Misuse of Antimicrobials and Selection of Methicillin-Resistant Staphylococcus pseudintermedius Strains in Breeding Kennels: Genetic Characterization of Bacteria After a Two-year Interval


Author’s address (for correspondence): A Rota, Dipartimento di Patologia Animale, Università di Torino, via Leonardo da Vinci 44, 10090 Grugliasco (TO) Italy. E-mail: ada.rota@unito.it


Methicillin-resistant Staphylococcus pseudintermedius (MRSP) strains have been isolated from dogs with increasing frequency; prolonged or excessive use of antimicrobials is associated with the selection of MRSP, and misuse of antimicrobials is frequent in breeding kennels. This study was carried out in two breeding kennels (A and B) in which we had isolated MRSP in 2008: the aim was to assess colonization of previously positive bitches and of other bitches sharing the same environment and to assess the genetic profile of both the old and the new strains [spa typing, staphylococcal cassette chromosome mec (SCCmec) typing and Pulsed-field Gel Electrophoresis (PFGE)]. Six animals from Kennel A (two from 2008) and eight from Kennel B (one from 2008) were tested: 16 MRSP strains were isolated only from bitches housed in Kennel B. Old and new isolates were mecA positive, resulted spa type t02 and carried SSCmec II–III. PGFE showed that all isolates were related and belonged to the main clone lineage dominating in Europe, ST71-J-t02-II–III. Kennels A and B differ in the use of antimicrobials, which has been reduced over time in Kennel A, while has remained excessive in Kennel B, where many agents belonging to different classes (third-generation cephalosporins, fluoroquinolones, macrolids) are administered to dogs, without veterinary supervision, especially around parturition. Misuse of antimicrobials is the key factor for the selection of MRSP strains in healthy dogs and for their persistence over time. Dog breeders should be aware that infections caused by multiresistant bacteria have very limited therapeutical options and represent a huge challenge for animal health.


Dogs are natural hosts of Staphylococcus pseudintermedius; S. pseudintermedius has recently been shown to be the species of the S. intermedius group that predominantly colonizes dogs and cats (Perreten et al. 2010). It is part of the normal cutaneous flora, although it is also considered an opportunistic pathogen and is frequently associated with secondary skin and ear infections (Fitzgerald 2009).

Methicillin-resistant S. pseudintermedius (MRSP) strains have been isolated with increasing frequency both from healthy dogs (Vengust et al. 2006; Griffeth et al. 2008; Rota et al. 2011) and from dogs showing pyoderma, otitis externa, wound infections or urinary tract infections (Loeffler et al. 2007; Rubin and Gaunt 2011; Van Duijkeren et al. 2011).

Resistance to methicillin in S. pseudintermedius, as well as in Staphylococcus aureus, is because of the presence of the mecA gene, which encodes the altered penicillin-binding protein 2a (PBP2a) that shows reduced affinity for all beta-lactam antimicrobials. The mecA gene is carried on the SCCmec (Lin and Davies 2007).

Prolonged or excessive use of antimicrobials is associated with the selection of MRSP strains. An investigation on dogs admitted to a small animal hospital revealed that the main factors associated with MRSP carriage were former hospitalization and antibiotic treatment during the previous 6 months (Nienhoff et al. 2011). We also found that the frequency of isolation of MRSP strains from healthy breeding bitches and dead septicaemic puppies was significantly higher in kennels where antimicrobials were misused (Rota et al. 2011).

Staphylococcus pseudintermedius rarely is an agent of human infections, and most often infections are associated with canine bite wound or immunosuppression (Frank et al. 2009).

People working or living with dogs are more likely to be colonized with S. pseudintermedius in their nasal cavity, also with MRSP strains (Sasaki et al. 2007; Hanselman et al. 2009; Youn et al. 2011); transmission to humans or to contact pets is more likely when dogs have clinical signs of infection (Van Duijkeren et al. 2011; Laarhoven et al. 2011). Although a recent longitudinal study showed that transmission of MRSP to humans was rare and sporadic, suggesting contamination instead of colonization (Laarhoven et al. 2011), the zoonotic potential and therapeutic challenge of MRSP should be taken into consideration (Epstein et al. 2009; Yoo et al. 2010; Youn et al. 2010; Paul et al. 2011).

Genetic characterization of various MRSP isolates in Europe showed that all of them belong to the same multiresistant clonal lineage (Perreten et al. 2010), suggesting that MRSP strains spread with relatively stable clones. Also a reported human infection was associated with the predominant clone (Stegmann et al. 2010).

So far, there have been no studies concerning the persistence and the pattern of MRSP colonization of dogs living in confined environments or the possible horizontal transfer from dog to dog. Loeffler et al. (2010) showed that methicillin-resistant S. aureus (MRSA) does not readily transmit between healthy dogs housed in shared kennels and that MRSA carriage was only transient in a regularly cleaned and disinfected environment. However, persistence and transmission of a Staphylococcus species more typical of dogs, like S. pseudintermedius, could be different.

This study was carried out in the same breeding kennels in which we had isolated MRSP strains 2 years before. The main objective was to assess colonization of previously positive bitches and of other bitches sharing the same environment. Secondly, we wanted to assess the genetic profile of both the old and the new strains, to verify the correlation with the main MRSP strains isolated in Europe.

Materials and Methods


Two breeding kennels (A and B) that were positive for MRSP in 2008 were included in this study. From 2008 to 2010, the average number of animals in Kennel A was 10 females and three males belonging to two different breeds (Boxer and Labrador), and there were 4–6 whelpings each year. The dogs lived in small groups outdoors during the day and were housed in pairs during the night. The kennels had concrete floor and were cleaned daily with ordinary cleaning chemical substances and disinfected monthly with bleach. Equipment such as food bowls was washed between use, but was shared among two or more animals.

In Kennel B, the average number of animals was 11 females and FOUR males, belonging to two different breeds (Boxer and English Bulldog), and there were 6–10 whelpings each year. The dogs were housed in single kennels and were grouped in outdoor pens during the day. The kennels had tiled floor and were cleaned and disinfected as in Kennel A; equipment and food bowls were treated as in Kennel A.

In Kennel A, no antimicrobial drugs had been administered without veterinary prescription since 2008, while in Kennel B, many antimicrobial agents of different classes (III generation cephalosporins, fluoroquinolones, macrolids) had been used without veterinary supervision, especially in the peripartum period.

‘Old’ Methicillin-resistant Staphylococcus pseudintermedius strains

The MRSP strains (N = 9) that had been isolated from the milk of six breeding bitches (four belonging to Kennel A and two to Kennel B) and from the organs of three dead puppies (all belonging to different litters of Kennel B) in 2008 were included in this study. The old strains were stored at −80°C degrees and revitalized by subculturing on nutritive broth and solid nutritive medium.


Only three of the six previously MRSP-positive bitches (two from Kennel A and one from Kennel B) could be included in the study because the other bitches no longer lived in the kennels. Four other bitches were included from Kennel A and seven from Kennel B, selecting the three lactating ones and the remaining ones at random, for a total of six bitches from Kennel A and eight from Kennel B.

All the animals were healthy and free of skin diseases, fresh wounds and ear infections at the moment of swab collection. Furthermore, they had not been treated with systemic antimicrobials during the previous 10 days.


Samples were collected from nostrils, mouth, axillary skin, nipples, vagina and milk (if the bitch was lactating), using sterile swabs (Copan Innovation®, Brescia, Italy). For nasal and oral mucosa testing, a sterile swab was introduced into both nostrils and in the oral cavity, respectively, and rotated for 5 s. For axillary skin and nipples collection, the swab was rotated for 30 s, and for vaginal collection, the swab was introduced into the cranial vagina. For milk culture, a drop was collected from each of the caudal mammary glands after local disinfection. Swabs were immediately placed inside the Amies transport medium provided with the swab, stored at 4°C and processed within 24 h. A total number of 73 samples were collected and cultured, 70 from body areas and three from milk.

Identification of S. pseudintermedius

Bacteriological cultures and primary identification of Staphylococcus sp. were performed using standard methods for isolation and identification of bacteria as previously described (Rota et al. 2011). Coagulase-positive strains were revealed by coagulase reaction on rabbit plasma (Istituto Zooprofilattico delle Venezie, Legnaro, Italy). API 20 Staph and API Staph ID 32 (BioMérieux, Marcy l’Etoile, France) were also used to identify bacterial Genus.

Definitive S. pseudintermedius identification was performed by PCR methods according to Bannoehr et al. (2009), modified by Rota et al. (2011).

Antimicrobial susceptibility and oxacillin resistance

Staphylococcus pseudintermedius strains were tested with some antimicrobials belonging to the principal pharmacological groups: β-lactams, potentiated sulphonamides, aminoglycosides, tetracycline, macrolides and fluoroquinolones. Criteria for the Kirby–Bauer disc diffusion method, for zone size limit interpretation and for quality control were drawn from Clinical Laboratory Standards Institute (2008, 2009). Oxacillin resistance was assessed on five randomly chosen colonies plated onto a selective medium, oxacillin resistance screening agar base (Oxoid, Basingstoke, UK) supplemented with oxacillin (4 μg/ml) (Clinical Laboratory Standards Institute, 2008).

All the oxacillin-resistant S. pseudintermedius strains were tested for the presence of the mecA gene by PCR (Louie et al. 2002; modified by Rota et al. 2011).

Spa typing and staphylococcal cassette chromosome mec typing

Spa typing was performed as previously described by Moodley et al. (2009). The repeats and spa types were defined using the scheme adopted by Moodley et al. (2009) and using sequence signatures 5′-AATAATTCA and 3′-GACAAGCG defined by Ruscher et al. (2010).

Staphylococcal cassette chromosome mec types I – V were determined using a multiplex real-time PCR assay (Valvatne et al. 2009). To differentiate between SCCmec III lacking SCC-Hg, SCCmec II–III and SCCmec VII-241, a conventional PCR was applied (Perreten et al. 2010).

Pulsed-field Gel Electrophoresis (PFGE)

Preparation of chromosomal DNA and plugs (Agarose Prep; Amersham Biosciences, Uppsala, Sweden) was carried out according to the Harmony protocol (Murchan et al. 2003). DNA fragmentation was performed using 20 U/μl SmaI (Fermentas, Vilnius, Lithuania), and separation of fragments was performed in a CHEF-DR II system (BIO-Rad Lab., Hercules, CA, USA) with a 1.2% agarose gel (Agarose NA; GE Healthcare, Uppsala, Sweden). The gel was run for 24 h at 5.6 V/cm with pulse time ramping from 1 to 5 s and at a temperature of 14°C. S. aureus NCTC 8325 was used as control.

Fragment patterns were analysed in BioNumerics® version 6.5 (Applied maths, Gent, Belgium) using DICE coefficient and UPGMA cluster analysis, and fragments between 9 and 117 kb were included in the analysis. Position tolerance and optimization were set at 1.2% and 1%, respectively. The following criteria were used for determining the relation between the different strains (Tenover et al. 1995): identical PFGE pattern, the same strain; 1 band difference, very closely related or possibly the same strain; 2–3 band difference, closely related; 4–6 band difference, possibly related; and over seven band difference, different strains.


All coagulase-positive staphylococci were identified as S. pseudintermedius. S. pseudintermedius was isolated from 42 of the 70 swabs that were collected from body areas and from the three milk cultures (Table 1).

Table 1.   Localization of Staphylococcus pseudintermedius isolates
KennelMouthNaresAxillary skinNipplesVaginaMilk
A N = 6 (1 lactating)223351
B N = 8 (2 lactating)657632
Tot = 458710983

Isolation of Methicillin-resistant Staphylococcus pseudintermedius

Sixteen of 45 S. pseudintermedius strains (35.5%) were oxacillin resistant and mecA positive; all of them were isolated in Kennel B, from five bitches that were colonized in one (one bitch) or more areas (Table 2).

Table 2.   Localization of Methicillin-resistant Staphylococcus pseudintermedius strains
KennelMouthNaresAxillary skinNipplesVaginaMilk
B N = 5 (2 lactating)323332

One of the bitches that resulted MRSP positive in the vaginal swab and in the nipple area is the mother of some puppies that died of MRSP septicaemia in 2008 and is also the oldest bitch of the tested group (10 years old).

Also the youngest animal (9 months old) in the examined group resulted colonized by MRSP in oral mucosa, axillary skin and nipple.

A 3-year old lactating bitch was MRSP positive in all the samples, and the bacterium was isolated in high number and alone as a single organism.

In other cases, methicillin-resistant and methicillin-susceptible S. pseudintermedius strains were isolated from the same localization.

The bitch that had MRSP in milk in 2008 and that was not lactating in 2010 was not colonized by MRSP in any of the sampled areas.

Antimicrobial susceptibility

The pattern of antibiotic susceptibility of both the old and the new MRSP isolates is reported in Table 3; all the strains show multidrug resistance.

Table 3.   Antibiotic-susceptibility profile of Methicillin-resistant Staphylococcus pseudintermedius strains isolated in 2008 and in 2010 *bacterial strains that showed the same susceptibility are grouped
  1. R, resistant; S, susceptible; I, intermediate [11].

Amoxycillin–clavulanic acidRRRRRR
First-generation cephalosporinsRIRRRR
Third-generation cephalosporinsRRRSRR

MRSP typing

All isolates from 2008 and one strain per bitch (in total five strains) from 2010 (Kennel B isolates 1–5) were typed using spa and SCCmec typing. The isolates were all assigned to spa type t02 and carried the SCCmec II–III.

All isolates that were spa-typed were also subjected to PFGE, and it was shown that all of them were related, differing in up to five bands (Fig. 1). In the old cases, the isolates from Kennel B were indistinguishable from each other with the single exception of isolate B10 that differed in one band. In Kennel A, it was shown that all isolates differed from each other in 2–4 bands and that the PFGE pattern from Kennel B was not found among isolates from Kennel A. The isolates from 2010 from Kennel B differed from the old cases in 2–4 bands, but were indistinguishable from each other, with one exception, isolate B5, which differed in one band.

Figure 1.

 Pulsed-field Gel Electrophoresis (PFGE)-based dendrogram determined using DICE coefficient and UPGMA cluster analysis with position tolerance 1.2% and optimisation 1%, showing the patterns identified among the Methicillin-resistant Staphylococcus pseudintermedius


The MRSP strains from 2008 were isolated in a study on breeding bitches around parturition; all the MRSP-positive bitches had been treated with antibiotics during the sampling period. The study design consisted in cultures only coming from vaginal and milk swabs. In this work, more sites were tested, also from non-lactating bitches and bitches not receiving any antibiotics upon sample collection. Regrettably, three of the six MRSP-positive bitches from 2008 could not be included in this study because of the fact that they were no longer in the kennels. Furthermore, milk could not be tested in the remaining three bitches from 2008 either because they had been spayed or were no longer used for breeding. Even with these limitations, some interesting observations can be drawn from the results, which are different between the two kennels. No MRSP strains were isolated in Kennel A, not even from the two bitches that had MRSP in milk in 2008 and from the milk of the single lactating bitch in 2010. On the contrary, more than half of the examined bitches of Kennel B were colonized by MRSP. A plausible hypothesis is that the difference is because of the different use of antimicrobials, which has been reduced over time in Kennel A, while has remained excessive in Kennel B, where bitches are routinely treated around parturition to reduce puppy loss. The presence of MRSP strains in both milk samples from Kennel B is likely due to the fact that recent antimicrobials administration must have selected resistant strains, as already observed in 2008 (Rota et al. 2011). One of the milk-positive bitches was indeed colonized in all the other sampled area. The concomitant presence of methicillin-resistant and methicillin-susceptible S. pseudintermedius strains in the same localization in other less recently treated animals confirms the hypothesis of an external factor that eliminates sensitive strains and selects resistant ones. A previous work (Saijonmaa-Koulumies et al. 2003) showed that S. pseudintermedius (S. intermedius: old classification) flora of untreated bitches was composed of one or two dominant and persistent clones that were transferred from the dam to her puppies immediately after birth; other types, that were not able to become established, were infrequently isolated. Antimicrobials are, in our case, the factor that biases bacterial competition towards resistant strains. Temporal studies on the duration of MRSP carriage are needed because data concerning MRSP persistence, should the selective pressure stop, are missing. All we know is that we did not isolate MRSP strains after 2 years of correct antimicrobials use in Kennel A. However, it is worth noting the fact that a 9-month-old bitch born in Kennel B was colonized by MRSP in three areas, suggesting either an early, still persisting, colonization from the mother or a subsequent contamination from other dogs. Contact among dogs is frequent in that kennel because they are grouped together in the outdoor pens. Exposure to a contaminated environment or contact with an infected dog can lead to MRSP contamination of cohabitating dogs, as shown in the study by Van Duijkeren et al. (2011); dogs, being natural hosts for S. pseudintermedius, rather easily become colonized with MRSP, but whether this is a long-term colonization remains to be investigated; persistence of environmental contamination appears to be short-lived when the infected dog becomes MRSP negative (Van Duijkeren et al. 2011), but in some cases, it can last for longer period of time (Laarhoven et al. 2011).

In this study, it was shown, on the basis of molecular methods, that all investigated isolates were probably related to each other. The result is not unexpected as earlier studies established that one major clone lineage dominates in Europe, the ST71-J-t02-II–III (Perreten et al. 2010; Ruscher et al. 2010). The spa type and SCCmec type identified show that our isolates belong to this lineage. This is also confirmed by comparing PFGE patterns in this study to European isolates belonging to ST71-J-t02-II–III included in Perreten et al. (2010) (Data not shown). Differences between the two kennels were also identified on the basis of PFGE results. In Kennel B, identical or nearly identical PFGE patterns were identified, whereas in Kennel A, we observed different PFGE patterns in all the ‘old’ isolates. These results show that one specific clone has spread in Kennel B, which is not the case for Kennel A. A likely explanation for the difference is the continuous use of antimicrobial agents in Kennel B, which may have led to the selection of this specific clone. It is interesting to note that the same PFGE pattern is not identified in Kennel B in the new cases compared with the old ones. Notwithstanding, we still only have one clone present in Kennel B. So, a first hypothesis is that the clone in the new cases may have evolved from the clone in the old cases. The difference in band patterns between old and new cases may be because of uptake of plasmids, inserts in the genome and/or point mutation driven by the continued use of several different classes of antimicrobial agents and time. This theory can be supported by the observation that the new bands identified are small in size (18–26 kB), suggesting that they might be the result of the uptake of a plasmid. Nonetheless, it is also possible that a new clone was introduced in Kennel B, as the only MRSP milk-positive dog from 2008 resulted negative in 2010.


Misuse of antimicrobials is the key factor for the selection of MRSP strains in healthy dogs and for their persistence over time. MRSP strains always show multidrug resistance, particularly resistance to virtually all antimicrobial agents approved for veterinary use in dogs and cats in Europe. Dog breeders should be aware that infections caused by multiresistant bacteria have very limited therapeutical options and represent a huge challenge for animal health and a concern for public health.


The authors would like to thank Dr Hasan Sontas, Dipartimento di Scienze Cliniche Veterinarie (Padua, Italy), and Mr Roberto Perin, Istituto Zooprofilattico Sperimentale delle Venezie (Padua, Italy), for their assistance with this study. Our acknowledgements also go to Annica Landén, Deparment of Animal Health and Antimicrobial Strategies (SVA), for her help with the PFGE.

Conflict of interest

None of the authors have any conflict of interest to declare.

Author contributions

A Rota designed the study and drafted the manuscript; C Milani was responsible for the sampling work; M Corrò performed bacteriological cultures; I Drigo identified MRSP; S Börjesson performed Spa typing, SCCmec typing and PFGE, and drafted the manuscript. All authors have contributed to interpretation of data and revision of the manuscript.