Abstract
The objective of this study was to identify the main extended-spectrum beta-lactamase (ESBL)-producing bacteria and to detect the frequency of the major genes responsible to trigger this resistance in hospitalized animals. We collected 106 rectal swabs from cats (n = 25) and dogs (n = 81) to detect ESBL-producing isolates. ESBL-positive samples were submitted to the antimicrobial susceptibility test, and polymerase chain reaction was performed to detect TEM, SHV, and CTX-M genes from different groups. We observed that 44.34% of these samples (11 cats and 36 dogs) were positive for ESBL-producing bacteria. Thirteen animals (27.66%—seven cats and six dogs) were hospitalized for elective castration (healthy animals). Only a single animal was positive for ESBL-producing bacteria at hospital admission (the animal also showed an ESBL-positive isolate after leaving the hospital), whereas 11 were positive only at the hospital discharge. Of the 73 ESBL-producing isolates, 13 were isolated from cats (8 sick and 7 healthy) and 60 from dogs (53 sick and 7 healthy). Escherichia coli was the major ESBL-producing bacterium isolated (53.42%), followed by Pseudomonas aeruginosa (15.07%), Salmonella sp., and Proteus mirabilis (5.48% each one). Antimicrobial resistance profile of ESBL-producing isolates showed that 67 isolates (91.78%) were resistant to 3 or more antibiotic classes, while 13 of them (17.81%—2 healthy cats and 11 sick dogs) were resistant to all tested antimicrobial classes. The blaTEM gene exhibited the highest frequency in ESBL-producing isolates, followed by the blaCTX-M group 8/25, blaCTX-M group 1 and blaCTX-M group 9 genes. These results are useful to assess the predominance of ESBL-producing isolates recovered from dogs and in cats in Brazil. Consequently, we draw attention to these animals, as they can act as reservoirs for these microorganisms, which are the major pathogens of nosocomial infections worldwide.
Introduction
One of the major resistance mechanisms developed by gram-negative bacteria is the hydrolysis of β-lactam ring through the action of enzymes. This process confers resistance to beta-lactam antimicrobial classes, including penicillin, cephalosporins (especially those from the third and fourth generations), and monobactams. 1 These enzymes do not hydrolyze cephamycins (cefoxitin) and carbapenems, but they can be inactivated by β-lactamase inhibitors (clavulanic acid, sulbactam, and tazobactam). 2 The aforementioned enzymes are known as extended-spectrum beta-lactamases (ESBLs), which are divided into three main groups: TEM, SHV, and, CTX-M. 3 OXA-1 genes (found in the plasmid and integron locations of gram-negative microorganisms) are frequently associated with the worldwide spread of CTX-M-15 ESBL, conferring reduced susceptibility to cefepime (CPM), and a combination of beta-lactams and beta-lactamase inhibitors.4,5
ESBL encoding genes are located in conjugative plasmids or integrons. Besides that, transference of these genes to other enterobacteria species can occur, which facilitates even more its spread. 6 In the last decades, resistance to cephalosporins has become a worldwide threat to health public,7,8 which is often responsible for nosocomial infections, and even community-acquired infections. 9
Companion animals, especially cats and dogs, are important sources of antimicrobial resistance gene transmission due to direct contact with humans, posing public health to great risk.10–12 Besides that, in the past years, studies have reported worldwide the presence of ESBL-producing Enterobacteriaceae isolates recovered from products of animal origin and companion animals. 13
Most research conducted in small animals aimed at the identification of ESBL-producing bacteria, especially Escherichia coli. However, other types of Enterobacteriaceae, such as Citrobacter spp., 14 Enterobacter spp., Klebsiella sp., 15 Pantoea agglomerans, 16 Serratia marcescens, Morganella morganii, and Providencia spp., have also been reported as ESBL-producing bacteria. 17 In addition, other investigations demonstrated that non-enterobacteria such as Pseudomonas aeruginosa, 16 Acinetobacter baumannii, 18 and Stenotrophomonas maltophilia are also able to produce ESBL in humans.
Epidemiological studies on molecular characterization of ESBL-producing bacteria in small animals admitted to veterinary hospitals are very scarce. For this reason, this study aimed to identify the main ESBL-producing gram-negative bacteria by detecting resistance genes in hospitalized animals.
Materials and Methods
Samples
The Brazilian veterinary medical teaching hospital of Universidade do Estado de Santa Catarina (CAV-UDESC) treats ∼5,000 animals per year. In this observational study, rectal samples were collected from consecutive hospitalized small animals admitted between September 2016 and December 2016. Sample collection occurred at admission and discharge of the patient. If hospitalization was longer than 3 days, additional samples were collected following our pre-established guidelines. These collections were then performed between admission and discharge days. Before this procedure, swabs were humidified with sterile water. Rectal swab specimens were stored in tubes containing 15 mL of brain/heart infusion (BHI) broth, and supplemented with 10 μg ceftriaxone (CRO) disk. Later, rectal samples were sent to Centro de Diagnóstico Microbiológico Animal (CEDIMA) of CAV-UDESC, where they were incubated at 37°C for 24 hr. This study was approved by the ethics committee of UDESC (2042290617), Brazil.
Phenotypic screening for ESBL-producing isolates
Samples were inoculated onto MacConkey agar at 37°C for 24 hr. Then, we screened positive samples for bacterial growth. To detect ESBL-producing isolates, we performed disk approximation testing according to the recommendations by Souza-Junior et al. 19 For the test, one amoxicillin/clavulanic acid (AMC) disk was centrally placed on an inoculated Mueller–Hinton agar plate. Then, additional disks of third- and fourth-generation cephalosporins and aztreonam (AZT) were placed at a distance of 20 mm from disc center to disc center of the plate. The third-generation cephalosporins used were ceftazidime (CAZ), cefotaxime (CTX), and ceftriaxone. Cefepime was the only fourth-generation cephalosporin used. Agar plate was incubated at 37°C for 24 hr. After this incubation period, we considered an ESBL-producing positive sample if there was a formation of a phantom zone, distortion of the inhibition zone edge toward the disk, or when the sample exhibited resistance to all antimicrobial agents.
Biochemical identification of ESBL-producing isolates and antimicrobial susceptibility testing profile
ESBL-producing bacteria were identified with the gram-negative Bactray system kit (Laborclin®, Vargem Grande, Brazil). To verify the susceptibility of isolates, we performed disk diffusion tests as described by the Clinical and Laboratory Standards Institute (CLSI). 20 Mueller–Hinton agar plates were incubated at 35°C ± 2°C for a period of 16–18 hr. The following disk diffusion disks were used: ampicillin (10 μg), amoxicillin/clavulanic acid (20 μg/10 μg), ticarcillin/clavulanic acid (75 μg/10 μg), cephalothin (30 μg), ceftriaxone (30 μg), ceftazidime (30 μg), cefotaxime (30 μg), ceftiofur (30 μg), cefepime (30 μg), cefoxitin (30 μg), aztreonam (30 μg), imipenem (10 μg), meropenem (10 μg), gentamicin (10 μg), streptomycin (10 μg), amikacin (30 μg), tobramycin (10 μg), chloramphenicol (30 μg), nitrofurantoin (10 μg), enrofloxacin (5 μg), norfloxacin (10 μg), ciprofloxacin (5 μg), marbofloxacin (5 μg), levofloxacin (5 μg), tetracycline (30 μg), doxycycline (30 μg), and trimethoprim/sulfamethoxazole (1.25/23.75 μg). The zones of growth inhibition around each antibiotic disk were measured and interpreted using the criteria established by the CSLI for human and animal specimens.21,22
Isolates that exhibited intermediate resistance were classified under a resistance pattern, since the therapeutic effect is uncertain, and antibiotic exposure might result in therapeutic failure. 18 E. coli ATCC 25922 reference strain was used as quality control to determine susceptibility to all antimicrobial agents.
Bacterial DNA extraction
Genomic DNA extraction of the isolates and reference strains was performed in accordance with a previous protocol for Parussolo et al. 23 DNA concentration measurements were performed on NanoDrop (Thermo Fisher, Waltham, MA). To perform polymerase chain reaction (PCR) tests, DNA concentration was adjusted to 15–100 ng.
Genotypic characterization of ESBL-producing isolates
Isolates that exhibited a positive ESBL phenotypic profile were further investigated to verify the presence of ESBL genes. For that, we performed two multiplex PCRs: one to detect the genes BlaTEM (and variants), BlaSHV (and variants), and BlaOXA-1; and the other to disclose BlaCTX-M genes of groups 1, 2, and 9. In addition, a conventional PCR was conducted to identify BlaCTX-M genes of groups 8 and 25, using specific primers (Table 1). 24 All amplification reactions were conducted in a 50 μL final volume containing PCR buffer (Tris-HCl—20 mM, KCl—50 mM), MgCl2 (1.5 mM), dNTP (200 mM of each), Taq DNA polymerase (1 U), primers (4 pmol of each), and bacterial DNA (2 μL). All PCR programs were carried out as follows: initial denaturation at 94°C for 10 min; 40 cycles of denaturation step at 94°C for 40 sec, annealing of the primer at 56°C for 40 sec, and 1 min of extension at 72°C. The last cycle extension was at 72°C for 7 min.
List of Primers Used in Polymerase Chain Reaction for Detection of Extended-Spectrum Beta-Lactamase-Producing Bacteria
All primers were designed and described by Dallenne et al. 24
f, forward; r, reverse.
Amplified fragments were subjected to electrophoresis in 2% agarose gel (100 V, 300 mA) for 1 hr, stained with GelRed™, and visualized in a transilluminator. Klebsiella pneumoniae CCBH5991 and K. pneumoniae CCBH15948 reference strains were used to assure quality control for the detection of BlaTEM+, BlaSHV+, BlaCTX-M+, and BlaOXA+ genes (Fundação Oswaldo Cruz–Fiocruz). E. coli ATCC 25922 reference strain was used as a negative control.
Results
Samples
A total of 106 samples of animals, 26.42% from cats and 73.58% from dogs, were collected. Of these animals, 28 (26.42%—15 cats and 13 dogs) were healthy and 78 (73.58%—13 cats and 65 dogs) were sick. Because eight of these animals required hospitalization longer than 3 days, it was necessary to perform more collections between admission and discharge days. Thus, the total number of samples analyzed was 224.
Phenotypic screening for ESBL-producing isolates
We detected the presence of ESBL-producing bacteria in 47 animals (44.34%), being 11 cats (23.40%) and 36 dogs (76.60%). Of the 47 animals that were positive for ESBL-producing isolates, 13 (27.66%), 7 cats and 6 dogs were hospitalized for elective castration (salpingo-oophorectomy in females and orchiectomy in males). Only a single animal was positive for ESBL-producing bacteria at hospital admission (the animal also showed an ESBL-positive isolate after leaving the hospital), whereas 11 were positive only at the hospital discharge. Tables 2 and 3 show the identification of types of microorganisms and the beta-lactam resistance genes from the isolates recovered from these animals.
List of Healthy Cats Admitted to CAV/UDESC That Revealed Positive for Extended-Spectrum Beta-Lactamase in Fecal Samples, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
AG, aminoglycosides; AMP, amphenicols; BLT, beta-lactam antibiotics; FQ, fluoroquinolones; NIT, nitrofurans; OE, orchiectomy; SUL, sulfonamide; TET, tetracycline; USO, salpingo-oophorectomy.
List of Healthy Dogs Admitted to CAV/UDESC That Revealed Positive for Extended-Spectrum Beta-Lactamase in Fecal Samples, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
Of the other 34 animals that were positive for ESBL producing, only 4 were cats (11.76%) and 30 were dogs (88.24%). Characterization of samples from sick hospitalized animals, including ESBL gene detection, is illustrated in Tables 4–7.
List of Sick Cats Admitted to CAV/UDESC That Revealed Positive for Extended-Spectrum Beta-Lactamase in Fecal Samples, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
FLUTD, feline lower urinary tract disease.
List of Dogs Admitted to HCV/CAV/UDESC That Revealed Positive Results for Extended-Spectrum Beta-Lactamase in Fecal Samples upon Admission to the Hospital, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
DND, diagnosis not defined.
List of Dogs Admitted to HCV/CAV/UDESC That Revealed Positive Results for Extended-Spectrum Beta-Lactamase in Fecal Samples at Two Points in the Collection, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
List of Dogs Admitted to HCV/CAV/UDESC That Revealed Positive Results for Extended-Spectrum Beta-Lactamase in Fecal Samples at Discharge from the Hospital, Associated with the Antimicrobial Susceptibility Profile and Detection of Resistance Genes
A total of 27 antimicrobial agents were used.
CFC, cerebrospinal fluid collection; LHL, left hind limb.
Of the 224 samples collected, those positive were as follows: 29 during admission to HCV, 9 between admission and discharge period, and 35 at hospital discharge, which resulted in 73 positive gram-negative ESBL-producing samples. Of the 73 ESBL-producing isolates, 13 were isolated from cats (8 sick and 7 healthy) and 60 from dogs (53 sick and 7 healthy).
Biochemical identification of ESBL-producing isolates and antimicrobial susceptibility testing profile
In the 73 ESBL-producing isolates, we observed the presence of E. coli (53.42%), P. aeruginosa (15.07%), Proteus mirabilis (5.48%), Salmonella spp. (5.48%), Klebsiella oxytoca (4.11%), Citrobacter freundii (2.74%), A. baumannii (2.74%), S. maltophilia (2.74%), Serratia liquefaciens (2.74%), Citrobacter braakii (1.37%), Citrobacter youngae (1.37%), S. marcescens (1.37%), and Hafnia alvei (1.37%).
Separately, bacterial identification proceeded as follows: (1) healthy cats: E. coli (4/7—57.14%), H. alvei (1/7—14.29%), S. maltophilia (1/7—14.29%), and C. braakii (1/7—14.29%); (2) healthy dogs: E. coli (3/7—42.86%), P. aeruginosa (2/7—28.57%), and Citrobacter freundii (2/7—28.57%); (3) sick cats: E. coli (5/8—65.50%) and Salmonella sp. (3/8—34.5%); and (4) sick dogs: E. coli (27/51—52.94%), P. aeruginosa (9/51—17.65%), P. mirabilis (4/51—7.84%), K. oxytoca (3/51—5.88%), A. baumannii (2/51—3.92%), S. liquefaciens (2/51—3.92%), C. youngae (1/51—1.96%), Salmonella sp. (1/51—1.96%), Serratia marcescens (1/51—1.96%), and S. maltophilia (1/51—1.96%).
We conducted antimicrobial susceptibility testing for seven different antibiotic classes on ESBL-positive samples to assess the multidrug resistance profile (multidrug resistance defined as phenotypic resistance to at least three of the tested antibiotic classes). Out of the 73 ESBL-producing bacteria samples, 67 (91.78%) were resistant to at least 3 classes, whereas 13 (17.81%—2 healthy cats and 11 sick dogs) were resistant to all of them. We only observed resistance values below 10% for meropenem (2.74%), imipenem (5.48%), and amikacin (9.59%), all of them in sick dogs. On the contrary, the other antimicrobial agents revealed a resistance value higher than 50%, except for tobramycin (39.73%) and nitrofurantoin (42.247%). Figure 1 shows the antimicrobial susceptibility profiles of ESBL-producing isolates for the antibiotics tested, excluding penicillins, aminopenicillins, monobactams, and cephalosporins.
Antimicrobial susceptibility profile of ESBL-producing bacterial isolates recovered from healthy and diseased animals admitted to HCV at CAV/UDESC. AMK, amikacin; CIP, ciprofloxacin; CLO, chloramphenicol; DOX, doxycycline; ENR, enrofloxacin; ESBL, extended-spectrum beta-lactamase; FOX, cefoxitin; GEN, gentamicin; IPM, imipenem; LVX, levofloxacin; MEM, meropenem; MFX, marbofloxacin; NIT, nitrofurantoin; NOR, norfloxacin; STR, streptomycin; SXT, trimethoprim/sulfamethoxazole; TET, tetracycline; TIM, ticarcillin/clavulanic acid; TOB, tobramycin; UDESC, Universidade do Estado de Santa Catarina.
Genotypic characterization of ESBL-producing isolates
Molecular characterization of ESBL-encoding genes revealed that all isolates harbored the blaTEM gene, which was identified alone in seven isolates (9.59%) or combined with the other genes. The combinations found were as follows: blaTEM+blaCTX-M-25 in 15 isolates (20.55%); blaTEM+blaCTX-M-25+blaOXA-1 and blaTEM+blaCTX-M-1 in 10 isolates (13.70%); blaTEM+blaCTX-M-9+blaOXA-1 and blaTEM+blaOXA-1 in 7 isolates (9.59%); blaTEM+blaCTX-M-9 in 6 isolates (8.22%); blaTEM+blaSHV and blaTEM+blaOXA-1 in 4 isolates (5.48%); blaTEM+blaSHV+blaCTX-M-1, blaTEM+blaSHV+blaOXA-1, and blaTEM+blaSHV+blaCTX-M-1+blaOXA-1 in 1 isolate (1.37%). Gene distribution can be seen individually in all the tables.
Discussion
Most studies underscore the detection of the major genes encoding the production of ESBLs in E. coli, but our study demonstrates that there are other very important microorganisms involved in the dissemination of the aforementioned genes. Surprisingly, as 46.58% (34/73) of the isolates were identified as non-E. coli, 44.12% (15/34) of them did not even belong to the Enterobacteriaceae family. Although this fact has been previously reported by other authors,6,16,17 there is still a lack of investigations pinpointing the association of bacterial identification in hospitalized animals (healthy and sick) with the occurrence of genes encoding ESBL production and antimicrobial susceptibility profile.
ESBL-producing bacteria were isolated from 47 (44.34%) animals, a value that is well above those found in other countries such as the United States (3.80%), Algeria (11.7%), Mexico, and Switzerland (17%), and even southeast of Brazil (28.50%).10,25–28 This is an alarming situation, since companion animals are constantly in contact with their owners and the human environment. Besides that, these results reinforce that the current conditions of this teaching veterinary hospital, where the flow of people is high, might promote the dissemination of this type of microorganism.
The results from the current study indicated that 91.78% of ESBL-producing isolates were considered multidrug resistant, a value similar to that described by Zogg et al. 28 (88.60%). In addition, Liu et al. 25 showed that 75.00% of ESBL-producing samples were resistant to at least 10 antimicrobial drugs. We interestingly observed that 89.00% of the isolates were resistant to the same number of antibiotics described by these authors and 39.73% were actually resistant to at least 20 antimicrobial drugs. The other concerning fact is that 17.81% of our resistant ESBL-producing samples were resistant to all antimicrobial classes tested (n = 7), which dramatically affects the choice of antibiotic therapy for these animals and impairs the safety of care by dissemination of opportunistic microorganisms to the environment.
On the subject of an antimicrobial drug of beta-lactam class (except for carbapenems), the drug resistance profiles varied from 71.23% to 98.63% (ceftriaxone and cefotaxime), which is in agreement with that presented by Liu et al. 25 ; however, they were higher than the ones reported by Carvalho et al. 10 (values between 9.50% and 33.30%). If we are going to evaluate the groups of animals individually, healthy cats showed 100% resistance to aztreonam, cefotaxime, ceftriaxone, cefepime, and cephalothin, while healthy dogs were 100% resistant to ceftriaxone, cefotaxime, cephalothin, ceftiofur, and ampicillin. In sick animals, all dogs were resistant to ceftriaxone, whereas cats were all resistant to aztreonam and ceftazidime. These high levels of resistance to the aforementioned antimicrobial classes increase the use of carbapenems (meropenem and imipenem), which usually are only approved for clinical treatment of severe infections, 29 consequently leading to the emergence and dissemination of carbapenemase-producing Enterobacteriaceae. With respect to carbapenems, only 2.74% of ESBLs were resistant to meropenem and 5.48% to imipenem, all of them were sick dogs, suggesting a decrease of permeability or even the presence of enzymes responsible for hydrolyzing antimicrobials of this group.
ESBL antimicrobial resistance in human and animal isolates is commonly associated with other antibiotic classes.28,30 Antimicrobial agents such as amoxicillin/clavulanic acid, chloramphenicol, trimethoprim/sulfamethoxazole, and fluoroquinolones are frequently used to treat infections caused by E. coli in cats and dogs, which led to an increase of multidrug-resistant strains.14,31 Similarly, the current study underscores this finding, as we observed more than 50% resistance to these antibiotics. In the groups of animals individually, fluoroquinolones showed more than 50% resistance only in sick animals (dogs and cats), with the exception of enrofloxacin, which was in all groups. This can be explained by the increased use of enrofloxacin in the study's veterinary hospital.
Baede et al. 32 and Schmidt et al. 33 reported the presence of ESBL genes in isolates recovered from both healthy and sick animals. Interestingly, these results are in agreement with ours, as we demonstrated that 27.66% of hospitalized animals for elective castration (without any signs of sickness or disease) had also positive samples for ESBL genes.
Except for two dogs, ESBLs retrieved from these patients were identified at the hospital discharge day, which suggests veterinary nosocomial colonization. This fact points out that after bacterial colonization, this animal will return to its natural environment and consequently will most likely disseminate these potential pathogenic and multidrug-resistant bacteria to the community.
Molecular characterization of ESBL-producing isolates in veterinary medicine assists in the surveillance, monitoring, and tracking of the most important genes encoded by these microorganisms and responsible for resistance to the most used antibiotic class in the veterinary praxis routine. According to Carvalho et al., 10 these pieces of information are also essential to identify similarity among bacterial human and animal isolates, since the prevalence of these genes in animals is yet to be well established in Brazil.
As stated by D'Andrea et al., 34 since the first sporadic detection of CTX-M in animals, The ESBLs of the CTX-M type are broadcast worldwide and are one of the most prevalent ESBLs, replacing the TEM and SHV types. In contrast to other studies that reported a predominance of CTX-M type of ESBL,10,26 our study showed that 100% of the samples harbored the TEM enzyme, whereas 78.02% were positive for different groups of CTX-M. Liu et al. 25 suggest that the historical aspects of animal treatment, geographical region, and phenotypic profile of ESBL-producing isolates might affect the occurrence of blaCTX-M gene. In Brazil, there are no studies showing the prevalence of each of these genes in pet animals
In relation to CTX-M enzyme variants, group 18/25 was the most isolated (43.86%), followed by group 1 (33.33%) and group 9 (22.81%). Other studies worldwide showed that the blaCTX-M-15 gene (from group 1) was the most predominant.13,25,26,28,35 Bevan et al. 13 confirmed that the CTX-M-2 gene is still predominant in ESBL-producing isolates recovered from human specimens in South America. Although we examined animal samples, none of them exhibited any variant of this gene.
Recently, one of the main studies in Brazil conducted in São Paulo on the identification of ESBL in companion animals showed that 31% of dogs and 6.9% of cats, all healthy, were colonized by ESBL/pAmpC-producing Enterobacteria. That same study also demonstrated colonization of 11.3% of dogs and 8.3% of cats with urinary problems. 36 Despite the differences in relation to the animals' environment and the conditions of sample collection, both our work and that carried out in São Paulo show concerns about pet animals hosting bacterial resistance genes and serving as vectors for the dissemination of these genes for humans.
Our findings showed that sick and healthy animals are potential carriers of ESBL-producing bacteria. Although E. coli is the most isolated ESBL-producing bacterium in animals, other species were also identified as carriers of resistance genes, which shows how inter- and intradissemination is eased among these microorganisms. Currently, TEM is not the most predominant gene in ESBL isolates worldwide. However, our results revealed that all samples harbored this gene, which reinforces the idea that antimicrobial resistance varies according to the period and geographic region of the study.
This study is the first molecular investigation of the predominance of ESBL-producing isolates that colonized healthy and diseased dogs and cats in South Brazil. Consequently, we draw attention to these animals, as they can act as reservoirs for these microorganisms, which are the major pathogens of nosocomial infections worldwide.
Footnotes
Acknowledgments
We thank Fundação Oswaldo Cruz (Fiocruz) and the Instituto Adolfo Lutz for providing the reference strains.
Disclosure Statement
No competing financial interests exist.
Funding Information
This work was supported by the Fundação de Amparo à Pesquisa e Inovação do Estado de Santa Catarina—FAPESC.