Discussion
In total, we examined 108 rectal swabs originating from a pool of six species of free-living animals and noted positive growth of E. coli on the selective MacConkey agar medium in over 50% of the samples. This is a lower percentage than that in a group of wildlife predators analysed previously, in which we observed resistance to at least one drug in approx. 70% of the tested animals (Osińska et al., 2020). However, the lower number of resistant isolates in herbivores and omnivores compared to carnivores is consistent with studies in mammals (Nhung et al., 2015) and birds (Smith, Wang, Fanning, & McMahon, 2014). These differences are related to the type of food consumed by these groups of animals. Carnivores and predators are at the end of the food chain and have more diverse food than herbivores, cumulating all potential resistance determinants from prey. In the case of free-living herbivores, the source of resistant bacteria is mainly assigned to the polluted environment, including plants as a source of food, water, and the ecological niche shared with other animal species (direct and indirect contact).
Using several types of plates supplemented with various antimicrobial substances, we were able to obtain up to 80 bacterial isolates from 55 single animals and, finally, using a two-step method, limit this number to 70 E. coli strains that are distinct in terms of their phenotype, genotype, or origin. This means that every third individual of the resistant E. coli -positive animals carries more than one strain, sometimes even three completely separate MDR clones, which may indicate significantly reduced resistance results when analysed according to the strategy “one randomly selected strain isolated from one animal”. We also isolated a large number of multidrug-resistant strains, estimated at over 70%, which confirms that regardless the species of host, type of diet, and behaviour, free-living animals are involved in the spread of resistant E. coli strains in Poland despite the lack of direct exposure to antimicrobials (Osińska et al., 2020; Nowakiewicz et al., 2020).
This fact is additionally confirmed by the finding of identical strain clones in two different pairs of the red deer. Unfortunately, we do not have information whether these animals were related to each other or occupied the same area, but this does not change the fact that the same strain was isolated from two different animals even in two cases. Red deer and roe deer live in herds; therefore, close contact may promote a horizontal transfer of resistance genes among animals sharing the same ecological niche (Dolejska & Literak, 2019). It is also probable that the statistically highest resistance to as many as four different antimicrobials and the multidrug resistance of all the roe deer strains may result from the herd lifestyle (swarm behaviour).
The high resistance to ampicillin, tetracycline, and sulfamethoxazole is also consistent with our previous studies on predatory animals (Osińska et al., 2020). This is also a typical result comparable to those reported in studies of livestock animals in the world (Nhung et al., 2015; de Alcântara Rodrigues, Ferrari, Panzenhagen, Mano, & Conte-Junior, 2020) and in the same wildlife animal or species groups: ungulates (Wasyl et al., 2018), squirrels (Jalal et al., 2019), and rats (Nkogwe, Raletobana, Stewart-Johnson, Suepaul, & Adesiyun, 2011; Himsworth et al., 2016).
We also recorded a high level of strain resistance to streptomycin and chloramphenicol, almost identical to that in other wildlife animals in Poland (Nowakiewicz et al., 2020; Osińska et al., 2020). As in the previous study, we found only one gentamycin-resistant strain, which confirms that resistance to gentamicin is rather sporadic among E. coli isolates from wild animals.
In contrast to our previous studies, lower kanamycin resistance was detected despite the use of this antibiotic for selective isolation of the strains.
The phenotypic resistance of the strains was reflected in most cases in the presence of at least one gene determining resistance to a given antimicrobial. As with E. coli strains isolated from other wildlife or domestic animals and strains isolated from nosocomial infections, the gene profile was very similar, with the predominance of the tetA (tetracycline-resistant strains) and sul 2 or bothsul 1 and sul 2 (sulfamethoxazole-resistant strains) (de Alcântara Rodrigues, Ferrari, Panzenhagen, Mano, & Conte-Junior, 2020; Hassan et al., 2020).
The nearly 40% level of chloramphenicol resistance was slightly higher than that recorded previously in wildlife animal studies in Poland (Osińska et al., 2020; Nowakiewicz et al., 2020) and as many as 59% of the chloramphenicol-resistant strains exhibited the presence of theflo R gene responsible for both chloramphenicol and florfenicol resistance. The latter antimicrobial is authorized for use in animals also in Poland; therefore, such a high level of resistance to chloramphenicol, although the drug is not approved for use in animals, is likely to be associated with cross-resistance following exposure of animals to florfenicol, residues of which were detected in environmental samples, mainly in water (Hanna et al., 2018). Moreover, florfenicol was considered as persistent against some water-treatment processes (Charuaud et al., 2019). The ESBL phenotype strain had thebla CTX-M -27 gene. This variant has become one of the most common contributors to the spread of blaCTX-M in humans in recent years (Cormier et al., 2019). Currently, the producer of the MDR, CTX-M-27, ST 131 C1 cluster is considered as a new epidemic clone (Fernandes et al., 2020). Although we did not analyse the strains with the MLST method, it is worth noting that the only strain with the ESBL phenotype isolated from the squirrel was also resistant to the largest number of drug groups (up to seven) among all strains tested similarly as the aforementioned cluster.
Virulence factors are an inherent attribute of pathogenic E. colistrains, unlike commensal strains, which naturally colonize the digestive tract of humans and animals without causing clinical symptoms. As many as 11.4% of the 70 strains with resistance to at least one antimicrobial exhibited the presence of specific virulence factors, qualifying them to specific pathotypes. However, the percentage of simultaneously pathogenic and drug-resistant strains seems to be lower than in other groups of free-living animals or other food-producing animals from which strains were isolated randomly and not selected for resistance (Nowak et al., 2017; Sarowska et al., 2019). Thus, a high level of resistance is not necessarily accompanied by increased virulence as in our study. We found mainly strains with the ExPEC pathotype. These strains were isolated from the squirrels, rat, and hedgehog samples. Due to the diverse and rich array of virulence factors, it is difficult to determine clearly the origin of ExPEC causing infections in humans. However, all the ExPEC strains in our study were characterized by the presence of genes encoding fimbrial adhesins (P, S, and F1C fimbriae) involved in adhesion to epithelial cells in the intestine, kidney, bladder, and lower urinary tract (Kuhnert, Boerlin, & Frey, 2000; Sarowska et al., 2019) as well as the synthesis of capsular polysaccharide with antiphagocytic activity and ferric aerobactin receptor involved in iron uptake transport (Chapman et al., 2006; Kuhnert, Boerlin, & Frey, 2000). All these factors were most often detected in UPEC or NMEC strains causing human infections (Sarowska et al., 2019). However, a similar panel of virulence factors was shown in strains from livestock (Bélanger et al., 2011; Wasiński, 2019) and ExPEC strains were isolated from pigs and cattle, where they were responsible for urinary tract infections (Chapman et al., 2006), pneumonia, mastitis, and meningitis (Tan et al., 2011). This pathotype was also found in faecal microbiota of cats and dogs. It was shown that dogs and cats in the same household can be carriers of the same UPEC strains (Johnson, Miller, Johnston, Clabots, & DebRoy, 2009). This indicates that the strains can overcome the species barrier and spread between animals and humans (Bélanger et al., 2011) (Johnson et al., 2008) (Johnson, Miller, Johnston, Clabots, & DebRoy, 2009).
Among other wild animal species, wild birds are the most widely studied group in terms of virulence (Borges et al., 2017). Borges et al. (2017) discovered the same virulence genes characteristic of APEC pathotype strains that were also found in poultry, wild mammals (Chapman et al., 2006), and humans (Maluta et al., 2014). This phenomenon suggests that wild birds may also transfer virulence genes to ExPEC strains in other hosts (Velhner, Suvajdžić, Todorović, Milanov, & Kozoderović, 2018) therefore, the virulence factors do not appear to be in any way correlated with the specific source of the strain.
We also found two isolates with the ETEC pathotype among samples collected from the squirrels. Enterotoxigenic E. coli strains are commonly isolated from farm animals and are responsible for diarrhoea, especially in young animals. Since ETEC strains have a narrow host species range due to specific adhesins, they are regarded as strains posing a lower risk to public health. Moreover, enterotoxins are encoded on plasmids, and ETEC strains are characterized by a large genotypic diversity making it difficult to determine their original source (Nagy & Fekete, 2005). In addition to specific adhesins, ETEC strains can produce two main classes of toxins: heat-stabile (ST) and heat-labile (LT) (Dubreuil, Isaacson, & Schifferli, 2016). Our strains contained one or two genes responsible for producing ST enterotoxins that are characteristic for ETEC of human, porcine, and calf hosts. Among livestock, ETECs are often isolated from pigs. The pig environment is one of the transmission pathways for ETEC strains. It has been confirmed that ETECs can persist for at least six months when they are protected by manure (Dubreuil, Isaacson, & Schifferli, 2016). In cattle, typical ETEC strains only produce the STa toxin (Nagy & Fekete, 2005). This is similar in dogs, but a small proportion of the strains were found to produce the STb toxin as well (Dubreuil, Isaacson, & Schifferli, 2016). The ETEC pathotype is very rarely isolated from wildlife animals and, due to the low percentage of isolation and insufficient research, the role of these strains in the aetiology of diarrhoea in this group of animals is unknown (Milton et al., 2019).
We also revealed two E. coli strains with the EHEC pathotype present in the squirrel and red deer samples; both of them produced Shigatoxin 2. Shiga-toxin-producing strains (STEC) are a large class ofE. coli pathogenic strains and are commonly found in farm and wild animals as well as humans. In humans, infection is mainly caused by the consumption of bacterium-contaminated food or contact with contaminated faeces (Milton et al., 2019). Cattle are the main reservoir of EHEC strains in animals. Unlike humans, EHECs found in adult cattle do not cause disease symptoms but may trigger life-threatening conditions in young animals (Rice, Sheng, Wynia, & Hovde, 2003). In our research, we only detected the presence of the stx2 gene, but thestx1 gene or both are detected equally often in animal-origin strains (Milton et al., 2019). Stx1 or both genes dominated in cattle (Tavakoli & Pourtaghi, 2017). Among wild animals, thestx2 gene is detected more frequently in birds (Sanches et al., 2017) and ungulates (Milton et al., 2019). In contrast, in iguanas, strains producing Shiga-toxins accounted for approx. 40% of pathotypes (Bautista-Trujillo et al., 2020).