Characterization of Extended-Spectrum β-Lactamase Producing- and Carbapenem–Resistant Escherichia coli Isolated from Diarrheic Dogs in Tunisia: First Report of bla_IMP Gene in Companion Animals

Abstract

Escherichia coli is an important opportunistic pathogen, causing several infections in dogs. The antimicrobial resistance of E. coli occurring in companion animals becomes an emerging problem. This study aimed to estimate the prevalence of ESBL-producing E. coli in diarrheic dogs, investigate the occurrence and molecular characterization of carbapenem-resistant isolates, and determine their virulence genes. Fecal samples were collected from 150 diarrheic dogs in Tunisia. E. coli isolates were screened for antimicrobial resistance against 21 antibiotics by the disk diffusion method. The characterization of β-lactamase genes, associated resistance genes, and virulence genes was studied using PCR. Among 95 E. coli strains, 25 were ESBL-producing, and most of them were multidrug-resistant. The most prevalent β-lactamase genes were bla_CTX-M1 (n = 14), bla_TEM (n = 3), and bla_CMY (n = 2). The bla_IMP carbapenemase gene was found in two carbapenem-resistant isolates, which showed that carbapenemase-producing E. coli spread to companion animals in Tunisia. Different virulence genes associated with extraintestinal pathogenic E. coli were detected. This is the first report of the characterization of carbapenem resistance and virulence genes in dogs in North Africa. Our study showed that diarrheic dogs in Tunisia can be a potential reservoir of ESBL- or carbapenemase-producing E. coli with a possible risk of transmission to humans.

1. Introduction

Escherichia coli (E. coli), a Gram-negative bacteria belonging to the Enterobacteriaceae family, is frequently found in the gastrointestinal tract of both humans and companion animals [1].

Nevertheless, this opportunistic pathogen can cause intestinal and extra-intestinal diseases and may be associated with diarrheal syndrome. In dogs, acute diarrhea is one of the most common gastrointestinal problems, potentially leading to severe dehydration and death [2]. Several pathogens can be responsible for enteritis with diarrhea, including canine parvovirus type 2, protozoa such as Giardia duodenalis and Cryptococcus spp., bacteria such as E. coli, Salmonella spp., Campylobacter spp., Clostridioides difficile, Clostridium perfringens, and Providencia alcalifaciens. Some bacteria are zoonotic, particularly as in the case of Campylobacter spp., for which the dog is considered a reservoir with a great risk of transmission to humans, like in a study in Lebanon reporting a fecal prevalence of 17% [3,4]. In veterinary medicine, beta-lactam antibiotics are highly prescribed antibacterial agents for dogs in order to treat infectious diseases [5]. This has led to the emergence and dissemination of resistant bacteria and allowed an increase in the Extended-Spectrum Beta-Lactamases (ESBL)-producing strains.

ESBL-producing Gram-negative bacteria are a serious threat to public health in human medicine as well as in the veterinary context. Worldwide, several studies have reported the transmission of multidrug-resistant bacteria between companion animals and their owners [6,7,8,9]. In addition, different studies all over the world have investigated ESBL-producing bacteria in companion animals [10,11]. In dogs and other companion animals, multidrug-resistant bacteria are frequently reported [12].

Recently, carbapenem resistance in companion animals has emerged, and different carbapenem resistance genes have been detected worldwide in companion animals, livestock, and sea food [13].

Antimicrobial resistance (AMR) among companion animals, particularly household pets, is a complex area that is becoming increasingly important. In this context, despite their close contact with humans, little attention has been given to the prevalence of antimicrobial resistance in companion animals. According to Dierikx et al. [14], the close contact between humans and companion animals such as dogs and cats makes the transmission of resistant organisms more likely to occur. In addition, dogs could increase the risk of transmitting zoonotic microorganisms, such as EPEC (Enteropathogenic E. coli) and STEC (Shiga-toxin producing E. coli) [15], and dogs have been suggested as a potential reservoir for ExPEC (Extra-intestinal pathogenic E. coli) [16]. ExPEC are facultative pathogens responsible for 80% of urinary tract infections, but also neonatal meningitis, surgical site infections, and pneumonia. ExPEC typically belongs to phylogroup B2, an E. coli genetic group that accumulates virulence factors, and occasionally to phylogroups D, F, or G. Food animals and pets are considered potential reservoirs of ExPEC [17]. It is noteworthy that contact with dogs or dog feces is considered a risk factor for the acquisition of resistant ExPEC [18].

The aim of this study was to determine the occurrence and molecular characterization of ESBL, carbapenemases, and virulence genes (associated with STEC and ExPEC) among ESBL-producing E. coli isolated from diarrheic dogs in Tunisia.

2. Materials and Methods

2.1. Sample Collection

From February 2018 to March 2019, 150 fecal samples were collected from diarrheic dogs at the National School of Veterinary Medicine of Sidi Thabet, Tunisia. All samples were obtained from the clinical department with diarrhea as the criteria of choice. The age of the dogs was between 3 months and 2 years. The number of males and females is 58 and 48, respectively.

2.2. Bacterial Isolation and Identification

MacConkey supplemented with cefotaxime (CTX, 1 μg/mL) was used to streak all samples after being enriched in water peptone buffer. Following a 24 h incubation period at 37 °C, colonies exhibiting a characteristic E. coli morphology and displaying a pink color were verified using chemical tests: triple sugar iron agar, Simmons’ citrate agar, and urea indole medium.

2.3. Antimicrobial Susceptibility Testing

The susceptibility of strains was detected using the inhibitory zone diameter interpretation standards according to the recommendations of the European Committee on Antimicrobial Susceptibility Testing (EUCAST, 2018E) [19].

The phenotypic detection of ESBL was realized by the double disk synergy test using a disk of amoxicillin and clavulanic acid placed in the center of a Mueller–Hinton agar plate surrounded by cefotaxime, ceftazidime, and cefepime disks. The enhanced inhibition zone of any of the cephalosporin disks on the side facing amoxicillin and clavulanic acid was considered an ESBL producer. Phenotypic testing for colistin was carried out using the Colispot test^®. The isolates were tested on 21 antibiotics (Mast Ltd.Group, Merseyside, UK). The following antimicrobial disks were used (µg per disk): amoxicillin (25), piperacillin (30), cefotaxime (30), cefoxitin (30), cefepime (30), ticarcillin–clavulanic acid (75/10), amoxicillin–clavulanic acid (20/10), aztreonam (30), cephalothin (30), cefuroxime (30), ceftazidime (30), ertapenem (10), gentamicin (10), streptomycin (10), nalidixic acid (30), enrofloxacin (5), chloramphenicol (30), florfenicol (30), tetracycline (30), and trimethoprim–sulfamethoxazole (1.25/23.75).

2.4. Genomic DNA Extraction

The boiling method was used to extract the genomic DNA from each sample. In short, bacterial colonies were suspended in 1 milliliter of distilled water, and the mixture was centrifuged for five minutes at 13,200 rpm. After that, 100 µL of distilled water was added after the supernatant was removed. The samples were heated for ten minutes to 95 °C. After being boiled, the cell suspensions were diluted and kept cold (−20 °C), intended for PCR (Polymerase Chain Reaction) usage.

2.5. Identification of Antibiotic Resistance Genes

PCR amplifications were used for the searched β-lactamase-encoded genes: bla_CTX-M, bla_TEM, bla_SHV, bla_CMY, bla_VIM, bla_IMP, bla_NDM-1, and bla_OXA-48 ESBL-producers’ strains. The presence of genes linked to resistance to several antibiotic classes, such as quinolones, aminoglycosides, sulfonamides, and tetracyclines, was assessed. Table 1 lists the primer sequences and PCR conditions.

2.6. Phylogenetic Grouping of the Isolates

According to the previous descriptions by Clermont et al. [20], the phylogenetic grouping (A, B1, B2, and D) was examined in the isolates. The determinations of phylogenetic groups A, B1, B2, and D depending on the existence of the chuA and yjaA genes and TSPE4.C2 determinants were detected by PCR.

2.7. Virulence Genes Identification

Each E. coli strain was examined for the presence of virulence genes. Multiplex PCR for Shiga-toxins (stx1 and stx2), intimin (eae), and enterohemolysin (ehxA) was used to identify STEC strains. Hemolysin (hly), cytotoxic necrotizing factor (cnf1), and cytolethal distending toxin (cdt3) were all tested using triplex PCR. Simplex PCR was used to identify the virulence genes linked to ExPEC, which include aer (aerobactin system), papA (P fimbriae), bfpA (bundle forming pilus), papG-III (P adhesin), fimH (type 1 fimbriae), traT (serum survival gene), ibeA (invasion of brain endothelium), and sfa/foc (S and F1C fimbriae). Duplex PCR was also used to analyze the virulence genes fyuA (gene encoding yersiniabactin) and iutA (ferric aerobactin receptor). Table 2 lists the virulence gene primers and PCR conditions.

2.8. Data Analysis and Interpretation

Using SPSS version 26 software, a Chi-square test was employed to examine any significant differences in risk factors, the associations of phylogroups, and the resistance and virulence genes (IBM Corporation, Somers, NY, USA). p < 0.05 was designated as the level of statistical significance.

3. Results

3.1. Bacterial Strains

Among 150 fecal samples obtained from diarrheic dogs, we isolated 95 E. coli isolates. A total of 25 E. coli isolates were ESBL-producing (26.6%), as shown in Table 3. The samples were collected from diarrheic dogs at the National School of Veterinary Medicine of Sidi Thabet, Tunisia. The age of the dogs was between three months and two years. Fifty-eight samples were from females and forty-eight were collected from males.

3.2. Antibiotic Resistance of ESBL-Producing Strains

The twenty-five ESBL-producing E. coli isolates were tested for their susceptibility to twenty-one antibiotic agents. We found that most of these strains were multidrug-resistant (MDR = 72%) since they were resistant to at least three families of antibiotics. Most ESBL isolates were resistant to amoxicillin (96%), tetracyclin (92%), and cefotaxime (84%). Rates of resistance for 25 ESBL-producing E. coli are listed in Table 4.

The isolation rate of ESBL-producing E. coli in this study was higher in males than females, but without statistical significance (p-value = 0.544). According to the dog’s age, the ESBL frequency was lower for the category superior to 24 months than other categories without statistical significance (p-value = 0.482). The ESBL rates were higher in summer and autumn than other seasons, with this difference being statistically significant (p-value = 0.011) (Table 5).

3.3. Beta-Lactam Resistance Genes

In this study, three different β-lactamase genes belonging to Ambler molecular class A (bla_TEM, bla_CTX-M, and bla_SHV) were detected. The most prevalent gene was bla_CTX-M1 (n = 14), bla_TEM (n = 3), and bla_CTX-M9 (n = 2), while only one isolate carried the bla_SHV gene. Otherwise, two isolates carried the bla_CMY gene (AmpC cephalosporinase).

3.4. Non β-Lactam Resistance Genes

Among the 25 ESBL-producing isolates, five harbored the tetA gene and sul2, three carried the sul1 gene, and only two isolates were aac (3)II positive.

3.5. Phylogroups

Most of the ESBL-producing E. coli isolates belonged to phylogroups A (16 isolates) and D (7 isolates), and only one isolate belonged to phylogroup B1.

3.6. Virulence Genes

Virulence genes implicated in ExPEC related to adhesins (fimH, papC, papG allele III, sfa/foc), toxins (hly, cnf1), and invasion factors (ibeA, aer, iutA, fyuA) were investigated among ESBL-producing isolates.

In regard to virulence factors associated with ExPEC, a high number of strains harbored the fimH virulence gene (21/25, 84%), 18/25 (72%) carried the traT gene and 14/25 (56%) carried the aer gene. In addition, eight strains harbored the iutA gene, four harbored the fyuA gene, and three isolates carried both the fyuA and the iutA genes. For the eae and bfpA virulence genes, they were both found in only one ESBL each.

3.7. Distribution of Virulence Genes and Resistance Genes in Phylogenetic Groups

The analysis of phylogroup distribution among the resistance genes revealed that group A was the most dominant that carried virulence factors, followed by group D. β-lactamase genes bla_CTX-M-1 and bla_CTX-M-15 were distributed among phylogenetic groups, whereas bla_CTX-M-9 and bla_TEM were present only in groups A and D. Resistance genes bla_SHV, bla_CMY, and bla_IMP were associated with group A, and the difference did not reach statistical significance (Figure 1).

The association between virulence genes and phylogenetic groups revealed that the genes traT, fimH, and iutA were widely disseminated in the groups. Genes traT, fimH, iutA, fyuA, and aer were more strongly associated with groups A and D than other groups. Group A was the dominant group that carried the virulence genes; BfpA and eae were found only in group A (Figure 2).

4. Discussion

In E. coli, ESBLs and carbapenemases are mainly responsible for the emerging resistance to the β-lactam antibiotics, especially the 3rd generation cephalosporins and carbapenems [17]. In the present study, we conducted the molecular detection and characterization of β-lactamase genes in ESBL-producing E. coli isolates from diarrheic dogs in Tunisia and also revealed the association between the phylogenetic groups and virulence gene profiles.

In this study, among 95 E. coli isolates obtained from 150 diarrheic dogs, 25 E. coli isolates were ESBL-producing (26.6%). The overall prevalence of E. coli was 62.6%, which is higher than results found in Brazil since they obtained a rate of 57.8% of E. coli isolated from dogs with diarrhea [21]. Our results are also higher than what was reported in Egypt [22], with a prevalence of E. coli of 23.7% (19/80) in the examined diseased dogs. In addition, our findings showed a rate of 26.6% of ESBL-producing E. coli isolated from diarrheic dogs, near the rate of 30% in India (Tudu et al. 2022), but lower than the rate of 55% reported in the Netherlands [23]. On the other hand, our results are higher than those in the United Kingdom, with only 1.9% [24].

The antimicrobial susceptibility of ESBL-producing E. coli isolates showed that these isolates had high rates of resistance. We found that 96% of the isolates were resistant to amoxicillin (24/25), 92% to tetracycline (23/25), and 84% to cefotaxime (21/25). These findings are higher than those found by Algammal et al., in which they reported that 100%, 84.2%, and 42.1% of the strains were resistant to tetracycline, amoxicillin, and cefotaxime, respectively, in 80 dogs suffering from bloody diarrhea [22].

In 1986, FEC-1 (Fujisawa E. coli-1) was the first CTX-M-type ESBL enzyme, discovered in a cefotaxime-resistant E. coli isolated from the feces of a laboratory dog in Japan [25]. In the year 2000, the first case of an ESBL-producing E. coli from animal origin was detected in Spain [26] when they reported for the first time the gene bla_SHV-12 from a sample of a dog with recurring chronic cystitis. A few years later, Carattoli et al. [27] detected the first cases of bla_CTX-M-1 producing bacteria in dogs and cats with and without pathology. Since then, different studies worldwide have reported increasing numbers of companion animals that are hosts for ESBL-producing E. coli [28,29].

We found three different β-lactamase genes belonging to Ambler molecular class A (bla_TEM, bla_CTX-M, and bla_SHV) among the strains. The most prevalent genes were bla_CTX-M1 (n = 14), bla_TEM (n = 3), bla_CTX-M9 (n = 2), and only one isolate carried bla_SHV (n = 1). Our results showed a total of 14 ESBL-producing E. coli (14/25, 56%) that harbored the bla_CTX-M-1 gene. In this context, many studies have reported a high prevalence of the bla_CTX-M-1 gene from ESBL-producing E. coli in companion animals [30,31]. Comparing our results to those found in the Netherlands [23], we found a higher number of ESBL harboring the bla_CTX-M-1 gene since they investigated 20 dogs with diarrhea and found the bla_CTX-M-1 gene in only two diarrheic dogs. In addition, our findings are higher than in a study in Germany [32], which showed that the bla_CTX-M-1 gene was less identified in dogs since they found this gene in 11 isolates among 67 dogs. For the bla_CTX-M-9, only two ESBL in our study carried this gene, and this is comparable to results found in Germany [32], where they found the bla_CTX-M-9 gene in only two isolates among 67 dogs. Our findings showed that two ESBL carried the bla_CMY gene, which is similar to results found in Germany [33].

It is noteworthy that resistance phenotypes, geographical regions, and the history of antimicrobial treatments on animals can affect the prevalence of bla_CTX-M genes. The high prevalence suggests a significant role for E. coli isolates from companion animals as ESBL gene reservoirs [31].

Given the absence of new commercialized antibiotics, numerous studies aim to develop alternatives such as the use of bacteriocins, essential oils, bacteriophages, or antimicrobials with other mechanisms of action such as metallophores. For example, the ability of metallophores to complex and transport metals into bacterial cells has also led to an alternative strategy for developing antimicrobial agents by complexing other antimicrobial metals and delivering potentially lethal concentrations of them into bacterial cells [34].

Concerning resistance to carbapenems, most reports of carbapenemase-producing Enterobacteriaceae have been mainly from humans [35,36], which may indicate a human-to-animal transfer through close contact [37]. In Enterobacteriaceae, the carbapenemases can be divided into three types: the class A carbapenemase group, which includes the KPC-type, the class B carbapenemase group (Metallo β-lactamases), and the class D carbapenemase group, which includes the OXA-48 type.

Worldwide, the most common carbapenemase type in E. coli is the OXA type. This type is endemic in Turkey, northern Africa, and India [38]. Worthy of note, carbapenemase genes have been identified in companion animals worldwide [33,39,40] and are associated with a high potential for dissemination [41]. In Tunisia, the first carbapenemase-producing isolate emerged from a Tunisian university hospital in 2006 [42]. Since then, different carbapenemase variants have been isolated from several origins (hospitals and wastewater effluents) in Tunisia [43,44]. Nevertheless, there has been some concern about carbapenemase-producing bacteria of animal origin all over the world [45,46], but studies on resistance to carbapenems in Tunisia are still limited [47,48]. In this context, we conducted the molecular detection and characterization of the carbapenemase genes (bla_IMP, bla_OXA-48, bla_NDM-1, bla_KPC, and bla_VIM) in ESBL-producing E. coli isolated from diarrheic dogs. It is important to highlight that our results demonstrated the presence of the bla_IMP gene in two isolates using PCR and sequencing and were confirmed by an E-test strip. To the best of our knowledge, this is the first report of carbapenemase-producing E. coli from diarrheic dogs in Tunisia. In contrast to our result, the bla_IMP gene was found in E. coli isolates from healthy rabbits in Tunisia [47] and also among Enterobacteriaceae isolated from wild boar (Sus scrofa) [49], but has never been reported before from dogs in this country.

It is noteworthy that carbapenemase genes are mainly localized on plasmids and are therefore highly transferable between different bacteria. The transfer of a plasmid carrying a carbapenemase gene suggests the possibility of dissemination in different fields (animal, human, and environmental) [50].

Carbapenemase-producing MDR bacteria were first described almost exclusively in humans, but since 2011, they have also been detected in livestock, companion animals, wildlife, and different environmental compartments, indicating their transfer to new hosts and reservoirs [51].

Worldwide, the prevalence of carbapenem-resistant Enterobacteriaceae among companion animals seemed to be low in Algeria, with a prevalence of 2.58% in dogs [52], as well as in France and Spain, with a prevalence of 0.6% [41,53]. This indicates that carbapenemase-producing isolates are an emerging problem and can constitute a serious threat to public health.

According to a recent study [54], there is no available data that identified the virulence genes of E. coli isolated from companion animals in Africa. To the best of our knowledge, this is the first report of virulence genes among E. coli isolated from diarrheic dogs in Tunisia. In this study, we found one ESBL-producing E. coli that carried the eae gene. In Brazil [55], the eaeA gene was identified in 12 (12.6%) isolates from 68 fecal samples of diarrheic dogs, which is higher than our findings. In contrast, a study in Iraq [56] showed that six (5.8%) isolates that were identified as E. coli O157:H7 and harbored the eaeA gene in 104 dogs with (11/16; 68.8%) and without diarrhea (7.9%; 7/88); this is higher than our results.

Our ESBL-producing E. coli with the positive eae gene can be classified as atypical enteropathogenic E. coli (EPEC) and did not harbor the stx and bfpA virulence genes of EPEC (eae+, stx^_, bfpA^_). However, in this study, we did not isolate the typical EPEC (eae+, bfpA+). Depending on their virulence genes, EPEC strains are subdivided as typical and atypical EPEC. For typical EPEC, humans are the only reservoirs, but for atypical EPEC, both animals and humans can be reservoirs, and atypical EPEC seems to be an important cause of diarrhea [57]. EPEC are characterized by the production of intimin (eae), and they can be classified as typical or atypical depending on the presence or absence of bundle-forming pili (bfp) [58]. In our data, only one isolate carried the bfpA gene.

Since data on virulence genes associated with dogs is not available in Africa, we compared our findings to those on other continents like Europe and America. Our findings of virulence factors associated with ExPEC showed a high rate (84%) of ESBL strains that harbored the fimH virulence gene (21/25), 72% (18/25) carried the traT gene, and 56% (14/25) carried the aer gene. In addition, 32% of ESBL strains harbored the iutA gene, 16% harbored the fyuA gene, and 12% carried both fyuA and iutA. These findings are higher than results found in the United States, where a rate of 69.1% for the fimH gene and 60.3% for the traT gene was reported among 68 ESBL strains isolated from dogs and cats [31]. In contrast, our rates of iutA and fyuA were 32% and 16%, respectively, lower than the findings of Liu et al. 2016.

Concerning the stx1, stx2, eae, and hly virulence genes, a recent study performed in Egypt reported that the prevalence of stx1, eaeA, and hlyA was 100% and 47.3% for stx2 in fecal samples of diseased dogs suffering from hemorrhagic diarrhea [22]. These findings are higher than our results, since we found only one isolate to be eae positive and all isolates did not harbor the stx1, stx2, or hly virulence genes.

It is considered that virulent extra-intestinal E. coli (ExPEC) strains usually belong to phylogenetic groups B2 and D in comparison with other phylogenetic groups. In contrast, commensal strains frequently belong to phylogenetic groups A and B1. In the present study, the isolates carrying virulence genes belonged mainly to phylogenetic groups A (16 isolates) and D (7 isolates), and only one isolate belonged to B1. Our results showed that 7 isolates belonged to phylogroup D in association with virulence factors of adhesion and invasion (fimH and aer) that are linked to ExPEC infections, so 32% of the isolates may be considered ExPEC, and the remaining isolates that belonged to phylogroup A might be commensal strains or intraintestinal pathogenic (InPEC). In contrast to our findings of one isolate that belonged to phylogroup B1, Coura et al. [21] obtained eight isolates from the B1 and E phylogroups, and EPEC was the more frequent pathovar identified, known as an important diarrheagenic agent in dogs.

5. Conclusions

Our study showed that diarrheic dogs in Tunisia can be a potential reservoir of ESBL-producing E. coli and demonstrated that carbapenemase-producing E. coli in companion animals is emerging. In addition, a high occurrence of antimicrobial resistance of E. coli was observed in fecal samples of dogs with diarrhea, highlighting the need for prudent use of antimicrobial agents in veterinary medicine in order to decrease the selection and spread of multi-drug resistant bacteria.

Our findings strongly suggest that isolates with carbapenem resistance are currently circulating among companion animals and can act as a reservoir of resistance genes transmitted between humans and animals. Fortunately, carbapenemases are still rare among isolates of companion animals, but the number of studies on resistance to carbapenems is increasing worldwide, and this resistance must be meticulously monitored.