To investigate the molecular epidemiology of ciprofloxacin-resistant CTX-M-15-producing Klebsiella pneumoniae epidemic clones (ECs) isolated from six nosocomial outbreaks and sporadic cases during 2005 in Hungary.
Two hundred and eighty-one extended-spectrum β-lactamase (ESBL)-producing K. pneumoniae clinical isolates collected from 41 centres were submitted to the National ESBL Reference Laboratory for further investigations. Of the 281 strains, 75 isolates proved to be SHV producers, whereas 6 isolates were ciprofloxacin-susceptible CTX-M-type ESBL producers. One hundred and ninety-six ciprofloxacin-resistant CTX-M-type β-lactamase-producing isolates collected from 35 centres were subjected to macrorestriction profile analysis. Furthermore, molecular typing was performed by PCR and sequencing of several antibiotic resistance genes, plasmid profile analysis, transfer of resistance determinants and multilocus sequence typing (MLST).
PFGE revealed the existence of three genetic clusters defined as ECs, where 129 isolates belonged to the previously described Hungarian EC (HEC), 46 isolates to epidemic clone II (EC II) and 21 isolates to epidemic clone III (EC III), respectively. All isolates harboured plasmids ranging from 2.0 to 230 kb. PstI digestion of plasmid DNA from transconjugants/transformants revealed diverse restriction patterns from distinct ECs. Sequence analysis of β-lactamase genes from 19 selected isolates detected blaCTX-M-15 and blaOXA-1 in strains from all three ECs and blaTEM-1 in EC III isolates located on large plasmids. ISEcpI associated with CTX-M-15 was detected only on a 50 kb non-conjugative plasmid from EC III. MLST identified three allelic profiles: ST 15 (HEC), ST 11 (EC III) and the novel ST 147 (EC II), which correspond to the PFGE clusters, respectively.
In 2005, 97% of all CTX-M-producing K. pneumoniae isolates detected across Hungary were highly ciprofloxacin-resistant CTX-M-15 producers and represented just three stable genetic clones.
During the past decades, cefotaximases have comprised the most rapidly growing group of extended-spectrum β-lactamases (ESBLs). They have been increasingly detected in Europe, Africa, America and Asia.1 CTX-M-15 recently emerged as the dominant type of cefotaximase in Gram-negative pathogens causing outbreaks in nosocomial as well as community settings.2
ESBL-producing Klebsiella pneumoniae constitute one of the most common Gram-negative bacteria showing multiple antibiotic resistance worldwide.1,2 These opportunistic pathogens are responsible for nosocomial infections primarily in intensive care units (ICUs) and pose a major risk especially to immunocompromised patients treated in these wards.3
Data collected by the National ESBL Survey, initiated by the National Center for Epidemiology in 2002, showed that K. pneumoniae was the most frequent ESBL-producing pathogen in Hungary, with an incidence of 65% to 75% of all ESBL-producing Enterobacteriaceae. Between 2002 and 2004, the predominant ESBL type in Hungary proved to be SHV,4,5 although the first 17 CTX-M-producing K. pneumoniae isolates—belonging to the Hungarian epidemic clone (HEC)—were already detected in 2003. This EC produced CTX-M-15 and was also resistant to ciprofloxacin. They were isolated from patients at eight geographically distant hospitals in five Hungarian counties and in the city of Budapest, primarily from post-operative wound infections. The widespread dissemination of HEC poses a serious threat in Hungarian healthcare institutions warranting continuous monitoring to control its spread.6
The aims of the study were the continuous monitoring of the incidence of HEC and the comprehensive molecular and epidemiological analysis of potential HEC isolates and HEC infections.
Materials and methods
In 2005, a total of 8124 K. pneumoniae strains were isolated from patients attending participating Hungarian hospitals. Antibiotic susceptibility testing in local laboratories detected 281 presumably ESBL-producing isolates, which were submitted to the National Center for Epidemiology for confirmation and typing. Of the 281 strains submitted from 41 centres, 75 isolates proved to be SHV producers. Six isolates were ciprofloxacin-susceptible CTX-M-type ESBL producers. Of the 200 isolates (four duplicates), 196 ciprofloxacin-resistant CTX-M-type β-lactamase-producing K. pneumoniae isolates collected from 35 Hungarian county and teaching hospitals were selected for further analysis. The selected isolates were obtained from 189 inpatients (one isolate/patient) from January to December 2005 from the following samples: urine (n = 70), blood (n = 35), lower respiratory tract (n = 17), wound (n = 29), stool (n = 8), upper respiratory tract (n = 5), catheters and other devices (n = 16), sputum (n = 5), bile (n = 2), decubitus (n = 1), CSF (n = 1) and environmental samples (n = 7).
Biochemical identification and susceptibility testing
Identification of the 196 isolates was carried out by the Micronaut E system (Genzyme Virotech GmbH, Ruesselsheim, Germany). Initial antibiotic susceptibility tests were performed by the Kirby–Bauer disc diffusion method in the local laboratories, according to the CLSI guidelines.7 The following antibiotic discs (Oxoid Ltd, Basingstoke, UK) were used: ceftazidime, cefotaxime, cefepime, ciprofloxacin, gentamicin, amikacin and tetracycline. E. coli ATCC 25922 was used as a control strain for antibiotic susceptibility tests. The MICs of antibiotics for 19 isolates selected on the basis of antibiogram, pulsotype, plasmid content, time and centre of isolation were determined by the Etest (AB Biodisk, Solna, Sweden), according to the manufacturer's instruction. The following Etests were used: ceftazidime (0.016–256 mg/L), cefotaxime (0.016–256 mg/L), cefepime (0.016–256 mg/L), gentamicin (0.016–256 mg/L), tobramycin (0.016–256 mg/L), amikacin (0.016–256 mg/L), ciprofloxacin (0.002–32 mg/L), tetracycline (0.016–256 mg/L), trimethoprim/sulfamethoxazole (0.002–32 mg/L), imipenem (0.002–32 mg/L) and cefoxitin (0.016–256 mg/L). The putative production of an ESBL was tested by Etest ESBL (AB Biodisk). K. pneumoniae ATCC 700603 was used as an ESBL-producing control strain, and it was included in all experiments.
Plasmid DNA analysis and transfer of resistance determinants
The mating assays and electroporation were carried out with 19 CTX-M-producing K. pneumoniae isolates selected according to the plasmid content, centre of isolation and pulsotype (PT), as described previously.10E. coli K12J5-3Rif and E. coli DH5α were used as recipients. For fingerprinting analysis, plasmid DNA from transconjugants or transformants was obtained by using a QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany), digested with PstI (Biolabs, Beverly, MA, USA) and detected by gel electrophoresis in 0.7% agarose at 110 V for 2 h.
PCR amplification and sequencing of the antibiotic resistance genes
The blaCTX-M, blaSHV, blaTEM, blaOXA-1-like, tet(A), tet(B), tet(C), aac-(3)-II, aac-(6′)-Ib, qnrA, qnrB and qnrS genes were detected and characterized as described previously.4–6,11–13 ISEcp1 was sought with the primer pair: 5′-GCA GGT CTT TTT CTG CTC C-3′ and 5′-TTT CCG CAG CAC CGT TTG C-3′.14 ISEcp1 associated with CTX-M-15 was screened by a combination of ISEcp1 forward primer and CTX-M universal reverse primer (5′-CGA TAT CGT TGG TGG TGC CAT A-3′).6 The blaCTX-M genes were amplified and sequenced using primers 5′-ATG GTT AAA AAA TCA CTG CG-3′ and 5′-CAG CGC TTT TGC CGT CTA AG-3′.6
Analysis of ciprofloxacin resistance by sequencing of the quinolone resistance-determining regions (QRDRs)
Alterations in the QRDRs of gyrA and parC encoding subunits of gyrase and topoisomerase IV enzymes in 10 selected isolates were determined by sequencing using primers described previously.15,16 The nucleotide substitutions resulting in amino acid changes were identified on the basis of the previously published sequences.15
The PFGE method was performed in line with the standardized CDC protocol.17 Gels were interpreted with Fingerprinting II Informatix Software (Bio-Rad). Levels of similarity were calculated with the Dice coefficient, and UPGMA (‘unweighted pair group method with arithmetic averages’) was used for the cluster analysis of the PFGE patterns. PTs were defined at 85% similarity between macrorestriction patterns and marked by letters according to the criteria established by Tenover et al.18 Clonally related isolates were supposed if they belonged to the same PT.
Multilocus sequence typing (MLST)
MLST with seven housekeeping genes was performed on 10 selected isolates according to Diancourt et al.19 Allele sequences and sequence types (STs) were verified at the http://pubmlst.org/kpneumoniae web site.
A total of 196 K. pneumoniae isolates were collected from six nosocomial outbreaks reported to the National Center for Epidemiology in 2005, as well as from related and sporadic cases. The isolates were obtained from clinical and screening samples of adult inpatients mainly from ICUs (n = 82) and medical (n = 34), traumatological (n = 18), surgical (n = 17) and urological wards (n = 15). The isolates were recovered in 36% from urine samples, 18% from blood, 15% from wound and 9% from lower respiratory tract.
Antibiotic resistance phenotypes
Based on the results of initial antibiotic susceptibility tests, all isolates were highly resistant to ceftazidime, cefotaxime and ciprofloxacin. Seventy-six percent proved resistant to gentamicin, 72% to tetracycline and 3% to amikacin.
Macrorestriction profile analysis
|Isolate||Centre ofisolation||Time ofcollection||Hospital ward||Specimen||Pulsotype||CAZ||CTX||FEP||GEN||TOB||AMK||CIP||TET||SXT||IPM||FOX|
|Isolate||Centre ofisolation||Time ofcollection||Hospital ward||Specimen||Pulsotype||CAZ||CTX||FEP||GEN||TOB||AMK||CIP||TET||SXT||IPM||FOX|
CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; GEN, gentamicin; TOB, tobramycin; AMK, amikacin; CIP, ciprofloxacin; TET, tetracycline; SXT, trimethoprim/sulfamethoxazole; IPM, imipenem; FOX, cefoxitin.
One hundred and twenty-nine isolates from 31 centres were assigned to PT N (the original PT of HEC), which comprised the major cluster. Of these strains, 70 ESBL-producing K. pneumoniae strains were isolated from three nosocomial outbreaks as follow: 45 isolates from a large nosocomial outbreak in the FJ2 centre (central Hungary), 15 isolates from a nosocomial outbreak in the BP18 centre (Budapest) and 10 isolates from a nosocomial outbreak in the BR1 centre (southern Hungary). The nosocomial outbreaks in hospitals FJ2 and BP18 progressed in two stages, in April and October and April and August 2005, respectively.
Forty-six isolates submitted from nine centres including those from two parallel nosocomial outbreaks in centres BP11 and BP14 (both in Budapest) comprised the second cluster designated PT R. Twenty-one isolates collected from four centres and including one nosocomial outbreak in centre BP14 (Budapest) comprised the third cluster designated PT S. Macrorestriction profiles, PTs, origin and distribution of 196 isolates in different centres are presented in Figure 2.
Plasmid profile analysis of isolates
All isolates harboured plasmids ranging from ∼2.0 to 230 kb grouped in distinct plasmid profiles. The plasmid profiles of isolates within PT N were diverse: 45 isolates collected from a nosocomial outbreak in the FJ2 centre carried a large plasmid of ∼140 kb in size, known from HEC from 2003, whereas the 15 isolates from the BP18 outbreak and the 10 isolates from the BR1 outbreak carried a plasmid of ∼90 kb in size. Another large plasmid of ∼90 kb in size was found in the 46 isolates assigned to PT R, whereas a plasmid of ∼50 kb was found in the 21 isolates assigned to PT S.
PCRs for CTX-M-, TEM-, SHV- and OXA-related β-lactamase genes
CTX-M-, TEM-, SHV- and OXA-specific PCRs were performed with 196 isolates. Positive results with CTX-M-, OXA- and SHV-specific primers were obtained for all the isolates. The blaTEM-specific amplicons were detected in isolates of EC III.
MICs for and CTX-M-types of selected isolates and their transconjugants or transformants
Conjugation studies were performed on 19 isolates. A plasmid of 140 kb from three HEC isolates, a plasmid of 90 kb from nine HEC isolates and another plasmid of 90 kb from four EC II isolates were transferred to the J53 recipient strain, whereas a plasmid of 50 kb from the three EC III isolates could be transformed to the DH5α recipient strain. Plasmids from the 19 transconjugants were digested with the PstI (Figure 3) restriction endonuclease. Identical or very similar fingerprints were obtained within individual ECs (EC II and EC III); however, restriction patterns of the two different plasmids originating from HEC, as well as of the plasmids from EC II and EC III, diverge from each other.
In the transconjugants, the cefotaxime and the tetracycline MICs remained similar to those of the donor strains. In contrast, MICs of the other examined cephalosporins and tobramycin for the transconjugants and transformants proved at least 10 times above those for the two recipient strains. The MICs of ciprofloxacin showed a slight elevation in the transconjugants and transformants, compared with the recipient strains (Table 1).
Transconjugants and transformants were screened by PCR for blaCTX-M, blaTEM and blaOXA genes and also for aac-(3)-II, aac-(6′)-Ib, tet(A), tet(B) and tet(C) genes, respectively (data not shown), and the products were sequenced. Sequencing revealed that plasmids from all three ECs equally harboured blaCTX-M-15, blaOXA-1, aac(6′)-Ib-cr and aac-(3)-II genes; the latter gene was not found in the transconjugants from gentamicin-susceptible isolates of HEC. Only the large non-conjugative plasmid of 50 kb—transformed from EC III—carried an additional blaTEM-1 gene and ISEcp1 associated with blaCTX-M-15. tet(A) and tet(C) genes were detected on plasmids of 140 and 90 kb originating from HEC.
The QRDRs of gyrA and parC genes were analysed in 10 isolates selected from the donor strains used in the conjugation experiments. Based on the alterations of nucleotide sequences in the QRDRs of GyrA and ParC in the HEC and EC III isolates, two amino acid changes (Ser-83→Phe and Asp-87→Ala) and one amino acid change (Ser-80→Ile) were detected, respectively. The EC II isolates exhibited a single nucleotide mutation in both gyrA (resulting in Ser-83→Ile substitution) and parC genes (resulting in Ser-80→Ile substitution). PCR screening for qnrA, qnrB and qnrS was also performed and showed negative results.
The 10 isolates and the HEC prototype isolate from 2003 were also subjected to MLST with seven housekeeping genes. Three distinct allelic profiles were obtained: ST 15 (allelic profile: 1-1-1-1-1-1-1) corresponding to HEC including the prototype isolate, ST 11 (allelic profile: 1-3-1-1-1-3-4) corresponding to EC III and the novel ST 147 (allelic profile: 4-3-6-1-7-4-38) corresponding to EC II (Figure 2). The analysis of STs by eBURST (http://pubmlst.org) showed that ST 15 is a single-locus variant of ST 14, ST 11 is a double-locus variant of ST 12 and ST 22, and ST 147 is a double-locus variant of ST 120.
After the first description of the HEC in 2003, a continuous monitoring system was implemented covering ciprofloxacin-resistant and CTX-M-producing K. pneumoniae isolates. Although in 2004 only 4 of the 183 ESBL-producing K. pneumoniae isolates submitted to the National Center for Epidemiology proved to be ciprofloxacin-resistant CTX-M-15 producers (I. Damjanova, Á. Toth, M. Jakab and J. János Topf, unpublished results), in 2005, an eruptive and extensive dissemination of ciprofloxacin-resistant CTX-M-15-producing K. pneumoniae was observed: 196 isolates, including those from six nosocomial outbreaks, were collected mainly from ICUs from 35 hospitals in 13 counties across Hungary.
PFGE analysis of all isolates revealed the existence of just three different genetic clusters named ECs: the HEC, EC II and EC III. The HEC could be subdivided into three main subclusters, indicating its relative heterogeneity. In contrast, EC II and EC III showed quite a homogeneous structure. The heterogeneity of HEC could be explained by its long-term and nationwide dissemination since 2003, whereas EC II and EC III were newly detected clones with a relatively small dissemination area.
Although PFGE analysis is an excellent tool for epidemiological typing of bacterial isolates, it does not provide comparable results for spatially and temporally non-related isolates. In contrast, the MLST technique that is based on indexing the genetic variation in housekeeping genes has been successfully employed for studying the longitudinal epidemiology of many bacterial species.20 Thus, MLST was chosen to confirm our PFGE results and to better understand the epidemiology and the population genetic structure of ESBL-producing K. pneumoniae in Hungary. It was found that the three genetic clusters assigned by PFGE really represented three different genetic clones. ST 11 (http://pubmlst.org/kpneumoniae) and ST 1519 were previously described in Europe, whereas ST 147 was a newly detected ST.
The interpretation of MLST results suggests a stable clonal structure for several K. pneumoniae strains. Interhospital ECs were detected, and the same ECs persisted for a long time in the same hospital. These findings suggest that the dissemination of CTX-M-15-type ESBLs among K. pneumoniae in Hungary has been confined primarily to ciprofloxacin-resistant stable bacterial clones exclusively in adult inpatients.
According to our investigation, ∼70% of the ESBL-producing K. pneumoniae isolates in 2005 belonged to the three large ECs (HEC, EC II and EC III). Moreover, the three ECs accounted for 36% of the nosocomial bloodstream infections caused by ESBL-producing K. pneumoniae in 2005.
It was assumed that ISEcp1 is a normal promoter for blaCTX-M-15 and plays a dominant role in its spread and mobility.11 Interestingly, in our study, ISEcp1 was found only on non-conjugative plasmids from EC III. Plasmids originating from HEC and EC II proved to be ISEcp1-negative with PCR, even though they were transferable at high frequencies and conferred high-level resistance to cefotaxime to transconjugants. A similar observation was made in the characterization of a blaCTX-M-15-carrying element in Cameroon, where ISEcp1 interrupted by IS26 consequently showed negative PCR results.21 Cloning experiments are planned to clarify the genetic environment of the blaCTX-M-15 gene in HEC and EC II.
The vast majority of isolates showed resistance to four antibiotic groups. Genes encoding resistance to cephalosporins, aminoglycosides and tetracycline coexisted on large mostly transferable plasmids. blaCTX-M-15, blaOXA-1 and aac(6′)-Ib-cr genes were equally detected on plasmids from all three ECs, whereas tet(A) and tet(C) genes were found exclusively on the two distinct large plasmids from HEC. A similar combination of resistance genes has been detected on R-plasmids from CTX-M-15-producing E. coli clonal strains A and D from the UK,11 as well as on multidrug resistance plasmids from CTX-M-15-producing E. coli clonally related strains from Canada, India, Kuwait, France, Switzerland, Portugal and Spain.22
Cumulative mutations in QRDRs of gyrA and parC are responsible for the high-level resistance to ciprofloxacin in all three ECs. Development of resistance to fluoroquinolones is a stepwise process,23 whereas higher fluoroquinolone MICs in Gram-negative bacteria need gyrA and parC mutations simultaneously.24,25 In the case of plasmid-mediated ESBL resistance, the process may be quite short. The exact process of CTX-M-15 acquisition and the source of the gene remain to be established.
We have previously shown5 that the high number of nosocomial outbreaks caused by SHV-type ESBL-producing K. pneumoniae seen in several Hungarian neonatal ICUs in 2002 and 2003 was the consequence of extensive and effective dissemination of identical allodemic R-plasmids. Conversely, the nationwide dissemination of cefotaximases and especially those of CTX-M-15 can be explained by the rapid and efficient expansion of a few K. pneumoniae ECs with surprisingly similar ‘resistance equipment’.26
As ESBL-producing klebsiellae, usually displaying resistance to ciprofloxacin, pose a serious nosocomial threat in Hungary, a pro-active infection control agenda was initiated in the affected hospitals.
Six of the seven K. pneumoniae-related nosocomial outbreaks reported to the National Center for Epidemiology in 2005 were caused by the three ECs. Such events illustrate the high epidemic potential of these ECs and their exceptional adaptation to the hospital environment. This phenomenal situation, in which 97% of the CTX-M-producing K. pneumoniae isolates detected across Hungary in 2005 were highly ciprofloxacin-resistant and represented just three stable genetic clones that persisted for a long time in the hospital settings, suggests a convergent population evolution in the K. pneumoniae species as in the case of the S. aureus species (MRSA versus MSSA).27 Are these ECs the new ‘MRSAs’?
No specific funding was received.
None to declare.
Results of this work were presented in part at the Seventeenth European Congress of Clinical Microbiology and Infectious Diseases, Munich, Germany, 2007 (O481). We are grateful to Mrs János Topf and Ms Andrea Torma for the antimicrobial susceptibility testing and conjugation. We thank all the contributing laboratories that provided isolates for this study.
- polymerase chain reaction
- electroconvulsive therapy
- antibiotic resistance, bacterial
- clone cells
- disease outbreaks
- electrophoresis, gel, pulsed-field
- klebsiella pneumoniae
- multi-antibiotic resistance
- extended-spectrum beta lactamases
- transfer technique
- epidural cortical stimulation