Establishment of Epidemiological Cut-Off Values and the Distribution of Resistance Genes in Aeromonas hydrophila and Aeromonas veronii Isolated from Aquatic Animals

The emergence of antimicrobial-resistant bacteria is an enormous challenge to public health. Aeromonas hydrophila and Aeromonas veronii are opportunistic pathogens in fish. They exert tremendous adverse effects on aquaculture production, owing to their acquired antibiotic resistance. A few Clinical and Laboratory Standards Institute (CLSI) epidemiological cut-off values (ECVs) against Aeromonas spp. are available. We evaluated antimicrobial susceptibility by establishing 8 ECVs using two analytical methods, normalized resistance interpretation and ECOFFinder. We detected antimicrobial resistance genes in two motile Aeromonas spp. isolated from aquatic animals. Results showed that 89.2% of A. hydrophila and 75.8% of A. veronii isolates were non-wild types according to the oxytetracycline ECVCLSI and ECVNRI, respectively. The antimicrobial resistance genes included tetA, tetB, tetD, tetE, cat, floR, qnrA, qnrB, qnrS, strA-strB, and aac(6′)-1b. The most common tet gene in Aeromonas spp. isolates was tetE, followed by tetA. Some strains carried more than one tet gene, with tetA–tetD and tetA–tetE found in A. hydrophila; however, tetB was not detected in any of the strains. Furthermore, 18.6% of A. hydrophila and 24.2% of A. veronii isolates showed presumptive multidrug-resistant phenotypes. The emergence of multidrug resistance among aquatic aeromonads suggests the spread of drug resistance and difficult to treat bacterial infections.


Introduction
The genus Aeromonas comprises 36 species representing ubiquitous bacteria isolated from food, animal, and aquatic environments [1]. Among the salmonids, the genus Aeromonas is an enteric pathogen, which causes haemorrhagic septicaemia, fin rot, and soft-tissue rot companied by high mortality [2,3]. Aeromonas spp. produce a variety of toxins, including hemolysins, aerolysins, and cytotonic enterotoxins, which cause diarrhea, enteritis, and dysentery [4,5]. Aeromonas spp. are opportunistic bacteria commonly present in freshwater and marine environments, with Aeromonas salmonicida subsp. salmonicida, Aeromonas hydrophila, and Aeromonas veronii identified as causative agents of hemorrhagic skin ulcers and furunculosis in Nile tilapia, common carp, and channel catfish [1,[6][7][8][9]. Pathogenic Aeromonas spp. kills 80-100% of commercial carp within 1-2 weeks, resulting in the deterioration of production quality in fisheries [10]. The resulting unfavorable conditions, such as hypoxia or nitrogen-waste accumulation, induce a significant reduction in immune response leading to increased risk of pathogen translocation, infection, and disease [11]. β-lactam-, aminoglycoside-, and quinolone-resistant strains of Aeromonas spp. have been isolated from water and fish worldwide [12][13][14]. Resistant strains have been Table 1. MIC distribution of antimicrobial agents in 43 Aeromonas hydrophila isolates obtained from aquatic animals in Korea.

ECV Establishment Using Two Analytical Methods
We aimed to establish the ECVs for doxycycline, enrofloxacin, erythromycin, florfenicol, flumequine, gentamicin, neomycin, and oxytetracycline by testing 43 A. hydrophila and 33 A. veronii isolates from various diseased aquatic animals using the normalized resistance interpretation (NRI) and ECOFFinder methods. Figure 1 shows the histogram of MICs for eight antimicrobial agents against A. hydrophila using the NRI method. Based on the MIC distributions, the ECV NRI for doxycycline was 2 µg mL −1 . This categorized 23 (53.5%) isolates as non-wild type (NWT); they exhibited reduced susceptibility. The ECV NRI values for erythromycin and florfenicol were 64 µg mL −1 and 1 µg mL −1 , which categorized 23 (53.5%) isolates and 24 (55.8%) isolates as NWT, respectively. The ECV NRI values for enrofloxacin and flumequine were 32 µg mL −1 and 64 µg mL −1 , respectively; however, the standard deviation values of 1.2 log 2 indicated inadequate precision. The NRI calculations did not generate results for oxytetracycline. Figure 2 shows the histogram of MICs for eight antimicrobial agents and the 99.0% ECV (ECV 99 ), which was calculated using ECOFFinder software. The ECV 99 value for doxycycline was 128 µg mL −1 , indicating that no isolates could be considered NWT. The ECV 99 value for enrofloxacin and gentamicin was 16 µg mL −1 , which categorized 11 (25.6%) and six (14.0%) isolates as NWT, respectively. However, ECOFFinder failed to provide ECV 99 values for four antimicrobial agents (erythromycin, flumequine, neomycin, and oxytetracycline) revealing the lack of a normal distribution; this complicated the interpretation of the MIC distributions. Figure 3 shows the histogram of MIC for eight antimicrobial agents against A. veronii using the NRI method. The ECV NRI values for doxycycline and enrofloxacin were 1 µg mL −1 and 0.06 µg mL −1 , which categorized 10 (30.3%) and 25 (75.8%) isolates as NWT, respectively. Figure 4 shows the histogram of MICs for eight antimicrobial agents and the ECV 99 , The ECV 99 values for florfenicol and flumequine were 0.5 µg mL −1 and 2 µg mL −1 , which categorized seven (21.2%) and eight (24.2%) isolates as NWT, respectively. The ECV 99 values for gentamicin and neomycin were 8 µg mL −1 and 16 µg mL −1 , respectively.   Figure 2 shows the histogram of MICs for eight antimicrobial agents and the 99.0% ECV (ECV99), which was calculated using ECOFFinder software. The ECV99 value for doxycycline was 128 µg mL −1 , indicating that no isolates could be considered NWT. The ECV99 value for enrofloxacin and gentamicin was 16 µg mL −1 , which categorized 11 (25.6%) and six (14.0%) isolates as NWT, respectively. However, ECOFFinder failed to provide ECV99 values for four antimicrobial agents (erythromycin, flumequine, neomycin, and oxytetracycline) revealing the lack of a normal distribution; this complicated the interpretation of the MIC distributions.   Figure 3 shows the histogram of MIC for eight antimicrobial agents against A. veronii using the NRI method. The ECVNRI values for doxycycline and enrofloxacin were 1 µg mL −1 and 0.06 µg mL −1 , which categorized 10 (30.3%) and 25 (75.8%) isolates as NWT, respectively.     Oxytetracycline did not allow for ECV99 calculation. ECV, epidemiological cut-off value; MIC, minimum inhibitory concentration; WT, wild type.

Comparison of the ECVCLSI, ECVNRI, and ECV99
We compared the ECVs of eight antimicrobial agents for A. hydrophila and A. veronii isolates using the CLSI, NRI, and ECOFFinder methods. There is no breakpoint for the two Aeromonas spp. isolates; however, recently, the CLSI provided six ECVs for A. hydrophila [18]. The ECVCLSI and ECVNRI for erythromycin against A. hydrophila, was 64 µg mL −1 (Table 3). Additionally, the ECVNRI and ECV99 for gentamicin was 16 µg mL −1 , which

Comparison of the ECV CLSI , ECV NRI , and ECV 99
We compared the ECVs of eight antimicrobial agents for A. hydrophila and A. veronii isolates using the CLSI, NRI, and ECOFFinder methods. There is no breakpoint for the two Aeromonas spp. isolates; however, recently, the CLSI provided six ECVs for A. hydrophila [18]. The ECV CLSI and ECV NRI for erythromycin against A. hydrophila, was 64 µg mL −1 (Table 3).
Additionally, the ECV NRI and ECV 99 for gentamicin was 16 µg mL −1 , which was two-fold higher than ECV CLSI . The ECV 99 for enrofloxacin was 16 µg mL −1 , which was more than nine dilution steps from the ECV CLSI . Among the ECVs for the eight antimicrobials, the ECV for florfenicol was optimal, showing the least 1-fold dilution between ECV CLSI and ECV NRI or ECV 99 . We calculated values for flumequine and neomycin using only the NRI method. The CLSI has not provided the breakpoint or ECVs for A. veronii. The ECV NRI and ECV 99 values for enrofloxacin (0.06 µg mL −1 ) and erythromycin (32 µg mL −1 ) were the same (Table 4), whereas the ECV NRI values for florfenicol, gentamicin, and neomycin were one-fold higher than the ECV 99 values. Oxytetracycline was evaluated using only the NRI method with 0.5 µg mL −1 as the ECV NRI value. Table 3. Comparison of the ECVs of eight antimicrobial agents for Aeromonas hydrophila isolates based on the CLSI, NRI, and ECOFFinder methods.

Species
Antimicrobial

Presumptive Multidrug-Resistant (pMDR) Aeromonas spp. Isolates
A total of 18.6% (n = 8) of the isolates presented a pMDR phenotype, suggesting that multiple antimicrobial resistance is a common phenomenon in A. hydrophila (Table 5). All isolates from Anguilla japonica, Silurus asotus, Salmo salar, and Misgurnus mizolepis were resistant to three or more classes of antimicrobials. One isolate was resistant to seven antimicrobial agents, and five isolates were resistant to six agents. Additionally, 24.2% (n = 8) of A. veronii isolates presented the pMDR phenotype, and were highly resistant to enrofloxacin, florfenicol, and oxytetracycline. None of the isolates were resistant to all the eight antimicrobial agents.

Distribution of Antimicrobial Resistance Genes (ARGs)
We analyzed four tet genes (tetA, tetB, tetD, and tetE) encoding proteins involved in tetracycline efflux ( Figure 5). In A. hydrophila, all the tet-positive isolates (35 isolates) were oxytetracycline NWT at ECV CLSI ( Figure 5A). The most common tet gene was tetE, which was found in 14 (40%) NWT isolates, followed by tetA, which was found in 12 (34.3%) NWT isolates. Some of the isolates carried more than one tet gene, with tetA-tetD (three isolates) and tetA-tetE (five isolates) related to the oxytetracycline MICs ranging from 32 µg mL −1 to 256 µg mL −1 and demonstrating high resistance to oxytetracycline. The tetB gene was not detected in any of the strains. We analyzed the four tet genes in A. veronii ( Figure 6). In A. veronii, all the tet-positive isolates (25 isolates) were oxytetracycline NWT at ECV NRI ( Figure 6A), and the most common tet gene was tetE, which was found in 13 (52%) of the NWT isolates. Additionally, A. veronii isolates with MICs of 64 µg mL −1 (two strains) and 128 µg mL −1 (one strain) carried two tet genes, (tetA-tetE and tetD-tetE, respectively). The tetB gene was not detected in any of the strains.

Quality Control (QC)
Eight antimicrobial agents of QC MICs for Escherichia coli ATCC 25922, Aeromonas salmonicida subsp. salmonicida ATCC 33658, and Enterococcus faecalis ATCC 29212 were within the acceptable range (94.3 to 100%) for the standard broth-microdilution method, as stipulated by the CLSI documents, M45, M7, and VET04 [18,24,25]. The results for doxycycline and neomycin against A. salmonicida ATCC 33658 were excluded from the QC, because of the lack of an established acceptable range in CLSI document VET04. Table S1 shows the MICs for the QC strains.

Discussion
The development of multiple antibiotic resistance strains of A. hydrophila and A. veronii in recent years is a serious public health concern, because of the possibility of their transmission from infected fish or water sources to humans and the subsequent infections [26]. In this study, we established eight ECVs for 43 A. hydrophila and 33 A. veronii isolates from aquatic animals and evaluated their ARG distributions. Some ECV CLSI values were suggested for A. hydrophila [18]. The lack of clinical breakpoints or guidelines to interpret ECVs for A. veronii prompted the use of two methods for determining ECVs and interpreting the antimicrobial susceptibility of A. hydrophila and A. veronii.
Three antimicrobials (doxycycline, enrofloxacin, and oxytetracycline) exhibited bimodal MIC distributions, which revealed two clearly distinct populations of A. hydrophila and A. veronii. Based on these distributions, the calculated MIC 50 (4 µg mL −1 ) for gentamicin against A. hydrophila and A. veronii was higher than 1 µg mL −1 . This is in line with that reported for 138 Aeromonas spp. isolates recovered from European rivers [27]. The MIC 50 and MIC 90 values for oxytetracycline were 34.97 µg mL −1 and 149.26 µg mL −1 , respectively, for 64 pathogenic Aeromonas strains isolated from ornamental fish [28]. Similarly, the MIC 50 values were ≤2 µg mL −1 for florfenicol, 8 µg mL −1 for oxytetracycline, and 0.5 µg mL −1 for ciprofloxacin for 72 aeromonads isolated from koi carp [29]. These findings suggested that the isolates obtained 10 years ago were more susceptible to these drugs.
Tetracycline classes, including oxytetracycline and doxycycline, are broad-spectrum agents extensively used to treat bacterial infections and prevent infections in aquaculture. However, oxytetracycline is poorly absorbed in the fish gut; therefore, it must be administered at high doses [30]. This study showed that 89.2% of A. hydrophila could be categorized as NWT upon applying an oxytetracycline ECV CLSI of 0.25 µg mL −1 ; 75.8% of A. veronii were determined as NWT upon applying an oxytetracycline ECV NRI of 0.5 µg mL −1 . This confirmed the high resistance rate in Aeromonas spp. However, 33 Aeromonas isolates (14.2%) recovered from 16 rivers were considered NWT for tetracycline (23), and 39 Aeromonas isolates (40.6%) from different fish species with reduced susceptibility to tetracycline were classified as NWT [23]. Additionally, A. hydrophila isolates from tilapia, carp, and channel catfish were more susceptible to doxycycline than to oxytetracycline [31]. Aeromonas spp. easily develop single or multiple antibiotic resistance phenotypes and are generally resistant to tetracyclines, quinolones, and β-lactams [5,32]. Moreover, tetracycline-resistant Aeromonas isolates are observed in wastewater discharge, lakes, and carp ponds [32][33][34][35]. In this study, we found that 62.8% of A. hydrophila isolates and 75.8% of A. veronii NWT isolates harbored tetA, tetD, tetE, or more than one tet gene, indicating that the WT isolates did not possess any tet genes. Aeromonas spp. isolates predominantly carried tetE, followed by tetA. However, 37% of A. veronii isolates recovered from channel catfish carried tetE, and 3.8% of isolates carried tetA [36]. Furthermore, A. hydrophila isolates showing oxytetracycline MICs ranging from 32-256 µg mL −1 harbored more than one tet gene (tetA-tetE and tetA-tetD), indicating that the degree of oxytetracycline resistance was associated with the number and type of tet genes present. E. coli isolates harboring tetA and tetB or tetA and tetC exhibited high MICs for tetracycline (256 µg mL −1 ) or oxytetracycline (512 µg mL −1 ) [37]. The ECV CLSI for A. hydrophila and ECV NRI for A. veronii might account for the correlations between the NWT isolates and the distribution of resistance genes.
In Korea, florfenicol is approved for use against bacterial diseases in Oncorhynchus mykiss, A. japonica, and Seriola quinqueradiata [38]. The ECV NRI for florfenicol is 1 µg mL −1 for A. hydrophila (55.8%) and A. veronii (21.2%), which were categorized as NWT with reduced susceptibility. However, 2.1% isolates of Aeromonas spp. are NWT considering the ECV NRI (2 µg mL −1 ) [21], and 25.5% are NWT considering the ECV NRI (4 µg mL −1 ) [23]. The high frequency of NWT isolates from Korea could be associated with the excessive use of antimicrobial agents in aquaculture; the recorded florfenicol sales was approximately six tons in 2019 [39]. Additionally, we detected cat and floR in A. hydrophila and A. veronii NWT isolates; both genes are associated with high MICs. A total of 7.5% A. veronii isolates harbored floR, which conferred resistance to florfenicol [36]. A resistance cassette, carrying the floR gene in A. salmonicida enables mobilization [40]. The first floR-containing plasmid was discovered in Aeromonas bestiarum [41]. Interestingly, the presence of cat was related to a low MIC for florfenicol (0.25 or 0.5 µg mL −1 ). These results indicated a higher correlation between the presence of floR and NWT categorization, compared to that with the presence of cat.
Enrofloxacin is a member of the fluoroquinolone family of antibiotics and exhibits strong bactericidal activity against aerobic and facultative anaerobic bacteria [42]. For A. hydrophila, the ECV CLSI of 0.03 µg mL −1 was lower than the ECV 99 of 16 µg mL −1 , indicating that lowering the ECV would increase the likelihood of identifying resistance genes or mutants while increasing the risk of misclassifying the number of WT isolates. Based on our findings, an ECV CLSI of 0.03 µg mL −1 would misclassify 58.1% of NWT (25 isolates), compared to an ECV 99 of 16 µg mL −1 . We mostly detected qnrS in A. hydrophila and A. veronii NWT isolates; therefore, ECVs should be established in detail based on the ARG distributions. qnrS was the most prevalent, with its presence in 68% of aeromonad isolates that demonstrated high levels of resistance to nalidixic acid and ciprofloxacin; no amplicon was detected for qnrA [43]. The detection of the factors enabling plasmidmediated quinolone resistance indicated that the complex Aeromonas mobilome increases the possibility of horizontal gene transfer, including that of qnrS and qnrB.
Erythromycin is not approved for use in the USA; however, Aeromonas strains highly resistant to erythromycin have been isolated from foreign countries [44]. Additionally, Aeromonas spp. are resistant to penicillin, cephalosporins, vancomycin, and erythromycin [45,46]. In this study, 53.5% and 15.2% of A. hydrophila and A. veronii, respectively, were categorized as NWT upon application of the erythromycin ECV NRI . Similarly, 50% and 53% of aeromonads isolated from lakes and chickens, respectively, showed resistance to erythromycin [47,48]. Furthermore, harboring macrolide MacB ABC transporter genes confers erythromycin resistance; the MacA gene regulates the drug-binding and ATPase activity of MacB [49]. We did not investigate the distribution of macrolide resistance genes; further studies are required to elucidate the cause underlying the acquisition of erythromycin resistance, owing to the high prevalence of erythromycin NWT Aeromonas spp. isolates.
The results showed that 3% of A. veronii was classified as NWT upon application of the gentamicin ECV NRI and ECV 99 . Consistent with these findings, 2% of Aeromonas spp. exhibited gentamicin resistance; however, no A. veronii isolate was resistant to gentamicin [50]. We did not detect any aminoglycoside-resistance genes among the 31 A. veronii isolates (94%). However, in an earlier report, all Aeromonas spp. isolates recovered from marketed cockles harbored aac(6 )-1b, with strA-strB found in 41% of the isolates [43]. The recommended first-line therapeutic options for Aeromonas infections are aminoglycosides and fluoroquinolones. We identified gentamicin as an aquatic medicine that can be inoculated orally to prevent Aeromonas infection. Its appropriate use could potentially prevent the emergence of new resistant strains.
The resistance phenotypes varied among isolates. The pMDR of A. hydrophila, which was resistant to three or more classes of antimicrobials, was 18.6%; this was lower than that observed in a previous study conducted on tilapia where 64% of isolates were resistant to six to eight drugs [31] and that in 95 motile pMDR aeromonads isolated from freshwater [46]. Additionally, multi-antibiotic resistant Aeromonas spp. isolates harbored a tripartite AheABC efflux pump, and the use of phenylalanine-arginine-β-naphthylamide contributed to intrinsic resistance [51]. Among the Aeromonas spp. isolates identified as pMDR, the most common resistance was against oxytetracycline (100%). Oxytetracycline is among the most commonly used antibiotics in humans and animals, and these results are consistent with those of a previous study [52]. The distribution of strains resistant to oxytetracycline has increased with the global use of antibiotics; the emergence of pMDR strains complicates the selection of available therapeutics.
This study provides eight putative ECVs for classifying WT and NWT isolates; however, the findings should not be used as Aeromonas-pathogen-treatment guidelines. These ECVs were derived from one laboratory; therefore, it is essential to evaluate different sources and a large number of isolates for reliably establishing ECVs for each Aeromonas strain [53]. The results from this study can be used as a foundation to establish clinical breakpoints for each Aeromonas strain. Additionally, it is necessary to study the NWT bacterial transcriptome and the mechanism of antibiotic resistance transmission between humans and fish to determine the cause of resistance acquisition.

Collection and Isolation of Aeromonas spp.
Aeromonas spp. isolates were collected between 2008 and 2020 from eight Korean provinces (Chungbuk, Chungnam, Gyeongbuk, Gyeongnam, Gyeonggi, Jeonbuk, Jeonnam, and Gangwon), with 43 A. hydrophila isolates recovered from A. japonica (n = 25), Carassius carassius (n = 3), S. asotus (n = 3), Cyprinus carpio nudus (n = 2), Sebastes schlegelii (n = 2), and others (n = 8); and 33 A. veronii isolates recovered from A. japonica (n = 13), C. carpio nudus (n = 9), C. carassius (n = 4), S. asotus (n = 3), and others (n = 4) (Figure 7). The bacterial strains are listed in Tables S2 and S3. The fish species were sampled from among seemingly healthy, clinical-subclinical, and moribund fish that differed by the year and region of collection. Samples were taken from the lesions, kidneys, and spleens of fish. All experiments were performed in accordance with Directive 2010/63/EU of the European Parliament and the Council (22 September 2010) on the protection of animals used for scientific purposes. Aeromonas spp. isolates were grown on Aeromonas agar (MB cells, Los Angeles, CA, USA), incubated at 37 • C for 24 h. Presumptive aeromonad colonies showing typical dark-green opaque color with a dark center were chosen and subjected to molecular identification. Genomic DNA was extracted from a single colony using a QIAmp DNA blood mini kit (Qiagen, Milan, Italy), according to the manufacturer instructions. DNA concentration and purity were quantified using a Nano Drop R 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) DNA was stored at −80 • C until use. Aeromonas spp. isolates were stored at −80 • C in tryptic soy broth (Merck, Kenilworth, NJ, USA) supplemented with 20% glycerol until further use. scientific purposes. Aeromonas spp. isolates were grown on Aeromonas agar (MB cells, Los Angeles, CA, USA), incubated at 37 °C for 24 h. Presumptive aeromonad colonies showing typical dark-green opaque color with a dark center were chosen and subjected to molecular identification. Genomic DNA was extracted from a single colony using a QIAmp DNA blood mini kit (Qiagen, Milan, Italy), according to the manufacturer instructions. DNA concentration and purity were quantified using a Nano Drop R 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) DNA was stored at-80 °C until use. Aeromonas spp. isolates were stored at −80 °C in tryptic soy broth (Merck, Kenilworth, NJ, USA) supplemented with 20% glycerol until further use.

Antimicrobial Susceptibility Test
Antimicrobial susceptibility tests were performed according to the broth microdilution method described in the CLSI guidelines VET04 [17,18]. The antimicrobial agents for Aeromonas spp. isolates are licensed and commonly used for aquatic animals in Korea [38]. The MICs of 43 A. hydrophila and 33 A. veronii isolates were tested using Sensititre CAMPY2 and KRAQ1 plates (Trek Diagnostics System, Cleveland, OH, USA). MICs for erythromycin (0.03-64 mg L −1 ), florfenicol (0.03-64 mg L −1 ), and gentamicin (0.12-32 mg L −1 ) were tested using CAMPY2; and those for doxycycline (0.25-64 mg L −1 ), enrofloxacin (0.03-32 mg L −1 ), flumequine (0.12-128 mg L −1 ), neomycin (0.5-64 mg L −1 ), and oxytetracycline (0.25-256 mg L −1 ) were tested using KRAQ1. Isolates were cultured on tryptic soy agar for 24 h at 28 • C, after which a suspension was prepared in sterile saline solution, adjusted to 0.5 McFarland standard, and diluted to reach a final inoculum concentration of 5 × 10 5 CFU/mL using a Nephelometer ® (V3011, Thermo Scientific, Roskilde, Denmark) to standardize inoculum density/turbidity. Microplates were incubated at 28 • C for 24 h for A. hydrophila and A. veronii. MICs were defined as the lowest drug concentrations that inhibited growth, compared to that in the drug-free growth control. E. coli ATCC 25922, A. salmonicida subsp. salmonicida ATCC 33658, and E. faecalis ATCC 29212 were included in the susceptibility test as QC strains. Recently, additional MICs of ECVs were made available for A. hydrophila in the updated CLSI guidelines [18]. We compared the A. hydrophila and A. veronii isolates among WT and NWT populations, according to the CLSI guidelines and the provisional ECVs proposed in this study.

Determination of Provisional ECVs
ECVs were calculated using two methods: NRI [55] and ECOFFinder [56]. The NRI method is a fully automatic and freely available Excel spreadsheet calculator (last updated in 2019; http://www.bioscand.se/nri) (accessed on 3 May 2021). The ECOFFinder method (v.2.1; last updated in 2020) is available from the EUCAST website (https://www.eucast. org/mic_distributions_and_ecoffs) (accessed on 3 May 2021). In this study, ECV determination was based on the distribution of antimicrobial MICs for each drug against A. hydrophila and A. veronii. ECV allows isolates to be categorized as WT at ≤x mg L −1 and NWT as >x mg L −1 . A 99.0% cut-off was applied, which means that approximately 99.0% of the WT MIC distribution was less than the identified ECV. pMDR was defined as resistance to more than three antimicrobial agents, classes, or subclasses of antimicrobial categories [57]. The number of pMDR Aeromonas was determined for eight antimicrobial agents (doxycycline, enrofloxacin, erythromycin, florfenicol, flumequine, gentamicin, neomycin, and oxytetracycline) in the clinical samples.

Terminology
When referring to the categorization of isolates based on their susceptibility, we followed the recommendations, which suggested that when isolates are categorized by applying ECVs, the terms "sensitive" and "resistant" should not be used [58]. WT is defined, for a fully susceptible species, as the absence of acquired-and mutational-resistance mechanisms to the drug, and NWT is defined as the reduced susceptibility to the presence of an acquired-or mutational-resistance mechanism to the drug. However, when referring to studies that used the term "resistant", we did not change their terminology. The CLSI uses the abbreviation "ECV" for epidemiological cut-off values, whereas EUCAST uses the ECOFF. This study used "ECV" to prevent confusion when comparing the ECOFF values using the two analytical methods.

Analysis of ARGs
We tested 43 A. hydrophila and 33 A. veronii isolates for the presence of ARGs, including tetA, tetB, tetD, and tetE for tetracycline; cat and floR for phenicol; qnr-type pentapeptide proteins encoded by qnrA, qnrB, and qnrS for quinolone; and strA-strB and aac(6 )-1b for aminoglycosides ( Table 7). The primers used to detect these genes were selected from previous studies. The PCR cycling conditions were as follows: 94 • C for 5 min, followed by 35 cycles of 95 • C for 30 s, annealing for 30 s at different temperatures, 72 • C for 30 s, and 72 • C for 5 min. The PCR products were separated using electrophoresis on a 1% agarose gel and purified for sequencing. Sequence identities were confirmed using the sequence information in the NCBI database (on https://www.ncbi.nlm.nih.gov/) (accessed on 22 June 2021).

Conclusions
This is the first study to establish ECV NRI and ECV 99 values for eight antimicrobials against 43 A. hydrophila and 33 A. veronii isolates recovered from aquatic animals in Korea and to detect ARGs in Aeromonas strains. A total of 89.2% A. hydrophila isolates and 75.8% A. veronii isolates were classified as NWT against oxytetracycline; they harbored tet genes; Aeromonas spp. isolates predominantly carried tetE, followed by tetA. Additionally, the distribution of floR and qnrS was prevalent in NWT isolates, whereas no aac(6 )-1b or strA-strB was detected in the 31 A. veronii isolates. The emergence of antibiotic-resistant strains of Aeromonas spp. reduces the choice of currently available therapeutic agents and it could lead to prolonged Aeromonas infections. Therefore, these results can potentially help aquaculture managers and researchers alleviate Aeromonas infections in aquaculture systems and raise awareness of the appropriate use of antimicrobials in aquaculture. Furthermore, these findings encourage the application of vaccination or herbal therapy, to reduce antibiotic resistance and public health problems.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics11030343/s1, Table S1: CLSI-approved broth microdilution MIC QC ranges determined for eight antimicrobial agents against selected reference strains; Table S2: Isolate year, fish species, disease outbreak, isolation source, and geographical location of the 43 A. hydrophila strains; Table S3: Isolate year, fish species, disease outbreak, isolation source, and geographical location of the 33 A. veronii strains.