High Prevalence of Beta-Lactam-Resistant Escherichia coli in South Australian Grey-Headed Flying Fox Pups (Pteropus poliocephalus)

The emergence of antimicrobial-resistant Escherichia coli in wildlife is concerning—especially resistance to clinically important beta-lactam antibiotics. Wildlife in closer proximity to humans, including in captivity and in rescue/rehabilitation centres, typically have a higher prevalence of antimicrobial-resistant E. coli compared to their free-living counterparts. Each year, several thousand Australian fruit bat pups, including the grey-headed flying fox (GHFF; Pteropus poliocephalus), require rescuing and are taken into care by wildlife rescue and rehabilitation groups. To determine the prevalence of beta-lactam-resistant E. coli in rescued GHFF pups from South Australia, faecal samples were collected from 53 pups in care. A combination of selective culture, PCR, antimicrobial susceptibility testing, whole-genome sequencing, and phylogenetic analysis was used to identify and genetically characterise beta-lactam-resistant E. coli isolates. The prevalence of amoxicillin-, amoxicillin-plus-clavulanic-acid-, and cephalosporin-resistant E. coli in the 53 pups was 77.4% (n = 41), 24.5% (n = 13), and 11.3% (n = 6), respectively. GHFF beta-lactam-resistant E. coli also carried resistance genes to aminoglycosides, trimethoprim plus sulphonamide, and tetracyclines in 37.7% (n = 20), 35.8% (n = 19), and 26.4% (n = 14) of the 53 GHFF pups, respectively, and 50.9% (n = 27) of pups carried multidrug-resistant E. coli. Twelve E. coli strain types were identified from the 53 pups, with six strains having extraintestinal pathogenic traits, indicating that they have the potential to cause blood, lung, or wound infections in GHFFs. Two lineages—E. coli ST963 and ST58 O8:H25—were associated with human extraintestinal infections. Phylogenetic analyses determined that all 12 strains were lineages associated with humans and/or domestic animals. This study demonstrates high transmission of anthropogenic-associated beta-lactam-resistant E. coli to GHFF pups entering care. Importantly, we identified potential health risks to GHFF pups and zoonotic risks for their carers, highlighting the need for improved antibiotic stewardship and biosafety measures for GHFF pups entering care.


Introduction
The emergence and global dissemination of antimicrobial-resistant Escherichia coli is accelerating in human, domestic animal, and wildlife populations [1,2]. Of particular concern are multidrug-resistant E. coli exhibiting resistance to clinically important antibiotics, including broad-spectrum and extended-spectrum beta-lactams, carbapenems, and fluoroquinolones [3,4]. The rapid spread of antimicrobial resistance (AMR) in E. coli is largely attributed to mobile genetic elements (MGEs), such as plasmids and transposons, which facilitate the horizontal transfer of antimicrobial resistance genes between E. coli and other Gram-negative bacteria [5]. Class 1 integrons are capable of carrying diverse antimicrobial resistance genes and, via their association with MGEs, also play an important role in the emergence and dissemination of AMR in E. coli [6]. abandoned, or heat-stress-affected pups are rescued by South Australian wildlife rescue and rehabilitation groups, and are taken into care until they reach juvenile age and are deemed suitable for release back into the wild. This study investigated the prevalence of betalactam-resistant E. coli in GHFF pups from the ABP colony that were taken into care during the 2018-2019 pup season. Beta-lactam-resistant E. coli underwent genetic characterisation to identify antimicrobial resistance genes and VFs associated with the ExPEC and IPEC pathotypes, along with phylogenetic analysis to identify anthropogenic-associated lineages.

Collection of Faecal Samples and Associated Metadata from GHFF Pups
Faecal samples (n = 58) were collected from 53 orphaned, abandoned, or heat-stressaffected GHFF pups entering care, with 5 pups sampled twice-i.e., at entry to care and at exit from care (n = 10 faecal samples)-and 48  . Faecal samples were collected using the FecalSwab TM system (COPAN, Brescia, Italy) from clean materials placed under flying fox pups (n = 18) or by rectal swab during veterinary examination (n = 40). FecalSwab samples were stored at 4 • C and then delivered to the laboratory for processing at room temperature. Received FecalSwab samples that were not immediately processed were stored at −80 • C with 35% glycerol until cultured. Metadata were collected for all 53 pups, including date of entry to care, animal weight and forearm length (FAL) measurement, medical history, reason for entry to care, and sampling data (Supplementary Data S1). GHFF pup age was estimated by correlating FAL with the Pinson and Kerr ageing chart [43].

Detection of Amoxicillin-Resistant E. coli in Faecal Samples
Faecal samples (n = 58) were cultured in media supplemented with amoxicillin to detect beta-lactam-resistant E. coli. FecalSwab medium (0.2 mL) was inoculated into 5 mL of Luria-Bertani (LB) broth (Difco Laboratories, Detroit, MI, USA) supplemented with 10 mg/L amoxicillin (Sigma, St. Louis, MO, USA) and incubated at 37 • C overnight. The LB broth was subsequently streaked onto Chromocult Coliform Agar (Merck Millipore, Burlington, MA, USA) supplemented with 10 mg/L amoxicillin (Sigma, St. Louis, MO, USA) and incubated at 37 • C overnight. Dark blue or purple colonies were presumptively identified as E. coli, and the corresponding faecal samples were deemed positive for the presence of amoxicillin-resistant E. coli. A single E. coli colony was selected from each positive agar plate, except where morphologically distinct colonies were observed (as determined by colony and halo colour), in which case one representative isolate was selected for each colony type observed.

Phylotyping and Class 1 Integron Screening of Amoxicillin-Resistant E. coli Isolates
GHFF isolates (n = 49) were inoculated into 10 mL of LB broth (Difco Laboratories, Detroit, MI, USA) and incubated at 37 • C overnight. Genomic DNA was extracted using the ISOLATE II Genomic DNA Kit (Bioline, London, UK), and DNA concentrations were determined using a Qubit dsDNA BR assay kit (Invitrogen, Waltham, MA, USA).
The E. coli isolates were assigned to phylogroups (A, B1, B2, C, D, E, F, or cryptic clade I) using the Clermont quadruplex phylotyping PCR method [44]. Phylogroups were assigned based on the combination of four genes (chuA, yjaA, tspE4.C2, and arpA) [44]. PCR amplification was performed using GoTaq ® Green Master Mix (Promega, Madison, WI, USA), with all primers at 1 µM each, and cycling conditions of 94 • C for 4 min; 35 cycles at 94 • C for 20 s, 59 • C for 20 s, and 72 • C for 20 s; and a final extension at 72 • C for 4 min.
Isolate DNA was screened for the class 1 integron integrase gene (intI1) using PCR with the primers HS463a and HS464, as previously described [19]. All intI1-positive isolates were then amplified using the primers HS458 and HS459 (to target the attl1-conserved and 3 qacE∆1-conserved regions of the class 1 integron) [45] and the primers HS458 and JL-D2 (to target attl1 and IS26 transposase-an alternate 3 sequence to the qacE∆1-conserved region) [46], using cycling conditions of 94 • C for 3 min; 35 cycles at 94 • C for 30 s, 55 • C for 30 s, and 72 • C for 1 min 30 sec; and a final extension at 72 • C for 5 min. The HS458/HS459 and HS458/JL-D2 PCRs both amplified the entire class 1 integron gene cassette array.

ERIC and CGG PCRs of Amoxicillin-Resistant E. coli Isolates
To determine the strain types of all 49 E. coli isolates, genome fingerprints were obtained using enterobacterial repetitive intergenic consensus (ERIC) PCR [47] and CGG-PCR [48]. ERIC PCR amplification was performed using GoTaq ® Green Master Mix (Promega, Madison, WI, USA), ERIC1 and ERIC2 primers [47] at 0.5 µM each, 10 ng of DNA template, and cycling conditions of 95 • C for 7 min; 30 cycles of 94 • C for 1 min, 53 • C for 1 min, and 56 • C for 4 min; and a final extension at 65 • C for 16 min. CGG-PCR amplification was performed using GoTaq ® Green Master Mix (Promega, Madison, WI, USA), 0.5 µM N 6 (CGG) 4 primer [48], 3.5 mM MgCl 2 , 10 ng of DNA template, and cycling conditions of 95 • C for 3 min; 35 cycles of 95 • C for 1 min and 72 • C for 3 min; and a final extension at 72 • C for 8 min. The PCR products underwent electrophoresis, and DNA fingerprints were visualised on 2% agarose gels (1 × TBE buffer) with SYBR Safe DNA Stain (Invitrogen, Waltham, MA, USA).

Whole-Genome Sequencing of E. coli Strains
All 49 E. coli isolates were assigned to one of 12 strain types using a combination of phylotyping, ERIC, CGG, and class 1 integron PCR data (Supplementary Table S1). For each E. coli strain type identified (n = 12), one representative isolate was selected to undergo whole-genome sequencing (WGS) (Supplementary Table S1). WGS was performed on an Illumina MiSeq system using the MiSeq reagent kit v2 (2 × 150 bp paired-end reads) at the Ramaciotti Centre for Genomics (Sydney, NSW, Australia). Raw sequence reads were assembled as de novo genome sequences using SPAdes Assembler 3.15.2 [49] in Geneious Prime 2022.0.2 (Biomatters Limited, Auckland, New Zealand). Raw sequence reads for all 12 E. coli isolates were uploaded to EnteroBase (http://enterobase.warwick.ac.uk/species/ index/ecoli (accessed on 13 February 2021)) and the NCBI Sequence Read Archive (SRA) under BioProject PRJNA814458 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA814458 (accessed on 10 March 2022)). Individual isolate EnteroBase Barcodes and SRA accession numbers are provided in Supplementary Table S2.

Phenotypic Antimicrobial Susceptibility Testing of E. coli Strains
The 12 representative isolates selected for WGS-i.e., one isolate for each of the 12 E. coli strain types-underwent additional phenotypic antimicrobial susceptibility testing against a panel of 15 antibiotics representing 11 antimicrobial categories (Supplementary Table S3), using the European Committee on Antimicrobial Susceptibility Testing (EUCAST) criteria [60]. Isolates were evaluated as susceptible or resistant using the EUCAST breakpoint criteria (v 9.0 available at http://www.eucast.org/clinical_breakpoints/ (accessed on 22 October 2019)), or the Clinical and Laboratory Standards Institute (CSLI) breakpoint criteria (CLSI M100 ED29:2019 available at https://clsi.org/standards/products/free-resources/access-ourfree-resources/ (accessed on 22 October 2019)) when EUCAST breakpoints were unavailable. A GHFF negative control isolate-E. coli FF1170 (EnteroBase Barcode ESC_JA9915AA)was included as a quality control and to assist in evaluating zone diameters, as previously described [19]. Where breakpoint data were unavailable, resistance was defined as growth up to the edge of the disk, and intermediate resistance was evaluated as inhibition zone diameters smaller than the negative control isolate FF1170, but not to the edge of the disc. Multidrug resistance was defined as acquired phenotypic resistance to at least one antimicrobial in three or more categories [61].

Phylogenetic Analysis of E. coli Strains
Isolates phylogenetically related to the 12 GHFF E. coli strains were identified in EnteroBase by searching sequence types according to the Achtman seven-gene MLST scheme (available at https://enterobase.warwick.ac.uk/species/index/ecoli (accessed on 15 February 2022)) [52]. Phylogenetic comparison of EnteroBase and GHFF isolates was performed in EnteroBase using GrapeTree to construct rapid neighbour joining (RapidNJ) minimum spanning trees based on the core genome's multilocus sequence typing (cgMLST) V1 + hierarchical clustering (HierCC) V1 scheme [62]. Trees containing >500 isolates were refined, with all resulting GrapeTrees containing between 27 and 439 isolates. GitHub URL links for interactive versions of all GrapeTree cgMLST phylogenetic trees are provided in Supplementary Table S4.

GHFF Pup Metadata
GHFF pups (n = 53) ranged in estimated age from 4 days to 15 weeks (mean age, 75 days) at entry to care. Pups entering care in November or December 2018 were either found alone on the ground (n = 11) or with their deceased mother (n = 2), and ranged in age from 4 to 56 days (mean age = 30.2 days). Pups entering care on 24 January 2019 (n = 40) were affected by a heat stress event, and ranged in age from 76 to 105 days (mean age = 89.5 days). Of the 53 pups entering care, 15 were housed in small groups with one of four wildlife carers, and 38 pups (all heat-stress-affected) were co-housed in a single large outdoor aviary, as they were old enough to self-feed on provided fruit and water. Two pups were administered prophylactic antibiotics (amoxicillin plus clavulanic acid), one pup received topical antimicrobial ointment to treat wing burns, and the forty heat-stressed pups received subcutaneous fluids to treat dehydration. Sample collection data and associated metadata for individual GHFF pups are available in Supplementary Data S1.
Of the five pups that were sampled at both entry to care (day 0) and just prior to exit from care (day 28-60), amoxicillin-resistant E. coli were detected in three pups (60%) at entry and in four pups (80%) at exit (Figure 1a). For the GHFF pups co-housed in one large group (n = 38), the prevalence of amoxicillin-resistant E. coli was lowest in pups sampled after one day in care (75%, 15 of 20), and highest in pups sampled after three days in care (100%, 4 of 4) (Figure 1b).

Strain Typing of Amoxicillin-Resistant E. coli Isolates
A total of 49 amoxicillin-resistant E. coli isolates were cultured from 58 faecal samples that were collected from 53 individual GHFF pups (Supplementary Data S1). PCR phylotyping of the 49 isolates identified four different phylogroup pattern types, namely, A, A/C, B1, and D/E (Supplementary Table S1). Nineteen isolates were class 1 integron integrase (intI1)-positive and harboured one of two integron types based on their gene cassette array patterns identified in HS458/JL-D2 PCRs (Supplementary Table S1). The genome fingerprints (ERIC and CGG PCRs) were correlated with phylogroup PCR patterns and integron status, and all 49 isolates were assigned to one of 12 distinct E. coli strain types (Supplementary Table S1).

Whole-Genome Sequencing of Amoxicillin-Resistant E. coli Strains
One representative isolate from each of the 12 distinct strain types identified was selected to undergo whole-genome sequencing (WGS) (Supplementary Table S1). WGS identified 10 different sequence types (STs), with three strains belonging to the ST10 clonal complex (ST10 and ST48) and two strains to ST58 ( Table 1). The 12 strains all belonged to different serotypes, with 11 H-antigen types and seven O-antigen types identified, and four strains were O-non-typable (ONT) ( Table 1). Phylotyping of WGS assemblies in EnteroBase, using ClermonTyping, determined that the 12 strains belonged to three phylogroups: A (n = 6), B1 (n = 5), and D (n = 1) ( Table 1). Of the five pups that were sampled at both entry to care (day 0) and just prior from care (day 28-60), amoxicillin-resistant E. coli were detected in three pups (6 entry and in four pups (80%) at exit (Figure 1a). For the GHFF pups co-housed in on group (n = 38), the prevalence of amoxicillin-resistant E. coli was lowest in pups sa after one day in care (75%, 15 of 20), and highest in pups sampled after three days i (100%, 4 of 4) (Figure 1b).

Strain Typing of Amoxicillin-Resistant E. coli Isolates
A total of 49 amoxicillin-resistant E. coli isolates were cultured from 58 faecal sa that were collected from 53 individual GHFF pups (Supplementary Data S1). PCR typing of the 49 isolates identified four different phylogroup pattern types, name A/C, B1, and D/E (Supplementary Table S1). Nineteen isolates were class 1 integro grase (intI1)-positive and harboured one of two integron types based on their gene ca array patterns identified in HS458/JL-D2 PCRs (Supplementary Table S1). The ge fingerprints (ERIC and CGG PCRs) were correlated with phylogroup PCR pattern integron status, and all 49 isolates were assigned to one of 12 distinct E. coli strain (Supplementary Table S1).

Phenotypic Antimicrobial Resistance Profiles of Amoxicillin-Resistant E. coli Strains
Antimicrobial susceptibility testing of the 12 representative isolates selected for WGS against a panel of 15 antibiotics (Supplementary Table S3) identified additional phenotypic resistance to a total of 10 antibiotics from 8 antimicrobial categories, across the 12 E. coli strain types (Supplementary Table S5). Two of the twelve strains-ST963 and ST2144-also exhibited resistance to amoxicillin plus clavulanic acid, and ST963 was also resistant to firstand third-generation cephalosporins ( Table 1, Supplementary Table S5). Seven strains were resistant to tetracycline, three strains showed resistance to one aminoglycoside (spectinomycin or streptomycin), two strains were resistant to trimethoprim/sulfamethoxazole, and one strain was resistant to chloramphenicol (
All GHFF pups sampled in November or December 2018, and resampled in January 2019, carried only the E. coli strain ST2144 O166:H49 (n = 10), which was not detected in any of the 40 heat-stressed pups entering care on 24 January 2019 ( Figure 2). The 40 heat-stressed pups sampled in late January 2019 carried 11 different amoxicillin-resistant E. coli strains (Figure 2, Supplementary Data S1).

Phylogenetic Analysis of E. coli Strains with ExPEC Characteristics
GrapeTree phylogenetic analysis of GHFF E. coli strains with ExPEC characteristics (n = 6 strains), as well as closely related isolates identified in EnteroBase, placed two strains (ST963 ONT:H18 and ST361 O9:H30) in lineages predominantly composed of humansourced isolates, and four strains (ST58 O8:H25, ST58 ONT:H37, ST641 O70:H10, and ST1727 ONT:H14) in lineages composed of isolates sourced from animals and humans. All GrapeTree phylogenetic lineages containing the GHFF strains with ExPEC characteristics also contained human-and/or animal-sourced ExPEC isolates.
Phylogenetic analysis of ST963 ONT:H18 isolates (n = 184) placed the GHFF strain within an Australian-sourced cluster (Figure 3a). The Australian cluster contained ST963 ONT:H18 isolates predominantly cultured from human blood or urine (n = 11) and wild silver gull cloacal swabs (Chroicocephalus novaehollandiae) (n = 8) (Figure 3b). SNP tree analysis of the cluster determined that the GHFF ST963 ONT:H18 strain was most closely related to four human-sourced ExPEC isolates from blood (n = 2) and urine (n = 2) (Figure 3b).   Analysis of ST58 O8:H25 isolates (n = 439) showed that the isolates were sourced from diverse countries and were not clustered according to geographical location (Figure 4a). The GHFF ST58 O8:H25 fimH34 strain was the only Australian-sourced isolate in a large cluster (n = 35) belonging to the ST58 BAP2 sub-lineage (Figure 4a). The GHFF cluster contained isolates from diverse host species and source types, including human ExPEC (n = 4), canine ExPEC (n = 2), poultry and swine meat products (n = 7), and faeces/sewage (n = 12) (Figure 4b). In SNP tree analysis of the cluster, the GHFF ST58 O8:H25 strain was most closely related to an isolate cultured from human faeces in New Zealand (Figure 4b).
Microorganisms 2022, 10    The single ST58 ONT:H37 fimH32 GHFF isolate belonged to an ST58 ONT:H37 BAP4 sub-lineage (n = 130) composed predominantly of poultry-, wild-bird-, and livestocksourced isolates, plus five human-associated isolates (Figure 5a). The ST58 sub-lineage contained numerous APEC; however, the GHFF ST58 ONT:H37 isolate belonged to a small cluster containing isolates from one wild bird and goat faeces (Figure 5a).  Phylogenetic analysis of ST361 O9:H30 isolates (n = 323) determined that the majority were sourced from humans (n = 168), including ExPEC cultured from urine and blood, along with isolates associated with livestock and companion animals (Figure 5b). The GHFF ST361 O9:H30 strain belonged to a divergent lineage containing only two isolates, with the second being a canine ExPEC cultured from urine collected from a dog in Melbourne, Australia (Figure 5b).
The ST1727 phylogenetic tree (n = 153) was composed of isolates sourced from diverse animals, humans, and food products, and included ExPEC cultured from human and canine urine (Figure 5d). The GHFF ST1727 ONT:H14 strain belonged to a small cluster (n = 11) of ST1727 O-variable:H14 isolates from diverse sources, including dairy, meat, fish, and plant products, as well as human faeces (Figure 5d).

Figure 7.
Phylogenetic and metadata analysis of three low-pathogenicity E. coli isolates from greyheaded flying fox (GHFF) pups and closely related isolates from human, animal, and environmental sources identified in EnteroBase. GrapeTree phylogeny constructed using a rapid neighbour joining (RapidNJ) minimum spanning tree based on the cgMLST V1 + hierarchical clustering (HierCC) V1 scheme. Scale bars indicate the number of cgMLST allelic differences. GitHub URL links for interactive versions of all GrapeTrees are provided in Supplementary Table S4

Discussion
This study determined that the prevalence of amoxicillin-resistant E. coli in GHFF pups was extremely high (77.4%), and was not consistent with the low levels previously detected in adult GHFFs from the ABP colony (4.1%) [19]. Additionally, 24.5% of pups carried E. coli strains that were also resistant to amoxicillin plus clavulanic acid, and 11.3% carried E. coli resistant to first-and third-generation cephalosporins, neither of which have been detected in E. coli from adults in the ABP colony [19]. Amoxicillin, amoxicillin plus clavulanic acid, and 1st-3rd generation cephalosporins are considered "veterinary critically important antimicrobial agents", due to the wide variety of diseases they treat in a broad range of animal species [4]. Overall, 50.9% GHFF pups carried multidrug-resistant E. coli strains, in contrast to only 0.81% of adults from the ABP colony [19]. The amoxicillin-resistant E. coli strains from GHFF pups did not exhibit resistance to fluoroquinolones, fourth-and fifth-generation cephalosporins, or carbapenems, which are deemed critically important antimicrobials for human and veterinary medicine [3,4].
The high prevalence of amoxicillin-resistant E. coli in pups sampled at entry to care (60%) and within 24 h of entering care (75%) indicates that they acquired resistant E. coli prior to entering care. The increase in prevalence of E. coli ST2144 observed in five cohoused pups between entry to care and exit from care most likely occurred via horizontal transmission between GHFF pups. This finding is consistent with reports on adult GHFFs that found that the occurrence of beta-lactam-resistant E. coli and class 1 integrons was higher in captive adults (6.5% and 41.2%, respectively) compared to wild adults (3.5% and 5.3%, respectively) [19,45].
Several factors are known to influence the gut microbiota composition in young mammals of different species, and these may have contributed to the high prevalence of amoxicillin-resistant E. coli observed in GHFF pups. Firstly, E. coli is typically one of the first bacterial species to colonise the gut in newborn mammals [63], including humans [64], resulting in a higher prevalence and abundance of E. coli in babies compared to adults [18,[64][65][66]. Secondly, a higher prevalence of antimicrobial-resistant E. coli has been observed in the gut microbiota of pre-weaned young compared to adults-including in human infants [65,67], calves [68,69], and piglets [70]-and is mostly likely attributable to the greater abundance of E. coli in pre-weaned young compared to adults [65]. Thirdly, the loss and reacquisition of antimicrobial-resistant E. coli over time has been observed in co-housed young animals-for example, in calves, where antimicrobial-resistant E. coli are most likely lost as the gut microbiome matures, and then reacquired via horizontal transmission from other calves or from new environmental sources [69,71].
Environmental contamination with human and animal effluent, wastewater, stormwater, and agricultural runoff has been suggested as possible acquisition sources of antimicrobialresistant bacteria for wildlife [29]. As flying foxes are either in flight or arboreal, they display a "dipping" behaviour, which involves skimming across the surface of a freshwater body during flight to wet their belly fur and acquire drinking water. Dipping can potentially expose a flying fox, and the young pup clinging to its mother during the first month of life [42], to antimicrobial-resistant bacteria in water sources, when the wet fur is licked or when a pup suckles milk from its mother. Although wild GHFF pups begin to fly short distances and forage for food from about 3 months of age, they are not fully weaned until approximately 4-5 months old [42,72]. Female GHFFs form maternity roosts within a colony, where pups are left in crèche groups from approximately one month old while their mothers forage at night [42]. This close roosting behaviour of pups in crèches is also likely to facilitate the horizontal transmission of antimicrobial-resistant faecal bacteria between pups [34,73].
In pups that entered care during November and December 2018 (n = 13), only one strain of amoxicillin-resistant E. coli was detected (ST2144). In contrast, the GHFF pups that entered care on 24 January 2019 following a heat-stress event (n = 40) carried a high diversity of amoxicillin-resistant E. coli, with 11 distinct strains identified. These findings suggest that the E. coli strains present in the January heat-stressed pups were acquired from different environmental sources, and/or there was an increase in faecal bacterial contamination of environmental sources in January. During November and December 2018, the reported monthly rainfall totals in Adelaide were 36 mm and 20 mm, respectively; however, between 20 December 2018 and 24 January 2019, zero rainfall occurred [74]. Additionally, a three-daylong heatwave occurred in January 2019, with maximum temperatures reaching 38.9 • C and 40.9 • C on 22 and 23 January 2019, respectively, and 46.6 • C on 24 January, when the heatstress event occurred. These factors may have contributed to increased E. coli contamination in Adelaide water bodies utilised by GHFFs [75]. During heatwaves, nursing female GHFFs may preferentially utilise water and food sources that are located close to the colony, potentially exposing them to sources with increased levels of E. coli contamination [76].
In a study utilising Global Positioning System (GPS) telemetry to track the flight movements of adult GHFFs from the ABP colony, they were observed to visit several water sources, and presumably dipping to obtain drinking water [40]. These water sources included a lake in the Adelaide Botanic Gardens, and the River Torrens-both in close proximity to the GHFF colony-along with suburban drainage ponds and a quarry dam [40]. The River Torrens has been reported to have poor water quality, with faecal contamination identified in multiple reports [77][78][79]; thus, it is a potential source of amoxicillin-resistant E. coli for the ABP GHFF colony.
An additional consideration is that wildlife entering captivity-even for short periods of time-may acquire antimicrobial-resistant E. coli whilst in care, either directly from their human carers, or indirectly from environmental contamination by humans, or from other wildlife and domestic animals in close proximity [22,30,80]. However, given that 10 amoxicillin-resistant E. coli strains were detected in GHFF pups within the first 24 h of entering care, it would be unlikely that these were all acquired from the captive environment. Additionally, the same E. coli strains were detected in pups taken into care at different times and/or at different locations, further suggesting acquisition from wild environmental sources prior to entering care.
The prevalence and abundance of antimicrobial-resistant E. coli in mammals is reported to decrease with age [65,69]; however, it is currently unknown whether the amoxicillinresistant E. coli identified in pups' gut microbiota persist into adulthood. Notably, carriage of amoxicillin-resistant E. coli was maintained throughout the time in care (up to 60 days) in four GHFF pups. The tendency for bats to aggregate may facilitate horizontal transmission and maintain the long-term circulation of antimicrobial-resistant E. coli strains within a co-housed group or colony [34,73]. The E. coli ST48 O4:H26 strain detected in five GHFF pups was almost identical to an isolate detected in an adult GHFF from the ABP colony 11.5 months earlier [19]. Notably, both the pup and adult were recovering from heat stress and co-housed in the same rehabilitation facility, suggesting that the E. coli ST48 O4:H26 strain acquired by adults and pups originated from the same source, and has been circulating for some time in this GHFF population.
Phylogenetic analysis indicated that all 12 amoxicillin-resistant E. coli strains detected in GHFF pups originated from anthropogenic sources, with six strains known to cause extraintestinal infections in humans and/or animals [10]. Of particular concern was the detection of the human-associated ExPEC strain ST963 ONT:H18 in six GHFF pups, which also exhibited resistance to extended-spectrum beta-lactams. This GHFF strain belongs to an Australian E. coli ST963 clonal cluster, which is typically characterised by a chromosomal bla CMY-2 gene and an IncF F29:A-:B10 plasmid [81]. Australian ST963 clonal isolates have been widely detected in humans and silver gulls (Chroicocephalus novaehollandiae) across Australia [23,81], with genomic analysis indicating that interspecies transmission is occurring between humans and gulls [81]. Although phylogenetic analyses suggested that the GHFF pup ST693 strain was most closely related to Australian human-sourced isolates from blood and urine, rather than to gull-sourced isolates, it is likely that GHFFs are also part of the ST963 interspecies transmission pathways between humans and wildlife in Australia [81]. ST58 O8:H25 was the most frequently detected E. coli strain in the heat-stressed GHFF pups (n = 14). Although this ST58 BAP2 sub-lineage harbouring a ColV plasmid appears to have originated in poultry and swine, it is increasingly associated with blood and urinary tract infections in humans [53]. The high prevalence of amoxicillin-resistant E. coli strains with ExPEC characteristics in GHFF pups poses a significant health risk to pups, as these strains have the potential to cause pneumonia, as well as wound or blood infections [10]. These ExPEC strains-especially ST963 and ST58 O8:H25-also pose a significant zoonotic risk, and have the potential to cause extraintestinal infections in human carers [53,82].
Amoxicillin and amoxicillin plus clavulanic acid are probably the most commonly administered antibiotics to GHFF pups in care [43,83]. Given the high prevalence of resistance to these antibiotics in GHFF pups (77.4% and 24.5%, respectively), the need for adherence to antimicrobial stewardship guidelines, and only administering antibiotics when there is clinical infection, is paramount [84]. Two GHFF pups in this study received prophylactic doses of amoxicillin plus clavulanic acid upon entry to care, and both pups carried E. coli ST2144, which harbours a bla TEM-33 gene conferring resistance to amoxicillin plus clavulanic acid. In this scenario, there is the potential for selective pressure and risk of overgrowth of the resistant ST2144 strain in the gut microbiota, and although it was classified as having low pathogenicity, ST2144 does carry two significant virulence factors (astA and lpfA) that can contribute to extraintestinal disease [10]. Additionally, antimicrobial selection pressure, combined with other stressors-such as separation of the pup from its mother and changing the diet to artificial milk formula-creates increased potential for horizontal transfer of antimicrobial resistance genes and virulence genes between bacterial strains in the pups' gut microbiome [85]. Consequently, horizontal gene transfer may drive the emergence of new strains of multidrug-resistant bacterial pathogens in the pups' gut microbiome [84]. Given that heat-stress events and pup abandonments are likely to increase in the future due to climate change, there are likely to be increasing numbers of GHFF pups entering care each year [86,87], and many of these will require antimicrobial therapy. Administration of antibiotics to injured or sick GHFF pups carrying antimicrobial-resistant isolates may result in a poor response to treatment and a worsened prognosis for recovery and release back into the wild.
Implementing biosafety measures for GHFFs in captive settings is important for pup health and human health. Husbandry recommendations to limit the horizontal transmission of resistant bacterial strains between co-housed pups in care, and to minimise the biological risks to both pups and their carers, include (i) individual housing of GHFF pups receiving antimicrobial therapy, (ii) reducing the number of pups in co-housing cages, (iii) preventing direct contact between pups from different co-housed groups, (iv) dedicated equipment for each pup cage, (v) frequent cleaning and disinfection of pup-associated equipment and cages, (vi) wearing gloves and changing them between cages/pups, plus frequent hand washing, and (vii) excluding contact between pups and other wildlife or domestic animals.
Further work is required to examine the transmission dynamics and persistence of antimicrobial-resistant bacteria in GHFF pups as they are transitioned from care to placement in larger groups in "crèche" cages, in preparation for release back into wild colonies when they reach juvenile age [43]. Of the thousands of GHFF pups entering care each year, approximately 75% will be successfully reared in captivity and released back into the wild [87], along with any antimicrobial-resistant bacteria that they carry, which then have the potential for dissemination to other animals, humans, and into the environment.