Intercontinental spread of OXA-48 beta-lactamase- producing Enterobacteriaceae over a 11-year period, 2001 to 2011

OXA-48 beta-lactamase producers are emerging as an important threat mostly in the Mediterranean area. We report here the molecular epidemiology of a collection of OXA-48 beta-lactamase-positive enterobacterial isolates (n=107) recovered from European and northAfrican countries between January 2001 and December 2011. This collection included 67 Klebsiella pneumoniae, 24 Escherichia coli and 10 Enterobacter cloacae. Using the EUCAST breakpoints, ninety-eight isolates (91.6%) were of intermediate susceptibility or resistant to ertapenem, whereas 66% remained susceptible to imipenem. Seventy-five per cent of the isolates coproduced an extended-spectrum beta-lactamase, most frequently CTX-M-15 (77.5%). Susceptibility testing to non-beta-lactam antibiotics showed that colistin, tigecycline, amikacin, and fosfomycin remain active against most of the isolates. Multilocus sequence typing indicated that the most common sequence types (ST) were ST101 and ST38 for K. pneumoniae and E. coli, respectively. The blaOXA-48 gene was located on a 62 kb IncL/M plasmid in 92.5% of the isolates, indicating that a single plasmid was mainly responsible for the spread of that gene. In addition, this study identified multiple cases of importation of OXA-48 betalactamase producers at least in Europe, and spread of OXA-48 beta-lactamase producers giving rise to an endemic situation, at least in France.


Introduction
Currently, an emergence of carbapenem resistance in Enterobacteriaceae is reported, mostly related to the spread of carbapenemases [1]. Those carbapenemhydrolysing beta-lactamases belong to the Ambler class A (e.g. KPC), class B (e.g. IMP, VIM and NDM) [1], and class D (e.g. OXA-48 and its variants possessing weaker but significant carbapenemase activity) [2]. OXA-48 had first been identified from a clinical Klebsiella pneumoniae isolate recovered in Istanbul, Turkey, in 2001 [3]. The corresponding gene, namely bla  , was then also identified in Escherichia coli and Citrobacter freundii, still in Turkey [4]. For several years, OXA-48 was identified only in Turkey, and almost all OXA-48 beta-lactamase producers were reported from patients hospitalised in Turkey or with a link to that country [4,5]. Since 2008, this gene has been identified in many other countries, most often in K. pneumoniae isolates [2,[5][6][7][8][9][10]. OXA-48 is now identified in the Middle East and in North African countries, and those countries are considered as reservoirs of OXA-48 beta-lactamase producers [2]. In addition to sporadic cases, an increasing number of outbreaks due to OXA-48-producing K. pneumoniae are currently observed, not only in Turkey but also in Belgium, France, Greece, the Netherlands and Spain [2,[11][12][13]. K. pneumoniae strains belonging to specific sequence types (ST), such as ST395 and ST101, have been involved in those outbreaks [12,14].
In order to gain further understanding of that phenomenon, our study aimed at comparing the genetic features of OXA-48 beta-lactamase-producing strains recovered from various countries by analysing an existing collection of 107 bla OXA-48 -positive enterobacterial isolates. The genetic context and the location of the bla OXA-48 gene were investigated, as well as resistance to broad-spectrum beta-lactams and non-beta-lactam antibiotics.

Bacterial isolates
A total of 107 OXA-48 beta-lactamase-producing enterobacterial isolates were investigated retrospectively. Enterobacteriacae producing OXA-48-like beta-lactamases were not included in this study. All isolates had been recovered from clinical specimens except a single isolate (one Serratia marcescens strain from an environmental water sample in Morocco), and had been received between January 2001 and December 2011 in our National Reference Laboratory which is also used as an International Reference Laboratory by many colleagues worldwide who send us their isolates for further characterisation. Of identical strains in an outbreak, only one was included in this work. The distribution of clinical samples was as follows: rectal swabs (n=33), urine samples (n=24), blood samples (n=12), wound samples (n=7), respiratory specimens (n=4), catheters (n=4), bone specimens (n=2), peritoneal fluids (n=2), and placenta specimen (n=1). One sample per patient was included. Detailed information could not be obtained for 18 clinical samples. The isolates were identified to species level using the API 20E system (bioMérieux, La Balme-les-Grottes, France).

Susceptibility testing
Routine antibiograms were determined by disk diffusion method on Mueller-Hinton (MH) agar (Bio-Rad, Marnes-la-Coquette, France) and interpreted using the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (updated 2012) and of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for tigecycline and colistin [15,16]. In addition, MICs were determined for imipenem, meropenem, ertapenem, cefotaxime, and ceftazidime using E-test (bioMérieux, La Balme-les-Grottes, France). The production of extended-spectrum beta-lactamases (ESBL) was evidenced by a double-disk synergy test performed with cefepime, ceftazidime, and ticarcillin/clavulanic acid disks [17] and more recently by using the rapid ESBL NDP test [18].

PCR and sequencing of betalactamase-encoding genes
Whole-cell DNA was extracted using the QiaAmp DNA minikit and following the manufacturer's recommendations (Qiagen, Courtaboeuf, France). All isolates were screened by PCR for the Ambler class A and B carbapenemase-encoding genes bla KPC , bla IMP , bla VIM , bla NDM [19][20]. For each isolate, the bla OXA-48 gene was amplified using primers preOXA-48A and preOXA-48B, and subsequently sequenced [21]. Detection of other betalactamase genes such as bla TEM , bla SHV , bla CTX-M , bla AmpClike, and bla OXA-1 was performed with internal primers, as described previously [19,22]. PCR products were analysed on agarose gel and sequenced by using the amplification primers with an automated sequencer (ABI PRISM 3100; Applied Biosystems). The nucleotide and deduced protein sequences were analysed using software from the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov).

Strain typing
Multilocus sequence typing (MLST) with seven housekeeping genes (rpoB, gapA, mdh, pgi, phoE, infB and tonB) was performed for K. pneumoniae isolates according to Diancourt et al. [23]. Allele sequences and STs were verified at http://pubmlst.org/Kpneumoniae. Fragments of seven housekeeping genes (adk, fumC, gyrB, icd, mdh, purA and recA) were amplified and sequenced for E. coli isolates as described on the following website http://mlst.ucc.ie/mlst/dbs/Ecoli. A different allele number was given to each distinct sequence within a locus, and a distinct ST number was attributed to each distinct combination of alleles. E. coli isolates were assigned to the major E. coli phylogenetic groups (A, B1, B2, and D) by multiplex PCR, as described [24]. The genetic relationship between the Enterobacter cloacae isolates was studied using Diversilab, a semi-automated typing system based on repetitive sequence-based PCR (rep-PCR) following the manufacturer's instructions (bioMérieux).

Plasmid DNA analysis, transformation and mating-out assays
Plasmid DNA was extracted from the isolates using the Kieser technique [25]. E. coli NCTC50192, harbouring four plasmids of 154, 66, 48 and 7 kb, was used as plasmid size marker. Plasmid DNAs were analysed by agarose gel electrophoresis. Direct transfer of the carbapenem resistance markers was attempted by liquid mating-out assays at 37°C, using E. coli J53 as recipient, or by electrotransformation of plasmid DNA, using E. coli TOP10 as recipient as reported [3,4]. Selection was performed on agar plates supplemented with ertapenem (0.5 µg/ml) and azide (100 µg/ml) for mating-out assays. In order to search for a possible chromosomal location of the bla OXA-48 gene in E. coli isolates 19 to 24, restriction with endonuclease I-CeuI followed by pulsed-field gel electrophoresis (PFGE) analysis was performed as described [26].

Replicon and transposon typing
PCR-based replicon typing (PBRT) of the main plasmid incompatibility groups reported in Enterobacteriaceae was performed as described [27] and using the specific primers designed from plasmid pOXA-48a [28]. Genetic structures surrounding the bla OXA-48 gene were determined according to the Tn1999-like PCR-mapping scheme as described [29].

Table 1a
Genetic features associated with OXA-48 beta-lactamase producers
The results in this Table are from E-tests.

Phylogenetic groups of the Escherichia coli isolates
More than half of E. coli isolates belonged to the phylogenetic group D (14 of the 24 E. coli isolates), seven E. coli isolates belonged to the phylogenetic group A, two belonged to the phylogenetic group B2, and one isolate belonged to the phylogenetic group B1 (Table 1).

Mulilocus sequence typing
The distribution of the sequence types among the K. pneumoniae and E. coli isolates is shown in Figures  1 and 2, respectively. ST101 was the most commonly observed ST for the K. pneumoniae isolates, accounting for 17 out of 67 isolates (25.4%), followed by ST395 and ST15 (7 isolates, 10.5%) (Figure 1). Six isolates (9%) belonged to ST147 (9%) and the other isolates to diverse STs, namely ST14 (n=4), ST45 (n=4), ST25 (n=2), ST392 (n=2), and one to other STs (Figure 1 (Figure 2). Since no MLST system has been developed for typing the E. cloacae species, these isolates were genotyped using the DiversiLab method. E. cloacae 7 and 9 recovered from Morocco were closely related, and E. cloacae 1, 2 and 5 (also from Morocco) belonged to the same cluster. The other E. cloacae isolates were distinct (data not shown).

Genetic location the bla OXA-48 gene
Using the specific primers designed from the reference plasmid pOXA-48a of K. pneumoniae 11978 [27] to amplify its replicase gene

Genetic environment of the bla OXA-48 gene
The bla OXA-48 gene was flanked by two copies of IS1999. In 21 isolates (19.6%), the upstream copy remained intact. This structure corresponded to transposon Tn1999, whereas 84 isolates (78.5%) had a Tn1999.2 transposon structure in which the IS1999 is disrupted by insertion of an IS1R element [4]. In two isolates ESBL: extended-spectrum beta-lactamases.
a The Providencia rettgeri and the Serratia marcescens isolates were excluded because of their natural resistance to colistin.
The percentages are rounded so as to add up to 100%.

Discussion
We have analysed here many different features of 107 known OXA-48-positive enterobacterial isolates which are widely distributed at least in several European and North African countries, and also in Turkey.
Noticeably, 25% of the OXA-48 beta-lactamase producers remained susceptible to broad-spectrum cephalosporins, which therefore present possible therapeutic options. At least positive therapeutic outcomes have been obtained using an animal model of infection and broad-spectrum cephalosporins [33]. Those ESBLnegative isolates were most often susceptible to the other classes of antibiotics, which is in line with the fact that the epidemic plasmid encoding the bla OXA-48 gene does not carry additional resistance determinants [28].
However, 75% of the OXA-48-positive isolates in our study harboured an additional ESBL-encoding gene that confers resistance to broad-spectrum cephalosporins. We have recently reported the genetic association of the bla CTX-M-15 and bla OXA-48 genes on the same transposon, indicating that this combination of multidrug-resistance genes may spread further in the future [32]. In addition, most of those ESBL-producing isolates were resistant to non-beta-lactam antibiotics, due to other resistance mechanisms. It is worth mentioning that 70 isolates (65%) were susceptible to imipenem and meropenem according to CLSI guidelines, further complicating the detection of OXA-48-producing isolates in laboratories. Conversely, most isolates showed intermediate susceptibility or resistance to ertapenem. Ertapenem may thus be the most appropriate carbapenem molecule for detecting OXA-48 producers. Therefore, a selective medium containing ertapenem has recently been developed for the detection of all types of carbapenemase producers including the OXA-48 beta-lactamase producers [34]. Taking into account the fact that 75% of the OXA-48 isolates were ESBL producers and the level of resistance to non-beta-lactam molecules, treatment options for infections caused by OXA-48 beta-lactamase producers may be limited. The efficacy of carbapenems in treating infections due to OXA-48 beta-lactamase producers with susceptibility or low-level resistance to several carbapenems remains debatable, because carbapenems have been shown to be an inefficient therapy for treating mice with induced peritonitis caused by an OXA-48-producing K. pneumoniae [33]. Also, imipenem-containing therapy failed to treat several OXA-48 infections in humans [4,11]. A single report described imipenem as efficient treatment against bacteraemia due to an OXA-48 K. pneumoniae isolate [35]. Controlled trials are needed to evaluate the real clinical efficacy of carbapenems in treating infections due to OXA-48 beta-lactamase producers.
The clonal distribution of OXA-48 beta-lactamasepositive isolates is interesting because a quarter of the K. pneumoniae isolates belonged to ST101. OXA-48-positive K. pneumoniae isolates belonging to ST101 have recently been implicated in an outbreak in Spain, and have also been detected in Tunisia [11,12]. We report here that the ST101 isolates were recovered from Tunisia, Morocco, and from South Africa and France from patients who did not travel abroad, suggesting that this ST has now widely spread in European countries and in Africa. Seven K. pneumoniae isolates belonged to ST395, a ST implicated in clonal outbreaks in Europe [11,14]. Interestingly, we detected seven ST15 among K. pneumoniae isolates recovered from patients who had a link with Morocco. That sequence type corresponds to an internationally occurring clone and has been associated with different ESBL genes, but also with the metallo-beta-lactamase genes coding for NDM and VIM [36,37]. The occurrence of OXA-48 betalactamase in a ST15 K. pneumoniae isolate had been reported only once, in 2012, in an isolate from Finland [38]. Those data are likely to indicate that a novel OXA-48 K. pneumoniae clone belonging to ST15 may emerge in Morocco. K. pneumoniae isolates belonging either to ST392 or ST147 (differing at a single locus) were identified in a total of eight isolates, with the two ST392 collected in Morocco and the six ST147 collected in Belgium, Turkey and France, and also from patients originating from Tunisia or Libya. This result highlights the dissemination of another OXA-48-producing clone, mainly in the Mediterranean area. The other K. pneumoniae isolates belonged to diverse ST, supporting the hypothesis of the widespread dissemination of a single bla OXA-48 -positive IncL/M plasmid among various genetic backgrounds. Overall, there is no association between ST type and ESBL type among OXA-48 producers.
Among the 24 E. coli isolates, seven were of ST38, showing that this clone is widely disseminated, as previously suggested [7,39]. Interestingly, the bla OXA-48 gene was chromosomally located in six of those isolates, as was speculated for the ST38 E. coli isolates recovered in the United Kingdom [7]. Such chromosomal location of the bla OXA-48 gene in E. coli may be associated to a lower level of resistance (a single gene copy). The other 17 E. coli isolates were genetically distinct. Furthermore, it is interesting to note that 16 of the 24 OXA-48-positive E. coli belonged to phylogenetic group D or B2, which mainly include virulent strains. The E. cloacae isolates were overall clonally diverse.
As suggested previously, the bla OXA-48 gene was located on a 62 kb IncL/M plasmid in most of our isolates (n=99, 92.5%), indicating that current spread of OXA-48 betalactamase producers is mainly related to the diffusion of this plasmid. The dissemination of the bla OXA-48 gene is also associated with the spread of different clones.
Interestingly, 20% of OXA-48-producing isolates collected in France were considered to be autochthonous, indicating that the bla OXA-48 gene has already spread in the community in France. This latter result indicates ongoing diffusion of OXA-48-type genes in Europe.