The proliferation of antibiotic-resistant bacteria in the environment has potential negative economic and health consequences. Thus, previous investigations have targeted wild animals to understand the occurrence of antibiotic resistance in diverse environmental sources. In this critical review and synthesis, we summarized important concepts learned through the sampling of wildlife for antibiotic-resistant indicator bacteria. These concepts are helpful for understanding dissemination of resistance through environmental pathways and helping to guide future research efforts. Our review begins by briefly introducing antibiotic resistance as it pertains to bacteria harbored in environmental sources such as wild animals. Next, we differentiate wildlife from other animals in the context of how diverse taxa provide different information on antibiotic resistance in the environment. In the third section of our review, we identify representative research and seminal works that illustrate important associations between the occurrence of antibiotic-resistant bacteria in wildlife and anthropogenic inputs into the environment. For example, we highlight numerous investigations that support the premise that anthropogenic inputs into the environment drive the occurrence of antibiotic resistance in bacteria harbored by free-ranging wildlife. Additionally, we summarize previous research demonstrating foraging as a mechanism by which wildlife may be exposed to anthropogenic antibiotic resistance contamination in the environment. In the fourth section of our review, we summarize molecular evidence for the acquisition and dissemination of resistance among bacteria harbored by wildlife. In the fifth section, we identify what we believe to be important data gaps and potential future directions that other researchers may find useful toward the development of efficient, informative, and impactful investigations of antibiotic-resistant bacteria in wildlife. Finally, we conclude our review by highlighting the need to move from surveys that simply identify antibiotic-resistant bacteria in wildlife toward hypothesis-driven investigations that: 1) identify point sources of antibiotic resistance; 2) provide information on risk to human and animal health; 3) identify interventions that may interrupt environmentally mediated pathways of antibiotic resistance acquisition and transmission; and 4) evaluate whether management practices are leading to desirable outcomes.
Antibiotic resistance evolved in environmental bacteria in the absence of human inputs and long before the clinical use of antibiotics (D'Costa et al. 2011). However, selection pressures driving the occurrence of antibiotic resistance and dissemination of resistance determinants (i.e., genes or specific mutations conferring resistance) have changed considerably within the past century. For example, the selection for antibiotic-resistant bacteria as a result of the use of antibiotics to treat bacterial pathogens in humans (Rammelkamp and Maxon 1942; Spink 1954), for prophylactic and therapeutic treatment of food animal herds (Berghash et al. 1983), and as growth promoters in domestic animals (Starr and Reynolds 1951) was recognized within decades of the discovery and development of penicillin as the world's first antibiotic compound (Fleming 1929). More recently, it has been established that very low concentrations of antibiotic compounds (Gullberg et al. 2011) and heavy metals (Baker-Austin et al. 2006) may also select for and induce (Kohanski et al. 2010) antibiotic resistance. Contamination of the environment with antibiotic resistance determinants (henceforth, antibiotic resistance contamination) and drivers of resistance such as antibiotic residues, heavy metals, and biocides has become extensive, resulting from activities such as agricultural production (Sengeløv et al. 2003; Peak et al. 2007; Chee-Sanford et al. 2009), solid waste disposal (Wu et al. 2015; Song et al. 2016), and wastewater management (Bouki et al. 2013; Rizzo et al. 2013; Karkman et al. 2019). Thus, resistance that evolved as a function of cell signaling or through the biologic arms race among bacteria and fungi (Davies 2006) has become increasingly common in clinical and veterinary settings and now appears to be widespread throughout the environment.
The proliferation of antibiotic resistance in the environment is pertinent to both human and domestic animal health because the environmental resistome (i.e., the summation of antimicrobial resistant determinants in the environment) may contribute to the emergence and proliferation of pathogens that are difficult or impossible to treat (Wellington et al. 2013; Larsson et al. 2018). That is, in addition to risk posed by antibiotic-resistant pathogens maintained in environmental sources, antibiotic resistance determinants in the environmental resistome have the potential to be incorporated into human and domestic animal bacterial communities through horizontal gene transfer (Dolejska and Papagiannitsis 2018). Given the potential negative economic and health consequences of the proliferation of antibiotic-resistant bacteria in the environment, many researchers are embracing One Health approaches for investigating antibiotic resistance (Robinson et al. 2016), including assessments of resistance in environmental sources such as soils, water, and wild animals (Allen et al. 2010).
Scientific investigation of antibiotic resistance in wild animals has included the sampling of peridomestic species, zoo animals, and free-ranging wildlife. Investigations have employed cross-sectional survey and molecular approaches, resulting in data from diverse taxa and regions, which have been summarized in numerous review products on antibiotic resistance in wild animals (Bonnedahl and Järhult 2014; Greig et al. 2015; Arnold et al. 2016; Vittecoq et al. 2016; Wang et al. 2017; Dolejska and Papagiannitsis 2018) and in the environment (Allen et al. 2010; Berendonk et al. 2015; Huijbers et al. 2015; Surette and Wright 2017; Bueno et al. 2018). However, despite rather extensive efforts to identify antibiotic resistance in wildlife, and the publication of numerous review products summarizing previous findings, our perception is that additional synthesis of existing information would be useful for evaluating the utility of using wildlife as indicators of antibiotic resistance in the environment and guiding future research efforts.
In subsequent sections of this critical review and synthesis, we first distinguish among previous sampling efforts targeting antibiotic-resistant bacteria in peridomestic animals, animals in zoologic collections, and free-ranging wildlife in the context of the way in which we may differentially interpret information based upon associations with humans, other animals, and the environment. Next, we summarize results from representative research and seminal works that illustrate associations between the prevalence of antibiotic-resistant bacteria harbored by free-ranging wildlife and anthropogenic inputs into the environment. Subsequently, we review results of investigations employing molecular approaches to assess how wildlife acquires, exchanges, and disseminates antibiotic-resistant bacteria and present our perspective as to how such investigations may provide valuable mechanistic inference. Next, we identify what we believe to be important data gaps and potential future directions that we hope other researchers will find useful toward the development of efficient, informative, and impactful investigations of antibiotic-resistant bacteria in wildlife. Finally, we conclude our review by highlighting the importance of developing hypothesis-driven approaches for elucidating environmental pathways of antibiotic resistance dissemination, assessing risk to human and animal health, identifying potential intervention strategies for curbing the spread of resistance, and evaluating the effectiveness of management actions to limit dissemination. Our intent was not to provide an exhaustive summary of all published literature on antibiotic-resistant bacteria in wild animals in this review, as has been previously attempted (e.g., Radhouani et al. 2014; Greig et al. 2015; Ljubojević et al. 2016; Vittecoq et al. 2016; Wang et al. 2017; Dolejska and Papagiannitsis 2018). Rather, we sought to highlight representative research and seminal works that illustrate what we perceive to be important trends and themes.
PERIDOMESTIC ANIMALS, ZOO ANIMALS, AND FREE-RANGING WILDLIFE
Peridomestic animals, animals in zoologic collections, and free-ranging wildlife may all be referred to, at times, as wild animals. However, these diverse types of animals may have very different relationships with humans and often occupy habitats with varying levels of anthropogenic inputs. Therefore, these different types of animals likely have generally different levels of exposure to anthropogenic antibiotic resistance contamination. Thus, we might expect peridomestic animals, zoo animals, and wildlife to provide different information on the local environmental resistome.
Peridomestic animals, defined here as vertebrate taxa reliant upon human-dominated habitats (e.g., urban areas) or habitats intensively managed by humans (e.g., agricultural fields and feedlots), may be directly and indirectly exposed to anthropogenic antibiotic resistance contamination through agricultural activities, livestock production practices, waste disposal, and wastewater management. For example, black rats (Rattus rattus) inhabiting swine farms in The Netherlands were identified as carrying methicillin-resistant Staphylococcus aureus (MRSA) sequence types associated with human and swine infections (van de Giessen et al. 2009). Similarly, Norway rats (Rattus norvegicus) inhabiting the inner city of Vancouver, Canada, harbored MRSA lineages previously described in humans and livestock (Himsworth et al. 2014) and a methicillin-resistant Staphylococcus pseudintermedius sequence type previously found in domestic dogs (Himsworth et al. 2013). Small mammals such as mice (Mus musculus and Peromyscus spp.), rats (Rattus spp.), voles (Microtus pennsylvanicus), and shrews (Soricidae) sampled in Canada and Vietnam were five to eight times more likely to carry antibiotic-resistant Escherichia coli when sampled on or in the vicinity of livestock farms as compared to small mammals inhabiting more natural habitats or rice fields (Kozak et al. 2009; Nhung et al. 2015). Furthermore, peridomestic animals in Nairobi, Kenya, were found to harbor a high prevalence (52%) of clinically relevant multidrug-resistant E. coli, which statistical models suggested could be explained, in part, by exchange with humans, human refuse, livestock, and livestock manure (Hassell et al. 2019). Thus, peridomestic animals appear to be indicators of antibiotic resistance contamination in the anthropogenically dominated habitats they occupy.
Zoo animals have a different, and perhaps more complex, relationship with humans and human inputs. These animals are kept in captivity for educational and breeding purposes, typically monitored for health and well-being, and provided veterinary care when necessary. Thus, bacteria harbored by zoo animals may face direct selection pressures driven by therapeutic doses of antibiotics or subtherapeutic doses excreted by cohoused animals receiving treatment. Furthermore, zoo animals may be exposed to antibiotic-resistant bacteria (both pathogens and commensals) via husbandry practices (Dobiasova et al. 2013). While zoonotic infections of humans with pathogens originating from zoologic collections have been infrequently reported (Bender et al. 2004), as has reverse zoonotic transmission of pathogens from humans to zoo animals (Messenger et al. 2014), there is little evidence that either transmission pathway is common or serves as an important source of antibiotic resistance in either human or zoo animal populations. Thus, zoo animals primarily provide information on resistance in and among zoologic collections rather than the greater human community or the surrounding natural environment. Examples of studies highlighting the infection of zoo animals with antibiotic-resistant bacteria include a study of Salmonella among a zoologic collection in Trinidad in which 66% of isolates were resistant to one or more antibiotics (Gopee et al. 2000), an investigation of gram-negative bacteria among a zoologic collection in Japan in which 21% of isolates were resistant to two or more antibiotic compounds (Ahmed et al. 2007), and an investigation of E. coli in captive nonhuman primates from six zoos in China in which 32% and 15% of isolates exhibited extended-spectrum b-lactam and plasmid-mediated quinolone resistance, respectively (Wang et al. 2012).
In contrast to peridomestic and zoo animals, free-ranging wildlife, defined here as wild vertebrates with no history of domestication or subjection to animal husbandry practices, and which often occupies habitats lacking human habitation, generally has less direct interaction with humans. Therefore, wildlife typically has less obvious direct exposure to anthropogenic antibiotic resistance contamination. There are important exceptions to this latter generalization in that numerous wildlife species have become increasingly dependent on anthropogenic food sources or otherwise synurbic (Fedriani et al. 2001; Duhem et al. 2008). Furthermore, many wild animals are now less able to avoid anthropogenic inputs because of the expansion of human settlements and agricultural areas into previously unfragmented wildlife habitats. Free-ranging wildlife generally does not receive antibiotics or veterinary care, although exceptions apply. These may include animals receiving antibiotic treatment during disease outbreaks, interventions involving endangered species (Spelman et al. 2013), and treatment following capture and handling as part of translocation efforts (Weiser et al. 2009). Thus, the prevalence of antibiotic-resistant bacteria harbored by wildlife generally appears to be primarily influenced by the relative level of exposure to anthropogenic antibiotic resistance contamination and through associated selection pressures imparted within the environment, with the exception of animals receiving veterinary care. The remainder of this review will focus on antibiotic resistance in bacteria harbored by free-ranging wildlife.
CROSS-SECTIONAL STUDIES OF ANTIBIOTIC-RESISTANT BACTERIA IN WILDLIFE AND APPARENT TRENDS
Numerous investigations have employed cross-sectional surveys to explore the prevalence of antibiotic-resistant bacteria in wildlife inhabiting a range of environments (summarized in Bonnedahl and Järhult 2014; Greig et al. 2015; Arnold et al. 2016; Wang et al. 2017; Dolejska and Papagiannitsis 2018). Collectively, these studies provide support for the premise that free-ranging wildlife likely experiences differential exposure to antibiotic-resistant bacteria, or alternatively, variable selection pressures driving the acquisition of antibiotic-resistant determinants by commensal bacteria (Fig. 1). Furthermore, the prevalence of antibiotic-resistant bacteria among wildlife appears to be dependent on a variety of factors, such as habitat use and foraging strategy of the species sampled, particularly as they relate to anthropogenic inputs into the environment (Fig. 1).
Cross-sectional studies have found that wildlife inhabiting remote regions of Alaska (Ramey et al. 2018), Antarctica (Bonnedahl et al. 2008; Rahman et al. 2008), including sub-Antarctic islands (Palmgren et al. 2000), and the Galapagos Islands (Thaller et al. 2010; Wheeler et al. 2012) exhibits limited or no evidence of carriage of antibiotic-resistant indicator bacteria. In contrast, wildlife sampled at more anthropogenically influenced sites in Alaska (Atterby et al. 2016), on the Antarctic Peninsula (González-Acuña et al. 2013) and sub-Antarctic Islands (Cerdà-Cuéllar et al. 2018), and on the Galapagos Islands (Wheeler et al. 2012) more commonly harbored antibiotic-resistant bacteria. When considering less remote regions of the world, wildlife sampled in habitats impacted by livestock production (Cole et al. 2005; Guenther et al. 2010; Sulzner et al. 2014; Mercat et al. 2015), biosolid fertilizers (Rogers et al. 2018), solid waste disposal (Rolland et al. 1985; Gómez et al. 2015), and wastewater treatment (Furness et al. 2017; Marcelino et al. 2019) has generally been found to have a higher prevalence of antibiotic-resistant bacteria as compared to less impacted habitats, although several exceptions have been reported (Routman et al. 1985; Katakweba et al. 2015; Swift et al. 2019). While numerous factors may influence the prevalence of antibiotic-resistant bacteria in wildlife, evidence overwhelmingly supports the premise that anthropogenic inputs into the environment contribute to the prevalence of antibiotic-resistant bacteria harbored by wildlife occupying those habitats. As such, wildlife may be good indicators of the burden of resistance within the local environment and may therefore be useful for identifying potential point sources of anthropogenic antibiotic resistance contamination.
Another factor associated with the prevalence of antibiotic-resistant bacteria harbored by wildlife is the foraging strategy of the species sampled. Foraging represents a mechanism by which wildlife may be exposed to anthropogenic antibiotic resistance contamination in the environment and could therefore be useful in predicting those wildlife species that may be sensitive indicators of the environmental resistome (Fig. 1). For example, in southern France, 19% (18/93) of samples from Yellow-legged Gulls (Larus michahellis), omnivorous birds commonly associated with foraging at landfills, yielded carbapenem-resistant E. coli isolates, but carbapenem-resistant isolates were not obtained from marine-feeding Slender-billed Gulls (Chroicocephalus genei; Vittecoq et al. 2017). Other surveys targeting synurbic Yellow-legged Gulls throughout Europe (Bonnedahl et al. 2009; Stedt et al. 2014; Migura-Garcia et al. 2017; Vergara et al. 2017) have consistently identified antibiotic-resistant E. coli and other indicator bacteria (e.g., Salmonella and Campylobacter). Another cross-sectional study targeting gulls at a remote colony in the Gulf of Alaska (Ramey et al. 2018) found antibiotic-resistant E. coli isolates in 15% (10/65) of samples collected from Glaucous-winged Gulls (Larus glaucescens), another omnivorous species commonly associated with landfill foraging, but did not find evidence for resistance in 65 samples collected from pelagic-feeding Black-legged Kittiwakes (Rissa tridactyla) inhabiting the same small island. Furthermore, Glaucous-winged Gulls sampled at or near a landfill in the same region of Alaska had a higher prevalence of antibiotic-resistant bacteria as compared to gulls inhabiting the remote Gulf of Alaska colony site (Atterby et al. 2016). Other investigations sampling landfill-foraging gulls in Australia, Canada, and Chile also found a relatively high prevalence (30–47%) of antibiotic-resistant indicator bacteria (Hernandez et al. 2013; Bonnedahl et al. 2015; Dolejska et al. 2015). Collectively, these studies support landfill foraging as a mechanism by which omnivorous birds may be exposed to anthropogenic antibiotic resistance contamination.
Consumption of livestock carcasses or peridomestic animals appears to be another foraging strategy that may be associated with the acquisition of antibiotic-resistant bacteria by wildlife. Cross-sectional studies of Egyptian Vultures (Neophron percnopterus), Eurasian Griffon Vultures (Gyps fulvus), and Turkey Vultures (Cathartes aura) have found evidence for antibiotic exposure, and the prevalence of antibiotic-resistant bacteria was associated with scavenging of livestock carcasses (Sulzner et al. 2014; Casas-Díaz et al. 2016; Blanco 2018). Additional investigations of avian scavengers and other raptors have also consistently found evidence for carriage of antibiotic-resistant bacteria (Pinto et al. 2010; Radhouani et al. 2010; Molina-López et al. 2011, 2015; Guenther et al. 2012; Radhouani et al. 2012; Mora et al. 2014). Furthermore, a survey of diverse wildlife species in Botswana found the prevalence of antibiotic-resistant E. coli was highest in carnivores (62.5%) and animals using urban and peri-urban habitats (25.6%) as compared to herbivores (9.1%) and animals using protected habitats (9.0%) within a national park (Jobbins and Alexander 2015). Collectively, these studies support the premise that scavenging of carcasses or consumption of peridomestic prey may promote exposure to antibiotic resistance contamination.
In summary, cross-sectional studies provide extensive evidence for associations between the prevalence of antibiotic-resistant bacteria in wildlife and habitat use, foraging strategy, and anthropogenic inputs into the environment (Fig. 1). Therefore, studies that fail to report the ecologic context in which wildlife are sampled provide limited information regarding the environmental pathways through which antibiotic resistance may be acquired and dispersed.
GENETIC ANALYSES OF ANTIBIOTIC-RESISTANT BACTERIA IN WILDLIFE AND INFERENCE REGARDING ACQUISITION AND DISSEMINATION
As discussed in previous text, assessing trends in prevalence of antibiotic-resistant bacteria among different wildlife host species using cross-sectional approaches may be informative. However, additional inference regarding the acquisition, maintenance, and spread of antibiotic-resistant bacteria harbored by wildlife can be obtained with genetic data, particularly when information is combined with other data streams. For example, using a combination of phenotypic and genotypic characterization of E. coli, numerous studies have reported evidence for the exchange of antibiotic-resistant bacteria between humans and wildlife. Rwego et al. (2008) examined isolates from humans, livestock, and mountain gorillas (Gorilla gorilla beringei) in Uganda and found correlations between habitat overlap and the prevalence and genetic similarity of antibiotic-resistant gastrointestinal bacteria. That is, the proportion of mountain gorillas harboring antibiotic-resistant E. coli and the genetic relatedness of gorilla-origin bacteria relative to those detected in livestock and humans were positively correlated to habitat overlap. Banded mongoose (Mungos mungo) living in close proximity to humans in Botswana frequently harbored antibiotic-resistant E. coli that shared a high degree of genetic similarity with isolates recovered from human waste (Pesapane et al. 2013). A study examining E. coli isolates from multiple sources in Portugal found quinolone resistance to be more common in human clinical isolates and those from gulls as compared to E. coli from urban streams, wastewater treatment plants, or birds of prey (Varela et al. 2015). Furthermore, genetic similarity was highest among isolates from clinics, gulls, and wastewater, suggesting a plausible environmental pathway of dissemination. Another investigation in Sweden similarly assessed the prevalence of antibiotic-resistant E. coli, the occurrence of specific genes conferring resistance, and the degree of genetic similarity of bacterial genotypes among samples from gulls, humans, livestock, and surface water (Atterby et al. 2017). Though the authors found limited evidence for clonal transmission of antibiotic-resistant bacteria between gulls and humans, they identified common E. coli sequence types and resistance genes among gulls, humans, and surface waters, also suggesting a plausible route of indirect environmental transmission. Collectively, all of these studies provide further evidence that antibiotic-resistant bacteria are exchanged between humans and wildlife, presumably via environmental pathways, although specific transmission routes have yet to be elucidated.
High-resolution genomic sequencing approaches, combined with additional data streams, have extended inference regarding how wildlife may acquire and disseminate resistance through space and via contact with sympatric animals. For example, Ahlstrom et al. (2018) compared genomes of antibiotic-resistant E. coli isolates recovered from sympatric gulls and Bald Eagles (Haliaeetus leucocephalus) inhabiting a south-central Alaska landfill and found evidence of shared resistance genes and bacterial strains. Clonal E. coli isolates with conserved core genomes and identical antibiotic resistance gene profiles among gulls and bald eagles indicated cross-species transmission (e.g., via fecal-oral transmission or predation) or exposure via the same point source (e.g., the landfill). Furthermore, through a combination of satellite tracking of landfill-foraging gulls and whole genome sequencing of E. coli isolates, the genetic population structure of antibiotic-resistant bacteria from gull feces collected at numerous locations within this same region of Alaska was found to generally mirror gull movements, providing evidence for local dispersal of antibiotic-resistant E. coli by gulls (Ahlstrom et al. 2019a). The highest genetic similarity was found between isolates from gull feces collected at the landfill site and the mouth of a nearby river, the same two sampling locations between which gulls made most local movements. Miller et al. (2019) used a combination of whole genome sequencing of antibiotic-resistant E. coli isolates recovered from giraffes (Giraffa camelopardalis) sampled in Kenya and social network analysis to assess bacterial dissemination among hosts. They found poor support for transmission of antibiotic-resistant E. coli among giraffes, suggesting that carriage of such bacteria within this system may be driven by independent acquisition events from point sources rather than circulation among the population after primary introduction. Additional investigations incorporating whole genome sequencing of antibiotic-resistant bacterial isolates, high-resolution animal tracking or habitat-use data, information on contact among individuals, and data derived from medical and veterinary clinical settings would be useful toward improving our collective understanding of environmental transmission networks and the mechanisms driving them (Arnold et al. 2016).
PERCEIVED DATA GAPS AND POTENTIAL FUTURE DIRECTIONS
While the quantity, resolution, and accessibility of information regarding antibiotic-resistant bacteria in wildlife have improved tremendously in recent decades, we contend that considerable data gaps remain. Important data gaps include better assessment of the applicability of wildlife to serve as efficient and meaningful sentinels of a community-relevant environmental resistome, the utility of wildlife in source attribution studies, and the role of wildlife in the dissemination of resistance through environmentally mediated pathways. We believe that these perceived data gaps will be most efficiently bridged through future hypothesis-driven approaches incorporating information from multiple data streams.
Surveillance of antibiotic-resistant bacteria in wastewater has already been used as an early warning system to monitor resistance trends in human populations (Kwak et al. 2015). However, it is unclear whether wildlife could also serve as efficient sentinels of resistance pertinent to either domestic animal or human health. The finding that the prevalence of antibiotic-resistant indicator bacteria in gulls may be higher than in the local human population in countries within Europe, North America, and South America (Hernandez et al. 2013; Stedt et al. 2014; Bonnedahl et al. 2015; Atterby et al. 2017) suggests that synurbic gulls could be amplifiers of resistance and therefore informative surveillance targets. Therefore, hypothesis-driven approaches to assess if resistance determinants may be amplified among bacterial communities maintained by landfill-foraging gulls or other synurbic wildlife surveillance targets (e.g., animals using wastewater treatment plants) would be informative, particularly if they were to incorporate high-resolution genomic comparisons with bacteria isolated from either local domestic animal or human populations.
The detection of resistance to critically important antimicrobials, such as colistin and carbapenems, in wildlife (Fischer et al. 2013; Dolejska et al. 2015; Liakopoulos et al. 2016; Ruzauskas and Vaskeviciute 2016; Sellera et al. 2016; Papagiannitsis et al. 2017; Vittecoq et al. 2017; Bachiri et al. 2018a, 2018b; Bouaziz et al. 2018; Köck et al. 2018; Ahlstrom et al. 2019b, 2019c; Mukerji et al. 2019) suggests that there may be promise for using wild animals as indicators of a clinically relevant environmental resistome (Dolejska and Papagiannitsis 2018). However, the significance of these detections in wildlife is still yet to be determined from a public health perspective. While there is evidence for wildlife origin zoonoses (Ejidokun et al. 2006; Gilsdorf et al. 2006; Kwan et al. 2014) and introductions of pathogens from wildlife to domestic animals (Crispell et al. 2017; Li et al. 2018), directional estimates of the transmission of antibiotic-resistant pathogens across the wildlife–domestic animal and wildlife-human interfaces are rare. Additionally, directionality of transmission (e.g., wildlife to humans or humans to wildlife) is difficult to decipher, even with the highest resolution molecular tools (Biek et al. 2012; Webster et al. 2017). Thus, the utility of using wildlife as an early detection system for emergent antibiotic-resistant pathogens of clinical importance is limited in the absence of additional information on associated risk to humans and domestic animals (Ashbolt et al. 2013; Martínez et al. 2015). This highlights a need for longitudinal investigations that relate information on the prevalence and genomic characteristics of antibiotic-resistant bacteria in synurbic wildlife to data derived from spatiotemporally proximate clinical human and veterinary populations.
Information on antibiotic-resistant bacteria in wildlife may also be valuable to future source attribution investigations. Often applied in the context of food safety, source attribution aims to estimate the proportion of infections in a population of interest attributable to a variety of potential sources based on the prevalence and characteristics of the infectious agent found in each source (Hald et al. 2004). However, such models can be extremely complex, particularly for antibiotic resistance, because they rely upon surveillance data from diverse sources, and the appropriate target to investigate (e.g., an antimicrobial resistant bacterial strain, gene, or plasmid) is not always clear (Sheppard et al. 2016; Manaia 2017; Pires et al. 2018). Recent progress has been made toward incorporating information on antibiotic-resistant bacteria in wildlife into comparisons of antimicrobial resistance determinants among different sources (Dorado-García et al. 2017; Thomas et al. 2017; Jamborova et al. 2018; McCann et al. 2019; Mughini-Gras et al. 2019); however, data are currently insufficient to detect clear trends in prevalence and genetic similarity among samples originating from wildlife, humans, agricultural commodities, and the environment. Therefore, there is a need for additional high-resolution genomic information on antibiotic-resistant bacteria from wildlife and other diverse sources that may be involved in transmission pathways to facilitate interpretation in source attribution investigations. High-resolution data from diverse hosts may also facilitate better understanding of genetic determinants favoring successful colonization, associations between plasmids and clinically relevant resistance genes, and cocarriage of resistance and determinants of biocide resistance or virulence.
While efficient use of wildlife as indicators of a clinically relevant environmental resistome or for resolving pathways of antibiotic resistance dissemination through source attribution studies may still be elusive in the short and medium term, researchers may immediately be able to strengthen inference derived from investigations through the simultaneous sampling of wildlife and the environments they occupy. That is, it may be informative to identify and sample potential point sources of antibiotic resistance contamination in the environment and to relate data derived from such samples to information on the prevalence and genetic characteristics of antibiotic-resistant indicator bacteria among the wildlife population investigated. For example, though landfill-foraging gulls tend to harbor relatively high levels of antibiotic-resistant indicator bacteria, it has not yet been clearly demonstrated whether landfills serve as an important source of antibiotic resistant determinants or whether this foraging strategy is an indicator of other behaviors or point sources that mediate acquisition of resistance determinants. Agricultural settings (Cole et al. 2005; Mathys et al. 2017) and wastewater treatment facilities (Alroy and Ellis 2011; Masarikova et al. 2016; Amos et al. 2014; Vredenburg et al. 2014; Akiba et al. 2016) are other obvious examples of potential point sources of antibiotic resistance contamination to which wildlife may be exposed. As such, the application of hypothesis-driven research approaches incorporating the sampling of wildlife and these types of environments would be valuable for improving the rigor of inference regarding dissemination pathways.
In the absence of environmental sampling, we encourage researchers to report results regarding the prevalence of antibiotic-resistant bacteria in wildlife in the context of the overlap of habitat use with local human populations and anthropogenic inputs into the environment. Given the overwhelming evidence that these variables influence the prevalence of antibiotic resistance in free-ranging animals, such information is requisite for deriving meaningful inference regarding dissemination of resistance determinants via environmentally mediated pathways (Chamosa et al. 2017) and for assessing whether management actions may help to control resistance spread (Hagedorn et al. 1999; Pruden et al. 2013). Furthermore, this information will facilitate evaluation of the effectiveness of management practices to curb the dissemination of resistance through environmental pathways.
There is overwhelming evidence that wildlife often harbors antibiotic-resistant indicator bacteria, particularly when animals occupy habitats in close proximity to humans and exhibit behaviors that facilitate exposure to anthropogenic antibiotic resistance contamination in the environment (i.e., foraging at landfills or scavenging of livestock carcasses). However, it is less clear what risks antibiotic-resistant bacteria may pose to wild animal health (e.g., protected populations that may require veterinary care) and the health of human and domestic animal populations that coinhabit wildlife habitat. As such, there is a need to move from surveys that simply identify antibiotic-resistant bacteria in wildlife toward investigations that employ One Health approaches to: 1) identify specific mechanisms and point sources from which wildlife is acquiring antibiotic-resistant bacteria, 2) provide information on risk associated with the maintenance, amplification, or spread of antibiotic resistance clinically relevant to human and animal health, 3) identify interventions (i.e., specific management practices) that may interrupt environmentally mediated pathways of antibiotic-resistant bacterial acquisition/transmission, and 4) evaluate whether management practices are leading to desirable and measurable outcomes.
Financial support for this project was provided by the US Geological Survey through the Contaminants Biology Program of the Environmental Health Mission Area and through the Wildlife Program of the Ecosystems Mission area. We appreciate reviews of prior versions of this article provided by J. Bonnedahl, J. Pearce, D. Mulcahy, and two anonymous reviewers. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government.