Antimicrobial Resistance in Escherichia coli from the Broiler Farm Environment, with Detection of SHV-12-Producing Isolates

Antimicrobial resistance is an important One Health challenge that encompasses the human, animal, and environmental fields. A total of 111 Escherichia coli isolates previously recovered from manure (n = 57) and indoor air (n = 54) samples from a broiler farm were analyzed to determine their phenotypes and genotypes of antimicrobial resistance and integron characterization; in addition, plasmid replicon analysis and molecular typing were performed in extended-spectrum-beta-lactamase (ESBL) producer isolates. A multidrug-resistance phenotype was detected in 46.8% of the isolates, and the highest rates of resistance were found for ampicillin, trimethoprim–sulfamethoxazole, and tetracycline (>40%); moreover, 15 isolates (13.5%) showed susceptibility to all tested antibiotics. None of the isolates showed imipenem and/or cefoxitin resistance. Twenty-three of the one hundred and eleven E. coli isolates (20.7%) were ESBL producers and carried the blaSHV-12 gene; one of these isolates was recovered from the air, and the remaining 22 were from manure samples. Most of ESBL-positive isolates carried the cmlA (n = 23), tet(A) (n = 19), and aac(6′)-Ib-cr (n = 11) genes. The following genetic lineages were identified among the ESBL-producing isolates (sequence type-phylogroup-clonotype): ST770-E-CH116–552 (n = 12), ST117-B2-CH45–97 (n = 4), ST68-E-CH26–382/49 (n = 3), ST68-E-CH26–49 (n = 1), and ST10992-A/B1-CH11–23/41/580 (n = 4); the latter two were detected for the first time in the poultry sector. At least two plasmid replicon types were detected in the ESBL-producing E. coli isolates, with IncF, IncF1B, IncK, and IncHI1 being the most frequently found. The following antimicrobial resistance genes were identified among the non-ESBL-producing isolates (number of isolates): blaTEM (58), aac(6′)-Ib-cr (6), qnrS (2), aac(3)-II (2), cmlA (6), tet(A)/tet(B) (22), and sul1/2/3 (51). Four different gene-cassette arrays were detected in the variable region of class 1 (dfrA1-aadA1, dfrA12-aadA2, and dfrA12-orf-aadA2-cmlA) and class 2 integrons (sat2-aadA1-orfX). This work reveals the worrying presence of antimicrobial-resistant E. coli in the broiler farm environment, with ESBL-producing isolates of SHV-12 type being extensively disseminated.


Introduction
Antimicrobial resistance (AMR) is one of the biggest problems that face health authorities worldwide. For a long time, AMR has been focused on the clinical setting, but nowadays, it is clear that this is a problem that involves humans, animals, and the environment, and AMR bacteria can be disseminated throughout the environment, which requires a One Health approach. The emergence and the spread of AMR are closely related to the were cleaned with mobile machinery. Before loading new flocks, the farm was disinfected. Broiler farm samplings were conducted at 15-16 days after chick entry, and temperature and relative humidity inside the building were 27 ± 1 • C and 70%, respectively. Air samples were taken inside the farm by stationary and mechanical sampling methods (Air Ideal), and non-antibiotic supplemented Chromocult coliform agar (CCA, Merck) plates were used for E. coli recovery. The manure samples were collected aseptically at the same time and place. Free mobility of the chickens in the enclosure ensured the homogeneity of the litter. Microbiological analysis was performed on 10 g of each litter sample homogenized with 90 mL of sterile peptone water. Serial dilutions of homogenized litter samples were performed and spread on the surface of CCA agar plates. By this procedure, 54 isolates of indoor air and 57 of manure from the farm were obtained. These isolates were identified by MALDITOF-MS (Matrix-Assisted Laser Desorption/Ionization) (Bruker Daltonik, Bremen, Germany) and preserved at −80 • C, and they were included and characterized in the present study.

Characterization of Antimicrobial Resistance Genes, Integrons, and Plasmids
DNA extraction from E. coli isolates was performed using the boiling method; briefly, one colony of an overnight culture was suspended in 1 mL of MilliQ water, and later, it was boiled for 8 min to break down the cell wall and centrifuged at 12,000 r.p.m. for 2 min to remove the pellet. DNA concentration was checked using the Nanodrop spectrophotometer.
In addition, the presence of integrase genes (int1 and int2) of class 1 and class 2 integrons and their variable regions were analyzed by PCR "primer-walking" strategy and subsequent sequencing [11,25].
Plasmid incompatibility (Inc) groups were determined by the PCR-based replicon typing (PBRT) [26] in all the ESBL-producing E. coli isolates.

Molecular Typing
All the isolates of the collection were assigned to different phylogroups by multiplex PCR assay [27]. Multilocus sequence typing (MLST) of the ESBL-producing isolates was performed by PCR and sequencing of seven housekeeping genes, according to the guidelines available at https://pubmlst.org/bigsdb?db=pubmlst_ecoli_achtman_seqdef (accessed on 10 January 2022), to determine the sequence type (ST) and the clonal complex (CC). Clonotype identification was performed by sequencing of fumC and fimH genes (CH) [28].

Statistical Analysis
Statistical analysis was carried out in IBM SPSS 26.0 using the chi-square test to compare pairwise differences in resistance rates to each antimicrobial agent between indoor air and manure samples and also among ESBL-and non-ESBL producers. A p-value < 0.05 is considered a statistically significant difference. Statistical analysis of the correlation between phenotypic and genotypic resistance was performed by calculating Pearson's correlation coefficient with R Study (SAS Institute Inc., Cary, NC, USA).

Antimicrobial Susceptibility Testing
Fifteen of the one hundred and eleven E. coli isolates (13.5%) showed susceptibility to all antimicrobial agents tested, and the remaining isolates (86.5%) exhibited resistance to at least one agent, with 52 isolates being considered as MDR (46.8%). High resistance rates were obtained for AMP, SXT, TET, and CIP (75.7-38.7%) and lower rates for CHL, CAZ, CTX, AMC, and GEN (27-10%). No IMP or FOX resistances were found in this collection of isolates. In addition, significant differences were observed between indoor air and manure isolates for AMP, AMC, CAZ, CTX, CIP, CHL, SXT, and TET resistance rates ( Figure 1).

Statistical Analysis
Statistical analysis was carried out in IBM SPSS 26.0 using the chi-square test to compare pairwise differences in resistance rates to each antimicrobial agent between indoor air and manure samples and also among ESBL-and non-ESBL producers. A p-value < 0.05 is considered a statistically significant difference. Statistical analysis of the correlation between phenotypic and genotypic resistance was performed by calculating Pearson's correlation coefficient with R Study (SAS Institute Inc., Cary, NC, USA).

Antimicrobial Susceptibility Testing
Fifteen of the one hundred and eleven E. coli isolates (13.5%) showed susceptibility to all antimicrobial agents tested, and the remaining isolates (86.5%) exhibited resistance to at least one agent, with 52 isolates being considered as MDR (46.8%). High resistance rates were obtained for AMP, SXT, TET, and CIP (75.7-38.7%) and lower rates for CHL CAZ, CTX, AMC, and GEN (27-10%). No IMP or FOX resistances were found in this collection of isolates. In addition, significant differences were observed between indoor air and manure isolates for AMP, AMC, CAZ, CTX, CIP, CHL, SXT, and TET resistance rates ( Figure 1). Figure 1. Percentages of antibiotic resistance among the total E. coli isolates (n = 111) and from those obtained from manure (n = 57) and indoor air (n = 54). Abbreviations: AMP-ampicillin; AMCamoxicillin clavulanate; CAZ-ceftazidime; CTX-cefotaxime; CIP-ciprofloxacin; GEN-gentamicin; CHL-chloramphenicol; SXT-trimethoprim/sulfamethoxazole; TET-tetracycline. No resistance for imipenem and cefoxitin was identified among the isolates. * p-value < 0.05. Resistance rates for CIP, CHL, and TET were higher among ESBL-positive isolates, but the pattern was different for GEN and SXT. It is important to highlight that significant differences were detected between ESBL-and non-ESBL-producing isolates for CIP, CHL, SXT, and TET, in addition to, as expected, broad-spectrum cephalosporins ( Figure 2).
Resistance rates for CIP, CHL, and TET were higher among ESBL-positive isolates, but the pattern was different for GEN and SXT. It is important to highlight that significant differences were detected between ESBL-and non-ESBL-producing isolates for CIP, CHL, SXT, and TET, in addition to, as expected, broad-spectrum cephalosporins ( Figure 2).

Molecular Characterization and Typing of ESBL-Producing E. coli
A total of 23 of the 111 E. coli isolates showed an ESBL phenotype, hosting diverse AMR phenotypes and genotypes, as well as genetic lineages. Most of these isolates were obtained from manure samples, but interestingly, one of the ESBL producers (X2583) was obtained from an indoor air sample. All these isolates carried the gene encoding the betalactamase SHV-12, with one of them (X2685) also carrying the gene encoding TEM-1.
Most of the ESBL producers showed CHL and TET resistance and carried the cmlA and tet(A) genes (with few exceptions). Eleven out of fourteen CIP-resistant isolates harbored the aac(6′)-Ib-cr gene. Two additional CIP-resistant isolates presented point mutations in the QRDR region in the gyrA and parC genes: two amino acid changes were detected in GyrA (S83L, D87N) and one in ParC (S80I). In addition, the sul1 and sul3 genes were detected in two and four isolates, respectively.
The 23 SHV-12-producing E. coli isolates were typed by MLST, and four different ST were detected: (a) twelve isolates were typed as ST770 and phylogroup-E (52.2%); the isolate from indoor air was included in this group; (b) three isolates were typed as ST68 and phylogroup-E (13%); (c) four isolates were typed as ST117 and phylogroup-B2 (17.4%); and (d) four isolates were typed as ST10992 (17.4%), two of them ascribed to phylogroup-A and the other two to phylogroup-B1. The CH typing identified seven clonotypes, with CH116-552 being the most prevalent, present in 12 ESBL producers associated with ST770. Other prevalent clonotypes were CH45-97 (two ST117 isolates), CH11-41 (two ST10992

Molecular Characterization and Typing of ESBL-Producing E. coli
A total of 23 of the 111 E. coli isolates showed an ESBL phenotype, hosting diverse AMR phenotypes and genotypes, as well as genetic lineages. Most of these isolates were obtained from manure samples, but interestingly, one of the ESBL producers (X2583) was obtained from an indoor air sample. All these isolates carried the gene encoding the beta-lactamase SHV-12, with one of them (X2685) also carrying the gene encoding TEM-1.
Most of the ESBL producers showed CHL and TET resistance and carried the cmlA and tet(A) genes (with few exceptions). Eleven out of fourteen CIP-resistant isolates harbored the aac(6 )-Ib-cr gene. Two additional CIP-resistant isolates presented point mutations in the QRDR region in the gyrA and parC genes: two amino acid changes were detected in GyrA (S83L, D87N) and one in ParC (S80I). In addition, the sul1 and sul3 genes were detected in two and four isolates, respectively.

Characterization of Antimicrobial Resistance Genes among Non-ESBL-Producing Isolates
AMR genes were analyzed in the collection of 88 non-ESBL-producing isolates. Fiftyeight of the sixty-one AMP-resistant isolates carried the bla TEM gene (95.1%). Several acquired resistance genes, such as aac(6)-Ib-cr and qnrS, were detected in six and two isolates out of twenty-nine CIP-resistant isolates, respectively. Most of the 28 TET-resistant isolates carried the tet(A) and/or tet(B) genes, either alone or combined (78.6%). Moreover, 51 of the 57 SXT-resistant isolates (89.5%) carried sul genes (sul1, sul2, or sul3), either alone or associated. The aac(3)-II and cmlA genes were detected among GEN-and CHL-resistant isolates, respectively ( Table 2). All isolates were negative for mcr-1, a gene associated with colistin resistance. Table 2. Antimicrobial resistance genes detected among the 88 non-ESBL-producing E. coli isolates in relation to their phenotype of resistance.

Characterization of Integrons among STX-Resistant E. coli Isolates
The presence of intl1 and intl2 genes was analyzed in the 62 SXT-resistant E. coli isolates detected in this study. The intI1 gene was identified in 26 isolates, intI2 in 7 isolates, and both genes in 5 additional isolates ( Table 2). The variable region of class 1 and/or class 2 integrin's could be identified in 14 of these isolates, and their gene cassettes arrays are shown in Table 3. In this regard, three types of gene-cassettes arrays (GC) have been detected in the class 1 integrons among the 11 isolates which could be characterized: (1)

Characterization of Integrons among STX-Resistant E. coli Isolates
The presence of intl1 and intl2 genes was analyzed in the 62 SXT-resistant E. coli isolates detected in this study. The intI1 gene was identified in 26 isolates, intI2 in 7 isolates, and both genes in 5 additional isolates ( Table 2). The variable region of class 1 and/or class 2 integrin's could be identified in 14 of these isolates, and their gene cassettes arrays are shown in Table 3. In this regard, three types of gene-cassettes arrays (GC) have been detected in the class 1 integrons among the 11 isolates which could be characterized: (1) dfrA1-aadA1 (n = 8); (2) dfrA12-aadA2 (n = 2) linked to sul1; and (3) dfrA12-orfX-aadA2-cmlA linked to sul3 (n = 1). The variable region of class 2 integrons was identified in five isolates, and the GC found was sat-aadA1 (Table 3).

Discussion
The AMR of E. coli isolates from the broiler farm environment was analyzed in this study, with a special focus on the characterization of ESBL-producing isolates. Of relevance is the high incidence (20.7%) of ESBL producers detected among the collection of E. coli isolates. It is important to note the detection of ESBL-producing E. coli isolates of the same genetic lineage and ESBL type in manure (n = 11) and indoor air samples (n = 1) from the broiler farm, showing dissemination of these resistant bacteria in different niches of the farm.
Among the most relevant results observed concerning the AMR patterns of the antibiotics studied, the highest resistance rates were detected for CHL, TET, and CIP (in ESBL-producing isolates, in addition to β-lactams), and for AMP, CIP, TET, and SXT (in non-ESBL-producing isolates) coinciding closely with the results reported by EFSA [1]. This could be associated with the high use of these antimicrobials in the poultry sector. The statistical analysis revealed a close correlation between antibiotics from different families, which might be explained by the location of the implicated resistance genes in the same genetic structures that could be co-selected.
ESBL genes could be transferred via plasmids between bacteria of diverse origins. The F, FIB, K, and HI1 plasmids were frequently detected among the SHV-12-producing isolates in poultry in other studies [38,39]. Moreover, IncI1 plasmids are also found associated with bla SHV-12 in the human and animal environment [17,40] and even in food isolates [10,30], where IncK and IncX3 replicons were also detected, in lower proportions. The IncF, IncF1B, IncK, IncHI1, and IncI1 replicon types have also been detected among our ESBL-producing isolates. Additional research is necessary to analyze in greater depth the genetic information associated with the ESBL and/or MDR phenotypes.
The ST770 lineage has been previously reported in CTX-M-1-or CMY-2-producing E. coli from poultry or chicken meat in different European countries [39,41] and associated with ESBL-negative avian pathogenic E. coli (APEC) [42]. Moreover, ST770 has also been detected in CTX-M-14-and CTX-M-2-producing E. coli isolates implicated in urinary tract infections [43]. Interestingly, this lineage has also been reported in ESBL producers from litter and poultry farm air (SHV and TEM-types), which reinforces the importance of cleaning and disinfection protocol to remove the microbial load before introducing the litter into the farms [44], according to the "all-in-all-out" strategy to avoid the dissemination of MDR bacteria. Furthermore, this lineage has also been detected in isolates carrying very relevant resistance genes, such as mcr-1 (colistin resistance), in that case in pets [45]. Data collected from studies in several countries show an increased association of the ST770 lineage with ESBL-and pAmpC-producing E. coli isolates.
Three of the lineages detected in this study (ST68, ST770, and ST117) have been detected in SHV-12, CTX-M-1, CTX-M-14, and/or CTX-M-15-positive E. coli isolates from food-producing animals [17]. In other origins, such as wildlife and companion animals, these three genetic lineages were found associated with the production of CMY-2 [46,47] and ST117 with the expression of CTX-M-1. [48]. In addition, ST68 has been reported in carbapenemase-positive isolates (OXA-48) [49]. Nevertheless, to our knowledge, the lineage ST68 has not been detected in Europe in SHV-12-producer isolates of poultry origin.
The ST117 is the most widespread genetic lineage in the poultry sector, being also the most frequent in our study. This lineage has been detected in SHV-12 producers from poultry slaughterhouse wastewater in Germany [34], and its high prevalence continues in the broiler production pyramid [50]. This lineage has also been detected in avian farming in ESBL-positive isolates carrying genes of the CTX-M family (CTX-M-1, CTX-M-14, CTX-M-15) or in CMY-2 producers [51]. In this sense, ST117 is considered a reservoir for ESBL and pAmpC encoding genes in poultry and is also closely associated with APEC isolates [52,53].
The unusual ST10992 lineage was firstly reported in this study, to our knowledge, in the poultry farm environment, among SHV-12-producing isolates.
Phylogenetic grouping provides epidemiological and ecological information related to the virulence capacity of E. coli isolates. Several studies revealed that phylogroups B2, D, E, and F are strongly associated with extraintestinal infections in humans (EXPEC strains), whereas A and B1 are generally associated with commensal bacteria in humans and animals, including those ESBL producers [43,51]. In this sense, phylogroups A, B1, and F have been frequently found among ESBL-producing E. coli isolates from poultry farms in different European studies, with phylogroups B2 and D scarcely detected [50,54]. Studies from the poultry environment in Africa reveal the predominance of phylogroups A and D [55]. In our study, the presence of A and B1 was evident among ESBL-producing isolates, although phylogroups E and B2 were also frequently detected.
Integrons are mobilizable genetic elements whose study in recent years has gained relevance due to their role in the acquisition and dissemination of AMR. Four types of gene cassette arrays were detected among the eleven characterized isolates in our study, associated to class 1 (dfrA1-aadA1, dfA12-aadA2, and dfrA12-orfX-aadA2-cmlA) and class 2 integrons (sat2-aadA1). They carried resistance genes to old antibiotics, such as trimethoprim (dfrA1, dfrA12), streptomycin (aadA1, aadA12), chloramphenicol (cmlA), or streptothricin (sat2), whose use in animals or humans could contribute to the selection of integron-positive isolates. The presence of integrons in poultry E. coli isolates containing the classical dfrA1-aadA1 structure has been previously reported [52,56].
Potential sources of AMR in poultry production consist of air, dust, soil, feed, and rodents or other animals. From these sources, AMR bacteria and genes can be transmitted to humans through the food chain or indirectly by the environment, posing a threat to public health [1]. Several studies have shown the role of air as a disseminating vehicle and the ability to persist both inside and outside a farm, and poultry manure is considered to be the main vector for the spread of AMR bacteria [55].

Conclusions
ESBL-producing isolates frequently contaminate the poultry farm environment, and these resistant bacteria could be present in the farm air. In our case, SHV-12 was the only ESBL type detected, which has been observed in different E. coli clones, some of them very frequent in the poultry sector in both ESBL-producing and APEC isolates. Furthermore, the detection of the same genetic lineage in both poultry manure and airborne isolates provides further evidence of the transmissibility of these resistant bacteria and their ability to survive and adapt to any niche. Close surveillance on the dissemination of ESBL-producing isolates in the farm environment should be performed to avoid the dissemination of these isolates through the food chain or indirectly through the air or dust, among others.
This study shows that both manure and airborne dust particles are important sources of resistant bacteria and AMR genes due to the high survival of these resistant microorganisms, making it necessary to apply appropriate cleaning protocols to reduce and avoid as much as possible the survival of bacteria. Based on these results, it is suggested that broiler farms could be considered as important vehicles for the transmission of multidrug-resistant bacteria favoring their spread into the natural environment and in turn posing a risk to public health.
AMR is probably the clearest example of a global and growing problem that requires solutions to deal with the "One Health" approach. Coordinated and precise actions are needed to reduce its impact now and in the future, as well as measures to ensure the economic viability of the poultry sector together with public health assurance.