Sharing More than Friendship — Nasal Colonization with Coagulase-Positive Staphylococci (CPS) and Co-Habitation Aspects of Dogs and Their Owners

Background Since the relationship between dogs and their owners has changed, and dogs moved from being working dogs to family members in post-industrial countries, we hypothesized that zoonotic transmission of opportunistic pathogens like coagulase positive staphylococci (CPS) is likely between dogs and their owners. Methodology/Principal Findings CPS- nasal carriage, different aspects of human-to-dog relationship as well as potential interspecies transmission risk factors were investigated by offering nasal swabs and a questionnaire to dog owners (108) and their dogs (108) at a dog show in 2009. S. aureus was found in swabs of 20 (18.5%) humans and two dogs (1.8%), and spa types which correspond to well known human S. aureus lineages dominated (e.g. CC45, CC30 and CC22). Multilocus sequence typing (MLST) of the two canine strains revealed ST72 and ST2065 (single locus variant of ST34). Fifteen dogs (13.9%) and six owners (5.6%) harboured S. pseudintermedius, including one mecA-positive human isolate (MRSP). Pulsed field gel electrophoresis (PFGE) revealed that one dog/owner pair harboured indistinguishable S. pseudintermedius- isolates of ST33. Ten (48%) of the 21 S. pseudintermedius-isolates showed resistance towards more than one antimicrobial class. 88.9% of the dog owners reported to allow at least one dog into the house, 68.5% allow the dog(s) to rest on the sofa, 39.8% allow their dogs to come onto the bed, 93.5% let them lick their hands and 52.8% let them lick their face. Bivariate analysis of putative risk factors revealed that dog owners who keep more than two dogs have a significantly higher chance of being colonized with S. pseudintermedius than those who keep 1–2 dogs (p<0.05). Conclusions/Recommendations In conclusion, CPS transmission between dog owners and their dogs is possible. Further investigation regarding interspecies transmission and the diverse adaptive pathways influencing the epidemiology of CPS (including MRSA and MRSP) in different hosts is needed.


Introduction
In post-industrial countries, the socio-economic relationship between many dogs and their owners has changed dramatically during the last decades [1]. Instead of having working dogs (watch dogs, sheepdogs etc.) living in stables or kennels, people keep pets who live in the household almost as family members. In particular people who feel disconnected from society tend to substitute social contacts by pets, including supportive anthropomorphic traits (''humanisation'' of non-human beings) [2,3]. As a result of this development, transmission of microorganisms between humans and their dogs may increase [4]. Especially microorganisms that reside on the skin or mucosal surfaces of dogs might easily be transmitted to pet owners either by direct contact or by sharing the same environment in the household, e.g. sofas, beds and bath tubs.
Commensals usually consist of a wide range of different bacterial species, including coagulase-positive staphylococci (CPS) like S. aureus and S. pseudintermedius. While S. aureus is a common colonizer in humans and widely distributed among mammalian hosts in general, S. pseudintermedius seems to be predominantly associated with small animals (i.e. dogs, cats) and was found only occasionally in human specimens in the past [5,6,7,8].
Opportunistic pathogens like CPS seem to be of special interest, since reports about the general transferability of these bacteria between pets and humans have increased during recent years, including methicillin resistant variants of S. aureus (MRSA) and S. pseudintermedius (MRSP) [9,10,11,12,13,14].
MRSA have evolved to one of the major nosocomial pathogens in healthcare systems, causing an increased mortality risk for patients and enormous costs for health insurance companies worldwide [15]. In the field of veterinary medicine, nosocomial outbreaks caused by MRSA have recently gained attention, e.g. in equine and small animal clinics [16,17]. Nasally colonized veterinary personnel seem to have an impact on rising infection rates in animal patients, just like health care workers in hospitals [14].
As far as CPS are concerned, S. pseudintermedius seems to be more common among dogs and cats than S. aureus [10]. Moreover, certain strains of S. pseudintermedius have become a major threat in veterinary medicine due to their frequent methicillin-and multidrug resistant appearance [18,19].
In general, data about the colonization of dogs and their owners with CPS in the community are scarce. Hanselman et al. (2009) reported that S. aureus, S. pseudintermedius and S. schleiferi ssp. coagulans (including methicillin-resistant variants) were harboured by dogs living in households. In addition, concurrent human and animal colonization by indistinguishable S. aureus and S. pseudintermedius strains have been observed [10]. Against this backdrop, the question has arisen how changes in the human-to-dog relationship influence inter-species CPS transmission. Therefore, the aim of this study was to assess the nasal colonisation of dogs and their owners with CPS and to gain insights into potential risk factors for CPS transmission among them.

Questionnaires and ethical considerations
The study procedure and protocols were based on informed consent of the participants.
The sampling procedures in dogs had been approved by the Ethical Committee of the veterinary department of the Free University of Berlin (date: 03.03.2009).
Sampling of dog owners was approved by the Ethical Committee of the Charité, Campus Virchow-Clinic (date: 10.10.2008, Berlin).
After the dog owners had been informed about the study, they were given a questionnaire to collect demographic data on owners (e.g. gender, age and profession), number and keeping conditions of the dogs and information on the quality of contacts they shared (e.g. frequency and type of dog handling, including physical contact). Questionnaires were completed by the owners just before or after the nasal swabs were taken.

Sampling procedure and bacterial growth
Nasal swabs (MASTASWAB TM ; MAST Diagnostica GmbH, Rheinfeld Germany), were simultaneously collected from voluntarily participating pet owners (by physicians) and dogs (by veterinarians) as a convenience sample on a single day at a dog show event in Berlin in March 2009.
Until processing, all nasal samples were stored for approximately 36 h at 4uC. Swabs were plated directly on columbia agar containing 5% sheep blood and BBL TM CHROMagar TM MRSA II (Becton Dickinson GmbH, Heidelberg, Germany). All incubated plates were investigated twice, first after 18 h and a second time after 36 h of incubation at 37uC. CPS suspected colonies were picked (at least two per plate) and subcultured.

Identification and typing of isolates
Phenotypic characterisation of CPS was performed as previously reported [8,24]. Briefly, staphylococcal isolates were identified as CPS based on colony morphology, a positive catalase reaction, gram stain appearance and a positive tube coagulase reaction. All CPS were further screened for mecA by PCR and nuc for species verification in case of S. aureus [25]. Species verification of S. pseudintermedius was performed by the MboI restriction analysis method of pta as described before [26].
MLST analysis on S. pseudintermedius was performed according to a scheme published by Bannoehr et al. (2007) including pta, cpn60, tuf, 16S rRNA and agrD [30]. Allele numbers and new STs were assigned according to the curator of the MLST database V. Perreten (vincent.perreten@vbi.unibe.ch).
PFGE analysis on S. pseudintermedius was carried out using endonuclease SmaI according to Ruscher et al. 2010 [19]. PCRbased detection of the bicomponent leukotoxin Luk-I (lukS and lukF) was carried out as described previously [31]. The presence of the exfoliative toxin ExpA (formerly named EXI) was determined by PCR with the primers expA-F GCGCGTCCTTCTGATCCA-GAACT and expA-R AACGTCCCCCTTTACCTACGTGAAT (according to AB489850.1, [32] at an annealing temperature of 58uC (25 cycles) including strain IMT21652 (JN604832) as a positive control. Detection of the gene encoding ExpB was performed as described previously [33].

Statistical analysis
Statistical analysis was conducted using STATA, version 11 (StataCorp). For bivariate analysis of risk factors for human CPS colonization, Odds Ratios (OR) were calculated and potential associations checked for statistical significance using the 2-tailed Fisher exact test. A value of p,0.05 was considered significant. Results were adjusted for age and sex.

Description of study population
In total, 108 dog/owner pairs participated in this study. All owners completed the questionnaire and permitted nasal screening, resulting in 108 swabs of human and 108 of canine origin. Blank information with regard to individual questions (unanswered questions) was marked and excluded from statistical analysis of risk factors.
The mean age of the study participants was 48.5 years (range: 23-87 years, median 49 years), 75% were female and all but one were German residents. Analysing the zip codes (first three digits) of the places where the dog owners lived revealed a wide geographic distribution over Germany (data not shown). 50.9% of the participants declared to own only one dog, while 47.2% of them possessed two or more dogs (median: one dog, mean 2.2 dogs, range 1-11 dogs). Nine of the dog owners worked in a healthcare associated area (hospital, nursing home, medical practise) and three were veterinarians. On the day of sampling, only one dog that accompanied the owner was included in the study.

Behavioural aspects of human-pet cohabitation
The majority of the study subjects (96/108; 88.9%) reported to share their habitation with one or more dogs, while ten (9.3%) kept the dogs in a kennel or garden. Seventy-four dog owners (68.5%) allowed their dogs to rest on the sofa, 43 (39.8%) permitted them to come onto their beds, 101 (93.5%) let their dogs lick their hands, 57 (52.8%) allowed them to lick their face, and 35 (32.4%) reported washing their dogs in their own bath tubs (Table 1).

S. aureus nasal carriage in healthy dogs and their owners and strain characterisation
Nasal carriage of S. aureus was found in 20 (18.5%) dog owners and two dogs (1.8%). With regard to S. aureus-positive individuals, no simultaneous colonization of dog and dog owner was observed in this study. One of the three veterinarians and one of the nine participants with a known healthcare association were found to be S. aureus carriers. None of the S. aureus-isolates harboured the genes encoding for PVL or methicillin resistance.
Spa-typing of all 22 S. aureus isolates showed a broad heterogeneity and wide distribution, including two new spa variants. According to the spa database, the twenty S. aureus isolates of human origin showed the following results: Four strains were assigned to spa types which presumptively correspond to CC45, three strains to CC59 and three others to CC7. Two strains may correspond to CC30 and two more to CC22. Another six spa types t346, t495, t156, t1294, t8641 (new) and t8640 (new) remained unassigned (Table 2).
MLST-analysis of the two canine S. aureus-isolates revealed that one belonged to the genetic lineage ST72 (IMT18888), and the other to ST2065 (IMT18809), a single locus variant of ST34 (Table 2: bold letters).

S. pseudintermedius nasal carriage in healthy dogs and their owners and strain characterisation
Six owners (5.6%) and fifteen dogs (13.9%) were carriers of S. pseudintermedius. In addition, one of the six human isolates (IMT18885) proved to be mecA-positive (MRSP). None of them were veterinarians or healthcare workers.
MLST analysis of all 21 S. pseudintermedius isolates of human and canine origin revealed a heterogenic distribution of already known and 9 new sequence types (Fig. 1).  Concurrent owner and dog S. pseudintermedius-colonization was demonstrated for the isolates IMT18797 (human) and IMT18798 (canine) by PFGE pattern analysis, revealing one pulsotype for both isolates, which belongs to ST33 (Fig. 1, pair 36). Two further S. pseudintermedius-isolates had indistinguishable PFGE patterns but were obtained from unrelated canine and human individuals of pair 11 and 74, respectively ( Fig. 1; IMT18789 and IMT18828).
All 21 S. pseudintermedius isolates were positive for LukI, one yielded a positive signal for ExpA and two for ExpB (Fig. 1).

S. schleiferii ssp. coagulans nasal carriage in healthy dogs and their owners
A single strain of S. schleiferii ssp. coagulans (IMT18886) was isolated from the nostrils of one dog. This isolate was not further analysed.

Statistical analysis of potential risk factors associated with S. pseudintermedius carriage in humans
Dog owners who kept more than two dogs had a significantly higher chance of being colonized with S. pseudintermedius than those who kept 1-2 dogs (p,0.05). No other suspected risk factor proved statistically significant, but elevated chances for S. pseudintermedius colonization were observed for dog breeders (OR = 3.9), for owners who allowed dogs to lick their faces (OR = 3.5), to rest on the sofa (OR = 2), or to sleep in their bed (OR = 1.25). A summary of the risk factor analysis is provided in Table 3. The result remained unchanged when adjusted for age and sex.

Discussion
This study aimed at elucidating aspects of the relationship and behavioural patterns between dog owners and their pets with regard to nasal colonisation with CPS. In addition, comparative analysis of CPS-carriage among dogs and humans allowed insights into the proportions of CPS carriage for each group, with a special focus on CPS as zoonotic pathogens.
Substantial proportions of the participating dog owners not only shared their house (88.9%) with their pets, but also their sofa (68.5%) or their beds (39.8%). Many dogs had obviously gained a status nearly equal to that of a (human) family member, i.e. the animals were given privileges formerly exclusively meant for humans, which are behavioural aspects typical for anthropomorphism.
Zoonotic transmission of CPS has been described [11,13,34] and it appears likely that the intensive daily contact between dogs and their owners may increase the likelihood of getting colonized by interspecies-transmission of CPS for both sides.
It is noteworthy that individual refusals to answer certain questions reached up to 11%, which may have biased the study results (Tables 1 and 3, columns ''no data'' and ''valid response''). Furthermore, almost half of the participating dog owners (47.2%) stated that they keep more than just the one dog that was available for sampling on the day of sampling. While the CPS-colonisation status of the other dog(s) remains unknown, it can be assumed that all dogs of one owner are likely to live in similar co-habitation conditions.
Longitudinal studies regarding S. aureus nasal carriage in humans report approximately 20% (range 12-30%) of individuals who are persistently colonized and 50% (range 16-69%) noncarriers depending on the study protocols and methods (e.g. use of enrichment broth, kind of study population) [35]. A recent study demonstrated 19.1% nasal colonisation in 278 healthy humans in Spain [36]. Eventually, the observed 18.5% S. aureus-colonization in the human participants of our study is marginally lower but in overall agreement with formerly published results and probably reflecting the abandonment of enrichment broth.
At present, data on the colonization of healthy dogs with S. aureus gained in a community environment is very limited. In this study, 1.9% nasal S. aureuscarriage was found among those dogs. When studies on S. aureus nasal carriage in dogs are compared, care should be taken with regard to individual study backgrounds, as the kind and selection of study participants, the use of enrichment broth, sampling in a clinical or in a non-clinical environment must be expected to influence the study results. However, Rubin et al. found 13 of 167 (7.8%) nasal S. aureuscolonization among clinically healthy dogs presented for vaccination in a veterinary college [37], while Fazakerley et al. reported 2.3% (1 of 43) nasal colonization in healthy dogs [38]. In addition, we reported about S. aureus-(nasal) carriage of 5.7% in dogs on admission to an animal hospital [14].
Spa-typing of all human and canine S. aureus revealed a high diversity of spa types as previously observed for community associated S. aureus [36]. The majority of spa types of human isolates (16/20) seemed to be associated with well known human lineages (e.g. CC45, CC30, CC22) and the remaining six had not been described in any context to animal associated lineages before. Therefore, a risk factor analysis for S. aureus-carriage in humans could not be conducted.
Interestingly, one of the two canine MSSA belonged to ST72, a genetic background also displayed by the USA700 type strain (CA-MRSA) and further MSSA and MRSA of human origin (http:// saureus.mlst.net). The second canine strain is a single nucleotide variant of ST34, an ST that had also previously been associated with MRSA of human origin [39]. In conclusion, both canine S. aureus strains isolated in this study may have once originated from a human host. These strains could belong to extended host spectrum genotypes (EHSGs), which were previously reported for MRSA strains of equine origin [40]. Consequently, dogs can be colonized with EHSG S. aureus strains (including MRSA) and represent a potential source of (re-) infection for humans [11].
Surprisingly, 5.5% of dog owners were found to be positive for S. pseudintermedius nasal carriage, including MRSP in one case. A recent study by Paul and co-workers found a nasal colonisation in 3.9% of 128 veterinarians attending a conference for animal dermatologists [41]. All strains were proven to be MRSP and their nasal carriage persisted for at least one month. While human infections with S. pseudintermedius have occasionally been reported in the past, an increased awareness of these CPS in human medicine ensued very recently because of the multi-drug resistant phenotype which is frequently associated with MRSP [19,41,42]. We found one dog/owner pair contemporarily colonized with one particular S. pseudintermedius strain, as has previously been reported for MRSP in clinical cases [13,43]. Moreover, another person and a dog who had no relationship according to their pair ID's were colonized with strains displaying an indistinguishable PFGEpattern, which emphasizes the transferability of S. pseudintermedius.
While S. pseudintermedius is considered as a common colonizer in the first line [7], the antibiotic resistance phenotypes demonstrated that ten of the 21 isolates showed resistance towards four or more antimicrobials. Among these isolates was strain IMT18884, which showed resistance towards 7 agents (Fig. 1).
It is striking that all six dog owners who carried S. pseudintermedius kept their dogs in the house, and had three or more dogs (one person did not answer the respective question). Statistical analysis supports the result by revealing that keeping two or more dogs is significantly associated with S. pseudintermedius colonization in humans (p = 0,0006). Other suspected risk factors were not statistically significantly associated with S. pseudintermedius carriage in our analysis. Interpretation of statistical results needs to take the small number of cases into account. As a matter of fact, more dogs will result in more saliva as well as more dog scurf and hair harbouring commensals (including CPS) in the household environment shared by both, dogs and owners. In addition, domestic contamination due to S. pseudintermedius has been reported for households of MRSP-infected pets before, and it was assumed that dust particles may play a role in this finding [34].
Furthermore, it was thought that certain MRSP lineages (e.g. ST71) may have a greater ability to adapt to the human host [41] than MSSP. However, it is important to note that the MRSP and MSSP strains from human origin detected in this study comprised six different STs (Fig. 1). Regarding the exfoliative toxins in MSSP reported on here, one human isolate yielded a positive PCR result for ExpA, while two canine strains were found positive for ExpB, indicating that these virulence factors may be more frequently associated with cases of superficial dermatitis in dogs [32,33].
The single MRSP (IMT18885) isolated during this study was of human origin, showed an ST41 background and was susceptible to most of the tested antimicrobials. In addition, IMT18885 had an oxacillin MIC of 2 mg/l and expressed a comparatively low MIC (4/2 mg/l) towards ampicillin-sulbactam (Fig. 1). Interestingly, this ST was originally described for a canine isolate (mecA negative) in the USA [30]. It seems likely that acquisition events of mecA harbouring staphylococcal chromosomal cassettes will occur in accessible strains over time.
In conclusion, exposure of humans to pet dogs seems to be associated with the possibility of getting colonized by S. pseudintermedius, regardless whether these bacteria are methicillin-resistant or not. Methicillin-resistant CPS attract more scientific interest because of the clinical relevance of MRS, but according to our findings the inter-species transferability does not seem to be necessarily linked with methicillin-resistance at all. However, it remains unknown whether S. pseudintermedius colonization in humans is transient or permanent. Further investigations regarding inter-species transmission and the diverse adaptive pathways influencing the epidemiology of CPS (including MRSA and MRSP) in the community are needed. Especially the close relationship between pets and their owners underline the importance of interdisciplinary research including human and veterinary microbiologists as well as epidemiologists as demonstrated by our study results.