Multiple and High-Risk Clones of Extended-Spectrum Cephalosporin-Resistant and blaNDM-5-Harbouring Uropathogenic Escherichia coli from Cats and Dogs in Thailand

Resistance to extended-spectrum cephalosporins (ESC) and carbapenems in Escherichia coli (E. coli), increasingly identified in small animals, indicates a crisis of an antimicrobial resistance situation in veterinary medicine and public health. This study aimed to characterise the genetic features of ESC-resistant E. coli isolated from cats and dogs with urinary tract infections in Thailand. Of 72 ESC-resistant E. coli isolated from diagnostic samples (2016–2018), blaCTX-M including group 1 (CTX-M-55, -15 and -173) and group 9 (CTX-M-14, -27, -65 and -90) variants were detected in 47 isolates (65.28%) using PCR and DNA sequencing. Additional antimicrobial resistance genes, including plasmid-mediated AmpC (CIT and DHA), blaNDM-5, mcr-3, mph(A) and aac(6′)-Ib-cr, were detected in these isolates. Using a broth microdilution assay, all the strains exhibited multidrug-resistant phenotypes. The phylogroups were F (36.11%), A (20.83%), B1 (19.44%), B2 (19.44%) and D (4.17%), with several virulence genes, plasmid replicons and an integrase gene. The DNA fingerprinting using a repetitive extragenic palindromic sequence-PCR presented clonal relationships within phylogroups. Multiple human-associated, high-risk ExPEC clones associated with multidrug resistance, including sequence type (ST) 38, ST131, ST224, ST167, ST354, ST410, ST617 and ST648, were identified, suggesting clonal dissemination. Dogs and cats are a potential reservoir of ESC-resistant E. coli and significant antimicrobial resistance genes.


Introduction
Urinary tract infection (UTI) is commonly caused by opportunistic bacteria and requires systemic antimicrobial treatment in small animal practice. Escherichia coli (E. coli) is a major pathogen causing UTIs in humans and animals, with a prevalence of 35-69% in dogs and cats [1]. According to the treatment guidelines developed by the International Society for Companion Animal Infectious Diseases (ISCAID), extended-spectrum cephalosporins (ESC), such as cefpodoxime, cefovecin and ceftiofur, are considered second-line antimicrobials for bacterial cystitis based on antimicrobial susceptibility results [2]. Due to the emerging spread of extended-spectrum β-lactamases (ESBLs), such as cefotaximase (CTX-M) and plasmid-mediated AmpC cephalosporinases (pAmpC) associated with multidrug resistance (MDR) [3][4][5], the available antimicrobial options have become ineffective for treatment. Extensively drug-resistant and pan drug-resistant Gram-negative bacteria, which are resistant to last-resort carbapenems and colistin, are a serious burden in human medicine [6] and have emerged in small animals and veterinary settings [7,8]. Mobilisable carbapenemase-encoding genes such as bla NDM-5 (New Dehli metallo-β-lactamase) and bla OXA-181 (oxacillinase) and mcr (mechanism of colistin resistance) genes in E. coli have been characterised and increasingly reported by animal sources worldwide [7,9,10]. Pets and people associated with pets having identical ESBL-producing E. coli strains support evidence of interspecies transmission of resistant bacteria between humans and animals [11].
Regarding zooanthroponotic transmission, high-risk E. coli clones in terms of sequence type (ST) by multilocus sequence typing (MLST) and virulent phylogroups are regarded as a public health issue. [5,12,13]. Uropathogenic E. coli (UPEC) is an extraintestinal pathogenic E. coli (ExPEC) pathotype mainly classified into virulent-phylogroup B2, containing numerous virulence factors for colonisation, invasion and survival outside intestine. High-risk human-associated ESC-resistant ExPEC ST38, ST131, ST410 containing bla CTX-M group 1 and 9 have been identified in human clinical samples from Thailand [14]. Of these, pandemic ESBL-producing E. coli ST131 and carbapenemase-producing E. coli ST410 distributed in companion animals have been increasingly reported worldwide [9,15], but information about the genetic background and relatedness of E. coli to a present linkage from the companion animal sector as a part of the One Health concept is unavailable in Thailand. Based on our routine diagnostic antimicrobial susceptibility testing results, ESC and carbapenem resistance in E. coli associated with UTIs in dogs and cats have frequently been observed, and questions about the molecular epidemiological relationships among these isolates and with the clones distributing in animals and humans from different regions were raised. Therefore, the aims of this study were to characterise the genetic features and antimicrobial genotypes/phenotypes of canine and feline ESC-resistant UPEC and to present multiple human-associated ExPEC clones from dogs and cats in Thailand.

Discussion
This study showed the heterogenicity of ESC-resistant UPEC isolated from canine and feline clinical specimens in terms of MDR, virulence gene carriage and genetic background. Various ESBL-and pAmpC-encoding genes, including bla CTX-M variants and CIT and DHA AmpC groups, were prevalent in these isolates, and bla NDM-5 was first detected in animal sources in Thailand. All the detected bla CTX-M variants belonging to group 1 (bla CTX-M-15 , bla CTX-M-55 and bla CTX-M-173 ) or group 9 (bla CTX-M-14 , bla CTX-M-27 , bla CTX-M-65 and bla CTX-M-90 ) are the predominant ESBL genes in E. coli from human clinical specimens in Thailand [14]. In this study, bla CTX-M-14 was the most prevalent ESBL gene, followed by bla CTX-M-55 and bla CTX-M-15 . Here, bla CTX-M-55 was first characterised in Thailand, and four major variants, including bla CTX-M-14, -15, -27 and -55 , which also widely disseminate in ESC-resistant E. coli from human sources [18]. In companion animals, bla CTX-M-14 is mostly carried by feline UPEC [19], and bla CTX-M-15 and bla CTX-M-55 are the most frequently occurring variants found across the globe, especially bla CTX-M-55 in Asian countries [5,11,[20][21][22][23][24][25][26]. Similar studies from Portugal identified faecal ESC-resistant E. coli containing several bla CTX-M variants in healthy and sicked dogs and cats that might be a potential source for zoonotic transmission [27,28]. We provide further evidence that the widespread bla CTX-M genes were shared among E. coli from human and animal origins. Besides bla CTX-M , the CIT and DHA groups of pAmpC and bla TEM were found in half of the studied isolates. The bla CMY-2 , which is a major CIT member, is more prevalent than bla CTX-M , which disseminates among ESC-resistant E. coli from canine stools in South Korea [11], but pAmpC and other bla genes were not specifically identified by sequencing in this study. Due to the presence of various classes of β-lactamase-encoding genes, the effectiveness of different β-lactam generations, which is the widest antibiotic group used in veterinary and human medicine, is limited by the accumulation and development of resistance.
The canine and feline UPEC populations carried plasmid-borne antimicrobial resistance genes, such as aac(6 )-Ib-cr, mph(A), bla NDM-5 and mcr-3, promoting additional resistance to higher generations and different groups of antimicrobials. In addition to stepwise mutations of gyrase and topoisomerase genes, aac(6 )-Ib-cr is usually co-harboured in fluoroquinolone and ESC-resistant E. coli [15,29], which is of concern due to the horizontal spread among Gram-negative pathogens. To our knowledge, this is the first report of carbapenemase bla NDM-5 and plasmid-mediated colistin resistance mcr-3 in E. coli from companion animals in Thailand. The bla NDM-5 is a frequent variant carried by high-risk human-associated ExPEC clones [30,31], and isolation from animal sources has been increasingly reported worldwide [32]. These alarm a crisis of resistance to last-resort antimicrobials in veterinary medicine. With the emergence of MDR, carbapenems are introduced for extra label treatment of serious Gram-negative infections in small animal practices, whereas colistin is not available because of the lack of recommendation from pharmacokinetic data in dogs and cats. The increased spread of resistance is possible if carbapenem is continuously being used without strict regulations in small animal practices. As colistin is not used in companion animals, livestock or humans are probable sources of mcr-3. The mcr-3 gene from Enterobacteriaceae has been also reported from Thai people and swine, suggesting a common prevalent type shared among different hosts in Thailand [6,33]. Companion animals living in complex proximities to humans, animals and food are prone to acquiring a variety of resistance genes and microorganisms [22].
Various clones and genetic linkage of the UPEC isolates have been illustrated by STs, phylogroups and DNA fingerprints. They belonged to phylogroups A, B1, B2, D and F, of which the latter is the most common phylogroup, followed by phylogroup A in our study. Recent studies found the intestinal carriage of the phylogroups A, B1, B2 and D to be associated with ESBL and pAmpC production in healthy and diseased dogs and cats [27,28]. Previous evidence supports the assumption that uropathotypes are mainly members of the virulent phylogroup B2, and the commensal isolates are phylogroup A [12]. An Australian study also found that phylogroups A and B1 were dominant among canine and feline fluoroquinolone-resistant E. coli, causing extraintestinal infections [15]. Clonal expansion and genetic relatedness could be deduced using DNA fingerprint cluster analysis, demonstrating an association of the major clusters II, III, IV and VIII with the phylogroups B2, A, B1 and F, respectively. A variety of virulence genes associated with uropathogenesis were found in the UPEC isolates and in all the phylogroups identified. However, the isolates in phylogroup B2, theoretically, carried the highest numbers of virulence genes. Based on genetic and virulence characteristics and clinical sources, our results support the uropathogenic potential of non-B2 E. coli phylogroups in canine and feline urinary tract infections.
We identified multiple high-risk, human-associated ExPEC STs with a significant public health implication, such as ST410-F (ST-phylogroup), ST131-B2, ST354-F, ST648-F and ST38-D, [34]. Of these, ST38, ST131 and ST410 are the major STs of ESBL-producing E. coli causing human infections in Thailand [14]. The frequently identified high-risk STs from canine and feline clinical specimens are ST38, ST131, ST224, ST354, ST648 and ST2179 [15,35]. Interspecies transmission and clonal dissemination are presumable by shared STs and the high similarity of genetic characteristics of strains from different sources [36]. The global spread of bla CTX-M-15 -carrying ST131, which mainly belongs to the virulent phylogroup B2, is considerably worrying due to its MDR and virulent traits associated with UTIs and fatal outcomes in humans [37]. The expansion of ST131 in small animals has been increasingly proposed, emphasising its clinical significance [19,20]. Shared high-risk clones between pets and humans in the same environment are a public health and occupational risk for pet-associated people [38]. Indeed, the identification and virulence determinants of human-associated ExPEC clones from diseased dogs and cats in this study confirmed the pathogenic potentials in animals, and infected pets or carriers may be a source of reverse transmission to humans. Environmental contamination in veterinary settings and colonisation in personnel should be further monitored to improve preventive measures against infection in animals.
Companion animals could be a potential niche for multiple high-risk ExPEC clonal expansion in the community. Clonally related bla CTX-M -carrying ST410 is spread among human and veterinary clinical settings, livestock and wild avian species, but the transmission scenarios cannot be clearly clarified [36,39]. Dogs and cats play a role as a reservoir of ST648 harbouring bla CTX-M group 1, which is sporadically found in humans [35,[40][41][42]]. An accumulation of multiple β-lactamase-encoding genes is more common in the high-risk Ex-PEC lineages. One study has reported an abundant co-harbouring of bla CTX-M-15 , bla TEM-1 , bla OXA-1 , bla CMY-2 and aac(6 )-Ib-cr in ESC-resistant ST410 strains from infected animals [20]. Besides bla CTX-M-15 , the most prevalent bla CTX-M-14 and bla CTX-M-55 were carried by ST131 isolates in our study. Moreover, ST167, ST354, ST410 and ST641 carrying bla NDM-5 found in this study confirm the capability to evolve by additional gene acquisition in these lineages. Nearly identical resistance genotypes with an additional bla NDM-5 co-carriage in the proliferative high-risk clones imply a more critical situation of antimicrobial resistance to limit options and reduce the effectiveness of last-line antimicrobials in small animal practices in Thailand.
The various plasmid replicons identified in the MDR UPEC isolates support the assumption that the plasmids could play a role in resistance and virulence contribution. The number of replicons identified in each isolate may not relate to the number of the carried plasmids. Closely related multi-replicon plasmids containing resistance and virulence genes in E. coli from human and dogs have been characterised by whole genome sequencing (WGS) [7,30]. However, transmissibility was not assessed in this study. Virulence and antimicrobial resistance genes may be co-harboured for the co-selection of the resistant and virulent traits, and plasmid recombination can enhance higher pathogenicity and create hybrid E. coli pathotypes [43]. Resistance determinants may be alternatively integrated in the plastic bacterial chromosome, which is supported by variable chromosomal regions or plasmids for bla CTX-M-15 localisation in ExPEC ST410 [39]. The bla CTX-M-14 , bla CTX-M-15 , bla CTX-M-27 and bla CTX-M-55 , found in multiple STs, support the active mobility of the elements for spreading among ExPEC from human and animal origins. Further analyses of gene localisation and the characterisation of mobile genetic elements can increase our understanding of the modes of transmission and co-selection of the resistance elements and strains and are useful for monitoring and creating interventions to limit the spread.
The bla NDM-5 in ST167, ST354, ST410 and ST641 supports the evolution of resistance, which is more successful in extensively drug-resistant ExPEC populations. In a previous study, WGS analysis was used to identify bla NDM-5 localisation on the IncFII plasmid in ST167 from dogs and cats in the USA [29]. A narrow host range of IncX3 plasmid is a vehicle for bla NDM-5 spread in ST167 in China [31], but this IncX3 was not detected in our carbapenem-resistant isolates. All the bla NDM-5 -carrying isolates were intI1-positive. The integron collecting bla NDM-5 and other resistance elements associated with plasmids have been characterised [7]. Although mcr and bla NDM co-carriage was not detected in this study, co-harbouring bla NDM-5 and mcr-1 in ST167 on different plasmids has been reported in a hospitalised patient with pneumonia in Japan [44]. Potential development is likely by gene acquisition when the strains will be introduced in the environment with antimicrobial exposure. Carbapenems are extra-labelly prescribed in animal cases with life-threatening MDR bacterial infections in Thailand; however, the history of carbapenem administration in dogs was not assessed. The NDM-5-producing ExPEC have emerged in hospitalised companion animals without carbapenem treatment within a 1-month period in the USA [29]. Nevertheless, antimicrobial use is a key factor contributing to resistance selection and spread. Due to the emergence of carbapenem and colistin resistance, the selection of these drugs, which are considered last-line antimicrobials in human medicine, should be avoided if the causative bacteria are still susceptible to available drugs for veterinary purposes.
The urinary tract is a body site that is mostly affected by ExPEC infection. Almost half of the E. coli (151/344) in our collection were from a UTI. It is likely that UPEC were indigenous E. coli in the intestinal tract, which can ascendingly be translocated through the defective urinary pathway. Both ESC and amoxicillin/clavulanic acid exposure promotes the intestinal carriage of ESBL-producing Enterobacteriaceae, which can be a source of extraintestinal infections [32]. Huang et al. found that 22.8% (65/283) of E. coli from companion animal specimens produce ESBLs, and more than 70% of those derive from urine, pus in the uterus and wounds [21]. In the present study, a high proportion (43.06%) of the ESC-resistant UPEC expressed the ESBL phenotype by CDT, but the expression in the remaining bla CTX-M -positive isolates might be masked by other co-existing genes, such as AmpC cephalosporinase, carbapenemase or others. Thus, we recommend including various surrogate cephalosporin generations, amoxicillin/clavulanic acid and carbapenems, together with ESBL phenotype detection in β-lactam panels, for antimicrobial susceptibility testing for surveillance purposes. The MDR of the ESC-resistant UPEC strains negatively impacted antimicrobial uses in small animal practices, especially the empirical selection of first-and second-line antimicrobials. To encourage diagnostic and antimicrobial stew-ardships, AST panels customising the drugs used for UTIs are strongly recommended for effective treatment. With a low nitrofurantoin resistance rate, this drug could still be an effective choice. However, AST is strongly recommended for the most reasonably appropriate treatment. This study focused on only UPEC, which is a subtype of ExPEC, from dogs and cats. In addition, ExPEC is a major pathogen associated with other extraintestinal diseases in small animals, such as pneumonia, endometritis and prostatitis [29,45,46], but the samples were difficult to obtain for routine clinical diagnosis because of the invasive sample collection procedures. Further studies on the virulence and genetic background of ExPEC in companion animals of other clinical relevance may help to understand the pathogenesis of ExPEC in small animals and the epidemiological linkage with strains from humans. Shedding in the environment through faecal excretion is also an important indirect mode of transmission, especially in animal healthcare settings where antimicrobials are intensively consumed. Control and prevention of spread should be undertaken in veterinary hospitals by prescription restriction, strict hygiene and decontamination. The development of resistance and spread can be impeded, and the effectiveness of antimicrobials can be prolonged by the implementation of antimicrobial stewardship programs in veterinary practices.

Bacterial Isolates, Species Identification and DNA Isolation
The E. coli isolates were obtained from clinical specimens submitted to the Veterinary Diagnostic Laboratory of the Faculty of Veterinary Science, Chulalongkorn University, between 2016 and 2018. The criteria used for selection were E. coli presenting resistance to 3GCs that were isolated from canine and feline samples associated with UTIs. Species and 3GC resistance were initially identified by GN-ID and AST-GN65 of the Vitek ® automated system in routine diagnostic activity (bioMerieux, Marcy l'Etoile, France). Bacteria were preserved in tryptic soy broth with 20% glycerol stock and kept at −80°C. Species were confirmed using matrix-assisted laser desorption and ionisation time-of-flight mass spectrometry (MALDI-TOF MS) (Microflex ® Biotyper) (Bruker, Dalnotik, Bremen, Germany), and genomic DNA was extracted by using the Nucleospin Tissue DNA extraction kit (Macherey-Nagel, Düren, Germany) from pure cultures after recovery on 5% sheep blood agar. The DNA was stored at −20°C to be used as DNA template for PCR-based assays.

Detection of ExPEC Virulence Genes
The ExPEC virulence genes associated with uropathogenesis that encode adhesins (iha, papC, crl, sfa/foc and afa/dra), cytotoxins (hlyA, cnf1/2, sat and vat) and the iron acquisition system (iucD and iroN) were detected using M-PCRs [62]. The presence of fimH was identified using a single specific PCR [63]. The PCR mixture was composed of reagents as described above.

Plasmid Replicon Typing and Integrase Detection
Plasmid replicons were investigated using five M-PCRs, including multiplex 1 for HI1, HI2 and I1-Iγ, multiplex 2 for X, L/M and N, multiplex 3 for FIA, FIB and W, multiplex 4 for Y, P and FIC and multiplex 5 for A/C, T and FIIAs, as well as three simplex PCRs for F, K and B/O [64]; intI1 was detected using a specific PCR [65].

Molecular Typing
The E. coli phylogroups including A, B1, B2, C, D, E, F and Escherichia cryptic clade I were characterised by the Clermont phylo-typing scheme [66]. Genetic relatedness was illustrated by a dendrogram construction based on rep-PCR fingerprint analysis. The repetitive regions were amplified by 1 µM REP-1 primer (5 -GCGCCGICATCAGGC-3 ) and 1 µM REP-2 primer (5 -ACGTCTTATCAGGCCTAC-3 ) in 50 µL of PCR reaction mixture [67]. The DNA fingerprint patterns were illustrated by 1.5% agarose gel electrophoresis, and the dendrogram was constructed using the Bionumeric software (Applied Maths, Sint-Martens-Latem, Belgium) with 1% position tolerance. Isolates presenting more than 70% similarity were categorised in the same cluster. Representative isolates of identical phylogroups and rep-PCR patterns that contained each bla CTX-M variants and the strains having bla NDM-5 or mcr-3 genes were selected for MLST. The STs were identified by the Achtmann MLST scheme by comparing them to sequences and alleic profiles on the Enterobase website (https://enterobase.warwick.ac.uk/species/ecoli/) (accessed on 15 March 2021) [13].

Conclusions
This study highlights that canine and feline ESC-resistant UPEC are genetically linked to high-risk, MDR, human-associated ExPEC clones. Various genes corresponding resistance to critically and highly critically important antimicrobials, including bla CTX-M types, pAmpC, bla NDM-5 and mcr-3, were first detected in E. coli from small animal origins in Thailand. This evidence supports the emergence of resistance to last-resort antimicrobials in companion animals, highlighting a "spillover" of the bacteria that may be bidirectionally transmitted between humans and pets.
Funding: This study was supported by Chulalongkorn University to the 90th Anniversary of Chulalongkorn University Fund (Ratchadaphiseksompoch Endowment Fund), Chulalongkorn Academic Advancement into its 2nd Century Project, the Center of Excellence for the Emerging and Re-emerging Infectious Diseases in Animals (CUEIDAs) and the One Health Research Cluster (7640002-HE02).

Institutional Review Board Statement:
This study was conducted according to the faculty regulations and approved by the Chulalongkorn University-Faculty of Veterinary Science Biosafety Committee (CU-VET-BC) (Protocol code 2031058) on 21 January 2021.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data available from the authors upon reasonable request.