Prevalence of carbapenem resistance encoding genes and corresponding MIC90 in enterobacteriaceae at a tertiary medical care center in Lebanon

Background: The aim of this study was to correlate genes involved in carbapenem resistance to MIC levels among clinical ESBL and non-ESBL producing carbapenem resistant Enterobacteriaceae (CRE) isolates of Escherichia coli and Klebsiella pneumoniae. Materials and Methods: E. coli (n=76) and K. pneumoniae (n=54), collected between July 2008 and July 2014, were analyzed. The MICs were determined against ertapenem (ERT), imipenem (IMP) and meropenem (MER). PCR was performed on all 130 isolates to amplify the resistance and outer membrane proteins (OMPs) encoding genes: blaOXA-48, blaNDM-1,blaTEM-1,blaCTX-M-15, ompCand ompF.Sequencing was performed on selected isolates. Results: The prevalence of blaOXA-48, blaNDM-1, blaTEM-1 and/or blaCTX-M-15 among E. coli isolates were 36%, 12%, 20% and 80%, respectively, while among K. pneumoniae they were 37%, 28%, 28% and 72%, respectively. K. pneumoniae isolates positive for any of these genes had an MIC90> 32μg/ml against ERT, IMP and MER, while in E. coli isolates there was a variation in the MIC90 values. Porin impermeabilitieswere due to mutations in ompCand ompF genes in E. coli, and loss of ompCand ompF genes in K. pneumoniae,andincreased MIC90 values. The presence of more than one carbapenem resistance encoding gene and/ or ESBL encoding genedid not have an effect on the MIC90 value in K. pneumoniae isolates, while in E. coli it showed higher MIC90 values. THE INTERNATIONAL ARABIC JOURNAL OF ANTIMICROBIAL AGENTS


Introduction
The dissemination of carbapenem resistant Enterobacteriaceae (CRE) globally is an alarming crisis, which attributes to a serious public health, infection control and society threat. Its fast-paced spread intensifies the clinical challenge further, requiring urgent interventions [1][2][3][4][5].
Several risk factors for infections with CRE have been identified including prolonged hospitalization, mechanical ventilation, exposure to broad spectrum antibiotics and previous colonization with these strains. In addition, high mortality rates of around 40% have been associated with CRE, mostly of bloodstream infections and pneumonia [4].
In the Middle East in general and in Lebanon in particular, there is a paucity of scientific reports compared to other regions of the world, concerning CRE and underlying mechanisms of resistance. In a previous pilot study [6], we determined in a limited number of carbapenem resistant isolates the mechanism of carbapenem resistance at a tertiary medical center in Lebanon to be due to bla  and/or bla NDM-1 and bla CTXM-15 along with outer membrane impermeabilities and to a significantly lesser effect of efflux pump activity.
This study was undertaken to screen for the prevalent carbapenem resistance (bla CTXM-15 , bla OXA-48 , bla NDM-1 and bla TEM-1 ) and outer membrane porin(OMPC and OMPF) encoding genes and correlate their role singly or in combination to MIC levels among CRE producing clinical isolates of E.coli and K.pneumoniae, from Lebanon.

Clinical isolates
Available phenotypically determined non-duplicate ESBL producing CRE E. coli (n=76) and K. pneumoniae (n=54) were collected at the Clinical Microbiology Laboratory of the Department of Pathology and Laboratory Medicine at the medical center, between July 2008 and July 2014. The CRE isolates were cultured on MacConkey agar (Scharlau, Spain) and stored in Brucella broth (BBL, USA), enriched with 15% glycerol for later use. The origins of each of the 130 isolates are listed in Table 1.

Conclusion:
Levels of MIC in CRE may largely depend on the type of resistance encoding genes, and porin impermeabilities. These resultsmay provide information for antibiotic regimen selection and epidemiological monitoring of resistance.

Minimum inhibitory concentrations (MIC)
Minimum inhibitory concentrations (MIC) for ertapenem (ERT), imipenem (IMP) and meropenem (MER) were determined by Etest (bioMérieux, France) following the manufacturer's guidelines and the results were interpreted using the CLSI standards for MIC interpretation. An MIC value ≥ 2 µg/ ml, against ERT, IMP or MER is considered resistant reflecting potential CRE of these isolates [8].

DNA extraction
DNA was extracted from the 130 isolates using the QIAamp DNA Mini Kit (QIAGEN, Germany), following the manufacturer's instructions for DNA extraction from bacteria.

Polymerase chain reaction (PCR)
Polymerase chain reaction (PCR) for the metallobeta-lactamases were previously done on tested isolates (data unpublished). However, due to the absence of thesemetallo-beta-lactamase encodinggenes, only the following resistance encoding genes: bla TEM-1 , bla CTX-M-15 , bla OXA-48, bla NDM-1 , ompC, and ompF were performed on the extracted DNA of each isolate. Primer sequences are listed in Table 1. Positive and negative controls used for the PCRare mentioned in Table 2.
Amplification was achieved using the PCR Sprint Thermal Cycler (Thermo Fisher Scientific, Waltham, MA, USA). The PCR cycling conditions are listed in supplementary table 2. PCR amplicons were electrophoresed on 1.5% agarose gel using SeaKem® LE Agarose (FMC Bio-Product, Rockland, ME, USA). A 50 base pair ladder (Fermentas Life Sciences, Burlington, Ontario, Canada) was run in parallel to the samples and served as a size marker. Amplicons were then observed using ULTRA LUM, Dual Light Transilluminator (Claremont, CA), and photographed using the Digi-Doc It program (Ultra Violet Products Ltd., Cambridge, UK).
Amplified PCR products were extracted from the gel after electrophoresis and purified using a QIAquick Gel Extraction Kit (QIAGEN, Valencia, CA) following the manufacturer's instructions. Purified DNA was seen on a 1.5% agarose gel. Sequencing reactions were performed using a BigDye® Terminator v.3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and purified using a DyeEx 2.0 Spin Kit (QIAGEN). Purified sequences were detected using an ABI 3500 Genetic Analyzer (Applied Table 1. PCR primers used for the detection resistance and outer membrane porin encoding genes

Statistical Analysis
Data statistical analysis was carried out, when applicable, using GraphPad. Chi Square with Yates' correction, one tailed statistics was carried out. The differences were considered to be statistically significant when the p-value obtained was less than 0.05.
Upon calculating the MIC 90 values, the results showed that among the E. coli isolates positive for the resistance encoding genes, MIC 90 was ≥ 32 µg/ ml for ERT. IMP had MIC 90 of 6 µg/ml for bla NDM-1 and bla CTX-M-15 positive isolates. As for MER, MIC 90 was 1.5 µg/ml and 4 µg/ml for bla NDM-1 and bla CTX-M-15 positive isolates, respectively. However, among K. pneumoniae the MIC 90 for the ERT, IPM and MER were all ≥ 32µg/ml for the isolates harboring either gene: bla OXA-48 , bla NDM-1, bla CTX-M-15 or bla TEM-1 (Table 4).
Comparing resistance encoding genes with the porin encoding genes it appeared that in E. coli the isolates negative for ompC or ompFor both were 4% (1/27) for bla OXA-48 positive isolates, 11% (1/ 9) for bla NDM-1 positive isolates, 5% (3/61) for bla CTX-M-15 positive isolates and 0% ( 0/15) for bla TEM-1 positive isolates. As for K. pneumoniae, the isolates negative for ompCor ompF or both were 35% (7/20) for   (Table 5). Furthermore, these differences in the results between E. coli and K. pneumoniae isolates were statistically significant (Tables 5 & 6). The presence of more than one resistance encoding gene did not have an effect on the MIC 90 value in K. pneumoniae isolates against the tested carbapenems (results remained >32 µg/ml). However, in E. coli isolates the presence of resistance encoding genes bla NDM-1 +ve &bla CTX-M-15 +ve showed higher MIC 90 values when compared to the MIC 90 values of isolates positive for only one of those genes. Furthermore, when the three genes bla NDM-1 , bla OXA-48 &bla CTX-M-15 coexisted in E. coli isolates the MIC 90 values against IMP and MER increased however those against ERT decreased. However, it is important to note that since the number of isolates (N) is small (range from 2 to 14) when co-production of genes is concerned the conclusions may not be statistically significant in this regard (Tables 3 & 4).
Sequence analysis of the resistance-encoding genes confirmed that the PCR amplicons were positive for the gene in question. Sequence analysis of the OmpF and OmpCporins showed considerable variability for selected carbapenem-resistantK. pneumoniae andE. coli, (both carbapenemase-and/ or ESBL-producers) in relation to both of the porins when their sequences were compared with those of the reference strain E. coli K-12 and with sequences of other susceptible isolates of E. coli and K. pneumoniae. A number of resistant isolates showed nucleotide substitutions, deletions and insertions, sometimes of multiple nucleotides, throughout the coding region.

Discussion
In our medical center, the prevalence of CRE among E. coli and K. pneumoniae showed an increase between 2008 and 2014 from 0% in both to 1% and 4%, respectively [9]. In this study, the association between genes involved in CRE resistance and the level of MICs against carbapenems was determined to reveal a possible potential clinical application in this approach for treating patients with CRE infections.
The molecular investigation on these isolates showed that bla CTXM-15 is the most common resistance encoding gene in both E. coli and K. pneumoniae isolates followed by bla TEM-1 ,bla OXA-48 and bla NDM-1 . This is in concordance with other studies  [11] and bla NDM-1 in 2012 [12], studies have shown a trend of elevated cases [6,9 &13-17]. Also concerning the geographic region of the Middle East, there are reports implying the spread of bla CTXM-1 , bla TEM-1 ,bla OXA-48 and bla NDM-1 genes. In 2011, it was reported that the prevalence of bla CTXM-15 along with other ESBLs to be predominant in Egypt, Kuwait, Lebanon and the United Arab Emirates [9]. Also in 2013, a study done in Jordan reported high incidence of CTX-M ESBL-producing E. coli to be found associated with fluoroquinoloneresistance and Class I integrons colonizing the intestine of Jordanian infants [10].Plasmid that harbors OXA-48 among various Enterobacteriaceae, was first identified in Turkey in 2003, then in European countries and in the Middle East [19 &20]. As for NDM-1, it was first identified in India in 2008 and subsequently reported also from the Arabian Peninsula: United Arab Emirates, Sultanate of Oman, Kuwait and Saudi Arabia [21]. Determining the simultaneous levels of carbapenem MICs and the genes involved was partly in search of an association pertaining to treatment potential of CRE infections. This approach was not reported in any study from this region, however conducted elsewhere [22& 23] . In our study the MIC 90 revealed that E. coli isolates positive for bla OXA-48 and bla TEM-1 were resistant to the three carbapenems tested (ERT, IMP and MER), while those positive for bla NDM-1 and bla  showed to be resistant only to ERT and IMP, and susceptible to MER. On the other handK. pneumoniae isolates positive for any one of the three resistance encoding genes, showed resistance to all three carbapenems. Thereby, it may not be advisable to treat CRE infections due to isolates harboring these genes, subsequently with carbapenems. It is suggested that doripenem is more stable against emerging resistance than the other carbapenems [4].
ERT has shown to have higher MIC levels, in both E. coli and K. pneumoniae isolates positive for any of the resistance encoding genes. This phenomenon was explained by Tangden et al., where ERT has shown to have higher MIC levels in ESBL producing isolates as compared to other carbapenems being less affected by the action of the enzymes [4]. Whereby, resistance to IMP and MER is less common. ERT resistance is indeed more frequently reported than resistance to other carbapenems, especially in ESBL-producing strains. In addition, ERT has been heavily used for UTI infections caused by ESBL-producing bacteria and proven effective for these infections, which might explain the generated resistance to this agent [4].
It seems that ERT is more affected by the presence of bla OXA-48 and the other resistance encoding genes, thus ERT may be a good indicator for their detection [12]. Metallo-β-lactamases, such as bla NDM-1 , may result in low-grade reduced susceptibility to carbapenems in some instances [4]. This was also observed in our data, and may present a challenge in the detection of the CRE, if based on only the MIC results [16].
Knowing that K. pneumoniae isolates had MIC values ≥ 32µg/ml against the three carbapenems regardless of the type of the resistance encoding genes, it may be concluded that the lack of the porin genes (OMPC and/or OMPF), which was significantly higher in K. pneumoniae than in the E. coli isolates, plays a major role in increasing the MIC levels. Similarly, it has been shown that carbapenem resistance levels intensify when resistance encoding genes such as bla OXA-48 are combined with permeability defects [13].Yang et al. demonstrated similar results, whereby they confirmed that the combination of SHV-1, CTXM-3, CTXM-14, TEM-1 or OXA-11 production reduced the expression of OMPK-35/36 resulting in increased MIC, which may or may not lead to increased resistance clinically [23].
Furthermore, PCR and sequencing analysis of the K. pneumoniae and E. coliOmp encoding genes, revealed mutations associated with carbapenem resistance. On the other hand a number of K. pneumoniaecarbapenem resistant isolates lacked the OMP encoding genes. Similarly, Little et al. reported the development of carbapenem resistance in K.pneumoniae due to porin disruption by insertion sequence ISEcp1in CTXM-15 positive isolate [24]. Therefore, the E.coli isolatesin this study harboring any of the porin encoding genes, have shown mutations by sequence analysis. This may indicate that High MIC values detected in the TEM-1 positive isolates harboring both porin encoding genes,are mainly due to mutated porins.
The first report of a K. pneumoniae strain coproducing NDM-1, VIM-1 andOXA-48 carbapenemases isolated in Morocco was reported by Barguiga et al in 2012 [25]. Further, the first K. pneumoniae isolate co-producing OXA-48 and NDM-1 in Turkey was reported by Kilic et al [26].The emergence of carbapenem-resistant Gram-negative bacilli in the Mediterranean region, which is in fact the cradle of western civilization representing nearly 475 million inhabitants (6.3% of world population), is of importance, since intermingling of this population may explain the prominence of the dissemination of carbapenemase producers more rapidly in this region [27]. Thereby, these isolates showing co-production of resistance genes may be the interest of future studies .
In conclusion, levels of MIC in carbapenem resistance may largely depend on the type of carbapenemase and/or ESBL encoding genes in combination with the porin encoding genes. Such information may provide information for antibiotic regimen selection, epidemiological monitoring of resistance.
Funding: Medical Practice Plan (MPP), Faculty of Medicine, American University of Beirut