Bacteria from Animals as a Pool of Antimicrobial Resistance Genes

Antimicrobial agents are used in both veterinary and human medicine. The intensive use of antimicrobials in animals may promote the fixation of antimicrobial resistance genes in bacteria, which may be zoonotic or capable to transfer these genes to human-adapted pathogens or to human gut microbiota via direct contact, food or the environment. This review summarizes the current knowledge of the use of antimicrobial agents in animal health and explores the role of bacteria from animals as a pool of antimicrobial resistance genes for human bacteria. This review focused in relevant examples within the ESC(K)APE (Enterococcus faecium, Staphylococcus aureus, Clostridium difficile (Klebsiella pneumoniae), Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacteriaceae) group of bacterial pathogens that are the leading cause of nosocomial infections throughout the world.


Introduction
The discovery of antimicrobial agents in the mid-twentieth century revolutionized the management and therapy of bacterial infections. Infections that would normally have been fatal became curable. Ever since then, the antimicrobial agents have saved the lives of millions of people. However, these gains are now seriously jeopardized by the rapid emergence and spread of antimicrobial-resistant bacteria [1]. Antimicrobial resistance (AMR) is a major health problem rapidly spreading across the world. The Review on Antimicrobial Resistance report [2] estimates that at least 700,000 annual deaths are due to infections by drug-resistant strains of common bacterial infections, human immunodeficiency virus (HIV), tuberculosis and malaria. Numbers suggested that up to 50,000 lives are lost each year due to antibiotic-resistant infections in Europe and the US alone [2]. The inappropriate use of antibiotics in food animals, as well as in the medical practice has potentiated the risk of untreatable infections. Due to the free movement of people and goods between countries, and the intensive international transport of livestock, the problem of AMR is becoming by nature a global problem. Moreover, the AMR emergence is accompanied with a decline in the discovery of new antimicrobial agents. It has been estimated that most of the antibiotics used presently for common human and animal infections will be useless within five to ten years, turning back the clock to the pre-antibiotic era [1].
Antimicrobial agents are principally used for therapy and prevention of human and animal diseases, but they are still used in some countries for growth-promotion in food animal productions [3]. Interactions between groups. Antimicrobial-resistant bacteria can spread to humans either by the food supply, direct contact with food or companion animals or, more indirectly, through environmental pathways, including waterways, soils and vegetables contaminated with human or animals waste, and vectors such as rodents, insects, and birds. Based on da Costa et al. [8] and McEwen et al. [9] with modifications.
The antimicrobial use in animals selects for AMR in commensal and zoonotic bacteria [9]. Soil treated with manure represents a "hot spot" of bacteria carrying AMR-genes [10]. However, soil itself is also a natural reservoir for antimicrobial-resistant bacteria [10]. The fecal wastes from animals contaminate groundwater, streams and other waterways, contributing to the spread of bacteria carrying AMR-genes [9]. Human wastes from homes, hospitals and offices also contribute to contaminate rivers and waterways with antimicrobial-resistant bacteria [9]. In fact, treated wastewater and lake water have been shown to contain AMR-genes and antimicrobial-resistant bacteria [10]. Soils and irrigation water are contamination sources for vegetables and fruits, in which resistant bacteria have been detected [10]. Antimicrobial-resistant bacteria may also spread between Figure 1. Interactions between groups. Antimicrobial-resistant bacteria can spread to humans either by the food supply, direct contact with food or companion animals or, more indirectly, through environmental pathways, including waterways, soils and vegetables contaminated with human or animals waste, and vectors such as rodents, insects, and birds. Based on da Costa et al. [8] and McEwen et al. [9] with modifications.
The antimicrobial use in animals selects for AMR in commensal and zoonotic bacteria [9]. Soil treated with manure represents a "hot spot" of bacteria carrying AMR-genes [10]. However, soil itself is also a natural reservoir for antimicrobial-resistant bacteria [10]. The fecal wastes from animals contaminate groundwater, streams and other waterways, contributing to the spread of bacteria carrying AMR-genes [9]. Human wastes from homes, hospitals and offices also contribute to contaminate rivers and waterways with antimicrobial-resistant bacteria [9]. In fact, treated wastewater and lake water have been shown to contain AMR-genes and antimicrobial-resistant bacteria [10]. Soils and irrigation water are contamination sources for vegetables and fruits, in which resistant bacteria have been detected [10]. Antimicrobial-resistant bacteria may also spread between farms via infected carrier animals, companion animals or wildlife vectors [9]. Finally, there is a flow of patients and bacteria between community and hospital environments ( Figure 1).
These complex transmission routes within farm animals, between farm animals and humans and the transfer of AMR-genes among bacteria, make it challenging to prove whether a reservoir of AMR-genes in livestock poses a risk for animal or human health [10]. However, some antimicrobial-resistant bacteria are zoonotic agents or can colonize and/or infect several hosts. In this sense, the current approach to evaluate the reservoir of AMR-genes in farm animals is to study the AMR-level of commensal bacteria and zoonotic agents in healthy farm animals and slaughter [10]. Reports from the European Food Safety Authority (EFSA) and the European Centre for Disease Prevention and Control (ECDC) monitoring AMR in animals are increasing. Yet, there are some limitations in the current data, and Thanner et al. [10] have recently suggested a voluntary monitoring program by researchers.
In order to underline the importance of the current available antimicrobials classes, the World Health Organization (WHO) started categorizing the most important antimicrobials in human medicine. The last revision of the list was done in 2016 [11]. The importance of each antimicrobial group is based on two criteria: C1, "the antimicrobial class is the sole, or one of limited available therapies, to treat serious bacterial infections in people"; and C2, "the antimicrobial class is used to treat infections in people caused by either: (i) bacteria that may be transmitted to humans from nonhuman sources, or (ii) bacteria that may acquire resistance genes from non-human sources". Antimicrobials that meet both criteria are considered "critically important" in human medicine, antimicrobials that meet one of the criteria are considered "highly important", and antimicrobials that meet none of the two criteria are considered "important" [11].
Similarly to the WHO, the World Organization for Animal Health (also named Organisation mondiale de la santé animale, OIE) [12] has developed a list of the antimicrobial agents of veterinary importance [13]. Since in veterinary medicine, many different species have to be treated the criteria to classify the antimicrobials were different than for the human medicine. The OIE criteria were based on a questionnaire prepared by the ad hoc group, which was sent to the OIE delegates of all member countries and international organizations which had signed a co-operation agreement with the OIE. The responses were analyzed by the ad hoc group and discussed in some international committees. The criterion C1 was based on the response rate to the questionnaire: "This criterion was met when a majority of the respondents (more than 50%) identified the importance of the antimicrobial class in their response to the questionnaire". The criterion C2 was based on the treatment of each serious animal disease and the availability of alternative antimicrobial agents: "This criterion was met when compounds within the class were identified as essential against specific infections and there was a lack of sufficient therapeutic alternatives". Similarly to the WHO list, antimicrobials that meet both criteria are considered "critically important" in veterinary medicine, antimicrobials that meet one of the criteria are considered "highly important", and antimicrobials that meet none of the two criteria are considered "important".
After the ban of antimicrobial growth promoters, antimicrobial agents are still allowed with veterinary prescription [10]. Around 37% of the antimicrobials (mainly ionophores) used in food animal production do not have equivalent drugs used for human therapeutic purposes [14]. Similarly, tetracyclines, that are not considered a first-line antimicrobial therapy in human medicine, make up another 44% of total antimicrobial used in animal agriculture [14]. While not all antimicrobial agents used in animal health are used in human medicine, most antimicrobials used in food animals are analogs to those used in human medicine [10]. In both WHO and OIE lists, substances belonging to certain groups (aminoglycosides, cephalosporins of third generation, macrolides, penicillins, and quinolones) were considered critical important antimicrobial groups [11,13]. In fact, some specific antibiotics are critically and/or highly important in both human and veterinary medicine (Table 1). Interestingly, the antibiotic streptomycin is also used in plant agriculture in the prevention of fire blight disease in apple and pear tree caused by the phytopathogenic Erwinia amylovora [10]. The use of antimicrobial agents in human medicine is restricted to therapy and prophylaxis. However, the antimicrobials in farm animals have therapeutic, prophylactic, metaphylactic and sub-therapeutic uses [9,15]. Therapeutic treatments are planned for individual animals that are diseased, but in food animal productions it is often more efficient to treat entire groups by medicating feed or water [9,15]. This metaphylactic use is particularly common and it implies the use of antimicrobials in the whole herd or flock for disease prophylaxis and/or therapy in case of presence of clinical illness in one individual of the group [3,15]. Moreover, for some animals (poultry and fish) this mass medication is the only feasible means of treatment [9]. This metaphylactic use results in the frequent exposition of entire groups of animals, healthy and diseased, to antimicrobial agents [10]. Prophylactic antimicrobial treatments are typically used during high-risk periods for infectious disease such as after weaning or transport [9]. Antimicrobials may also be administered in relatively low (sub-therapeutic) concentrations to food animals to promote growth and to enhance feed efficiency [9,15]. A ban on the use of growth promoters was implemented in Europe in 2006, but it has not led to any consistent decrease in antimicrobials consumption since this ban has been compensated by metaphylactic and prophylactic uses [3].
In addition to antimicrobials, metals compounds, such as zinc and copper, are also used to supplement animal (mainly in pigs) feed for the prevention of post-weaning diarrhea and the stimulation of growth. Resistance to these compounds is often associated to resistance to antimicrobial drugs such as methicillin in staphylococci, or macrolides and glycopeptides in enterococci [16]. It has been shown that these metal resistance genes are frequent in animal-associated bacteria [16][17][18][19]. The use of antimicrobials, biocides and metal compounds in animal productions in sub-therapeutic doses and with long exposure periods, may promote that bacteria fix genes that confer AMR [6,10]. These resistance genes can subsequently be transmitted to human-adapted pathogens or to human gut microbiota via people, contaminated food or the environment [6,10].
The multiple pathways involved in AMR-genes dissemination and exchange within the agriculture, the environment and the food processing industry ( Figure 1) make difficult to track the movement of these genes in vivo [10]. In this regard, Thanner et al. [10] underlined some gaps regarding our current knowledge about AMR in plant and animal agriculture, and proposed a worldwide surveillance program of soil, plants, animals, water and wastewater treatment plants using the same methods that for AMR monitoring of human hospitals isolates. As for AMR in human bacteria, a better knowledge of the AMR in animal bacteria will help to achieve an effective and controlled use of antimicrobials in animals, thereby avoiding the dissemination of known and novel AMR-genes [10].

Presence of AMR-Genes in Animals: The Metagenomics Evidence
Although some bacterial species (such as Mycobacterium tuberculosis and Streptococcus pneumoniae) are specialist human pathogens, a larger number of species are opportunistic pathogens (such as Escherichia coli) causing disease in humans and other hosts including livestock and wildlife species, and are also present in the wider environment [3]. The interplay of these ecologies is important, since animals and the environment represent a major AMR reservoir. Through evolution, microorganisms have synthetized antibiotics and/or develop resistance methods for microbial competition in the environment [20]. In fact, AMR-genes have been found in soils not exposed to antimicrobials [21,22]. However, the human activity is increasing and changing this environmental resistome [20]. Indeed, animal microbiomes have acquired genes over years of exposure to antimicrobial agents and heavy metals compounds used as therapeutics, metaphylactics, prophylactics and growth promoters [3,23].
Current studies based on metagenomics and/or real-time polymerase chain reaction (PCR) approaches have given diverse results regarding the human, animal and environmental resistome. These novel technologies offer the possibility to elucidate the presence of AMR-genes in human, animal and environmental microbiomes and to identify the factors causing their persistence, selection and spread [10].
Some studies have suggested that human and animal microbiomes are different [21]. Agga et al. [21] compared environments related to animal (cattle and swine) and municipal (human) waste and saw that antimicrobial-resistant bacteria populations associated with animal agriculture were distinct from those associated to human activity. However, regarding the gene content, 25 out of 61 unique AMR-genes identified were common between municipal waste and animal samples. The half of the AMR-genes detected in another study [24], were only found in external environments. These genes from external environment microbiomes were mainly related to biocide and metal resistance [24]. Human microbiota had the highest abundance and diversity of AMR-genes, and the lowest taxonomic diversity [24]. Nevertheless, it was seen that tetracycline resistance genes dominated in both human and animal microbiomes [24]. Moreover, 20.5% of the AMR-genes detected were found in human, animal and environmental samples [24]. A recent study based on the comparison of published data on metagenomics made similar conclusions [25]. This study showed that the environment is a reservoir of the basic forms of resistance genes (such as bla TEM ), while both the human and mammalian gut microbiomes contained the widest diversity of clinically relevant resistance genes [25].
Some studies have identified AMR-genes in animal samples regardless of the antibiotic exposure. Sequence-based metagenomics analysis of conventionally raised cattle without therapeutic antibiotics exposure revealed that 3.7% of the sequences encoded resistance to antibiotics and toxic compounds, and nearly half of these genes encoded multidrug resistance efflux pumps [26]. A similar metagenomics study in chicken ceca, revealed that around 2% of the sequences encoded resistance to antibiotics and toxic compounds [27]. At least one quarter of these genes were related to tetracycline and fluoroquinolones resistance [27]. Studies in swine fecal samples revealed the existence of at least 149 AMR-genes in non-medicated animals [28,29].
Although non-exposed animals may already carry bacteria with resistance genes, some studies have underlined that their resistome can change after antibiotic exposure. Diverse studies have shown that antibiotic treatment increased diversity of antibiotic resistance genes [28,29]. Moreover, some enriched genes, such as an aminoglycoside O-phosphotransferase, confer resistance to antibiotics that were not administered, demonstrating the potential for indirect selection of resistance to classes of antibiotics not fed [28]. The effects of administering sub-therapeutic concentrations of antimicrobials to beef cattle were investigated in a recent study, showing that the antimicrobial treatment differentially affected the abundance of certain resistance genes in fecal deposits, but not their persistence [30]. In other study, the administration of the third generation cephalosporin ceftiofur in dairy cows increased the β-lactam and multidrug resistance genes in feces [31]. Nevertheless, another study identified approximately the same number (21)(22)(23)(24)(25)(26) of unique AMR-genes in manure samples of four dairy cows despite different prior exposure to antibiotics [32].
Some metagenomics studies have underlined the dominance of tetracycline resistance genes in animal microbiomes [24,33]. These results may be partially explained by the historical and current exposure to tetracyclines in the animal husbandry. Yet, a metagenomics study in pigs reared in an antibiotic-free environment revealed also the presence of diverse tetracycline resistance genes including novel genes, as well as a gene [tet (40)] not previously observed outside the human gut microbiome [34].
Although the metagenomics studies seem promising, they have some limitations. Since, some annotated genes (ex. efflux pumps) that confer AMR, perform other basic functions, more research on gene expression and functional analysis is needed to determine whether these genes can confer phenotypical resistance to antimicrobials [23]. Moreover, the sequence-based metagenomics approach does not provide information about the genetic context [23]. Nevertheless, some promising plasmid metagenomics studies have showed a broad dissemination of plasmid carrying AMR-genes in pig and bovine samples [35,36]. The functional genomics study about the tetracycline resistome of the pig gut has shown that most of the genes resided on putative mobile genetic elements (MGEs), which may contribute to the maintenance and dissemination of antibiotic resistance in antibiotic-free environments [34]. Nevertheless, other studies have probed that the presence of MGEs is also affected by antibiotic exposure [31,37]. It has been seen that the antimicrobial exposure increased the abundance of phage integrase-encoding genes (that may carry virulence or AMR-genes) in the viromes of swine, demonstrating the induction of prophages with antibiotic treatment [37]. In the study based on the administration of ceftiofur in dairy cows, an increase in gene sequences associated with phages, prophages, transposable elements and plasmids was observed [31].
Some metagenomics studies have proven that the animal microbiome is different of the human microbiome, but, interestingly, their shared a part of their resistome. Acquired resistance genes seem disseminated in the absence of selective pressure, but their abundance is affected by antibiotic exposition. These findings confirm that continue antimicrobial selective pressure in both humans and animals may benefit the dissemination of acquired resistance genes [23]. As other authors have concluded, a prudent use of antibiotics in human and veterinary health is needed to slow down the AMR spread and prevent the emergence of novel AMR-genes [23].

AMR-Genes in Gram-Positive Bacteria from Animals
Three Gram-positive species are considered members of the ESC(K)APE group: C. difficile, E. faecium, and S. aureus. This section focuses on the current knowledge about these pathogens regarding their AMR in animals, particularly on the genes common within these species (Table 2).

Clostridium difficile
C. difficile is an ubiquitous environmental organism widespread in rivers, lakes and soils. It is also found in the hospital environment where it is difficult to eradicate, as well as in meat products and animals including diverse species (calves, ostriches, chickens, elephants, dogs, horses and pigs) [49].
Although there are some current genomic updates of the clostridial phylogeny [50,51], in this review we use the classical nomenclature of Clostridium difficile.
C. difficile is recognized as the major cause of healthcare antibiotic-associated diarrhea [46]. Over the last decade, an alarming increase in incidence of C. difficile infection was observed across the USA, Canada and Europe, and it has been associated with the emergence of the highly virulent (hypervirulent) clone BI/NAP1/027 [named according to its restriction enzyme analysis (REA), pulsed-field gel electrophoresis (PFGE) and PCR Ribotype (RT)] [46,52]. This dramatic increase in C. difficile infections was associated to the fluoroquinolone resistance of this clone [52]. Fluoroquinolones are broad-spectrum antimicrobials highly effective for the treatment of bacterial infections in animals and humans. Most RT027 isolates harbored mutations in the quinolone-resistance determining region (QRDR) of the DNA gyrase subunit gene, gyrA, that confer resistance to fluoroquinolones [52]. Nevertheless, clinical C. difficile strains acquire fluoroquinolone resistance due to alterations in the QRDR of either GyrA or GyrB DNA gyrase subunits [46].
C. difficile infection has also emerged as a cause of diarrhea in the community, especially in populations previously considered at low risk, such as young people, antibiotic-naive patients or people without healthcare exposure [46]. Studies in the community indicated that up to 13% healthy human subjects are asymptomatically colonized with C. difficile [52]. In addition to RT027, a number of emergent hypervirulent RTs have recently been identified, notably the hypervirulent RT078 recognized as infection cause in hospitals, the community and animals [46]. The use of fluoroquinolones in the pork industry may have also contributed to the emergence of the multidrug-resistant RT078 [53]. In a study by Keessen et al. [53], most human and porcine isolates were resistant to ciprofloxacin (96%), and some were also resistant to moxifloxacin (16% for both human and porcine isolates). Resistance to moxifloxacin in this study was associated with a gyrA mutation [53]. Moreover, the use of fluoroquinolones [ciprofloxacin in humans, and enrofloxacin in pigs] was significantly associated with isolation of moxifloxacin-resistant isolates in both populations [53]. The authors proposed that the increased fluoroquinolone use could have contributed to the spread of C. difficile RT078 [53].
RT078 is commonly isolated from swine and other food animals [52]. It is likely that RT078 is an important pathogen of piglet diarrhea worldwide [49,54]. Molecular genotyping has suggested that RT078 isolates of human and swine origin are highly related and may therefore represent a potential zoonotic transmission [52]. Several studies have suggested that transmission from swine to humans may occur in farms or in large integrated swine operations, and moreover, the aerial dissemination of C. difficile from pig farms has been shown [52,54]. Due to the increased incidence of C. difficile infection outside the hospital environment and the presence of the same genetic lineage in food animals and its products, some authors have suggested that C. difficile may be considered as a foodborne pathogen [52]. Although RT078 is predominant in studies on swine, other RTs have been described in animals [52]. For example, RT046 has been found in both piglets and humans in Sweden [55].
AMR in C. difficile has been less intensively investigated than in other Gram-positive pathogens (such as S. aureus). In fact, few AMR-genes [such as erm(A), erm(B), tet(M), tet (44), tet(W), ant(6)-Ib, catD and cfr(B)] have been characterized in C. difficile isolates [46]. Resistance to cephalosporins is still uncharacterized, although most clinical C. difficile strains are resistant to these antibiotics. Similarly, some C. difficile human and animal isolates with reduced susceptibility to metronidazole have been found, although the resistance mechanism is not completely understood [46].
The most widespread mechanism of resistance to the antibiotics of the macrolide-lincosamide-streptogramin B (MLS B ) group in C. difficile is ribosomal methylation due to the erythromycin ribosomal methylases (erm) genes of class B [46]. It is to note, that erm(B) is the erm gene with the widest bacteria host range, and it has been found in both Gram-positive and Gram-negative bacteria, aerobic, and anaerobic genera and in most ecosystems [56]. However, erm-negative C. difficile strains resistant to both erythromycin and clindamycin or only to erythromycin have been also described [46]. Alterations in the 23S rRNA or ribosomal proteins (L4 or L22) have been found in some of these strains, but the presence of these changes in susceptible isolates had excluded their role in resistance [46]. Interestingly, a multidrug resistant cfr-like gene, cfr(B), that modifies the 23S rRNA has been recently found in clinical C. difficile isolates [57,58] (see Section 4.3.2).

Enterococcus faecium
Enterococci are commensal bacteria of the gastrointestinal tract of mammals and other animals that can also be detected in the environment [60,61]. In adult people, enterococci account for about 1% of the intestinal microbiota, being E. faecium and Enterococcus faecalis the most prevalent enterococci in the human gastrointestinal tract [61]. These species are opportunistic human pathogens that are implicated in life-threatening hospital acquired infections such as bacteremia and infective endocarditis [60,61].
Enterococci are intrinsically resistant to a number of first-line antimicrobial agents. They have resistance against cephalosporins and cotrimoxazole, as well as low-level resistance to β-lactams and aminoglycosides [60]. Moreover, clinical and animal Enterococcus isolates with resistance to other antimicrobials such as macrolides, tetracyclines, streptogramins and glycopeptides have been described [60,62]. Antimicrobials such as linezolid, daptomycin and tigecycline can be used in the treatment of enterococcal infections [63,64]. Nevertheless, resistance against these former antimicrobials has been reported, mainly in the clinical setting [63,64].
Nowadays, the vancomycin-resistant enterococci (VRE) pose a major therapeutic challenge due to their difficult treatment [61]. VRE were found in the hospital setting in the 80s [61,64], while the first description of animal reservoirs of vancomycin-resistant E. faecium was published in 1993 [65]. Later on, VRE have been described in diverse animals and environmental sources [60]. As vancomycin has not been used in veterinary medicine, it was hypothesized that the use of another glycopeptide, the avoparcin, as additive in farm animals feed, has influenced the emergence of VRE in animals since the 70s [60]. This correlation has been criticized and several studies have shown that vancomycin-resistant E. faecium persisted in animals for an extended time after the banning of avoparcin in the 90s [54,60]. However, this persistence has been related to co-selection with other antimicrobials (such as tetracycline or the macrolide tylosine) and metals (such as copper sulphate) [54,60]. It has been suggested that VRE strains are predominately host-specific, being hospital isolates genetically different from animal strains [60]. However, some research, based on multilocus sequence typing (MLST) showed that certain isolates from diverse clonal complexes (CCs) are present in animals, healthy humans and patients [54,60,61]. Moreover, the same variants of Tn1546 carrying the glycopeptide-resistant gene vanA have been detected in enterococci from human and animal origin, underlying that E. faecium from animals can act as a donor of AMR-genes for other pathogenic enterococci [54,60].
Similar to glycopeptides, the use of other antimicrobials as grown promotors may have influenced the emerging of enterococci resistant strains in food animals. In 1999, the streptogramin combination quinupristin/dalfopristin (RP59500, Synercid) was approved for clinical use. The mixture of type B streptogramin (quinupristin) and type A streptogramin (dalfopristin) in a 30:70 ratio showed good activity against multiresistant E. faecium strains [66]. Nevertheless, before its approval for clinical use, in 1997, quinupristin/dalfopristin-resistant E. faecium isolates were found in both human patients and chicken samples [66]. The use of virginiamycin, a streptogramin licensed for growth promotion in animals (including chickens and other poultry), was associated to the development of resistance to streptogramin combinations by enterococci [66]. In this sense, diverse streptogramin A resistance genes [such as vat(D), vat(E), vga(D) and vat(H)] have been detected in E. faecium from animals in Europe, USA and/or Asia [60].
Regarding other antimicrobials, human and animal E. faecium isolates frequently harbored genes conferring resistance against aminoglycosides [aph(3 )IIIa], tetracycline [tet(M)], and macrolides [erm(B)] [61]. Recently, a novel gene, optrA, which encodes for an ABC transporter that confers resistance to oxazolidinones and phenicols has been found in E. faecalis and E. faecium of human and animal origin [67]. This gene has been found chromosomally and plasmid located in Enterococci [40], but also in Staphylococcus sciuri [41,68]. The optrA-carrying plasmids in enterococci have also other AMR-genes against antibiotics such as MLS B [erm(A)-like)] and/or phenicol (fexA) [40]. Retrospective analysis of genome sequences has revealed that optrA has a wide dissemination in Gram-positive bacteria and it has also been found in diverse Streptococcus species (including S. suis, S. agalactiae and S. pyogenes) [69].
Some genes (cfr and dfrK) first discovered in animal-related staphylococci have recently been found in enterococci (see Sections 4.3.2 and 4.3.3). The trimethoprim resistance gene dfrK, encoding a trimethoprim resistant dihydrofolate reductase, was recently detected in E. faecium [43] (see Section 4.3.3). The multi-resistance gene cfr encodes a RNA methyltransferase that modifies the 23S rRNA gene conferring resistance to ribosome-targeting antimicrobials [44]. It has been detected in E. faecalis from animals and humans, E. thailandicus from pigs and farm environment, E. casseliflavus and E. gallinarum from pigs [42,45]. A non-functional cfr, due to a deletion in the regulatory region upstream, has been detected together with optrA in a clinical E. faecium isolate [38]. As for C. difficile, a functional cfr(B) has recently been found in E. faecium clinical isolates [39] (see Section 4.3.2).

Staphylococcus aureus and Related Species
S. aureus is part of the normal and transit human microbiota, and it is usually present in the nasopharyngeal mucosa, but also in the skin and other corporal areas [70][71][72]. The general carrier rate in humans is estimated between 20 and 30 percent for persistent colonization [73]. S. aureus is an opportunistic pathogen, which produces a wide spectrum of diseases, ranging from minor's skin to deep infections, as well as potentially-fatal diseases such as diverse invasive infections and toxin-mediated diseases [74,75]. As in humans, S. aureus is a commensal bacterium for animals but it is also able to cause diverse infections including abscesses, chondronecrosis, dermatitis, mastitis, pyaemia, osteomyelitis, pneumonia, septicemia and skin and wound infections [76].
Staphylococci of animal origin harbor a wide variety of AMR-genes [48,[77][78][79][80][81][82][83]. Recent reviews by Wendlandt et al. [48,81] have underlined that at least 44 AMR-genes in staphylococci have been detected in both human and animal isolates (Table 3). Interestingly, some shared genes were firstly described in isolates from animal origin [such as tet(L), erm(T), dfrK, fexA, cfr] or possibly originated from Staphylococcus species related to animals (as the methicillin resistance gene mecA). Next sub-sections are focused on those genes originated from animal related staphylococci.

mec Genes in Staphylococci: Origin and Reservoirs
One of the most important acquired resistance in staphylococci is methicillin resistance. This resistance is mainly due to the acquisition of the mecA gene, encoding a β-lactam low affinity penicillin binding protein (PBP) called PBP2a. This gene is carried on a MGE termed staphylococcal cassette chromosome mec (SCCmec). This MGE has been more extensively studied for S. aureus, since this is the most important Staphylococcus species for the human health. To date, eleven major SCCmec types carrying mecA have been described in methicillin resistant S. aureus (MRSA), which have been assigned on the characterization of its two essential components the mec and ccr complexes [84][85][86]. The mec-gene complex corresponded to the mecA operon variants, which can include functional and/or truncated regulatory genes. The ccr complex includes the recombinase(s) involve in the excision and integration of the element into the chromosome. The remaining regions of the SCCmec on which subtypes are based are called junkyard or joining regions, and they can contain MGEs, as well as heavy metal [arsenate (ars operon), cadmium (cadD-cadX), cadmium and zinc (ccrC), copper (copB), mercury (mer operon)] and other AMR (amynoglicosides (aadD, aad9/spc, aacA-aphD), bleomycin (ble), erythromycin [erm(A)], tetracycline [tet(K)]) resistance genes. Many non-typable SCCmec cassettes exist, and non-S. aureus staphylococci carry similar SCCmec and novel mec-ccr combinations as those found in S. aureus [17,[87][88][89][90]. This suggests horizontal transfer and recombination events within Staphylococci species. In fact, the mecA is widely distributed among methicillin resistant coagulase negative Staphylococci (MRCoNS) [17,[87][88][89][90][91].  Closed related mecA allotypes have been described in members of the S. sciuri group [98][99][100]. The members of the S. sciuri group are mainly considered commensal animal-associated bacteria with a broad range of hosts, although they have also been found in the environment and occasionally causing disease in humans and other hosts [98][99][100]. The mecA variants found in members of the S. sciuri species group are chromosomal located without being part of a SCCmec, and are usually phenotypically susceptible to β-lactams or show a heterogeneous resistance [91]. In this sense, a member of the S. sciuri group, the commensal animal-related S. fleurettii, has been suggested as the highly probable origin of the mecA gene [101]. The mecA-containing regions of S. fleurettii strains recovered from animals or its products are similar to the mec region of the S. aureus SCCmec type II. Moreover, the analysis of the corresponding gene loci region (with mecA gene homologues) of S. sciuri and S. vitulinus (which evolved from a common ancestor with that of S. fleurettii), probed that the mecA gene of S. fleurettii descended from its ancestor and was not recently acquired [101]. These findings suggested that the SCCmec in S. aureus was generated by acquiring this intrinsic mecA region of S. fleurettii.
Recently, two mecA gene homologs (mecB and mecC) have been discovered [102][103][104]. The gene mecB (formerly named mecAm) was found in 2009 in a Macrococcus caseolyticus isolated from a chicken [102]. This gene is located in a typical mec operon (blaZb-mecB-mecR1b-mecIb) with similar regulatory genes but also accompanied of a blaZ homologue encoding for a putative β-lactamase [85]. The mecB operon has been found on a transposon (Tn6045) either plasmid or chromosomal located [102], as well as in SCCmec elements [102,105]. Nowadays, mecB has not been detected in staphylococci. However, there is a potential risk of transmission since staphylococci and macrococci are closely related members of the same family (Staphylococcaceae). Moreover, these bacteria share a common niche, since M. caseolyticus is a commensal bacterium colonizing animal skin as staphylococci. Additionally, this mec variant is located in different MGEs, including a plasmid, which is a more mobile vehicle within bacteria than the SCCmec [85].
The gene mecC (formerly named mecA LGA251 ) was discovered in 2011 in S. aureus by two different working groups [103,104]. It was described from isolates originating from mastitis in cows and from humans in the United Kingdom, Ireland and Denmark [103,104]. Even if this gene was first described in 2011, a retrospective analysis in Denmark has shown that mecC-positive MRSA strains have been circulating before [106]. The mecC gene has a wide distribution and positive strains have been found in humans from both infection cases and carriage state, as well as in animals including livestock (dairy cattle, beef cattle, sheep, farmed rabbits), companion (cats, dogs, guinea pigs), wildlife (birds, mammals) and zoo (mara) animals [91,107]. Generally, there is a low occurrence of mecC-positive human isolates (ex. less than 1% in clinical MRSA from Belgium, [108]). This gene has mainly been associated with S. aureus lineages related to infections and colonization in animals [91]. The gene mecC in S. aureus is located on the SCCmec XI, which as few other SCCmec types also carried heavy metal resistance genes [104]. mecC homologues have been found in other staphylococci (S. xylosus, S. sciuri, and S. stepanovicii), associated to cassettes similar to SCCmec XI or to composite SCCmec together with mecA [109][110][111]. New mecC allotypes have been identified: mecC1 in S. xylosus related to bovine mastitis [109] and mecC2 in Staphylococcus saprophyticus from common shrew [112].

The Multi-Resistance Gene cfr
Linezolid, the first member of the oxazolidinone class of antibiotics, is considered to be a last-resort antimicrobial for the treatment of infections caused by VRE, MRSA and penicillin-resistant pneumococci [45]. Several mechanisms conferring linezolid resistance have been described in staphylococci, including point mutations in genes encoding 23S rRNA and mutations in ribosomal proteins L3, L4 and L22 [113,114]. In 2000, the gene cfr was identified in a bovine S. sciuri recovered in 1997 [115]. This gene encodes an RNA methyltransferase that modifies the 23S rRNA gene conferring combined resistance to phenicols, lincosamides, oxazolidinones, pleuromutilins and streptogramin A antimicrobials (known as PhLOPSA A phenotype). It also confers decreased susceptibility to the 16-membered macrolides (josamycin and spiramycin) [45]. Afterwards, this gene has been identified in methicillin-susceptible S. aureus (MSSA), MRSA, various coagulase negative staphylococci (CoNS) and in the coagulase-variable S. hyicus [45].
Although linezolid resistance mediated by cfr is not frequent in the clinical environment [116], reports of clinical outbreaks by cfr-containing S. aureus strains have been reported [117,118]. Recently, a clinical case due to a cfr-positive livestock-associated (LA-) MRSA CC398 has been described [119]. This gene has been found in important MRSA pandemic lineages such as the sequence type (ST)22/SCCmec IV, the Panton-Valentine leukocidin (PVL) positive ST8/SCCmec IV/USA300 or the ST125-MRSA-IVc [114,120,121]. It has also been detected in clinical strains of S. capitis [122] and methicillin-resistant S. epidermidis (MRSE) [123].
Recent studies have described the cfr gene or a variant in other Gram-positive (including Bacillus spp., C. difficile, Enterococcus spp., Macrococcus caseolyticus, Jeotgalicoccus pinnipedialis and S. suis) and Gram-negative (Proteus vulgaris and E. coli) species (Table 4). With the exception of Clostridium and Enterococcus isolates, the cfr-positive isolates of the other non-Staphylococcus genera were obtained exclusively from livestock and related farm environments [39,45,57,58].
In Bacillus spp., the gene cfr has been identified in three related plasmids from pig isolates from China, one (PBS-01) of these plasmids being previously reported in S. cohnii [45]. The cfr-carrying plasmid PSS-03 has been detected in both M. caseolyticus and S. cohnii [45], while the plasmid pJP1 or variants type pJP1-like have been found in M. caseolyticus, J. pinnipedialis, and S. lentus isolates [45]. Interestingly, some of these cfr-plasmids found in Gram-positive bacteria carried additional resistance genes against aminoglycosides (aadD, aadY), bleomycin (ble), MLS B [erm(B), erm(C)] and/or phenicol (fexB). In fact, a new multiresistance plasmid pWo28-3 related to pJP1, harbouring aacA-aphD, aadD, ble, optrA, cfr and fexA, has been identified in S. sciuri [68]. The cfr has also been found associated to a novel IS (ISEnfa5) truncating a novel lincosamide resistance gene lnu(E) in the plasmid pStrcfr of S. suis [126].
Recently, the cfr variant gene, designated cfr(B), was identified in C. difficile and E. faecium clinical isolates [39,57,58]. The nomenclature and functionally of this gene has been extensively discussed [127]. The Cfr proteins detected in other Gram-positive (Bacillus, Macrococcus, Jeotgalicoccus, Staphylococcus and Streptococcus) and in Gram negative (Proteus and Escherichia) bacteria were indistinguishable or similar (99% identity) from the original S. sciuri Cfr protein [127]. However, the cfr(B) shared only 75% amino acid identity with the original Cfr protein [39,58,127]. Schwarz et al. [127] suggested that the protein encoded by cfr(B) was structurally distantly related to the original Cfr. However, further research proved that the cfr(B) product does function as a Cfr protein [58], and therefore may be considered as a variant. Regarding Enterococcus, cfr-carrying plasmids have been described in E. faecalis from animals and humans, as well as in pig-related isolates of E. thailandicus, E. casseliflavus and E. gallinarum [42,45]. The cfr gene has also been detected chromosomally located in E. casseliflavus from pigs [42]. As mentioned in Section 4.2, a non-functional cfr gene has been detected together with optrA in clinical E. faecium isolates [38]. However, recently a cfr(B) gene sharing 99.9% sequence identity with the corresponding gene in C. difficile has been identified in E. faecium clinical isolates [39,128]. Deshpande et al. [39] described two copies of cfr(B) chromosomally located and embedded in a Tn6218 similar to the cfr-carrying transposon described in C. difficile. While the cfr(B) described by Bender et al. [128] was found in Tn6218-like elements possible linked to plasmids.
It has been shown that the gene cfr is functionally active in Gram-negative hosts [45]. Furthermore, it has been identified chromosomally located (inserted into the chromosomal fimD gene) in P. vulgaris from pigs, and in diverse plasmids in E. coli from pigs [45,[130][131][132][133][134] or food of animal origin [135]. The IS26 appears to play an important role in the transfer of this multiresistance gene in Gram-negative bacteria, since it appears in the cfr-carrying plasmids detected in E. coli [45,[130][131][132][133][134][135]. One of these cfr-carrying plasmids from E. coli of swine origin carried also the extended-spectrum-β-lactamase (ESBL) gene bla CTX-M-14b [133].

Other Genes in Animal-Associated S. aureus
S. aureus from animals can be related to various CCs [90,136,137]. However, animals are considered the main reservoir of the specific S. aureus lineage CC398, which has been the subject of numerous studies during recent years [18,19,82,[138][139][140][141][142][143]. Studies based on the phylogenetic analysis of genome-wide single nucleotide polymorphisms (SNPs) supported the existence of two subpopulations in CC398: an ancestral human-adapted clade and an animal-associated clade [144,145]. Nevertheless, LA-MRSA CC398 isolates of the animal-associated clade are able to infect humans [119,145]. Analysis of genes present in CC398 has revealed numerous genes in common with clinical S. aureus and/or other staphylococci, but also some novel or rare resistance genes have been found in animal-related isolates from this lineage [79,81]. Some of these rare or novel genes are often on plasmids encoding for other AMR-genes such as the multi-resistance gene cfr (Table 5).
One interesting gene, found in animal-associated S. aureus, is the phenicol exporter gene fexA. It was first described in a bovine S. lentus isolate, but later it has been found in diverse plasmids of LA-MRSA CC398 from pigs, cattle and horses, as well as on plasmids from the also animal-related S. aureus CC9 [79,81]. This gene has been related to the non-conjugative transposon Tn558 that has been detected (partially or complete) in plasmids or in the chromosome of various CoNS of animal origin [79]. The fexA is co-located in plasmids with additional resistance genes (Table 5). Interestingly, the fexA gene has been described in cfr-carrying plasmids of clinical important MRSA pandemic lineages [114,120].
The tetracycline gene tet(L) was found initially in the 80s in diverse plasmids from Bacillus. In staphylococci, it was first described in the early 90s on a plasmid from a porcine S. hyicus. In addition, subsequently, it was detected on structurally diverse plasmids of staphylococci of animal origin [77,79]. This gene has been found co-located on plasmids from staphylococci with other additional antimicrobial and heavy metals resistance genes (Table 5). Regarding LA-MRSA CC398, the gene tet(M) is the most frequent tetracycline-resistant gene found in these isolates, although it is accompanied frequently by tet(K) and tet(L) [139]. The gene tet(L) has also been described in clinical S. aureus in the 90s [146], and recently in a cfr-carrying plasmid with additional antimicrobial genes [dfrK and aadD] in a clinical S. aureus ST125 strain [121].
The trimethoprim resistance gene dfrK was discovered co-located on a plasmid with tet(L) in LA-MRSA CC398 [79,81]. Similar to other genes in staphylococci, the dfrK has been found on diverse plasmids co-located with other antimicrobial and heavy metals resistance genes ( Table 5). The dfrK gene is widely disseminated in LA-MRSA CC398 and it has been found in isolates from pigs, cattle and poultry [79]; but it has also been found chromosomally located on the transposon Tn559 of MSSA CC398 [147]. Additionally, the Tn559 carrying dfrK has been found in E. faecium [43]. The erm(T) gene has been previously identified in other Gram-positive bacteria (such as Streptococcus and Lactobacillus) [79]. In staphylococci, this gene has been originally found in animal-associated CC398, but further research suggested that it is particularly associated to the MRSA and MSSA CC398 from the ancestral human clade [79,149,150]. Nevertheless, a recent study comparing erm(T)-carrying plasmids from S. aureus ST398 from pig and humans, showed that these plasmids are quite similar and all carried additional tetracycline [tet(L)] and heavy metal (cadD-cadX, copA, mco) resistance genes [148]. The plasmid (pUR1902) recovered from a pig isolate carried also aadD as a plasmid (pUR2941) from a human isolate, while the remaining plasmid (pUR2940) from a human isolate additionally carried erm(C) and dfrK [148].
A recent review by Wendlandt et al. [81] underlined the presence of diverse multidrug resistance (MDR) genes co-located together or with other resistance genes on plasmids in staphylococci from animal origin. These genes included MLS B [erm(A), erm(B), erm(C), erm(T), erm (33)], lincosamides-pleuromutilins-streptogramin A [lsa(E), vga(A), vga(C)] and PhLOPSA A (cfr) resistance genes [81]. The main resistance mechanisms conferred by these MDR genes included target modification by methylation (erm and cfr genes) and active efflux via ABC transporters [vga and lsa(E) genes] [81]. Their location on plasmids, co-located with other AMR-genes, may potentiate their co-selection and persistence in animal staphylococci, and furthermore represent an important pool of resistance against critically and highly important antimicrobial agents [81].

AMR-Genes in Gram-Negative Bacteria from Animals
The Gram-negative members of the ESC(K)APE group are A. baumannii, P. aeruginosa, and the members of Enterobacteriaceae. A large number of antimicrobial resistance genes have been described in Gram-negative bacteria [44,47]. Since this area is extremely large and complex, in this section, we summarize part of the current knowledge about AMR in A. baumannii, P. aeruginosa and Enterobacteriaceae from animals, with special attention to the emergence of carbapenemase-producing Gram-negative bacteria in animals (Table 6) and the emergence of the colistin resistance.

Acinetobacter baumannii
A. baumannii is an important nosocomial pathogen that usually affects immunocompromised patients suffering from various underlying diseases [159]. Nosocomial infection with this bacterium has been associated with increased morbidity, mortality and health care costs [160]. It is responsible of hospital outbreaks, and it has a remarkable ability to survive for prolonged periods throughout hospital environments [160]. It is well established that the members of the genus Acinetobacter are ubiquitous microorganisms. However, A. baumannii as a highly prevalent microorganism in nature, is a misconception because the difficulties encountered in its identification [159,160]. In fact, A. baumannii is phenotypically and genotypically closely related to other Acinetobacter species (A. pittii, A. nosocomialis and A. calcoaceticus), making the species identification challenging [159].
A. baumannii has been related with community-acquired infections (single events or case series) [160], and has been isolated from various environmental locations: soils contaminated with petroleum hydrocarbons, vegetables, inanimate surfaces in contact with humans, manured agricultural soil, pig slurry and aquaculture environments [160]. Multi-susceptible Acinetobacter isolates were commonly found in the 70s-80s in soil and in the hospital setting, clinical A. baumannii being easily treated with common antibiotics during this period [20,159]. However, this bacterium has an extraordinary ability to upregulate or acquire resistance determinants, and nowadays infections with multidrug or even pandrug resistant isolates are increasing [159].
During the last decade, A. baumannii strains have also been isolated from animals, mainly causing outbreaks in veterinary clinics or hospitals [160]. A. baumannii strains have been isolated from diverse animals including ducks, pigeons, chickens, donkeys, rabbits, pets (cats and dogs), mules, livestock (cattle, caws, goats, pigs), horses, lice and arthropods [161]. In a study based on PFGE, it was seen that isolates from livestock were different than A. baumannii human strains [162]. Yet, isolates recovered from pig fecal samples harbored bla OXA-51 , which has already been reported in human clinical isolates [162]. In the same study new bla OXA-51 -like genes (bla OXA-148 , bla OXA-149 and bla OXA-150 ), not previously detected in human isolates, were described in bacteria from cattle [162]. However, it is important to underline that the bla OXA-23 and bla OXA-51 -like genes may be naturally occurring in A. baumannii [163]. On the other hand, the studies in companion animals have reported A. baumannii isolates genetically similar to the nosocomial European (also called International) clones I, II and III, suggesting its spread from humans to animals directly or via the environment [159,160,164,165].
Regarding the distribution of AMR-genes in A. baumannii in animals, some studies have reported the emergence of carbapenemase producing A. baumannii in livestock and companion animals [153,159] ( Table 6). The bla OXA-23 is a wide distributed carbapenemase gene in A. baumannii isolates, which has been described recently in livestock and/or pets [151,153,154,159,165]. The bla OXA-23 gene has also been found in A. lwoffi from poultry [154]. In A. baumannii this gene has been found in STs not previously reported in humans, as well as in A. baumannii strains belonging to ST2, which has been previously related to hospitals outbreaks [151]. The bla OXA-58 has been found together with bla OXA-23 in A. baumannii from fowl [151]. Similarly, a bla NDM-1 positive A. baumannii isolate has been recovered from a pig suffering from pneumonia and sepsis [158]. This bla NDM-1 was harbored by a plasmid also carrying other AMR-genes [aphA6, ble and msr(E)-mph(E)] [158]. The bla NDM-1 has been found in other Acinetobacter species [A. junii (from a pig farm) and A. calcoaceticus (from around a cow farm)] recovered from environmental samples and farm animals [155,166]. A new carbapenemase (bla OXA-497 ) has been found in A. baumannii from dairy cattle [157]. A class 1 integron similar to an integron of a human isolate has been also identified in an equine A. baumannii isolate [167].
Currently, there are some reports underlying the emergence of A. baumannii strains resistant to both carbapenems and polymyxins (colistin and polymyxin B) in the clinical setting [168]. Resistance rates against polymyxins varied from 0.7% to 6.5% depending on the country [168]. Most polymyxin resistant strains from Europe have been recovered in Greece and Italy [168]. Interestingly, recent reports have also underlined the presence of colistin and polymyxin B resistant A. baumannii and other Acinetobacter spp. isolates in meat (chicken, turkey, beef and pork) [169,170]. Resistance to polymyxins in A. baumannii is mediated by mutations in the genes pmrA and/or pmrB [168]. Nevertheless, a new colistin resistance mechanism mediated by plasmids (the mcr-genes, see Section 5.3.4) has emerged in Enterobacteriaceae [168]. Although no clinical or environmental A. baumannii isolates carrying mcr-genes have yet been described, a recent study has proved that A. baumannii transformed with mcr-1 carrying plasmids had increased colistin resistance [171]. This finding highlights the threat of a possible dissemination of these mcr-genes to multidrug resistant A. baumannii [171].

Pseudomonas aeruginosa
P. aeruginosa is an opportunistic pathogen often found in water and soil that is pathogenic to plants, humans, farm animals and companion animals. In humans, it is a cause of community and nosocomial infections, especially in patients immunocompromised and/or with cystic fibrosis. In animals, it caused pyoderma, otitis and urinary tract infections in companion animals, mastitis in dairy cows, endometritis in horses and hemorrhagic pneumoniae in fur-bearing animals [172][173][174][175][176]. Due to the presence of several drug efflux systems and porins, as well as its cell wall with low permeability, P. aeruginosa is intrinsically resistant to a wide range of antimicrobials including benzylpenicillins, aminobenzylpenicillins, carboxypenicillins, first and second generation cephalosporins, chloramphenicol and tetracycline [172,173]. Moreover, this bacterium is able to form biofilms and to acquire diverse resistance mechanisms.
The studies about AMR in animals are scarce, and mainly have focused in companion animals. A study has shown that P. aeruginosa isolates recovered from diverse veterinary samples during 1994-2003 have high resistance rates against β-lactams (70-100%) and sulphonamides (80-90%), while resistance against quinolones an aminoglycosides are more variable, with resistances ranging from 5 to 98% depending of the antibiotic [173]. Similar results were found in a study with canine isolates recovered between 2003 and 2006, although an increase in quinolone resistance was seen [174]. Resistance rates to the former antimicrobials are also similar in current studies among companion animals, although higher resistance rates against the aminoglycoside gentamicin [177] and the quinolone enrofloxacin [176] have been found. In P. aeruginosa from companion animals, the resistance to quinolones has been related to point mutations in gyrA, gyrB, parC and/or parE genes [172,176], and the resistance to aminoglycoside has been associated to diverse resistance genes (such as aacA4 and aadA6) [172,176]. Resistance rates are generally lower in P. aeruginosa from livestock comparing to strains from companion animals [174,175].
Regarding its potential as zoonotic pathogen, a recent study by Haenni et al. [174] confirmed that the P. aeruginosa population in veterinary samples from France has a non-clonal epidemic structure. There was a poor association between an animal species and a specific clone, even though certain clones, possibly correlating with higher pathogenicity, seem to be more prevalent than others [174]. Clones associated to human outbreaks were detected, but not the most frequent epidemic clones associated to MDR in humans (also called "high risk clones") [174].
The presence of ESBLs (via phenotypic methods) has been confirmed in P. aeruginosa isolates from animals [176,178]. Moreover, some reports have underlined the emergence of P. aeruginosa isolates with carbapenemases in animals ( Table 6). The gene bla IMP-4 has been found in P. aeruginosa from dog [156], and the gene bla VIM-2 has been found in P. aeruginosa from cattle and fowl [151].
Similar to A. baumannii, recent reports have underlined the emergence of polymyxin-resistant P. aeruginosa [168]. Resistance rates are low (0.5% to 1.1%) in most countries, although, the situation is worrying in China where 22.2% of extensively drug-resistant bacteraemic P. aeruginosa isolates are resistant to polymyxin B [168]. Diverse mechanisms of polymyxin resistance have been described in P. aeruginosa, even though no clinical or environmental strains carrying mcr-genes have been reported [168]. Nevertheless, the recent study by Liu et al. [171] showed that P. aeruginosa isolates transformed with plasmids carrying mcr-1 only had moderate changes in colistin susceptibility.

Enterobacteriaceae
As some Enterobacteriaceae (Ex. Salmonella enterica, Yersinia enterocolitica) are typically food-borne pathogens, this section will mainly focus on some transmissible AMRs-genes that have emerged in Enterobacteriaceae from animals ( Table 7). The recent discover of the cfr gene in Enterobacteriaceae has been discussed at Section 4.3.2.

Emergence of Streptothricin-Resistant E. coli in the 1980s
Similarly to Gram-positive bacteria, the use of antibiotics as grown promotors in the animal husbandry may potentiate the emergence of resistant Enterobacteriaceae isolates in food animals. A good example is the emergence of streptothricin-resistant E. coli in the 1980s. The aminoglycosidic growth promoter nourseothricin (streptothricin) was used in farm animals in Germany during the 1980s, while no equivalent antimicrobials were used in humans over this period [49]. Resistance emerged in E. coli from pigs the second year after the introduction of this antibiotic. It was mediated by a plasmid containing a transposon coding for a streptothricin acetyltransferase [181]. Subsequently, this resistance was found in E. coli isolated from pig farmers, in the community and in other Enterobacteriaceae from humans (Salmonella and Shigella) [49]. Nowadays, streptothricin-resistance is extensively extended and it is also characteristic of Gram-positive bacteria such as enterococci and staphylococci [182].

ESBL/AmpC-Carrying Enterobacteriaceae in Animals
Bacteria carrying ESBLs are a worldwide clinical problem. ESBLs are mainly plasmid-encoded enzymes providing an extended resistance to β-lactam antibiotics, due to their ability to inactivate cephalosporins [49,183]. They can be produced by a variety of different bacteria including Enterobacteriaceae or no-fermenting bacteria (such as P. aeruginosa), E. coli and K. pneumoniae being the most frequently found ESBL-producing bacteria. ESBL-producing bacteria are known as nosocomial pathogens and since the late 1990s they have been increasingly found as a causal agent of infections in the community [183]. Until the 90s, the vast majority of ESBLs identified in human clinical isolates were SHV (sulfhydryl-variable) or TEM (named Temoneira for the first patient from whom the pathogen was isolated) types [184]. However, later, CTX-M (cefotaximase) β-lactamases have emerged and currently they are the most prevalent ESBLs in human Enterobacteriaceae [184].
The occurrence of ESBL-producing bacteria has been broadly recognized in veterinary medicine since the 2000s [183,184]. ESBL-producing bacteria have been found as disease agents and/or colonizers in livestock, companion animals, zoo animals and wild animals [54,[183][184][185][186]. E. coli and Klebsiella spp. are cause of mastitis in dairy cattle, but most often livestock animals are asymptomatic carriers of ESBLs producers [54,187]. The first ESBL-producing E. coli was isolated from a dog with urinary tract infection and it carried SHV-12 [188]. Later on diverse CTX-M, TEM and SHV types have been observed in E. coli, Salmonella spp. and K. pneumoniae from livestock and companion animals [54,155,183,185,186]. The β-lactamases CTX-M-1, CTX-M-14, CTX-M-15, SHV-12 and TEM-52 are the most frequent types in Enterobacteriaceae from animals [54,155,183,[185][186][187]. This distribution is similar in humans, where CTX-M types are the major β-lactamases in E. coli and Klebsiella spp. [54,183]. Host-range plasmids of different incompatibility groups (such as IncN, IncI, incF and IncK) have been related to these bla CTX-M genes [154].
Identical phylogenic lineages (such as ST131) have been found in E. coli isolates from humans and animals [184]. The risk of zoonotic transfer from livestock to people with close contact to these animals is still largely unknown, but some studies have implicated a transfer of ESBL-producing E. coli or ESBLs genes from poultry or pigs to farm workers [183] Besides this direct zoonotic transfer, other routes as foods of animal origin may be a risk factor for human colonization or infection. In a recent study 70.6% of tested farms were ESBL-positive [183]. In one case the same isolate was detected in human and cattle samples, indicating a zoonotic transfer. In few other cases, pig and human isolates shared the same ESBLs genes, although the isolates belonged to different lineages suggesting horizontal gene transfer.
In contrast to the situation in Europe, ESBL genes have not been commonly reported in animal isolates in North America [154]. In North America, the plasmid-encoded AmpC B-lactamases genes are more frequently found [154], but these genes have also been described in Europe [54]. The broad-spectrum cephalosporinases (AmpC) were the first β-lactamases described in E. coli [47], and nowadays they have emerged worldwide [184]. As for ESBLs, E. coli carrying AmpC (bla CMY-2 ) has been identified in animals during the 2000s [184]. These β-lactamases have been detected in E. coli and Salmonella spp. from livestock and companion animals, being CMY-2 the most frequent one [184,185]. Other AmpC β-lactamases encoding by bla ACC type genes have also been detected in Salmonella spp. from livestock [153,184].
The epidemiologically most important carbapenemases are class B metallo-β-lactamases (MBLs) such as VIM (Verona integron-encoded MBL), IMP (imipenemase) and NDM (New Delhi MBL), class A, such as, KPC (K. pneumoniae carbapenemase), and class D including OXA (Carbapenem-hydrolysing oxacillinase) carbapenemases [153]. Resistance due to carbapenemases is mainly linked to the nosocomial setting, since carbapenems are not approved for use in veterinary medicine. However, they may be use for companion animals under certain conditions [154]. Although they are not used in livestock, carbapenem-resistant bacteria have been isolated from farm animals in recent years [154]. In animals, the first carbapenemase gene, bla VIM-1 , was found in a porcine E. coli and it was linked to a multiresistance class 1 integron [152]. Later on, this and other carbapenemases have been found in livestock, wildlife and companion animals (Table 6), notably bla VIM-1 in S. enterica serovar Infantis and E. coli from livestock; bla NDM-1 in S. enterica sevorar Corvallis and E. coli from wild animals and/or livestock; bla OXA-48 in E. coli and K. pneumoniae from companion animals [153,155]. Carbapenemases have been detected in other Gram-negative bacteria including A. baumannii (see Section 5.1) and P. aeruginosa (see Section 5.2). However, the limited reports suggested that carbapenem-resistant bacteria are still at a very low prevalence in livestock [154].

mcr-Genes Mediating Colistin Resistance
Colistin, also named polymyxin E, is currently used as a last-line drug against MDR Gram-negative bacteria [168,179]. However, resistance against colistin has even emerged in humans without contact to this antibiotic [189]. Moreover, since the 60s colistin has been used in pig production for therapeutic (in monotherapy), prophylactic and grown promotions purposes [190]. In addition, to its use in pigs, polymyxins (especially polymyxin B), are used in some countries for the treatment of coliform and Pseudomonas mastitis in cows [190]. Polymyxins are used in companion animals (dogs and cats) for topical indications such as otitis and ophthalmic diseases [190].
Colistin resistance is frequently due to chromosomal mutations and it has been detected in P. aeruginosa, A. baumannii and Enterobacteriaceae [168,179]. Most mechanisms conferring resistance against colistin are related to modifications of its primary target, the lipid A moiety of lipopolysaccharide (LPS) [168,179]. Colistin resistance mechanisms are different across bacterial genera, but most implied lipid A modifications with 4-amino-4-deoxy-L-arabinose (L-ara4N) and/or phosphoethanolamine (PEtN) [168,179].
The chromosomal-related colistin resistance mechanisms have no possibly of horizontal transfer, but plasmid-related genes (mcr-1, mcr-2) coding for PEtN transferases have recently emerged. The first mcr-gene, corresponded to mcr-1. It was initially described in an IncI2 plasmid from an E. coli from a pig, but in the same study it was soon after found in E. coli and K. pneumoniae from patients [191]. This gene was discovered in November 2015, but recent research has identified this gene in E. coli collections from the 80s [179]. The amino-acid sequence of MCR-1 showed that it was closely related (63%) to the PEtN transferase EptA found in Paenibacillus sophorae and Enhydrobacter aerosaccus [191]. Interestingly, the polymyxin is biosynthesized in Paenibacillus spp. [191]. Structural analysis of the MCR-1 complete protein showed that it was closely related to the PEtN transferases LptA of Neisseria meningitidis and EptC of Campylobacter jejuni, which are also known to be intrinsically resistant to polymyxin [179,191]. Further research showed that the prevalence of the mcr-1 gene was 20% in animal strains and 1% in human strains in China, and nowadays it has been also detected in animal and human isolates recovered from other countries covering Europe, Africa and America [168,179,187].
A variant (with only a SNP of difference) of mcr-1, mcr-1.2 has been detected on a transferable IncX4 plasmid in a K. pneumoniae recovered from a patient [195]. This mcr-1.2 strain also carried β-lactamases genes (bla TEM-1 , bla SHV-11 , and bla KPC-3 ) [195]. Very recently, a novel plasmid mediated colistin resistance gene, mcr-2 has been discovered [198]. This mcr-2 has been found in an IncX4 plasmid of E. coli from porcine and bovine origin in Belgium [198]. This gene corresponded to a new mcr-gene, since MCR-1 and MCR-2 shared 80.7% of similarity [198]. It has been seen that the mcr-2 gene had higher prevalence than mcr-1 in E. coli isolates from porcine origin [198].
Food animals seem the main source of human contamination by the mcr-genes [190]. However, the mcr-1 gene has also been detected in E. coli carried or infecting humans without animal contact [190]. These findings underlined that this gene, which probably has emerged in the animal sector, is already widespread in the environment and it is transmissible via various routes to humans [190,199].

Other AMR-Genes in Enterobacteriaceae from Animals
Increasing levels of quinolone resistance among Enterobacteriaceae and other bacteria (Campylobacter spp.) have been a particular case of concern since the 90s [155]. The extended use of quinolones to treat poultry infections has increased the quinolone resistance among E. coli strains in poultry industry [200]. Although quinolone resistance is generally not highly frequent in animal farming, high and/or moderate levels of quinolone resistance have been reported in poultry [200].
The mechanisms of quinolone resistance in Enterobacteriaceae from animals are similar to those described in isolates from humans [201]. Mutations at gyrA (DNA gyrase) and/or parC (topoisomerase IV) genes were responsible of quinolone resistance in E. coli and Salmonella spp. isolates from animals [155,201]. Moreover, reports about plasmid-mediated quinolone resistance (PMQR) genes [qnr, acc(6)-Ib-cr, qepA, oqxAB] in bacteria from animal origin have been published since 2000s [155,180].
The Qnr proteins are encoded by several variants of qnr genes (qnrA, qnrB, qnrC, qnrD, qnrS), being the qnrS1 gene variant the most frequently reported and worldwide diffused in animal and human samples [200] (Table 9). Some qnr genes have been described in Enterobacteriaceae (mainly Salmonella spp. and E. coli) of animal origin [155,180,202]. Their wide distribution suggests an origin prior to the inclusion of quinolones in medicine [180]. It has been suggested that these genes have originated in bacteria from the natural environment [20]. In fact, qnr genes have been found in aquatic and waterborne organisms such as Shewanella, Aeromonas and Citrobacter species, and in the Vibrionaceae family [20,180].
As the qnr genes, PMQR efflux pump genes (qepA and oqxAB) have also a wide distribution, and they have been found in diverse Enterobacteriaceae from animals and humans [155,180] (Table 9). These genes are related to transmissible plasmids, although oqxAB is also commonly found in the chromosome of K. pneumoniae [180]. Other plasmid-mediated efflux pumps affecting quinolones have been described sporadically [180].
The cr variant of the aminoglycoside aac(6 )-Ib resistance gene was discover in 2006 [203]. The aac(6 )-Ib gene encodes a aminoglycoside acetyltransferase able to confer resistance against kanamycin, tobramycin and amikacin. The two mutations present at aac(6 )-Ib-cr confer low level ciprofloxacine resistance, with a slightly cost to the levels of aminoglycoside resistance [203]. Hence, it acts additively together with Qnr proteins to generate quinolone resistance [201]. This gene is widely distributed in Gram-negative bacteria (Table 9).
Regarding other aminoglycoside resistance genes, the gene aac(3)-IV, which confers cross-resistance between gentamicin and apramycin, need to be underlined [204]. It was originally isolated in 1981 from an E. coli recovered from farm animals in France. Although apramycin has only veterinary use, this gene has been detected in Enterobacteriaceae from human patients and wastewater from a residential area [204]. It has been suggested that apramycin consumption at farm level has increased the occurrence of aac(3)-IV positive E. coli in pigs [204].  [155] and Veldman et al. [202]. C, companion animals origin; E, environmental origin; F, food origin; H, human origin; L, livestock origin; W, wildlife origin; Z, zoo animals.
Enterobacteriaceae from animals carry other resistance genes. For example, a moderate incidence of chloramphenicol resistance is registered in E. coli from food animals in Europe, mainly mediated by genes such as catA1, floR and cmlA1 [200]. Other chloramphenicol resistance genes (such as catB) are more frequent in human E. coli isolates [200].

Conclusions
The continuous antibiotic selective pressure in human and animal health may contribute to the mobilization of acquired resistance genes. This is illustrated by several facts: (i) some studies have suggested that the mecA gene may have originated in animal related staphylococci; (ii) it has recently been observed that two mutiresistance genes (cfr and optrA) associated to MGEs such as plasmids along with other AMR-genes, have a wide dissemination in Gram-positive bacteria from animals and humans; (iii) typical nosocomial resistances linked to ESBLs and carbapenemases are emerging in Gram-negative bacteria from animals; and (iv) the mcr-genes, which may probably have emerged in the animal sector are currently spreading among human and animal E. coli isolates. These examples underline the fact that bacteria from animals represent an important pool of resistance genes for human pathogens.
Resistance to antibiotics is escalating, and at the same time the current pipeline of new antimicrobials is running dry, creating an ever increasing gap. Since a complete ban on the use of antimicrobials in farm animals would have serious repercussions for animal health, welfare and productivity, we need to use these agents more prudently in both human and animal medicine [3]. Rapid diagnosis tools are needed to determine therapy strategies more quickly and accurately, as well as the further examination of alternatives to antimicrobials for farm animals (such as phage therapy or vaccines) [3]. We are facing the possibility of a future without effective antibiotics for some infections and a scenario where infections that hitherto were considered harmless, are now a serious health problem and a major cause of morbidity, mortality, together with major financial and social repercussions. Acknowledgments: We thank members of our laboratories and our microbiologist colleagues for their support to the Belgian National Reference Centre-Staphylococcus aureus.

Conflicts of Interest:
The authors declare no conflict of interest.