First Microbiological and Molecular Identification of Rhodococcus equi in Feces of Nondiarrheic Cats

Rhodococcus equi is responsible for infections in multiple-host animals. In humans, the prevalence of rhodococcus has increased worldwide and represents an emergent risk. R. equi is a soil-borne opportunistic bacterium isolated from feces of a wide variety of domestic species, except cats; thus, there is no known potential risk of its transmission from humans. Here, the mono- and cooccurrence of Rhodococcus equi and other bacteria and selected virulence markers were investigated in feces of nondiarrheic cats from urban (n=100) and rural (n=100) areas. Seven (7/200=3.5%) R. equi isolates were recovered in ceftazidime, novobiocin, and cycloheximide (CAZ-NB) selective media, exclusively of cats from three distinct farms (p=0.01), and these cats had a history of contact with horses and their environment (p=0.0002). None of the R. equi isolates harbored hosted-adapted plasmid types associated with virulence (pVAPA, pVAPB, and pVAPN). One hundred seventy-five E. coli isolates were identified, and 23 atypical enteropathogenic E. coli (aEPEC), 1 STEC (Shiga-toxin producing E. coli), and 1 EAEC (enteroaggregative E. coli) were detected. Eighty-six C. perfringens type A isolates were identified, and beta-2 and enterotoxin were detected in 21 and 1 isolates, respectively. Five C. difficile isolates were identified, one of which was toxigenic and ribotype 106. The main cooccurring isolates in cats from urban areas were E. coli and C. perfringens A (26/100=26%), E. coli and C. perfringens type A cpb2+ (8/100=8%), and aEPEC (eae+/escN+) and C. perfringens type A (5/100=5%). In cats from farms, the main cooccurring isolates were E. coli and C. perfringens type A (21/100=21%), E. coli and C. perfringens type A cpb2+ 8/100=8%), and E. coli and R. equi (4/100=4%). We identified, for the first time, R. equi in nondiarrheic cats, a finding that represents a public health issue because rhodococcus has been reported in both immunosuppressed and immunocompetent humans, particularly people living with HIV/AIDS.


Introduction
The population of cats and dogs has increased worldwide, and close contact with their owners may favor the transmission of pathogens, including those with zoonotic potential, which is a public health concern [1].
Enteric pathogens are a complex group of agents represented mainly by a number of bacterial, viral, protozoal, and 2 BioMed Research International parasitic organisms related to infections in humans, companion animals, livestock, and wildlife [2]. These pathogens may be transmitted from cats to humans by ingestion of food, water, vegetables, and fruits contaminated with feces of their pets [1,3].
The majority of comprehensive studies with enteric pathogens from cats involve diarrheic animals, focused mainly on one pathogen [4,5], although diarrhea apparently is caused by a combined action of enteric pathogens [6]. Nevertheless, few studies have evaluated the complexity of the microbial population of feces from nondiarrheic cats [7,8], as well as main virulence markers of enteric pathogens [4]. In addition, nondiarrheic cats are recognized as reservoirs of enteric pathogens [1,4], including zoonotic agents, because these pet animals have hunting behavior and self-cleaning habits and establish broad territoriality and have close proximity with their owners [9].
Rhodococcus equi (R. equi) is a well-recognized bacterium that is opportunistic in nature. R. equi infection is characterized by diverse pyogranulomatous clinical disorders in humans, domestic animals, and wildlife [15]. It is a soilborne organism that is widespread in the environment and is eliminated through the feces of animals [16]. In humans, the prevalence of rhodococcus has increased globally, particularly among people living with HIV/AIDS [17], and disease poses an emergent risk in the coming years [18]. This pathogen has been isolated from feces of a wide variety of animal species, including equines, bovines, small ruminants, dogs, pigs, and some wildlife. Nevertheless, to date, R. equi has not been isolated from the feces of cats [16], suggesting that there is a lack of a potential risk of transmission of this pathogen from cats to humans. The pathogenicity of R. equi is attributed to the presence of plasmid-encoded virulenceassociated proteins (VAPs) [19]; however, the virulence plasmid profile of R. equi isolated from cats is poorly understood or neglected [20,21].
In this scenario, the aim of the present study was to investigate the mono-and cooccurrence of R. equi, E. coli (DEC pathotypes), Clostridium perfringens, Clostridium difficile, and Salmonella spp. and selected virulence markers of R. equi, E. coli, and Clostridium species in feces of nondiarrheic cats from urban and rural areas.

Cats.
A cross-sectional study was carried out using cats from a city and 13 farms located in the central region of the state of São Paulo, Brazil. The cats had different pedigrees, genders, and ages. The cats from rural areas had contact with bovines, equines, and small ruminants and had free access to the environment. All of the farms were similar in terms of the general management of the livestock, animal hygiene, and facilities. The exclusion criteria were as follows: (1) cats with clinical signs of diarrhea or other clinical manifestations and (2) cats undergoing antimicrobial therapy or those with history of antimicrobial use in the last 10 days.

Fecal Samples.
Fresh fecal material was collected directly from the rectums of cats using cotton swabs and immediately introduced in Stuart's media. The samples were transported to the laboratory in plastic containers under refrigeration (4-8 ∘ C). At the laboratory, the samples were diluted in 5 mL of sterilized distilled water, and aliquots were stocked frozen (-20 ∘ C) for further microbiological and molecular analyses. Fecal samples were collected from each cat only once.

Microbiological
Identification of E. coli, Salmonella spp., and R. equi. For E. coli isolation, fecal samples were plated on MacConkey agar and defibrinated sheep blood agar (5%), incubated in aerobic conditions at 37 ∘ C, and evaluated at 24, 48, and 72 hours [22]. For Salmonella isolation, fecal samples were inoculated into selective tubes containing tetrathionate broth and then incubated overnight in aerobic conditions at 37 ∘ C. After incubation, samples were cultured on Salmonella/Shigella (SS) agar plates and incubated aerobically at 37 ∘ C for 24-48 hours [22]. Colonies with black central areas in SS agar compatible with Salmonella were subjected to a commercial agglutination test using a polyvalent anti-Salmonella serum (Probac6, Sao Paulo, Brazil). For R. equi isolation, fecal samples were cultured in ceftazidime, novobiocin, and cycloheximide (CAZ-NB) selective media incubated in aerobic conditions at 37 ∘ C and evaluated at 24, 48, and 72 hours [23,24]. All microorganisms were classified using conventional phenotypic tests [22].

Molecular Identification and Virulence of R. equi.
Rhodococcus equi isolates were detected by using a polymerase chain reaction (PCR) [29]. Plasmid DNA from R. equi was obtained by the alkaline lysis method [29]. The extracted genetic material was analyzed by digestion with restriction enzymes (HindIII, BamHI, EcoRI, and EcoT22). Then, the plasmid samples were separated by electrophoresis and examined under UV light. PCR for the genes vapA, vapB, and vapN was performed as previously described [19,29].

Clostridium difficile Isolation, PCR, A/B Toxin Detection, and Ribotyping.
To isolate C. difficile spores, equal volumes of the diluted stool samples and 96% ethanol (v/v) were mixed; after incubation for 30 minutes at room temperature, 50 L aliquots were inoculated on plates containing cycloserinecefoxitin fructose agar supplemented with 7% horse blood and 0.1% sodium taurocholate (Sigma-Aldrich Co., St. Louis, MO, USA). After anaerobic incubation at 37 ∘ C for 96 hours, all colonies with suggestive morphology were subjected to a previously described multiplex-PCR for a housekeeping gene (tpi), toxins A (tcdA) and B (tcdB), and a binary toxin gene (cdtB) [30]. All toxigenic C. difficile isolates were submitted to PCR ribotyping. Intergenic spacer regions were amplified using Bidet primers as previously described [31]. Amplification products were separated by electrophoresis in 3% agarose gel (Bio-Rad, California, USA) for 5 hours at 2.5 V/cm, and the gel was analyzed with BioNumerics 7.00 (Applied Maths, Belgium). PCR ribotypes were designated by international Cardiff nomenclature.

Statistical Analysis.
Statistical analysis was performed using R (version 3.2.3) software. The chi-square test with Yates correction (or Fisher's exact test) was used to compare the frequencies of the different enteric pathogens between cats from urban and rural areas and to compare the presence of different virulence markers among the two groups of animals. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for positive results by area. The statistical significance level was set at 0.05.

E. coli.
We identified 175 (175/200=87.5%) E. coli isolates from cats from urban (93 isolates) and rural (82 isolates) areas (OR=2.91). The escN gene was detected in 13.0% (23/175) of the isolates. All 23 isolates were negative for the bfpA and bfpB genes and were thus classified as aEPEC. In addition, one STEC (harboring stx2 gene) and one atypical EAEC (aatA + /aggR − ) were identified. The risk of positive E. coli isolation in urban cats was 2.91 (OR=2.91) times greater than cats from rural areas.
3.4. R. equi. Seven (7/200=3.5%) R. equi isolates were isolated from feces in CAZ-NB selective media exclusively among cats from farms ( Table 1). All these cats had contact with horses and their breeding environment, showing a statistical association between R. equi isolation and rural areas (p=0.0002). None of R. equi isolates harbored three hostedadapted plasmid types associated with virulence (pVAPA, pVAPB, and pVAPN) and were considered plasmidless or avirulent isolates.

Discussion
The increase in the number of dogs and cats and their importance to the physical and mental health of their owners   has been highlighted globally. The close relationship and direct contact between these animals and their owners are essential to understanding the epidemiology and virulence of different pathogens that affect pet animals and to the implementation of control strategies to avoid transmission of diseases from pets to humans [1]. In particular, over 40 million cats share households in Brazil. In this scenario, we investigated the mono-and cooccurrence of different bacteria and selected virulence markers in feces from nondiarrheic cats from Brazil. The presence of E. coli, C. perfringens, C. difficile, and R. equi harboring different virulent markers and/or toxins from feces of nondiarrheic cats from urban and/or rural areas highlights the risk of transmission of these pathogens from cats to humans due to the close contact of owners and their cats. In addition, we report the first microbiological and molecular identification of R. equi from the feces of cats, where the virulence plasmid profile was assessed.
The majority of comprehensive studies that have investigated the prevalence of enteric pathogens among cats have involved diarrheic animals, particularly focused on only one pathogen [4,5]; however, diarrhea is apparently caused by the combined action of agents from the enteric microbiota of cats [3,6]. In addition, enteric disorders are influenced by diverse factors, such as age, pedigree, history of vaccination for some pathogens, and nutritional, environmental, and management conditions [6]. Conversely, few studies have evaluated the complexity of the microbial population from feces of nondiarrheic cats [7,8]. In the current study, E. coli (175/200=87.5%) and C. perfringens type A (86/200=43%) were the main enterobacteria isolated from nondiarrheic cats, whereas R. equi (7/200=3.5%) and C. difficile (5/200=2.5%) were isolated in lower frequencies, and no Salmonella species were identified. This result indicates that E. coli and C. perfringens type A are major agents of bacterial origin found in the enteric microbiota of nondiarrheic cats [4,7,8].
The most frequent cooccurring microorganisms (and selected virulence markers) in the nondiarrheic cats sampled were aEPEC and C. perfringens (type A and cpb2 + ). This result is not surprising because enterobacteria and clostridia are major groups of bacteria inhabitants of the enteric tract of domestic animals [22]. In addition, this finding reinforces the complexity of microorganisms present in feces of nondiarrheic cats, including zoonotic pathogens. This finding poses a public health concern due to the potential risk of transmission of fecal pathogens from cats to humans by ingestion of contaminated food, water, vegetables, and fruits [1,3], apparently favored by some behaviors of cats, for example, hunting and self-cleaning habits, establishment of broad territoriality, and close proximity with their owners [9].
The pathogenicity of enteric and extraenteric E. coli involved in human, livestock, pets, and wildlife infections is characterized by a diversity of virulence factors, particularly exotoxin production, adherence to epithelial cells, iron uptake, and serum resistance [36]. In the current study, 175 E. coli strains were isolated from nondiarrheic cats, with an OR=2.91 in urban (93 isolates) and rural (82 isolates) areas. These strains were subjected to diagnostics for selected virulence markers of typical EPEC (eae, bfpA, bfpB), atypical EPEC (escN), EHEC (stx1, stx2), ETEC (elt, est), and EAEC (aatA, aagR) pathotypes related to human infections. Of these, the genes eae, escN, stx2, and aatA were detected in the isolates, a finding that indicates the potential transmission of DEC pathotypes to humans because they may be harbored by domestic cats.
tEPEC and aEPEC are epidemiologically related to diarrhea in humans [27,28,36]. These strains contain a pathogenicity island (PAI) named locus enterocyte effacement (LEE). In our study, the eae and escN genes used for identification of the LEE region in the E. coli isolates were concomitantly identified in 13.0% of our nondiarrheic cats, indicating that these animals act as an important reservoir of aEPEC. In France, E. coli isolates harboring the eae gene were reported in 5% of feces from cats with and without diarrhea [37]. Likewise, in another study performed in Brazil, this gene was observed in isolates obtained from 2.5% of the nondiarrheic cats studied [5]. Nevertheless, the presence of aEPEC isolates in our nondiarrheic cats indicates a public health concern due to the association of this pathotype of DEC with enteric diseases in humans [36].
STEC serotype O157:H7 has become important as a cause of foodborne and waterborne disease in humans that is eliminated in feces of animals, although the impact of cats in the transmission of STEC to humans remains unclear [5]. This potentially life-threatening illness produces high morbidity and mortality rates, particularly in children, and causes acute to chronic thrombocytopenia, hemolytic anemia, and hemorrhagic colitis [38]. stx1 and stx2 (formerly verocytotoxin or Shiga toxin) are the main virulence factors of STEC. The gene stx2 was detected at low frequency among cats, in 0.6% of urban areas animals. Likewise, virulent stxproducing E. coli was reported in a cat from Argentina [39]. The detection of E. coli harboring stx genes without the LEE region in feces from our nondiarrheic cats is circumstantial evidence that these pet animals may represent potential carriers of STEC, which may be transmitted to their owners.
The genes aatA and aagR constitute virulence markers of the EAEC pathotype, which is related to persistent diarrhea in children, travelers, and human patients living with HIV/AIDS [40]. Nevertheless, the impact of the EAEC pathotype in companion animals remains unclear [41]. Indeed, only one E. coli strain was positive for the aatA gene, although this isolate lacks the aggR gene, which is responsible for encoding a member of the AraC/XylS family of bacterial transcriptional activators. None of the E. coli strains isolated from our 200 nondiarrheic cats were positive for the elt and est genes, which are considered virulence markers of the ETEC pathotype [36].
Clostridia are well-known Gram-positive spore-forming bacilli characterized by anaerobic requirements and the production of potent hemolytic, neurotoxic, and histotoxic entero-and exotoxins that affect humans and animals [22]. In companion animals, this bacterium causes hemorrhagic and necrotic enteritis. Clostridium species are widely distributed in soil and water, although many pathogenic species are normal inhabitants of the enteric microflora of humans and animals. In particular, C. perfringens stands out due to its capacity for toxin production. This bacterium is classified into five types (A to E) according to the production of four major toxins: alpha, beta, epsilon, and iota [14]. In addition to these major toxins, it can produce other virulence factors that are associated with the pathogenesis of some diseases in humans and animals, such as enterotoxin, which is responsible for foodborne disease in humans, and NetF, which is responsible for bloody diarrhea in dogs [35].
Among our 200 nondiarrheic cats, eighty-six (43%) C. perfringens type A isolates were identified, without a statistical association with urban or rural areas. In a cross-sectional study of 100 cats admitted to a municipal shelter in the USA, C. perfringens was identified in 42% and 50% of animals with and without diarrhea, respectively [7]. In the UK, a prevalence of 56.6% of C. perfringens was reported among 11151 samples from diarrheic cats submitted to a reference veterinary laboratory [8].
In the current study, the beta-2 encoding gene (cpb2) was the most prevalent and was detected in 24% of isolates from 10 and 11 cats from urban and rural areas, respectively. In previous studies, the prevalence of C. perfringens cpb 2 was between 10% and 30% in dogs and cats [14]. Although some authors suggest that beta-2 toxin is relevant for the pathogenesis of C. perfringens-induced infections in some species, its real role is still poorly understood [42]. One of our isolates (1.1%) from an urban cat was also positive for the enterotoxin-encoding gene (cpe). Although C. perfringens cpe+ is responsible for foodborne disease and nosocomial diarrhea in humans, the role of domestic animals as reservoirs of these strains is unclear (Uzal et al., 2014) [42]. C. perfringens cpe+ is also associated with bloody diarrhea in dogs when also positive for the recently described NetF-encoding gene (netF) [43]. All C. perfringens isolates were negative for NetF, and, thus far, these strains have been described only in dogs and foals with gastrointestinal disorders [35,43].
C. difficile is a highly pathogenic species in humans related to gastrointestinal manifestations secondary to nosocomial infections and prolonged use of antimicrobials. It is also linked to diarrhea in dogs, but its role in cats remains unclear [43]. Toxigenic C. difficile strains produce A and/or B toxins encoded by the genes tcdA and tcdB, respectively [44], as well as binary toxin (CDT), which is related to some ribotypes, including pet animals [45]. Five (2.5%) isolates of C. difficile were identified in the sampled cats, but only one was toxigenic (A+B+C+D). Notably, this strain was ribotype 106, a common ribotype related to C. difficile in humans worldwide [46][47][48]. The role of cats as reservoirs for toxigenic C. difficile is poorly understood, as opposed previous studies have suggested that dogs may be a potential source of this bacterium for humans [49].
Salmonellosis is one of the most common zoonosis globally. Salmonella enterica possesses different pathogenic serotypes in humans and domestic animals. This opportunistic bacterium may be eliminated in feces of domestic animals with and without enteric signs [50]. No Salmonella species were isolated from our 200 nondiarrheic cats. Other studies have reported a low prevalence [7,8] or absence [6] of isolation of Salmonella species in feces of diarrheic and nondiarrheic cats. Nevertheless, despite the lack of isolation of Salmonella in the sampled cats, due to the high pathogenicity of this agent in animals and humans as well as the risk of fecal elimination among apparent healthy or nondiarrheic animals, it is recommended that this pathogen be considered in the diagnosis of enteric agents from pet animals.
R. equi is a well-recognized Gram-positive bacterium that is opportunistic in nature, is characterized by various pyogranulomatous clinical disorders, and is able to infect humans, livestock, companion animals, and wildlife [15]. In humans, clinical rhodococcus has increased worldwide, and the majority of cases occur among immunosuppressed patients, especially people living with HIV/AIDS [20]. Nevertheless, clinical disease in immunocompetent humans has also been reported [17,29]. It is a soil-borne organism that is widespread in the environment and eliminated through the feces of different animal species. However, to date, R. equi has not been isolated from the feces of cats [16], and the risk of transmission of this pathogen from cats to humans remains unclear [16]. Curiously, fecal samples from seven of our 200 cats showed isolation of R. equi in selective culture media, which was confirmed by molecular methods. The successful isolation of R. equi may be attributed to the efficacy of ceftazidime, novobiocin, and cycloheximide, which are included in CAZ-NB media to inhibit the growth of other microflora found in the feces of cats because these selective media have been indicated to isolate pathogens in contaminated material from foals, pigs, and farm environments [23,24]. In addition, isolation of R. equi exclusively from cats from rural areas (p=0.0002) may be due to contact with livestock species because these pathogens are eliminated in feces (particularly equines and bovines) and are widely distributed in farm environments [16,17]. To our knowledge, this report describes, for the first time, the identification of R. equi in the feces of nondiarrheic cats, indicating that this pet animal, particularly cats from rural areas, may be a reservoir of this pathogen for human and other animal species.
The pathogenicity of R. equi is characterized by plasmidgoverned infectivity attributed to the presence of VAPs. Three host-adapted virulence plasmid types are recognized: pVAPA, pVAPB, and pVAPN [19]. The pVAPA type is harbored by isolates that cause typical life-threatening pyogranulomatous bronchopneumonia and ulcerative colitis in foals [17], as opposed to pVAPB type strains that are mainly recovered from the lymph nodes of pigs and wild boars [24] and human R. equi-induced infections, especially among immunocompromised patients [20]. The novel pVAPN, which was recently described, has been recovered from the lymph nodes of slaughtered cattle and humans, including people living with HIV/AIDS [51]. In addition, R. equi lacks the vapA, vapB, and vapN genes and is mainly isolated in the feces of livestock and farm environments. These strains are called "plasmidless" or "avirulent" [17,19]. Conversely, a restrict number of studies have focused on investigating the virulence plasmid pattern of R. equi isolated from cats [20,21]. In this context, characterization of the virulence of R. equi strains of nine cats and nine dogs presenting diverse clinical manifestations from Canada, USA, South Africa, Brazil, and New Zealand revealed that five feline isolates and one canine strain harbored pVAPA (formerly VapA), and the 12 remaining isolates were considered avirulent (plasmidless) [20]. A report of cutaneous lesions caused by R. equi in a 2-year-old male domestic cat from Brazil revealed that the isolate carried the pVAPA type (87-kb type I variant) [21]. In the current study, none of the R. equi isolates harbored hosted-adapted plasmid types associated with virulence (pVAPA, pVAPB, pVAPN). In addition to these seven cat isolates currently classified as plasmidless or avirulent, they hypothetically might harbor unknown virulence plasmid types [51]. Surveillance studies concerning the hosted-adapted plasmid pattern of R. equi strains isolated from domestic animals from different geographic areas and countries are important for investigating the virulence and transmission risks of this pathogen from animals to humans, particularly because clinical rhodococcus has been reported in both immunocompromised and immunocompetent human patients, including disease caused by plasmidless R. equi strains [20,52].

Conclusions
The presence of R. equi, C. perfringens, C. difficile, and diarrheagenic E. coli harboring virulent markers in feces of nondiarrheic cats from urban and/or rural areas highlights the risk of transmission of these zoonotic pathogens from cats to humans due to close contact of owners and their cats. Moreover, we report the first microbiological and molecular identification of R. equi from the feces of nondiarrheic cats, providing evidence that this pet animal, especially cats from rural areas, constitutes a reservoir of this pathogen for human and other animal species.

Data Availability
The data used to support the findings of this study are included within the article.

Conflicts of Interest
The authors declare that they have no competing interests.