Frequency of quinolone resistance genes among extended-spectrum β-lactamase (ESBL)-producing Escherichia coli strains isolated from urinary tract infections

Background As an opportunistic pathogen, Escherichia coli (E. coli) is widely recognized as the main cause of nosocomial infections as well as some disorders especially those associated with urinary tract infections (UTIs). This study, therefore, sets out to determine the extent of antibiotic resistance to quinolones and to measure the frequency of qnr genes (A, B, and S) within extended-spectrum beta-lactamase (ESBL) and non-ESBL-producing strains of E. coli isolated from UTI-diagnosed patients as well as to investigate their antimicrobial susceptibility patterns for some selected antibiotics in southwest Iran. Methods Two hundred E. coli strains were isolated from UTI-diagnosed patients, hospitalized in nine different wards of Ahvaz Golestan Hospital between November 2015 and March 2016. The isolates were confirmed through well-practiced phenotypical methods. Moreover, the antimicrobial susceptibility test was successfully performed using a disk diffusion method. ESBL production among the isolates was screened by double disk synergism test (DDST), and the qnr genes were identified using a multiplex PCR. Results Out of the 200 samples collected, 167 isolates were confirmed to be E. coli strains. Maximum and minimum resistance were reported against nalidixic acid and chloramphenicol with 65.3% and 17.4%, respectively. Most of the isolates were resistant to all three types of quinolones studied in this research. Using multiplex PCR, the qnr genes were found in 100 (59.88%) strains (qnrA = 10, qnrB = 21, qnrS = 41, qnrB-S = 21, qnrB-A = 1, qnrA-S = 3, qnrA-B-S = 3), 58% of which was found among ESBL-producing isolates. Conclusions Resistance to quinolones antibiotics was highest among ESBL-producing isolates harboring, especially qnrS among other determinants of the qnr gene. There is a need for sensitive antibiotic stewardship especially in hospitals of Ahvaz, Khuzestan province. Further research is needed to ascertain the gravity of quinolones resistance in Iran and to quickly act against its spread among other nosocomial pathogens.


Background
As an opportunistic pathogen, Escherichia coli (E. coli) is regarded as the main cause of nosocomial infections, especially such medical disorders as UTIs (urinary tract infections), sepsis, aspiratory infections, and meningitis among neonates. The outbreak of plasmid-mediated quinolone resistance (PMQR) has steadily been increasing in recent years, leading to a decline in optimal therapeutic options [1].
Quinolones are a group of broad-spectrum antimicrobial agents that are helpful in the treatment of several infections caused by gram-negative bacteria, especially E. coli. However, over the past few years, the gram-negative bacteria have increasingly turned to be resistant to quinolones in Iran [1,2].
Quinolones are a group of wide-spectrum antimicrobials that are often used to treat E. coli infections. Several genes and systems involved in quinolone resistance Enterobacteriaceae isolates. Quinolone resistance gene (qnr), a member of PMQR determinants, was initially discovered in a strain of Klebsiella pneumoniae (K. pneumonia).
The first member of this family (qnr s ) is qnrA that was discovered in the USA in 1998 [1,3]. Qnr proteins are responsible for protecting the target enzymes including topoisomerase IV and DNA gyrase against quinolones such as ciprofloxacin and norfloxacin [4].
Qnr genes such as qnrA, qnrB, and qnrS are three types of the main classes of qnr determinants which have been identified in several members of Enterobacteriaceae family such as E. coli and K. pneumonia [3,5,6].
Over the past few years, researchers have discovered some other classes of the gene namely qnrC and qnrD [6,7]. The qnr gene product is composed of pentapeptide repeat proteins which protect the strains against the effects of quinolone antibiotics. Moreover, qnr gene products may result in 8 to a 32-fold rise in minimum inhibitory concentration (MIC) against these groups of antibiotics, enabling these types of proteins to reduce the ciprofloxacin activity. Structural modification in its structure may follow antibiotic pressure, leading to resistance to quinolones thereafter [8].
The extent of quinolone resistance in clinical isolates such as E. coli is unexpectedly high in Iran, especially in those strains that produce ESBLs [1]. Moreover, ESBLs genes have often been reported as being co-associated with genes that encode PMQR (qnr, aac (6′)-Ib-cr and qepA.) and sometimes other families of ESBL genes (blaSHV and blaTEM) have been found located together with qnr genes on a special plasmid in a pathogen isolate [9,10]. The qnr genes are widespread all around the world, especially in hospitalized patients.
This study, therefore, is set out to determine the extent of antibiotic resistance to quinolones and to measure the frequency of qnr genes (A, B, and S) within non-ESBL and ESBL-producing strains of E. coli isolated from UTI-diagnosed patients as well as to investigate their antimicrobial susceptibility patterns for some selected antibiotics in southwest Iran.

Patients and bacterial strains
Two hundred bacterial isolates from the urine of UTI-diagnosed patients were collected from the clinical laboratory of Golestan Hospital (affiliated to the Jundishapur Medical University of Ahvaz, Khuzestan province, Iran). The isolates were from both male and female outpatients and inpatients diagnosed with UTI who presented at different wards of Golestan Hospital such as urology, nephrology, ICU, internal, pediatric, transplant, psychology, and outpatient department (OPD) from November 2015 to March 2016. The isolates were then transported to the Department of Medical Microbiology in the Jundishapur University of Ahvaz for further processing and analysis as outlined below. The cases were divided into nine age groups (1, 0-10; 2, 11-20; 3, 21-30; 4, 31-40; 5, 41-50; 6, 51-60; 7, 61-70; 8, 71-80; 9, 81-90). Out of the 200 clinical bacterial isolates collected from hospital laboratory, 167 isolates were confirmed as bacilli with the ability to grow in MacConkey agar and fermented lactose (overnight), tested negative with Gram's staining, oxidase test, triple sugar iron agar, SH 2 -indol-motility, Voges-Proskauer, urease test, and citrate utilization test but positive with motility, indole, and Methyl Red tests [11]. All 167 isolates were confirmed using the API20E identification kit (bioMerieux SA, Marcy l'Etoile, France) to E. coli [12].

ESBLs screening test
A DDST method was used to screen for ESBLs production among the tested isolates. To all isolates that were identified as resistant strains to ceftazidime or cefotaxime (on MHA media), a double disk (containing clavulanic acid and ceftazidime or clavulanic acid and cefotaxime) was placed (near the single disk). If there was ≥ 5 mm extra zone diameter difference between the single disk and the double disk, after overnight incubation, the test strain was reported as ESBL positive. K. pneumoniae ATCC 700603 and E. coli ATCC 25922 were used as positive and negative control strains respectively, in accordance with CLSI guidelines.
Detecting qnr (A, B, and S) genes by multiplex PCR technique At first E. coli isolates were cultivated in Luria-Bertani broth at 37°C overnight. At the second step, four to five colonies of the isolates were transferred to a microtube containing 1.5 ml DW (water), and finally, the bacterial DNA was extracted for molecular analysis [3] via boiling method. Based on the sequence arrangements of the target genes, six primers were used [14] to amplify the internal fragments of 580 bp, 264 bp, and 428 bp for qnrA, B, and S genes, respectively. Nucleotide sequence primers used for detection of qnr (A, B, and S) genes by the multiplex PCR method are presented in Table 1. The multiplex PCR was performed in a 50 μl reaction mixture, containing 2 μl of total DNA, 6 μl of the primers (1 μl for each), 22 μl of the DW, and 20 μl of the master kit (Sinaclon, made in Iran). Target fragments were amplified using a thermal cycler (Eppendorf, Germany) as follows: initial denaturation at 94°C for 4 min, 30 cycles of amplification consisting of denaturation at 94°C for 30 s, annealing at 53°C for 45 s, extension at 72°C for 30 s, and a final extension stage at 72°C for 6 min. The electrophoresis of the PCR products was performed on 2% agarose gel stained with ethidium bromide (0.5 μg/ml) as follows: 5 μl of PCR product was mixed with 1 μl of loading buffer, and then run at 110 V for 50 min in 1× TBE [tris 54 g(pH ± =8-8.3), 27.5 g boric acid, 20 ml EDTA] solution. All positive controls for qnrA, B, and S were obtained from Pastor Institute, Iran. The products of these genes (qnrA, B, and S) were sequenced (Macrogen Company, Korea) and confirmed in the NCBI database (http://blast.ncbi.nlm.nih.gov/).

Statistical analysis
The statistical analysis was done using SPSS software (IBM, Chicago, IL, USA) version 20. For the purposes of this study, Fisher's test and the non-parametric chi-squared test were performed and P value < 0.05 was considered statistically significant.

Results
Out of the 200 urinary bacterial isolates collected from the hospital, a total of 167 isolates was confirmed as E. coli. Most of the isolates were gathered from OPD (59%) (Fig. 1) Table 2). The results of ESBLs screening revealed that 64.07% (107/167) of the tested E. coli strains were ESBL positive, out of which 65.91% were isolated from the outpatients and the rest from the inpatients. The majorities of ESBL-producing bacteria were isolated from females (66.35%) and were also gathered via OPD, whereas the least proportion of the isolates (1.19%) was collected from psychology wards. The qnr genes were successfully amplified by multiplex PCR technique (Fig. 2), and the results of this technique revealed that 100/167 (59.88%) of the isolates harbored the qnr genes with the following distributions: qnrA, n = 10; qnrB, n = 21; qnrS, n = 41; qnrB-S, n = 21; qnrB-A, n = 1; qnrA-S, n = 3; qnrA-B-S, n = 3 (Table 3). Fifty-eight percent (58/100) of the qnr genes-carrying isolates were ESBL producers (Table 3). Among the ESBL and qnr positive strains, 9/58 (15.5%) were sensitive to all quinolone antibiotics (Fig. 3). The study also revealed that 65.85% of the 41 isolates carrying only qnrS gene were resistant to all the three types of quinolone antibiotics (NA, OF, Cip), suggesting that qnrS gene seems to result in more resistance to quinolones than qnrB and qnrA (Table 4). A significant correlation (P value < 0.05) was observed between ciprofloxacin and ofloxacin among the strains harboring qnrS genes (Table 4). Moreover, in cases where a combination of qnrS and B was detected, the resistance to quinolones escalated to over 95% ( Table 4). The relationship between ESBLs production, presence of qnr genes, and the quinolone resistance patterns among the tested strains is shown in Fig. 3.

Discussion
Over the last decades, E. coli has been recognized as one of the leading causes of nosocomial infections in the world. In principle, some genes located on the especial plasmids in such strains are regarded as being responsible for such problems [15,16]. By developing some proteins, the genes (qnrs, ESBLs) build up resistance to antibiotics [17]. Though qnr genes and ESBLs are the most significantly resistant agents (against β-lactam and quinolone drugs), recently, the rate of resistance to other antibiotics has sharply risen [1,17]. In our study, the majority (68.26%) of the UTI-diagnosed patients were women. This might be related to the anatomical position of women ' s urethral tract closer to the anal canal; in addition, most of these patients belong to the third and fourth age groups (aged between 21and 40 years)-the most sexually active age group. Thus, it could be concluded that sex and age are two important factors regarding the prevalence of UTIs, as reported by Rahn [18]. In the current study, the isolates showed the highest resistance to NA (65.3%) and least resistance to CL (17.4%). Therefore, it appears that the use of nalidixic acid for the treatment of E. coli associated with UTIs in this center might be ineffective. However, since chloramphenicol is not routinely used clinically, aztreonam can still be a drug of choice for treating UTIs. Furthermore, Fig. 1 The percentage of qnr genes among Escherichia coli isolates in nine wards. ICU, intensive care unit; OPD, outpatient department  the low resistance to chloramphenicol found in the study could be related to its low usage in routine treatments for UTIs. Interestingly, most of the isolates were resistant to all the quinolone-containing antibiotics, ( Table 2). This could be due to the reckless and inappropriate usage of the drugs [19]. Unfortunately, the rate of quinolone resistance in southwest Iran (Ahvaz) is significantly higher than other parts of Iran as reported by Pakzad et al. and even higher than what is found in other parts of the world [20]. According to the findings of this study, 64.07% of the isolates were ESBL producers. A lower rate of ESBL-producing strains has been reported in Japan and the USA [21,22]. In Korea, it was reported to be 17.7% and 84% of ESBL-positive E. coli strains in 2003 and 2009, separately [15,17]. Mood et al. reported that 42.5% of the E. coli isolates they studied were ESBL producers and this is lower than the finding of our study. The frequency of ESBL-producing E. coli was, therefore, higher in our study (except for the Korean study in 2009) than what has been reported from other countries and other Iranian metropolises [19]. Arbitrarily speaking, this high rate of ESBLs isolates in Iran and the lower frequency in some other countries could be attributed to several factors including but not limited to drug abuse pattern [23]. Studies have been conducted for qnr genes on E. coli strains in some Asian countries such as Korea (qnrB 5.6%), Japan (qnrA 6.5%), China (qnrA 8%) and Iran (qnrA 6%, qnrB 2.66%). Several pieces of research have been carried out in Europe (UK and Germany) and the USA on qnr genes among isolates of E. coli and other members of Enterobacteriaceae, but only Robicsec et al. reported qnrA (4%) in the USA [24,25]. Moreover, the total percentage of E. coli isolates harboring qnr genes (59.88%) reported in this study, as shown in Table 3, is significantly higher in number and diversity than those reported in Europe, Asia, and the USA [23,26,27]. It seems there is a high frequency of transconjugation or transformation mechanisms of the genes occurring in southwestern Iran, and this could be attributed to the hot and humid climate [23] of southwest Iran. Fifty-eight percent of the strains harboring qnr genes were also ESBL producers (Table 3)   and showed the highest level of resistance to quinolones (ESBL and qnr positive strains, as shown in Fig. 3). Moreover, the lowest level of resistance was seen among qnr and ESBL negative strains (Fig. 3). Contrary to our expectations, seven (4.19%) qnr and ESBL-negative isolates were resistant to all quinolones mentioned, which might be due to other different defensive mechanisms. On the other hand, nine (5.38%) qnr and ESBL-positive isolates were absolutely sensitive to all quinolones, most probably the resistance genes in those strains were silent [23,28,29]. Out of the total isolates harboring qnr genes, 10% (10/100) isolates contained qnrA, 21% (21/100) isolates contained qnrB, 41% (41/100) isolates contained qnrS, and 21%(21/100), 1% (1/100), 3% (3/100), and 3% (3/100) isolates contained a mix of qnrS-B, qnrB-A, qnrA-S, qnrA-B-S, respectively. The qnrS gene seems to play a more significant role than qnrA and qnrB with regard to quinolones resistance since 65.85% (27/41) of the isolates harboring the qnrS gene were resistant to all quinolones ( Table 4). As shown in Table 4, the combination of qnrS and B detected in some strain have increased the level of resistance to quinolone-containing antibiotics nearly to 100% (20 of 21), suggesting that the synergistic effects resulting from the combination of qnrS with qnrB in a strain may increase their defensive positions. Also, a significant correlation (P value < 0.05) was reported between qnrS gene and resistance to Cip and OF (Table 4). Attaining complete cure from UTIs caused by E. coli is highly challenging, and the presence of qnr genes and ESBLs production are said to be most responsible. However, some other factors such as the overexpression of efflux pump or declined intracellular concentration and mutations (in topoisomerase IV and DNA gyrase enzymes) could also be equally effective in this regard [12,30,31]. On the other hand, quinolones are the most prescribed antibiotics for UTI therapy in Iran. Today, one of the most important concerns is the risk of treatment failure among Iranian patients, especially those diagnosed with UTIs [20,32]. To the best of our knowledge, this is the first study to investigate the presence of qnr genes (A, B, and S) in both ESBL-producing and non-ESBLs strains of E. coli isolated from patients diagnosed with UTIs and their resistance patterns to quinolones antibiotics in southwest Iran. Finally, the major limitations were small sample size, study of only seven drugs in terms of antimicrobial susceptibility, and single center-based design of this study.

Conclusion
The frequency of qnr genes among clinical isolates of E. coli demonstrated in this study is higher than the usually reported frequency from other provinces of Iran. Resistance to quinolones antibiotics was highest among ESBL-producing isolates, especially those harboring qnrS among other determinants of qnr gene. There is a need for sensitive antibiotic stewardship especially in the health facilities of Ahvaz, Khuzestan province. Further research, with a wider scope, is needed to ascertain the gravity of quinolones resistance in Iran and to quickly act against its spread among other nosocomial pathogens.