Molecular characterization of Extended-spectrum β lactamase- producing E. coli recovered from community-acquired urinary tract infections in Upper Egypt

Treatment of community urinary tract infections (UTIs) caused by extended-spectrum β lactamase (ESBL)- producing Escherichia coli (E. coli) is more expensive than treating ESBL-negative opposites. Evaluation of the prevalence of ESBL-production among urinary E. coli isolates is crucial due to its great impact on the choice of proper antimicrobials. Accordingly, the aim of this work was to detect and characterize ESBL-producing E. coli isolated from outpatients with signs of UTIs in Upper Egypt. Urinary E. coli isolates were identified by 16S rRNA and their ESBL-production was confirmed by Modified Double Disc Synergy Test (MDDST) and ESBL- CHROMagar media. Isolates were then subjected to Polymerase Chain Reaction (PCR) for new Clermont phylogrouping, ESBL genes detection and CTX-M typing. The study enrolled 583 patients with clinically diagnosed UTIs. Uropathogens were found in 400 urine samples (68.6%) out of which 134 E. coli isolates were identified. Among the examined uropathogenic E. coli (UPEC), 80 (59.7%) were recognized as ESBL-producers. Greater than half of the ESBL-producers were multi-drug resistant (MDR) (62%). All of them were susceptible to meropenem. Most of the E. coli isolates were distributed in 4 phylogenetic groups: B2 = 42 (52.5%), F = 17 (21.25%) and Clade I or II = 10 (12.5%). The predominant gene types were TEM 60 (75%) and CTX-M gene 45 (56.25%). The CTX-M-1 group was the most prevalent (62.2%), including the CTX-M-15 enzyme, followed by the CTX-M-2 group, CTX-M-8 group and CTX-M-9 group. In conclusion, the results present alarming evidence of a serious spread of ESBL genes in Egypt, especially the epidemiological CTX-M 15, with the potential for the dissemination of MDR UPEC strains in the community.

www.nature.com/scientificreports www.nature.com/scientificreports/ The spread of ESBLs in Enterobacteriaceae has become an ever-increasing problem 6 , with a global rise of ESBL-producing Enterobacteriaceae 7 . One of the most frequently found Enterobacteriaceae harboring ESBL genes is Escherichia coli (E. coli), where multi-drug resistance (MRD) due to ESBL production is rapidly becoming a threat to the community 8 . Spreading rates of nosocomial ESBL producing E. coli are markedly variable, with flat trends in Europe ~15%, and increasing trends in North America from 7.8% in 2010 to 18.3% in 2014 9 . In fact, propagation of MDR and ESBL-producing E. coli strains reduces the treatment preferences. It is also mandatory to be informed with the predominant resistant pattern of any region, which could assist in proper antimicrobial therapy 10 .
A recent study done by our group 11 , showed a disturbingly high level of ESBL producers in the urine of asymptomatic pregnant women. Thus, we aimed at evaluating their incidence in patients suffering from UTIs in the community. To the best of our knowledge, there are no previous epidemiological data regarding the phylogenetic grouping of ESBL-producing E. coli causing UTIs in Egypt. Accordingly, a phenotypic and a genotypic evaluation of ESBL-producing E. coli was carried out, followed by phylogenetic grouping of the obtained isolates.

Results
isolation of E coli from urine samples. Out of 583 urine samples obtained from patients suffering from UTIs, 400 isolates were confirmed as positive cultures. 134 of these cultures were confirmed to be E. coli (33.5%) ( Figure S1). phenotypic detection for esbl production. Out of 134 E. coli isolates, 80 (∼60%) were ESBL producers by CHROMagar ESBL screening and 75 (56%) using Modified Double Disc Synergy Test (MDDST) (Fig. 1). There was no statistical significance between CHOROMagar and MDDST regarding detection of ESBL-production (P > 0.5; X² = 0.245). Different patterns of synergism of 3 rd generation cephalosporin and 4 th generation cephalosporin with Amoxicillin Clavulanate using MDDST were observed on testing 80 CHROMAgar ESBL producing E. coli isolates (Table 1).

Molecular detection of eSBL genes.
The predominant ESBL gene in this study was bla TEM , which was found in 60 isolates (75%) ( Figure S2). The bla CTX-M gene was found in 45 isolates (56.25%), whereas bla SHV was found in only 15 isolates (18.75%) ( Figure S3 & S4). Most of the isolates (66.25%) showed coexistences of more than one gene, with (28.75%) of ESBL-producing E. coli harboring bla TEM , bla SHV , and bla CTX−M . Coexistences of two genes was also observed where bla TEM and bla CTX−M were detected in (21.25%) of the isolates, bla TEM and bla SHV in (6.25%) of the isolates, while bla SHV and bla CTX−M were found in (10%) of the isolates. About 25% of the isolates harbored bla TEM alone, and 13.75% had bla CTX-M alone. None of the isolates had bla SHV alone. Distribution of tested ESBL genes among different groups is found in Table 4.  Figure S7). Group 2 enzymes were produced by 12 (26.7%) isolates. The rest of the isolates included: 10 isolates that produced group 8 CTX-M enzymes and one isolate that produced group 9 enzyme. On the other hand, 3 E. coli isolates harbored both group 1 and group 8 enzymes. There were no producers of group 25 CTX-M enzymes detected in this study.

Discussion
Investigating the prevalence of antimicrobial resistance rates is of great importance in both creating strategies for empirical treatment and in evaluating the existing guidelines. The frequencies and types of infections caused by ESBL-producing Enterobacteriaceae have increased dramatically in the past few decades with disparity between different institutions and countries. Since the beginning of the new millennium, E. coli has become the most commonly isolated ESBL-producing bacteria worldwide with CTX-M ESBLs being the most frequently isolated types 13 . This upsurge in ESBL-producing E. coli adds a great burden to the treatment of community-onset UTIs as such isolates are frequently multidrug-resistant, with increased chance of treatment failure [14][15][16] . Since there is no comprehensive surveillance of community-acquired UTIs caused by E. coli in Egypt, this study was aimed at evaluating the prevalence and the mechanisms underlying their ESBL production. To the best of our knowledge, this is the first article from Upper Egypt to report the characteristics of ESBL-producing E. coli from community-onset UTIs.
Out of a total of 134 clinical isolates of E. coli, 80 isolates (59.7%) were ESBL positive. This high frequency of ESBLs is comparable to those found in Egypt by Al-Agamy et al. and Abdel-Moaty et al. 17   www.nature.com/scientificreports www.nature.com/scientificreports/ the detected isolates were ESBL producers, however, their study did not demonstrate whether the isolates were community or hospital-acquired. On the other-hand, our reported incidence is much higher than that reported earlier in Egypt by Fam et al., with a lower prevalence of 17% among community-acquired UTIs 18 , suggesting an increasing trend in the incidence of ESBLs-producing E. coli in Egypt. Compared to countries in the same region, our studies were also higher than that found in Lybia (6.7%) 19 and Emirates (39%) 20 . However, comparable results (67%) were obtained by Zorgani et al. (in a different larger-scale study done in Lybia), who reported a high incidence of ESBL-producers among hospital isolates 21 .
Within different ESBLs, CTX-M enzymes are the most predominant in different epidemiological settings, which have outnumbered other ESBL enzymes such as TEM and SHV 22,23 , with more than 172 CTX-M variants reported to date. Al-Agamy et al. reported the first detection of CTX-M β-lactamase production by urinary nosocomial E. coli isolates in Egypt. They found a high incidence of ESBLs-producing isolates in a single hospital (60.9%) 24 . In this study the predominant gene types were bla TEM in 60 isolates (75%) and bla CTX-M in 45 isolates (56.25%), while bla SHV was found in only 15 isolates (18.75%). The relatively higher frequency of bla CTX-M among our community isolates is concurring with the fact that CTX-M ESBLs originate from environmental bacteria unlike TEM-or SHV-ESBLs 25 . These findings are in agreement with previous studies done in Egypt where bla CTX-M was prevailing 24,26 .
Interestingly, bla TEM was detected in most isolates although it is commonly found in hospital strains, this could probably be due to previous contact with hospitals.
Among bla CTX-M , bla CTX-M-15 was the most prevalent (37.8%) in our study, which concurs with various reports demonstrating the extensive worldwide dissemination of bla CTX-M-15 mediated by clonally related E. coli strains 27 .
Regarding the susceptibility profile of ESBLs-producing isolates; all of the isolates were resistant to cephems, which is concurring with previous studies 28,29 .
On the other hand, about 13% of the isolates were resistant to ciprofloxacin using CLSI disk breakpoints 32 , which is much lower than previously reported in Egypt by Abdel-Moaty et al. 17 or in the Middle East region 19,33 . This decreased resistance rate could be due to a better understanding and a wiser use of fluoroquinolones in UTI cases.    www.nature.com/scientificreports www.nature.com/scientificreports/ In addition, all the isolates were susceptible to meropenem, which is in consensus with a previous study done in the same region on asymptomatic urinary carriers of ESBL-producer strains 11 , who reported that all ESBL producers were sensitive to imipenem (100%).
Alteration in the phylogenetic types are important in identifying novel groups of emerging bacteria that are better recognized as a result of this analysis. Phylogenetic grouping in this study was done according to a modified Clermont method 12 , which was done for the first time on the Egyptian isolates. This new method used modified primers for chuA, yjaA and TspE4.C2, which eradicated some primer mismatches. The most imperative benefit of this method was its power to distinguish strains belonging to phylogroups C, E, F and clade I. In this study most of the isolates were in group B2 (52.5%), followed by group F (21.25%), Clade I or II (12.5%), while 13.75% were of unknown type. The high frequency of phylogenetic group of B2 (52, 5%) is comparable to previous studies [34][35][36] , where the B2 subgroup was the most common group especially among CTX-M15 strains, as well as phylogroup F, which is closely similar to phylogroup B2 37 .
The presence of Clade I or II isolates (cryptic lineages which are phenotypically similar to E. coli but genetically divergent 38,39 ) require further investigation as this is the first report for the presence of such clades among extra intestinal isolates in Egypt. Interestingly, cryptic isolates found in this study harbor a variety of ESBLs: TEM, SHV and CTX-M, suggesting a threatening horizontal gene transfer in our community. Environmental spread of ESBL-producing E. coli could be attributed to the release of wastewater into rivers 40,41 , where mobile genetic elements are allowed to transfer ESBL-production from environmental bacteria to human and animals 42 .
In conclusion, our data underscores the importance of continuous surveillance of antimicrobial resistance in community E. coli isolates and shows the alarming increases of ESBL-production among such strains. Public health efforts should focus on the correct use of antibiotics to limit their dissemination and further investigation of molecular epidemiology of ESBLs in various clinical samples would be promising to obtain a better database for ESBL-producing E. coli in Egypt.

Methods
Study design. This cross-sectional study was conducted to assess the prevalence and antimicrobial resistance pattern of ESBL-producing E. coli isolates from patients with suspected community-onset UTIs during the period of August 2016 to February 2018 from Minia General Hospital, Kidney Hospital, Suzan Mubarak Hospital and Liver Virus Unit in Minia Governorate (located in Upper Egypt). Community-onset infections were defined as infections that have an onset in less than 48 hours of hospital admission or that present in the outpatients' 43 . Study recruits in this work were patients ≥18 years with symptoms of suspected urinary tract infections, attending the outpatient's clinics or admitted to the inpatient's wards (within 48 hours of admission). Written informed consent was obtained from the patients prior to data collection. The methods were carried out in accordance with the relevant guidelines and regulations. All experimental protocols were approved by the Ethics Committee of the Faculty of Science, Beni-Suef University. Patients with history of antibiotics intake within the last 2 weeks were excluded.
Sample size. Before the study, the number of required patients was determined after a power calculation according to data obtained in a previous study carried in Assiut, Egypt 44 . In that study the frequency of ESBL-producing E. coli was about 6.8%. A sample size of 80 patients in the group was determined to provide 80% power and 5% type I error with precision of 5.5% using the following equation:   46 . Briefly, isolates were inoculated on a plate containing a disc of amoxicillin-clavulanate (20/10 μg) at the center along with three 3 rd generation cephalosporins; cefotaxime, ceftriaxone, cefpodoxime and a 4 th generation cephalosporin; cefepime placed at 15 mm and 20 mm, respectively, centre to centre to that of the amoxicillin-clavulanate disc 47 . Any increase in the zone towards the disc of amoxicillin-clavulanate was considered as positive for ESBL production.
Antibiotic susceptibility pattern of E. coli isolates. Antimicrobial susceptibility testing for phenotypically confirmed ESBL-isolates was determined by the disk diffusion method with reference to the standards of the Clinical and Laboratory Standards Institute 32 . The quality of antibiotic sensitivity was confirmed by using E. coli (2020) 10 phylogenetic grouping by quadruplex pcR. A quadruplex polymerase chain reaction (PCR) 12 modified from the original triplex PCR method by Clermont and colleagues 48 was used to group the E. coli isolates phylogenetically. This method was used to allocate the ESBL-producing E. coli isolates based on the presence or absence of 4 genes (arpA (400 bp), chuA (288 bp), yjaA (211 bp), TspE4.C2 (152 bp)) and allocating E. coli isolates into 1 of 8 phylo-groups (A, B1, B2, C, D, E, F and cryptic clade).
Statistical analysis. Analyses of data were performed by SPSS software (version 23) and proportions were compared using the Chi-square test (χ2 test) to determine the significant differences in resistance. Differences were considered significant at P < 0.05. Graphics were performed using Excel 2010. For description statistics, data are presented as mean standard deviation for continuous variables, as well as frequency and percentage for categorical variables.