Antimicrobial Resistance and Type III Secretion System Virulotypes of Pseudomonas aeruginosa Isolates from Dogs and Cats in Primary Veterinary Hospitals in Japan: Identification of the International High-Risk Clone Sequence Type 235

ABSTRACT This study aimed to investigate the current trends in antimicrobial resistance among Pseudomonas aeruginosa clinical isolates of canine and feline origin and the prevalence of their sequence types (STs) and type III secretion system (T3SS) virulotypes, which remains unknown in Japan. A total of 240 nonduplicate clinical isolates of P. aeruginosa from dogs (n = 206) and cats (n = 34) collected from 152 primary care animal hospitals between August 2017 and October 2019 were examined. PCR detection of T3SS genes (exoU and exoS) and carbapenemase genes, multilocus sequence typing, and whole-genome sequencing of the representative carbapenem-resistant isolates were performed. Resistance rates to imipenem and meropenem were 6.67% and 2.08%, respectively. A high resistance rate (17.92%) was encountered with ciprofloxacin. The exoU−/exoS+ was the predominant T3SS virulotype (195 isolates, 81.3%), followed by exoU+/exoS− (35 isolates, 14.6%), exoU−/exoS− (7 isolates, 2.9%), and exoU+/exoS+ (3 isolates, 1.3%). A high frequency of the high-risk clones ST235 and clonal complex 235 (CC 235) (28.9%), followed by ST357 (21.1%), were noted among these 38 exoU+ isolates. Seventeen carbapenem-resistant isolates comprising 2 exoU+ isolates, including an ST235 isolate, and 15 exoU−/exoS+ isolates belonging to non-ST235/CC235 were detected, of which all were carbapenemase negative. Different combinations of mutations among oprD, efflux pump regulatory genes, and AmpC β-lactamase regulatory genes were identified among representative isolates with high-level resistance to imipenem. This study emphasizes the occurrence of ST235 isolates among companion animals, which may represent a threat to public health because of the ability of this clone to acquire and spread resistance elements, including carbapenemase genes. IMPORTANCE Pseudomonas aeruginosa is an environmentally ubiquitous and important opportunistic human pathogen responsible for life-threatening health care-associated infections. Because of its extensive repertoire of virulence determinants and intrinsic and acquired resistance mechanisms, the organism could be one of the most clinically and epidemiologically important causes of morbidity and mortality. In recent years, worldwide spreading of multidrug-resistant high-risk clones, particularly sequence type 235 (ST235), has become a serious public health threat. Companion animals which share much of their living environment with humans could be important reservoirs and spreaders of antimicrobial-resistant bacteria and resistance genes of clinical importance in humans, such as extended-spectrum β-lactamase-producing Enterobacterales and methicillin-resistant Staphylococcus aureus. However, antimicrobial resistance, virulence, and genotyping of P. aeruginosa in companion animals remain largely unknown. This work sheds light on the potential spread of high-risk clones in companion animals.

their STs and T3SS virulotypes, which remains unknown in Japan, to better understand their epidemiological aspects.
The presence of 40 mg/liter phenylalanine-arginine b-naphthylamide (PabN) did not inhibit the growth of 43 ciprofloxacin-resistant P. aeruginosa isolates except 1, for which the PAbN concentration not affecting its growth (20 mg/liter) was used for the efflux pump inhibition. All 43 isolates, including 7 carbapenem-resistant isolates, showed more than or equal to a 4-fold reduction in the MIC of ciprofloxacin in the presence of PAbN (data not shown).
Multilocus sequence typing (MLST) of exoU+ isolates. The 35 exoU1/exoS2 isolates were assigned to 20 STs from which 2 new STs, namely, ST235-like and ST3654, were identified (Table 3). The 3 exoU1/exoS1 isolates belonged to different STs, including a new ST, ST3653. The high frequency of high-risk clone ST235 (8, 21.1%) and ST141 Phenotypic characteristics of carbapenem-resistant P. aeruginosa isolates. Among the 240 isolates, 17 (7.1%) were resistant to imipenem and/or meropenem. Detection of carbapenemase production by a modified carbapenem inactivation method using Tris-HCl (CIMTris) assay yielded negative results for all 17 isolates. In those isolates, the major carbapenemase genes bla IMP , bla NDM , bla VIM , and bla GES were not detected (Table 4).    Sequence type newly assigned in this study.
c PAbN concentration of 20 mg/liter was used.
The presence of 40-mg/liter PAbN did not inhibit the growth of 17 carbapenem-resistant P. aeruginosa, except for 2 isolates (no. 8 and 10), for which PAbN concentration not affecting their growth (20 mg/liter) was used for the efflux pump inhibition. A 4-fold or greater decrease in the MICs of imipenem and/or meropenem with the addition of PAbN was observed in 8 of 17 isolates (47.1%) ( Table 4). In the presence of 3-aminophenylboronic acid (APB), an inhibitor of AmpC b-lactamase, most isolates (15/16, 93.8%) exhibited a greater than or equal to an 8fold reduction in the MICs of imipenem, while MICs of meropenem for 5 isolates were not affected. It was noticed that APB and PAbN plus APB effectively change interpretive categories from imipenem resistant/intermediate (MIC, $8/4 mg/liter) to susceptible (MIC, #2 mg/liter) in 13 (81.3%) and 15 isolates (93.8%), respectively. Meropenem MIC of an isolate (no. 12) with the imipenem-susceptible, meropenem-resistant phenotype was affected by PAbN (4-fold reduction) but was not affected by APB.

DISCUSSION
The perspective of One Health emphasizes the importance of AMR monitoring and integrated actions across human, food, animal, and environmental health. However, there are few studies on the epidemiology of antimicrobial resistance in P. aeruginosa clinical isolates from companion animals that share a living environment with humans. The strength of our study is that we included 240 P. aeruginosa clinical isolates collected from 152 primary care animal hospitals across the country, allowing us to understand the current situation of their antimicrobial resistance in companion animals as well as to investigate the resistance trends over time in Japan by comparing the results with those in earlier studies (19,20). Namely, a trend toward a progressive increase in the frequency of imipenem resistance, from 0% in 2003 to 2010 (19), to 0.5% in 2014 to 2015 (20), to 6.67% in 2017 to 2019 (this study) is noted according to the breakpoints of the Clinical and Laboratory Standards Institute (CLSI) M100 30th edition guidelines (26).
The carbapenem resistance rates of isolates from companion animals were found to be similar to those from human outpatients in Japan, leading to concerns about the  (19,20). Constant amounts of veterinary fluoroquinolones such as enrofloxacin and orbifloxacin, at 0.90 tons per year during 2013 and 2018, have been used for dogs and cats as described in the Nippon AMR One Health Report (NAOR) 2020 (in Japanese, https:// www.mhlw.go.jp/content/10900000/000715546.pdf). However, consumption of fluoroquinolones which are approved for human use such as levofloxacin and ciprofloxacin in companion animals cannot be traced because their prescription is left to the discretion of veterinarians in Japan. Thus, attention is necessary to monitor for high levels of resistance to ciprofloxacin, which is an active metabolite of enrofloxacin in companion animals, because ciprofloxacin is an effective treatment option in many human diseases. In contrast, piperacillin, piperacillin-tazobactam, ceftazidime, and cefepime retained activity against .99% of animal isolates, which is greater than the activity against isolates from hospital inpatients and outpatients (Table 1) To our knowledge, only one study has analyzed the T3SS virulotypes in a limited number of P. aeruginosa isolates from canine, reporting that 3 of 19 isolates from ocular infections were exoU1 genotypes (18). This study revealed that the prevalence of T3SS virulence genes among 240 clinical isolates of canine and feline origin (15.8% for exoU and 82.5% for exoS) is more similar to that of 90 environmental isolates (17.8% for exoU and 82.2% for exoS) than that of 243 human isolates from bloodstream infections (20.6% for exoU and 76.1% for exoS) (12,27). There was no significant association between the exoU1/exoS2 virulotype and ciprofloxacin and carbapenem resistance, contrary to some other studies reporting such an association in human isolates (11,28). However, the exoU1 genotype has been an independent risk factor for early mortality of human bloodstream infections (11,12), and our results draw attention to the prevalence of exoU1 isolates in a greater number of animal isolates. Among global high-risk clones with MDR/XDR profiles, ST235 is highly associated with exoU, and this combination has been a predictor of a highly unfavorable prognosis (29). Thus, MLST analysis of exoU1 isolates is of clinical importance, as ST235 has not been confirmed among P. aeruginosa isolates from companion animals. MLST revealed clonal heterogeneity in the 38 exoU1 isolates, including exoU1/exoS2 and exoU1/exoS1 virulotypes, with the presence of 20 STs, including 3 newly identified STs. Notably, the global high-risk clone ST235 was detected in 7 exoU1/exoS2 isolates and in 1 exoU1/exoS1 isolate. Moreover, 3 CC235 isolates were found in 2 exoU1/exoS2 isolates and in 1 exoU1/exoS1 isolate. Thus, 11 ST235/CC235 isolates were detected in 38 exoU1 isolates (28.9%), resulting in an overall ST235/CC235 frequency of 20.8% in 53 isolates with the exoU1 genotype and/or carbapenem resistance that were typed by MLST. In contrast, ST235/CC235 isolates were not found in 15 exoU1/exoS2 isolates resistant to carbapenems. The CC235 has been linked with metallo-b-lactamase genes and, hence, is responsible for their dissemination worldwide, including to Japan (4,30). Our findings provide important insights into the role of companion animals as the potential reservoirs of P. aeruginosa high-risk clones, although they did not harbor metallo-b-lactamase genes.
In P. aeruginosa, mutational overexpression of a multidrug efflux pump, MexAB-OprM, leads to reduced susceptibility to fluoroquinolones and most b-lactams, including meropenem, while imipenem is not affected (31). MexAB-OprM overexpression combined with OprD inactivation produces high-level meropenem resistance (MIC, .32 mg/liter) (32). Mutational inactivation of OprD contributes mainly to imipenem resistance, even achieving high-level resistance (MIC, .32 mg/liter), but drives only moderate resistance to meropenem (32). In addition, OprD inactivation, in combination with AmpC b-lactamase overexpression, confers resistance to carbapenems (33). In the present study, 7.1% (17/240) of isolates from companion animals were found to be carbapenem resistant but carbapenemase non-IMP producers. MIC assays performed in the presence of an efflux pump inhibitor PAbN and/or AmpC b-lactamase inhibitor APB revealed the contribution of AmpC b-lactamase production to imipenem resistance in 10 of 12 imipenem-resistant meropenem-susceptible isolates. In the remaining 2 isolates, the active role of efflux pumps (isolate 8) or the synergistic role of AmpC b-lactamase production and efflux pumps (isolate 2, imipenem MIC of 64 mg/ml) in mediating imipenem resistance is estimated. Among 4 imipenem-and meropenem-resistant isolates, AmpC b-lactamase-mediated imipenem resistance is considered in 3 isolates, including 1 also exhibiting efflux pump-mediated meropenem resistance (isolate 13, imipenem MIC of 32 mg/liter, meropenem MIC of 8 mg/liter). The involvement of the synergistic effect of AmpC b-lactamase production and efflux pumps in imipenem resistance and that of the efflux pump mechanism in meropenem resistance could be considered in the remaining 1 isolate (isolate 10, imipenem MIC of 128 mg/ml, meropenem MIC of 16 mg/liter). Efflux pump-mediated meropenem resistance is suggested for a meropenem-resistant imipenem-susceptible isolate.
In conclusion, this study confirms the presence of high-risk clones in canine and feline isolates with the exoU1 genotype; ST235/CC235 isolates were the most predominant, followed by ST357 isolates. These animal isolates belonging to high-risk clones were found to be carbapenemase nonproducers, although international high-risk clones have been associated frequently with carbapenemase production. Nonetheless, the predominant occurrence of ST235 among companion animals may represent a threat to public health because of the ability of this clone to acquire and spread resistance elements, including carbapenemase genes. A comprehensive survey of P. aeruginosa needs to be conducted to better understand the spread of antimicrobial resistance elements, STs, and T3SS virulotypes at the human-animal-environment interface and to assess their clinical implications on humans and animals.
Antimicrobial susceptibility testing. MICs were determined by the broth microdilution method recommended by the CLSI using a custom-designed microtiter panel (Kyokuto Optopanel MP; Kyokuto Pharmaceutical, Tokyo, Japan), and the results were interpreted according to CLSI M100 30th edition guidelines for human (26). Escherichia coli ATCC 25922 and P. aeruginosa ATCC 27853 were used as quality-control strains. Detection of T3SS virulence genes (exoU and exoS). All isolates were subjected to PCR detection of T3SS effector toxin genes exoU and exoS using specific primers as shown in Table 6 (44,45). The amplification products from some arbitrarily selected isolates were sequenced to confirm the reliability of the results.
Analysis of carbapenem-resistant P. aeruginosa isolates. Detection of carbapenemase activity in P. aeruginosa isolates exhibiting resistance to imipenem and/or meropenem was performed using the modified carbapenem inactivation method CIMTris (46). Carbapenemase genes (bla IMP , bla NDM , bla VIM , and bla GES ) were detected by PCR assay using specific primers as shown in Table 6.
The imipenem and meropenem MICs were determined using microdilution panels prepared inhouse in the absence and in the presence of 20-or 40-mg/liter efflux pump inhibitor PabN (MP Biomedicals, Santa Ana, CA), 20-or 40-mg/liter PAbN plus 300-mg/liter APB (Sigma-Aldrich Japan, Tokyo, Japan), and 300-mg/liter APB in parallel. P. aeruginosa ATCC 27853 was used as the control strain. MIC reduction (4-fold or greater) in the presence of PabN and/or APB was considered the participation of efflux pump and/or AmpC b-lactamase in the resistance to those carbapenems.
MLST analysis. MLST for carbapenem-resistant isolates and exoU1 isolates was performed according to the MLST scheme described by Curran et al. except that Tks Gflex DNA polymerase (TaKaRa Bio, Otsu, Japan) and alternative PCR primers for mutL shown in Table 6 were used (47,48). Briefly, seven housekeeping genes (acsA, aroE, guaA, mutL, nuoD, ppsA, and trpE) were amplified and sequenced. Those sequences were submitted to the PubMLST database (https://pubmlst.org/organisms/pseudomonas-aeruginosa) for the assignment of allelic numbers and STs. Those STs were assigned to clonal complexes, as defined by closely related ST groups within the single-locus variant or double-locus variant linkages (49). WGS analysis of high-level imipenem-resistant P. aeruginosa. WGS and de novo assembly of three representative carbapenem-resistant isolates were conducted in a manner like that described previously (50). Briefly, the genome was sequenced using the 150-bp paired-end method with the NovaSeq6000 platform (Illumina, San Diego, CA), and the resulting raw reads were assembled de novo into scaffolds using the SPAdes v.3.14.1. The assembled scaffolds were queried with ResFinder 4.1 and PAst 1.0 that are available from the Center for Genomic Epidemiology (http://www.genomicepidemiology.org) for the prediction of antibiotic resistance genes and in silico serotyping, respectively. A wgMLST tree was constructed using PGAdbbuilder (http://wgmlstdb.imst.nsysu.edu.tw), with the addition of the genomes of 21 P. aeruginosa strains belonging to ST235-O11 obtained from the NCBI database (Table 7). The phylogenetic tree was visualized using iTOL v5 (http://itol.embl.de/). An in-depth exploration of heavy metal resistance and virulence genes was performed manually on WGS data.
Analysis of ciprofloxacin-resistant P. aeruginosa isolates. Efflux pump-mediated fluoroquinolone resistance was investigated in ciprofloxacin-resistant isolates using efflux pump inhibitor PAbN as described above. A decrease of more than 4-fold in MIC in the presence of PAbN was considered the participation of efflux pumps in the resistance to ciprofloxacin.
Statistical analysis. The Fisher's exact test was performed to assess the association between T3SS virulotypes (exoU1/exoS2 and exoU2exoS1) and antimicrobial resistance by using R v4.0.5. Resistant and intermediate isolates were grouped as resistant for statistical analyses. P value of #0.05 were considered statistically significant. Data availability. Genome assemblies for P. aeruginosa isolates no. 2 (CA12133), 10 (CA17343), and 13 (CA19603) have been deposited in the DDBJ/EMBL/GenBank database under the GenBank assembly accession numbers GCA_015704805.1, GCA_014596255.1, and GCA_014596275.1, and the BioProject accession number PRJNA648803.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, XLSX file, 0.03 MB.