ESBL/pAmpC-Producing Escherichia coli Causing Urinary Tract Infections in Non-Related Companion Animals and Humans.

Urinary tract infections (UTI) caused by Escherichia coli are frequently diagnosed in humans and companion animals. Extended-spectrum beta-lactamase (ESBL)- and cephalosporinase (pAmpC)-producing Escherichia coli are worldwide-disseminated and frequently multidrug-resistant, hence leading to treatment failure and public health concerns. This study aimed to characterize and compare ESBL/pAmpC-producing E. coli strains causing community-acquired UTI in companion animals and non-related humans. Third-generation cephalosporin (3GC)-resistant E. coli (companion animals n = 35; humans n = 85) isolated from patients with UTI were tested against 14 antimicrobials following CLSI guidelines. PCR-based assays were used to detect the major E. coli phylogenetic groups, pathogenicity associated-islands (PAIs), virulence genes, and ESBLs/pAmpC resistance genes. ESBL/pAmpC-producing E. coli isolates were typed by multi-locus sequence typing (MLST) and PCR. E. coli strains from companion animals and humans shared two MDR high-risk clonal lineages: ST131 and ST648. To the best of our knowledge, this study reports the first description of E. coli ST131 clade C1-M27 and the clonal lineage ST131 clade A in humans with community-acquired UTI in Portugal. Considering that companion animals with UTI are generally treated at home by the owners, measures should be implemented to avoid the spread of multidrug-resistant high-risk clones to humans and their household environment.

1. Introduction

Urinary tract infections (UTI) may be caused by the uropathogenic Escherichia coli (UPEC), which is one of the extraintestinal pathogenic E. coli pathotypes (ExPEC), and one of the most frequent etiologic agents of UTI worldwide both in humans and companion animals [1,2,3,4].

The increase in antimicrobial resistance caused by the dissemination of resistant and multidrug-resistant (MDR) bacteria, such as extended-spectrum beta-lactamase (ESBL), cephalosporinase (AmpC) and carbapenemase-producing E. coli, is a current global threat responsible for thousands of deaths each year [5,6]. ESBL/AmpC and carbapenemase-producing E. coli are frequently MDR, thus causing treatment failure due to their ability to hydrolyze third- and fourth-generation cephalosporins or carbapenems, which have been considered critically important antimicrobials to human and veterinary medicine [5]. Moreover, in recent decades, some studies have alerted us to the emergence of MDR high-risk clonal lineages of clinically significant bacteria in companion animals, raising public-health concerns, since infected and colonized companion animals may contribute to the spread of such bacteria among humans, domestic animals, wildlife, and the environment [4,7,8,9]. The need for a global action plan to address the antimicrobial resistance crises requires a One Health approach supported by scientific data that can be used to raise awareness of decision-makers and the general population [6].

The production of beta-lactamase (ESBL/AmpC and carbapenemase) enzymes is an important factor that promotes the dissemination of high-risk clonal lineages that are also frequently MDR. The most prevalent of these are the ESBL CTX-M-like enzymes, although others, such as TEM and SHV enzymes and the plasmid-mediated AmpC (pAmpC) are also common. There is a large body of research showing that the distribution of ESBL/pAmpC varies geographically and according to the animal species considered. Interestingly, CTX-M-15 is one of the most frequent ESBLs detected among E. coli isolated from companion animals and humans in Europe and the Americas [10].

Besides the association with various antimicrobial resistance determinants, the classification as an international MDR high-risk clonal lineage is also linked with the bacterial pathogenicity, global distribution, ability to colonize and persist in the hosts for more than 6 months, capacity for transmission between hosts, and the ability to cause recurrent infections [11]. However, information about the pathogenicity and virulence of E. coli strains isolated from companion animals is still scarce.

The population structure of ESBL/pAmpC-producing E. coli is currently dominated globally by several high-risk clonal lineages, including the ST131, ST648, ST69, ST393, ST405, and ST410 lineages [11,12]. ST131 is one of the most successful E. coli clonal lineages, it is disseminated worldwide and is frequently a CTX-M-15 producer [8]. Although seemingly more frequent in humans, the ST131 lineage is also being increasingly detected in companion animals [8]. E. coli ST131 has been grouped into different clades, which are usually associated with specific fimH (type-1 fimbriae) alleles: clade A (fimH41 ST131-O16), clade B (fimH22 ST131-O25b), and clade C (fimH30 including ST131-O25b fimH30-R/fimH30-Rx). E. coli ST131 clade C1-M27 is associated with blaCTX-M-27, and C2 with blaCTX-M-15 [11,12]. The ST648 is also an important pandemic clonal lineage as it is frequently an ESBL producer and commonly causes infection in humans and companion animals [13]. The frequency and geographic distribution of beta-lactamase enzymes and E. coli clonal lineages may change over time, thus creating the need for continuous monitoring in animals and humans.

With the growing contact between companion animals and humans, the risk of animal-to-human transfer of ESBL/pAmpC-producing bacteria is a concern [14]. The risk of inter-species transfer of antimicrobial resistant bacteria is supported by previous studies that have demonstrated that companion animals may share UPEC with the remaining human household members [15,16,17,18,19]. Interestingly, humans may also be a reservoir of UPEC for their companion animals as seen with E. coli O25b:H4-B2-ST131. This clonal lineage has dramatically spread during recent decades and some of its clades, such as the H30Rx C2 (C2/H30Rx), have been linked to the dissemination of ESBLs, especially the CTX-M-15 enzyme [15,16,17,18,19]. The dissemination of ESBL/pAmpC-producing E. coli is a complex issue since their emergence, ecology, and association with virulence are still poorly understood both in human and companion animal infections. Thus, this study aimed to characterize and compare the antimicrobial resistance, virulence profile, phylogenetic distribution, and predominant clonal lineages of ESBL/pAmpC-producing E. coli strains causing UTI in companion animals and non-related humans.

2. Results

2.1. Antimicrobial Resistance and Phylogenetic Group of 3GC-Resistant E. coli

From a total of 330 non-duplicate E. coli isolates included from companion animals with UTI, 10.6% (n = 35/330) were resistant to third-generation cephalosporin (3GC). A high proportion of these 3GC-resistant E. coli were also resistant to ciprofloxacin (74.3%, n = 26/35), norfloxacin (71.4%, n = 25/35), trimethoprim/sulfamethoxazole (71.4%, n = 25/35), gentamicin (40.0%, n = 14/35), and tobramycin (31.4%, n = 11/35). Overall, 71.4% (n = 25/35) of these strains were MDR. However, no resistance to carbapenems was detected. Regarding 3GC-resistant E. coli from humans with UTI, these showed higher resistance frequencies against fluoroquinolones (88.2%, n = 75/85), trimethoprim/sulfamethoxazole (74.1%, n = 63/85), gentamicin (37.6%, n = 32/85), and tobramycin (49.4%, n = 42/85). Overall, 84.7% (n = 72/85) of the E. coli strains were MDR (Table 1).

Table 1

Antimicrobial resistance of 3GC-resistant E. coli isolated from companion animals and humans with UTI.

Antimicrobials	Companion Animal(N = 35) a%R (n)	Human-CA(N = 85) a%R (n)	p Value *
Ampicillin/amoxicillin	100% (n = 35)	100% (n = 85)	N.s.
Amoxicillin/clavulanate	77.1% (n = 27)	27.1% (n = 23)	<0.0001
Cefoxitin	62.9% (n = 22)	8.2% (n = 7)	<0.0001
Cefotaxime	91.4% (n = 32)	100% (n = 85)	0.023
Ceftazidime	62.8% (n = 22)	42.4% (n = 36)	0.047
Imipenem	0.0% (n = 0)	0.0% (n = 0)	N.s.
Meropenem	0.0% (n = 0)	0.0% (n = 0)	N.s.
Ciprofloxacin	74.3% (n = 26)	88.2% (n = 75)	0.096
Norfloxacin	71.4% (n = 25)	88.2% (n = 75)	0.033
Nitrofurantoin	5.7% (n = 2)	2.4% (n = 2)	N.s.
Gentamicin	40.0% (n = 14)	37.6% (n = 32)	0.838
Amikacin	5.7% (n = 2)	10.6% (n = 9)	0.506
Tobramycin	31.4% (n = 11)	49.4% (n = 42)	0.105
Trimethoprim/sulfamethoxazole	71.4% (n = 25)	74.1% (n = 63)	0.822
Multidrug resistant	71.4% (n = 25)	84.7% (n = 72)	0.125

%R, percentage of resistant strains; Human-CA, human community-acquired UTI; n, number of strains; a the number shown (N) is the total number of strains tested. * p value < 0.05 was considered statistically significant. N.s. not significant to calculate.

Regarding the phylogenetic group, 3GC-resistant E. coli isolates from companion animals belonged mainly to group-D (48.6%, n = 17/35) and 3GC-resistant E. coli strains from humans belonged mainly to group-B2 (67.1%, n = 57/85) (Table 2).

Table 2

Phylogenetic groups of 3GC-resistant E. coli isolated from companion animals and humans with UTI.

Phylogenetic Group	Companion Animal(N = 35) a% (n)	Human-CA(N = 85) a% (n)	p Value *
Group A	22.9% (n = 8)	12.9% (n = 11)	0.187
Group B1	11.4% (n = 4)	14.1% (n = 12)	0.777
Group B2	17.1% (n = 6)	67.0% (n = 57)	<0.0001
Group D	48.6% (n = 17)	5.9% (n = 5)	<0.0001

Human-CA, human community-acquired UTI; n, number of strains; a the number shown (N) is the total number of strains tested. * p value < 0.05 was considered statistically significant.

The frequency of ESBL-producing E. coli was significantly higher in 3GC-resistant strains from humans with UTI than in companion animals (p < 0.0001). ESBLs were detected in E. coli belonging to all phylogenetic groups, including group-A which is considered less pathogenic. Nevertheless, group-B (including group-B1 and group-B2) predominated (Table 3).

Table 3

ESBL and pAmpC-producing E. coli by phylogenetic group.

PhylogeneticGroup	Companion Animal(N = 35) a		PhylogeneticGroup	Human-CA(N = 85) a		p Value *ESBLGenes	p Value *pAMPcGenes
	ESBL Genes% (n)	pAmpC Genes% (n)		ESBL Genes% (n)	pAmpC Genes% (n)
A (n = 8)	62.5% (n = 5)	0.0% (n = 0)	A (n = 11)	100% (n = 11)	0.0% (n = 0)	0.057	N.s
B1 (n = 4)	75.0% (n = 3)	50.0% (n = 2)	B1 (n = 12)	100% (n = 12)	0.0% (n = 0)	0.250	0.050
B2 (n = 6)	83.3% (n = 5)	33.3% (n = 2)	B2 (n = 57)	94.7% (n = 54)	5.3% (n = 3)	0.337	0.067
D (n = 17)	5.8% (n = 1)	94.1% (n = 16)	D (n = 5)	60.0% (n = 3)	40.0% (n = 2)	0.024	<0.0001
Overall	40.0% (n = 14)	57.1% (n = 20)	Overall	94.1% (n = 80)	5.9% (n = 5)	<0.0001	<0.0001

% percentage of strains; Human-CA, human community-acquired UTI; n, number of strains; a the number shown (N) is the total number of strains tested; ESBL, extended-spectrum β-lactamase; pAmpC, pAmpC β-lactamases. * p value < 0.05 was considered statistically significant. N.s. not significant to calculate.

pAmpC-producing E. coli strains belonged mainly to group-D in both groups, and were significantly more frequent in companion animals (94.1% in E. coli from companion animal; 40.0% in E. coli from humans, p < 0.0001) (Table 3).

Although blaSHV was detected among ESBL-producing E. coli, blaCTX-M-type ESBL clearly predominated in both groups. Several types of CTX-M enzymes were found showing high diversity of these ESBLs in UPEC, especially in strains from humans. Nevertheless, blaCTX-M-15 predominated in both groups (Table 4).

Table 4

ESBLs and pAmpC genes in 3GC-ressistant E. coli isolated from companion animals and humans with UTI.

Beta-Lactamase Gene	Companion Animal(N = 35) a% (n)	Human-CA(N = 85) a% (n)
bla SHV-12	0.0% (n = 0)	1.2% (n = 1)
bla CTX-M-1	5.7% (n = 2)	10.6% (n = 9)
bla CTX-M-1-type	5.7% (n = 2)	0.0% (n = 0)
bla CTX-M-15	20.0% (n = 7)	54.2% (n = 46)
bla CTX-M-15-type	0.0% (n = 0)	1.2% (n = 1)
bla CTX-M-32	8.6% (n = 3)	5.9% (n = 5)
bla CTX-M-9	2.9% (n = 1)	0.0% (n = 0)
bla CTX-M-9-type	0.0% (n = 0)	4.7% (n = 4)
bla CTX-M-14	0.0% (n = 0)	8.2% (n = 7)
bla CTX-M-27	0.0% (n = 0)	7.1% (n = 6)
bla CTX-M-2group	0.0% (n = 0)	1.2% (n = 1)
bla CMY-2	57.1% (n = 20)	5.9% (n = 5)

% percentage of strains; Human-CA, human community-acquired UTI; n, number of strains; a the number (N) shown is the total number of strains tested.

Regarding pAmpC, only blaCMY-2 was found, being the predominant antimicrobial resistance mechanisms responsible for 3GC-resistance in E. coli from companion animals. Interestingly, the cefoxitin resistance phenotype of four 3GC-resistant E. coli strains from companion animals could not be explained by any of the tested genes; thus, other mechanism of resistance were likely involved.

Moreover, carbapenemase genes were not detected in either 3GC-resistant E. coli collections, which is in line with the carbapenem-susceptible phenotype of these strains.

2.2. Pathogenicity Island Markers and Virulence Genotyping of 3GC-Resistant E. coli

Eight UPEC pathogenicity islands (PAIs) were screened in all 3GC-resistant E. coli strains. The most prevalent PAIs among strains from humans and companion animal were PAIIV536 (91.8%, n = 78/85; and 74.3%, n = 26/35, respectively), followed by PAICFT073 (78.8%, n = 67/85; and 54.3%, n = 19/35, respectively) (Table 5).

Table 5

Frequency of pathogenicity island markers (PAIs) among 3GC-resistant E. coli strains from companion animals and humans with UTI.

Detected PAIs	Companion Animal(N = 35) a% (n)	Human-CA(N = 85) a% (n)	p Value *
PAIIJ96	0.0% (n = 0)	0.0% (n = 0)	N.s.
PAIIIJ96	11.4% (n = 4)	41.2% (n = 35)	0.004
PAII536	8.6% (n = 3)	40.0% (n = 34)	0.0005
PAIII536	22.9% (n = 8)	35.3% (n = 30)	0.203
PAIIII536	0.0% (n = 0)	0.0% (n = 0)	N.s.
PAIIV536	74.3% (n = 26)	91.8% (n = 78)	0.017
PAIICFT073	54.3% (n = 19)	78.8% (n = 67)	0.013
PAIIICFT073	20.0% (n = 7)	69.4% (n = 59)	<0.0001

n, number of strains; Human-CA, human community-acquired UTI; a the number shown (N) is the total number of strains tested. * p value < 0.05 was considered statistically significant. N.s. not significant to calculate.

All E. coli strains were positive for the ecpA gene, the major pilin subunit of E. coli common pilus, in both groups. Furthermore, the papEF operon segment, iucD, hlyA, and cnf1 were also frequent in both groups (Table 6). However, the cytotoxic necrotizing factor-1 (cnf1 gene) and aerobactin siderophore (iucD gene) frequencies were significantly higher in E. coli strains from humans (p = 0.012 and p = 0.0002, respectively) (Table 6).

Table 6

Frequency of virulence genes among 3GC-resistant E. coli isolated from companion animals and humans with UTI.

Target Virulence Determinant	Target Gene	Companion Animal(N = 35) a	Human-CA(N = 85) a	p Value *
Pap fimbriae	papEF operon segment	45.7% (n = 16)	49.4% (n = 42)	0.841
Sfa fimbriae	sfa	20.0% (n = 7)	20.0% (n = 17)	N.s.
Afa afimbrial adhesin	afa	2.9% (n = 1)	9.4% (n = 8)	0.281
Alpha-hemolysin operon	hlyA	40.0% (n = 14)	42.4% (n = 36)	0.841
Cytotoxic necrotizing factor-1	cnf1	17.1% (n = 6)	41.2% (n = 35)	0.012
Aerobactin siderophore	iucD	48.6% (n = 17)	83.5% (n = 71)	0.0002
E. coli common pilus	ecpA	100% (n = 35)	100% (n = 85)	N.s.
Uropathogenic specific protein	usp	2.9% (n = 1)	0.0% (n = 0)	0.292

n, number of strains; a the number shown (N) is the total number of strains tested; Human-CA, human community-acquired UTI. * p value < 0.05 was considered statistically significant. N.s. not significant to calculate.

2.3. Clonal Lineages of ESBL/pAmpC-Producing E. coli

The companion animal strains were distributed in 15 sequence types (STs), while the human strains belonged to 19 STs (Table 7). The most common STs among ESBL/pAmpC-producing E. coli strains from companion animals with UTI were the ST648 (n = 11), ST131 (n = 5), ST539 (n = 2), and ST1775 (n = 2) (Table 7).

Table 7

Sequence types of ESBLs/pAmpC-producing E. coli isolated from companion animals and humans with UTI.

Phylogroup	Sequence Type	Clonal Complex	β-Lactamase (ESBL/pAmpC)	Species (n)
A	ST10	10	bla CTX-M-1	Human (2)
A	ST23	23	bla CTX-M-32	Cat (1)
A	ST88	23	bla CTX-M-1	Dog (1)
			bla CTX-M-15	Human (1)
			bla CTX-M-32	Human (1)
A	ST90	23	bla CTX-M-9like	Human (1)
			bla CTX-M-27	Human (1)
A	ST167	10	bla CTX-M-32	Human (1)
A	ST540	-	bla CTX-M-32	Human (1)
A	ST609	46	bla CTX-M-32	Dog (1)
A	ST617	10	bla CTX-M-1	Human (1)
A	ST5257	-	bla CTX-M-32	Human (1)
A	ST6023	-	bla CTX-M-14	Human (1)
A	Unassigned ST *	-	bla CTX-M-15	Dog (1)
B1	ST58	155	bla CTX-M-1	Human (2)
B1	ST224	-	bla CTX-M-32	Cat (1)
B1	ST453	86	bla SHV-12	Human (1)
			bla CTX-M-1	Human (1)
			bla CTX-M-15	Human (1)
			bla CTX-M-14	Human (3)
B1	ST533	-	blaCTX-M-15 + blaCMY-2	Dog (1)
B1	ST539		bla CTX-M-1like	Dog (1)
			bla CMY-2	Cat (1)
B1	ST847	-	bla CTX-M-14	Human (1)
B1	ST1196	-	bla CTX-M-1	Human (1)
B1	ST1725	-	bla CTX-M-15	Human (1)
B1	ND	-	bla CTX-M-2group	Human (1)
B2	ST131	131	bla CTX-M-1	Cat (1), human (2)
			bla CTX-M-15	Dog (2), cat (1), human (42)
			bla CTX-M-32	Human (1)
			bla CTX-M-9like	Human (3)
			bla CTX-M-14	Human (1)
			bla CTX-M-27	Human (5)
			bla CMY-2	Dog (1), human (3)
B2	ST372	-	bla CTX-M-15	Dog (1)
D	ST57	350	bla CMY-2	Dog (1)
D	ST117	-	bla CTX-M-15	Human (1)
D	ST354	354	bla CTX-M-14	Human (1)
			bla CMY-2	Dog (1)
D	ST405	405	bla CMY-2	Dog (1)
D	ST410	-	bla CTX-M-15	Human (1)
D	ST648	648	blaCTX-M-9 + blaCMY-2	Cat (1)
			bla CMY-2	Dog (3), cat (7), human (1)
D	ST778	38	bla CMY-2	Human (1)
D	ST1775	-	bla CMY-2	Dog (2)
D	ST3258	-	bla CMY-2	Cat (1)

ND, not done; -, not applicable; n, number of strains; * new ST allelic profile [4].

Regarding the E. coli ST648 clonal lineage from companion animals, the PAIIV536-PAIICFT073 (n = 5) combination was the most frequently detected. Different virulence gene profiles were found, but the most prevalent was ecpA-papEF (n = 7) (Table S1).

ST131 E. coli isolated from companion animals included two E. coli ST131 C2/H30Rx clades harboring the blaCTX-M-15 and one ST131C1/H30R1 (C1-non27) sub-clade harboring the blaCTX-M-1. The E. coli ST131 C1/H30R1 (C1-M27) clade harboring the blaCTX-M-27 was not detected in companion animals while the remaining ST131 E. coli strains remained unclassified by this assay (Table S1). For ST131 E. coli strains from companion animals, PAII536-PAIIIJ96-PAIICFT073-PAIIV536-PAIIICFT073 (n = 2) and PAIICFT073-PAIIV536-PAIIICFT073 (n = 2), were the combinations detected most frequently. However, different profiles of virulence genes were also found (Table S1). Among phylogroup-B1, E. coli from companion animals belonged to ST539 (n = 2), one harboring blaCMY-2 and one harboring blaCTX-M-1type; followed by ST533 (n = 1, harboring blaCTX-M-15 and blaCMY-2) and ST224 (n = 1, harboring blaCTX-M-32). Regarding phylogroup-A E. coli strains from companion animals, these belonged to an unassigned ST (n = 1, harboring blaCTX-M-15), ST609 (n = 1, harboring blaCTX-M-32), ST88 (n = 1, harboring blaCTX-M-1), and ST23 (n = 1, harboring blaCTX-M-32) (Table 7).

The most common clonal lineages among ESBL/pAmpC-producing E. coli strains from humans were ST131 (n = 57) and ST453 (n = 6) (Table 7). As in companion animals, E. coli strains isolated from humans and from different phylogenetic groups also harbored diverse ESBLs/pAmpC genes: group-A (n = 4, blaCTX-M-32; n = 3, blaCTX-M-1; n = 1, blaCTX-M-14; n = 1, blaCTX-M-27; n = 1, blaCTX-M-15; and n = 1, blaCTX-M-9like); group-B1 (n = 1, blaSHV-12; n = 4, blaCTX-M-14; n = 4, blaCTX-M-1; n = 2, blaCTX-M-15; and n = 1, blaCTX-M-2); group-B2 (n = 42, blaCTX-M-15; n = 5, blaCTX-M-27; n = 3, blaCTX-M-9like; n = 2, blaCTX-M-1; n = 1, blaCTX-M-14; n = 1, blaCTX-M-32; and n = 3, blaCMY-2); and group-D (n = 2, blaCTX-M-15; n = 1, blaCTX-M-14; and n = 2, blaCMY-2) (Table 7). Among group-B2, E. coli strains from humans belonged mainly to the ST131-C2/H30Rx clade harboring the blaCTX-M-15 gene (71.9%, n = 41/57). The ST131-C1-M27 E. coli sub-clade harboring the blaCTX-M-27 gene was detected in 8.8% (n = 5/57) of ST131 strains from humans. The clonal lineage O16-H5-ST131 (clade A) was also detected (1.8%, n = 1/57) and found to harbor blaCTX-9like ESBL. Finally, three strains isolated from humans belonging to the pandemic clone O25b:H4-B2-ST131 were found to be pAmpC-producers due to blaCMY-2 (Table S2). Regarding PAIs, the ST131 E. coli strains from humans were frequently positive for the following combinations: PAII536-PAIIIJ96-PAIII536-PAIICFT073-PAIIV536-PAIIICFT073 (n = 27) and PAIICFT073-PAIIV536-PAIIICFT073 (n = 23). The distribution of the different virulence genes showed that the most prevalent profiles in strains from humans belonging to group-B2 were ecpA-iucD (n = 17) and ecpA-papEF-sfaDE-hlyA-cnf1-iucD (n = 15) (Table S2).

Notably, ESBL/pAmpC-producing E. coli from companion animals and humans with UTI belonged to two MDR high-risk clonal lineages, namely the ST131 and ST648. Moreover, the ST88 and ST354 clonal lineages were also shared by companion animals and humans (Table 7, Tables S1 and S2).

3. Discussion

This study showed that ESBL/pAmpC-producing E. coli from companion animals and humans with UTI may harbor a big diversity of clinically relevant beta-lactamases and were associated with several virulence determinants.

E. coli strains isolated from companion animals were frequently associated with the presence of blaCTX-M-15 and blaCMY-2 genes, while those isolated from humans were associated with blaCTX-M-15 and blaCTX-M-1. The high prevalence and disseminations of blaCTX-M-15 in E. coli isolated from animals and humans agrees with studies conducted worldwide [10]. Moreover, CMY-2-producing E. coli strains belonged mainly to the phylogenetic group-D, which is consistent with a previous study conducted in the United States [20].

Despite the similarities between the E. coli strains isolated from both study groups with respect to mobile genetic determinants of antimicrobial resistance and virulence, the phylogenetic group-B2 and group-D was significantly more common in humans and companion animals, respectively. This finding may point to different E. coli host species adaptations contributing both to the global dissemination of overlapping antimicrobial resistance and virulence determinants.

The E. coli phylogenetic group-A, which is usually considered commensal and less pathogenic, was found to cause UTI in both companion animals and humans. Furthermore, this phylogroup was associated with a high diversity of globally disseminated blaCTX-M genes, such as blaCTX-M-15 and blaCTX-M-32. These findings highlight the high dissemination efficiency of plasmid-mediated beta-lactamases that may lead to therapeutic failure even in infections caused by less pathogenic strains.

Overall, 3GC-resistant E. coli strains belonging to the phylogenetic group-B2 had the higher number of PAI markers. This association of the group-B2 with several PAI markers is in line with previous reports of UPEC strains [21]. Interestingly, the most frequent PAI combination pattern was related to strains containing PAIIV536 and PAIICFT073. These PAI markers contain fimbrial adhesins and iron-uptake-system encoding genes that seem to be important for UPEC fitness and effective host colonization of the urinary tract. A high prevalence of fimbrial adhesin-encoding genes (such as, papEF operon segment) has been described in E. coli isolated from human patients diagnosed with UTI, thus highlighting the importance of these structures in the pathogenesis of UTI [22]. Furthermore, PAIICFT073 also carries the toxin hemolysin A (hlyA gene), that is responsible for the creation of pores in the host cell membranes leading to cell lysis [22]. Notably, PAIIV536 and PAIICFT073 were detected in 74.3% and 54.3% of E. coli strains from companion animals, respectively. Strains isolated from humans had an even higher prevalence of these PAIs (91.8% and 78.8%, respectively).

The uropathogenic-specific protein gene (usp) was detected only in one E. coli ST131 strain, which was isolated from a dog diagnosed with UTI in 2015. This protein is a genotoxin active against mammalian cells that can induce characteristics of apoptosis and has been associated with E. coli isolates from pyelonephritis, prostatitis, and bacteremia of urinary tract origin. It has been proposed that the usp gene provides immunity to its producer and enhances infectivity of the urinary tract [23,24].

Regarding the E. coli population structure, four sequence types were detected in companion animals and humans—the ST131 and ST648 MDR high-risk clonal lineages, and the ST88 and ST354 clonal lineages. The fact that the E. coli strains included in this study, from companion animals and humans, were collected in different years and from non-related patients is considered a study limitation, since it could have limited the detection of additional E. coli STs that are able to cause UTI in both groups. It should be noted that the results from this study may differ from studies including samples from a more recent timeframe as consequence of natural evolution and dissemination of beta-lactamases and E. coli clonal lineages. Nevertheless, the retrospective nature of this study is important as it contributes to the global understanding of the ecology of this common pathogen. Furthermore, this study includes data about virulence, which is still seldomly studied in strains from companion animals.

The ST131 clonal lineage harboring blaCTX-M-15 or blaCTX-M-14 has been detected in E. coli strains isolated from companion animals in many other countries [17,19,25]. However, the E. coli O25b:H4-ST131 harboring the blaCMY-2 gene has been rarely described [26,27]. In Japan, between 2005 and 2010, the blaCTX-M-14 gene was found to be the most common, followed by blaCTX-M-15 and blaCTX-M-2 [25]. In the present study, blaCTX-M-14 was also only detected in E. coli strains isolated from humans. The blaCTX-M-27 gene, a single-nucleotide variant of blaCTX-M-14, is being increasingly detected among E. coli strains isolated from humans and companion animals with UTI in the United States, Asia, and Europe [25,28,29,30]. In this study, the blaCTX-M-27 gene was only detected in one strain of human origin. To the best of our knowledge, this is the first description of the ST131 C1/H30R1 E. coli subclade C1-M27 and blaCMY-2-producing E. coli O25b:H4-ST131 in humans with community-acquired UTI from Portugal. Some reports have documented an increase in the number of the ST131 C1/H30R1 E. coli clade C1-M27 since the late 2000s. Isolates of this emerging clade have been reported in clinical samples from humans of Japan, France, Germany, Berlin, Geneva, Madrid, and Utrecht [25,31,32,33]; and also in companion animals, birds and urban seagulls [30,34,35]. Moreover, in northern Portugal, the C1-M27 clade has been isolated from fecal samples of healthy humans [36]. It is noteworthy that the ST131 C1/H30R1 E. coli clade C1-M27 has been shown to have a higher dissemination rate than the O25b:H4-ST131-H30Rx [37,38]. This dissemination rate may be accelerated by the expression of advantageous virulence determinants directed to the intestinal tract colonization, which is of particular importance to patients receiving antimicrobials or frequently admitted to hospital settings [33]. Thus, human colonization by clade C1-M27 should be monitored to improve preventive measures against infection and colonization of companion animals.

Previous studies about clinical E. coli strains showed that the O16-H5-ST131 clonal lineage (clade A) is globally distributed [39]. In this study, an E. coli O16:H5-ST131 harboring blaCTX-9-like was detected, which, to the best of our knowledge, is the first description in a human with community-acquired UTI from Portugal. The lower frequency of the O16-B2-ST131 in this study when compared to other countries [39,40,41], may reflect geographical and temporal differences in its distribution. Furthermore, according to Matsumura et al. [40], the O16-B2-ST131 clonal lineage has a lower prevalence of resistance to fluoroquinolones and ceftriaxone than O25b-B2-ST131 isolates [39,40,41]. This may also explain the low detection of O16-B2-ST131 in this study since it was focused on 3GC-resistant E. coli. Considering its high frequency in humans, the development of rapid molecular detection methods to identify ST131, while waiting for culture and antimicrobial susceptibility, could aid in a more effective initial antibiotic therapy and the reduction of its dissemination [39].

The E. coli ST648 high-risk clonal lineage is pandemic and globally reported in healthy and diseased humans and companion animals worldwide [13,20,42,43,44,45,46,47,48]. E. coli ST648 isolated from human infection may harbor several ESBL/pAmpC and carbapenemases, contributing to its dissemination [13,42,49]. As in ST131, the E. coli ST648 is frequently associated with CTX-M enzymes, namely CTX-M-15 [13]. The 3GC-resistant E. coli ST648 strains from this study, were all found to be CMY-2-producers. Furthermore, the CMY-2-producing ST648 clonal lineage was very common among isolates from companion animals, which is in line with previous studies [20,43]. Being a clonal lineage that is high-risk to humans, monitoring of ST648 in companion animals is of the upmost importance.

Interestingly, the E. coli ST88 and ST354 clonal lineages have been associated with poultry and broiler meat, suggesting that farm animals may be reservoirs of E. coli that are able to cause extraintestinal diseases in humans and companion animals [26,50,51,52,53,54]. E. coli ST354 has also been isolated from clinical samples of companion animals from Australia, and has been suggested to have a propensity to persist and circulate in animal-care facilities [55]. To the best of our knowledge, this is the first report of a CMY-2-producer E. coli ST354 in companion animals with UTI in Europe.

E. coli ST10 and ST410, other two important pandemic clonal lineages, were found in humans with UTI from this study [11,56]. E. coli ST410 seems to be a successful ExPEC clonal lineage like ST131 [57,58]. E. coli ST410 was first described, in 2016, in China [59] and since then has been reported worldwide in humans, companion animals, wildlife, and the environment [58,60,61,62,63]. Notably, in 2017, the ST410 was detected in companion animals with UTI from China [45]. Nevertheless, only a few studies have detected this clonal lineage in E. coli isolated from companion animals [45,62,63]. The E. coli ST10 clonal lineage has been reported in samples from animals (birds, swine, and sheep), in healthy human feces, and in humans with UTI. This clonal lineage is also usually associated with several 3GC-resistance genes such as blaCTX-M-14 and blaCTX-M-15 [64,65,66,67]. However, in the present study, the ST10 clone was associated with blaCTX−M−1. Although not detected in this study, ST410 has been previously described in isolates from companion animals, again, highlighting their possible role in the dissemination of high-risk clonal lineages.

The results from this study have high clinical relevance since it is shown that 3GC-resistant E. coli strains causing UTI in companion animals not only may belong to MDR high-risk clonal lineages, but are also likely to harbor critically important mobile genetic determinants associated with high pathogenicity or antimicrobial resistance.

4. Materials and Methods

4.1. Bacterial Isolates

Three hundred and thirty non-duplicate uropathogenic E. coli (UPEC) were isolated, between 1999 to 2015, from companion animals with UTI at the Laboratory of Antibiotic Resistance, the Faculty of Veterinary Medicine, University of Lisbon, Portugal. Furthermore, 85 non-duplicate 3GC-resistant E. coli isolates from humans with community-acquired UTI were obtained in 2013 from a diagnostic laboratory in the Lisbon area. 3GC-resistance was determined by antimicrobial susceptibility testing using cefotaxime or ceftazidime as surrogates, as described in Section 4.2.

Identification and confirmation of the isolate species was performed by a previously described PCR targeting the E. coli gadA gene [68].

4.2. Antimicrobial Susceptibility Testing

Antimicrobial susceptibility testing was carried by the disk-diffusion method according to Clinical Laboratory Standards Institute (CLSI) guidelines [69,70]. The antimicrobial agents tested were: amoxicillin 25 μg or ampicillin 10 μg, amoxicillin/clavulanate acid 30 μg, cefotaxime 30 μg, ceftazidime 30 μg, cefoxitin 30 μg, imipenem 10 μg, meropenem 10 μg, gentamicin 10 μg, tobramycin 10 μg, amikacin 30 μg, ciprofloxacin 5 μg, enrofloxacin 5 μg, norfloxacin 10 μg, nitrofurantoin 300 μg, and trimethoprim/sulfamethoxazole 25 μg. ESBL production was confirmed in all 3GC-resistant isolates by the double-disk synergy test and the results were interpreted according to the to CLSI guidelines. E. coli ATCC (American Type Culture Collection) 25,922 was used as a reference strain for antimicrobial susceptibility quality control.

Antimicrobial categories were used to characterize multidrug resistance as previously proposed by Magiorakos et al. [71].

4.3. Molecular Detection of Antimicrobial Resistance Genes

DNA extraction was conducted using a boiling method [72]. Antimicrobial resistance genes were investigated in resistant and intermediate resistant strains.

3GC-resistant E. coli were screened for blaCTX-M-type genes by PCR [73]. Positive isolates for blaCTX-M-type were further tested by PCR for blaCTX-M-group1, blaCTX-M-group2, and blaCTX-M-group9 [74] and positive amplicons were submitted to nucleotide sequencing. Cefoxitin-resistant E. coli isolates were tested using a multiplex-PCR with specific primers targeting plasmid-borne genes encoding AmpC β-lactamases (blaCIT, blaLAT, blaACT, blaMIR, blaFOX, blaMOX, and blaDHA), as previously described [75]. Positive samples for the group CIT were submitted to nucleotide sequencing after PCR amplification targeting the entire blaCMY-2 gene [76]. 3GC-resistant E. coli negative for blaCTX-M-type or AmpC genes were tested for the presence of blaTEM-type and blaSHV-type ESBL genes [72].

Strains were screened by PCR for the presence of common carbapenemase genes (blaIMP, blaOXA, blaVIM, blaNDM, and blaKPC), as previously described [77].

4.4. Uropathogenic Escherichia coli Phylogenetic Typing, Pathogenicity Island Markers, and Virulence Genotyping

Phylogenetic typing was performed in all 3GC-resistant E. coli strains to determine the main phylogenetic groups (A, B1, B2, and D) according to the amplification of chuA gene, yjaA gene, and TspE4C2 fragment [78].

Eight UPEC PAIs were screened by multiplex PCR assays, as previously described [28,29]. PCR reactions were split in three separate multiplex assays: multiplex A for PAIIII536, PAIIV536, and PAIIICFT073; multiplex B1 for PAIIIJ96 and PAII536; and multiplex B2 for PAIII536, PAIIJ96, and PAIICFT073. Negative and positive controls (E. coli CFT073, E. coli 536, and E. coli J96) were used in all PCRs [21,79].

3GC-resistant E. coli strains were screened by PCR for the presence of several virulence determinants: mediate adhesion (p-fimbrial adhesion genes papEF operon segment), Sfa fimbrial and Afa afimbrial adhesins (sfa and afa genes, respectively), toxin (α-hemolysin hlyA gene from the alpha-hemolysin operon), cytotoxic necrotizing factor 1 (cnf-1 gene), aerobactin siderophore (iucD gene) [80], the major pilin subunit of E. coli common pilus (ecpA gene), and the bacteriocin-like genotoxin uropathogenic specific protein (usp gene) [9]. Negative and positive controls (E. coli CFT073, E. coli 536, E. coli J96, and E. coli KS52) were used in all PCRs.

4.5. ESBL/pAmpC-Producing Escherichia coli Multi-Locus Sequence Typing

The ST131 clonal lineage, O16/O25b types, and the ST131-H30Rx clade were identified, as previously described by PCR [39,81,82].

E. coli strains not belonging to the ST131 clonal lineage were typed by MLST. Briefly, the seven housekeeping genes of the E. coli MLST scheme (adk, fumC, gyrB, icd, mdh, purA, and recA) were amplified by PCR using the primers and amplification conditions previously described in https://enterobase.warwick.ac.uk/ (accessed on 10 February 2020) [83]. PCR products were purified using the NZYTech Gel Pure Kit (NZYTech-Genes and Enzymes, Lisbon, Portugal) and sequencing was performed by Stabvida (Caparica, Portugal). Sequence quality was confirmed using Ugene Unipro software (Unipro, Novosibirsk, Russia) and the respective alleles and sequence types were retrieved using the publicly available E. coli MLST database.

4.6. Statistical Analysis

The SAS statistical software package for Windows v. 9.4 (SAS Institute Inc., Cary, NC, USA) was used for statistical analysis. The Fisher’s exact test was used for comparisons between groups (two by two analysis of contingency tables) with a p value of 0.05.

5. Conclusions

This study showed that 3GC-resistant E. coli from companion animals and humans with community-acquired UTI frequently belong to important pandemic high-risk clonal lineages and harbor clinically relevant antimicrobial resistance and virulence determinants that are easily disseminated. Considering the close contact between companion animals and humans in modern society, the dissemination of pandemic E. coli clones, such as ST131-C2/H30Rx (blaCTXM-15) and ST648, in patients with UTI requires the joint action of human and veterinary medicine. Although the degree of inter-species transmission and zoonotic/zooanthroponotic potential of such bacteria is complex to evaluate, the high frequency of common PAI markers and CTX-M enzymes reported in this study, highlights that the link between humans and companion animals goes beyond sharing specific E. coli clones. As appraised by the scientific community, a One Health approach is required to fully grasp the dissemination dynamics of such bacterial clones and/or their antimicrobial resistance and virulence mobile genetic determinants.