Materials and methods

Description of the setting:
The Veterinary University of Vienna [Veterinärmedizinische Universität Wien, (VUW)] consists of a large hospital with separate departments for small animals, horses, farm animals, reproduction and diagnostic imaging/laboratory diagnostics. On average, 23 000-24 000 domestic animals including horses, ruminants, pigs, dogs, cats and rodents are admitted to hospital for a variety of diseases each year. Within the equine department there are separate clinic buildings for orthopaedics, soft tissue surgery and internal medicine. When necessary for diagnostics and/or specialised treatment, animals are moved between different clinics. Furthermore, veterinarians undertaking postgraduate education are on duty in different departments, and move freely between the various clinic buildings.

Origin of MRSA from infections and nasal colonisation in horses
Clinical isolates (from 24 cases) were obtained from specimens for bacteriological diagnostics that were routinely submitted in cases of wound infections, infected joints and suspected infections of various organ systems from summer 2003 until spring 2005.
In order to investigate nasal colonisation, the both nostrils of 24 horses (4 with an MRSA infection, 20 without) that were treated by the orthopaedics department during the same time period in 2004 and 2005 were screened for MRSA by taking nasal swabs. Colonisation was found in only 1 of these animals.

MRSA from nasal colonisation of VUW staff and veterinarians:
Specimens originated from direct cultures of swabs taken from both nostril

Reference strains for healthcare-associated epidemic MRSA
These strains represent multilocus sequence types (ST) of the major clonal lineages of epidemic MRSA from Europe (ST22: 1678/96; ST05: 3391/02; ST247: 134/93; ST45: 1150/93, ST254: 993/93 and 1000/93).
The strains were initially isolated from outbreaks of healthcare-associated infections and were established by representative SmaI macrorestriction patterns and multilocus sequence types (MLST). These strains were included in the HARMONY collection of epidemic MRSA from Europe [18] and in the first MLST-based population study of MRSA from sources worldwide [3].
In the study described here, these reference strains were used for comparison of SmaI macrorestriction patterns.

Reference strains for community-acquired MRSA (CA-MRSA)
ST80: 3925/02; ST01: 2773/03; ST30: 1880/04.
These strains represent multilocus sequence types of community-acquired MRSA that are frequently isolated in central Europe [(6-8] and have been used in this study for comparison of SmaI macrorestriction patterns. They originate from deep-seated skin infections in the community without hospital association, and are positive for the Panton-Valentineleukocidin determinants (lukS-lukF).

Methodology of specimen processing
MRSA from infections in horses were obtained from direct cultures of swabs onto blood agar-plates. Colonies typical for S. aureus were subjected to species identification according to standard procedures [19] and were also evaluated for antimicrobial susceptibility [20]. Nasal colonisation swabs from the anterior of horses and of veterinary personnel were streaked onto blood agar plates and in parallel onto CHROM agar for MRSA from Becton-Dickinson. After incubation for 48 hours, at least five colonies that were suspected to be S. aureus were further subjected to species identification and antimicrobial susceptibility testing.

Susceptibility testing
First line testing in veterinary clinical microbiology was performed by disk diffusion assay [20]. All isolates exhibiting oxacillin resistance were subjected to microbroth assay for MIC determination [20] and to polymerase chain reaction (PCR) for the mecA gene.

Molecular typing
SmaI macrorestriction patterns were obtained by use of the standardised HARMONY protocol [18] with subsequent cluster analysis based on the soft ware
described by Claus et al [21]. For comparison of SmaI patterns, cluster analysis was performed by comparing gel images.
For multilocus sequence typing (MLST) primers used and conditions of the PCR reaction corresponded to those described by Enright et al [3]. Sequences were analysed by use of the MLST databank (http://www.mlst.net).

Characterisation of SCCmec elements by PCR
PCR for ccr-complexes, detection of type II and type III specific sequences and discrimination of type IV was performed as described by Witte et al [6].

Demonstration of antibiotic resistance and virulence associated genes by PCR
PCR for lukS-lukF was performed as described by Witte et al 2005 [5]. For PCR detection of genes conferring resistance to methicillin (mecA), oxytetracycline (tetK, tetM), macrolides (ermA, ermB, ermC) and gentamicin (aac6’-aph2”), primers used and conditions were as in previous studies (Braulke et al [22] and Werner et al [23]). For PCR for superantigen determinants (tst, eta, etb, etc) primers and conditions were used as described by Mehrotra et al 2000 [24].

Results

Emergence of MRSA infections in horses:
In 2003 there were 344 equine cases from which clinical specimens were submitted for bacteriological, diagnostics. S.aureus was isolated in 47 (14%) of these cases including 19 infections with MRSA. In 2004 samples from each of 29 among 259 cases were positive for S.aureus (11%) with 3 of them confirmed as MRSA infection. From January 2005 until April 2005 there were 21 S.aureus infections among 165 equine cases (13%), 2 of them were MRSA infections.
The time course, type of infection with MRSA and clinical department affected are shown in Figure 1. The index case occurred in surgery in mid 2003. Investigation into the introduction of MRSA from the community into the hospital via this patient was unsuccessful.
Currently, we have no information regarding cases of MRSA infections from other veterinary institutions in Austria. In this country the frequency of MRSA among S. aureus from healthcare-associated infections in humans is approximately 10%. This represents a relatively low incidence of infections when compared to the situation in other European countries [1]. Overall, the incidence of infections at the VUW with MRSA appears low considering the number of about 5000 horses admitted in 2004 and 2005, that means about 4.8 cases with an MRSA infection MRSA per 1000 admissions.

Typing and comparative characterisation to MRSA from humans:
All 24 isolates from infections horses exhibited similar SmaI macrorestriction patterns with only minor variations that are still in the range of variability during the course of an epidemic (25). This pattern is consistent with intrahospital spread of one particular MRSA clone. These fragment patterns were different from those exhibited by healthcare-associated epidemic MRSA disseminated in Europe and from those of community-acquired MRSA [FIGURE 2].

Furthermore, there was no congruence when SmaI-patterns of MRSA from horses were compared to patterns of 3680 MRSA isolates from healthcare-associated and community-acquired infections that were sent for typing to the author’s laboratory as the German National Reference Center for Staphylococci at the Robert Koch Institute between 2001 and 2004.
Five horse MRSA isolates that were subjected to MLST were identified as ST254.
PCR for typing of SCCmec elements that was performed on 5 isolates from horses revealed type IVd whereas IVc was found for MRSA of ST254 from humans [TABLE ]. None of the investigated horse MRSA contained lukS-lukF, tst1, eta, etb or etc.

Transmission to human nasal colonisation of personnel and veterinarians:
During the time periods of emergence of MRSA infections in horses in the surgery and orthopaedic clinics in 2004 and 2005, nasal swabs from 43 people that were directly involved in treatment of animals (veterinarians, veterinary assistants, animal keepers) were investigated. Two veterinarians were revealed as long term carriers (massive colonisation demonstrated in both a first investigation and follow-up sample 3 weeks later). The MRSA isolates exhibited the same SmaI macrorestriction patterns as isolates from infections in horses and contained SCCmec IVd elements.

Nasal colonisation of horses:
Data from human medicine indicates that nasal colonisation is an important reservoir with regard to infections of the primary carrier and to further dissemination [26,27]. A temporary colonisation (negative in a second investigation) was detected in only one among 24 horses. The MRSA isolate exhibited the same SmaI macrorestriction pattern as the isolates from infections and also contained a SCCmec IVd element.

Discussion

MRSA from infections in horses in a central European veterinary hospital exhibit MLST ST254. This type has also been identified in healthcare-associated epidemic MRSA. This strain was frequent in the 1990s but has subsequently decreased in prevalence [28]. Reference isolate 994/93 is a representative of ST254 that was disseminated in the hospitals of the Order of Holy Elisabeth in the south-west of Germany and west of Austria [28]. A direct relationship between human MRSA of ST254 to those from horses is however unlikely,as both exhibit different SmaI-macrorestriction patterns and contain different subtypes of SCCmec IV elements. Subtypes of SCCmec elements of type IV differ by various DNA sequences in the region downstream from mecA. At the present time, no acquisition or loss of these sequences has been observed during the time course of dissemination of epidemic MRSA. PCR typing based on subtype specific DNA sequences appears to be a reliable tool for discrimination of subtypes. Demonstration of different subtypes of SCCmec elements in the genomic background of ST254 does however, not exclude an exchange of MRSA between humans and horses in the past. Another possibility is that methicillin-susceptible S. aureus of ST254 that was already widely disseminated among humans [29] was transferred to horses and later acquired a SCCmec element that is different from those acquired by human MRSA of ST254.
Until now MRSA exhibiting typing patterns like those of ST254 from horses have not been detected among MRSA isolates from infections in humans. Furthermore, the human ST254 strain has so far only been associated with healthcare-associated infections and has not emerged in the community.
However, the finding of stable nasal colonisation of two veterinarians who had been in contact with animals affected by MRSA infections demands further investigation of potential animal to human transmission.

This is underlined by findings of MRSA among horses that were reported from Canada. In this report a single well-recognised MRSA clone exhibiting the so called CA-MRSA-05 typing pattern that previously had been identified in healthcare-associated settings from 5 different geographical sites in Canada was demonstrated to have the ability to colonise the nose of horses. This clone spread among both horses and humans on farms and among personnel in veterinary hospitals (14,15). When representative isolates of these MRSA strains were subjected to MLST, we found it to be ST 8 [TABLE]. MRSA strains exhibiting ST8 are widely disseminated in US hospitals and may also become more frequent in Canada since a population dynamics of S. aureus and in particular of MRSA is well known [30]. Furthermore, MRSA of this clonal lineage containing the lukS-lukF gene that confers enhanced virulence became prevalent as community-acquired MRSA in the US [31], and sporadic cases of infections in the community have also been reported from Norway [32].

Conclusion

Infections in horses with MRSA of MLST ST254 emerged independently of MRSA infections in humans. Although MRSA in horses may presently not represent a substantial reservoir for infections in humans in central Europe, further surveillance is needed with respect to human transmission and to emergence of new clonal lineages.

Acknowledgement: we are grateful to Magister Ewald Denner, Institute for Bacteriology at Vienna
Veterinary University, for proving MRSA isolates from 2003, and in part from 2004.