A total of 500 UPEC clinical strains were collected from 2010 to 2012 at the Hospital Infantil de México Federico Gómez (HIMFG) from pediatric patients with community- and nosocomial-acquired UTIs. The UPEC strains were selected considering an average bacterial count of ≥100,000 CFU/mL as determined by Kass-Sandford (Kass, 1957). UPEC strains associated with persistent UTIs were isolated from cultures from the same patient taken in different months. UPEC strains were cultured in trypticase soy broth (TSA), MacConkey medium (Mac), Luria-Bertani medium (LB), and Mueller-Hinton Broth (MHB) (Difco-Becton Dickinson, NJ, USA) at 37°C for 24 h according to assay type. Additionally, UPEC strains were stored in cryovials at −70°C with Brain Heart Infusion broth (BHI; Difco-Becton, Dickinson, NJ, USA) containing 25% glycerol (Sigma-Aldrich, St. Louis, MO, USA). Only one isolate from each infectious episode was included in the study, and isolates were considered to come from the same episode if they were collected <30 days apart. Thus, all MDR- and XDR-UPEC strains were isolated from 103 infectious episodes in 71 patients, and 8 patients experienced more than one episode.

The minimum inhibitory concentration (MIC) as determined by the MHB microdilution method was used to evaluate the antimicrobial susceptibility of 500 UPEC clinical strains according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI, 2016). MDR strains were defined as having acquired no susceptibility to at least one antibiotic in three or more classes. XDR strains were defined as having non-susceptibility to at least one agent in all but two or fewer antibiotic classes (Magiorakos et al., 2012). The MIC for each antibiotic was compared to the standard values of the CLSI. The antibiotic panel that was used included ampicillin (AM; Sigma-Aldrich, St. Louis, MO, USA), amoxicillin-clavulanate (AMC; Great West Road, Brentford Middlesex, UK), ticarcillin-clavulanate (TIM; Gold Biotechnology, Inc., Ashby Road, St. Louis, MO), piperacillin-tazobactam (TZP; Siemens Medical Solutions USA, Inc., Valley Stream Parkway, Malvern, PA, USA), cephalothin (CF; Eli Lilly and Company, S Harding St, Indianapolis, IN, USA), cefaclor (CEC; Phadia Laboratory Systems, Thermo Scientific, Wyman Street, Waltham, MA, USA), ceftazidime (CAZ; Roselle Rd, Schaumburg, IL, USA), aztreonam (ATM; Bristol-Myers Squibb Corporate, Park Avenue, NY, USA), norfloxacin (NOR), ofloxacin (OFX; MP Biomedicals, Solon, OH, USA), meropenem (MEM), imipenem (IPM; AstraZeneca Pharmaceuticals LP, Wilmington, DE, USA), gentamycin (GM; Schering-Plough Pharmaceuticals, Kenilworth, NJ, USA), ceftriaxone (CRO), trimethoprim-sulfamethoxazole (SXT; Roche, Basel, Switzerland), tetracycline (TE; Heritage Pharmaceuticals Inc., Fieldcrest Avenue, Edison, NJ, USA), and nitrofurantoin (F/M; McKesson Pharmaceutical, One Post Street, San Francisco, CA, USA). E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as controls.

The extended-spectrum beta-lactamases (ESBLs) were phenotypically detected as previously recommended by CLSI using the double-disc synergy test based on the synergistic effect between clavulanic acid (inhibitor of ESBLs) and β-lactam antibiotics (cefotaxime, CRO, CAZ, cefepime, cefpirome, and ATM). Additionally, ESBLs were detected using an individual disk that was tested with/without clavulanic acid (10 μg/mL) and by the Hodge test using Klebsiella pneumoniae ATCC 700603 (ESBL+) and E. coli ATCC 25922 (ESBL-) as control strains (CLSI, 2016).

DNA was extracted from the MDR- and XDR-UPEC clinical strains cultured in LB using the Wizard® Genomic DNA Purification Kit (Promega Corporation, Woods Hollow Road, Madison, WI, USA) according to the manufacturer's instructions. Multiplex polymerase chain reaction (PCR) assays were used to determine the presence of chuA, yjaA, and one genomic fragment (TspE4.C2) (Clermont et al., 2000) in MDR- and XDR-UPEC clinical strains (Table S1). The PCR products were resolved by electrophoresis on a 1.5% agarose gel (Promega Corporation, Woods Hollow Road, Madison, WI, USA) that was stained with ethidium bromide solution (Sigma-Aldrich, St. Louis, MO, USA) and visualized using a Gel Imaging System (ChemiDoc MP System, Bio-Rad; Mexico). According to the amplicon data, the strains were assigned to phylogenetic groups as follows: group B2 (chuA⁺/yjaA⁺), group D (chuA⁺/yjaA⁻), group B1 (chuA⁻/TspE4.C2⁺), and group A (chuA⁻/TspE4.C2⁻) (Clermont et al., 2000).

Eight virulence-associated genes [three variants (I, II, and III) of papG (PapG), fimH (FimH), bcsA (cellulose), csgA (CsgA), ecpA (EcpA), iutD (aerobactin), hlyA (α-hemolysin), and tosA (type 1 secretion A)] from MDR- and XDR-UPEC clinical strains were identified by multiplex PCR using specific primers (Table S1). CFT073 E. coli was used as a positive control.

Integrons in the MDR- and XDR-UPEC strains were detected by multiplex PCR, which amplified the conserved region of the integrase-encoded genes intI1, intI2, and intI3 using specific primers (Table S1). MDR-UPEC strains 502U1-0412 and 674U-0612 were used as positive controls for class 1 and 2 integrons, respectively.

The variable regions of class 1 and 2 integrons from the MDR-UPEC clinical strains were amplified by PCR using a high-fidelity Pfu polymerase of Thermo-Fisher Scientific (CA, USA) (Table S1). The amplicons were cleaned and concentrated using the Zymo DNA Clean and Concentrator of Zymo Research (CA, USA) and subjected to next-generation sequencing on a NexSeq500 System (Illumina, CA, USA), which was performed at “Unidad de Secuenciación del Instituto Nacional de Medicina Genómica” (CDMX, Mexico). The sequences were analyzed and assembled using ClustalO, ORF Finder (Open Reading Frame Finder), and BLAST (Basic Local Alignment Search Tool) from the NCBI (National Center of Biotechnology Information) (Sievers et al., 2011; Soleimani et al., 2014).

A phylogenetic analysis of MDR- and XDR-UPEC clinical strains was performed using pulsed-field gel electrophoresis (PFGE) following the specific modifications of the protocols established by the “Laboratorio de Investigación en Bacteriología Intestinal, HIMFG” (Ochoa et al., 2015). Briefly, the samples were digested with 2 U of Xbal (Invitrogen, Life Technologies, Wyman Street, Waltham, MA, USA) at 37°C for 4 h and separated by electrophoresis in a 1.2% ultrapure DNA agarose gel (Bio-Rad, Life Science Research, Hercules, CA, USA) that had been previously dissolved in 0.5X TBE. The samples were run on a CHEF Mapper system (Bio-Rad Life Science Research, Hercules, CA, USA) for 23 h at 200 V (6 V/cm) under two different linear ramped pulse times: 1–10 s and 10–40 s. The gels containing the macrorestriction products were stained for 40 min with 0.5 mg/mL ethidium bromide (Sigma-Aldrich St. Louis, MO, USA) and visualized under UV light using an Imaging System (ChemiDoc MP System®, Bio-Rad, Mexico City, Mexico). A lambda ladder PFGE marker (New England Biolabs, Hertfordshire, England, UK) was used as a molecular weight marker. PFGE patterns were analyzed using NTSYS-pc software (version 2.0, Applied Biostatistics, Inc., NY, USA) with the Sørensen-Dice similarity coefficient and a cluster analysis of the Unweighted Pair Group Method using Arithmetic averages (UPGMA) (Dice, 1945). The clonality of the UPEC strains was evaluated considering the Criteria of Tenover (Tenover et al., 1995).

Fisher's exact test and Fisher's PLSD ANOVA were used to identify significant differences in the data at p < 0.05.

According to the resistance profile to 10 classes of antibiotics, 16.40% (82/500) of the UPEC strains were MDR, and 4.20% (21/500) were XDR. Additionally, MDR- and XDR-UPEC strains were isolated from hospitalized patients with complicated UTIs from different wards of HIMFG, such as Nephrology, Urology, PICU (Pediatric Intensive Care Unit), NICU (Neonatal Intensive Care Unit), Oncology, and Pediatrics. From these, 82 MDR-UPEC strains were distributed as follows: 31.70% (26/82) were resistant to three classes of antibiotics, 28.04% (23/82) to four classes, 24.39% (20/82) to five classes, 9.75% (8/82) to six classes, and 6.09% (5/82) to seven classes. Furthermore, in eight patients, 40 MDR-UPEC strains were collected with five isolates from each patient. Moreover, 21 of the XDR-UPEC strains were distributed as follows: 52.38% (11/21) of were resistant to eight classes of antibiotics and 47.61% (10/21) to nine classes (Tables S2, S3). Interestingly, 76.19% (16/21) of XDR-UPEC strains had an ESBL phenotype (p ≤ 0.028) compared to 18.29% (15/82) of MDR-UPEC strains. Class 1 and 2 integrons were identified in 43.90% (36/82) and 2.43% (2/82) of MDR-UPEC strains, respectively. Additionally, 47.61% (10/21) of XDR-UPEC strains presented class 1 integrons, and 9.52% (2/21) presented class 2 integrons. These UPEC strains did not present class 3 integrons.

MDR-UPEC strains were distributed into three phylogenetic groups: 54.87% (45/82) to D, 39.02% (32/82) to B2, and 6.09% (5/82) to A; the XDR-UPEC strains were distributed as follows: 47.61% (10/21) to B2, 42.85% (9/21) to D, and 9.52% (2/21) to A. Moreover, phylogenetic group B1 was not identified in MDR- or XDR-UPEC strains. On the other hand, 28.19% (9/32), 11.12% (5/32) and 20% (1/32) of MDR-UPEC strains in groups B2, D, and A, respectively, were associated with the ESBL phenotype. In contrast, 80% (8/10), 66.67% (6/9), and 100% (2/2) of XDR-UPEC strains in groups B2, D, and A, respectively, produced ESBLs.

The tosA gene encoding the type 1 secretion A toxin was distributed at different percentages to phylogenetic groups A, B2, and D of MDR- and XDR-UPEC strains (Table 1). The tosA gene from MDR-UPEC strains associated with phylogenetic group B2 was significantly different (p ≤ 0.0001) from that from XDR-UPEC strains (Table 1). Fimbrial (ecpA, fimH, csgA, and papGII), iron uptake (chuA, iutD) and α-Hemolysin (hlyA) genes from MDR- and XDR-UPEC strains were highly distributed into phylogenetic groups B2 and D (Table 1). Additionally, the papGIII variant gene of the MDR-UPEC strains was only distributed into phylogenetic group B2 and D, while the papGI variant gene was not identified in any UPEC strain (Table 1). The bcsA gene was distributed into phylogenetic groups as follows: 51.10% (23/45) to D, 43.75% (14/32) to B2 and 40% (2/5) to A of MDR-UPEC strains; however, XDR-UPEC strains carrying the bcsA gene were only distributed into phylogenetic group B2 (Table 1).

MDR-UPEC strains with class 1 integrons were related to phylogenetic groups A, B2, and D [40% (2/5), 43.75% (14/32), and 44.43% (20/45), respectively]; under these same conditions, class 2 integrons were related to phylogenetic groups B2 and D [3.12% (1/32) and 2.22% (1/45), respectively] (Table 2). In the XDR-UPEC strains, class 1 integrons were related to phylogenetic groups B2, D, and A [30% (3/10), 55.56% (5/9), and 100% (2/2), respectively]; in addition, the class 2 integrons were associated with phylogenetic groups B2 and A in 10% (1/10) and 50% (1/2), respectively (Table 2).

A total of 56.52% (26/46) of MDR- and XDR-UPEC strains amplified the variable region of class 1 integrons with a product of ~1800 bp; in addition, 75% (3/4) of the MDR- and XDR-UPEC strains amplified the variable region of class 2 integrons with a product of ~2000 bp (Table 3). The sequencing analysis of ~1800-bp products showed the gene cassettes for aadA1 (aminoglycoside-3′-adenylyltransferase), aadB (aminoglucoside-2′-adenylyl-transferase), aacC (gentamicin-acetyl-transferase), ant1 (streptomycin-3′-adenylyltransferase), dfrA1 (dihydrofolate reductase type 1), and dfrA17 (streptomycin-3′-O-adenylyltransferase); their distribution is shown in Table 3. In contrast, the aadA1, dfrA1, and aadA4 (streptomycin-3′-adenylyl transferase) genes were identified in products of ~2000 bp (Table 3). Finally aadA1, aadB, aacC, ant1, dfrA1, and dfrA17, which encode modifying enzymes within the class 1 and 2 integrons from MDR- and XDR-UPEC strains, were associated with resistance to GM or SXT.

Patterns consisting of 11–21 DNA fragments ranging in size from 48.5 to 436.5 Kb obtained from the PFGE assay of the MDR- (51 pulsotypes) and XDR- (21 pulsotypes) UPEC strains were used to construct two dendrograms with cophenetic correlation coefficients of 0.8124 and 0.8037 (Figures 1, 2). Furthermore, 12.19% (10/82) of the MDR-UPEC strains were distributed into cluster I, 15.85% (13/82) into cluster II, 12.19% (10/82) into cluster III, 21.95% (18/82) into cluster IV, 20.73% (17/82) into cluster V, 10.97% (9/82) into cluster VI, and 5% (5/82) into cluster VII (Figure 1). Of the seven clusters from the dendrogram, four clusters (II, III, IV, and V) grouped MDR-UPEC strains within isolates that were clonally related, collected at different times, and considered persistent; however, in clusters I, VI, and VII, a clonal relationship was not identified. Finally, XDR-UPEC strains were distributed into four clusters: 23.80% (5/21) into cluster I, 23.80% (5/21) into cluster II, 33.33% (7/21) into cluster III, and 19.04% (4/21) into cluster IV (Figure 2).

The four clusters (II, III, IV, and V) grouping MDR-UPEC strains shared the presence of fimbrial genes (ecpA, fimH, csgA, and papGII) and an iron uptake gene (chuA). In clusters I, VI, and VII, a relationship between phylogenetic groups, integrons and virulence genes was not identified. In cluster II, two clonal groups associated with phylogenetic group D were identified: 557U (1–5 clones) characterized by the absence of the bcsA gene and 177U (1, 2, 4, and 5 clones) by the presence of the bcsA gene. In cluster III, two clonal groups associated with phylogenetic group B2 were identified: 532U (1 and 3 clones) characterized by the presence of the ESBL phenotype and 533U (2 and 4 clones) by the lack of the ESBL phenotype. In cluster IV, two clonal groups associated with phylogenetic group D were identified: 424U (2–5 clones) characterized by the presence of the iutD and hlyA genes and 118U (1–5 clones) by the presence of class 1 integrons and the bcsA gene. In addition, 424U4 lacked the csgA gene compared with 424U (2, 3, and 5 clones). In cluster V, four clonal groups [561U (1–5 clones), 440U (2 and 3 clones), 441U (4 and 5 clones), and 511U (1–5 clones)] associated with phylogenetic group B2 and with the tosA and hlyA genes were identified. Class 1 integrons and the bcsA gene were found in three clonal groups (561U, 440U, and 441U) and the iutD gene in three clonal groups (440U, 441U, and 511U). Interestingly, the papGIII gene was only found in clonal group 511U associated with the tosA and hlyA genes (Figure 1).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.