Clonality, outer-membrane proteins profile and efflux pump in KPC- producing Enterobacter sp. in Brazil

Background Carbapenems resistance in Enterobacter spp. has increased in the last decade, few studies, however, described the mechanisms of resistance in this bacterium. This study evaluated clonality and mechanisms of carbapenems resistance in clinical isolates of Enterobacter spp. identified in three hospitals in Brazil (Hospital A, B and C) over 7-year. Methods Antibiotics sensitivity, pulsed-field gel electrophoresis (PFGE), PCR for carbapenemase and efflux pump genes were performed for all carbapenems-resistant isolates. Outer-membrane protein (OMP) was evaluated based on PFGE profile. Results A total of 130 isolates of Enterobacter spp were analyzed, 44/105 (41, 9%) E. aerogenes and 8/25 (32,0%) E. cloacae were resistant to carbapenems. All isolates were susceptible to fosfomycin, polymyxin B and tigecycline. KPC was present in 88.6% of E. aerogenes and in all E. cloacae resistant to carbapenems. The carbapenems-resistant E. aerogenes identified in hospital A belonged to six clones, however, a predominant clone was identified in this hospital over the study period. There is a predominant clone in Hospital B and Hospital C as well. The mechanisms of resistance to carbapenems differ among subtypes. Most of the isolates co-harbored blaKPC, blaTEM and /or blaCTX associated with decreased or lost of 35–36KDa and or 39 KDa OMP. The efflux pump AcrAB-TolC gene was only identified in carbapenems-resistant E. cloacae. Conclusions There was a predominant clone in each hospital suggesting that cross-transmission of carbapenems-resistant Enterobacter spp. was frequent. The isolates presented multiple mechanisms of resistance to carbapenems including OMP alteration.


Background
Healthcare associated infections caused by Enterobacter spp. have increased in the last decade all over the world [1,2]. Carbapenems are frequently used to treat serious infections caused by multi-resistant Gram-negative bacilli, especially those caused by over production of AmpC cephalosporinases or extended-spectrum β-lactamases (ESBL), such as infections caused by Enterobacter spp. Thus, the emergence of carbapenems resistance, defined as resistance to ertapenem, imipenem and/or meropenem, is becoming a therapeutic challenge [1,2].
Therefore, the role of mechanisms of carbapenems resistance, such as OMP and efflux pump in Enterobacter spp, needs to be better addressed. The present study was conducted to investigate the clonality and mechanisms of carbapenems resistance in E. aerogenes and E. cloacae identified in three hospitals in Brazil.

Bacteria collection
One hundred and thirty Enterobacter spp. clinical isolates (105 E. aerogenes and 25 E. cloacae) identified in two hospitals (Hospital A and Hospital B whit 30 km of distance between them) in São Paulo in the state of São Paulo, and one hospital (Hospital C) 530 km distant, in Londrina in the state of Paraná, Brazil, were evaluated over a 7-year period, from 2005 to 2011.
Although, located in another state, the strains of E. cloacae identified in Hospital C in Londrina were evaluated in order to investigate the mechanism of carbapenems resistance and clonality in this species as well.
The identification of species was performed by API20 E (bioMérieux, France) and additional tests (Modify Rugai, Motility and Lysine).

Clinical data
The following clinical and demographic data from the medical records of patients hospitalized in Hospital A and Hospital B, were registered: age, gender, underlying diseases, site of infection, length of stay in the Intensive Care Unit and death. Definitions (CDC) for the infections were those used by the Centers for Disease Control and Prevention. An Epi Info™ database was built, and results were expressed as means (standard deviation) or median (interquartile range), depending on normality. All data were analyzed anonymously and confidentially, with approval by the Research Ethics Committee of the three hospitals.

Ethics statement
The study was performed in two hospitals located in São Paulo, Brazil, the Central Institute of Hospital das Clínicas of University of São Paulo (ICHC-FMUSP) and Hospital Itapecerica da Serra and one hospital in Paraná, the Universitary Hospital in Londrina. It was approved by the ethics committee of the hospitals. The approval number is 007/11.
PCR for all isolates (105 E. aerogenes and 25 E. cloacae) was performed to detect genes of efflux pump acrART using primers according to the study of Perez et al. [12].
DNA sequencing of genes of resistance was performed using the MegaBACE 1000, DNA Analysis System (Amersham Biosciences, UK. England), using DYEnamic ET Dye Terminator Kit (with Thermo Sequence™ DNA Polymerase II) US81090 code. The sequences were analyzed using the Sequence Analyzer software using the Cimarron Base Caller 3.12. The genetic sequence was compared with the database available on the Internet (BLAST -http://www.ncbi.nlm.nhi.gov/blast/).

Pulsed-field gel electrophoresis
DNA analysis of all carbapenems-resistant isolates, 44 E. aerogenes and 8 E. cloacae were performed by PFGE, after digestion with XbaI Fast (Invitrogen) and the electrophoretic run was made with the following parameters: 23 h with pulse times ranging from 5 to 60s at 6 V/cm, using the CHEF-DR III System (Bio-Rad Laboratories, Richmond, CA, USA). DNA relatedness was computationally analyzed using BioNumerics v.7.1 software (Applied Maths, Sint-Martens-Latem, Belgium). The banding patterns were compared by using the unweighted pair-group method with arithmetic averages (UPGMA), with the Dice similarity coefficient required to be >80% for the pattern to be considered as belonging to the same PFGE type (dendrogram).

Outer membrane protein profile
Based on clonality, the OMP of 22 E. aerogenes isolates and 5 E. cloacae isolates were analyzed. Bacterial outer membrane proteins (OMPs) were purified by treatment of the cell envelops with 2% sodium-N-lauryl sarcosinate (Sigma Chemical.St Louis., MO). The proteins were quantified by the Bradford method with the aid of Bradford reagent (Bio-Rad Laboratories, Brazil) by spectrophotometry at 595 nm (GeneQuant Pro, GE-Healthcare Life Sciences) and subsequently treated with beta-mercapto-ethanol at a ratio of 2 μl acid beta-mercaptoethanol acid to 10 μl protein. The proteins studied were applied to polyacrylamide gels prior manufactured 12.5% (GE Healthcare) at a concentration of 20 ng. Electrophoreses run was performed on the Multiphor II device (SG) at 600 V, 50 mA, and 30 W for about 60 min. The following molecular weights were used as markers: 97.4 KDa (phosphorylase B from rat muscle), 66,2KDa (bovine serum albumin), 45 kDa (egg albumin), 31,0KDa (bovine carbonic anhydrase), 21,51KDa (trypsin inhibitor) and 14,4KDa (lysozyme). Enterobacter aerogenes ATCC13048 was used as control. After electrophoresis, the gel was stained with Silver Staining Kit plus one Kit, Protein (GE Healthcare) according to the manufacturer's instructions. We classified the protein profile based on intensity of band in without lack or loss of protein (++++), very little reduction of protein (+++), reduction (++), major reduction (+) and absent of protein based on previously described by Mostachio et al. [13].

Efflux pump activity
The efflux pump activity was analyzed based on carbapenems's MIC in 5 E. cloacae that harbored efflux pump gene and 5 in E. aerogenes belonged to different clones. MICs of imipenem, meropenem and ertapenem with 50 and 100 mg/mL and without the efflux pump inhibitors Carbonyl-cyanide-m-chlorophenylhydrazone (CCCP-Sigma Chemical.St Louis., MO) were determined by agar dilution to investigate the role of efflux pump on carbapenems-resistant E. aerogenes and E. cloacae isolates. The influence of an efflux pump on the carbapenems's MIC for a given bacterial strain was determined by a reduction of at least four-fold of the respective MIC in the presence of CCCP [2].

Results
A total of 44/105 (41, 9%) E. aerogenes and 8/25 (32, 0%) E. cloacae were resistant to carbapenems. Thirtynine isolates of E. aerogenes resistant to carbapenems were identified in Hospital A, 5 isolates of E. aerogenes in Hospital B and 8 isolates of E. cloacae in  Table 2. Most of the isolates were from blood 12 (30.8%), followed by 9 (23%) from respiratory tract secretion and 6 (15.4%) from urine. Most of patients (N = 17) undergone surgery (4 liver and 2 kidney transplant), 2 patients received chemotherapy (1 acute leukemia and 1 bone marrow transplant) and 1 patient was HIV positive (Table 3). There was no clinical information regarding isolates identified in Hospital C, although, all isolates were from surveillance swabs.
The outer membrane proteins of 22 of the 44 isolates of E. aerogenes and 5 of 8 isolates of E. cloacae were analyzed. The intensity of the proteins of interest Omp F (39 KDa) OMP C (42 KDa), 35-36KDa were compared with the molecular weight and ATCC E. aerogenes 13048 was used as control (Table 4).  The protein profile was classified based on intensity of band in without lack or loss of protein (++++), very little reduction of protein (+++), reduction (++), great reduction (+) and absent ( Table 4). All isolates resistant to carbapenems in this study showed a decrease in 35-36 KDa protein.
We observed that the 42KDa protein in our analysis did not appear to be involved in carbapenems resistance.
E. aerogenes isolates with decrease of 39 KDa protein, presented CIM of 2-128 mg/mL to imipenem, 4-64 mg/mL to meropenem and 8 a ≥128 mg/mL to ertapenem. Isolates of E. aerogenes, which presented absence of 39 KDa protein showed MIC of 8 to 32 mg/ml for imipenem, 4 to 32 mg/ml for meropenem and 32-64 mg/mL to ertapenem.
E. aerogenes isolates with decrease of 35-36 KDa protein, presented MIC of 2-32 mg/ml for imipenem and 4 to 32 mg/ml for meropenem and 8-128 mg / mL to ertapenem. E. aerogenes isolates with absence of the protein of 35-36 KDa showed MIC of 128 mg/mL to imipenem, 64 mg/mL to meropenem and ≥128 mg/mL to ertapenem.
Moreover, all isolates of carbapenems-resistant E. cloacae showed decreased of 39KDa protein and absence of 35-36 KDa protein and their MICs to imipenem were 8 to 64 mg/ml, 2 to 16 mg/mL to meropenem and 8-64 mg/mL to ertapenem ( Table 5).
The PFGE showed 6 clones (A, B, C, E, F and G) among 39 isolates of carbapenems-resistant E. aerogenes identified in Hospital A. Clone A was present only in the first two year of study (2005 and 2006), and it was replaced by clone C that circulated during the entirely study. Twentyseven (69.2%) isolates belonged to this predominant clone C that harbored KPC, TEM-1 and CTXM, whose subtypes ranging from C1 to C19 over the 7-year study period, from 2005 to 2011 (Fig. 1). Two subtypes (C17 and C19) of clone C and clone F were KPC negative.
The 4 carbapenems-resistant E. aerogenes identified in Hospital B belonged to a predominant clone nominated as clone D with subtypes from D1 to D3. This clone harbored KPC and TEM-1, and showed decreased of 39 and 35-36 KDa proteins ( Fig. 1 and Table 5).
The efflux pump AcrAB-TolC gene was only identified in carbapenems-resistant E. cloacae. All 5 isolates positive for this efflux pump were resistant to imipenem, meropenem and ertapenem with MICs above 8 mg/mL for all carbapenems (Tables 4 and 5).
The 5 carbapenems-resistant E. aerogenes isolates belonged to different clones and showed efflux pump activity in presence of CCCP inhibitor. Among the carbapenems-resistant E. cloacae, 2 of the 5 isolates showed efflux pump activity for meropenem, imipenem and ertapenem, and 3 isolates showed no activity on presence of ertapenem (Tables 4 and 5).

Discussion
In the present study, we demonstrated that all E. aerogenes and E. cloacae isolates resistant to carbapenems were susceptible to polymyxin B, tigecycline and fosfomycin. KPC was the only carbapenemase identified. In contrast with previous studies [5,10,14], most of carbapenems-resistant E. aerogenes isolates co-harbored KPC and wide spectrum ESBL, such as blaTEM and blaCTX-M, in association with alteration of OMPs, mainly reduction or loss of 35-36 KDa and 39KDa proteins. In addition to alteration of OMPs, E. cloacae isolates harbored blaKPC; blaTEM and efflux pump gene acrART. E. aerogenes isolates were polyclonal, although a different predominant clone was finding in each hospital. On the other hand, E. cloacae belonged to two clones. The mechanisms of resistance to carbapenems E. aerogenes and E. cloacae are important agents of healthcare-associated infections in several countries and resistance to carbapenems has been increasing in the last decade, thus there is a need for in vitro studies showing alternatives for the treatment of infections due to these microorganisms [3,7,14]. Since our susceptibility results showed that all carbapenems-resistant isolates were susceptible to tigecycline, fosfomycin, and polymyxin B, we consider these findings quite promising. However, it is noteworthy that despite the excellent results in vitro, pharmacokinetics and pharmacodynamics of tigecycline and fosfomycin, and the limited clinical experience in serious infections, especially bloodstream infections, are obstacles to the routine use of these antimicrobials in the treatment of systemic infections. Despite of these drawbacks, these drugs have been used with success as combination therapy in the treatment of severe infections caused by Enterobacteriaceae resistant to carbapenems [15,16].
The demographic and clinical data from the studied hospitals showed that the most frequent site of isolation of carbapenems-resistant Enterobacter was blood, most of our patients had undergone surgical procedures, and had a high overall mortality, similar to previous studies [14,17]. The only carbapenemase identified in our study was KPC. Data regarding carbapenems resistance in Enterobacter spp. in Brazil are scarce, although KPC is the carbapenemase most often described in Enterobacter spp. isolates in Brazilians hospitals [5,7,14].
Regarding the clonality, we identified in Hospital A, in São Paulo, 6 clones of E. aerogenes, named A, B, C, E, F and G over the study period, from 2005 to 2011, with the predominance of clone C. One predominant clone was also identified in hospital B in São Paulo. An intriguing finding is that one isolate of hospital B seems to be related to a clone of hospital A, pointed to a possible inter-hospital spread. Two clones A (subtypes A1 to A5) and B were identified among E. cloacae isolates resistant to carbapenems in Hospital C, in Londrina, Paraná. These findings suggest that cross-transmissions have occurred over the study period in all hospitals and highlight the importance of the stringent enforcement of hand hygiene and contact precautions to control the spread of this agent. Another interesting finding is that the mechanism of resistance to carbapenems in E. aerogenes from hospital A differs among subtypes and CTX-M was only identified in this hospital. The profile of outer-membrane proteins of our isolates demonstrated that all isolates studied showed reduction of 35-36 KDa OMPs. These OMPs have been associated with β-lactam resistance in Enterobacter spp [18][19][20][21]. Studies in vivo and in vitro showed that diminished or loss of these OPMS were associated with increase on imipenem and meropenem MIC [18][19][20][21][22]. Other OMPs previously associated with carbapenemsresistance among E. aerogenes are OMPs of 42KDa and OMP of 39KDa [19]. However, in contrast with previous studies, we could not associate the alteration of the 42KDa OmpC with carbapenems resistance in our E. aerogenes and E. cloacae isolates. This finding could be explained perhaps by the clones that circulated in Brazil. Data regarding OMP in carbapenems-resistant Enterobacter in Brazil is limited. However, a previous Brazilian study demonstrated that OmpC and OmpF were present in only 6.6% of E. cloacae isolates resistant to ertapenem [5].
Few authors had evaluated the role of efflux pump on carbapenems resistance in Enterobacter spp [23,24]. Efflux pump AcrAB system has been described as associated with resistance to meropenem and imipenem in E. cloacae [2,22,23]. In carbapenems-resistant E. aerogenes there were descriptions of efflux pump associated with KPC and/or porin loss as well [2,23,24]. In our study, the efflux pump AcrART was found only in E. cloacae. However, eventhough E. aerogenes isolates did not present the efflux pump AcrART gene, carbapenem efflux pump activity was observed in the presence of CCCP inhibitor. Possibly, these E. aerogenes isolates have other efflux pump, not described yet. These findings reinforce the importance of association of mechanism of resistance in carbapenems-resistant Enterobacter spp.
This study has several limitations: besides being retrospective, we evaluated only three hospitals and we could access the clinical data from only two hospitals. However, we were able to demonstrate that there was an association of mechanism of resistance such as β-lactamase, alteration of outer-membrane protein and efflux pump in carbapenemsresistant Enterobacter spp. in Brazil.

Conclusion
In conclusion, we observed that there was a predominant clone in each hospital, suggesting that cross-transmission of carbapenems-resistant Enterobacter spp. was frequent. The isolates presented multiple mechanisms of resistance to carbapenems, such as KPC, ESBL and alteration of the outer membrane protein. E. cloacae presented as well the efflux pump AcrART, and in vitro activity of efflux pump inhibitor. These findings can be useful to control the spread of carbapenems resistance among Enterobacter spp. in the hospital setting.