Broad-Spectrum Cephalosporin-Resistant and/or Fluoroquinolone-Resistant Enterobacterales Associated with Canine and Feline Urogenital Infections

The aim of the present study was to characterize Enterobacterales resistant to 3rd and 4th generation cephalosporins, carbapenems and/or fluoroquinolones, isolated from dogs and cats with urogenital infections. In total, 36 strains (Escherichia coli (n = 28), Klebsiella pneumoniae (n = 3), Serratia marcescens, Raoultella ornithinolytica, Proteus mirabilis, Citrobacter portucalensis and Enterobacter cloacae (each n = 1)) were included in the present study, 28 from Austria and 8 from Serbia. Isolates were characterized by a polyphasic approach including susceptibility pheno- and genotyping and microarray-based assays. Escherichia (E.) coli isolates were additionally characterized by two-locus (fumC and fimH) sequence phylotyping and multi-locus sequence typing (MLST) of selected isolates. MLST of carbapenem-resistant Enterobacter cloacae isolates was also performed. Among E. coli, the most dominant phylogenetic group was B1 (27.8%), followed by C, (16.6%), A and Clade II (5.5% each), B2 and F (2.77% each). The most predominant β-lactam resistance genes were blaTEM (70%) and blaCTX-M (38.8%), blaCMY (25%). blaNDM was detected in one carbapenem-resistant Enterobacter cloacae ST114. The most common ST among selected E. coli was 744 (10.7% isolates). The pandemic clones ST131 and ST648 carrying CTX-M-15 were also detected. Remaining STs belonged to 469, 1287, 1463 and 1642. E. coli clonotyping revealed 20 CH types. Based on the presence of certain virulence genes, three isolates were categorized as ExPEC/UPEC. The most prevalent virulence factors were fimH detected in 61%, iucD and iss both in 55%, iroN in 27.8%, papC in 13.8% and sat in 8.3% isolates.


Introduction
The most common causative agents of urogenital infections (UGI) are the members of the order Enterobacterales, amongst them primarily Escherichia (E.) coli with a prevalence in companion animals between 35-70% [1] and in humans between 75-95% [2]. UGI include infections of the bladder, kidneys

Results
In Austria, isolates belonging to the order Enterobacterales were detected in 550 of 2398 samples resulting in a carriage rate of 22.94% whereas 28 isolates were broad-spectrum cephalosporin-resistant and/or fluoroquinolone-resistant resulting in a carriage rate of 1.17% over all tested animals. We observed higher carriage rates in Serbia. In 45 out of 144 samples (carriage rate 31.25%), isolates belonging to Enterobacterales were detected whereas eight were broad-spectrum cephalosporin-resistant and/or fluoroquinolone-resistant resulting in a carriage rate of 5.56% over all tested animals. Twenty-eight isolates were identified as E. coli and the remaining isolates were Klebsiella pneumoniae (n = 3), Raoultella ornithinolytica (n = 1), Citrobacter portucalensis (n = 1), Proteus mirabilis (n = 1), Serratia marcescens (n = 1) and Enterobacter cloacae (n = 1) (Tables 1 and 2). All isolates were resistant to at least one antibiotic from at least three different classes and accordingly, were categorized as MDR [37]. They originated from dogs (n = 30) and cats (n = 6), wherefrom 58.3% (n = 21) were isolated from urine (obtained via cystocentesis, which is routine), 27.8% (n = 10) from the vagina, 2,7% (n = 1) from an ejaculate, 5.5% (n = 2) from the prostate gland and 5.5% (n = 2) from the bladder wall.
All isolates examined were susceptible to amikacin and only one (Enterobacter cloacae) was resistant to meropenem. Low resistance to nitrofurantoin and fosfomycin (in one Klebsiella pneumoniae and Enterobacter cloacae isolate) was observed. In fourteen isolates (E. coli (n = 11), Klebsiella pneumoniae (n = 1), Citrobacter portucalensis (n = 1) and Enterobacter cloacae (n = 1)), genes from the bla CTX-M family were detected, alone or in combination with other bla genes. In nine E. coli isolates exhibiting the AmpC/ESBL phenotype, bla CMY alone or together with mutation events in the ampC promoter/attenuator region were found. In two isolates displaying the AmpC phenotype, these mutations were detected without the presence of bla CMY or other bla genes (Tables 1 and 2). New Delhi metallo-β-lactamase (bla NDM-1 ) was found in one carbapenem-resistant Enterobacter cloacae from Serbia. The same isolate harbored bla CTX-M- 15 , bla NDM and bla OXA-1 . Ciprofloxacin resistance was found in 86.1% (n = 31) isolates. Plasmid mediated quinolone resistance (PMQR) genes were identified in 33.3% (n = 12) isolates, of which nine isolates harbored qnrB and 3 isolates qnrS genes. In all but three ciprofloxacin-resistant isolates, amino acid substitutions were detected in the QRDR of gyrA and/or parC. The genes sul1, sul2 and tet(A) were the most prevalent non-β-lactamase genes and non-PMQR detected.

Discussion
Among Austrian E. coli isolates, clones belonging to B2 (phylotype)-O25b:H4 (serotype)-ST131 is widely recognized as one of the most important pandemic pathogens. This clone has been shown to cause various infections in humans and animals ranging from asymptomatic bacteriuria to complicated UGIs, septicemia, meningitis, and pneumonia. In several publications E. coli ST131 has been unanimously classified as ExPEC and/or UPEC [1,2,4,7,8,11,12,14,15,17,35,[38][39][40]. According to origin and site of infection, E. coli ST131 strains were additionally categorized as FUPEC, NEMEC, APEC or SEPEC. It is believed that the global dissemination of E. coli ST131 as well as E. coli ST648 has contributed to the worldwide emergence of CTX-M-15 producing E. coli [14,41]. Anyway, CTX-M-15 is a mostly plasmid-transferable gene and not all ST131 E. coli isolates harbor it. Thus, E. coli ST131 may be susceptible to cephalosporins [42]. ST131 is mostly (but not necessarily) associated with fluoroquinolone resistance [35,43]. So far, CTX-M-15 harboring and fluoroquinolone-resistant E. coli B2-O25b:H4-ST131 has been described as the cause of UGI in dogs in Portugal, the Netherlands, Switzerland, the United Kingdom and France [24,33]. It has also been isolated from dogs and cats in the USA, from dogs, cats and horses in Denmark, Germany and Spain and from dogs in China [35,44]. E. coli B2-O25b:H4-ST131-CTX-M-27 variant has been isolated from urine specimen of dogs and cats in Japan [31]. Non CTX-M-15, fluoroquinolone-resistant E. coli B2-O25b:H4-ST131 has also been detected in the urinary tract of dogs and cats in the USA and its transmission from companion animals to humans has been suspected suggesting dogs and cats as possible reservoir of this clone [14]. However, E. coli ST131 has never been reported to be isolated from the prostate of a dog, as was the case in our study. Zoonotic transmission of E. coli ST131 is still the subject of scientific debates. Transmission of E. coli ST131 from poultry to humans and rats has been observed and transfer of E. coli ST131 via meat products, particularly from chicken meat to humans causing UGI has been proven [2,13,17]. In Austria, this pandemic clone has only been isolated from wildlife so far [45].
Another clone, E. coli F (D)-ST648 (CTX-M-15) which also has readily been classified as ExPEC or/and UPEC was recognized as the zoonotic pandemic agent of various types of infections most common in humans but also in livestock and companion animals [40]. It has been detected in pigs, calves and other livestock in Germany and the Netherlands, dogs and cats in the USA, Germany and Italy as well as in wild birds in Mongolia and Germany [34]. It was first isolated as a cause of cystitis in a cat in 2013 in Switzerland [15], later it has been detected in urine samples of dogs with UGI, also in Switzerland [33]. It was a frequently isolated clone (D-ST648 [CMY, TEM]) from intestinal samples originating from stray dogs reported in a study from South Korea [23]. Another significant finding in this study is, to our knowledge the first isolation of a CTX-M-15 and NDM-1, together with OXA 1, TEM producing Enterobacter cloacae ST114 from a vaginal swab of a dog from Serbia (strain ESBL40). Carbapenems-resistance in bacteria of animal origin is a subject of great concern due to their importance in therapy of life-threatening infections in human and in veterinary medicine [26,43,46]. Although several publications state that the risk of carbapenem resistance in animals is low due to the limited use of this class in veterinary medicine [26,43,46], this is the second report on the detection of carbapenemases carrying bacteria associated with canine UGI from Serbia [36] and bla NDM , bla VIM , bla IMP , bla OXA and bla KPC carbapenemases have been detected worldwide in companion animals. Unlike carbapenemases, CTX-M-15 is highly prevalent in Enterobacter cloacae and CTX-M-15-producing ST114 Enterobacter cloacae has been recently reported as a high-risk clone for public health which was predominant in cats, dogs and horses [47,48].
Of the other ST found in our study, E.  [33] and E. coli ST744-CTX-M-1 was also reported in livestock [43].
In our study, one AmpC/ESBL E. coli C-O9:H9-ST1287-CH11-54 (TEM, CMY) was detected in a vaginal swab of a dog. Fluoroquinolone resistant, non-AmpC, non-ESBL E. coli ST1287 categorized as APEC has been reported in a parrot pet from Brazil [49] and in a blood sample from a patient with septicemia as well as in the feces of a healthy individual [21,50]. This ST seems to be rare and, to our knowledge, AmpC/ESBL/FQR E. coli ST1287 has been isolated from a dog for the first time.
E. coli B1-ST469 and B1-ST1463, which have also been found in our study, have rarely been reported, mostly in humans and turkey [22]. To the best of our knowledge, these two STs have not been isolated from cases of UGI in companion animals, so far.
One E. coli isolated from the vagina of a dog has been typed as B1-ONT:H7-ST1642-CH11-54 (TEM). The same ST (B1-O8:H7-ST1642) harboring CTX-M-55 has been isolated from a urine specimen of a cat in Japan [31], the CTX-M-14 harboring variant has been detected in human blood samples [40], the CTX-M-15 variant has been found in feces of healthy lambs in Brazil [30] and the E. coli B1-ST1642 harboring bla TEM and bla CMY has been detected in an intestinal specimen from a stray dog in South Korea [23].
In CTX-M harboring isolates, bla CTX-M-1 was more common which is also inconsistent with findings in other studies mostly reporting CTX-M-15 being predominant among ESBL producing enterobacteria isolated from humans and animals [24]. In addition, CTX-M-14 and CTX-M- 55 harboring Enterobacterales have been increasingly reported in companion animals [23,24], which were not detected in our investigation. The least prevalent CTX-M variant was bla CTX-M-9 , detected in one AmpC/ESBL/FQR E. coli from Austria (ID 4) (Clade II-CH10-58 [CTX-M-9, TEM]). In one large study among companion animals in China, out of 107 CTX-M producing E. coli, only one was CTX-M-9 [51]. In another study from UK, out of 876 investigated and 164 cephalosporin-resistant E. coli originating from companion animals, neither one harbored CTX-M-9 [24].
Based on the published surveys [19,[26][27][28], the use of critically important antimicrobials (amongst them 3rd and 4th generation cephalosporins and fluoroquinolones) is cited as being frequently prescribed for dogs and cats in a number of countries including Belgium, Spain, France, the United Kingdom and Germany. Carbapenems are mostly prescribed in complicated cases (MDR infections) based on results of susceptibility testing and treatment durations are longer than typically recommended in human medicine [26].
In our study, MDR (ESBL/FQR) E. coli ST131 (555) and Klebsiella pneumoniae (3824) have been isolated from two canine prostate samples. Not all antibiotics are capable of penetrating the prostate tissue, so the therapy choice is initially limited to enrofloxacin, trimethoprim and fosfomycin [18]. Both prostate isolates from our study were susceptible only to fosfomycin which has not been licensed for the use in companion animals, neither in Serbia, nor in Austria. Therefore, this antibiotic is not included in routine susceptibility testing, though out of label use would be feasible. Besides, CLSI does not provide interpretative categories for fosfomycin in its document [29] and human-related interpretative categories may be used, although interpretation should then be performed with caution. The finding of MDR bacteria in the canine prostate (including human-related clones such as E. coli ST131) is becoming a reality and one should expect more such cases in the future. Besides, out of 36 MDR strains included in this study, 35 were susceptible to fosfomycin. The only fosfomycin resistant strain was a carbapenem resistant Enterobacter cloacae ST114 isolated from the vagina of a Serbian dog. Since fosfomycin is a critically important human-related antibiotic according to WHO [52], the emergence of bacterial strains susceptible only to fosfomycin in companion animals may lead to an undesired, more frequent use of this antibiotic in veterinary medicine in the future.
Ten E. coli and the Citrobacter portucalensis isolate (9 originating from urine samples) from our study exhibited an ESBL/AmpC/FQR resistance phenotype and all were susceptible to nitrofurantoin and fosfomycin. Nitrofurantoin has also been listed in the ISCAID for the treatment of MDR related UGIs in companion animals. Interpretative categories for nitrofurantoin are presented in the CLSI VET01, Ed4E document, though nitrofurantoin has not been licensed for use in companion animals, neither in Austria nor in Serbia.
Although it has been proven that ExPEC/UPEC clones that share the same antibiotic resistance patterns may cause UGIs in humans as well as in animals, recommendations for the treatment of UGIs in humans differ significantly from guidelines related to animals. The Infectious Diseases Society of America (IDSA) in collaboration with the European Society for Microbiology and Infectious Diseases (ESCMID) removed beta-lactams and fluoroquinolones from the list of first-line antibiotics for the treatment of uncomplicated UGIs in adult humans [53]. Instead, nitrofurantoin, trimethoprim-sulfamethoxazole and fosfomycin are recommended. Amoxicillin and ampicillin have been removed because of their relatively poor efficacy and the high prevalence of antimicrobial resistance to these agents worldwide [53].
There are no defined criteria for strict categorization of E. coli strain as ExPEC or UPEC or any other subcategory (APEC, NEMEC, SEPEC), because all these subcategories mostly share the same virulence genes [1,4,[7][8][9]13,17,38,39,[54][55][56]. Some authors assume that the presence of at least two of the following genetic determinants indicates categorization of an isolate as ExPEC/UPEC: papA and/or papC (P fimbriae structural subunit and assembly), sfa/focDE (S and F1C fimbriae), afa/draBC (Dr binding adhesins), iutA (aerobactin system) and kpsM II (group 2 capsule) [1]. This minimal predictive set of ExPEC virulence genes has not been thoroughly validated and thus is considered incomplete as strains meeting this definition do not always cause disease. Accordingly, drawbacks of such a scheme were shown in a study [57] in which E. coli ST648 has been categorized as pandemic ExPEC but results demonstrated that out of 47 ST648 isolates only 17 were assignable to ExPEC based on the aforementioned scheme. Thus, we did not use this scheme in our study. However, since we considered iucD and iutA to have the same function [40] encoding aerobactin ferric receptors, 3 of our investigated isolates that were shown to harbor papC/iucD could potentially be classified as Another approach is to associate virulence factor(s) (VF(s)) with ST, CH types, phylogroups, antibiotic resistance and thus to establish pathogenic and zoonotic potential of the suspected ExPECs. Therefore, we included VFs mostly associated with ExPEC or UPEC [17] amongst them papC, fimH, iucD, sat, cnf 1, iss, iha, ast, nfa, iroN, tsh and hly. Based on this expanded list of VFs our 18 isolates investigated can be assigned to ExPEC/UPEC (containing 3 or more VFs from the expanded VF list).
The fimH gene was found in 61% (n = 20 Austrian and n = 2 Serbian) isolates. It encodes FimH adhesins/type 1 fimbriae, which are essential virulence factors for uropathogenic E. coli and their successful colonization of the urinary bladder epithelium [58] but are also referred as a major virulence factor of E. coli belonging to the ExPEC category [5,6,58]. Anyway, this type of adhesin is not only typical for uropathogenic but also for intestinal as well as enterohaemorrhagic (EHEC) isolates. Reports show that fimH is equally represented in E. coli isolated from both humans and animals [59].
The gene iucD encodes for an enzyme included in the aerobactin biosynthesis pathway [60]. The aerobactin siderophore system is an important virulence factor and it is absent in avirulent E. coli strains, thus, finding iucD in 55% (n = 15 Austrian, n = 5 Serbian) may serve as evidence for pathogenicity of the isolates [60]. The secreted autotransporter toxin SAT is a virulence factor of UPEC and/or diffusely adhering E. coli (DAEC) both associated with pyelonephritis and urosepsis [61]. Sat is reported to be predominant in E. coli of human origin [57]. In our study, it was found in 3 isolates, two from Austria (555: B2-O25b:H4-ST131-CH40-30; 3056: O25:H4-CH40-30) and one from Serbia (1srb: C-CH11-0 [TEM, OXA-2]). The latter also harbored the ast gene encoding heat-stable enterotoxin 1 which is mostly found in ETEC, thus isolate 1srb probably belonged to diarrheal E. coli. P fimbriae (pap) are the second common virulence factor of UPEC, which plays an important role in the pathogenesis of ascending UGIs and pyelonephritis in humans [61], it was found in 13.8% (n = 5) isolates in our study.
Gene tsh encoding temperature-sensitive haemagglutinin was found in two Austrian (3441: C-ONT:H9-ST744; 3448: C-ST744) and one Serbian strain (MDR56: A-O128:H26). It is predominant in E. coli isolated from animals [57] and mostly present in APECs as previously reported with a prevalence of more than 50% in APEC, 4.5% in UPEC and 11.5% in NMEC isolates tested [5]. Strain MDR56 C-STNT probably was of intestinal origin. This is supported by the finding that it belongs to phylogroup A to which commensal E. coli from the gastrointestinal tract mostly belong [38]. Besides, MDR56 was typed as O128 (H26) which is a serotype often reported as shiga toxin-producing E. coli (STEC) mostly found in sheep and humans [59]; however, MDR56 was shown to be negative for stx.
The increased serum survival gene iss has been reported in E. coli isolated from humans and animals [57] and in approximately 60% of UPEC and NMEC strains. It has also been associated with APEC strains [60]. In our study it was found in 55% (n = 20) of isolates.

Materials and Methods
During the study period between October 2010 and March 2015, 2398 samples from Austria as well as between January 2012 and March 2015, 144 samples from Serbia taken from the canine and feline urogenital tract were considered in the present study. If the microbiological examination yielded bacterial growth, isolates were identified to the species level by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry (Bruker Daltonik, Heidelberg, Germany). Antimicrobial susceptibility testing was performed by agar disk-diffusion according to standards of the Clinical and Laboratory Standards Institute (CLSI) [62]. Escherichia coli ATCC ® 25922 served as quality control strain. The following antimicrobials were used-ampicillin, cefotaxime, ceftazidime, cefepime, aztreonam, meropenem, gentamicin, amikacin, tobramycin, ciprofloxacin, trimethoprim-sulfamethoxazole, tetracycline fosfomycin, nitrofurantoin and chloramphenicol (Becton Dickinson, Heidelberg, Germany). In addition, isolates were checked for extended-spectrum β-lactamase (ESBL) production by ESBL-test via agar disk diffusion combination disk tests using cefotaxime and ceftazidime with and without clavulanic acid [62]. Furthermore, cefoxitin (30 µg) was added to this test, to detect AmpC phenotypes. Isolates belonging to Enterobacterales and were broad-spectrum cephalosporin-resistant and/or fluoroquinolone-resistant were included in the present study.
All isolates were obtained from samples during routine bacteriological diagnostics at the Institute of Microbiology, University of Veterinary Medicine, Vienna, Austria as well as the Department for Microbiology, Faculty of Veterinary Medicine University of Belgrade, Serbia. All these clinical samples were received from third parties and therefore not subject to reporting obligations of the Ethics and Animal Welfare Commission of the University of Veterinary Medicine in Vienna and not subjected to Ethics Commission for the experimental animals welfare protection of the Faculty of Veterinary Medicine in Belgrade. All isolates were stored at −80 • C until further examination.
The serogenotyping and the detection of resistance genes and major virulence genes of E. coli isolates was performed using the miniaturized microarray-based E. coli PanType AS-2 kit (Alere, Jena, Germany). The same assay was used to detect resistant determinants in non-E. coli Enterobacterales. Resistance genes in one carbapenem-resistant Enterobacter cloacae-complex isolate were analyzed by CarbDetect-AS-2 Kit microarray (Alere, Jena, Germany) [63,64]. PCRs were performed to detect tet(A) and tet(B) in carbapenem resistant Enterobacter cloacae as well as to amplify bla NDM and bla CTX which were then sequenced [65,66]. Sequences were aligned with BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and compared with reference sequences available in GenBank and the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/pathogens/beta-lactamase-data-resources/). The quinolone resistance-determining regions (QRDR) of gyrA and parC in ciprofloxacin-resistant isolates were amplified by PCR and sequenced [67]. E. coli isolates displaying the AmpC phenotype were also analyzed for mutations in the chromosomal ampC promoter/attenuator region as described previously [68]. Allele-specific PCR was performed to identify the O25 clone of selected E. coli isolates [69]. The E. coli isolate was assigned to a phylogroup using the quadruplex assignment method [70]. Clonal relatedness of E. coli isolates was assessed by two-locus sequence typing of combined data of fumC and fimH sequences as described by Weissman et al. [71]. Allele and CF clonotype numbers were used for goeBURST analysis using PHYLOViZ [72]. Selected E. coli isolates were further subjected to multilocus sequence typing (MLST) [73]. Allelic profiles and sequence types (ST) were determined by querying the E. coli MLST website (http://enterobase.warwick.ac.uk/species/ecoli/allele_st_search). The carbapenem-resistant Enterobacter cloacae isolate was characterized by MLST according to the protocol recommended by Miyoshi-Akiyama et al. [74] and ST was determined using Enterobacter cloacae MLST database (https: //pubmlst.org/ecloacae/).

Conclusions
The study discovered E. coli B2-O25b:H4-ST131-CH40-30-CTX-M-15 and FQR in a prostate specimen from a dog and E. coli F (D)-ONT:H6-ST648-CH4-27-CTX-M-15 and FQR in a canine vaginal sample, which are important pandemic pathogens that have been shown to cause various infections in humans and animals. Furthermore, carbapenem-resistant Enterobacter cloacae ST114 harboring bla NDM , bla CTX-M-15 , bla OXA-1 and bla TEM was detected in a canine vaginal swab from Serbia, which is considered to be the first report of NDM-1 harboring Enterobacter cloacae isolated from a dog. The findings in this study suggest a possible reverse zoonotic transmission of ExPEC/UPEC-companion animals may be infected by human-related clones because of their close co-existence and then may become an important source of those clones for the human population.