Antibiotic-resistant Escherichia coli isolated from urban rodents in Hanoi, Vietnam.

Antimicrobial resistance (AMR) is a global public health concern for both clinical and veterinary medicine. Rodent feces are one of the major infectious sources of zoonotic pathogens including AMR bacteria. So far, there are limited studies reported focused on Escherichia coli isolated in rodent feces from rural and suburban areas in Vietnam. In this study, we investigated the prevalence of antimicrobial resistance in E. coli isolated from feces samples of 144 urban rodents caught in Hanoi, Vietnam. A total of 59 AMR E. coli was isolated from urban rodents of which 42 were multidrug-resistant (MDR) isolates (resistance to at least three classes of antimicrobial agents), four were extended-spectrum β-lactamase (ESBL) producing isolates and five were colistin-resistant isolates. The highest prevalence of the resistance was against ampicillin (79.7%: 47/59), followed by tetracycline (78.0%: 46/59), nalidixic acid (67.8%: 40/59), sulfamethoxazole-trimethoprim (59.3%: 35/59), chloramphenicol (45.8%: 27/59), ciprofloxacin (44.1%: 26/59), cefotaxime (30.5%: 18/59), cefodizime (23.7%: 14/59), amoxicillin-clavulanate (22.0%: 13/59), and gentamicin (22.0%: 13/59). With regard to the virulence genes associated with diarrheagenic E. coli (DEC), only aaiC gene found in one AMR isolate. In general, the use of antimicrobials does not aim to treat rodents except for companion animals. However, our findings show the carriage of AMR and MDR E. coli in urban rodents and highlight the potential risk of rodents in Hanoi acting as a reservoir of transferable MDR E. coli, including ESBL-producing, colistin-resistant E. coli, and virulence-associated with DEC.

Rodents are known as reservoir animals for the transmission of a variety of zoonotic pathogens in urban locations [14, 22]. Some bacteria, viruses and parasites have been carried by rodents, namely Escherichia coli, Salmonella spp., Campylobacter spp., Leptospira spp., Orientia tsutsugamushi and hantavirus [18, 26, 28, 41]. Rodent-borne diseases spread to humans through the consumption of contaminated food, the inhalation of the aerosol, rodent bites and arthropod vectors from rodents [33]. Feces are one of the significant sources for pathogens including E. coli, with the potential risk for causing intestinal diseases in humans and animals [20]. In comparison to livestock, a few studies have shown a profile of AMR E. coli carried by rodents in both urban and rural areas [7, 18, 23, 40]. Most of the previous studies focused only on the resistant phenotype of AMR E. coli isolated from rodents [7, 17, 41]. Recently, the research has shifted to AMR genes in combination with the resistant phenotype [18]. Although a discrepancy sometimes occurs in the results between the phenotype and genotype, it is able to conduct broader analysis predictions.

In the zoonotic perspective, AMR E. coli and diarrheagenic E. coli are of significant concerns. The recent reports studied on the potential zoonotic E. coli were indicated the high prevalence of AMR E. coli including ESBL-producing E. coli and virulence genes carried E. coli from urban rodents [6, 18, 19, 24]. In addition, colistin-resistant bacteria have gained much more attention since the first mobile colistin resistance gene (mcr-1) was identified in E. coli in China 2015 [31]. Colistin-resistant bacteria have been reported from humans, livestock and environments in various locations in the world [9, 50]. AMR genes including the colistin resistance genes carried by E. coli can be transferred to other bacterial populations, rising a potential risk of the expansion to humans and the environment. Rodents inhabiting close to humans and livestock may involve in the transmission of AMR genes including the colistin resistance genes. There was a study which investigated the prevalence of AMR E. coli including ESBL-producing E. coli isolated from small mammals in the Mekong Delta, Vietnam [40]. The prevalence of MDR was approximately eight times higher in E. coli isolated from small mammals (rodents and shrews) trapped on farms than those trapped in forests or rice fields. The results strongly suggested that MDR in small mammals were obtained from MDR-carried livestock and their environment.

Since rodents have the ability of the adaption to their habitat particularly with regards to a density of human population, rodents are in contact with human waste and infrastructure (e.g., garbage and sewage) in the urban areas. In this study, we investigated the prevalence of putative pathogenic E. coli in rodents trapped in Hanoi, Vietnam and then, examined the possible association and distribution of antimicrobial resistant profiles, antibiotic resistance genes, ESBL-producing, virulence genes and colistin-resistance in E. coli isolates from urban rodents.

MATERIALS AND METHODS

Rodent trapping and sample collection

Rodents were captured using live traps at eight locations of Hanoi, Vietnam in October 2017, March and June 2018 (Table 1). The identification of rodent species was conducted by DNA sequencing of the mitochondrial cytochrome b gene [60]. Rodents were euthanized by isoflurane inhalation, as recommended by the American Veterinary Medical Association (AVMA) guidelines. Rectal swabs and/or rectal feces were collected and frozen using dry ice and then stored at −80°C until use [11].

Table 1.

No. of antimicrobial-resistant Escherichia coli isolated from urban rodents in Hanoi, Vietnam

Location		GTVT hospital			Ha Dong hospital		Dong Tam market	Thanh Cong market		Ha Dong market		Phung Khoang market	Thai Ha market		Giap Bat cargo station
Latitude, Longitude		21°1’33.50”N,			20°58’17.16”N,		20°59’48.52”N,	21°1’21.51”N,		20°58’11.61”N,		20°59’11.32”N,	21°0’49.40”N,		20°58’48.77”N,
		105°48’11.17”E			105°46’30.66”E		105°50’42.56”E	105°48’54.17”E		105°46’46.86”E		105°47’37.87”E	105°49’20.76”E		105°50’29.24”E
Rat species		Rna)	Rab)	Rrc)	Rna)	Rrc)	Rna)	Rna)	Rrc)	Rna)	Rrc)	Rna)	Rna)	Rrc)	Rna)
No. of resistant isolates / no. of samples (%)	Oct. 2017	0/9	0/1	0/1	-	-	3/12	1/17	-	-	-	-	-	-	3/16
	Mar. 2018	-	-	-	12/13	0/2	-	6/15	-	3/12	1/2	-	-	-	6/11
	Jun. 2018	2/5	-	-	-	-	-	8/9	0/1	2/2	-	10/10	2/4	0/2	-
	Subtotal	2/14	0/1	0/1	12/13	0/2	3/12	15/41	0/1	5/14	1/2	10/10	2/4	0/2	9/27
	Location total	14/31 (45.2)					36/86 (41.9)								9/27 (33.3)
	Total	59/144 (41)

a) Rattus norvegicus, b) Rattus argentiventer, c) Rattus rattus.

E. coli isolation

Rectal swabs and/or feces were soaked in 1 ml of Luria-Bertani (LB) broth (Dickinson and Co., Franklin Lakes, NJ, USA). The samples were plated using full loops (10 mm inoculate loop) onto Deoxycholate Hydrogen Sulfide Lactose (DHL) Agar (Eiken Chemical Co., Ltd., Tokyo, Japan), and incubated at 37°C overnight. One colony showing typical E. coli morphology from each rat sample was identified by the following biochemical tests: Triple Sugar Iron (TSI) agar (Becton, Dickinson and Co.), Lysine Indole Motility (LIM) test (Eiken Chemical Co., Ltd.), Voges-Proskauer (VP) test (Eiken Chemical Co., Ltd.), citrate utilization tests, and oxidase test (Merck KGaA, Darmstadt, Germany) [55]. The confirmation of E. coli was conducted by detecting the yaiO gene by PCR [35].

Antimicrobial susceptibility test

Antimicrobial susceptibility tests were conducted on Mueller Hinton II agar (Becton, Dickinson and Co.) plates according to the Kirby-Bauer disc diffusion method using disc and titer details as follows: Ampicillin (ABP, 10 µg), Cefodizime (CDZ, 30 µg), Gentamicin (GM, 10 µg), Tetracycline (TC, 30 µg), Ciprofloxacin (CIP, 5 µg), Cefotaxime (CTX, 30 µg), Amoxicillin-Clavulanate (ACV, 20 µg and 10 µg, respectively), Nalidixic acid (NA, 30 µg), Chloramphenicol (CP, 30 µg), Sulfamethoxazole −Trimethoprim (ST, 1.25 µg and 23.75 µg, respectively) (Eiken, Chemical Co., Ltd). Potential production of extended-spectrum β-lactamase (ESBL) was confirmed by the double-disk synergy test using CTX, ACV and Ceftazidime (CAZ, 30 µg). The resistance phenotype was interpreted accurately according to the Clinical and Laboratory Standards Institute (CLSI) guidelines 2011 and the manufacturer guideline [12]. In this study, E. coli JCM 384 was used as a quality control strain.

Detection of antimicrobial resistance genes

Colonies of isolated E. coli were dissolved in 100 µl of distilled water and boiled at 95°C for 15 min. The major resistant genes for β-lactamase (blaTEM, blaSHV, blaCTX-M and blaCMY-2), sulfonamides (sul1, sul2 and sul3), quinolone (qnrA) and tetracycline [tet(A), tet(B) and tet(C)] were tested by single or multiplex PCR. To classify blaCTX-M genes into the five major groups, CTX-M-1, CTX-M-2 (TOHO), CTX-M-8/25 and CTX-M-9 groups, the detection of blaCTX-M genes were used four primer sets designed by Pitout et al. to detect [43]. PCR conditions and primers (Life Technologies Japan Ltd.) for other resistance genes described by Kozak et al. and Mammeri et al. [29, 32] were used in this study.

Identification of diarrheagenic E. coli

Primers of multiplex PCR used in this study were described in Table 2. The target genes associated with diarrheagenic E. coli (DEC) include elt (heat labile enterotoxin), est (heat stable enterotoxin), bfpA (bundle-forming pilus), eae (E. coli attaching and effacing), stx (Shiga toxin), ipaH (invasion plasmid antigen H), aatA (outer membrane protein), aaiC (secreted protein) and aggR (transcriptional activator).

Table 2.

List of primers for the identification of diarrheagenic Escherichia coli

Pathogena)	Primer name	Target gene	Primer sequence (5′-3′)	Amplicon (bp)	Reference
ETEC	LT-F	elt	CACACGGAGCTCCTCAGTC	508	Panchalingam et al. 2012
	LT-R		CCCCCAGCCTAGCTTAGTTT
	ST-F	est	GCTAAACCAGTAG/AGGTCTTCAAAA	147	Panchalingam et al. 2012
	ST-R		CCCGGTACAG/AGCAGGATTACAACA
EPEC	BFPA-F	bfpA	GGAAGTCAAATTCATGGGGG	367	Panchalingam et al. 2012
	BFPA-R		GGAATCAGACGCAGACTGGT
	SK1	eae	CCCGAATTCGGCACAAGCATAAGC	881	Toma et al. 2003
	SK2		CCCGGATCCGTCTCGCCAGTATTCG
STEC	VTcom-u	stx	GAGCGAAATAATTTATATGTG	518	Toma et al. 2003
	VTcom-d		TGATGATGGCAATTCAGTAT
	SK1	eae	CCCGAATTCGGCACAAGCATAAGC	881	Toma et al. 2003
	SK2		CCCGGATCCGTCTCGCCAGTATTCG
EIEC	ipaIII	ipaH	GTTCCTTGACCGCCTTTCCGATACCGTC	619	Toma et al. 2003
	ipaIV		GCCGGTCAGCCACCCTCTGAGAGTAC
	CVD432F	aatA	CTGGCGAAAGACTGTATCAT	630	Panchalingam et al. 2012
	CVD432R		CAATGTATAGAAATCCGCTGTT
EAEC	AAIC F	aaiC	ATTGTCCTCAGGCATTTCAC	215	Panchalingam et al. 2012
	AAIC R		ACGACACCCCTGATAAACAA
	aggRks1	aggR	GTATACACAAAAGAAGGAAGC	254	Toma et al. 2003
	aggRks2		ACAGAATCGTCAGCATCAGC

a) ETEC: Enterotoxigenic E. coli, EPEC: Enteropathogenic E. coli, STEC: Shiga toxin-producing E. coli, EIEC: Enteroinvasive E. coli; EAEC: Enteroaggregative E. coli.

The PCR mixture was adjusted to 25 µl containing with of 12.5 µl of Gotaq (Promega Corp., Madison, WI, USA), 0.3 µM of each primer, distilled water and 2 µl of DNA template. Multiplex PCR was performed as follows; 95°C for 5 min, 30 cycles of 95°C for 20 sec, 52°C for 40 sec, 72°C for 30 sec and 72°C for 7 min.

Detection of colistin-resistant E. coli

Detection of colistin resistance genes was performed by PCR using the primers for mcr-1, mcr-2 and mcr-3 genes as described in the previous studies [31, 58, 59]. Amplified fragments were confirmed by DNA sequencing. The sequences were deposited in the GenBank (accession numbers MN519790–MN519794). Susceptibility to colistin was examined by the broth microdilution and macro dilution methods. The determination of minimal inhibitory concentration (MIC) was >2 mg/l as resistant according to the breakpoint for Enterobacteriaceae in the European Committee on Antimicrobial Susceptibility Testing (EUCAST).

RESULTS

Prevalence of resistant E. coli isolates from urban rodents

A total of 144 rodents (ricefield rat (Rattus argentiventer); 1, brown rat (R. norvegicus); 135, black rat (R. rattus); 8) were captured at eight locations in urban areas of Hanoi in 2017 and 2018 (Table 1). Fifty-nine isolates (58 isolates from R. norvegicus and one isolate from R. rattus) were identified to AMR E. coli. AMR E. coli were obtained from 14 out of 31 (45.2%) from hospitals, 36 of 86 (41.9%) from markets and 9 of 27 (33.3%) from a cargo station. There was no statistical difference in the prevalence of AMR E. coli between hospitals, markets and a cargo station. The most common AMR was resistance to ABP, which was detected in 79.7% (47/59), followed by TC (78.0%: 46/59), NA (67.8%: 40/59), ST (59.3%: 35/59), CP (45.8%: 27/59), CIP (44.1%: 26/59) and CTX (30.5%: 18/59) (Table 3). Multidrug resistance (MDR; resistance to three or more antimicrobial classes) were identified in 42 isolates out of 59 (71.2%) from R. norvegicus.

Table 3.

Prevalence of antimicrobial-resistant Escherichia coli isolated from urban rodents in Hanoi, Vietnam

Antimicrobial agentsa)	Antimicrobial classes	No. of resistant isolates (%)
		GTVT hospital	Ha Dong hospital	Dong Tam market	Thanh Cong market	Ha Dong market	Phung Khoang market	Thai Ha market	Giap Bat cargo station	Subtotal	Total
		n=2	n=12	n=3	n=15	n=6	n=10	n=2	n=9		n=59
ABP	Beta-lactams	2 (100)	11 (91.7)	2 (66.7)	12 (80)	5b) (83.3)	8 (80)	1 (50)	6 (66.7)	47 (79.7)	50 (84.7)
ACV		0	2 (16.7)	0	4 (26.7)	2 (33.3)	4 (40)	0	1 (11.1)	13 (22.0)
CDZ		0	3 (25)	0	2 (13.3)	1 (16.7)	7 (70)	0	1 (11.1)	14 (23.7)
CTX		0	4 (33.3)	1 (33.3)	1 (6.7)	2 (33.3)	7 (70)	0	3 (33.3)	18 (30.5)
CIP	Quinolone	1 (50)	7 (58.3)	0	7 (46.7)	3 (50)	7 (70)	0	1 (11.1)	26 (44.1)	43 (72.9)
NA		0	10 (83.3)	1 (33.3)	11 (73.3)	4 (66.7)	9 (90)	1 (50)	4 (44.4)	40 (67.8)
CP	Chloramphenicol	1 (50)	7 (58.3)	1 (33.3)	6 (0.4)	3 (50)	7 (70)	0	2 (22.2)	-	27 (45.8)
GM	Aminoglycoside	0	5 (41.7)	0	5 (33.3)	1 (16.7)	2 (20)	0	0 (0)	-	13 (22.0)
ST	Sulfonamide	2 (100)	9 (75)	1 (33.3)	9 (60)	3 (50)	9 (90)	0	2 (22.2)	-	35 (59.3)
TC	Tetracycline	2 (100)	9 (75)	2 (66.7)	14 (93.3)	5b) (83.3)	9 (90)	0	5 (55.6)	-	46 (78.0)
Multi-drug resistant		2 (100)	10 (83.3)	1 (33.3)	12 (80)	4 (66.7)	9 (90)	0	4 (44.4)	-	42 (71.2)

a) ABP: Ampicillin, ACV: Amoxicillin−Clavulanate, CDZ: Cefodizime, CTX: Cefotaxime, CIP: Ciprofloxacin, NA: Nalidixic acid, CP: Chloramphenicol, GM: Gentamicin, ST: Sulfamethoxazole−Trimethoprim, TC: Tetracycline. b) Including an isolate from Rattus rattus.

Detection of AMR genes

The most frequent resistance gene was blaTEM (69.5%: 41/59) followed by tet(A) (64.4%: 38/59), sul2 (35.6%: 21/59), sul3 (23.7%: 14/59), sul1 (18.6%: 11/59), and tet(B) (11.9%: 7/59) (Table 4). On the other hand, genes of blaSHV, qnrA and tet(C) were not detected in any isolates. Nine of the 50 isolates phenotypically resistant to β-lactams (ampicillin, amoxicillin-clavulanate, cefodizime, and/or cefotaxime) had none of β-lactamase genes (Tables 3 and 4). In isolates showing quinolone-resistant phenotypes (26 and 40 isolates resistant to ciprofloxacin and nalidixic acid, respectively), no quinolone-resistance gene (qnrA) was found. Four out of the 35 isolates phenotypically resistant to sulfamethoxazole-trimethoprim had no genes of sulfonamide resistance, and vice versa. Two out of the 46 isolates phenotypically resistant to tetracycline possessed no tetracycline resistance genes.

Table 4.

No. of antimicrobial resistance genes detected in 59 antimicrobial-resistant Escherichia coli isolated from urban rodents in Hanoi, Vietnam

AMR genes	Antimicrobial classes	No. of isolates positive for AMR genes (%)
		GTVT hospital	Ha Dong hospital	Dong Tam market	Thanh Cong market	Ha Dong market	Phung Khoang market	Thai Ha market	Giap Bat cargo station	Subtotal	Total
		n=2	n=12	n=3	n=15	n=6	n=10	n=2	n=9		n=59
blaTEM	Beta-lactams	2 (100)	10 (83.3)	1 (33.3)	12 (80)	4a) (66.7)	8 (80)	0	4 (44.4)	41 (69.5)	41 (69.5)
blaCTX-M-1 group		0	2 (16.7)	0	0	1 (16.7)	0	0	0	3 (5.1)
blaCMY-2		0	0	0	1 (6.7)	0	0	0	0	1 (1.7)
blaSHV		0	0	0	0	0	0	0	0	0 (0)
qnrA	Quinolone	0	0	0	0	0	0	0	0	-	0 (0)
sul1	Sulfonamide	0	3 (25)	0	7 (46.7)	0	1 (10)	0	0	11 (18.6)	35 (59.3)
sul2		2 (100)	5 (41.7)	1 (33.3)	7 (46.7)	3 (50)	2 (20)	0	1 (11.1)	21 (35.6)
sul3		1 (50)	5 (41.7)	1 (33.3)	2 (13.3)	1 (16.7)	2 (20)	0	2 (22.2)	14 (23.7)
tet (A)	Tetracycline	2 (100)	8 (66.7)	2 (66.7)	9 (60)	4a) (66.7)	9 (90)	0	4 (44.4)	38 (64.4)	44 (74.6)
tet (B)		0	1 (8.3)	0	5 (33.3)	1 (16.7)	0	0	0	7 (11.9)
tet (C)		0	0	0	0	0	0	0	0	0

a) Including an isolate from Rattus rattus.

Detection of ESBL-producing E. coli

According to the double-disk synergy test using CTX, ACV and CAZ, four out of the 59 isolates were confirmed as ESBL-producing E. coli (Table 5). Three out of the four ESBL-producing E. coli were isolated from Ha Dong hospital and one from Ha Dong market. Furthermore, all four ESBL-producing isolates were identified as MDR and carried the blaTEM gene, and three out of these isolates harbored blaCTX-M-1 group.

Table 5.

Characteristics of extended-spectrum β-lactamase-producing Escherichia coli isolated from urban rodents in Hanoi, Vietnam

ID	Location	Antimicrobial resistant phenotypea)	β-lactamase gene
100	Ha Dong hospital	ABP-CDZ-CTX-CIP-NA-CP-GM-TC-ST	blaTEM, blaCTX-M-1 group
101		ABP-CDZ-CTX-CIP-NA-CP-ST	blaTEM, blaCTX-M-1 group
105		ABP-CDZ-CTX-CIP-NA-TC	blaTEM
133	Ha Dong market	ABP-CDZ-CTX-CIP-NA-CP-GM-TC-ST	blaTEM, blaCTX-M-1 group

a) ABP: Ampicillin, ACV: Amoxicillin−Clavulanate, CDZ: Cefodizime, CTX: Cefotaxime, CIP: Ciprofloxacin, NA: Nalidixic acid, CP: Chloramphenicol, GM: Gentamicin, ST: Sulfamethoxazole, TC: Tetracycline.

Detection of diarrheagenic genes

Isolates were analyzed for the presence of virulence genes associated with enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), Shiga toxin-producing E. coli (STEC), enteroinvasive E. coli (EIEC) and enteroaggregative E. coli (EAEC) by multiplex PCR (Table 2). There was no detection of ETEC, EPEC, STEC and EIEC in this study. Only one out of the 59 AMR isolates carried aaiC gene (data not shown), indicating the potential pathogen of EAEC. The isolate carried aaiC gene was MDR (ABP-NA-ST-TC) and carried AMR genes of blaTEM, sul2 and tet(A).

The mcr-1 genes were amplified from five isolates, 250 bp from IDs 71 and 109 and 259 bp from IDs 102, 120 and 137 (accession nos. MN519790– MN519794). These sequences were identical with the mcr-1 gene from E. coli strain SHP45 (accession no. KP347127) isolated from swine in China, 2015, which was the first report of colistin-resistant E. coli. All of the isolates carrying mcr-1 gene exhibited MIC of 4 µg/ml and showed 5 to 6 antimicrobial classes in the resistant phenotypes. The isolate ID 71 from Thanh Cong market showed resistance to all 10 antimicrobial agents tested in this study.

DISCUSSION

The present study demonstrated that 59 out of 144 (41%) rodents carried AMR E. coli in the urban city of Hanoi, Vietnam (Tables 1 and 2). The prevalence of AMR E. coli was no significant difference between hospitals, markets and the cargo station in Hanoi (Table 2). Among 59 AMR E. coli, the most common resistant phenotype was observed to β-lactams (84.7%), followed by tetracycline (78%), quinolone (72.9%), and sulfonamide (59.3%). The resistant to ampicillin was the most prevalent in Hanoi (79.7%) which was similar to the previous study in the Mekong Delta (85.9%) [40]. The resistance to tetracycline was much higher in Hanoi (78%) than in Mekong delta (34.5%), indicating environmental contamination with tetracycline in the urban area [54]. MDR was identified in 42 out of 59 AMR isolates (71.2%). The prevalence of MDR isolates from small mammals in Hanoi was higher than those observed in Vancouver, Canada (41.5%), Berlin, Germany (58.2%), Nairobi, Kenya (66.7%), or Mekong Delta, Vietnam (27.2%) [17, 18, 23, 40]. Since rodents in urban areas have opportunities to contact with human sewage systems and garbage dumps, they might take up with MDR E. coli including colistin-resistant via these routes. In recent years, colistin-resistant bacteria have been observed in wildlife and investigated to understand the mechanism of spreading colistin resistance globally. Although colistin-resistant E. coli were found in some wildlife, Algerian hedgehog and Barbary macaques, birds and rodents were concerned as potential carriers and/ or spreaders of colistin-resistant E. coli [1, 2, 4, 13, 34, 45, 47]. Our data revealed that urban rodents could be a carrier and spreader of colistin-resistant E. coli (Table 6). The prevalence of colistin-resistant E. coli in urban rodents (5/59, 8.5%) was much lower than those in healthy persons (80.6%) in Vietnam [59]. Besides colistin-resistant E. coli, ESBL-producing E. coli were isolated in urban rodents (Table 5). Prevalence of ESBL-producing E. coli in urban rodents (4/59, 6.8%) was also lower than livestock (20%) and healthy persons (35.2%) in Vietnam [39]. Since the role of wildlife has been documented in spreading AMR bacteria, the urban rodents may play as the maintenance host or vector for ESBL-producing and colistin-resistant isolates [53, 57].

Table 6.

Characteristics of colistin-resistant Escherichia coli isolated from urban rodents in Hanoi, Vietnam

ID	Location	Colistin resistance gene	MICa) (µg/ml)	Other resistant phenotypeb)
71	Thanh Cong market	mcr-1	4	ABP-ACV-CDZ-CTX-CIP-NA-CP-GM-ST-TC
102	Ha Dong hospital	mcr-1	4	ABP-CIP-NA-CP-GM-ST-TC
109	Ha Dong hospital	mcr-1	4	ABP-CTX-NA-CP-GM-ST-TC
120	Thanh Cong market	mcr-1	4	ABP-CIP-NA-CP-ST-TC
137	Phung Khoang market	mcr-1	4	ABP-CIP-NA-CP-ST-TC

a) Minimal Inhibitory Concentration (MIC). b) ABP: Ampicillin, ACV: Amoxicillin−Clavulanate, CDZ: Cefodizime, CTX: Cefotaxime, CIP: Ciprofloxacin, NA: Nalidixic acid, CP: Chloramphenicol, GM: Gentamicin, ST: Sulfamethoxazole, TC: Tetracycline.

Regarding blaCTX-M genes, there are several reports of the geographically major groups. The blaCTX-M-1 group is more common in E. coli isolated from human and livestock samples in Europe [8, 15]. The samples from rodent feces were also detected the blaCTX-M-1 group in Germany [15]. In Southeast Asian countries such as Thailand, Laos and Vietnam, the blaCTX-M-1 and blaCTXM-9 groups are the most predominant groups carried by E. coli isolated from humans [37]. On the other hand, the blaCTX-M-2 group is present in E. coli isolated from humans and broiler chickens in South American countries [16, 44, 49], cattle in Japan [48] and humans in Israel [10]. The blaCTX-M-8 group was found in E. coli isolated from chicken meats in Brazil [36], humans in the United States [30] and gulls in Spain [52]. The blaCTX-M-25 group was reported from clinical samples in Israel [56] and broiler chickens in Japan [61]. In this study, we found the blaCTX-M-1 group in three ESBL-producing E. coli isolated from rodent feces. This is the first report of the blaCTX-M-1 group among rodents in Asian countries.

There were reports of the discrepancies between AMR phenotypes and the presence or absence of resistance genes [3, 29]. In this study, there were almost good agreements between AMR phenotypes and the presence or absence of resistance genes in regard to the resistance to β-lactams, sulfamethoxazole-trimethoprim and tetracycline resistance. Isolates resistant to β-lactams but without β-lactamase genes (blaTEM, blaSHV, blaCTX-M, blaCMY-2) may be caused by other genes because β-lactamase genes were assigned to nine distinct structures including TEM, SHV, CTX-M, PER, VEB, GES, BES, TLA and OXA based on the comparison of their amino acid sequences [5]. AMR genes are encoded in plasmids which are responsible for the spread of resistance genes [27, 51]. Possible of horizontal transmission of β-lactamase gene (blaCTX-M-1 group, blaCMY-2) and colistin resistance gene (mcr-1) poses a risk to public health, food safety, and water source. However, there was no quinolone resistance gene (qnrA) in quinolone-resistant phenotypes, 26 and 40 isolates resistant to ciprofloxacin and nalidixic acid, respectively. These results may reflect the fact that resistance to these antimicrobials can be acquired by different AMR genes or another resistance mechanism such as mutations in the quinolone resistance-determining regions [18] or in efflux pumps [25].

With regard to virulence genes associated with DEC, the aaiC gene was detected from one MDR isolate. The aaiC gene is a virulence gene typically associated with enteroaggregative Escherichia coli (EAEC) [42]. The EAEC group have been found in diarrhea patients in several regions of the world [21]. While one of 59 isolates (1.7%) carried a virulence gene in this study, the prevalence of virulence genes in urban rodents were 0% and 3.8% in Berlin, Germany and Vancouver, Canada, respectively [18, 23]. In contrast, isolates from livestock showed higher carriage of virulence genes, 31.3% and 10.9% from calves and chicken in Vietnam, respectively [38, 46]. The low prevalence suggests a low risk of pathogenic E. coli infection from urban rodents, but continuous monitoring should be implemented for the risk assessment.

In conclusion, our data suggest that urban rodents could be a reservoir and spreader of AMR E. coli including multidrug-resistant, ESBL-producing and colistin-resistant isolates in Hanoi, Vietnam. The prevalence of AMR E. coli in rodents from urban areas (41%) was almost same with that in rodents from farms (45%) and much higher than forest and rice fields (5.8%) in Vietnam [40]. The risk of urban rodents carrying AMR E. coli is still unclear. Applying antibiotic usage strategies and improving hygiene is necessary for these locations. Further studies may help understand the interaction of AMR bacteria and resistance genes in all environments.