High prevalence and diversity of extended-spectrum β-lactamase and emergence of OXA-48 producing Enterobacterales in wildlife in Catalonia

Most of the studies focused on antimicrobial resistance (AMR) performed in wildlife describe Escherichia coli as the principal indicator of the selective pressure. In the present study, several species of Enterobacterales with a large panel of cephalosporin resistant (CR) genes have been isolated from wildlife in Catalonia. A total of 307 wild animals were examined to determine the prevalence of CR enterobacteria, AMR phenotypes and the presence of common carbapenem and CR genes. The overall prevalence of CR-phenotype was 13% (40/307): 17.3% in wild mammals (18/104) and 11.5% in wild birds (22/191) (p<0.01). Hedgehogs showed the highest prevalence (13.5% of 104) of the mammal specimens, and raptors the highest in bird specimen (7.3% of 191). Although CR E. coli was the most frequently isolated (45%), other CR- Enterobacterales like Klebsiella pneumoniae (20%), Citrobacter freundii (15%), Enterobacter cloacae (5%), Proteus mirabilis (5%), Providencia spp (5%) and Serratia marcescens (2.5%) were also isolated. A high diversity of CR genes was identified among the isolates, with 50% yielding blaCMY-2, 23% blaSHV-12, 20% blaCMY-1 and 18% blaCTX-M-15. Additionally, resistance to carbapenems associated to OXA-48 gene was found. Most of the CR isolates, principally K. pneumoniae and C. freundii, were multi-resistant with co-resistance to fluoroquinolones, tetracycline, sulphonamides and aminoglycosides. This study reports high prevalence of Enterobacterales harbouring a variety of CR genes and OXA-48 mediated-carbapenem resistance, all of them frequently associated to nosocomial human infections, for the first time in wild mammals and wild birds. Implementation of control measures to reduce the impact of anthropogenic pressure in the environment is urgently needed.


Study population
Wild animals attended at the Wildlife Rehabilitation Centre (WRC) of Torreferrusa (Catalonia, North-East Iberian Peninsula) were analysed between November 2016 and May 2017. This is a public WRC under the direction of the Catalan Wildlife-Service ("Direcció General de Polítiques Ambientals, Departament de Territori i Sostenibilitat of the Generalitat de Catalunya"). Sampling methods and handling protocols of animals were in agreement with the Catalan Wildlife Service who stipulates the management protocols and Ethical Principles according to the Spanish legislation [24]. All animals were examined and tested using cloacal or rectal swabs on arrival at the centre before receiving any pharmacologic or antimicrobial treatment. The most frequent cause of hospitalization was related to anthropogenic origin due to direct persecution (gunshot, poisoning, illegal captivity or traps) to involuntary human induced threats (collisions with vehicles, fences or electric lines and electrocution).

Microbiological analysis
Rectal and cloacal swabs were plated in MacConkey agar supplemented with ceftriaxone (1mg/L). Single colonies growing on the plate were subculture and identified biochemically using API (bioMérieux, Marcy l'Etoile, France) or VITEK 2 (bioMérieux, Spain) systems.

Molecular characterization of antimicrobial resistance genes
Sanger DNA sequencing was done for bla TEM , bla SHV , bla CTX-M , and bla OXA PCR products at the Genomic and Bioinformatics Service of the Universitat Autònoma de Barcelona (Spain). Sequences and chromatograms were manually explored to trim bad-quality bases with BioEdit 7.2. Once the assembly of the consensus sequences was done, partial sequences were aligned using Clustal Omega program, and finally blasted against the public database (National Center for Biotechnology Information, NCBI). Allelic variants of the ESBL-resistance genes were determined based on these partial sequences, and AmpC genes were classified according to the CMY-1 and CMY-2 groups. Statistical analysis. Descriptive analysis was performed under 95% confidence, using SPSS Advanced Models TM 15.0 (SPSS Inc. 233 South Wacker Drive, 11th Floor Chicago, IL 60.606-6412). The Chi-square test or Fisher exact test was used for comparison between proportions when appropriate. Statistically significant results were considered for unadjusted pvalue < 0.05.

Results
The sample size comprised 307 wild animals belonging to 67 different species grouped as birds (62%), mammals (34%) and reptiles (4%) (Fig 1). Animals came from different regions of Catalonia with a high density of urban areas and pig farming production.
A high genetic diversity of CR encoding genes was observed in all Enterobacterales, with 40% (16/40) of the isolates harbouring 2 to 5 different resistance genes in the same isolate (Table 1). Furthermore, the carbapenemase-encoding gene OXA-48 was detected in E. coli and P. mirabilis isolated from European hedgehog and Barn owl, respectively (Table 1). Other carbapenemase-encoding genes tested were not found.
Most of the ESBL/AmpC Enterobacterales isolated (92%), with the exception of E. cloacae, were multiresistant with a common resistance phenotype comprising β-lactams-quinolonestetracycline-sulfamethoxazole/trimethoprim (Table 1). K. pneumoniae and C. freundii isolates both presented a multi-drug resistance profile including the resistance to aminoglycosides ( Table 2). Moreover, 90% of the K. pneumoniae isolates were resistant to ciprofloxacin and  sulphametoxazole, 70% to kanamycin, 55% to streptomycin, and 10% to florfenicol. Additionally, 83% of the tested C. freundii isolates exhibited resistance to trimethoprim and nalidixic acid and 67% to tetracycline ( Table 2). Although none of the mcr-genes were detected in this study, the colistin resistant phenotype was observed in Klebsiella spp isolated from a European greenfinch and Algerian hedgehog, and in a Providencia spp isolated from a common buzzard.

Discussion
This study identifies for the first time a high percentage of wild mammals and wild birds as carriers of potential nosocomial Enterobacterales harbouring diverse ESBL, CMY and OXA-48 genes. Moreover, most of the isolates principally K. pneumoniae and C. freundii, presented a high prevalence of resistance also to fluoroquinolones. In general, E. coli is the most reported ESBL/pAmpC-producing enterobacteria worldwide, with increasing frequency from animals, food, environmental sources and humans. In recent years, CR-E. coli transmission has been reported in different hosts, demonstrating a close human-animal ESBL/pAmpC gene similarity between livestock (broilers and pigs) and personnel working at the farms [13]. Additionally, similar CR genes have been reported between isolates from the community and those from human clinical settings, sewage water and wild birds [13].
Although ESBL transmission has been studied extensively in Enterobacterales from humans and livestock, data on antimicrobial resistance in the environment is still limited [2]. Moreover, most of the studies related to ESBL-carrying bacteria in wildlife are focused on the wild bird population and mainly restricted to E. coli species [32]. Several studies conducted in E. coli from avian species have identified bla CTX-M-1 , bla CTX-M-14, bla CTX-M-15 and bla SHV-12 as the predominant ESBL types circulating in Spain [15,[33][34][35][36], Portugal [37], Tunisia [20], The Netherlands [38], Poland [39] and the Czech Republic [40]. Contrarily, in the present study, bla CTX-M-1 and bla CTX-M-14 were not detected in our avian species, but bla SHV-12 and bla CTX-M-15 were the most frequent ESBL types identified not only in E. coli but also in K. pneumoniae and C. freundii isolates. K. pneumoniae has been described in low prevalence (average 1.5%) in wild gulls from different European countries [41][42][43], including wild migratory birds from Spain, which exhibited bla CTX-M-15 ESBL-producing K. pneumoniae [36]. Interestingly, both bla CTX-M-15 and bla  are also currently the most predominant genes in human clinical specimens from community and health care-associated infections in Spain [44,45]. Thus, the human community could potentially be a source of ESBL environmental contamination, through water contaminated with human sewage from urban areas and hospital settings. In this study, bla CMY-1 group was principally detected in E. coli, K. pneumoniae, Proteus and Providencia spp from avian wildlife, like hawks, owls and small forest birds. Although this is an unusual variant in Spain, the presence in the present study might be explained by those species feeding habits. Raptors are predators occupying the top of the food chain; therefore, they can acquire AMR from a wide variety of preys (mammals, birds, reptiles or scavenging livestock). Moreover, some of these raptors are migratory species, being exposed to different environmental habitats in their migratory movements. In consequence, the role of migratory raptors as disseminators of these AMR traits is a serious concern to be further investigated.
Regarding bla CMY-2, it is the most common CMY type reported worldwide [46]. In this study, bla CMY-2 group was highly detected in E. coli and K. pneumoniae from hedgehogs and wild birds. Plasmid mediated genes can spread easily to other organisms. C. freundii, Enterobacter and Serratia spp in this study presented genes of the CMY-2 family. Since these types of AmpC genes are chromosomally encoded in some of these bacteria species, we cannot conclude the plasmidic nature of such enzymes. However, for epidemiological studies, it is important to report this type of resistance since these Enterobacterales can be involved in severe nosocomial infections and they all presented a MDR profile, except for E. cloacae.
Surprisingly, European hedgehogs represented an important reservoir of ESBL/AmpC-producing E. coli and other Enterobacterales, especially for bla CMY-2 (67%) and bla SHV-12 (25%) in this study. Our results are in agreement with previous studies conducted in Spain reporting low to moderate (1.3%-10%) prevalence of bla CMY-2 and bla SHV-12 E. coli variants in hedgehogs, deer and minks [22,47]. It is important to highlight that hedgehogs are in close contact with humans (home range including gardens), but also with livestock in the countryside, which could explain their acquisition of these AMR types.
Plasmid-mediated colistin resistance by mcr-1 has been reported worldwide in Enterobacterales isolated from humans, livestock, companion animals, food and wildlife [48]. Colistin has been used in veterinary medicine during the last decades for the treatment of gastrointestinal infections in livestock, principally in pigs and poultry [49]. Consequently, livestock is considered the main reservoir of mcr-1 selection and dissemination worldwide. Recent works disclosed the relationship among mcr-1-harbouring E. coli isolates recovered from the environment, pig production and human clinical isolates, demonstrating the rapidly evolving epidemiology of plasmid-mediated colistin-resistant E. coli strains worldwide and the importance of the One Health approach [50,51]. In our study, some Klebsiella and Providencia spp isolates were phenotypically resistant to colistin, but no mcr-associated genes were detected in any of the examined isolates. Information about carbapenem-resistant Enterobacterales is very scarce in wildlife and has only been reported in avian species [36,52]. In this study, we report the presence of bla  in E. coli and P. mirabilis isolates from a European hedgehog and a Barn owl, respectively. The presence of bla OXA-48 in wild mammals and birds in Catalonia is highly indicative of the wide environmental pollution of this variant, commonly reported in hospitals in Spain [53].
To our knowledge, there are no reports in wildlife, especially in wild mammals, describing the presence of ESBL genes in such a variety of Enterobacterales, like Klebsiella spp, Citrobacter spp, Serratia spp, or Enterobacter spp in Spain. C. freundii, is considered an opportunistic pathogen, associated with nosocomial infections, especially in patients who have been hospitalized for a prolonged period of time. In the last years, this bacterium has been classified as an emerging health problem associated to urinary tract infections commonly diagnosed in healthcare settings [54]. E. cloacae has been reported as important opportunistic and multi-resistant pathogen involved in outbreaks of hospital-acquired infections worldwide [55,56], including Spain [57]. ESBL-S. marcescens has also been classified as one of the top ten priority pathogens causing infections in intensive care units [58].
The high prevalence of CR Enterobacterales encountered in this study is really concerning, since wildlife is not directly exposed to any antimicrobial agent. Therefore, faecal contamination of water or soil with MDR bacteria and/or antimicrobial residues can lead to a selection pressure. Wastewaters from urban areas and hospitals have been identified as one of the major sources of AMR environmental contamination [2]. High prevalence of bla SHV-12 but also bla-TEM-1 and bla CTX-M-1 alleles have been reported in aquatic environments (urban waters, natural or artificial water reservoirs, seawater or drinking water) in several countries worldwide, likely due to their relatively easy transmission to surface water through waste water treatment plant discharges [2,59]. In our study, wildlife in close contact with urban and farming areas of Catalonia carried a large variety of zoonotic/nosocomial bacteria genetically resistant to β-lactams-quinolones-tetracycline-sulfamethoxazole/trimethoprim-aminoglycosides with similar resistant genes to those found in livestock and clinical settings. However, further studies are needed to assess clonal relatedness among different cephalosporin and carbapenem resistant enterobacteria at the human-animal-environment interface.

Conclusions
This study describes for the first time a high prevalence of Enterobacterales harbouring a large variety of ESBL in addition to carbapenem resistant OXA-48 genes in wild mammals, remarkably in hedgehogs, and wild birds in Catalonia (northeast Spain). AmpC CMY-2 group and the ESBL genes bla SHV-12 and bla CTX-M-15 were the most frequent types identified in E. coli, K. pneumoniae and C. freundii isolates. These results support the concept that wildlife is a good sentinel of AMR environmental contamination and underline the importance of the One Health approach since wildlife can contribute indirectly to the dissemination of resistance genes into other natural environments.
Supporting information S1