Characteristics of extended-spectrum β-lactamase–producing Escherichia coli isolated from fecal samples of piglets with diarrhea in central and southern Taiwan in 2015

Abstracts Background The production of extended-spectrum β-lactamases (ESBLs) confer resistance to the commonly used beta-lactam antimicrobials and ESBL–producing bacteria render treatment difficulty in human and veterinary medicine. ESBL–producing bacteria have emerged in livestock in recent years, which may raise concerns regarding possible transfer of such bacteria through the food chain. The swine industry is important in Taiwan, but investigations regarding the status of ESBL in swine are limited. Results We collected 275 fecal swab samples from piglets with diarrhea in 16 swine farms located in central and southern Taiwan from January to December 2015 and screened them for ESBL–producing Escherichia coli. ESBL producers were confirmed phenotypically by combination disc test and genotypically by polymerase chain reaction and DNA sequencing. The occurrence rate of ESBL–producing E. coli was 19.7% (54 of 275), and all were obtained in swine farms located in southern Taiwan. bla CTX-M-1-group and bla CTX-M-9-group were the two bla CTX-M groups found. bla CTX-M-55 (34 of 54; 63.0%) and bla CTX-M-15 (16 of 54; 29.6%), which belong to the bla CTX-M-1-group, were the two major bla gene types, whereas bla CTX-M-65 was the only type found in the bla CTX-M-9 group. Twenty-seven strains contained bla TEM-1, and the other 27 strains contained bla TEM-116. One strain found in Pingtung harbored three bla genes: bla TEM-116, bla CTX-M-55, and bla CTX-M-65. ESBL–producing E. coli exhibited a multidrug-resistant phenotype, and multilocus sequence typing revealed that the ST10 clonal complexes, including ST10, 167, 44, and 617 accounted for 35% (19 of 54) of these strains. Conclusions ESBL-producing E. coli from piglets with diarrhea were isolated from swine farms located in southern Taiwan. The most commonly detected bla were bla CTX-M-15 and bla CTX-M-55. The ST10 clonal complexes comprised most of our ESBL-producing E. coli strains. Fecal shedding from swine may contaminate the environment, resulting in public health concerns; thus, continued surveillance of ESBL is essential in swine and in other food animals.


Background
Diarrhea is a common clinical syndrome in the swine industry and may be classified into three entities. Sucking piglets usually exhibit neonatal diarrhea a few days after birth, and young piglet diarrhea occurs from the first week after birth to weaning [1]. Older piglets, commonly 2 weeks after weaning, may contract post-weaning diarrhea. Neonatal and postweaning diarrhea are caused by pathogenic Escherichia coli, and causative agents of young piglet diarrhea may include transmissible gastroenteritis virus, rotavirus, coccidia, and E. coli [1]. Occurrence of diarrheal disease can be reduced by the vaccination of sows to let piglets obtain maternal antibodies. However, measures such as the use of antibiotic supplements in feed are also frequently practiced along with vaccination to reduce the incidence of diarrhea. If prudent usage of antibiotics is not taken into consideration, the massive, indiscriminate, and long-term use of antibiotics in veterinary practice may contribute to the selection and spread of drug-resistant bacteria [2,3].
Production of extended-spectrum β-lactamases (ESBLs) confers resistance to the frequently used beta-lactam antimicrobial agents, including the thirdgeneration cephalosporins such as ceftriaxone, ceftazidime, and ceftiofur. However, ESBLs are inhibited by the β-lactamases inhibitors clavulanic acid, sulbactam, and tazobactam [4]. TEM, SHV, and CTX-M-types are the three major families of ESBL [4]. All CTX-Mtypes enzymes are ESBLs, whereas the TEM-and SHV-types of ESBL arise by point mutation at specific residues from the natural TEM-1/TEM-2 and SHV-1 β-lactamase [5]. The production of ESBLs is mainly plasmid mediated, and such plasmids often carry genes that encode resistance to other classes of antimicrobials, such as fluoroquinolones and aminoglycosides [6]. ESBLs are widely distributed in Enterobacteriaceae, particularly in E. coli, and the rapid emergence and spread of ESBL-producing E. coli have been reported in food animals globally [7]. Such findings raise concerns about the possible transfer of ESBL producers through the food chain, thus presenting a hazard to public health [8].
The status of ESBL-producing E. coli in food animals in Taiwan has only been reported in cows [9]. Although a foot and mouth disease outbreak in 1997 had a great impact on the swine industry [10], swine are still among the most important agricultural products in Taiwan. The objective of this study is to analyze the fecal carriage of ESBL-producing E. coli isolated from piglets with diarrhea in 16 pig farms located in central and southern Taiwan. It is important to screen for the ESBL producers from food animals such as swine from a public health perspective.

Sample collection
A total of 275 fecal swab samples were collected from the piglets with diarrhea before weaning from 16 swine farms in Taiwan (one in Taichung, one in Nantou, one in Chunghua, two in Yunlin, four in Chiayi, one in Tainan, and six in Pingtung) from January to December 2015. These 16 farms belong to the same swine industry corporation. Isolating E. coli from feces in piglets with diarrhea and preparing "tailored vaccine" has been routinely practiced in these farms. These E. coli were cultured, inactivated and used as a vaccine component to feed pregnant sows. Neonatal piglets will presumably obtain maternal antibodies when sucking colostrum. Occurrence of neonatal diarrhea due to E. coli infection may be reduced as long as piglets have enough maternal antibody. We shared these fecal swab samples and inoculated on CHROMagar ESBL (CHROMagar, Paris, France) to screen for ESBL-producing E. coli. Any pink colony that appeared on the agar after incubation at 37°C for 16-18 h was initially designated as ESBL-producing E. coli since they were resistant to cefotaxime and/or ceftazidime, and its identity as E. coli was confirmed with the RapID TM ONE System (RapID TM , Lenexa, KS, USA). Confirmed E. coli strains were stored at −80°C for further study.

ESBL testing
E. coli isolates were tested phenotypically for ESBL production by combination disc tests with cefotaxime and ceftazidime (30 μg), with and without clavulanic acid (10 μg), as stated by the guidelines of the Clinical and Laboratory Standards Institute [11]. The tested E. coli strains were plated on Muller-Hinton agar at a concentration of 0.5 McFarland standards and grown at 35°C for 16-18 h. A difference of 5 mm or more in the inhibition zones for at least one cefotaxime or ceftazidime/ clavulanic acid combination versus the corresponding cefotaxime or ceftazidime alone was used to define an ESBL producer. Klebsiella pneumoniae ATCC 700603 and E. coli ATCC 25922 were used as the positive and negative controls, respectively [11].

Detection of bla genes
The E. coli strains that were phenotypically confirmed to be ESBL producers were examined with polymerase chain reaction (PCR) to detect their bla genes. The tested strains were cultured for 16-18 h at 37°C on tryptic soy agar plates (Difco/Becton Dickinson, Franklin Lakes, NJ, USA), and a loopful of bacterial cells was resuspended in 200 μL ddH 2 O and boiled for 10 min [12]. After centrifugation at 12,000 g for 10 min, the supernatant was saved as the source of template DNA for PCR. The primer sequences used to amplify bla TEM , bla SHV , bla CTX-M-1-group , bla CTX-M-2-group , bla CTX-M-8group , bla CTX-M-9-group , and bla CTX-M-25-group , the annealing temperature, and the expected PCR product sizes are specified in Table 1. The PCR cycling program was set as follows using a LifeEco thermocycler (Bioer Technology, Hangzhou, China): initial denaturation at 95°C for 5 min, followed by 35 cycles at 95°C for 30 s, then the annealing temperature specified in Table 1 for 40 s, and 72°C extension for 1 min. The reaction was then maintained at 72°C for 10 min. Ten microliters of each PCR sample was loaded onto a 1.2% agarose gel and electrophoresed at 100 volts for 40 min. The gels were then stained with a fluorescent nucleic acid dye (Biotium, Hayward, CA, USA) and examined under a blue light LED illuminator (Smobio, Hsinchu City, Taiwan). The PCR products were then sliced from the agarose gel and subjected to further purification and sequenced by ABI 3130 x1 Genetic Analyzer (Applied Biosystems, Foster, CA, USA) in Center for Genomic Medicine, National Cheng Kung University, Tainan, Taiwan. The results were analyzed with MEGA 6.0 and examined with the NCBI BLAST program (http://www.ncbi.nlm.nih.gov/ blast/) and β-lactamase database (http://www.ncbi.nlm.nih.gov/pathogens/submit-beta-lactamase).

Results
Fifty-four samples exhibited pink colonies on CHROMagar ESBL, initially indicating an identity of ESBL-producing E. coli. All of these strains were then confirmed biochemically with RapID TM ONE System as E. coli, and they were not hemolytic when grown on blood agar. These 54 strains exhibited the ESBL phenotype when assayed by combination disc tests. From our results, we did not detect any ESBL-producing E. coli in diseased piglets from any of the five swine farms in Taichung, Nantou, Chunghua, and Yunlin, which were located in central Taiwan. ESBL-producing E. coli were all obtained in swine farms in southern Taiwan, including Chiayi, Tainan, and Pingtung, with the exception of one farm in Chiayi (farm ID CY-3) and one in Pingtung (farm ID PT-3). Geographic distribution of the swine farms and occurrence of ESBL were indicated in Fig. 1. Overall, the occurrence rate of ESBL-producing E. coli was 19.7% (54 of 275). Table 2 lists the occurrence of ESBL-producing E. coli in 16 swine farms.

Discussion
The ESBL-producing E. coli were all obtained from the swine farms located in southern Taiwan. There were five swine farms in central Taiwan (Taichung, Changhua, Yunlin, and Nantou) that participated in our study, and only 43 fecal samples (43 of 275; 15.6%) were collected from piglets with diarrhea and screened for ESBL. Although the scale of these farms was similar to that of those located in southern Taiwan, the hygienic procedures or disease control management of individual farms may contribute to such differences in diarrheal cases. The specificity of this chromogenic agar was 100% because all of the pink colonies, indicative of ESBLproducing E. coli, were phenotypically and genotypically positive for ESBL. A previous report also suggested the high sensitivity and specificity of CHROMagar ESBL in the detection of clinical ESBL-producing Enterobacteriaceae [14]. However, our strategy may also lose some ESBL producers that could grow on blood agar or Mac-Conkey agar but not on CHROMagar ESBL.
The occurrence of ESBL-producing E. coli in food animals has been increasing around the world [2]. For example, more than 40% of the ESBL-producing E. coli were detected from piglets with post-weaning diarrhea in Heilongjiang Province, China [15]. The authors also compared their findings with those of a similar study in healthy pigs in China and concluded that ESBLproducing E. coli were more commonly found in sick animals [16]. Because we did not investigate the prevalence of ESBL in a healthy swine population, there was no basis of comparison for healthy and diseased swine in Taiwan. Although diseased pigs are not likely to enter slaughter or market, fecal shedding from such pigs can contaminate the piggery environment and provide a reservoir for the exchange of drug-resistance genes [17].
TEM-116-producing E. coli was first identified in Korean hospitals in a nationwide survey in 2002 [18]. Consequently, a high prevalence of TEM-116 was also reported in Spain [19]. In animals, TEM-116-producing E. coli has been detected in dogs [20,21]. Our results, to the best of our knowledge, demonstrate for the first time the presence of bla TEM-116 genes in the ESBL-producing E. coli from porcine origin. Although most of our TEM-116-containing strains also had CTX-M, we did find four E. coli strains that harbored only TEM-116 that exhibited an ESBL phenotype. ESBL producers within the CTX-M group are becoming more common [22,23]. In Europe, CTX-M-1 is broadly disseminated in animals, whereas CTX-M-14 is most prevalent in animals in Asian countries [8]. The most frequently found CTX-M types in our study were CTX-M-15 and CTX-M-55, whereas CTX-M-14 was reported in healthy and Table 3 The bla genes and sequence type of ESBL-producing E. coli    [15,16,24]. CTX-M-55 was first isolated from patients in a hospital in Thailand; this novel CTX-M type was derived from CTX-M-15, with only a single substitution of valine instead of alanine at residue 77 [25]. The incidence of CTX-M-55 has been reported to exceed that of CTX-M-15 in outpatient infection cases in Chinese county hospitals [26]. Thus, the authors of that study hypothesized an animal-human transfer of CTX-M-55 because most of these outpatients in county hospitals live in rural areas and thus have more chances to come into contact with infected food animals and farm sewage [26]. In our study, CTX-M-55 was the predominant ESBL type isolated from swine with diarrhea in southern Taiwan. CTX-M-55-producing E. coli has also been detected in the milk of cows with clinical mastitis in the same region [9]. It is conceivable that ESBL-producing E.  [15,24,27]. These findings underscore the importance of screening and investigation of the genotypes of ESBL producers in food animals on a regular basis. Our investigation did not detect any CTX-M-8 or CTX-M-25, which were also not detected in previous studies [15,28]. Antimicrobial susceptibility test revealed a high frequency of the resistance of ESBL-producing E. coli to most antimicrobial agents. Inappropriate use or overuse of antimicrobial agents, including third-generation cephalosporins, may be associated with the emergence of ESBL-producing E. coli in swine [29]. The selection of CTX-M-producing E. coli in swine by treatment with ceftiofur has been documented [30]. It is worthwhile to consider banning the use of third-generation of cephalosporins such as ceftiofur in food animals to decrease the occurrence of ESBL producers. For example, the occurrence of ESBL-producing E. coli was reduced when third-generation cephalosporins were banned in the Danish pig industry [31]. Forty-two percent of ESBLproducing E. coli were resistant to amoxicillin/clavulanic acid. Possible reasons that account for this phenotype may include hyper production of chromosomal class C β-lactamase, possession of plasmid-mediated TEM enzymes, production of oxacillinases, or production of inhibitor-resistant TEM by these isolates [32]. In addition, plasmid mediated AmpC may also cause resistance to amoxicillin/clavulanic acid [8].
The ST10 clonal complexes (ST10, 167, 44, 617) comprised most of our ESBL-producing E. coli strains. There were 13 strains that did not match any ST in the current database; however, six of these had only a one to threeallele difference from ST10 clonal complexes. It is fair to say that ST10 was the dominant clonal complex in our study. A recent investigation indicated that ESBLproducing E. coli were commonly isolated from river waters in southern Taiwan and that ST10 and ST58 was the most frequently found clonal complexes [33]. The authors of that study also observed a substantial association of these ESBL-producing E. coli with the presence of chicken farms at that region. Geographically, food animal farms, including swine, chicken, and cattle, are primarily situated in southern Taiwan. It is conceivable that livestock may spread ESBL-producing E. coli from feces, thus contaminating the environment. We did not detect ESBL-producing E. coli ST131 (O25:H4) that possessed CTX-M-15, a leading cause of urinary tract infections and bacteremia in human medicine globally, in our study. However, swine and other food animals may play a role as vectors in the transmission of bacteria to humans [34], so continued surveillance of food animals for ESBL-producing E. coli is essential.
Our study has some limitations. Fecal samples from healthy piglets were not collected and there was no comparison for the occurrence of ESBL-producing E. coli between the healthy and diseased populations. The virulence factors like K88, K99 or 987 P fimbriae genes in our ESBL-producing E. coli isolates were not screened and they were not hemolytic when grown on blood agar. It is possible that the ESBL-producing E. coli in the present study was not the causative agent for the diarrhea of these piglets. We did not detect if these E. coli isolates produced AmpC-β-lactamases, which also hydrolyze third-generation cephalosporins. AmpCproducing E. coli were also found in increasing numbers in food-producing animals [8]. In addition, profiles of resistant plasmids were not characterized in our study. Plasmid analysis methods like PCR-based replicon typing could assign the incompatibility (Inc) groups [35], whereas replicon sequence typing could discriminate IncF plasmid variants [36]. Inclusion of plasmid characterization could have provided insights into the epidemiology of the ESBL plasmid in our study.

Conclusions
ESBL-producing E. coli from piglets with diarrhea were isolated from swine farms located in Chiayi, Tainan, and Pingtung. bla CTX-M-15 and bla CTX-M-55 were the most commonly detected bla genes. The ST10 clonal complexes comprised most of our ESBL-producing E. coli strains. Fecal shedding from swine may contaminate the environment, from a public health perspective, continued surveillance of ESBL is essential in swine and in other food animals.