An evaluation of multidrug-resistant Escherichia coli isolates in urinary tract infections from Aguascalientes, Mexico: cross-sectional study

Background Uropathogenic Escherichia coli (UPEC) are one of the main bacteria causing urinary tract infections (UTIs). The rates of UPEC with high resistance towards antibiotics and multidrug-resistant bacteria have increased dramatically in recent years and could difficult the treatment. Methods The aim of the study was to determine multidrug-resistant bacteria, antibiotic resistance profile, virulence traits, and genetic background of 110 E. coli isolated from community (79 isolates) and hospital-acquired (31 isolates) urinary tract infections. The plasmid-mediated quinolone resistance genes presence was also investigated. A subset of 18 isolates with a quinolone-resistance phenotype was examined for common virulence genes encoded in diarrheagenic and extra-intestinal pathogenic E. coli by a specific E. coli microarray. Results Female children were the group most affected by UTIs, which were mainly community-acquired. Resistance to trimethoprim–sulfamethoxazole, ampicillin, and ampicillin–sulbactam was most prevalent. A frequent occurrence of resistance toward ciprofloxacin (47.3%), levofloxacin (43.6%) and cephalosporins (27.6%) was observed. In addition, 63% of the strains were multidrug-resistant (MDR). Almost all the fluoroquinolone (FQ)-resistant strains showed MDR-phenotype. Isolates from male patients were associated to FQ-resistant and MDR-phenotype. Moreover, hospital-acquired infections were correlated to third generation cephalosporin and nitrofurantoin resistance and the presence of kpsMTII gene. Overall, fimH (71.8%) and fyuA (68.2%), had the highest prevalence as virulence genes among isolates. However, the profile of virulence genes displayed a great diversity, which included the presence of genes related to diarrheagenic E. coli. Out of 110 isolates, 25 isolates (22.7%) were positive to qnrA, 23 (20.9%) to qnrB, 7 (6.4%) to qnrS1, 7 (6.4%) to aac(6′)lb-cr, 5 (4.5%) to qnrD, and 1 (0.9%) to qnrC genes. A total of 12.7% of the isolates harbored blaCTX-M genes, with blaCTX-M-15 being the most prevalent. Conclusions Urinary tract infection due to E. coli may be difficult to treat empirically due to high resistance to commonly used antibiotics. Continuous surveillance of multidrug resistant organisms and patterns of drug resistance are needed in order to prevent treatment failure and reduce selective pressure. These findings may help choosing more suitable treatments of UTI patients in this region of Mexico.


Background
Escherichia coli is an important cause of extra-intestinal infections, enteric disease, and systemic infections in humans and animals. Uropathogenic Escherichia coli (UPEC), one of the members of the extra-intestinal pathogenic E. coli (ExPEC) is a predominant pathogen causing urinary tract infections (UTIs) [1]. UPEC is one of the main causes of community (80-90%) and nosocomialacquired UTIs (30-50%) [2]. These strains harbor a variety of virulence factors that allow them to establish an infection, including adhesins, toxins, host defense avoidance mechanisms and multiple iron acquisition systems [3].
UTIs are one of the most common bacterial infections worldwide. It has been estimated that 150 million UTIs occur per year worldwide [4]. ExPEC strains are responsible for an estimated 40,000 deaths and at least $2.6 billion of expenses in healthcare treatment in the United States alone [5]. In Mexico, 3'648,784 cases were reported on 2015 [6]. Although these infections are treatable, the increasingly accelerated rates or multi-drug resistant organisms (MDRO) lead to complication, treatment failure and increased rates of mortality and morbidity [7].
The enhanced prevalence and global spread of extended-spectrum beta-lactamase (ESBL) genes [8] such as CTX-M enzymes that are associated to multidrug-resistant (MDR) phenotypes and fluoroquinolones resistance, has become a major concern [9,10]. Quinolones are broad-spectrum antibiotics and very important in the treatment of a wide range of diseases, especially urinary tract infections. Quinolone resistance is mainly caused by chromosomal mutations in the quinolone resistance-determining region (QRDR) of gyrase (gyrA and gyrB) and topoisomerase (parC and parE) genes, but also is caused by plasmid-mediated quinolone resistance (PMQR) which include: the pentapeptide repeat family Qnr proteins (QnrA, QnrB, QnrS, QnrC and QnrD) [11], the variant of the aminoglycoside-acetyltransferase modifying enzyme, AAC(6′)-lb-cr [12], and the efflux pumps QepA and OqxAB [13][14][15]. Although PMQR genes confer only low-level of resistance to quinolones [16] they could spread horizontally among Enterobacteria and facilitate the selection of resistant mutants as well as the selection of additional resistance mechanisms, enabling bacteria to become fully resistant [8,17].
The emergence of MDR E. coli causing urinary tract infections with high virulence potential is alarming. In Mexico, a high rate of resistance against quinolones has been observed in environmental, diarrheagenic [18], and pediatric clinical isolates [19]. MDR phenotype of UPEC has also been shown [20,21]. However, lack of sufficient data on virulence spectrum and MDR-UPEC isolates from community and hospital infections hinders the infections control and management efforts. The aim of this study was to determine the occurrence of MDR-phenotype, antibiotic resistance profile, virulence traits, and genetic background of E. coli isolates from community and hospital-acquired UTIs from Mexico.

Escherichia coli isolation
This study was conducted on a total of one hundred and ten urine cultures collected from patients suspected to have an UTI (urine samples contained bacterial counts to ≥ 10 5 CFU/ml) and sought attention in the Centenario Hospital Miguel Hidalgo, in Aguascalientes, Mexico during the period June to November 2012. Patient's ages ranged from neonatal to 91 years. Only one nonduplicated E. coli isolate per culture was considered. Patients from whom E. coli was isolated at least 48 h after admission were considered to have a hospitalacquired infection, all other infections were considered to be community-acquired [22]. To isolate E. coli, serial decimal dilutions of the sample were prepared in 0.85% NaCl, and were used to inoculate MacConkey agar plates which were incubated overnight at 37 °C. All the isolates were screened by PCR for the uidA gene as confirmation ( Table 1). The E. coli isolates were stored at − 80 °C in Luria-Bertani broth and 20% (vol/vol) glycerol.

Detection of virulence factors
Escherichia coli isolates from urinary tract infections were screened for the presence of commonly detected UPEC virulence genes including: fyuA, fimH, sfaS, afa/ dra, papC, agn43, yfcV, vat, hlyA, cnf1and kpsMTII. Primers and controls used in this study are listed in Tables 1 and 2.

Characterization of resistance genes and mutations
Screening for plasmid mediated quinolone resistance (PMQR) genes, including qnrA, qnrB, qnrS, qnrC, qnrD, qepA, and acc(6′)-lb genes was performed as previously described [13,[25][26][27][28]. The qnrC, qnrD and aac (6′)-Ib cr variants were identified sequencing the PCR products. The quinolone resistance determining regions (QRDR) of gyrA and parC genes were amplified and sequenced on both strands and predicted peptide sequences were compared to the corresponding gene from the MG1655 genome as described by Namboodiri et al. [29]. Beta-lactams genes detection including bla TEM , bla SHV , bla OXA , bla PSE and bla CTX-M were carried out by PCR amplification using specific primers [30]. Amplicons from all of the bla CTX-M positive isolates were sequenced. The nucleotide sequences obtained were aligned and further analyzed by comparison to sequences from a catalogue of beta-lactamases (http://www.lahey .org/studi es/).

DNA microarray analysis
Microarray hybridizations were performed using E. coli maxivirulence version 3.1 microarray as previously described [31]. It allows the detection of 348 virulence genes and 98 antibiotic resistance genes and variants. DNA extraction and hybridizations were performed as described previously [32]. Each isolate was assigned to a specific E. coli pathotype according to its virulence gene profile and based on the classification published in previous papers [31,33].

Statistical analysis
Categorical variables were compared using Pearson's Chi square or Fisher's exact test as appropriate. Continuous variables were compared using Mann-Whitney U test (STATISTICA V. 10, StatSoft, United States). p values less than 0.05 were considered statistically significant.

Patient demographics
During the period of study, females (86 samples, 78.2%) were the most affected group of patients (p < 0.05) as compared to males (24 samples, 21.8%). Most of the UTIs were detected in community obtained samples (79 samples, 71.8%). Thirty-one samples were recovered from hospital-acquired infections (28.2%). E. coli isolated from female patients were correlated with communityacquired UTIs (p = 0.0297). No differences among the prevalence of E. coli isolated from male patients among community-and hospital-acquired infections were found (54.2% vs 45.8%, respectively). Maximum number of cases was found in female children in the age group of 2-10 years (34 samples, 39%, p = 0.0372, Table 3).

Identification of virulence, and antimicrobial resistance genes by microarray analysis
Microarray analysis was done on 18 E. coli isolates belonging to the phylo-groups B2, D and F. Overall, 145 virulence and 40 antimicrobial resistance genes among the 315 and 82 genes and variants investigated were detected at least once in one or more of the isolates. The total number of virulence genes per isolate ranged from 18 to 63 genes. The median number of the virulence genes per isolate was 35. Microarray hybridizations demonstrated that all the E. coli isolates tested possessed virulence genes related to a pathotype including ExPEC and diarrheagenic E. coli (DEC). The total number of antimicrobial resistance genes per isolate ranged from 0 to 21 genes with a median number of 6.5 genes. Since the microarray carries a large set of virulence factors, numerous incomplete ExPEC that would normally be missed in a PCR-based assay were found. Thus, various unusual gene combinations were discovered, such as ExPEC pathogenic profiles with assorted EPEC genes coding for the type III secretion system 2 proteins ErpJ and SpaS (gene eprJ and epaS, respectively). Moreover, detection of virulence (mean 39.0 ± 16.9) and antimicrobial resistance (mean 9.0 ± 1.9) genes were higher on hospital-acquired UTIs isolates compared to community-acquired infections ( Table 6). Higher frequency of virulence genes was found among isolates retrieved from females compared to those from males (36.25 ± 17.04
Overall bla CTX-M genes and PMQR genes as well as mutation in QRDR were found in both community-and hospital-acquired infections as well as in male and female genders. The PMQR qnrA (21.4%), and qnrB (18.6%), were highly distributed on MDR-phenotype (Table 7). Isolates from males had higher prevalence of aac(6′)-Ib-cr gene compare to female isolates (16.6% vs 3.5%, p = 0.0194). Quinolone-resistance isolates also presented higher prevalence to bla CTX-M compared to quinolonesensitive phenotype isolates (21% vs 5.2%, p = 0.012).

Discussion
In accordance with global trends, our results reveal higher prevalence of urinary tract infections in female patients than in males [7,[34][35][36]. The emergence of high rates of antibiotic resistance and MDR-phenotype from urinary tract infections related bacteria becomes a public health concern worldwide. In this study, more than 70% of the isolates showed resistance to trimethoprim-sulfamethoxazole (TMP-SXT) that is recommended as a first choice for UTI treatment [36]. Previous Table 7  studies reported similar results [20,36]. In agreement with Bouchillon et al. [34] in this study, UPEC isolates presented more than 40% resistance to ciprofloxacin and levofloxacin antibiotics. Moreover, resistance rates to antimicrobial drugs were higher in isolates from male than in those from female patients [37]. Interestingly, fluoroquinolones are widely used for the treatment of UTIs in male patients [36,38]. It is possible that male infections may be more difficult to eradicate because of the higher rates of antibiotic resistance observed in  [35,36,39]. Susceptibility analysis of isolates to antibiotics prior to treatment choice is recommended. Multidrug-resistant strain prevalence (63.3%) was higher in this work than in other study from Mexico, which reported only 16.4% of MDR strains [21]. While work of Paniagua-Contreras et al. [20], showed 97% of MDR strains. It is well known that susceptibility patterns may vary in different geographical regions and can be change over time [36]. In addition, an unexpected result was found, as the percentage of resistance to the antibiotics tested and MDR-phenotype were similar in hospital and community-acquired infections, indicating an important reservoir of resistance in both settings [34]. This could be due to household use of antibiotics before 2010, which would were acquired without prescription then. Moreover, resistance to cephalosporin and nitrofurantoin was associated to hospital-acquired infections. Co-resistance with fluoroquinolones and TMP-SXT was also observed in both hospital and community.
As detected by Ochoa et al. [21], phylo-group D, an important phylo-group in pathogenic ExPEC was associated to MDR-phenotype. Additionally, FQ-resistant strains were correlated to phylo-group D strains. Strains from hospital-acquired infections exhibited a greater number of virulence genes than those from community. In addition, E. coli isolates from hospital-acquired UTIs are correlated to the presence of kpsMTII gene that has been associated to pyelonephritis, a more severe infection of the upper urinary tract [40]. In agreement with other studies [20,41], our results showed an important frequency of fyuA-encoding yersiniabactin receptor and, fimH encoding type 1 fimbrial adhesion.
Detailed analysis of a strains subset using microarrays, revealed that some of them from both, hospital-and community-acquired infections carried virulence related genes of enteroaggregative and diffusely adherent E. coli (EAEC and DAEC), including capU (cap locus protein, hexosyltransferasa), deoK (deoxyribokinase), shf (putative virulence factor, plasmid pAA2, similar to Shigella flexneri Shf ), virK (similar to Shigella flexneri virulence protein VirK, plasmid pAA2) and astA gene (EAEC heatstable enterotoxin 1). This was also reported in previous works [42][43][44][45], which noticed that EAEC virulence related genes were among the most frequent markers of diarrheagenic E. coli reported in ExPEC strains. The deoK operon is frequently associated with strains isolated from infected urine and blood and is part of a large genetic island carrying genes contributing to the strain intrinsic virulence and/or adaptive properties [43]. These strains, by acquiring the pAA plasmid, could become a potential diarrheal agent [42].
Another finding was the presence of enteropathogenic E. coli (EPEC) related genes on five UPEC isolates. These genes included the eae (coding for intimin, a protein involved in attaching and effacing lesions), as well as virulence related genes such as espG (protein secreted by the type III secretion system and translocated into host epithelial cells), eprJ (E. coli type III secretion system 2 protein ErpJ), epaS (E. coli type III secretion system 2 protein EpaS), and eivG (E. coli type III secretion system 2 protein EivG) [3,46]. T3SS genes are not common in UPEC isolates, however genes encoding components of T3SS have been found [47,48]. These unusual gene combinations illustrate genome plasticity displayed by UPEC strains that may result in new hybrids pathotypes [45,49,50].
In this study, UPEC strains carrying PMQR genes associated to quinolones resistance including qnrA, qnrB, qnrS, qnrC and qnrD as well as aac(6′)-Ib-cr were detected. The most common PMQR-genes identified were qnrA and qnrB, which is in contrast to previous studies were aac(6′)-Ib-cr gene detection was more frequent on ExPEC isolates [19,51]. Moreover, qnrA and qnrB were frequently found as a part of MDR-phenotype. qnrD gene was found in 4.5% of the isolates. Previous papers reported qnrD in bacteria isolated from rooks [51], pigs [15] and humans clinical isolates of Salmonella enterica serovar Kentucky, serovar Bovismorbificans, Proteus mirabilis and Morganella morganii [52,53], as well as in E. coli from dogs [54]. In accordance to others studies [19,55,56], qepA was not identified. Although PMQR genes provide a low level of FQ-resistance, they have been reported to favor the selection of additional chromosome-encoded resistance mechanism [57]. In our study, QRDR mutations in gyrA and parC were prevalent in community-acquired infections (89% vs 11%). Overall, mutations in QRDR were equally present among isolates from males and females patients.

Conclusions
Our study describes significant antimicrobial resistance in E. coli UTI isolates from hospital and community in Mexico. Diverse genotypes and phenotypes including multidrug-resistance, fluoroquinolone resistance and carriage of virulence genes related to several enteropathogenic E. coli were found among UPEC isolates. Thus, continuous surveillance for antimicrobial resistance of UPEC is needed in order to prevent treatment failure and to improve the strategies to mitigate the occurrence of antibiotic resistance organisms and ensure the best treatment to UTI patients. This study highlights the importance of antimicrobial resistance of virulent E. coli from urinary tract infection in Mexico.