Antimicrobial resistance and genetic diversity of Escherichia coli isolated from humans and foods

Abstract

Antibiotic resistance has increased in recent years, raising the concern of public health authorities. We conducted a study of Escherichia coli isolates obtained from human and food samples to assess the prevalence of antimicrobial resistance and to determine the genotype and clonal relationship of 84 E. coli isolates (48 from humans and 36 from foods). An antimicrobial susceptibility test was performed using the disk diffusion method. Virulence factors were evaluated by multiplex PCR, and the clonal relationship among the resistant isolates was studied by Pulsed Field Gel Electrophoresis (PFGE). All isolates were susceptible to ceftriaxone. Overall, 26%, 20.2%, 15.4% and 6% of the isolates were resistant to tetracycline, ampicillin, sulfamethoxazole/trimethoprim and cephalotin, respectively. Twenty two percent of the isolates exhibited resistance to more than one antimicrobial agent. Multiple-drug resistance was mostly observed in the human isolates and involved the antibiotics ampicillin and tetracycline. None of the six virulence genes were identified among the isolates. Analysis of genetic diversity by PFGE of 31 resistant isolates, revealed 29 distinct restriction patterns. In conclusion, E. coli from humans and foods are resistant to commonly used antibiotics and are highly genetically diverse. In this setting, inappropriate use of antibiotics may be a cause of high resistance rate instead of clonal spread.

Introduction

Escherichia coli are a common inhabitant of the intestinal tract of humans and animals and the most common cause of nosocomial- and community-acquired infections (von Baum and Marre, 2005). Within the gastrointestinal tract, commensal E.coli can transfer its antibiotic-resistant genes to diverse microorganisms, such as pathogenic bacteria, especially when exposed to antimicrobials (Smith et al., 2007). The transfer of antibiotic-resistant genes was first described by Smith (1969) who isolated E.coli strains from the digestive tract of humans and animals. Over the years, this finding has been confirmed by numerous studies (Aaerestrup and Wegener, 1999; Winokur et al., 2001; Angulo et al., 2004; Wang et al., 2006).

Humans become colonized and/or infected through physical contact, occupational exposure or food consumption. Food, especially of animal origin, is an important vehicle of antibiotic-resistant pathogens (von Baum and Marre, 2005; Riaño et al., 2006). Some studies have shown epidemiological relationships among different E.coli strains isolated from humans and foods and an increased number of resistant isolates (Voltattoni et al., 2002 Ramchandani et al., 2005; Manges et al., 2007; Johnson et al., 2007; Thorsteinsdottir et al., 2010).

The aims of this study were to evaluate the antimicrobial susceptibility of E.coli strains isolated from humans and foods, to identify potential pathogenic strains and to determine the possible epidemiological relationship among them by pulsed-field gel electrophoresis (PFGE).

Materials and Methods

Escherichia coli Strains

A total of 84 E.coli isolates from human (48) and food (36) samples were evaluated. Human isolates were obtained from stool (39) and urine (9) samples of outpatients at two major hospitals in the city of Salvador, Brazil. At the Food Microbiology Research Laboratory of the Federal University of Bahia (UFBA) and at the Central Laboratory of Public Health Prof. Gonçalo Moniz (LACEN/BA), food isolates were obtained from meat (8), chicken meat (1), milk (2), dairy products (6) and ready-made meals (19) according to Downes and Ito (2001).

Antimicrobial susceptibility profile

Susceptibility testing was carried out by a disc diffusion method according to the Clinical and Laboratory Standard Institute guidelines (CLSI, 2010). The tested antibiotics (Laborclin, Paraná, Brazil) included ampicillin (10 μg), cephalothin (30 μg), cefotaxime (30 μg), ciprofloxacin (5 μg), gentamicin (10 μg), tetracycline (30 μg), chloramphenicol (30 μg), trimethoprim/sulfamethoxazole (1.25/23.75 μg), amoxicillin/clavulanic acid (20/10 μg), levofloxacin (5 μg) and ceftriaxone (30 μg). E. coli ATCC 25922 was used as a quality control. Multidrug resistance was defined as resistance to at least two classes of antimicrobial drugs (Knezevic and Petrovic, 2008).

Investigation of pathogenic serotypes

The E. coli isolates were screened by multiplex-PCR (Tornieporth et al., 1995) with primers (IDT, USA) specific to the virulence genetic markers ipaH, eae, stx, bfp, LT and ST for E. coli enteroinvasive (EIEC), enteroaggregative (EAEC), enteropathogenic (EPEC), enterohemorrhagic (EHEC) and enterotoxigenic (ETEC), respectively, as described by Tornieporth et al. (1995) and Meng et al. (1998). The primers, the size of the amplification products (i.e., amplicons) and the references used in the detection of the six virulence gene markers are shown in Table 1. The non-pathogenic E. coli K12 HB101 was used as a PCR negative control, and EPEC O111: H2, H34x ST ETEC, ETEC LT 52593, EIEC O144: H25, and EHEC EDL933 were used as positive controls.

Table 1. Primers and amplicons size of the virulence genes searched in the 84 E.coli isolates studied.

Virulence Gene	Serotype	Primers	Sequence (5′-3′)	Size of product PCR	Reference
BFP gene	EPECa	EP1	CAA-TGG-TGC-TTG-CGC-TTG-CT	324	Tornieporth et al., 1995
		EP2	GCC-GCT-TTA-TCC-AAC-CTG-GT
LT gene	ETEC (LT)b	ETLT1	GCG-ACA-AAT-TAT-ACC-GTG-CT	708	Tornieporth et al., 1995
		ETLT2	CCG-AAT-TCT-GTT-ATA-TAT-GT
Sta gene	ETEC (ST)c	ETST1	CTG-TAT-TGT-CTT-TTT-CAC-CT	182	Tornieporth et al., 1995
		ETST2	GCA-CCC-GGT-ACA-AGC-AGG-AT
ipaH	EIECd	EI-1	GCT-GGA-AAA-ACT-CAG-TGC-CT	424	Tornieporth et al., 1995
		EI-2	CCA-GTC-CGT-AAA-TTC-ATT-CT
eae	EHECe	EHeae1	GTG-GCG-AAT-ACT-GGC-GAG-ACT-A	435	Meng et al., 1998
		EHeae2	GAT-CGT-AAC-GGC-TGC-CTG-ATA-TAA
hly	EHECe	EHhly1	AGC-CGG-AAC-AGT-TCT-CTC-AG	526	Meng et al., 1998
		EHhly2	CCA-GCA-TAA-CAG-CCG-ATG-T

^aenteropathogenic;

^benterotoxigenic the heat-labile enterotoxin (LT)-producing;

^centerotoxigenic the heat-stable enterotoxin(STa)-producing;

^denteroinvasive;

^eenterohemorrhagic.

Pulsed-Field Gel Electrophoresis (PFGE)

The Center for Disease Control and Prevention protocol for molecular subtyping of E. coli O157:H7 by PFGE was used (Bender et al., 1997). PFGE of XbaI-digested DNA (Sigma-Aldrich, USA) was performed on the isolates that were resistant to an antibiotic. Isolates exhibiting identical PFGE patterns were considered to be genetically indistinguishable, those exhibiting 1-3 band differences were considered to be closely related, and those exhibiting 4-6 band differences were considered to be possibly related.

Data Analysis

The XbaI (Sigma-Aldrich, USA) fingerprints were analyzed using GelCompar II software (Applied Maths, Kortrijk, Belgium). A similarity dendrogram was constructed by the unweighted pair group (UPGMA) method using the Dice similarity coefficient; the relatedness between the isolates was interpreted according to the method described by Tenover (1995).

Ethics

This project was approved by the Research Ethical Committee of Department of Health of the State of Bahia (No. 0028.0053.000-09).

Results and Discussion

Susceptibility profile

The prevalence of antimicrobial resistance among the E. coli strains isolated from humans and foods is shown in Table 2. Overall, 29 (34.6%) of the isolates were resistant to at least one of the 11 antibiotics tested, two (2.4%) had an intermediate susceptibility, and 53 (63%) were pan-sensitive. Among those isolates recovered from humans (n = 48), 30 (62%) were susceptible to all the tested antibiotics, 13 (27%) were resistant to ampicillin, 13 (27%) were resistant to tetracycline, and 11 (23%) were resistant to trimethoprim/sulfamethoxazole. Among those isolates recovered from food, 23 (64%) were susceptible to all the antibiotics tested, 9 (25%) were resistant to tetracycline, 4 (11%) were resistant to ampicillin, and 2 (5%) had an intermediate resistance to cephalotin (Table 2).

Table 2. Prevalence of antimicrobial resistance among Escherichia coli isolates from human and food source.

Antibiotics	% Resistance
	Human (N = 48)	Food (N = 36)	Overall (N = 84)
Ampicillin	27.0	11.0	20.2
Amoxicillin/Clavulanic acid	0	5.5	2.3
Cephalotin	4.0	8.3	6.0
Cefotaxime	2.0	0	1.2
Ceftriaxone	0	0	0
Ciprofloxacin	6.2	2.7	4.7
Chloramphenicol	6.2	2.7	4.7
Gentamicin	2.0	2.7	2.3
Levofloxacin	6.2	0	3.5
Tetracycline	27.0	25.0	26.0
Trimethoprim/sulfamethoxazole	23.0	5.5	15.4

These results are consistent with those reported by Jakobsen et al. (2010), who found that human isolates had a higher resistance percentage to ampicillin, trimethoprim/sulfamethoxazole and tetracycline than to other drugs. Furthermore, Pires et al. (2007) reported that the highest resistance percentage was to ampicillin, followed by trimethoprim/sulfamethoxazole.

In this study, the prevalence of tetracycline resistance was similar in E. coli isolates obtained from humans and foods, contrary to the results reported by Meyer et al. (2008), who investigated E. coli isolates obtained from foods, animals and humans in Germany. The authors found a higher rate of tetracycline resistance in E. coli isolated from animals and foods than in E. coli isolated from humans. However, in both studies, the rate of resistance to ampicillin was the second highest in the strains isolated from dietary sources. Ampicillin resistance has been widely observed even when the use antimicrobial agents is restricted, as reported by Hoyle et al. (2006), who analyzed E. coli isolates obtained from animal feces living in organic farms.

In a similar study conducted by Jakobsen et al. (2010) in Denmark, the rate of tetracycline resistance was higher in E. coli isolates obtained from food than in those obtained from humans. In Denmark, tetracycline is used mainly in swine farms (Vieira et al., 2009). Meyer et al. (2008) attributed this high rate of tetracycline resistance in E. coli isolates from food to the widespread use of antimicrobials during food production. According to the Brazilian Ministry of Agriculture, Livestock and Supply, the use of antimicrobials has not been allowed since 1998 (Ordinance No. 193 of 12.05.1998 repealed by the Normative Instruction number 26, 07.09.2009) (Brazil, 2009).

The rate of antimicrobial resistance observed in this study among the E. coli isolates obtained from food was lower than that reported by Van et al. (2007), Van et al. (2008) and Altalhi et al. (2000), who reported resistance rates close to 100%. Because of the use of antimicrobial agents in animal feed and the selective pressure that such drugs exert on the microorganisms, it is assumed that isolates obtained from foods of animal origin have a resistance profile that is superior to those from other food sources.

The percentage of human isolates that exhibited multidrug resistance was 77.7% (14/18); the percentage of food isolates with multidrug resistance was 45.5% (5/11) (Table 3). Similar results were described by Thorsteinsdottir et al. (2010), who found a higher proportion of multidrug resistant E. coli in human isolates than in food isolates.

Table 3. Multiple antimicrobial drug resistance among Escherichia coli isolates by source.

Antimicrobial	% resistant
	Human (N = 48)	Food (N = 36)	Total (N = 84)
AMP	4.2 (2)	2.7 (1)	3.5 (3)
TET	4.2 (2)	13.8 (5)	8.3 (7)
CFL	0	0	0
AMP TET	2.0 (1)	0	1.2 (1)
AMP, SUT	6.2 (3)	0	3.5 (3)
AMC, CFL	0	2.7 (1)	1.2 (1)
TET, CLO	0	2.7 (1)	1.2 (1)
TET, SUT	4.2 (2)	0	2.4 (2)
AMP, TET, SUT	4.2 (2)	2.7 (1)	3.5 (3)
AMP, TET, SUT, CLO	6.2 (3)	0	0
AMP, AMC, CFL, TET	0	2.7 (1)	1.2 (1)
GEN, TET, CIP, LVX	2.0 (1)	0	1.2 (1)
AMP, CFL, TET CIP, LVX	2.0 (1)	0	1.2 (1)
AMP, CFL, GEN, TET, CIP, SUT	0	2.7 (1)	1.2 (1)

AMP-ampicillin; TET-tetracycline. SUT- trimethoprim/sulfamethoxazole; CLO-chloramphenicol. CFL-cephalothin. AMC-amoxicillin/clavulanic acid; LVX-levofloxacin. CIP-ciprofloxacin;GEN-gentamicin. CRO-ceftriaxone. CTX-cefotaxime.

Multiplex PCR

The analysis of the genes encoding important virulence markers of E. coli revealed that none of the isolates carried the ipaH, eae, hly, bfp, LT and ST genes. Paneto et al., 2007, and Gonzalez et al., 2000, reported low prevalence rates of pathogenic E. coli in food samples.

Genetic diversity of E. coli isolates using pulsed-field gel electrophoresis

A genetic diversity analysis by PFGE of 31 resistant isolates revealed 29 distinct restriction patterns (Figure 1).

Figure 1. Dendrogram based on XbaI PFGE types identified among 31 E. coli strains. The data were sorted by the UPGMA method. AMP-ampicillin, TET-tetracycline, SUT- trimethoprim/sulfamethoxazole, CLO-chloramphenicol, CFL-cephalothin, AMC-amoxicillin/clavulanic acid, LVX- levofloxacin, CIP-ciprofloxacin, GEN-gentamicin, CRO-ceftriaxone, CTX-cefotaxime. *Isolates that show intermediate susceptibility to cephalotin.

There were no strains with an indistinguishable electrophoretic pattern that could characterize a clone. The highest similarity percentage was 84.2% between the human and food isolates. Two groups (F and M) contained one human isolate and one food isolate that differed by 3 bands (i.e., close related isolates). Despite the limitations, we cannot exclude the hypothesis that a food reservoir exists and that food-borne transmission of E.coli is common.

The genetic diversity among the E. coli isolates in this study can be explained by the fact that the isolates were not epidemiologically related; they had been isolated from different locations. The isolates were obtained in the same period and had some resistance characteristics. Moreover, the human isolates were obtained from outpatients who were epidemiologically unrelated. Even when studying geographically related strains, an absence of clones is possible, as reported by Campos et al. (2009), who observed an absence of an endemic clone and a high diversity of E.coli strains isolated from humans and foods.

In practice, the typing of microorganisms is more effective as an aid during the investigation of outbreaks when applied to a low number of epidemiologically related isolates (Tenover et al., 1995). However, it is necessary to study strains of different origin, especially those not involved in the outbreaks with antimicrobial resistance to assess possible modifications of their genome. Such an analysis is crucial for understanding their origin and their impact on public health and for establishing possible control strategies.

Conclusion

This study, which was performed in the third largest city of Brazil, found that 34.6% of the E. coli isolates were antimicrobial resistant. The most frequent antibiotic resistance observed was tetracycline in both human and food isolates. In addition, we found a high percentage of multidrug resistance among these isolates, particularly in isolates obtained from humans and foods of animal origin. None of the strains was considered to be pathogenic, and the molecular typing by PFGE found a high degree of genetic diversity among the isolates.