Presence and characterization of Escherichia coli virulence genes isolated from diseased pigs in the central region of Argentina

Background: The main pathogen of neonatal and post weaning diarrhea and edema disease (ED) is Escherichia coli and pathotypes involved are enterotoxigenic, enteropathogenic, and shiga toxigenic (ETEC, EPEC, and STEC, respectively). Those diseases cause economic loss in pig production. Aim: The aim of this work was to evaluate the presence of strains expressing virulence markers genes and the antibiotic susceptibility profiles of E. coli from clinical cases of post weaning diarrhea and ED in farms in the central area of Argentina. Materials and Methods: Intensive pig farms from the central region of Argentina were sampled. Intestinal mucosa swabs from pigs with diarrhea were taken, seeded on MacConkey agar plates, biochemically typified and tested by polymerase chain reaction (PCR). Antibiograms were made by disk-diffusion method. Results: A total of 54 strains from clinical cases studied showed PCR findings: 88.88% (48/54) expressed at least one gene coding for a virulence factor. Colonization factors found were: 39.58% of strains had F18, 33.33% were F4 and 31.25% adhesin involved in diffuse adherence-I; 29.17%, 25%, and 2.1% expressed LT, STb, and STa, respectively. 25% were STx and 16.67% were eae positive. Only 2.1% were STx2. The most active antibiotics against most strains were gentamicin and ceftiofur, but resistance profiles against many antibiotics were found. Conclusion: High circulation of pathogens strains of E. coli among pigs with diarrhea with an extended antibiotic resistance profile.

ETEC enterocyte adhesion is mediated by fimbriae (F) such as F4 (K88), F5 (K99), F6 (P987), F18, and F41 and is capable of producing enterotoxins which act on enterocytes. According to their thermal stability, enterotoxins are classified as heat-stable (STa, STb, and EAST1) and heat-labile (LTI and LTII) [5]. STEC strains produce shiga toxin (STx1 and STx2) and an adhesion protein called intimin (encoded by the eae gene) which is responsible for microvilli attaching and effacing on enterocytes surface [6], this protein is also important in EPEC pathotype. Some strains expressing F18 and the adhesin involved in diffuse adherence (AIDA) are associated to ED and produce STx2e toxin [2,7,8]. Other strains compatible with EPEC produce intimin without expression of any other associated toxins but are capable of inducing a typical lesion of attaching and effacing microvilli on enterocytes surface [9].
Intensification of the porcine industry has led producers to carry out widespread feeding practices such as incorporation of antibiotics to feed, as a disease therapeutic strategy or as growth promoters [11]. Many antibiotics used for pig production are also important in human medicine. Dissemination of antibiotic resistance to humans might ensue from these practices [12,13]. Therefore, it is very important to know which the antibiotics resistance profiles are prevailing in the farm's E. coli strains, to prevent usage of those antibiotics which are also used in human medicine but are still active against the pathogen. Authors from different countries reported the prevalence of E. coli pathotypes for different categories of pigs [14][15][16][17][18]. However, not enough information about confirmatory diagnosis of virulent strains of E. coli from clinical cases is available in Argentina. Knowledge of E. coli pathotypes infecting the farms plus antibiotic susceptibility profiles is very important tools for pig producers to build solid therapeutic and preventive strategies.
The central aim of this work was to evaluate the presence of strains expressing virulence markers genes and the antibiotic susceptibility profiles of E. coli from clinical cases of post weaning diarrhea and ED occurred in farms in the central area of Argentina.

Ethical approval
In order to guarantee a safe, correct and carefully handling of pigs, authors proceeded according to specifications of international Ethic Guidelines (Internal Ethical Committee CICUAE Res. 533/16). However, the procedures performed do not cause discomfort to animals.

Farms
Intensive pig production farms in the central region of Argentina (states of Córdoba, Santa Fe, Buenos Aires, Entre Ríos and San Luis, the country's most important production area) were sampled according to the clinical cases emergency. All farms included in this study had the following characteristics: High health status, 200-500 sows, automatized feed systems, parturition control, reduced waste contact, controlled environmental temperature, and permanent veterinarian assistance.

Bacterial isolates processing
Intestinal mucosa swabs samples were collected from clinical cases of piglets with post weaning diarrhea and/or with ED. Samples were inoculated on MacConkey agar plates (Britannia) and 5% equine blood agar plates (Britannia) to check β-hemolysis. After 24 h at 37°C Gram-staining was performed on organisms of emerging colonies to corroborate bacterial morphology. For biochemical characterization [15] hydrogen Sulfide-Indole-Motility (SIM) (Oxoid), Methyl Red-Voges Proskauer (MR-VP) (Britannia), citrate (Britannia), amino acids decarboxylation (arginine, lysine, and ornithine) (Britannia), oxidase (Britannia), catalase, three sugar iron (Oxoid) tests were carried out. All E. coli isolates were lyophilized until molecular characterization was made.

Molecular characterization
A total of 54 E. coli isolates were tested by polymerase chain reaction (PCR). DNA was obtained using a commercial kit (Fermentas) according to manufacturer instructions for Gram-negative bacteria. These extracts were used as reaction templates. Fimbriae and toxins genes primers studied were: STa, STb, LT, F4, F18, VT1 (STx1), VT2 (STx2), eae, EAST1, AIDA-I, and Rfb (O157) [1,15], sequences are listed in Table- Reactions were subjected to one initial denaturalization cycle of 94°C, followed by 30 cycles of 30 s at 94°C (denaturalization), 30 s at 58°C (primers annealing), 30 s at 72°C (extension); finally 2 min at 72°C (final extension). Reference strains of E. coli were used as pathotype and positive control reaction [1] ( Table-1).  AIDA=Adhesin involved in diffuse adherence A standard electrophoresis in 2% of agarose in tris-acetate EDTA buffer stained with ethidium bromide performed at 100 V during 30 min was used to reveal the amplified products. Visualization of amplicons was made under transilluminator (UV -λ 300 nm), and a 100 pb ladder was used. Gels were photographed with a Kodak Easy Share Z7590 camera system and evaluated with Kodak Digital Science 1D software [1].

Statistical analysis
To evaluate associations between variables (genes, clinical signs, states, and season) contingency tables were used, through exact Fisher's test. When statistical association was revealed, multivariate test "multiple correspondence analysis" was applied to know levels of variable relation. SAS version 9.2 software and InfoStat version 2015 [21] were employed.

Antimicrobial profiles
Table-4 summarizes the results of antibiograms. According to these findings, the most suitable antibiotics for a therapeutic approach were GEN and CTF.
The correspondence between isolates and antibiotics resistance was shown in Figure-1 (Table-5). Six groups of strains were observed (50% of distance), three of each included one strain distanced from the other three groups. Among these groups, one of each included two principal clusters with 23 strains; the other group was formed by two clusters with two strains each and finally one group with one strain.

Discussion
This work revealed the frequency of pathotypes of E. coli isolated from post weaning diarrhea and ED cases in swine from the central region of Argentina, and their relation to antimicrobial susceptibility. We found combinations of virulence factors strongly associated with severity of disease [1][2][3]14]: F4/ EAST1 (16.7% of strains), STb/LT/F18/Stx1/East1/ AIDA-I (8.3%), and STb/LT/F4/EAST1 (6.3%); similar findings were reported by Moredo et al. [1,14]. However, strains expressing only one virulence factor were also found on diseased pigs, therefore those  virulence factors could also be regarded as important virulence markers (e.g., EAST1, STa, F4, F18, Stx1, Stx2, and eae). Limited information about E. coli pathotypes is available in our country and it is not updated [1,14,15,22]. The results presented do not agree with those informed by Parma et al. [22], who reported not having found characteristic virulence genes of ETEC and STEC strains among their isolates. However, Moredo et al. [14] showed similar findings to ours, but in pigs without clinical signs. Genes reported by Moredo et al. were STx1 (STEC) and EAST1 (ETEC), they did not find Rfb O157, which was found in our work. Related to colonization factors and toxins F18 (40.4%), F4 (4.2%), STb (25%), and STa (2.1%), respectively, Moredo et al. [1] reported frequencies similar to ours, but we found differences regarding LT, which we found more frequently. These genes coding for important virulence factors have high pathogenic potential and may trigger piglet diarrhea, under adequate conditions. Therefore, the absence of disease reported by Moredo et al. might be related to the application of adequate environmental, hygienic, and sanitary measures [23]. Another difference was in gene coding for intimine (eae), not found by Moredo et al. [14], but detected (16.67%) in this work. In agreement with Choi et al. [24] EAST1 was the most frequent virulence factor found both individually or in association with F4. STb and LT were genes found without any association by Choi et al. [24], but contrarily to that report, in our work STb and LT were found related to EAST1. A high proportion of EAST1-STa in pigs with diarrhea was detected [25], but we only found one strain with this association (Table-2). We found F4 and F18 in both diseases and LT in post weaning disease (PD) and one case of edema disease (ED).
The shiga toxins (STx1, STx2 and its variants), intimine protein and the enterohemolysin are major virulence factors of STEC [26]. Although no O157 gene was found, we obtained a prevalence of 25% for STx1, 16.67% for eae (gene encoding intimine protein), and 2.1% for STx2. This fact suggests that pigs with post weaning diarrhea or ED could be reservoirs of pathogenic strains of E. coli [9]. Pigs carrying hemolytic uremic syndrome (HUS)strains  could potentially generate an outbreak of that disease not only by possible environmental contamination but also because of participation of pork in the food chain. Fortunately, such strains were not found in this study.
Bacteriological studies of clinical cases of diarrhea in swine production should be standard practice. However, in developing countries, diarrhea treatment is frequently intuitive or empiric with concomitant multiresistance dissemination [20]. Our results showed important antibiotic resistant profiles among isolates from sick animals such as Okello et al. [12], Luppi et al. [13], and Mathew et al. [11]. GEN and CTF elicited low amounts of resistant strains in accordance with Moredo et al. [1], who found a few antibiotics inactive for ETEC strains from healthy pigs. Dendrograms are useful tools to evaluate distinct drugs or groups of antibiotic resistance in a same pathotype of bacteria.

Conclusion
In conclusion, we provide information about an important circulation of strains and pathotypes of E. coli among pigs with diarrhea. The strains studied showed high resistance profiles, these results could be useful for swine industry.