Klebsiella species are important opportunistic pathogens causing community-associated and nosocomial infections (Brisse et al., 2006). β-lactam antibiotics are widely used in treatment  of  many bacterial infections. The  persistent exposure of bacterial strains to β-lactams induces production of β-lactamases which are encoded by several genes and selections for mutations (Chaudhary and Aggarwal, 2004). Some of β-lactamases have activity even against 3rd and 4th generation cephalosporins and monobactams, these are called the extended spectrum β-lactamases (ESBLs). ESBLs are a heterogeneous group of plasmid-mediated enzymes (Bush and Jacoby, 2010), which are prevalent in Klebsiella pneumoniae. They are frequently isolated from patients with septicemia, pneumonia, or urinary tract infection (Nordmann et al., 2009). It is clearly known that prevalence of ESBL producers in any hospital depends on various factors; antibiotic policies and types of disinfectants used, especially in the Intensive Care Unit (ICU) (Wollheim et al., 2011). More than 300 different ESBL gene variants have been described (Paterson and Bonomo, 2005).

AmpC β-lactamases are a group of enzymes widely found in K. pneumoniae. They inactivate cephalosporins such as ceftazidime and cefotaxime. Plasmid-mediated AmpC enzymes have been reported since 1980s (Bradford et al., 1997). Most ESBLs arise as a result of mutations in the genes of TEM, SHV and CTX-M (Bradford, 2001). Carbapenems (for example, imipenem and meropenem) are considered one of the few last resort therapies for serious infections caused by multidrug-resistant (MDR) Gram-negative bacteria. The emergence of novel β-lactamases with direct carbapenem-hydrolyzing activity has contributed to the increased prevalence of carbapenem resistant Enterobacteriaceae (Gupta et al., 2011). Carbapenemases are enzymes that vary in their ability to hydrolyze carbapenems, and they are poorly inhibited by clavulanate (Robledo et al., 2011).

Carbapenemases can be divided into two types based on their mechanism of action. The first (for example, KPC and OXA types) uses a serine residue active site that covalently attacks β-lactam ring. The second (for example, VIM and NDM types) are the Metallo-β-lactamases (MBLs) that use Zn atoms to break the β-lactam ring (Walsh et al., 2005). OXA family encoded carbapenemases are most commonly produced by Acinetobacter spp. but have also been reported in K. pneumonia and E. coli (Mendes et al., 2009). The most recent MBL, the New Delhi MBL (NDM) showed hydrolytic activity against all β-lactams and has originated in India in 2009 and rapidly spread worldwide (Yong et al., 2009; Li et al., 2014).

In the present study, the antimicrobial susceptibility of Klebsiella clinical isolates was evaluated using disc diffusion method, and the causes of increased resistance to β-lactam antibiotics were determined both phenol- typically and genotypically using polymerase chain reaction (PCR).

Bacterial isolation

A total of 100 Klebsiella isolates were isolated from clinical specimens obtained from wounds, surgical wounds, urinary catheters, diabetic foot, burns, blood and sputum of patients admitted to Mansoura and Zagazig University Hospitals during the period from April, 2011 to November, 2011. All the isolates were collected under approved ethical standards. The isolates were identified and verified using standard biochemical reactions including: growth on MacConkey agar, citrate utilization, Voges Proskauer, methyl Red, and motility tests (Koneman et al., 1997).

Antibiotic susceptibility tests

The antibiotic resistance pattern of isolates was determined using the disk diffusion method according to the Clinical and Laboratory Standard Institute guidelines (CLSI, 2013). The antibiotic discs were obtained from Oxoid, UK. The tested antibiotics were ofloxacin (OFX, 5 µg), gentamicin (CN, 10 µg), aztreonam (ATM, 30 µg), imipenem (IPM, 10 µg), sulphmethoxazole/trimethoprim (SXT, 25 µg), amikacin (AK, 30 µg), cefepime (FEP, 30 µg), cefoperazone/sulbactam (SCF, 105 µg), cefrtiaxone (CRO, 30 µg), meropenem (MEM,10 µg), pipracillin/tazobactam (TPZ, 100/10 µg) and tetracycline (TE, 30 µg).

Detection of ESBLs producing isolates

Phenotypic detection of ESBL was done by modified double disc synergy test (DDST) according to Jarlier et al. (1998). Briefly, a disc of amoxicillin-clavulanate (20/10 μg) was placed on the surface of Muller Hinton agar (MHA) plates then discs of cefotaxime (30 μg) and ceftazidime (30 μg) were kept 20 mm apart from amoxicillin-clavulanate disc. The plates were incubated aerobically at 37°C overnight. The enhancement of the zone of inhibition of the cephalosporin discs towards amoxicillin-clavulanate disc was taken as evidence of ESBL production.

Determination of AmpC producing isolates

Phenotypic detection of AmpC was done according to Singhal et al. (2005). Briefly, E. coli ATCC 25922 was streaked on MHA plate. One colony of Klebsiella isolates were inoculated on a sterile moistened 6 mm disc. A cefoxitin disc was placed next to this disc (almost touching). The plates were incubated overnight at 37°C. A flattening of growth or the indentation of the cefoxitin inhibition zone was considered a positive test for the presence of AmpC mediated resistance.

Determination of MBLs producing isolates

Phenotypic detection of Metallo β-lactamases (MBLs) was done by combined disk test according to Yong et al. (2002). Briefly, MHA plates were surface inoculated with overnight broth culture of Klebsiella isolates. Two imipenem discs (10 μg) and two ceftazidime (30 μg) discs were placed on the plate surface. Five (5) μl of 0.5 M EDTA solution was added one imipenem and one ceftazidime disc. The inhibition zones of imipenem and ceftazidime and their EDTA-impregnated discs were compared after overnight incubation at 37°C. A zone size difference of ≥ 7 mm was taken as indicative of MBLs production.

Determination of carbapenemases producing isolates

Phenotypic detection of carbapenemases was done using modified hodge test [MHT] according to Lee et al. (2001). Briefly, a culture of E. coli ATCC 25922 was inoculated onto the surface of MHA plate. A 10 μg imipenem disc was placed at the center of the plate, the Klebsiella strains were streaked from edge of the disk to the periphery of the plate. The plate was incubated overnight at 37°C, presence of distorted inhibition zone ‘cloverleaf shaped’ of E. coli was considered as positive result for carbapenemase production by the test strain.

Genotypic detection of resistance genes

Colony PCR was done for detection of resistance genes in Klebsiella isolates, including the ESBLs genes (TEM, SHV, CTX-M), the Metallo- β- lactamase gene (NDM-1) and the carbapenamases genes (OXA-9 and KPC-1). The primers were designed and supplied from IDT (Intrgrated DNA Technologies, Coralville, Iowa, USA). The primers for NDM-1 and OXA-9, KPC were designed according to Geyerand Hanson (2013).  All primers are listed in Table 1. The gDNA was prepared according to Nair and Venkitanarayanan (2006). The PCR mixture was prepared in a final volume of 50 μl and contained: 25 μl of 2× DreamTaq TM Green Master Mix (Thermo Fisher Scientific, http://www. thermoscientific.com/), One μl of each primer, 5μl gDNA and nuclease-free water to 50 μl. The cycling conditions were: initial denaturation at 95°C for 3 min followed by 30 cycles of denaturation at 95°C for 5 s, annealing for 30 s at temperature (5° below primer Tm) as indicated in Table 1 and extension at 72°C for 1 min and a final extension cycle at 72°C for 5 min.

The resistance gene New Delhi Metallo β-lactamase (NDM-1) was amplified from gDNA of the seven imipenem and meropenem resistant Klebsiella isolates. It was observed that 3 isolates only harbored NDM-1 gene of ampilcon size 209 bp (Figure 3a). The resistance genes of carbapenemase (OXA-9 and KPC) were amplified from gDNA of the 7 imipenem and meropenem resistant Klebsiella isolates. OXA gene was observed in two isolates only (amplicon size 300 bp) as shown in Figure 3b, while KPC gene was not detected in any of the seven imipenem resistant isolates.

The Klebsiella species have been shown to be important opportunistic pathogens causing serious infections (Traub et al.,   2000).   Epidemics   caused   by MDR Klebsiella species have led to closures of hospital units or even whole hospitals (Casewell and Phillips, 1981).

In the present study Klebsiella species were isolated from various clinical specimens. The prevalence of Klebsiella in blood, urine, sputum, catheter, swabs and diabetic foot specimens were 56, 49.4, 45.7, 45.7, 44 and 20, respectively. Thosar and Kamble (2014) reported prevalence rate of 45 and 42.9% in sputum and urine, respectively. Our results were higher than those reported by Podschun and Ullmann (1998) and Acheampong et al. (2011). The relative high prevalence of Klebsiella may have been due to the increased resistance of these isolates which decreased the possibility of elimination of these isolates, leading to nosocomial outbreaks by these resistant isolates (Brisse et al., 2006).

In the present study, the result of AMS revealed that Klebsiella isolates exhibited high resistance to tetracycline (97 %), SXT (88.2%), ceftriaxone (85.3%), TPZ (73.5%), and aztreonam (73.5%). This high resistance rate is more than that reported previously (Acheampong et al., 2011; Midan et al., 2012; El- sharkawy et al., 2013). We concluded that these agents have a limited capacity in treating Klebsiella infections. The tested strains exhibited resistance to ofloxacin by 63.7%, while resistance to gentamicin was 43.1%. The data reveals that resistance to quinolones and gentamicin was less than that reported for 3^rd generation cephalosporins in this study (ceftriaxone, 87%). The excessive and over use of cephalosporins antibiotics lead to emergence of resistance and therefore shifting to alternative quinolones and aminoglycosides therapy (Sekowska et al., 2002). Less use of quinolone and structural dissimilarities is one of the reasons of comparatively low resistances than cephalosporins. The emergence of resistance to 4th generation cephalosporins (cefepime) was detected in our study by (53%). This susceptibility rate could be attributed to the empirical use of this antibiotic in the treatment of nosocomial infections in Egyptian hospitals.

Resistance to amikacin was 14.7%, which correlated with results obtained by Yasmin (2012). Resistance to imipenem and meropenem were 6.86% but higher than those reported by El-sharkawy et al. (2013). Yasmin (2012) reported that all Klebsiella isolates were sensitive to imipenem and meropenem. According to Rahal et al. (2008), carbapenems are the most effective therapy for ESBL bacterial infections. However, their routine use can select resistant strains leading to the emergence of imipenem-resistant K. pneumonia strains.

In the present study, 50% of Klebsiella isolates were ESBLs producing. This result was quite similar to that reported by Haque and Salam (2010) and Aladag et al. (2013) who reported 57.9 and 55%, respectively, but was higher than the 40% detected in France (Branger et al., 1998) and 13% in Hong Kong (Ho et al., 2000). High occurrence of ESBLs in Klebsiella spp is of great concern since they spread easily, and strains become more efficient at acquiring more resistance plasmids (Chaudhary and Aggarwal, 2004).

In the present study, 32.4% of Klebsiella isolates were AmpC producers. This result was higher than that reported by Shivanna and Rao (2014), but lower than that .found by Akujobi et al. (2012) and Fam et al. (2013). The increased presence of plasmid mediated AmpC β-lactamases worldwide is becoming of great concern (Jacoby, 2009). Infections caused by AmpC-positive bacteria cause higher patient morbidity and mortality (Livermore, 2012), demanding new measures of Klebsiella infections management. 14.7% of our isolates were both ESBL and AmpC producers, a result that underlines the need for new measures of management of Klebsiella infections.

In the current study all Klebsiella isolates that exhibit resistance to imipenem and meropenem were MBL and carbapenemase producers by phenotypic detection tests. These results come in complete accordance with the results of Bora et al. (2014). It is reported that carbapenemase-producing K. pneumoniae exhibit resistance to important antibiotics, such as amino-glycosides and fluoroquinolones (Livermore et al., 2011). Uncontrolled clinical utility of carbapenems is a serious worldwide threat.

PCR was used in this study to detect the presence of TEM, SHV and CTX-M, these were detected in all the ESBL isolates (100%). Our  results  come  in  accordance with Kiratisin et al. (2008) who detected in ESBL-producing K. pneumoniae, CTX-M (99%), SHV (87.4%) and TEM (71.7%); but, were higher than that reported in India, where of the 64 ESBL-Klebsiella isolates, 48% had TEM and 20% had SHV (Jain et al., 2008). Another study in USA showed 93% of SHV production and 20% of TEM production among ESBL Klebsiella isolates (Bradford et al., 2004).

The incidence of Carbapenem resistant entero-bacteriaceae has been increasing worldwide (Bhaskar et al., 2013). The main mechanism of resistance to carbapenem is by enzyme production. The most commonly identified enzymes are the carbapenemases OXA, KPC and the MBL enzyme NDM. They have been identified in different countries (Pfeifer et al.,  2012). In the seven imipenem resistance isolates, OXA-9 was detected in 2 isolates (28.5%). Kiratisin et al. (2008) reported OXA-10 in 11.8% of Klebsiella isolates in Thailand. However OXA genes were not detected in any of 15 Klebsiella isolates from Turkey (Bali et al., 2010).

None of our isolates have the carbapenem resistance gene KPC. Our results differed from those reported in USA  by  Bradford  et al. (2004), who found KPC in 100% of isolates. Also, Robledo et al. (2011) found KPC in 73% of 457 MDR Klebsiella isolates. In this study, the MBL NDM-1 was detected in 42.8% of imipenem resistant isolates. Chaudhary and Payasi (2013) detected NDM in 6.25% (in 24/150) of Klebsiella isolates.

We recommend that strict antibiotic policies and measures to limit unnecessary use of cephalosporins and carbapenems should be enforced to minimize the emergence of resistance to these antibiotics. Also rapid routine molecular detection of resistance determinants is essential to optimize therapy, improve outcomes, and limit the spread of such resistance.

The authors have not declared any conflict of interest