Phenotypic Investigation of Florfenicol Resistance and Molecular Detection of floR Gene in Canine and Feline MDR Enterobacterales

Simple Summary Florfenicol is a synthetic analog of chloramphenicol, which is used mostly in livestock animals. Its potential as an alternative agent in companion animals is strong because of its beneficial properties, such as its extensive wide spectrum and relatively increased, compared to chloramphenicol, safety during administration. However, the emergence of resistant strains can overcome this potential. The distribution of florfenicol-resistant bacteria was investigated among multidrug-resistant (MDR) Enterobacterales isolated from diagnostic samples of companion animals in Greece, which were collected throughout the country. Data regarding sample origin, type of infection, bacterial species, and resistance profiles were assembled and compared. The presence of specific florfenicol-associated antibiotic resistance genes (ARGs) was examined, and the results were interpreted in comparison with those of antibiotic susceptibility testing (AST). The results of this study indicate that the distribution of florfenicol-resistant MDR Enterobacterales in pets throughout the country is considerable (17.9%) and mainly attributed to the plasmid-mediated floR gene. Thus, there is an increased risk of co-acquisition of florfenicol-specific ARGs through horizontal transfer, along with several other resistance genes. Even though the potential of florfenicol to constitute an alternative antibiotic in companion animals seems high, continuous monitoring of antibiotic resistance profiles is needed. Abstract Florfenicol is a promising antibiotic for use in companion animals, especially as an alternative agent for infections caused by MDR bacteria. However, the emergence of resistant strains could hinder this potential. In this study, florfenicol resistance was investigated in a total of 246 MDR Enterobacterales obtained from canine and feline clinical samples in Greece over a two-year period (October 2020 to December 2022); a total of 44 (17,9%) florfenicol-resistant strains were recognized and further investigated. Most of these isolates originated from urine (41.9%) and soft tissue (37.2%) samples; E. coli (n = 14) and Enterobacter cloacae (n = 12) were the predominant species. The strains were examined for the presence of specific florfenicol-related resistance genes floR and cfr. In the majority of the isolates (31/44, 70.5%), the floR gene was detected, whereas none carried cfr. This finding creates concerns of co-acquisition of plasmid-mediated florfenicol-specific ARGs through horizontal transfer, along with several other resistance genes. The florfenicol resistance rates in MDR isolates seem relatively low but considerable for a second-line antibiotic; thus, in order to evaluate the potential of florfenicol to constitute an alternative antibiotic in companion animals, continuous monitoring of antibiotic resistance profiles is needed in order to investigate the distribution of florfenicol resistance under pressure of administration of commonly used agents.


Introduction
Florfenicol is a synthetic, fluorinated derivative of chloramphenicol, which was successfully developed in the 1980s, for usage in veterinary medicine [1,2].It is currently used worldwide in cattle, poultry, pigs and fish [3].
This agent possesses several beneficial properties.Initially, it has an extensively wide spectrum, including most Gram-negative and Gram-positive bacteria, chlamydiae, and rickettsiae [3].Therefore, it could constitute an alternative treating option in several cases, where available narrow-spectrum agents are not effective, mostly due to acquired antibiotic resistance.Moreover, it is not affected by several chloramphenicol-resistance mechanisms, such as the enzymatic modification by bacterial chloramphenicol acetyltransferases (CATs), as it contains a terminal fluorine instead of a hydroxyl group in its molecule [2,4].Additionally, it possesses an obvious advantage in safety compared to chloramphenicol, due to the substitution of the nitro group (-NO 2 ), which was incriminated for the drug's serious side effects, mainly dose-unrelated aplastic anemia [2,3].
Resistance to florfenicol can be conferred by various antibiotic resistance genes (ARGs), such as floR, cfr, fexA, fexB, and optrA, whereas multidrug efflux pumps have also been implicated [4].The FloR gene was first reported in an epidemic strain of Salmonella enterica (serovar Typhimurium) in 1999, correlated with a multiple resistance genetic locus of the emerging DT104 strain [5].It is the predominant determinant of resistance in Gram-negative bacteria [6].The rRNA methylase-encoding gene cfr was first described in a Mammaliicoccus sciuri isolate of bovine origin in 2000 [7].It is commonly distributed through plasmids among bacteria of different genera and species [4].
Florfenicol's usage is currently limited in pets.Various studies have examined its potential as a wide-spectrum alternative agent for companion animals [8,9].Nevertheless, there are some difficulties with its administration, with the exception of otic gel formulations [10] and certain limitations must be investigated to obtain sufficient data for safe usage [11].Therefore, further research is needed to evaluate its effectiveness in vitro at different sites of infection.However, since multidrug resistance has emerged, the armamentarium of veterinarians is limited, and evaluation of novel alternatives is necessary.
The objective of this study was to investigate the distribution and the characteristics of florfenicol-resistant Enterobacterales, with a focus on MDR isolates, since florfenicol is mostly used as a "second line" antibiotic in veterinary medicine.Moreover, we sought to identify the molecular basis of this resistance and indicate possible factors enhancing the emergence of relevant strains and concerns regarding public health.

Isolation and Selection of the Bacterial Strains
The bacterial strains included in this study were isolated from clinical canine and feline samples over a two-year period (October 2020 to December 2022).The samples were obtained during routine veterinary practice in veterinary hospitals throughout Greece.Isolation and biochemical identification were initially performed using conventional techniques and a commercial identification kit (Invitrogen GN-ID A, Gold Standard Diagnostics, Budapest, Hungary).Routine susceptibility testing was then performed.
All Enterobacterales isolates demonstrating a MDR profile, according to a previously described classification [12], were collected.In cases where more than one bacterial species were obtained from a specific sample in considerable populations, all Enterobacterales were collected, regardless of the assumption about the causative agent of the infection.The strains used in this study were subsequently selected from the aforementioned group by exhibiting a phenotype resistant to florfenicol using the disc diffusion method (Kirby Bauer).The current CLSI document provides only minimum inhibitory concentration (MIC) values for Enterobacterales of swine origin [13].As there were no specific zone diameter breakpoints for florfenicol, those provided for bovine respiratory pathogens were used (S ≥ 19 mm, I: 15-18 mm, R ≤ 14 mm).The results of the disc diffusion test were then confirmed by the MIC method (VITEK ® 2, bioMérieux, Craponne, France).All isolates were maintained in Brain Heart Infusion Broth supplemented with 20% glycerol at −80 • C.

Antimicrobial Susceptibility Testing
The disk diffusion method was initially performed to evaluate the susceptibility of the selected strains to the following 19 antimicrobial agents: florfenicol, chloramphenicol, ampicillin, amoxicillin + clavulanic acid, cefaclor, cefuroxime, cefoxitin, ceftazidime, piperacillin + tazobactam, amikacin, gentamicin, tobramycin, ciprofloxacin, enrofloxacin, sulfamethoxazole + trimethoprim, doxycycline, minocycline, fosfomycin, and nitrofurantoin.Both the zone diameter breakpoints and the contents of the disks, as specified by the relevant CLSI documents [13,14] (with the exception of florfenicol as previously noted), are listed in Table 1.Colonies from pure culture of each strain were added to saline in order to achieve a McFarland turbidity of 0.5 in the resulting suspension.Subsequently a sterile swab was used to inoculate a quantity of this suspension on the surface of a Mueller-Hinton agar plate.After the addition of susceptibility discs, the plates were incubated at 35 • C for 16-18 h.The results the AST were then confirmed with the MIC method, using VITEK ® 2 (bioMérieux, Craponne, France).The MIC for florfenicol was evaluated with the break-points provided by the relevant CLSI document for swine isolates (S ≤ 4, I:8, R ≥ 16) [13].All respective breakpoints are presented in Table 1.
Whole genomic DNA was extracted from all strains, using a commercial spin column kit (IndiSpin Pathogen Kit, INDICAL BIOSCIENCE GmbH, Leipzig, Germany).All procedures were performed according to the manufacturer's instructions.To perform PCR, previously described primers were evaluated [3].Primers used are listed in Table 2.For each reaction, a 25 µL mix was created per strain by adding 12.5 µL of Xpert Fast Mastermix (2X) with dye (GRiSP Research Solutions, Porto, Portugal), 2 µL (10 pmol) of each primer, 0.5 µL of bacterial DNA and 8 µL of PCR-grade water.The conditions were as follows: 95 • C for 1 min, followed by 40 cycles of 95 • C for 15 s (denaturation), 60 • C for 15 s (annealing), and 72 • C for 3 s (elongation), followed by a final extension at 72 • C for 3 min.DNA products for each gene were identified after electrophoresis in 0.5 Tris-borate-EDTA using a 1.5% agarose gel stained with ethidium bromide solution.

Isolates, Origin of the Samples and Co-Infections
A total of 246 MDR Enterobacterales isolates were obtained from clinical samples of pets during the study period, 151 (61.4%) of which were of canine origin and 95 (38.6%) of feline.A group of 44 of them (17.9%) were resistant to florfenicol.Thirty of these isolates (68.2%) originated from canine infections, whereas the remaining 14 (31.8%) were from cats.Therefore, the resistance rate was 19.9% and 14.7% in canine and feline strains, respectively.The site of infection and distribution of the respective samples are presented in Table 3. Regarding the geographical distribution, florfenicol resistant isolates were obtained from samples collected throughout the country (Table S1).In particular, they originated in Athens (n = 18), Thessaloniki (n = 16), Volos (n = 4), Serres (n = 2), Giannitsa (n = 1), Heraklion (n = 1), Ierapetra (n = 1) and Kozani (n = 1).Most isolates were obtained from urine and soft tissue samples.Feline samples mainly originated from cases of urinary tract infections (UTIs, 69.2%), while canine samples from soft tissue infections, followed by UTIs (46.7% and 30% respectively).Regarding soft tissue samples, the majority of them were wounds and skin lesions in dogs, while two samples from a surgical wound and a skin lesion were obtained from cats.

Antimicrobial Susceptibility Testing
As only MDR isolates were included in this study, high rates of resistance were unsurprisingly detected against most of the agents tested.Data regarding the resistance rates by the disc diffusion test and MIC methods are presented in Tables 6 and 7, respectively.Detailed data for each strain are provided in Tables S1 and S2.  1 The rates are presented only for the number of strains tested for each agent by VITEK 2 (Table S2).
The results of the disc diffusion and the MIC method were mostly coincident, with a few exceptions.Regarding florfenicol, all isolates were resistant in both tests.

Detection of Florfenicol Resistance Genes
The floR gene (primers: F: 5′-ACGTTTATGCCAACCGTCCT-3′, R: 5′-CATTACAA-GCGCGACAGTGG-3′) was detected in most of the isolates (31/44, 70.5%).In total, 22 of the 30 canine (73.3%) and 9 of the 14 feline (64.3%) isolates were floR-positive.In contrast, none of the strains carried cfr (Table 8, Figure 1).Moreover, floR-positive bacteria exhibited a significantly limited or totally absent inhibition zone for florfenicol when tested by the disc diffusion method (Table S1).The mean inhibition zone in bacteria carrying floR was approximately 7.5 mm (the total absence of inhibition zone was measured at 6 mm, which is the diameter of the discs), whereas it was significantly wider (12.5 mm) in isolates where the gene was not detected (unpaired t-test, t = 8.2911, df = 42, p < 0.0001).Comparable results were observed for chlo- Moreover, floR-positive bacteria exhibited a significantly limited or totally absent inhibition zone for florfenicol when tested by the disc diffusion method (Table S1).The mean inhibition zone in bacteria carrying floR was approximately 7.5 mm (the total absence of inhibition zone was measured at 6 mm, which is the diameter of the discs), whereas it was significantly wider (12.5 mm) in isolates where the gene was not detected (unpaired t-test, t = 8.2911, df = 42, p < 0.0001).Comparable results were observed for chloramphenicol (6.9 mm for floR-positive and 10.6 mm for the rest) (unpaired t-test, t = 4.7952, df = 42, p < 0.0001).Thus, a correlation between the phenotype (inhibition zone diameter) and genotype (floR gene presence) was identified.
In reference to the respective MIC results, isolates FC12, FC14 and FC25 had an MIC of 16 for florfenicol, whereas all other isolates exhibited higher values (≥32).These three bacteria were all floR-negative.Correspondingly, the MIC for chloramphenicol for strains FC10, FC12, FC32 and FC39 was 32, while all other strains exhibited values ≥64.Only FC39 of the aforementioned group was floR-positive.Thus, a correlation between the presence of floR gene and higher MIC values is also indicated.

Discussion
The results of this study demonstrated characteristics of florfenicol-resistant Enterobacterales and the respective canine and feline clinical samples in Greece.Moreover, the floR gene is the most prevalent molecular resistance determinant.To our knowledge, this is the first report of florfenicol-resistant Enterobacterales and the first investigation of associated ARGs in companion animals in the country.In addition, although several relevant studies have been conducted worldwide on livestock animals, data regarding pets are limited.
The resistance rate of florfenicol in MDR isolates was 17.9%, which is definitely considerable for a second-line antibacterial agent.However, in comparison with the rates documented for other classes of antibiotics, available for usage in companion animals, this percentage could be considered adequate to indicate florfenicol's potential.Additionally, its usage entails a lower risk for human medicine than other treatment options for MDR bacteria [28].The results of this study are in accordance with most previous reports, as relatively low percentages of resistant isolates were identified in canine and feline Enterobacterales [29][30][31][32][33][34].Nevertheless, significant percentages of resistance were documented in a four-year surveillance study in China [35].The prevalence of resistance in specific areas can be enhanced by antibiotic consumption in livestock animals [6].In conclusion, even though current data about pets are limited, florfenicol demonstrated good efficacy against canine and feline isolates in most cases.Therefore, these results support the hypothesis that it could constitute an alternative "second-line" antibiotic for companion animals.
However, the presence of isolates demonstrating a florfenicol resistance profile raises concerns regarding their further distribution under the pressure of a significant limitation of available treatment options against MDR infections, which has been documented in veterinary medicine in recent decades.In addition, florfenicol resistance genes including floR, which was the only one detected in this study, are regularly plasmid-encoded [2,4].Thus, the danger of horizontal gene transfer under selection pressure is definitely not negligible, especially in cases of wide administration.Furthermore, the isolation of relevant strains from infection sites indicates the clinical importance of this issue, owing to the requirement for effective antibiotic treatment.Concludingly, even though the documented resistance rate could partially support its usage against MDR infections, this usage entails an important risk of distribution of resistance among bacterial populations.
In addition, certain limitations have to be taken into consideration regarding florfenicol's toxicity, administration, pharmacokinetic and pharmacodynamic properties.Particularly, the risk of bone marrow suppression, anaphylactic reactions and side effects from the gastrointestinal tract are notable concerns, especially for non-topical treatment [11].Moreover, its solubility in water is relatively low and thus the formulation of concentrated aqueous solutions in organic solvents (suitable for oral dosing) is difficult [36].Consequently, its administration in pets is challenging with the exception of the otic gel.Further research is undoubtedly essential for obtaining sufficient data for its safe and efficient usage.The application of new technologies, like nanotechnology, could be beneficial.The use of nanoemulsions for example increased its bioavailability in pigs [36], while more types of nanostructures could be evaluated in the future to improve its effectiveness, stability and drug delivery or reduce its dosage, frequency of administration and toxicity [11].
In this study, floR was the only identified gene, detected mostly in E. cloacae (n = 8), E. coli (n = 7), Proteus mirabilis (n = 7), and Klebsiella pneumoniae (n = 5).In relevant articles, including bacteria from pets, the floR gene has occasionally been detected, mainly in E. coli and Proteus mirabilis isolates [31,34,37].However, in one study, the gene was widely disseminated, possibly through nosocomial E. coli strains that caused infections in dogs [38].Its predominance in florfenicol-resistant Enterobacterales has also been demonstrated in studies on livestock animals [3,39].
The distribution of floR through plasmids between different species, such as Enterobacter cloacae and Citrobacter freundii, has also been highlighted [39].Both species were identified in this study.Moreover, horizontal transfer through plasmids has been identified in isolates of both human and animal origin, and the potential of spreading between animal and human pathogens has been pointed out [6,40].Accordingly, the susceptibility profiles documented in this study included extremely high resistance rates for specific classes of antibiotics, such as β-lactams, tetracyclines, and folate pathway inhibitors, which could be mediated by ARGs acquired through plasmids.In addition, co-localization of the respective genes has already been identified in canine gut bacteria [41].Therefore, concerns have arisen regarding the distribution of relevant mobile genetic elements (MGEs) among canine and feline Enterobacterales, conferring a multi-resistant profile to the associated strains.The possibility that these strains will spread further under the pressure of wide administration of agents commonly used in veterinary medicine is all but negligible.Thus, surveillance measures and sparing and targeted antibiotic consumption should be priorities.
The results of the disc diffusion and the MIC test in this study were associated with the presence or absence of a floR-specific exporter.Positive isolates regularly exhibited an extremely limited inhibition zone or a higher MIC value, respectively, for both florfenicol and chloramphenicol, compared to negative ones.This was anticipated because this gene confers high-level phenicol resistance and is commonly correlated with high MICs in AST [3,4].Thus, an indication of the basis of resistance is possibly provided by AST, but molecular investigation is essential for identification.
Despite the results of the susceptibility test, no ARGs (floR and cfr) were detected in a group of thirteen isolates.Therefore, the molecular basis of the florfenicol-resistant phenotypes is diverse.More mechanisms that could confer florfenicol resistance have been described, such as multidrug transporter systems, and several of them have been associated with E. coli [2,4].Thus, the high percentage (7/14, 50.0%) of florfenicol-resistant E. coli in the aforementioned group (Table 8) could be explained by acquisition of the respective mechanisms.Molecular investigations of additional ARGs, including multidrug efflux pumps, could provide sufficient data regarding the determinants of resistance and their distribution.
This study has certain limitations.Initially, only MDR bacteria were included, since florfenicol is not commonly used in companion animals, and thus not routinely tested in the AST in cases of susceptibility to "first line" antibiotics.Even though clinical samples were examined, data regarding each case history and previous medication were not evaluated.Furthermore, only species of the Enterobacterales order were selected and examined, whereas more species are regularly implicated in phenicol-resistance, such as staphylococci, enterococci or Acinetobacter spp.Additionally, the interpretation regarding the causative agent of the infection in some samples (especially soft tissue ones) was difficult and thus the possibility that some strains could constitute contamination is considerable.For example, the pathogenicity of some species like Pluralibacter gergoviae for dogs and cats is debatable.Moreover, in vitro susceptibility was estimated using zone diameter and MIC breakpoints from bovine respiratory pathogens and swine Salmonella enterica, respectively.The lack of specific breakpoints for canine and feline pathogens creates doubts regarding the reliability of the AST results and their actual evaluation in clinical practice.Finally, the presence of two florfenicol-specific ARGs was investigated, whereas more genes, like multidrug efflux pumps, could be implicated in the resistance.
Nonetheless, a large number of clinical samples originating from veterinary clinics throughout the country and collected during a two-year period was included in the initial selection phase.Moreover, all isolated bacteria could constitute opportunistic pathogens for both animals and humans and also contribute to the horizontal distribution of resistance and affect the outcome of antibiotic treatment.Therefore, these results are clinically significant and highly representative of the current situation in the country.
In reference to future research perspectives, an extensive surveillance study is essential in companion animals in Greece, since current data are limited, while the risk of the distribution of MDR strains through pets is definitely significant.This study should include data regarding the medication history of the animals in order to provide comprehensive epidemiological results and the association between preceding antibiotic treatment and emerging resistance.Moreover, an extensive molecular investigation of MGEs and horizontal gene transfer could provide sufficient data regarding the acquisition of resistance to several agents, including florfenicol.Finally, the potential of florfenicol as an alternative agent in veterinary medicine, alone or in combination, should be thoroughly examined in order to obtain sufficient information about its efficacy against various types of infection in companion animals, its toxicity and possible side effects during therapy; the application of new technologies should be also investigated as it could potentially improve the pharmacodynamic or pharmacokinetic properties.

Conclusions
Florfenicol could constitute a promising alternative antibiotic for companion animals, under the pressure of the emerging multi-drug resistance.However, specific aspects need to be thoroughly investigated for completely safe and well-targeted administration.In our results, even though florfenicol resistance among MDR Enterobacterales isolates was found to be relatively low, it was mostly mediated by the plasmid-located floR gene.This gene was identified in bacteria of different genera and species throughout the country.Hence, the possibility that floR is co-transferred through plasmids along with ARGs against other classes of commonly used antibiotics is significant.Therefore, even though the potential of florfenicol as an alternative agent is demonstrated, the establishment of continuous monitoring of its efficacy and antibiotic resistance profiles in companion animals is also necessary.

Figure 1 .
Figure 1.PCR gel electrophoresis images: (a) product of approximately 400bp size detected in isolates FC1, FC2, FC3, FC4, FC6, and FC9, indicative of the presence of floR; (b) product at 580 bp not detected, indicating the absence of cfr in isolates FC1 to FC12.

Figure 1 .
Figure 1.PCR gel electrophoresis images: (a) product of approximately 400bp size detected in isolates FC1, FC2, FC3, FC4, FC6, and FC9, indicative of the presence of floR; (b) product at 580 bp not detected, indicating the absence of cfr in isolates FC1 to FC12.

Table 2 .
Primers used in this study.

Table 3 .
Collection site and origin of the samples included in this study.
1 FC12 (E.coli) and FC17 (Enterobacter cloacae) strains were obtained from the same cat urine sample.

Table 4 .
Canine samples: sampling site, isolate species and other bacteria detected in the same samples.

Table 4 .
Cont.Isolates included in this section were obtained from the same sample as the selected bacteria.MRSP: Methicillinresistant Staphylococcus pseudintermedius.SDR: Single-drug resistant: The isolate that exhibits a resistant phenotype against antibiotics of one or two different classes, not counting the intrinsic resistance mechanisms of each species.MRSA: Methicillin-resistant Staphylococcus aureus.NDR: No drug resistance: The isolate that exhibits full susceptibility to the tested agents, with the exception of intrinsic resistance.ND: Not detected. 1

Table 5 .
Feline samples: sample site, isolate species and other bacteria detected in the same samples.

Table 6 .
Resistance rates of the 44 isolates included in this study by the disc diffusion method.

Table 7 .
Resistance rates of the 44 isolates included in this study by the MIC method.

Table 8 .
Distribution of floR-positive bacteria.

Table 8 .
Distribution of floR-positive bacteria.