KPC-3-Producing Klebsiella pneumoniae in Portugal Linked to Previously Circulating Non-CG258 Lineages and Uncommon Genetic Platforms (Tn4401d-IncFIA and Tn4401d-IncN)

KPC-3-producing bacteria are endemic in many countries but only recently became apparent their wide distribution in different Portuguese hospitals. The aim of this study is to characterize genetic backgrounds associated with blaKPC−3 among Klebsiella pneumoniae isolates recently identified on non-hospitalized patients in Portugal. Twenty KPC-producing K. pneumoniae identified between October 2014 and November 2015 in three different community laboratories were characterized. Isolates were mainly from patients from long-term care facilities (n = 11) or nursing homes (n = 6), most of them (75%) previously hospitalized in different Portuguese hospitals. Standard methods were used for bacterial identification and antibiotic susceptibility testing. Carbapenemase production was assessed by the Blue-Carba test, and identification of bla genes was performed by PCR and sequencing. Epidemiological features of KPC-producing K. pneumoniae included population structure (XbaI-PFGE, MLST and wzi sequencing), genetic context (mapping of Tn4401), and plasmid (replicon typing, S1-PFGE, and hybridization) analysis. All K. pneumoniae isolates produced KPC-3, with two MDR K. pneumoniae epidemic clones representing 75% of the isolates, namely ST147 (wzi64/K14.64, February–November 2015) and ST15 (two lineages exhibiting capsular types wzi19/K19 or wzi93/K60, July-November 2015). Other sporadic clones were detected: ST231 (n = 3; wzi104), ST348 (n = 1; wzi94) and ST109 (n = 1, wzi22/K22.37). blaKPC−3 was identified within Tn4401d in all isolates, located in most cases (80%) on cointegrated plasmids (repAFIA+repAFII+oriColE1;105-250 kb) or in 50 kb IncN plasmids. In conclusion, this study highlights a polyclonal structure of KPC-3-producing K. pneumoniae and the predominance of the ST147 clone among non-hospitalized patients in Portugal, linked to platforms still unnoticed in Europe (blaKPC−3-Tn4401d-IncFIA) or firstly reported (blaKPC−3-Tn4401d-IncN). This scenario underlines the recent penetration of successful mobile genetic elements in previously circulating MDR K. pneumoniae lineages (mainly ST147 and ST15) in Portugal, rather than the importation of the global lineages from clonal group 258.

The bla KPC genes are commonly located on Tn4401, a 10 kb Tn3-like transposon delimited by two 39-bp imperfect inverted repeat sequences harboring bla KPC , transposase and resolvase genes, and insertion sequences ISKpn7 (upstream bla KPC ) and ISKpn6 (downstream bla KPC ; Chen et al., 2014b). It is recognized as a highly active transposon enhancing the spread of bla KPC genes to different plasmid scaffolds (Cuzon et al., 2011). To date, six Tn4401 isoforms have been described with variable deletions between ISKpn7 and bla KPC providing different promoter regions to the gene (a, −99 bp; b, no deletion; c, −215 bp; d, −68 bp; e, −255 bp; g, equal to isoform c but with one single nucleotide mutation on P2 promotor), and consequently different expression levels of the bla KPC gene (Naas et al., 2012;Chmelnitsky et al., 2014). Besides its genetic environment, other factors are known to have greatly contributed to the spread of KPC producers in many countries, leading to an increasing challenge in the design of effective infection control measures. First, the introduction and subsequent expansion of bla KPC−2 and bla KPC−3 on multidrug resistant (MDR) K. pneumoniae lineages from clonal group (CG) 258 [sequence types (ST) 11, 258, 512] (Munoz-Price et al., 2013;Chen et al., 2014b), followed in a few countries (e.g., Israel, Italy, Colombia) by subsequent dispersion to other clonal backgrounds (Baraniak et al., 2015;Bonura et al., 2015;Ocampo et al., 2015). Second, the acquisition of bla KPC by plasmids from different incompatibility groups (IncFII K2 , IncFIA, IncI2, IncN, IncX3, ColE), favored a quick intra-and inter-species dissemination (Chen et al., 2014b).
In Portugal, KPC-2 was identified only in an environmental Escherichia coli isolate in 2010 (Poirel et al., 2012), while KPC-3 producers were first detected in 2009 in a central hospital (Machado et al., 2010). However, only recently became evident the widespread distribution of KPC-3 among K. pneumoniae isolates in different Portuguese hospitals (Silva et al., unpublished data;Manageiro et al., 2015). In this study, we aim to trace the landscape of KPC-3-producing K. pneumoniae isolates recently identified outside hospital boundaries in Portugal by detailed characterization of clonal and plasmid genetic backgrounds.

Bacterial Isolates and Epidemiological Data
Thirty K. pneumoniae isolates showing reduced susceptibility to carbapenems were identified between October 2014 and November 2015 in three different community laboratories in the North of Portugal, one of them receiving samples from all over the country. Twenty of them were identified as KPC producers and further characterized in this study. They were detected in urine samples (n = 19) or sputum (n = 1) of patients between 61 and 89 years old (mean age = 83; 16 females, 4 males; Table 1). Most of these patients were institutionalized in longterm care facilities (LTCFs) [n = 11; 55%; six different LTCFs (A-F)] or nursing homes (NH) [n = 6; 30%; five different NH (A-E)], while some were identified in ambulatory (n = 3; 15%; Table 1). Most of them (n = 13; 65%) had been hospitalized in the previous month in different hospitals from the North or Centre of Portugal, although in three cases no hospitalization or older hospitalization events (4-9 months) were detected ( Table 1). Travel history abroad was discarded for 60% of the patients (n = 12/20), or considered improbable for the remaining patients due to their clinical conditions (impaired mobilization and chronic underlying diseases).

Carbapanemase Production and Variable Antibiotic Resistance Phenotypes
All isolates produced KPC-3 and demonstrated resistance or intermediate phenotypes to ertapenem (MIC = 1 to 16 mg/L), and susceptible, intermediate or resistance phenotypes to imipenem (MIC = 2 to 16 mg/L) and meropenem (MIC = 1 to 8 mg/L), with colonies growing within the inhibition zone of all carbapenems tested, a hetero-resistance phenotype usually observed for KPC producers (Nordmann et al., 2009; Table 2). Although for some isolates the MIC values for imipenem and meropenem were interpreted as susceptible by the clinical breakpoints defined by EUCAST, in all cases they were above the epidemiological cut-off values (ECOFFs) defined for K. pneumoniae (http://www.eucast.org/mic_distributions_and_ecoffs/; Table 2).
All isolates were defined as multidrug resistant (MDR) in accordance with the definition of MDR for Enterobacteriaceae (non-susceptible to ≥1 agent in ≥3 antimicrobial categories; Magiorakos et al., 2012), although some of them exhibited a less extensive resistance profile to non-β-lactams, being resistant only to ciprofloxacin ( Table 2). All isolates were susceptible to colistin (MIC = 0.25-2 mg/L; Table 2).

DISCUSSION
In this study, we highlight a polyclonal structure of KPC-3 producing K. pneumoniae isolates among patients outside hospital boundaries in Portugal consistent with nosocomial acquisition, and unveil novel or uncovered plasmid backbones carrying bla KPC−3 in Europe.
The first clinical cases of KPC-3 producers in Portugal were detected in 2009 in a pediatric unit of a hospital from the Lisbon and Tagus Valley region and involved 2 K. pneumoniae belonging to ST11 (Machado et al., 2010). Months later and until 2011, an outbreak involving 41 KPC-3-producing isolates, most of them (n = 29) assigned to ST14, was reported (Calisto et al., 2012). More recently, a nationwide study reported 22 K. pneumoniae producing KPC-3 (mainly ST11, ST14, ST15, and ST147 clones) in several hospitals between 2010 and 2013 but plasmid backgrounds had been poorly characterized (Manageiro et al., 2015). However, the situation concerning carbapenemase-producing Enterobacteriaceae in Portugal was only recently recognized in the EuSCAPE survey, where our country appeared in level 2b (sporadic hospital outbreaks) mainly due to the expansion of KPC producers (Albiger et al., 2015). The data presented in this study strengthens an ongoing dissemination of KPC-3 producers in Portugal, where the identification of these bacteria among such a high diversity of healthcare institutions other than hospitals might potentiate their impact for both hospital and community settings. It is thus advisable a reinforcement of infection control measures, surveillance, and tracking of isolates resistant to carbapenems in clinical institutions, and a coordinated action between clinicians, epidemiologists and national reference laboratories for guidance and harmonization of protocols (Albiger et al., 2015).
The ST147 clone [clonal group (CG) 147] exhibiting capsular type K14.K64 was identified in patients from diverse LTCFs and NHs for a long period of time and seems to be the predominant lineage among KPC-3-producing K. pneumoniae in different healthcare settings (Silva et al., unpublished data;Manageiro et al., 2015). In fact, identical KPC-3-producing ST147 isolates were recently involved in outbreaks in hospitals where some of the patients had been previously hospitalized. Although, nosocomial acquisition is the most probable origin for most KPC-3 producers identified in patients included in this study, it is of notice that in three cases no obvious hospitalization link could be established. Indeed, considering the frequent displacement of these patients between institutions (integrated network of LTCFs in Portugal) and hospitals (we had only access to the last hospitalization event) and that intestinal colonization might be persistent in time (Feldman et al., 2013), we cannot completely discard cross transmission events in the units analyzed.
The ST231 K. pneumoniae clone (CG231) had already been linked to GES-5 plus SHV-12 production in Portugal (Manageiro et al., 2015) and its association with community invasive infections (sepsis, lethal pneumonia, or meningitis), and high content in virulence and antimicrobial resistance genes have recently been highlighted (Holt et al., 2015). The ST109 clone is rarely reported (http://bigsdb.web.pasteur.fr/klebsiella/klebsiella.html) and it is described for the first time in Portugal. However, it belongs to the CG17, associated with the expansion of CTX-M-15 and different carbapenemases worldwide (Rodrigues et al., 2014;Holt et al., 2015).
The bla KPC−3 was linked to Tn4401d isoform in all characterized KPC-3-producing Enterobacteriaceae from Portugal (this study; Manageiro et al., 2015). In this study, we show that in most cases (80%) Tn4401d-bla KPC−3 was located within cointegrated FIA, FII, and ColE1 plasmids (105-250 kb; Table 1) corroborating the strong association between Tn4401d-bla KPC−3 with IncFIA plasmids pointed out previously in large collections from the USA (Chen et al., 2014a;Deleo et al., 2014;Bowers et al., 2015;Chavda et al., 2015). We detected FIA and FII replicons identical to those of pBK30683 plasmid (GenBank accession number KF954760) plus an additional ori ColE1 gene identical to that of the ColE1 plasmid pKBuS13 (GenBank accession number KM076933), supporting the role of these mobilizable plasmids in the assembly of MDR plasmids (Chen et al., 2014a;Garbari et al., 2015). These and other (IncFIA plus IncA/C 2 or IncFIA plus IncX 3 ) cointegration forms seem to play an important role in the intra-and inter-species spread of carbapenem resistance genes (Chen et al., 2014a,b;Chavda et al., 2015). To the best of our knowledge, we unveil for the first time a cointegrate IncFIA platform carrying Tn4401d-bla KPC−3 in Europe, characterized and highly represented by far only in isolates from the USA (mainly among non-ST258 and non-K. pneumoniae isolates; Chen et al., 2014a). In the remaining isolates (20%, 4 ST147), bla KPC−3 -Tn4401d was located on ca. 50 kb IncN plasmids, an association primarily described in this study.
In conclusion, this study highlights a polyclonal structure among KPC-3 producers identified in geographically dispersed non-hospitalized patients in Portugal, not always linked to nosocomial acquisition, a situation that deserves close monitoring due to its high clinical or epidemiological impact. In all cases, a common platform (bla KPC−3 -Tn4401d) was identified in plasmids still unnoticed in Europe (bla KPC−3 -Tn4401d-IncFIA) or firstly reported here (bla KPC−3 -Tn4401d-IncN). Their identification in previously circulating MDR K. pneumoniae lineages in our area (ST147, ST15, ST231, ST348) underlines the recent penetration of successful mobile genetic elements in locally circulating clonal backgrounds, rather than the importation of the most common global lineages from CG258.

AUTHOR CONTRIBUTIONS
CR and JB performed the experiments and contributed with the acquisition of molecular data. CR and ÂN wrote the article and performed analysis and interpretation of molecular data. EM and JA contributed with epidemiological data and revision of the manuscript. ÂN and LP contributed with the design of the study and final revision of the manuscript. All authors read and approved the final version of the manuscript.