Molecular epidemiology of carbapenem resistant Enterobacteriaceae in Valle d’Aosta region, Italy, shows the emergence of KPC-2 producing Klebsiella pneumoniae clonal complex 101 (ST101 and ST1789)

Background The spread of carbapenem resistant Enterobacteriaceae (CRE) is an emerging clinical problem, of great relevance in Europe and worldwide. The aim of this study was the molecular epidemiology of CRE isolates in Valle d’Aosta region, Italy, and the mechanism of carbapenem resistance. Results Sixty consecutive CRE samples were isolated from 52 hospital inpatients and/or outpatients from November 2013 to August 2014. Genotyping of microbial isolates was done by pulsed-field gel electrophoresis (PFGE) and multi-locus sequence typing (MLST), carbapenemases were identified by PCR and sequencing. Carbapenem resistance gene transfer was performed by filter mating, plasmids from parental and transconjugant strains were assigned to incompatibility groups by PCR-based replicon typing. Molecular characterization of CRE isolates assigned 25 Klebsiella pneumoniae isolates to PFGE types A1-A5 and sequencing type (ST) 101, 17 K. pneumoniae isolates to PFGE type A and ST1789 (a single locus variant of ST101), 7 K. pneumoniae isolates to PFGE types B or C and ST512, 2 K. pneumoniae isolates to PFGE type D and ST405, and 5 Escherichia coli isolates to PFGE type a and ST131. All K. pneumoniae ST101 and ST1789 isolates were extended-spectrum beta-lactamase (ESBL) producers and carried blaCTX-M-1 group gene; 4 K. pneumoniae ST101 isolates were resistant to colistin. Molecular analysis of beta-lactamase genes identified blaKPC-2 and blaCTX-M-group 1 into conjugative plasmid/s assigned to IncFII incompatibility group in ST101 and ST1789 K. pneumoniae isolates, blaKPC-3 into conjugative plasmid/s assigned to IncF incompatibility group in ST512 and ST405 K. pneumoniae isolates, blaVIM-1 into conjugative plasmid/s assigned to IncN incompatibility group in ST131 E. coli isolates. Conclusions The spread of CRE in Valle d’Aosta region was caused by the selection of KPC-2 producing K. pneumoniae ST101 and ST1789 epidemic clones belonging to clonal complex 101, KPC-3 producing K. pneumoniae epidemic clones assigned to ST512 and ST405, and VIM-1 producing E.coli ST131 epidemic clone. Carbapenem resistance, along with blaKPC-2, blaKPC-3 and blaVIM-1 carbapenemase genes, was transferred by conjugative plasmids assigned to IncFII, IncF, and IncN incompatibility groups, respectively, in filter mating experiments. The emergence of colistin resistance was observed in KPC-2 producing K. pneumoniae ST101 isolates. Electronic supplementary material The online version of this article (doi:10.1186/s12866-015-0597-z) contains supplementary material, which is available to authorized users.


Background
The spread of resistance to carbapenems among Enterobacteriaceae has become a major public health problem. In fact, epidemics or endemic conditions of patients infected or colonized by Enterobacteriaceae non susceptible to carbapenems has been reported in the hospital and in the community setting [1][2][3]. Carbapenem resistant Enterobacteriaceae (CRE) are resistant to most β-lactams, including the carbapenems, and frequently carry additional antimicrobial resistance genes to other non-β-lactam antibiotics, which render them resistant to most antibiotics [1][2][3].
To control the spread of CRE in health care facilities and the transmission of CRE among patients, guidelines have been established from European Centre for Disease Prevention and Control for the screening and identification of CRE isolates among colonized patients [15], which have been adopted in Italy, both at regional and national level [16].
The aim of the current study was to analyze the molecular epidemiology of 60 consecutive CRE in the hospital wards and long-term care facilities of the USL Valle d' Aosta, and the mechanism of carbapenem resistance.

Setting and study period
Valle d' Aosta is the smallest region of Italy (3262 sq Kms), entirely located in the north western Alps region; there is one health agency and only one public Hospital, with 480 beds: there is one microbiology laboratory that processes samples from both Hospital and the territory. Moreover, a network of about 45 long-term care residencies for elders is distributed over the territory, accounting for another 1000 beds. The present study analyzed 60 consecutive isolates of CRE from 52 hospital inpatients and/or outpatients who were admitted to the USL Valle d'Aosta between November 2013 and August 2014.

Surveillance of CRE Enterobacteriaceae
Surveillance of CRE was performed according to European Centre for Disease Prevention and Control [15] and the Ministero della Salute, Italy [16]. Surveillance and infection control measures adopted for CRE positive patients included: rectal swab screening of both hospital inpatients and outpatients; isolation and cohorting of CRE positive hospital inpatients; recall and screening of CRE positive outpatients.

Bacterial strains and microbiological methods
All the Enterobacteriaceae isolates were identified using the Vitek 2 automatic system and the ID-GNB card for identification of Gram-negative bacilli according to the manufacturer's instructions (bio-Merieux, Marcy l'Etoile, France). Two Raoultella spp., and one Enterobacter spp. were further identified as R. planticola, R. ornithinolytica and E. aerogenes by amplification and sequencing of the rpoB gene as previously described [17].

Antimicrobial susceptibility testing
Carbapenem resistance of Enterobacteriaceae was screened using the meropenem disk alone as previously described [18]. Susceptibility tests were performed using the Vitek 2 system and the AST-GN card (bioMe'rieux). Breakpoints values were those recommended by the EUCAST 2014 [19]. Escherichia coli ATCC 25922 was used as quality control strain. ESBL activity was evaluated using the double-disk synergy test as previously described [20]. Colistin susceptibility was evaluated by broth microdilution in Mueller-Hinton broth II (MHBII) according to Clinical and Laboratory Standards Institute guidelines [21].

Molecular analysis of antimicrobial resistance genes
All the isolates were subjected to polymerase chain reaction (PCR) to screen for the presence of bla KPC , bla IMP , bla VIM , bla NDM and bla OXA-48 , as described previously [22]. Acinetobacter baumannii AC 54/97 [23], A. baumannii 161/07 [24] and Klebsiella pneumoniae D001 [18] were used as positive quality control strains for bla IMP-2 , bla NDM-1 and bla OXA-48 carbapenemase genes, respectively. The presence of the correct size PCR product was confirmed by agarose gel electrophoresis. The full-length alleles of bla KPC and bla VIM were amplified using primers KPC-F2: GTATCGCCGTCT AGTTCTGC and KPC-Rm: CTTGTCATCCTTGTTA GGCG, primers 5'VIM1: ATGTTAAAAGTTATTAGT AGTT and 3'VIM1: CTACTCGGCGACTGAGCGATT, respectively. Direct sequencing of PCR products was performed using the ABI Prism BigDye Terminator v3.1 Ready Reaction cycle sequencing kit and the 3730 DNA analyzer (Applied Biosystems, Foster City, CA). Nucleotide sequences obtained were compared with the deduced amino acid sequence available in GenBank using tblastx [25]. The presence of bla CTX-M extendedspectrum beta-lactamase genes was investigated by multiplex PCR as described previously [26]. Escherichia coli SGH 20-ADM was used as positive quality control strain for bla CTX-M -15 gene [27]. The presence of the armA methylase gene was investigated by PCR using specific primers as described previously [28].

PFGE typing and dendrogram analysis
XbaI DNA macrorestriction of Enterobacteriaceae isolates was performed as previously reported [29]. PFGE conditions were as follows: pulse times ranged from 5 to 40 s over 24 h at6 · 0 V/cm and at 14C. The PFGE profiles obtained were converted to TIFF files and subjected to cluster analysis using the GelCompare II v. 4.6 software package (AppliedMaths, Sint-Martens-Latem, Belgium). Clustering was based on the unweighted pairgroup method with arithmetic averages (UPGMA). The Dice correlation coefficient was used to analyze the similarities of the banding patterns with a tolerance of 1.5 %. The interpretation of chromosomal DNA restriction patterns was based on the criteria of Tenover et al. for closely related isolates [30]. Briefly, strains showing more than three DNA fragment variations and a similarity of <85 % at dendrogram analysis were considered to represent different PFGE types, while one to threefragment differences and a similarity of >85 % upon dendrogram analysis were considered to represent PFGE pattern subtypes.

MLST analysis
MLST analysis of K. pneumoniae isolates was performed using the Institut Pasteur's MLST scheme [31]. The allele sequences and sequence types (STs) were identified at http://bigsdb.web.pasteur.fr/klebsiella/klebsiella.html. MLST analysis of E. coli isolates was performed as previously described [32,33], with the following minor modifications for annealing temperatures of the primers, respectively 54°C for adk, fumC and icd genes, 58°C for recA and mdh genes, 60°C for purA gene, and 64°C for gyrB gene. Allele sequence and MLST profile definitions were identified at http://mlst.warwick.ac.uk/mlst/dbs/Ecoli.

Conjugative transfer of carbapenem resistance and plasmid typing
Conjugation experiments were performed by filter mating experiments as previously described [34] using sodium-azide resistant E. coli J53 [35] as recipient strain on BHI agar plates in the presence of 16 mg/l imipenem and 100 mg/l sodium-azide. The frequency of transfer was calculated as the number of transconjugants divided by the number of surviving recipients. Plasmids from parental and transconjugant strains were assigned to incompatibility groups by PCR-based replicon typing (PBRT-kit) (Diatheva s.r.l., Fano, Italy) performed on total DNA using previously described primers and conditions [36,37].

Ethical approval
The study was approved by the ethics committee of the Aosta Regional Hospital (protocol number 836/2015). All microbiological samples were taken as part of standard care procedures. No written informed consent was necessary for this type of study.

CRE isolates included in the study
The molecular epidemiology of carbapenem-resistant Enterobacteriaceae was studied in the Valle d' Aosta region from November 2013 to August 2014. During this period, 60 consecutive CRE were isolated from 52 different patients, both from clinical samples and from surveillance screening. Of these patients, 6 were outpatients, 4 were admitted to nursing homes, while the remaining 42 were admitted to hospital's internal wards (in particular, 19 in geriatrics, 4 in intensive care, 4 in neurology, 10 in medical departments and 5 in surgical wards). The isolated materials were 32 urine culture, 20 rectal swabs, 3 respiratory secretions, 2 blood culture, 1 liquid drainage, 1 peritoneal-fluid, and 1 wound swab. Fifty-one isolates were identified as K. pneumoniae, 6 were E. coli, 1 R. planticola, 1 R. ornithinolytica, and 1 E. aerogenes (Figs. 1 and 2 and Additional file 1: Table S1).

Antimicrobial susceptibility patterns of Enterobacteriaceae isolates
Both K. pneumoniae and E. coli isolates showed a multidrug-resistant antibiotype. In particular, they were resistant to imipenem, meropenem, ertapenem, beta-lactam/ beta-lactamase inhibitor combinations (clavulanic acid/ amoxicillin, piperacillin/tazobactam), third and fourth generation cephems, fluoroquinolones (ciprofloxacin), and trimethoprim-sulfamethoxazole; 38 K. pneumoniae isolates were resistant but 13 K. pneumoniae isolates were susceptible to aminoglycosides (amikacin and gentamicin); 46 K. pneumoniae isolates were resistant but 5 K. pneumoniae isolates were susceptible to fosfomycin; all E. coli isolates were susceptible to aminoglycosides and fosfomycin. Also, 42 K. pneumoniae isolates were identified as ESBL producers (Table 1 and Additional file 1:  Table S1). Interestingly, all E. coli isolates and the majority of K. pneumoniae isolates were susceptible to colistin, but 4 K. pneumoniae isolates were resistant to colistin, minimal inhibitory concentrations (MICs) 16-125 mg/l (Table 1 and Additional file 1: Table S1). Raoultella spp and E. aerogenes isolates were resistant to carbapenems, beta-lactam/beta-lactamase inhibitor combinations, third and fourth generation cephems, fluoroquinolones and trimethoprim-sulfamethoxazole but susceptible to aminoglycosides, fosfomycin and colistin (Additional file 1: Table S1).

Molecular analysis of antimicrobial resistance genes of CRE isolates
PCR and sequence analysis of carbapenemase genes identified bla VIM-1 in 5 E.coli isolates, bla KPC-2 in 42 K. pneumoniae and 1 R. planticola isolates, and bla KPC-3 in 9 K. pneumonia isolates. No carbapenemase genes were identified in 1 R. ornithinolytica, 1 E. coli, and 1 E. aerogenes isolate (Figs. 1 and 2 and Additional file 1: Table S1). Also, bla CTX-M-1 group gene was amplified from all 42 ESBLproducing K. pneumoniae isolates; armA gene was amplified from all 38 K. pneumoniae isolates resistant to aminoglycosides but not from 13 K. pneumoniae isolates susceptible to aminoglycosides (Additional file 1: Table S1).

Molecular epidemiology of CRE isolates in the USL Valle d'Aosta
To investigate whether the isolation of CRE during the study period was caused by the spread of epidemic clones, all 60 CRE isolates were genotyped. Molecular typing by PFGE identified four major PFGE types in 51 K. pneumoniae isolates, which we categorized from A to D. Of the 42 K. pneumoniae isolates assigned to major PFGE type A, 17 isolates showed an identical macrorestriction pattern (PFGE type A), while 25 strains differed in the migration of one or two DNA fragments and were classified into subtypes A1-A5 (Fig. 1). PFGE types E and F were assigned to R. ornithinolytica and R. planticola isolates, respectively (data not shown). Also, PFGE analysis identified identical macrorestriction pattern (PFGE type "a") in 5 E. coli isolates and PFGE types "b" and "c", which differed in the migration of more than 3 bands, in 1 E. coli and 1 E. aerogenes isolate, respectively ( Fig. 2). Genotype analysis using MLST identified ST101 for K. pneumoniae isolates assigned to PFGE subtypes A1-A5 and a novel ST1789 for K. pneumoniae isolates assigned to PFGE type A. Because ST101 and ST1789 were single locus variants, they were clustered in the same clonal complex (CC) 101 by eBURST analysis. On the contrary, MLST identified ST512 for K. pneumoniae isolates assigned to PGFE types B and C and ST405 for K. pneumoniae isolates assigned to PGFE type D. ST512 and ST405 differed by at least two loci from all other profiles found in the population under analysis and were regarded as singletons (Fig. 1). Based on all the above data, four distinct epidemic genotypes were identified in K. pneumoniae isolates, which we named ST1789/A, Trimethoprim-Sulfamethoxazole ≤20 ≥320 ≤20 -≥ 320 ≥320 ≥320 ≤20 -≥ 320 Colistin ≤0,5 ≤0,5 ≤0,5 -125 ≤0,5 ≤0,5 ≤0,5 a MIC, minimal inhibitory concentration ST101/A1-A5, ST512/B-C, and ST405/D, respectively. MLST analysis identified ST131 for all E. coli isolates assigned to PFGE type "a" and subtype "a1", which we named ST131/a epidemic genotype (Additional file 1: Table S1). All K. pneumoniae isolates assigned to ST101/ A1-A5 and ST1789/A epidemic genotypes were ESBL producers and carried bla CTX-M-1 group gene. Also, 23 of 25 K. pneumoniae isolates assigned to ST101/A1-A5 epidemic genotype, 8 of 17 K. pneumoniae isolates assigned to ST1789/A epidemic genotype, and all K. pneumoniae isolates assigned to ST512/B-C epidemic genotype carried armA gene and were resistant to aminoglycosides (Tables S1). Four K. pneumoniae isolates assigned to ST101/A1-A5 epidemic genotype were resistant to colistin (Additional file 1: Table S1). To further study the mechanism responsible for carbapenem resistance, we asked whether carbapenem resistance might have been transferred through conjugation. Resistance to imipenem and meropenem (MIC ≥16 mg/l) and ESBL activity along with bla KPC-2 and bla CTX-M-group 1 was transferred from K. pneumoniae isolates assigned to genotypes ST1789/A and ST101/A1 to E. coli J53 strain at frequencies ranging from 8.3 × 10 −4 to 1.4 × 10 −6 cfu/recipient cells. Also, resistance to imipenem and meropenem (MIC ≥16 mg/l) along with bla KPC-3 or bla VIM-1 was transferred from K. pneumoniae isolates assigned to genotypes ST512/B, ST512/C, ST405/D and E. coli isolate assigned to genotype ST131/a, respectively, to E. coli J53 at frequency ranging from 2.5 × 10 −2 to 2.5 × 10 −7 cfu/recipient cells. All transconjugants were resistant to sodium-azide and showed a PFGE profile identical with that of the recipient strain (Table 2). Also, we asked whether conjugative plasmids were responsible for the horizontal transfer of carbapenem resistance. PCR-based replicon typing identified FII replicon and IncFII incompatibility group plasmid/s in ESBL positive and carbapenem-resistant K. pneumoniae isolates and transconjugants producing bla KPC-2 and bla CTX-M-group 1 genes, while FIIk replicon and IncF incompatibility group plasmid/s in K. pneumoniae isolates and transconjugants producing bla KPC-3 gene. Moreover, the E. coli donor strain expressing bla VIM-1 gene and its transconjugant were shown to carry plasmid/s with N replicon and IncN incompatibility group (Table 2).

Discussion
In the present report, we studied the molecular epidemiology and the genetic basis of carbapenem resistance in CRE isolates in the USL Valle d' Aosta between November 2013 and August 2014. In accordance with previous data [1-7, 9, 10], the spread of CRE in Valle d' Aosta region was caused by the selection of few epidemic distinct genotypes of KPC-2 producing K. pneumoniae, KPC-3 producing K. pneumoniae, or VIMproducing E. coli. The majority of the isolates (42 of 60) were assigned to K. pneumoniae ST101/A1-A5 and ST1789/A genotypes, which belonged to the same CC101 and carried bla KPC-2 into IncFII conjugative plasmid/s. KPC-2 producing K. pneumoniae ST101/A1-A5 were isolated from patients in geriatrics, neurology, other medical wards, and from outpatients; KPC-2 producing K. pneumoniae ST1789/A were isolated from patients in geriatrics, medical wards, nursing homes, and one outpatient. The isolation of identical genotypes of CRE in long-term care facilities and adjacent hospital wards is in agreement with a previous study showing the isolation of identical carbapenemase among Enterobacteriaceae in an Italian tertiary-care hospital and adjacent long-term care facilities [8]. KPC-2 producing K. pneumoniae ST101 has already been reported as sporadic isolation in Brazil [38] and sporadic or epidemic isolation in Italy [9,[39][40][41], while KPC-2 producing K. pneumoniae ST1789 and nosocomial outbreak caused by KPC-2 producing K. pneumoniae ST101 and ST1789 belonging to CC101 were described for the first time herein. Also, sporadic isolation of KPC-3 producing K. pneumoniae ST101 has been reported in Southern Italy [42], thus suggesting that bla KPC-2 might have been acquired through horizontal gene transfer. In further support of this hypothesis, we demonstrated herein that K. pneumoniae ST101 and ST1789 isolates carried bla KPC-2 into IncFII conjugative plasmid/s. We also observed the spread of KPC-3 producing K. pneumoniae ST512/ B and ST512/C genotypes. Because K. pneumoniae ST512 is a single locus variant of ST258, KPC-3 producing K. pneumoniae ST512 isolates can be assigned to the multidrug-resistant high risk CC-258, which was responsible of several epidemics in Europe and Italy [6,9,11,42]. Our data showed also the spread of VIM-1 producing E. coli ST131 among CRE isolates. This is consistent with previous studies showing the emergence of hypervirulent E. coli ST131 clonal lineage [6] and VIM-like producing CRE in Europe and Northern Italy [4,[6][7][8]10]. Several studies demonstrate that epidemic clones of CRE are selected in the community and hospital setting because of their multiple antimicrobial resistances [1][2][3][4][5][6][7][8][9][10][11]. The presence of carbapenemases and the acquisition of carbapenemases through horizontal gene transfer have been frequently reported [1][2][3][4][5][6][7]. Accordingly, our data showed that carbapenem resistance, along with bla KPC-2 , bla KPC-3 and bla VIM-1 carbapenemase genes carried by conjugative plasmids assigned to IncFII, IncF, and IncN incompatibility groups, respectively, was transferred by filter mating experiments. Interestingly, bla KPC-3 carried by conjugative plasmid/s assigned to IncF incompatibility group was found in both K. pneumoniae ST512 and ST405 epidemic genotypes isolated in the study. Our data are in agreement with previous studies showing bla VIM-1 carbapenemase into conjugative plasmids assigned to IncN incompatibility group [43], bla KPC-2 into conjugative plasmids assigned to IncFII incompatibility  [5], bla KPC-3 into conjugative plasmids assigned to IncF incompatibility group [37]. However, the presence of bla KPC-2 carbapenemase into conjugative plasmids assigned to IncF incompatibility group has been described also [41]. The above all data suggest that conjugation was responsible for carbapenem resistance acquisition in CRE included in the study. The analysis of antimicrobial resistance profiles of CRE epidemic genotypes showed that they were microbialdrug resistant. In particular, we identified CTX-M1 group ESBL in KPC-2 producing ST101 and ST1789 K. pneumoniae isolates and corresponding transconjugants. In agreement with our data, the coexistence of KPC-2 carbapenemase and CTX-M-15 group 1 ESBL has been demonstrated [12][13][14]. Also, all 7 K. pneumoniae isolates assigned to ST512, 23 out of 25 isolates assigned to ST101, and 8 out 17 isolates assigned to ST1789 epidemic genotypes were resistant to aminoglycosides and showed the presence of armA gene (Additional file 1: Table S1). This is consistent with previous studies showing the isolation of extensively drug-resistant ArmA-and KPC-2-producing ST101 K. pneumoniae isolates in Italy [40,41]. An interesting finding of our study was the occurrence of colistin resistance in four KPC-2 producing K. pneumoniae isolates assigned to ST101/A2-A4 genotypes. Our data are in agreement with recent studies showing the emergence of colistin resistance in KPC-2 producing K. pneumoniae isolates from the Netherlands [44] and KPC-2 or KPC-3 producing K. pneumoniae isolates from Italy [42,45,46]. Because KPC-2 or KPC-3 producing colistin resistant K. pneumoniae isolates from Italy and the Netherlands belong to ST258 or ST512 single locus variant [42,[44][45][46], the data shown herein suggest that colistin resistance may develop in different genetic backgrounds.

Conclusions
The spread of CRE in Valle d'Aosta region was caused by the selection of KPC-2 producing K. pneumoniae ST1789 and ST101 isolates, which were single locus variants and could be assigned to CC101. KPC-2 producing K. pneumoniae CC101 were isolated from patients in geriatrics and other hospital medical wards and in outpatients in the USL Valle d' Aosta. KPC-3 producing K. pneumoniae ST512 and ST405, and VIM-1 producing E.coli ST131 were isolated also. Carbapenem resistance, along with bla KPC-2 , bla KPC-3 and bla VIM-1 carbapenemase genes, was transferred by conjugative plasmids assigned to IncFII, IncF, and IncN incompatibility groups, respectively, in filter mating experiments. The emergence of colistin resistance was observed in KPC-2 producing K. pneumoniae ST101 isolates.

Additional file
Additional file 1: Table S1