mix. The PCRs were carried out in a final volume of 25 μl containing 25 ng of chromosomal DNA, 0.025 μg of each primer (except for tpi_up and tpi_dn, of which 0.25 μg was added), 0.2 mM deoxynucleoside triphosphate (Invitrogen, Carlsbad, CA), 0.5 U Platinum Taq DNA polymerase (Invitrogen), 2.5 μl 10× buffer, and 1.5 mM MgCl2 (both supplied with Platinum Taq polymerase). The amplified products were purified using the GFX96 PCR purification kit (Amersham Biosciences, Buckinghamshire, United Kingdom), followed by sequencing reactions performed using the Big Dye Terminator v3.1 cycle sequencing kit (Applied Biosystems). Unincorporated dye terminators were removed from the extension products by isopropanol precipitation as described in the ABI Prism BigDye Terminator cycle sequencing ready reaction kit protocol (Applied Biosystems). Sequences of both strands were analyzed using an ABI Prism 3100 DNA genetic analyzer (Applied Biosystems). The sequencing primers were the same as in the initial PCR amplification.
ccr sequences were submitted to the EMBL/GenBank database and assigned accession numbers AY669512, AY254749, AY254756, and AY254758.
Sixty-six strains were resistant to oxacillin, while one strain had an oxacillin MIC of 1 μg/ml. All 67 strains tested positive for the presence of the mecA gene by PCR and were therefore considered MRSA isolates (Table (Table1).1). There was a 98.5% correlation between phenotypical expression of oxacillin resistance (oxacillin MIC) and the presence of the mecA gene. Forty-four isolates had oxacillin MICs of <96 μg/ml, while 23 isolates had an oxacillin MIC of ≥96 μg/ml.
Among 66 S. aureus strains, 15 PFGE profiles were obtained (Fig. (Fig.1;1; Table Table2).2). One strain was not PFGE typeable, and it was not included in the PFGE analysis. A major PFGE type, F, was detected among 26 isolates with 19 different subtypes. Isolates with PFGE type F were characterized by sequence type 8 (ST8) and SCCmec type IV (ST8-IV) (Fig. (Fig.1,1, left), whereas isolates of PFGE types M and N were characterized by ST80 and SCCmec type IV (Fig. (Fig.1,1, right). The strains N315 and 85/2082 and the Brazilian clone, Hungarian clone, and Iberian clone did not show any relationship to any of the clinical isolates in this study. NCTC 8325 showed a two-band difference from isolates with PFGE patterns F1 and ST8-IV. They were considered closely related (Fig. (Fig.1,1, left). MRSA strain NCTC 10442 (ST250-I) showed a four-band difference from strains harboring PFGE patterns F1 and F4. The pediatric clone (ST5-IV) clustered with the strains containing ST5-IV and showed a three-band difference from strains with PFGE pattern G4. They were considered possibly related (data not shown). The New York/Japan clone (ST5-II) was considered possibly related to strains with PFGE pattern G4, showing a six-band difference (data not shown).
Sixty-one (91%) out of 67 isolates was of SCCmec type IV (Table (Table2).2). Two isolates were of SCCmec type I, two were of SCCmec type II, one isolate was of SCCmec type III, and one isolate was of SCCmec type IIIA. Three out of 67 strains were ccrAB PCR untypeable. Positive Southern blot hybridization with ccrAB probes, the mec complex PCR, and the IS1272-mecA PCR indicated that they were of SCCmec type III, IIIA, or IV. The ccrA2 gene was sequenced in three ST8-IV isolates, three ST80-IV isolates, two ST5-IV isolates, and one ST12-IV isolate. The ccrA2 gene in strains with ST80 (GenBank accession no. AY669512), ST5, and ST12 showed 100% identity to ccrA2 in strain MR108 containing SCCmec type IVc (GenBank accession no. AB096217). ccrA2 in ST8-IV MRSA strains (GenBank accession no. AY254749, AY254756, and AY254758) (19) showed 96% and 99% identity to ccrA2 in strains N315 (SCCmec type II) (GenBank accession no. D86934) and MR108 (SCCmec type IVc) (GenBank accession no. AB096217), respectively. ST8-IV strains contained a local variant of the ccrA2 gene.
MLST identified 13 STs among the 67 isolates (Table (Table2).2). No new STs were detected. Six STs were each represented by a single MRSA isolate only. ST8 and ST80 were the most frequent, and they were identified in 40% (27/67) and 19% (13/67) of the isolates, respectively. All MRSA isolates with ST80 were of SCCmec type IV. Two different SCCmec types were identified among ST8 isolates: ST8-I (one isolate) and ST8-IV (26 isolates). One ST5 isolate was associated with SCCmec type I, and four isolates of ST5 were associated with SCCmec type IV. They showed close relationship to four isolates containing ST125-IV on PFGE. The 67 strains were classified into seven clonal complexes (Table (Table2).2). Combining the ST and SCCmec type, we identified 15 clonal types, out of which clones ST8-IV, ST80-IV, ST45-IV, ST125-IV, and ST5-IV were each represented by four or more isolates. Of the 15 clonal types, only 1 was identical to one of the five pandemic MRSA clones. ST5-IV (pediatric clone) was represented by four isolates. Other international clones observed were ST45-IV (Western Europe, Scandinavia) (http://www.mlst.net) (n = 4) and ST8-IV (EMRSA-2, -6, and NY/V) (13) (n = 26); ST30-IV and ST36-II (EMRSA-16) (13, 30), represented by three isolates and one isolate, respectively; and ST22-IV (EMRSA-15) (1, 12) and ST5-I (EMRSA-3), represented by two isolates and one isolate, respectively. ST225 and ST125 are both single-locus variants of ST5, and ST39-IV is a double-locus variant of ST239 (14) (Table (Table22).
Resistance to fusidic acid, erythromycin, and kanamycin was detected in 48%, 33%, and 37% of the isolates, respectively. Only a few of the isolates showed resistance to ciprofloxacin (13%), doxycycline (19%), streptomycin (24%), rifampin (3%), gentamicin (5%), clindamycin (10%), and chloramphenicol (8%) (Table (Table1).1). All isolates were susceptible to trimethoprim-sulfamethoxazole, mupirocin, arbekacin, and vancomycin (Table (Table1).1). Constitutive production of beta-lactamase was found in 84% of the isolates. Inducible clindamycin resistance was detected in 18% of the isolates. Of the 67 strains, 18 (26%) were multiresistant (i.e., resistant to three or more antibiotic groups in addition to beta-lactams). The strains were resistant to an average of 2.7 antibiotics.
All ST80-IV strains were resistant to aminoglycosides (kanamycin, streptomycin) and oxacillin. In most cases, erythromycin, clindamycin, fusidic acid, ciprofloxacin, and chloramphenicol resistance were linked to ST80-IV strains (Table (Table2).2). The resistance profile typical for ST8-IV was fusidic acid and oxacillin, which was observed in 17 (65%) out of 26 strains (Table (Table22).
In an attempt to determine the molecular epidemiology of MRSA strains in northern Norway from 1995 to 2003, 67 isolates were characterized by different molecular typing methods. The majority of isolates belonged to two predominant clones, i.e., ST8-IV and ST80-IV, whereas the remaining isolates showed great genetic diversity.
The most frequent MRSA clone in northern Norway was ST8-IV. This clone is represented by EMRSA-2 and -6 and has been identified in several European countries and the United States (2, 13). Chung et al. (8) reported ST8-IV as the second most frequent MRSA clone in Miami, FL. They called this clone the New York clone V, and the ST8 genetic background was proposed to represent the predicted ancestor of the very first European MRSA strain. This clone was multiresistant. This is in contrast to the results of our study, where only fusidic acid and oxacillin resistance was observed in 17 out of 26 ST8-IV isolates. ST8-MRSA clones are proposed to have emerged by multiple independent introductions of SCCmec into a successful ST8-methicillin-susceptible S. aureus (MSSA) clone (13). In our study, strains with ST8-IV were highly related, suggesting clonal expansion in northern Norway. However, variation in the resistance pattern and local variants of ccrAB sequences found in northern Norwegian ST8-IV strains suggest that they have evolved for some time separated from their origins and possibly acquired SCCmec from local staphylococcal strains (19).
The second most frequent MRSA clone was ST80, accounting for 19% of all MRSA isolates, indicating that strains with ST80-IV have established themselves outside hospitals in northern Norway between 1999 and 2003 and are probably emerging. Two of the ST80 strains were recovered from the same hospital, while the remaining 11 isolates were from outpatients. ST80 has been associated with the central European CA-MRSA isolates reported from France and Switzerland (39). Aires de Sousa et al. (2) described MRSA strains with ST80 in Greece. They all contained a distinct spaA type, SCCmec IV, and were resistant to oxacillin, erythromycin, and fusidic acid. Erythromycin and fusidic acid resistance was associated with ST80 in 9 out of 13 strains in our study, while kanamycin, streptomycin, and oxacillin resistance was observed in all 13 strains. Interestingly, the ST80 strains were resistant to an average of 4.1 antibiotics, in contrast to, e.g., ST8 strains, which were resistant to 1.6 antibiotics. PVL and SCCmec type IV are common among CA-MRSA isolates of ST80 (11, 16). In our study, the ST80-MRSA strains were not tested for the presence of PVL. It remains to be elucidated if PVL has contributed to the success of the ST80 MRSA strains. At present the origin of clone ST80-IV is not clear. It remains to be elucidated what has facilitated the spread of ST8 and ST80, if these STs are representative of the whole country, and if the numbers are increasing in the region. Today, it is unknown which clone(s) is responsible for the increase in the number of reported MRSA cases in 2003 in Norway. This is a subject for further studies.
In contrast to the MRSA isolates represented by the two dominant clones, the rest of the isolates were more genetically diverse. Twenty-eight out of 67 isolates (41.8%) represented STs with less than five isolates, and we considered them sporadic cases. Andersen et al. (4) reported that most cases of MRSA infection in Norway occur unexpectedly or sporadically and that the epidemic ends shortly afterward, e.g., within 1 month. Most sporadic cases of MRSA strains in northern Norway appear to be due to the continuous introduction of new international clones into the country or the establishment of new MRSA clones.
It is noteworthy that strains with ST80-IV, ST5-IV, and ST12-IV in this study contained ccrA2 genes with 100% identity to ccrA2 in SCCmec type IVc. According to Ito et al. (22), this is the first report of CA-MRSA associated with SCCmec type IVc.
The isolates in this study were generally susceptible to most of the antibiotics tested, apart from fusidic acid. Fusidic acid resistance was observed in 47.8% of the isolates. This has also been reported by Tveten et al. (37), who describe a specific clone of S. aureus with resistance to fusidic acid that causes impetigo bullosa in outpatients in Norway. Other studies indicate that the Norwegian MRSA types are similar to the old classical, sensitive MRSA types, often with a relatively low level of methicillin resistance (4). There was no correlation between antibiogram and SCCmec type in our study; i.e., multiresistance was observed both in strains containing SCCmec type IV and in strains harboring the other SCCmec types. We observed 24 isolates with SCCmec type IV that were resistant to three or more classes of antimicrobial agents. This is in accordance with other studies suggesting that some strains with SCCmec type IV and a high level of resistance have acquired resistance to non-beta-lactam antibiotics in order to be able to survive in the hospital environment (1) or through exposure to the antibiotics (29).
There was a high correlation between ST and PFGE type in this study. Strain types that were grouped by MLST had similar PFGE profiles, whereas distinct MLST strain types had very different PFGE profiles. This has also been reported by others (13, 17).
In the present study, 61 out of 67 isolates (91%) containing SCCmec type IV were associated with nine different STs. All isolates containing SCCmec type IV belonged to 12 different PFGE types. It has been suggested that the relatively small SCCmec type IV strains may have an increased mobility compared with their larger SCCmec counterparts and therefore greater propensity for transfer to diverse S. aureus genetic backgrounds (1, 9, 11). There is a less intensive selective pressure outside hospitals, advantageous for SCCmec type IV, as SCCmec type IV gives a smaller fitness cost for the bacteria than SCCmec types I, II, and III (1). The presence of SCCmec type IV in distinct genetic backgrounds could represent either an independent acquisition event or horizontal gene transfer of the SCCmec element from one S. aureus strain to another (25, 30).
Seventy-four percent of the MRSA strains were isolated from persons in the community. According to our epidemiological information, they did not have any known link to health care facilities or known risk factors for the acquisition of MRSA. Most of the strains were associated with type IV SCCmec (91%), nonmultiresistance (74%), low levels of oxacillin resistance (oxacillin MIC, <96 μg/ml) (73%), and genetic diversity. These are all characteristic traits associated with CA-MRSA (29). However, the same clonal types (e.g., ST8-MRSA-IV and ST80-MRSA-IV) were observed in both hospitals and communities, and therefore it is impossible to decide where they originated. It is likely that some health care-associated MRSA cases were misclassified as community-associated MRSA cases and vice versa. The true site of acquisition of MRSA is rarely known with certainty (33). Our findings are consistent with community-acquired, rather than nosocomial, MRSA infection. Further studies among randomly selected healthy members of the community will probably give the answer if we have “true CA-MRSA” strains in northern Norway.
It is assumed that SCCmec type IV has repeatedly integrated into MSSA backgrounds (32) and that this element may be acquired from coagulase-negative staphylococci (26). The molecular typing performed in this study has shown that the ST8 isolates are closely related, so they could recently have diverged from a common ancestor, for example, by horizontal gene transfer of the SCCmec type IV to a local MSSA strain. MSSA NCTC 8325 showed a close relationship to the north Norwegian MRSA isolates. This illustrates the need for MLST analyzes of MSSA strains in Norway in order to say something about the origin of the MRSA strains.
We have shown that two successful MRSA clones predominate in northern Norway. The majority of strains are CA-MRSA and SCCmec type IV, and there is generally a low-level of antibiotic resistance. It remains a puzzle why such a large proportion of MRSA disease is caused by only two clonal types of bacteria, and it is unknown what factors determined the ease of transmission of these MRSA strains. The frequent findings of SCCmec type IV along with heterogeneous genetic backgrounds may indicate horizontal transfer of SCCmec IV among staphylococcal strains in northern Norway in parallel to the clonal spread of successful MRSA strains. Monitoring the geographic distribution of epidemic clones may contribute to our understanding of why certain MRSA clones are spreading between countries, whereas others are limited to a single country, city, or hospital.
This work was supported by grants from the University of Tromsø, Tromsø, Norway.
We thank K. Hiramatsu of Juntendo University, Tokyo, Japan; Y. Tveten of Telelab, Skien, Norway; and the department of Microbiology, University Hospital of North Norway, for kindly providing strains. We thank T. Tessem, I. Sæther, and L.-H. Henriksen for excellent technical assistance. Thanks go to S.-H. Barkhald for extracting clinical data for all the strains. We also thank R. Jureen, C. Klingenberg, A. Sundsfjord, and G. Simonsen for critical reading of the manuscript.