1. Introduction
The rise of antimicrobial resistant (AMR) bacteria is a serious threat to human and animal health, as increasing resistance to commonly used antibiotic therapies have created a burden on treatment options [
1]. AMR bacteria can arise in nature and are commonly found in food producing animals like poultry [
2]. For instance, the gastrointestinal tract (GIT) of chickens and the facilities that house these poultry serve as reservoirs for AMR resistant bacteria [
3,
4]. Of these AMR bacteria, extended-spectrum β-lactamase (ESBL) producing
Enterobacteriaceae, carbapenem-resistant
Acinetobacter (CRA), and colistin resistant
E. coli (CRE) are emerging AMR bacteria found in the poultry environment [
5,
6,
7]. The Centers for Disease Control and Prevention (CDC) lists ESBL-producing
Enterobacteriaceae and CRA as serious and urgent threats, respectively [
8]. Although not explicitly listed as an antibiotic resistant threat by the CDC, colistin resistance is clinically relevant given its use as a last resort antibiotic for treating multidrug resistant (MDR) infections [
9]. As poultry is one of the most consumed meat sources globally [
10], it is crucial to identify the presence and prevalence of these AMR populations in the poultry environment and understand the spread of these resistances to other bacteria.
Bacteria primarily acquire AMR genes by horizontal gene transfer (HGT), a leading contributor to bacterial coevolution [
11]. Conjugative plasmids are responsible for HGT of virulence and AMR genes, which has led to the rapid rise of AMR in bacterial pathogens [
12,
13]. Recently, the transfer of mobile colistin resistance (
mcr) and ESBL-producing genes have been linked to a variety of plasmid types and bacterial hosts in the poultry environment [
14,
15]. As AMR populations persist in the poultry environment, there is an increased risk that pathogens might acquire AMR genes.
Although
E. coli and
Acinetobacter are commensal gut bacteria in poultry and are detected in the feed, feces, and environment of poultry facilities, both
E. coli and
Acinetobacter have the potential to cause extraintestinal diseases in both humans and poultry [
16,
17,
18,
19,
20,
21,
22]. Importantly, extraintestinal pathogenic
E. coli (ExPEC) infections are often highly fatal in humans and poultry and are increasing worldwide, imposing a major burden on public health [
23]. As pooling of these potential pathogens and AMR genes are taking place in poultry, it is important to detect both AMR genes and associated virulence markers that can identify potential pathogens. There are limited studies that investigate the role of production environments (i.e., conventional cage [CC] versus cage-free [CF]) and maturity stages (i.e., early, peak, and late lay) on AMR emergence. In commercial farms, layer hens can be categorized by the period in which egg production begins (early), is at its highest (peak), and later diminishes due to age (late). Studies have shown that the maturity stage of layer hens can impact the colonization and shedding of particular bacteria and the diversity of bacteria inhabiting the GIT [
24,
25]. Recently, our study has shown that different layer maturity stages exhibit differing levels of
Enterobacteriaceae in CC and CF conditions [
26]. We thus hypothesized that both environment and maturity may play a role in AMR diversity and potential virulence detection.
In this study, we examined ceca contents from hens in commercial CC and CF environments as potential reservoirs for CRA, CRE, and ESBL-producing Enterobacteriaceae. Potential spread of these resistances was examined as well as the virulence potential of E. coli isolates. We were able to identify MDR Acinetobacter, ESBL-producing E. coli, and widespread presence of CRE in both CC and CF environments. Phylogenetic and virulence screening identified possibly MDR ExPEC isolates in both environments. Finally, AMR was demonstrated to be transferable in the CC environment via plasmid mediated HGT.
4. Discussion
Commensal bacteria like
Enterobacteriaceae and pathogens isolated from the chicken GIT are commonly resistant to multiple antibiotics [
40]. Commonly,
E. coli isolates sourced from chickens have shown resistance to tetracycline, chloramphenicol, ampicillin, and streptomycin due to the indiscriminate use of these antibiotics as growth promoters in poultry production [
41,
42]. Concurrent with these studies, our results show the extensive antibiotic resistance patterns associated with
E. coli from both CC and CF farms in all maturity stages. Furthermore, the amount of CRE isolated from each maturity stage did not differ. Alarmingly, numerous CRE isolates from both CC and CF showed MDR to multiple antibiotics tested. Recent research has shown that the maturity of layer hens plays a significant role in the diversity of the ceca microbiota [
26]. Although the detection of CRE is increasing in the poultry environment [
15,
43,
44], there is little information on how maturity can impact the presence of these AMR bacteria. Factors such as disease and antibiotic intervention can drive the emergence of AMR bacteria in the poultry environment [
45]. As the presence of MDR bacteria from chickens increases, the detection of specific resistance genes illustrates a relationship between resistant bacteria and the spread of resistance mechanisms. Future studies should investigate the role that the environment and possible intervention strategies that can limit the emergence of colistin resistant bacteria.
Our study highlights the importance of detection methods for colistin resistant bacteria using media like CHROMagar COL-APSE™. Notably, CHROMagar COL-APSE™ adheres to the European Committee on Antimicrobial Susceptibility Testing (EUCAST). The EUCAST breakpoint for colistin resistance is strictly set at 2 mg/L, whereas CLSI resistance cutoff in the United States is 4 mg/L. All isolates that were initially selected from CHROMagar COL-APSE™ were identified as resistant on MHA according to the EUCAST breakpoint of 2 mg/L [
46]. The number of isolates initially selected from the CHROMagar COL-APSE™ media was greatly reduced after screening using MHA with 4 mg/L colistin. Using both CHROMagar COL-APSE™ media and further selection using CLSI guidelines can be useful for accurate detection of resistant isolates. Although use of CHROMagar COL-APSE™ and further confirmation with either CLSI or EUCAST guidelines can identify multiple resistant bacteria, these methods do not accurately detect the presence of
mcr genes [
47]. PCR investigation of all confirmed colistin resistant bacteria should be performed to confirm the presence of the emerging
mcr genes.
The emergence of
mcr genes in multiple settings have become an increasing global threat [
14,
48]. Notably, the principle
mcr-1 gene has been reported in over 40 countries to date [
49]. None of the
mcr genes tested in our study (
mcr-1,
mcr-2,
mcr-3,
mcr-4, and
mcr-5) were detected in any of our CRE isolates. Although our isolates did not carry the
mcr genes that are often spread through conjugative plasmids [
50], our study highlights the ability of AMR
E. coli from chickens to exhibit spontaneous resistance to colistin. There are several mechanisms that can contribute to the ability of bacteria to become spontaneously resistant to colistin. For instance, the most common resistance strategy involves the modification of the bacterial outer membrane through alteration in the lipopolysaccharide (LPS) and reduction in its negative charge as this negative charge is the target of colistin [
51]. Furthermore, overexpression of efflux-pump systems and overproduction of capsule polysaccharides can enhance the resistance to colistin in
Enterobacteriaceae [
52,
53]. The possibility of threatening AMR bacteria like ESBL-producing
E. coli and MDR
Acinetobacter to exhibit resistance to colistin spontaneously is concerning because of colistin’s use as a last resort antibiotic [
54].
As evidenced in this study, the detection of ESBL-producing
E. coli and MDR
Acinetobacter can be detected in the poultry environment. In order to understand the genetic attributes that contributed to the resistance of these isolates, we investigated multiple β-lactamase and carbapenemase producing genes in both groups. Specifically, we attempted to detect the
blaTEM,
blaSHV,
blaCTX-M, and
blaCMY β-lactamase genes as these β-lactamases are the most commonly identified [
55]. Furthermore, we attempted to detect carbapenemase genes
blaOXA,
blaKPC, and
blaVIM that are commonly found in CRA [
56,
57,
58]. As none of the β-lactamase and carbapenemase genes were detected in the CRA isolates, other resistance mechanisms are likely responsible for the resistance. Carbapenem resistance in
Acinetobacter can be mediated by reduced drug permeability through porin loss or modification and the overexpression of efflux pump [
59,
60]. Interestingly, we successfully identified the AmpC Class C β-lactamase encoding
blaCMY gene in only the early lay period of both CC and CF environments.
blaCMY has been increasingly reported in chickens as a source for the spread of AmpC to both human and avian pathogenic bacteria [
61,
62]. Although
blaCMY is not classified as an extended-spectrum β-lactamase, the AmpC β-lactamase is resistant to otherwise useful β-lactamase inhibitors like clavulanic acid that is often paired with antibiotics, like ceftazidime, to limit the effects of β-lactamases [
63,
64]. Strikingly, in our study the ESBL gene
blaTEM was detected in all cephalosporin resistant CC
E. coli, illustrating the ability of AMR
E. coli to harbor multiple β-lactamase producing genes.
To evaluate whether CRE and ESBL-producing
E. coli detected in this study can be pathogenic, we tested our isolates using
E. coli phylogenetic typing, a common technique utilized to sort
E. coli isolates into groups that differ in ecological niches, life-history characteristics, and propensity to cause disease [
65]. Namely, four main
E. coli phylotypes (A, B1, B2, and D) are extensively used for classification [
66,
67,
68]. We identified A, B2, and D groups in all maturity stages in both environments. Historically, phylogenetic group A is more commonly associated with commensal
E. coli, whereas phylogenetic groups B2 and D are associated with virulent ExPEC infections [
69,
70]. For instance, phylogenetic groups B2 and D have been associated with
E. coli isolates that cause urinary tract infections (UTI) and show increased presence of virulence [
71] factors associated with UTI when compared to the B1 and A phylogenetic groups.
Additionally, because of their plasmidic location, we attempted to detect four ExPEC-associated virulence factors (
cvaC,
iroN,
iss, and
iutA) among our isolates [
33].
cvaC and
iroN were not identified in any of our isolates. However, there was wide distribution of both
iss and
iutA in isolated from both environments. Interestingly, the late lay period of the CC isolates exhibited markedly less
iss and
iutA than the early and peak lay periods. Encoding the receptor for the siderophore aerobactin,
iutA is an important factor in urinary pathogenic
E. coli (UPEC) infections that allow the bacteria to competitively acquire
iroN that would otherwise be acquired by the host [
72]. Furthermore, the
iss virulence factor encodes a specific outer membrane protein that increases serum survival for ExPEC during extraintestinal infection [
73]. In addition to their role in the pathogenesis, these virulence factors could confer a competitive advantage to isolates that express them compared to their counterparts in the GIT.
The ability for
E. coli to spread AMR genes through plasmids via HGT plays a significant role on the distribution of resistance in different bacteria. Several factors play a role in the ability of plasmids to be transferred between bacteria. For instance, incompatibility restriction, host genetics, and strain-specific factors can influence the transfer of plasmids [
38,
74,
75]. In this study, we investigated a total of 18 plasmid incompatibility groups in all
E. coli isolates. We were able to identify several isolates that carry IncN, P, HI1, and I1 incompatibility groups that have been shown to be associated with a broad-host range [
76,
77,
78]. Interestingly, in our study, all ESBL-producing
E. coli in the CC environment carried IncK/B type plasmids. Due to their multiple plasmids visualized in plasmid extraction, and presence of the common IncK/B plasmids, the CC ESBL-producing
E. coli were selected for conjugation assays. CF
E. coli that carried the
blaCMY gene did not yield any of the plasmid replicon types investigated in this study, and conjugation was not observed under any conditions. Furthermore, CRA isolates did not yield any plasmid replicon types investigated in this study, and no plasmids were visualized using the method described or using IBI-Scientific (Dubuque, IA, USA) and Qiagen (Hilden, Germany) commercially available plasmid extraction kits. This suggests that the AMR genes of these isolates could be chromosomal rather than plasmidic.
In this study, we demonstrated the ability of the ESBL-producing
E. coli from CC to transfer AmpC β-lactamase via a large 106 kb IncK/B plasmid (pAmpC) to a variety of hosts, including the APEC strain χ7122, its avirulent plasmid-cured derivative χ7368, and
E. coli K-12 χ6092, indicating the ability of the plasmid to transfer to both virulent and non-virulent bacteria. Interestingly, numerous studies have highlighted the presence of pAmpC associated with both IncK plasmids and the poultry environment [
79,
80,
81,
82]. Although transfer of AMR genes and plasmidic virulence factors on the same plasmid has been shown [
83,
84], the ESBL-producing donors in our study did not transfer the virulence factors
iss and
iutA to the avirulent strains. This suggests that the plasmidic virulence factors are located on different plasmids than the pAmpC in the donor strains. Nonetheless, the presence of pAmpC in populations of MDR bacteria with the propensity to cause disease is concerning. Finally, increased detection of pAmpC can cause an eventual decline in the efficacy of cephalosporin antibiotics in both humans and poultry.