Multidrug Resistant Escherichia coli Harboring Extended-spectrum β-lactamase-encoding Genes Isolated from Clinically-healthy Pigs

Extended-spectrum β-lactamase (ESBL)-producing Escherichia coli is a serious-global public health issue. A total of 292 E. coli isolates obtained from fecal samples of pigs in Central (n = 103) and Northeastern (n = 189) provinces of Thailand were included in this study. Eighty-six E. coli isolates were phenotypically confirmed to be β-lactamase producers (29.5%) and screened for the presence of β-lactamase genes. The genes in CTX-M family was most frequently found (90.7%). The blaCTX-M-15 gene (59.3%) was predominantly identified CTX-M genotype, followed by blaCTX-M-14 (31.4%) and blaCTX-M-4 (25.6%). The blaTEM-1 gene was prevalent (75.6%). The blaCTX-M-4 and blaCTX-M-14 genes were located on conjugative plasmid. The results highlight healthy pigs as reservoirs of ESBLproducing E. coli carrying ESBL genes that could be horizontally transferred.


Introduction
The rise in the occurrence and spread of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli is currently a global emergency. These pathogens constitute a significant threat to the efficacy of many antimicrobials including 3 rd and 4 th generation cephalosporins [1,2]. Evidently, acquisition of antimicrobial resistance (AMR) genes arises due to continuous and indiscriminate use of antimicrobials in food animal production [3]. The dissemination of CTX-M type ESBLs in the environment, humans, companion animals and food animals (including pigs, cattle and chickens) is on the increase [4,5] thus paving the way for animal to human transmission and vice versa [6,7].
E. coli acquire resistance to extended-spectrum cephalosporins through transfer of conjugative plasmids carrying genes for ESBLs or Ampicillin class C (AmpC) β-lactamases [8]. These ESBL genes are commonly found in E. coli isolated from humans and animals, possibly because of plasmid transfer or spread of unique clonal lineages [9]. Occurrence of ESBL-producing Enterobacteriaceae from stool and fecal samples collected from apparently healthy individuals and pigs in Thailand has been reported [10,11], highlighting the possible animal to human transmission and vice versa. ESBLproducing E. coli were initially observed in human clinical isolates, but are now increasingly found in food animals [12,13]. It was suggested that food-producing animals may be an important source of ESBL-producing bacteria that might spread through the food chain [11]. The prevalence of ESBLproducing E. coli isolated from food animals have been reported in France, Denmark, Spain, China, and Nigeria [3,14,15]. These studies highlighted role of food animals, especially pigs, as a reservoir of ESBL-producing E. coli and their resistance determinants that may be transferred to humans and contaminate the environment.
Spread of ESBL-producing E. coli is a worldwide phenomenon. The burden of ESBL-producing E. coli from most developing countries Thailand inclusive is until recently receiving much attention.
Over the past few years, awareness into the threat of evolution and spread of AMR in Thailand has considerably increased. However, knowledge on ESBLs in bacteria from food animals and its distribution is still limited. This study aimed to explore the importance of food-animals as sources of infection to humans by characterizing ESBL phenotype and genotype in E. coli isolated from healthy pigs.

Results and Discussion
Antimicrobial susceptibilities and ESBL production are shown in Table 3. The results showed that 29.5% (86/292) of E. coli isolates were ESBL producers (Table 3). This was lower than other similar studies carried out in Thailand. Boonyasiri et al. [16] reported a 40% and 76.7% prevalence of ESBLproducing E. coli from healthy chickens and pigs in an eastern and a northern province in Thailand.
Nuangmek et al. [17] reported occurrence of 36.7% (n = 588) EBSL-producing E. coli from 107 pig farms and 89 layer farms in the country. The difference in these two studies may be attributed to many factors such as the density of pigs in each farm, difference in sensitivity of detection methods, farm management practices and frequency of antimicrobial usage. Taken together, it is suggested that pig farms in Thailand might serves as a hub for ESBL-producing E. coli and concerted efforts towards prudent use of antimicrobial agents are needed.
Among ESBL-producing E. coli, 98.8% (85/86) were resistant to ampicillin and tetracycline, followed by streptomycin (95.3%, 82/86), gentamicin (93%, 80/86), chloramphenicol (70.9%, 61/86), sulphonamide (64%, 55/86), ciprofloxacin (57%, 49/86) and trimethoprim (55.8%, 48/86) ( Table 3). Statistically-significant association (p < 0.05) between ESBL-producing E. coli and resistance to four antimicrobial agents was observed (Table 3). There was a positive association between the occurrence of ESBL-producing E. coli and resistance to ampicillin, gentamicin and streptomycin. The positive correlation between ESBL production and ampicillin resistance supports that ESBL producers displayed significantly higher rates of ampicillin resistance, in agreement with a previous study [18]. Co-resistance to gentamicin and streptomycin suggests that ESBL genes and genes encoding resistance to gentamicin and streptomycin may be co-located either on the same plasmid or on different plasmids with different incompatibility groups within the same isolate. In contrast, a negative association was observed between ESBL-producing E. coli isolates and resistance to trimethoprim. Interestingly, resistance to tetracycline (p=0.13) and chloramphenicol (p=0.13) were not significantly associated with ESBL producers. However, the resistance phenotype to these two antibiotics was horizontally transferred in conjugation experiment. It is likely that ESBL genes and the genes encoding tetracycline and/or chloramphenicol resistance were located on different plasmids in different Inc groups and co-selected by ampicillin. Therefore, limiting or suspending the use of a single antimicrobials does not effectively lower the selective pressure of the development and spread of AMR. Actions to contain AMR must be strengthened to ensure stewardship of all antibiotics.
ESBL-producing E. coli showed multidrug resistance (MDR) against at least three different classes of antimicrobial agents. This agrees with a previous study demonstrating the occurrence of MDR-ESBL producing E. coli from hospitalized patients, livestock wastewater and the environment [19]. Most ESBL genes are usually located on plasmids with various resistance determinants [18]. This is supported by the observations in the conjugation experiment demonstrating that transconjugants were additionally resistant to non-cephalosporin antibiotics (i.e. chloramphenicol, streptomycin, gentamicin and tetracycline) ( Table 5). Therefore, the frequent usage of extended-spectrum cephalosporins for prophylaxis or treatment purposes can facilitate the emergence of MDR E. coli.
The most common type of ESBL gene identified was blaCTX-M (90.7%). The CTX-M group I gene was detected in 59.3% of the E. coli isolates, all of which were identified as blaCTX-M-15, followed by blaCTX-M-14 (31.4%) and blaCTX-M-4 (25.6%) ( Table 4). This agrees with previous studies conducted in food animals, humans and environment in Thailand [20,21]. However, previous studies reported CTX-M-14 as the most prevalent ESBL gene identified from food animals, healthy and hospitalized individuals and environmental waste water in China, Korea, and Thailand [14,22,23]. Interestingly, another study reported CTX-M-1 as the most predominant ESBL gene in food animals in UK, Germany, Tunisia and Switzerland [24]. These results showed that genotypes of ESBL vary across different geographical location. However, it is just a matter of time, before genotypes that were usually restricted to one geographical area will be isolated from areas and sources that they have not previously been reported to occur. On the other hand, the occurrence of CTX-M-15 in pigs reported in this study, indicated the zoonotic potentials of ESBL producing E. coli. A previous study showed that CTX-M-15 is commonly associated with ESBL producing E. coli isolated from humans [24]. Both CTX-M-1 and CTX-M-15 belongs to the same lineage, indicating the widespread of the CTX-M group in different sectors.
The blaTEM gene was observed in 75.6% of the ESBL-producing E. coli isolates and all were identified as blaTEM-1. This is not surprising because TEM is the most common β-lactamase produced by Gram-negative bacteria particularly the Enterobacteriaceae family. TEM-1 is an enzyme that confers resistance to narrow and broad spectrum β-lactams such as penicillin and amoxicillin, two of the most commonly used antimicrobial agents in pig production and whose resistance has been frequently observed [25]. A previous study in Thailand reported the occurrence rate of blaTEM among ESBL producing E. coli in clinically healthy pigs (87.9%, 282/321) [25]. Our finding showed that most ESBL-positive isolates (83.7%) carried more than one bla gene. Production of more than one type of ESBLs in an isolate is not uncommon and has been previously reported [21]. More than ninety percent (90.8%) of the blaTEM-1 positive E. coli isolates also carried CTX-M type ESBL, in agreement with a previous study [23]. The fact that it is a common finding highlights that it may have considerable impact on public health.
In addition, one ESBL-producing E. coli isolate is susceptible to ampicillin (Table 3) and this isolate was negative to all ESBL genes tested in this study. This phenomenon agrees with previous studies that were conducted in ESBL-producing E. coli from various clinical samples [26][27][28][29]. However, it is still unclear about genetics underlying ESBL production in this ampicillin-resistant ESBL producing strain. Taken together, these results demonstrated that ESBL-producing E. coli can harbor various types of ESBLs with different substrate (antibiotic) profiles and do not always exhibit ampicillin-resistance phenotype.
Based on the conjugation experiment, only 5 out of 13 ESBL producers carrying blaCTX-M-4 and blaCTX-M-14 yielded ampicillin-resistant Salmonella transconjugants (Table 4) with the conjugation efficacy of 10 -7 -10 -8 . The Salmonella transconjugants were confirmed to be ESBL producers and harbor the corresponding ESBL encoding genes (blaCTX-M-4 and blaCTX-M- 14). Therefore, the blaCTX-M-4 and blaCTX-M-14 gene in these 5 isolates were horizontally transferred to the Salmonella recipient strains, in consistent with previous studies [29,30]. This supports the predominance of CTX-M type among the isolates in this study and the spread of these genes among bacterial strains that are naturally shared between humans and animals. The occurrence of ESBL-producing E. coli from stool samples of people in close contact with pigs was previously demonstrated [31]. The finding of this study further suggests that CTX-M type ESBL from E. coli strains isolated from pigs can be transferred to humans and vice versa.
Two randomly-selected transconjugants from each donor/recipient combination were additionally tested for their antimicrobial susceptibilities ( Table 5). All transconjugants were resistant to streptomycin, supporting the positive association that was observed between the occurrence of ESBL-producing E. coli isolates and resistance to streptomycin. However, transconjugants of only one donor/recipient combination exhibited resistance to gentamicin, suggesting that gentamicin-resistance and ESBLs (blaCTX-M-4 and blaCTX-M-14) genes were not located on the same plasmid. The genes are likely to coexist on plasmid in different incompatibility groups.
In conclusion, this study revealed the occurrence of a significant proportion of ESBL-producing E. coli in healthy pigs in Thailand. It also represents one of the first report of CTX-M type ESBL including blaCTX-M-15, blaCTX-M-4 and blaCTX-M-14 in E. coli isolates from healthy pigs in Thailand with CTX-M-15 subtype of CTX-M group I being the most predominant ESBL genotype. It is indicated that pigs serve as reservoirs of ESBL genes that will perpetuate the spread and maintenance of ESBL genes in Thailand. OR > 1 represent positive association and OR < 1 represent negative association. Table 4. Genotypes of β-lactamase-producing Escherichia coli from healthy pigs (n=86).

Bacterial isolates
A total of 292 stock cultures of E. coli isolates collected from fecal samples of clinically healthy pigs were reconfirmed and used in this study. The pig farms were in different provinces in central (n = 104) and northeastern Thailand (n = 188) ( Table 1). All E. coli strains were isolated as previously described [32]. Presumptive E. coli isolates were cultured on MacConkey agar (Difco, Sparks, MD, USA). Five distinct colonies resembling E. coli from each culture plate were confirmed on eosin methylene blue agar (EMB) and by Indole test. A single colony from each positive sample was collected and stored in 20% glycerol at -80°C.

Detection of ESBL-encoding genes
All ESBL-producing E. coli were screened for the presence of nine bla genes including blaTEM, blaSHV, blaCMY-1, blaCMY-2, blaPSE, blaCTX-M group I, II, III and IV using PCR (Table 2). DNA template was whole cell DNA freshly prepared from overnight cultures [20] (Luvsansharav et al. 2012). PCR reaction was performed using GeNei TM mastermix (Merck, Munich, Germany) according to manufacturer's instruction. PCR products were purified using PCR DNA and Gel Band Purification Kit (Nucleospin™; McCherey-Nagel, Duren, Germany) prior to DNA sequencing. The resultant nucleotide sequences were compared to previously established sequences from the GenBank Database using BLAST (http://www.ncbi.nlm. nih.gov/).

Conjugation experiments
Conjugation transfer of resistance genes was tested by biparental mating method described by Lay et al. [12]. All ESBL-producing E. coli (n=86) were used as donors and the spontaneous rifampicin resistant derivatives of Salmonella Enteritidis strain SE12rif R (rifampicin MIC = 256 µg/ml) were used as recipients. Briefly, the overnight culture of the donor and recipient strains were grown in fresh Luria Bertani broth (Difco) at 37 o C to log phase. The donor and recipient culture were mixed at 1:1 ratio (750 µl of each) in a microcentrifuge. The bacterial cells were collected and placed on a 0.45-µmpore-size filter (Millipore, Merck, MS, USA) on LB agar and incubated at 37°C overnight. Transconjugants were selected on LB agar containing rifampicin (32 µg /ml) and ampicillin (100 µg/ml) and confirmed as Salmonella on xylose lysine deoxycholate (XLD) agar (Difco/BD). The conjugation efficiency was estimated as the number of transconjugants per donor cell ( Table 5). Number of transconjugants was calculated by counting the colonies in LB and XLD containing rifampicin (32 µg /ml) and ampicillin (100 µg /ml) plates. All transconjugants were examined for ESBL production and ESBL genes as described above.

Data analysis
Occurrence of ESBL-producing E. coli and AMR was analyzed by descriptive statistics. Pearson's Chi Square test was performed to determine the association between ESBL producers and AMR using SPSS version 22.0. A p-value of < 0.05 was statistically significant. The odds ratios (OR) and their 95% confidence intervals (CI) were also calculated.