Background

Urinary tract infection (UTI) is a common occurrence in paediatric patients. Approximately, 8% of children experience at least one episode of UTI during some period of their childhood [1]. Lack of proper diagnosis and treatment may cause significant morbidity due to progressive destruction of the renal structure and may lead to chronic renal failure [2].

Global emergence of extended-spectrum beta-lactamases (ESBLs) producing uropathogens in community settings is of a great concern [3, 4]. Rapid dissemination of most of the beta-lactamases is facilitated by transferable plasmids that carry resistance genes to several other antibiotics [5]. Infections caused by ESBL-producing organisms are often difficult to treat because of the resistance to beta-lactams and coexisting resistance to other groups of drugs [6]. Escherichia coli among the Enterobacterales is the major uropathogen in the paediatric age group reported from both community and hospital settings [7,8,9]. E. coli producing ESBLs is a global concern due to limited available therapeutic options. There is a rising trend reported in the occurrence of community-acquired UTI from the western world caused by ESBL-producing E. coli among the paediatric population [2, 10]. However, there is scarce knowledge about the prevalence, and molecular epidemiology of resistance genes associated with paediatric UTI in the community setting, especially in the South Asian region including Indian subcontinent [11, 12]. This study reports the draft genome sequence and analysis of 19 uropathogenic E. coli (UPEC) (15 ESBL positive and 4 ESBL negative) isolates from paediatric patients with strictly defined community-acquired UTI from four different geographical locations of India. The study also aimed to determine the distribution of antimicrobial resistance (AMR) genes and the plasmid replicons that help in their replication. This information would be helpful in understanding the epidemiology and development of control and prevention strategies in community settings of developing nations like India.

Methods

This was a prospective multicentric cross-sectional study conducted at the Community Health Centers (CHC) from four different geographical regions of India over a period of three years. It was funded by Indian Council of Medical research in 2019 and conducted by the CAUTION-ED study group. Ethical clearance for the study was obtained by the nodal center (IEC-192/05.04.2019, RP-28/2019). Informed consent for these patients was obtained from their guardians. Consecutive patients with age > 1.5 years to 18 years from the defined community health centers presenting to OPD with increased frequency, urgency, burning, or unexplained fever as chief complaints were included in this study. Patients with vesicoureteral reflex and on history of antibiotic prophylaxis were excluded from the study. Clean catch mid-stream urine sample was collected for processing. Growth of single organism with colony count ≥ 103 (CFU/ml) was considered significant bacteriuria. Identification of isolates to species level was done using Matrix-Assisted Laser Desorption Ionization Time of Flight Mass Spectrometry (MALDI-TOF MS) (Biomeriux, Germany) and Antibiotic Susceptibility Testing (AST) was performed by VITEK-2 (Biomeriux, Durhum, US). The different cards used for antimicrobial susceptibility testing by VITEK -2 were GN (N-235) for Lactose Fermenting (LF) Bacteria, GN (N-281) for Non Lactose Fermenting (NLF) Bacteria and GP (N-628) for Gram positive Bacteria (GPC). MIC50 and MIC90 for different antimicrobial agents were also calculated.

Whole genome sequencing, assembly and annotation

Nineteen UPEC isolates (15 ESBL + ve and 4 ESBL − ve) from the total 54 UPEC isolates were randomly selected for whole genome sequencing (WGS) from all four different geographical regions. WGS was performed to determine the molecular distribution of AMR genes, transposable genetic elements like plasmids, transposons and replicons and their relationship with sequence types.

Genomic DNA from 19 UPEC isolates were extracted using QIAamp DNA Mini Kit (Qiagen, Germany) using manufacturer’s protocol. Library for Oxford Nanopore sequencing were prepared using the PCR-free Native Barcode Expansion kit (EXP-NBD104) and Ligation Sequencing kit (SQK-LSK109). Reads with Q score ≥ 8 were used to find antibiotic resistance markers in ARMA (antimicrobial resistance mapping application) workflow from the EPI2ME analysis platform (https://epi2me.nanoporetech.com). The reads were also used to find plasmids in PlasmidFinder 2.0- https://cge.food.dtu.dk/services/PlasmidFinder/ (accession date: 22/11/21) [13, 14]. Assemblies were made using Minimap2 [15] and NC_000913.3 E. coli str. K-12 substr. MG1655 was used as reference. Samtools and Bcftools [16] were used to find alignment statistics, perform variant calling and obtain consensus sequences. Chromosomal mutations associated with antibiotic resistance were assessed using ResFinder 4.1- https://cge.food.dtu.dk/services/ResFinder/ (accession date: 16/11/21) [17]. The assemblies were annotated using Prokka [18] and pan-genome analysis was performed with Roary [19]. Accessory genome presence/absence was used for hierarchical clustering in Pvclust [20] with 1000 bootstrap iterations. The assemblies were also analysed for the identification of Multi Locus Sequence types (MLST) in MLST 2.0- https://cge.food.dtu.dk/services/MLST/ (accession date: 18/11/21) [21,22,23]. The mobile genomic elements were predicted using Mobile Element Finder-https://cge.food.dtu.dk/services/MobileElementFinder/ (accession date: 19/07/2022).

Phylogenetic analysis of core genomes

Snippy-core [24] was used to perform core-genome alignment for all 19 isolates. A rooted Maximum Likelihood phylogenetic tree was created using MEGA11 [25] with 1000 bootstrap iterations. NC_011740.1 E. fergusonii ATCC 35,469 was used as an outgroup. Annotation of the tree with MLST and plasmids was done in Interactive Tree Of Life (iTOL) [26].

Results

Demographic features

A total of 658 urine specimens were tested following standard operative protocol during the study period. More than fifty percent of patients belonged to age group of 11–18 years (55%), followed by a similar number in both 6–10 years and 0–5 years age groups (22.4%) with equal affection of male and female. Total of 73 uropathogens (11%) were isolated causing significant bacteriuria confirming UTI. E. coli species was observed as the most prominent with a prevalence rate of 86% (n = 54), followed by Klebsiella pneumoniae (n = 7, 11%), Proteus spp. (n = 2, 3%) Acinetobacter spp. (n = 2, 2.7%), Pseudomonas aeruginosa (n = 1, 1.4%), Enterococcus spp. (n = 6, 8.2%), and Staphylococcus aureus (n = 1, 1.3%). 63% (34/54) of E. coli isolates were found ESBL-positive. Among the E. coli isolates, the highest resistance was observed against ticarcillin (74%), followed by ampicillin (72.3%), cefalothin (70.4%), ciprofloxacin (68.6%), cefixime (63%) and ceftriaxone (61.2%). More than one-third isolates were resistant to amoxicillin-clavulanic acid (50%), piperacillin-tazobactam (50%), ceftazidime (48.2%), cotrimoxazole (47%), cefoxitin (37%) and gentamicin (32.5%). Approximately, 20% of isolates were found resistant to amikacin (20.4%). The resistance against ertapenem and nitrofurantoin was observed as 9.3% and 12.2%. All the isolates revealed 100% susceptibility against fosfomycin. MIC50 for E. coli isolates against ceftriaxone was observed 16 times that of its resistance breakpoint. (Table 1).

Table 1 Antimicrobial susceptibility of major uropathogens causing community-acquired urinary tract among the paediatric patients
Full size table

Genomic features

All nineteen isolates (15 ESBL + ve and 4 ESBL − ve) were identified as E. coli by EPI2ME analysis platform from Oxford Nanopore Technologies. Genome assembly was done by mapping the sequencing reads with Q score ≥ 8 to E. coli MG1655 (Accession number: NC_000913.3) as described in the methods section. The chromosome lengths of the assembled genomes for each of the 19 strains ranged from 4.67 to 4.8 Mbp. MLST analysis revealed that 16 isolates belonged to 11 different MLST types while the remaining three did not match any known Sequence Type (ST) (Fig. 1). Plasmid identification from raw reads showed several isolates harboured multiple plasmids that carry antibiotic resistance genes (ARGs) and they have been annotated in Fig. 1. ST131 was the most prevalent phylogroup identified among the isolates. Plasmid IncFIB was found to be present in 63.1% (12/19) followed by IncFIA that was detected among 52.6% (10/19) of the sequenced isolates. Hierarchical cluster analysis of all the 19 E. coli isolates was performed after constructing an accessory genome presence–absence matrix.

Fig. 1
figure 1

Phylogenetic analysis of core genome of Escherichia coli isolates from paediatric patients. A rooted maximum likelihood phylogenetic tree constructed using SNPs across the core-genomes of the 19 E. coli isolates, reference (NC_000913.3) and outgroup (E. fergusonii, NC_011740.1). The phylogram has been annotated with the MLST types of the 19 isolates as well as the plasmids identified in each of them. Number of bootstrap iterations: 1000

Full size image

Phylogenetic and cluster analysis

Core genome alignment of the 19 E. coli isolates and the reference genome (E. coli MG1655, Accession number: NC_000913.3) was performed using Snippy-core (https://github.com/tseemann/snippy) to identify a total of 51,314 single nucleotide polymorphisms (SNPs) between them. A rooted phylogenetic tree was constructed using E. fergusonii ATCC 35,469 (Accession number: NC_011740.1) as an outgroup (Fig. 1). Two distinct clades were observed in the phylogenetic tree (Clade A and Clade B), which further branched into subclades. Based on hierarchical cluster analysis, the 19 E. coli isolates were observed in two distinct clusters, C1 and C2 (Fig. 2). It was observed that all the isolates from clade A constitute the core-genome phylogeny cluster together in cluster C1 and the isolates from clade B form cluster C2. Isolates ID6, ID17 and ID18 of cluster B2 were further branched into a subclade in clade B and were identified to be ST131. ID15 being part of the same subclade was not identified as ST131. Six ESBL-positive isolates from Clade B were observed to carry lncFIAplasmid. Isolates ID1, ID10 and ID8 formed a separate subclade of which ID1 and ID10 belonged to ST14 and ID8 belonged to ST1193. Similarly, isolates ID19 and ID2 from the same subclade branching from Clade B belonged to ST12 carrying plasmids lncFIB and lncFII. In Clade A, a similar sub-branching was observed in cluster C1 for isolates ID23 and ID4. Both of these isolates were found to carry plasmid lncFIB. Isolates ID5 and ID7 from Clade A were identified as ST167 and clustered together in cluster C1 of the dendrogram.

Fig. 2
figure 2

Hierarchical cluster dendrogram based on the accessory genome. The dendrogram branches into two major clusters, labelled as C1 and C2, in red. The probability value for significant clusters is ≥ 0.95

Full size image

Correlation of genetic antibiotic resistance markers to phenotype

Genetic antibiotic resistance profiling of all 19 isolates using EPI2ME ARMA workflow [CARD database [27]] and ResFinder 4.1 [28] showed multiple drug resistance genetic markers against β-lactams, aminoglycosides, sulphonamides, quinolones, macrolide, and co-trimoxazole (Table 2). All fifteen isolates phenotypically identified as ESBL producers harbour at least one of the β-lactamase genes, including blaTEM (− 4, − 1, − 76, − 33, − 166), blaCTX-M (− 15, − 101, − 55, − 33), blaOXA (− 1, − 320, − 224), blaACT (− 5, − 14), and blaNDM (− 5, − 4, − 9). The majority of isolates were observed to carry blaCTX and blaTEM types of ESBL genes. blaCTX-M-15 gene was detected in 60% (9/15) followed by blaCTX-M-55 (53.3%, 8/15) and blaCTX-M-33 (53.3%, 8/15). Among blaTEM genes, blaTEM-4was the most prevalent present in 60% (9/15) of isolates. Among the carbapenemases, 5 isolate were found to harbour blaNDM-5 followed by 3 isolates with blaNDM-4. blaOXA gene was detected in 53.3% (8/15) of the isolates, with blaOXA-1 being the predominant one. blaACT-5 was found in 17 isolates followed by blaACT-14. Chromosomal point mutations in the genes associated with quinolone resistance was observed for gyrA (n = 17, 89%), parC (n = 12, 63%) and parE (n = 10, 53%). Plasmid-mediated resistance markers for quinolone were present in 3 of the isolates, ID2 (qnrS1, qnrS3 and qnrS4), ID14 (qnrS1) and ID19 (qnrB4). Twelve isolates (63.2%) carried resistance genes for trimethoprim including dfrA1, dfrA12, dfrA14 and dfrA17. Genetic markers for aminoglycoside resistance including aminoglycoside phosphotransferases (APHs) (n = 10, 53%), aminoglycoside adenylyltransferases (aadA) (n = 11, 58%) and aminoglycoside acetyltransferases (AAC) (n = 6, 32%) were observed in several isolates. Most of the isolates from Clade A were observed to carry multiple aminoglycoside adenylyltransferases whereas isolates from Clade B were observed to carry multiple aminoglycoside acetyltransferases genes. APH(6)-Id was the most prevalent phosphotransferase followed by APH(3ʺ)-Ib and APH(3ʹ)-IIa. Macrolide resistance gene mphA was found to be present in 47.3% of all the isolates sequenced in the study. However, we did observe few discordant results between the phenotype and genotype for example: isolate 22 and 23 found ESBL-negative despite of presence of beta lactamase genes blaCTX-M-15 and blaTEM. Similar discordant observation were also found incase of cefoxitin susceptibility in isolate no 7, 16, 1, 2 ,6, 9, 10. These discordant observations might be explained by the possibility that mere presence of the resistance genes does not always warrant expression and phenotypic exhibition.

Table 2 Phenotypic resistance profiles with corresponding genotypic resistance markers of the Escherichia coli isolates
Full size table

Mobile genetic elements carrying AMR gene cassettes

A composite transposon, Tn2680 was commonly found in seven out of 19 isolates (ID4, ID5, ID6, ID3, ID18, ID2, ID7). The genetic context of this transposon was found to be varying, carrying different AMR genes cassettes. In four isolates (ID7, ID3, ID4, ID5), this transposon was found carrying blaNDM-5, aadA2, sul1, dfrA1, and qacE genes together with IS26 insertion sequence (Fig. 3). It was also found to carry blaTEM-1B and rmtB genes in two isolates (ID4 and ID5) whereas in two other isolates (ID18 and ID4), it harboured blaOXA-1, catB3 and aac(3)ld genes. Since this transposon was observed with similar genetic context in several isolates, it has the potential to disseminate the multi-drug resistance in the same or even the other bacterial strains. A 14 kb unit transposon (Tn7) was found harbouring dfrA1 and aadA2 genes in ID11. Another 4 kb unit transposon (Tn2) known to have originated from plasmid p838B-R observed carrying blaTEM-1B in ID14 isolate.

Fig. 3
figure 3

Comparison of mobile genetic elements among the Escherichia coli isolates. a The figure shows the presence of the mobile elements carrying similar AMR gene cassettes and draws the comparison of these elements between isolates ID5, ID4, ID7 and ID3. All these 4 isolates were seen carrying Tn2680 with the same AMR genes with high similarity to each other bracketed by a set of IS26. The horizontal lines in the background represent the transposons with forward orientation (blue) and reverse orientation (green). The squares represent the insertion sequences that form the transposon and the arrows represent the genes carried by the transposons. The dark gray shaded area indicates the regions with high identity between the isolates. b The region inside the box has been zoomed in to show the co-integrated mobile genetic elements. The partial fragments of ISEcp1(105 bp, 717 bp) were seen lying in close proximity to the insertion sequences (IS26) of Tn2680 and surrounding the gene cassette containing blaNDM-5, aadA2, sul1, dfrA12 and qacE. Other resistance genes such as aac(6ʹ)-lb-cr, blaOXA-1, catB3 and blaCTX-M15 with a complete ISEcp1 fragment was found downstream of the 717 bp partial fragment

Full size image

Two composite transposon carrying virulence genes (iutA, iucC, sat, papA, papC and iha responsible for siderophore, aerobactin, autotransporter toxin) were found in ID3 and ID4 isolates. In ID3, it was associated with insertion sequence IS629 carrying iutA and iucC genes. In ID4, two transposons with ISEc43 sequences at both ends were found overlapping each other (one IS carrying iucC, iutA and sat gene while the other was carrying papA_F7-2, papC and iha virulence genes). No specific pattern of dissemination of AMR genes in different geographical regions was observed in our study.

Acquisition of bla NDM-5, bla OXA and bla CTX-M15 into E. coli

blaNDM-5 was the only variant of blaNDM observed to be associated with mobile elements in our study. It was seen to be carried commonly with a composite transposon, Tn2680 in close association with other AMR genes such as aadA2, sul1, dfrA1, qacE and mphA in few isolates.

Another composite transposon, originally associated with plasmid pST01 belonging to IS1380 family with insertion sequence ISEc9 was found with 100% coverage and 100% identity in isolate ID4. It was seen carrying a similar gene cassette formed with blaNDM-5, aadA2, sul1, dfrA1, qacE carried by Tn2680 in addition to few other genes such as blaOXA-1, blaNDM-5, blaCTX-M-15, aac(6ʹ)lb-cr and catB3. The end of the transposon was seen with complete insertion sequence of ISEcp1, however, partial insertion sequence fragments of 105 bp and 717 bp of ISEcp1 were found upstream and downstream of gene cassette. Insertion sequence IS26, also was observed to be present in close proximity to the 717 bp ISEcp1 fragment. It seems plausible that two transposons carrying different gene cassettes have been co-integrated together and can be mobilised as one whole mobile genetic element. However, the mobility of the integrated element has to be investigated further. blaOXA-1 was seen being carried on Tn2680 in two isolates (ID18 and ID4). blaCTX-M15 was found associated with Tn2680 in two isolates (ID2, ID6) and ISEc9 insertion sequence in isolate 4. Isolate 3 harbored blaCTX-M-15. However, no association with any mobile element was seen.

Discussion

The present study focuses on paediatric population reporting the molecular epidemiology and various resistance markers of E. coli isolates from various community centers of India. Out of 63 Enterobacterales, 86% were identified as E. coli followed by 11% K. pneumoniae. Girls were most commonly affected than boys with a predominance of age group above 5 years. Studies from Israel offer a unique perspective on ESBL-producing infections in children from the Middle East, where a significant uptrend was found in the yearly incidence of paediatric ESBL-positive UTI infections from 1.2 to 5.2% during the study period 2008 to 2011.The ESBL positivity among E. coli in our study was around 63%. Similar findings of E. coli isolates have been obtained from other studies conducted in community settings in different geographical regions. [4, 29,30,31].

Earlier, studies have shown that CTX-M-type ESBLs have replaced TEM- and SHV-type ESBLs in Europe, Canada, and Asia as the most common ESBL type among various members of the Enterobacterales. Additionally, earlier molecular epidemiology studies have reported that one specific E. coli clone i.e. ST131 has been highly associated with the production of CTX-M resistance in gram-negative infections in paediatric populations [7].

In the current study, ST131 phylogroup was the most prevalent (15.8%) among the identified sequence types of E. coli isolates, which has been shown to be a high-risk clone [32]. More than 66% of the E. coli isolates belonging ST131 were carrying blaCTX-M-15 gene. ST167, another high-risk clone was also identified (10.5%) harboring blaNDM-5 along with multiple blaNDMs [33]. Three of our isolates (15.8%) could not be assigned to any known sequence type based on the seven-gene Achtman scheme described at https://pubmlst.org/ [34]. In the present study, IncF plasmids were known to carry ESBL genes especially blaNDM, genes for aminoglycoside modifying enzyme and plasmid-mediated quinolone resistance (PMQR) [35]. The emergence of a variety of CTX-M-positive E. coli isolates in paediatric populations poses a serious threat, as beta-lactams are often the first line of therapy for UTIs and fluoroquinolones are not routinely used in these populations. The increasing prevalence of CTX-M-15-harboring E. coli ST131 strains in children as demonstrated by this study has important clinical and public health implications due to the risk of treatment failure. Particular to community-acquired ESBLs, UTI rates ranged from 3.8% to as high as 43% in European children with multiple genitourinary comorbidities [31]. European studies on the molecular epidemiology of ESBL-producing isolates in children are primarily small, single-center cohort studies. However, overall trends are consistent with the global shift to CTX-M–type ESBL dominance [31]. Another report from French children highlights 44% of community-acquired ESBL positive clinical isolates harbouring blaCTX-M-15followed byblaCTX-M-14. More than 60% of isolates of the B2 phylogroup strains in both healthcare and community-acquired infections belonged to the ST131 clone [31].

A study in 2017 from South India investigated the prevalence of ESBL levels and associated genes present in E. coli from paediatric UTI [36]. The prevalence of ESBL producer was observed as 37.5% in which blaCTX‑M gene being the most prevalent (87.5%), followed by blaTEM (68.4%) and blaSHV (3.1%). 63% (35/54) ESBL positivity was observed in the current study among the circulating E. coli isolates of which blaCTX-M (66.6%; 10/15) as the most prevalent followed by blaTEM (60%; 9/15).

In the present study, various plasmid-mediated AMR genes along with chromosomal point mutations and virulence genes were found in the E. coli isolates. The presence of high-risk clones in these isolates from paediatric patients suffering with UTI in the community setting of India is worrisome.

Conclusions

The incidence of antibiotic resistance in the community-onset UTIs caused by ESBL-producing carrying high-risk resistant clones of E. coli among children seems to be increasing in India; thereby management of UTI in paediatric patients might be challenging for the clinicians in near future. This study demonstrates a need for heightened awareness regarding the increasing frequency of these resistant isolates in the paediatric population in the community and their potential impact on disease management. Rapid detection with One health approach is the need of the hour for proper management and prevention of the spread of drug-resistant pathogens in the community. Regular surveillance at different community settings will be helpful to understand the genotypes, their transmission dynamics and further implementation of preventive measures.