Emergence of Enterobacter cloacae Complex Co-Producing IMP-10 and CTX-M, and Klebsiella pneumoniae Producing VIM-1 in Clinical Isolates in Japan

Background: Carbapenemase-producing Enterobacteriaceae (CPE) are an emerging threat in healthcare settings worldwide. Objectives: We evaluated the presence of carbapenemase genes in CPE in a tertiary care university hospital in Tokyo, Japan. Methods: Carbapenem-resistant clinical isolates were collected in 2018 at Teikyo University Hospital (Tokyo, Japan). Bacterial species were identified using MALDI-TOF MS. Carbapenemase production was evaluated using a carbapenemase inactivation method. The presence of carbapenemase genes was confirmed by multiplex PCR and DNA sequencing. Results: Four CPE isolates were identified: two Enterobacter cloacae complex strains and Klebsiella oxytoca and Klebsiella pneumoniae strains. Three of the isolates (E. cloacae complex and K. oxytoca) were IMP-1-type producers, including IMP-10 in their produced metallo-β-lactamase, and are epidemic in East Japan. The IMP-10-producing E. cloacae complex strain also produced CTX-M ESBL. The other CPE isolate (K. pneumoniae) is a VIM-1 producer. VIM-1-producing K. pneumoniae is epidemic in Europe, especially in Greece. Accordingly, the VIM-1 producer was isolated from a patient with a medical history in Greece. Conclusions: This study revealed the emergence of E. cloacae complex co-producing IMP-1-type carbapenemase and CTX-M ESBL, and K. pneumoniae producing VIM-1 carbapenemase in clinical isolates in Japan. Metallo-β-lactamase was the most prevalent type of carbapenemase at Teikyo University Hospital, especially IMP-1-type carbapenemase. The detection of VIM-1-producing K. pneumoniae suggests that epidemic CPE from overseas can spread to countries with low CPE prevalence, such as Japan, highlighting the need for active surveillance.


Introduction
Carbapenem-resistant Enterobacteriaceae (CRE) are often multidrug-resistant (MDR) and pose severe problems in terms of clinical treatment and infection control. Recently, carbapenemase production has become a common mechanism of carbapenem resistance in Enterobacteriaceae, which has given rise to carbapenemase-producing Enterobacteriaceae (CPE). The limited options for the treatment of MDR CPE infections significantly increases morbidity and mortality [1].

Multilocus Sequence Typing (MLST)
MLST was performed using the protocol developed by the Institut Pasteur (https://bigsdb.pasteur. fr) [10] and provided by PubMLST (https://pubmlst.org) [11,12]. Genomic DNA was isolated using a DNeasy Blood & Tissue kit (Qiagen, Hilden, Germany). DNA fragments were amplified by PCR using KOD FX Neo (Toyobo, Osaka, Japan) and specific primers (Supplemental Material) in a final volume of 50 µL. The amplification began with an initial denaturation at 94 • C for 2 min followed by 35 cycles of amplification using a step program (10 s at 98 • C, 30 s at the melting temperature (T m ), and 1 min at 68 • C). The PCR products were treated with ExoSAP-IT Express (Affymetrix, Santa Clara, CA, USA) and directly sequenced as described in our previous report [2].

Results and Discussion
We collected 70 carbapenem-resistant clinical isolates during 2018. Among them, we identified four CPE as MBL-producing Enterobacteriaceae: two E. cloacae complex isolates, and Klebsiella oxytoca and K. pneumoniae isolates. CRE accounted for <0.2% of the bacterial isolations, and CPE accounted for <6% of the CRE infections, consistent with the results of a previous national surveillance report in 2014 [3]. The antimicrobial susceptibility profiles of the MBL-producing Enterobacteriaceae isolates, strain numbers, sample characteristics, MLST, carbapenemase production, and ESBL are summarised in Table 1. All the MBL-producing Enterobacteriaceae isolates were MDR.
The carbapenemase and ESBL genes harboured by the isolates were identified by multiplex PCR. Strain TK1601 harboured bla IMP-1 ; strain TK1602 harboured bla VIM and bla SHV ; strain TK1603 harboured bla IMP-1 , bla CTX-M-1 , and bla TEM ; and strain TK1604 harboured bla IMP-1 . Sequencing analysis identified the bla VIM gene in K. pneumoniae and bla IMP-1 gene in an isolate of E. cloacae complex as bla VIM-1 and bla IMP-10 , respectively.
The four Enterobacteriaceae isolates produced MBLs representing IMP-1-type and VIM-1 enzymes. IMP-1 is an epidemic MBL in Japan [15]. The distribution of IMP-producing Gram-negative bacteria in Japan is biased, such that, as reported, IMP-1 occurs more frequently in eastern Japan and IMP-6 occurs more frequently in western Japan [15,16]. The detection of IMP-1 at Teikyo University Hospital was consistent with such an IMP distribution in Japan. The detection of IMP-1-producing K. oxytoca at the University of Tokyo Hospital was previously reported in 2009 [17]. The detected strain may be similar to the one detected in the current study, as their drug sensitivity patterns, as reported, are identical ( Table 2). The MLST of previous K. oxytoca was not performed. In this study, IMP-1-producing K. oxytoca and IMP-1-producing E. cloacae complex were found to belong to ST88 and ST252, respectively. E. cloacae complex ST252 was initially isolated as a KPC-3 producer from a liver-transplant patient in a hospital (PA, USA) in 2009 [18]. E. cloacae complex ST252 producing GES-5 was isolated from the leg wounds of a diabetic patient in a Czech hospital in 2016 [19]. IMP-1-producing E. cloacae ST252 and IMP-1-producing K. oxytoca ST88, such as the isolate characterised in this study, have never been reported. Therefore, IMP-1-producing E. cloacae ST252 and IMP-1-producing K. oxytoca ST88 might have emerged locally in Japan. This finding suggests that E. cloacae ST252 is an international clone of a carbapenemase reservoir, although this will require further investigation. Furthermore, in Tokyo, the in-hospital mortality of patients colonised by IMP-1-producing E. cloacae complex is high [20]. Therefore, it is necessary to continue to screen hospitalised patients for the presence of IMP-1-producing E. cloacae complex as an essential measure for infection control.  Antimicrobial susceptibility profiles of IMP-1-producing K. oxytoca from this study and a previous study [17].

Strain Number TK1604 K27
Year On the other hand, the current study identified an E. cloacae complex isolate co-producing IMP-10 and CTX-M-1 group enzymes. Of note, IMP-1-producing E. cloacae complex ST78 is a successful (prevalent and persistent) clone in Tokyo [15]. Interestingly, we identified IMP-10-producing E. cloacae complex ST78, suggesting a selection of an ST78 strain variant producing IMP-1 with a V49F (145 G to T) substitution [21,22]. The distribution of IMP-1 and IMP-10 enzymes was reported in other CPE species, such as Serratia marcescens in Tokyo [23]. IMP-1-and IMP-10-producing S. marcescens have Class I integrons similar to In316 located in the plasmid pMTY11043_IncHI2 harboured by IMP-1-producing E. cloacae complex ST78 isolate TUM11043 [15,23]. Therefore, plasmids carrying Class I integrons encoding IMP-1 and IMP-10 might be widely distributed in CPE in Tokyo. Because IMP-10-producing P. aeruginosa was initially reported in Japan and Class I integrons encoding IMP-10 in P. aeruginosa isolates have a similar structure to the Class 1 integron in S. marcescens, the mobile genetic element carrying bla IMP-10 might have been transferred initially from P. aeruginosa to Enterobacteriaceae in Japan [21,[23][24][25]. The MBL-producing Enterobacteriaceae isolates in this study share the incompatibility plasmids IncFIB and IncHI2. The IMP-10-producing E. cloacae complex and VIM-1-producing K. pneumoniae harboured the unique incompatibility plasmids IncL/M and IncHI1, respectively.
We isolated VIM-1-producing K. pneumoniae from a patient who had been previously hospitalised in Greece. Hospitalisation in Greece might be a risk factor for infection or colonisation by VIM-producing Enterobacteriaceae. For instance, an outbreak of VIM-1-producing K. pneumoniae from Greece in a hospital in France was reported in 2006 [26]. Many VIM-producing K. pneumoniae strains are being isolated in Greece [27]. The dominant clone of VIM-1-producing K. pneumoniae belongs to ST147 [28]. K. pneumoniae ST70 was characterised as an NDM-1 producer in Greece [29]. However, a VIM-1-producing K. pneumoniae ST70, such as the isolate characterised in the current study, has never been reported. The emerging clone of K. pneumoniae ST70 may have evolved to host a mobile genetic element encoding VIM-1.
VIM-2-producing P. aeruginosa has spread throughout Asia, including Japan [30]. By contrast, the VIM-producing Enterobacteriaceae isolate from Japan was reported as a VIM-2-producing E. cloacae, isolated from a patient in a paediatric ward [31]. However, the most notorious nosocomial pathogen, K. pneumoniae, producing VIM-1 has not yet been reported in Japan. VIM-2 is a significant epidemic carbapenemase harboured by P. aeruginosa in Japan. Therefore, the possibility that the VIM-1-encoding plasmid might be transferred from P. aeruginosa to K. pneumoniae in Japan may be excluded. On the other hand, VIM-1-producing K. pneumoniae is epidemic in Europe [1].
The findings presented herein suggest that a medical history in Greece might be linked to the CPE colonisation described in this study. This should serve as a warning about the possibility that doctors and other medical staff can come into contact with CPE when admitting patients colonised by CPE, and then spread the infection to other inpatients at the hospital. To reduce the spread of CPE in low-prevalence countries, such as Japan, the screening of all patients, especially those with a history of hospitalisation and travel abroad, is inevitable.
Given the limitations imposed by this surveillance study, the carbapenemase production by CRE may have been underestimated because of the relatively short screening period and the exclusion of CPE isolates but not resistant isolates. Although we detected MBL as a major carbapenemase in a CRE at Teikyo University Hospital in 2018, it strongly depended on the yearly screened patients ( Table 3). The screening methods used to identify CPE also affected the results. Nevertheless, the implementation of active surveillance has helped to prevent the spread of CPO in the last decade at Teikyo University Hospital. Improvements in the methods for screening CPE may help to more precisely reveal the status of CPE colonization or infection in healthcare settings in the future. In conclusion, we report here the isolation of E. cloacae complex co-producing IMP-1-type MBL and CTX-M ESBL, E. cloacae complex producing IMP-1 MBL, K. oxytoca producing IMP-1 MBL, and K. pneumoniae producing VIM-1 MBL in Japan. This study contributes to the delineation of recent epidemic trends and treatment options for CPE in Tokyo, Japan, and highlights the possibility of resistance gene transfer among Enterobacteriaceae as well as resistance gene spread from travellers returning from abroad. The study also highlights the need for active surveillance and analysis of the medical history of patients returning from travel abroad.