The role of wildlife (wild birds) in the global transmission of antimicrobial resistance genes

Antimicrobial resistance is an urgent global health challenge in human and veterinary medicine. Wild animals are not directly exposed to clinically relevant antibiotics; however, antibacterial resistance in wild animals has been increasingly reported worldwide in parallel to the situation in human and veterinary medicine. This underlies the complexity of bacterial resistance in wild animals and the possible interspecies transmission between humans, domestic animals, the environment, and wildlife. This review summarizes the current data on expanded-spectrum β-lactamase (ESBL), AmpC β-lactamase, carbapenemase, and colistin resistance genes in Enterobacteriaceae isolates of wildlife origin. The aim of this review is to better understand the important role of wild animals as reservoirs and vectors in the global dissemination of crucial clinical antibacterial resistance. In this regard, continued surveillance is urgently needed worldwide.


INTRODUCTION
Over several decades, antimicrobial resistance has become a global clinical and public health threat against the effective treatment of common infections caused by resistant pathogens, resulting in treatment failure and increased mortality (WHO, 2014). The development of bacterial resistance is a natural evolution of microorganisms, but the widespread use and misuse of antibacterial agents in humans and animals has accelerated this process (WHO, 2014). Furthermore, the increasing frequency of global travel and trade has also contributed to the rapid worldwide spread of antimicrobial resistance (Laxminarayan et al., 2013). Some resistant clones, such as Escherichia coli ST131, Klebsiella pneumoniae ST258 and ST11, and methicillin-resistant Staphylococcus aureus (MRSA) USA 300, which are involved in the spread of resistance to crucially significant antibiotics in human medicine, have been widely disseminated (Lee et al., 2016;Mathers et al., 2015;Nimmo, 2012). Antimicrobial resistance is a complex and multifaceted problem involving humans, animals, and the environment. However, the role of wildlife in the emergence of antibacterial resistance might be underestimated. The first report of antibacterial resistance in wildlife revealed chloramphenicol resistance in E. coli isolates obtained from Japanese wild birds (Sato et al., 1978). Since then, the occurrence of resistant bacteria in wild animals has been increasingly reported within diverse animal species across different geographical areas. In addition, several important antimicrobial resistant pathogens, such as MRSA (Loncaric et al., 2013a;Porrero et al., 2014), vancomycin-resistant enterococci (Drobni et al., 2009;Sellin et al., 2000), Salmonella spp. (Lee et al., 2011a), Vibrio cholerae (Aberkane et al., 2015), and Campylobacter spp. (Weis et al., 2016), have been described in wild animals, highlighting the importance and complexity of wildlife, not normally exposed to antibiotics directly, in the transmission of resistant bacteria. 1 This review gives a brief overview of the emergence and prevalence of expanded-spectrum β-lactamase (ESBL), AmpC β-lactamase, carbapenemase, and colistin resistance genes in Enterobacteriaceae strains from wild animals, all of which have significant public health impact. Furthermore, this review aims to better understand the role of wildlife in the transmission of clinically significant antimicrobial resistance in Enterobacteriaceae.

ESBL-PRODUCING ENTEROBACTERIACEAE FROM WILDLIFE
The global dissemination of ESBL-producing Enterobacteriaceae in human clinics is an urgent problem that poses a serious challenge to the treatment of infectious diseases, particularly the worldwide emergence of CTX-M-15-producing ST131 E.
Since the first report on ESBL-producing E. coli isolates from wild animals in Portugal in 2006(Costa et al., 2006, ESBLproducing Enterobacteriaceae of wildlife origin have so far been reported in Europe, Africa, Asia, South America, North America, and Australia (Table 1). Although ESBLs have been found in various Enterobacteriaceae, most ESBL-producing bacterial pathogens in wild animals are E. coli, followed by K. pneumoniae (Table 1). To date, at least 80 wildlife species have been found to be carriers of ESBL-producing Enterobacteriaceae, most being wild birds (Table 1). Similar to that among isolates from human and veterinary medicine, the CTX-M family is the most prevalent type of ESBL-producing Enterobacteriaceae found in wild animals (Table 1) (Table 1). Significant geographical differences have been observed in the occurrence of CTX-M enzymes. As summarized in Table 1, CTX-M-15 is the only reported CTX-M-type β-lactamase in Africa to date, and is the most common CTX-M-type enzyme reported in Bangladesh. In Canada and the US, CTX-M-14 is dominant, followed by CTX-M-15. Diversity in CTX-M β-lactamases has been reported in European countries, with the predominance of CTX-M-1 and CTX-M-15. Interestingly, CTX-M-15 is also reported to be the most common CTX-M enzyme in Franklin's gulls (Leucophaeus pipixcan) in northern Chile (Báez et al., 2015), although CTX-M-1 was previously reported to be dominant in the same gull species in central Chile (Hernandez et al., 2013). Báez et al. (2015) hypothesized that, based on their migratory habits, Franklin's gulls from the north acquired resistant CTX-M-15-producing ST131 and ST10 E. coli clones, which are highly prevalent in humans in the US and Canada but scarce in Chile. However, this hypothesis, though possible, needs further investigation.
More than 170 different sequence types (STs) have been identified in ESBL-producing E. coli isolates of wildlife origin (Table 1). Among them, ST131 is the most commonly detected clone. The dominant ST131 clone identified in wild animals, which has been frequently described in humans, companion animals, food products, and the environment, is involved in the international dissemination of bla CTX-M-15 and bla CTX-M-14 (Alghoribi et al., 2015;Bogaerts et al., 2015;Ewers et al., 2010;Hu et al., 2013;Hussain et al., 2014;Kawamura et al., 2014;Kim et al., 2017;Mathers et al., 2015). Additionally, other STs described in wild animals, such as ST10, ST69, ST405, ST410, and ST648, have also been reported in various sources and are responsible for the intercontinental distribution of CTX-M (Fischer et al., 2014Hansen et al., 2014;Hu et al., 2013;Liu et al., 2016a;Su et al., 2016;Wang et al., 2016). However, some STs found in wildlife, such as ST1340, ST1646, ST2687, ST3018, and ST3056, have been identified as new types and have not yet been reported in human or veterinary isolates (Bonnedahl et al., 2010;Hasan et al., 2014;Jamborova et al., 2015).
In summary, the ESBL gene types identified in wild animals are the same as those in human and veterinary medicine. Thus, interspecies transmission mediated by successful pandemic ESBL-producing clones and plasmids in humans, domestic animals, and wildlife might occur.

PLASMID-MEDIATED AmpC β-LACTAMASE-PRODUCING ENTEROBACTERIACEAE FROM WILDLIFE
Plasmid-mediated AmpC β-lactamases among Enterobacteriaceae in human and veterinary medicine are of considerable global concern because they confer resistance to clinically important cephalosporin antibiotics and β-lactamase inhibitors (Jacoby, 2009;Smet et al., 2010). CMY-2 is the most prevalent AmpC βlactamase and has been globally disseminated among Enterobacteriaceae in humans, companion animals, foodproducing animals, and retail meat (Bogaerts et al., 2015;Carmo et al., 2014;Hansen et al., 2016;Jacoby, 2009;Ma et al., 2012;Smet et al., 2010;Vogt et al., 2014;Wu et al., 2015). As shown in Table 2, AmpC β-lactamases have been reported in E. coli, K. pneumoniae, and Enterobacter cloacae isolates of wildlife origin in Europe, North America, and Asia, and particularly in central Europe, similar to ESBL-producing Enterobacteriaceae (Table 2). To date, 20 different wild animal species, mostly birds, have been identified as bla CMY-2 carriers ( Table 2). Like that in human and veterinary medicine, bla CMY-2 is the most commonly detected AmpC-type β-lactamase among wild animals, though its identification has been limited to E. coli isolates.
The bla CMY-2 gene is mainly located on IncA/C and IncI1 plasmids from Enterobacteriaceae isolates of human and veterinary origin (Bogaerts et al., 2015;Bortolaia et al., 2014;Carattoli, 2009;Guo et al., 2014;Sidjabat et al., 2014). Among wild animals, bla CMY-2 has been reported to be mainly associated with IncI1 plasmids in E. coli from Dutch wild birds and wild seagulls in the US (Poirel et al., 2012;Veldman et al., 2013). Other plasmid types, such as IncB/O, IncK, and IncF, have also been identified as carriers of bla CMY-2 in wildlife (Poirel et al., 2012;Veldman et al., 2013), as previously reported in humans, companion animals, broiler chickens, and retail meat (Bortolaia et al., 2014;Hansen et al., 2016;Hiki et al., 2013;So et al., 2012;Vogt et al. 2014). Thus far, the IncA/C plasmid, a major carrier of bla CMY-2 , has not yet been identified in wild animals. The absence of bla CMY-2 -bearing IncA/C plasmids could simply reflect the limited studies on the characterization of bla CMY-2 -carrying plasmids in wildlife, or might indicate that bla CMY-2 -harbouring IncI1 plasmids are more successful among wildlife.
The first reported carbapenemase-producing bacteria in wild animals were isolated from black kites (Milvus migrans) in Germany (Fischer et al., 2013). Among 184 cefotaximeresistant Salmonella spp. isolates, only one Salmonella Corvallis isolate belonging to ST1541 has shown reduced susceptibility to carbapenem, and carries the carbapenemase gene bla NDM-1 located on ~180 kb IncA/C conjugative plasmid pRH-1738 (Fischer et al., 2013). The broad-host-range IncA/C plasmids are among the most predominant plasmids associated with bla NDM-1 in humans (Carattoli, 2013). Fischer et al. (2013) supposed that the bla NDM-1 -bearing Salmonella Corvallis isolate might have originated from non-European countries and was transferred to Germany through the black kite migratory route, since Salmonella Corvallis was prevalent in South-East Asia and was emerging in North Africa and Nigeria, rather than in European countries. The complete sequence of plasmid pRH-1738 further confirms this hypothesis. Plasmid pRH-1738 exhibited high relatedness with plasmid pMR0211 obtained from human Providencia stuartii isolate in Afghanistan, but showed distinct differences from other sequenced NDM-1-IncA/C 2 plasmids from Western countries (Villa et al., 2015). In addition, fosfomycin resistance gene fosA3, which has been rarely detected in Europe but is prevalent among CTX-Mencoding E. coli and K. pneumoniae isolates in Asia (i.e., China, Japan, and South Korea), has also been identified on NDM-1producing plasmid pRH-1738. bla NDM-1 transferred with fosA3 on IncA/C plasmid has only been described in clinical E. coli and Citrobacter freundii isolates in China (Qin et al., 2014). Taken together, these findings suggest that the origin of this plasmid might be in the Asiatic region.
Large-scale transmission of IMP-producing bacteria into wildlife was first reported in 2015. In total, 120 carbapenemaseproducing Enterobacteriaceae of 10 species were obtained from silver gulls in Australia, mainly E. coli (n=85), carrying bla IMP-4 , bla IMP-38 , or bla IMP-26 (Dolejska et al., 2016). The bla IMP-4 gene has been found in 116 isolates, and is the most commonly detected gene among carbapenemase-producing Enterobacteriaceae isolates in human clinics in Australia (Bell et al., 2016;Sidjabat et al., 2015). bla IMP-4 in gulls is carried by various conjugative plasmids, mostly IncHI2-N plasmid type, followed by IncA/C plasmids, as well as IncL/M and IncI1, and is associated with a class 1 integron-containing bla IMP-4 -qacG-aacA4-catB3 array in most positive strains (Dolejska et al., 2016). The same array carried by IncA/C and IncL/M plasmids is also reportedly responsible for the dissemination of bla IMP-4 in clinical isolates in Australia (Espedido et al., 2008), and by the IncHI2 plasmid in Salmonella Typhimurium from a cat in Australia (Abraham et al., 2016). Furthermore, 19 different STs have been detected in IMP-4-producing E. coli isolates, including five prevalent lineages (ST216, ST58, ST354, ST167, and ST224), in which ST58, ST354, and ST167 are clinically relevant clone lineages (Ben Sallem et al., 2015;Fernández et al., 2014;Huang et al., 2016).
Although carbapenem resistance reported in wild animals is rare, the emergence of NDM-1 and IMP carbapenemases in wild birds is of concern.

RESISTANCE GENE mcr-1 IN ENTEROBACTERIACEAE FROM WILDLIFE
Colistin is widely applied in food-producing animals and is currently used as the last resort for treating infections caused by multi-resistant gram-negative bacteria (Kaye et al., 2016). Since the first identification of the plasmid-mediated colistin resistance gene mcr-1 in China in 2015 (Liu et al., 2016b), it has been identified in Enterobacteriaceae isolates from foodproducing animals, companion animals, food products, the environment, and humans worldwide (Anjum et al., 2016;Doumith et al., 2016;Hasman et al., 2015;McGann et al., 2016;Xavier et al., 2016;Zhang et al., 2016;Zurfluh et al., 2016).
The role of wild birds as reservoirs and vectors for the global distribution of mcr-1 should be considered. Recently, mcr-1 was described in an E. coli strain isolated from European herring gull (Larus argentatus) feces collected from the Kaunas (Lithuania) city dump (Ruzauskas & Vaskeviciute, 2016). However, the emergence of mcr-1 in wildlife could be traced back to E. coli strains isolated in 2012 (Liakopoulos et al., 2016). Five extended-spectrum cephalosporin-resistant E. coli isolates obtained from kelp gulls in Ushuaia, Argentina in 2012 were found to carry mcr-1 and bla CTX-M-2 (n=1) and bla CTX-M-14 (n=4) and exhibited elevated colistin MICs (4-8 mg/L). The mcr-1 gene was located on a ~57 kb IncI2 plasmid without bla CTX-M in all five isolates. IncI2 plasmids, which have been detected in E. coli and Salmonella isolates from food, food-producing animals, and humans in China, Great Britain, the US, Venezuela, and Denmark, have been reported to be associated with the transmission of mcr-1 (Anjum et al., 2016;Delgado-Blas et al., 2016;Doumith et al., 2016;Hasman et al., 2015;Meinersmann et al., 2016;Yang et al., 2016). Notably, four mcr-1-carrying isolates, which belong to ST744, have been previously described in Denmark and carry the mcr-1-bearing IncI2 plasmid (Hasman et al., 2015;Liakopoulos et al., 2016).

CONCLUSIONS
Clinically relevant resistance, such as ESBL, AmpC cephalosporinase, carbapenemase, and colistin resistance, has been detected in wild animals, particularly wild birds, from distinct geographical areas. Thus, wild animals could serve as important reservoirs of resistant bacteria. Although the origin of bacterial resistance genes in wild animals remains unclear, as wildlife are not exposed to antibiotics directly, contact with sewage or animal manure might be one possibility (Wellington et al., 2013). Additionally, the potential of wild animals as vectors of resistant bacteria or genetic determinants should not be underestimated. Wildlife, especially migratory birds with their instinctive mobility, can carry resistant bacteria over long distances, even between continents; thus, this might be a new transmission route and partly responsible for the global dissemination of bacterial resistance. Contamination of food or water by wildlife is recognized as an important risk factor for the transmission of antimicrobial resistance or pathogens to food animals and humans (Greig et al., 2015).
Wild animals might play a vital role in the worldwide spread of clinically relevant pathogens or resistance genes. Pandemic ESBL-producing E. coli clones or plasmids shared by humans, domestic animals, and wildlife further strengthen this hypothesis. Thus, continued surveillance of multi-resistant bacteria in wild animals is warranted.