INTRODUCTION
Methicillin-resistant
Staphylococcus aureus (MRSA) was first discovered in the 1960s and is thought to have developed independently from several clones of methicillin-susceptible
Staphylococcus aureus (MSSA) (
1). In human medicine, it was initially recognized as a nosocomial infection; however, in the 1990s, an increasing incidence of hospital-independent human MRSA infections was observed (
2). These so-called community-acquired MRSA (CA-MRSA) cases have since been reported in many countries.
With the detection of MRSA in pigs and farmers, MRSA has become a One Health issue (
3). These livestock-associated MRSA (LA-MRSA) strains are not only associated with disease in animals, particularly horses, but also human infections (
4,
5). Pigs, in particular, can be heavily colonized with MRSA and are thought to constitute the main reservoir (
4,
6–8). Therefore, humans with regular close contact with these animals, e.g., farmers, slaughterhouse workers, and veterinarians, have a higher risk of being colonized and thus of developing infections (
6,
9–12). Furthermore, LA-MRSA can be found in horses, veal calves, and poultry (
1). In horses, LA-MRSA can also be an agent of nosocomial infection in an animal clinic environment and can be transmitted to clinical staff in this context (
13,
14).
LA-MRSA strains mainly belong to clonal complex 398 (CC398). They are thought to have emerged from ancestral human MSSA, gaining resistance genes but losing human specific virulence factors in the process (
15). The human innate immunomodulatory genes (chemotaxis inhibitory protein [
chp], staphylococcal complement inhibitor [
scn], and the plasminogen activator staphylokinase [
sak]) often carried on a phiSa3 prophage are usually not present in animal isolates (
15). However, it is possible for CC398 MRSA to regain these virulence factors, which might enhance its potential as a human pathogen. The acquisition can occur by lateral gene transfer during cocolonization in the host (
16). Other known virulence factors, like leucocidins, e.g., Panton-Valentine leucocidin (PVL) or enterotoxins, are also rarely found in LA-MRSA (
1,
17).
MRSA strains are defined by their methicillin resistance, mainly mediated by
mecA, but they frequently also carry additional resistance genes. For LA-MRSA, this is most frequently a tetracycline resistance [e.g.,
tet(M)] gene which is present in virtually all strains, probably reflecting the widespread use of this antibiotic on farms (
17,
18). The
vga(E) gene conferring resistance to streptogramin A, pleuromutilin, and lincosamide was also originally detected in an LA-MRSA isolate from a Swiss pig (
19). It was found to be located on Tn
6133 together with
erm(A) (macrolide, lincosamide, and streptogramin B resistance) and
ant(9)-Ia (spectinomycin resistance) (
19,
20).
Various typing methods have been used to analyze the phylogeny and epidemiology of MRSA. A classical multilocus sequence typing (MLST) scheme based on seven housekeeping genes was developed by Enright et al. in 2000 (
21).
spa typing based on repeats in the staphylococcal protein A sequence allows higher discrimination than does MLST and remains a popular typing method (
22). However, whole-genome sequencing-based methods are now rapidly becoming standard since they allow much higher discrimination as well as detection of virulence and resistance genes (
23). The MLST scheme has been expanded to core-genome MLST (cgMLST), which allows a more detailed analysis, while single-nucleotide polymorphism (SNP)-based methods allow the highest discrimination and are most useful for closely related isolates (
23).
The current study aims to evaluate the spread of LA-MRSA (with a special focus on pig isolates) to farmers and veterinarians. In this framework, traditional
spa typing was compared to cgMLST and SNP analyses of selected isolates. A previous MRSA prevalence study in Swiss veterinarians in 2012 revealed a prevalence of 3.8%, with LA-MRSA only detected in veterinarians treating farm animals (
24). Since then, routine monitoring data indicated a greater than 100% increase in MRSA prevalence in pigs (
25). Therefore, we wanted to analyze the impact of this development on MRSA carriage in farmers and veterinarians. To address this question, farmers and veterinarians were sampled at two major Swiss conventions. The collected MRSA isolates (
n = 23), as well as selected isolates from pigs (
n = 12), pork (
n = 2), poultry meat (
n = 3), and horses (
n = 3), were subjected to whole-genome sequencing.
DISCUSSION
In the current study, the prevalence of MRSA in Swiss slaughter pigs was investigated, and the isolates were compared to those from Swiss veterinarians, farmers, and horses.
MRSA sequence type 398 (ST398)-
spa t011 is responsible for the steep increase in MRSA in Swiss slaughter pigs in recent years. This
spa type has also been reported from pigs in other countries. A recent study from Spain looking at indoor-housed pigs found over 80% of strains belonging to this type (
27), while a study looking at isolates in Cameroon and South Africa found only this type (
28). In Spain, an MRSA ST398-t011 isolate was even found in a wild boar (
29).
spa typing results for MRSA strains from pigs at slaughter for 2017 were available from Finland (MRSA prevalence, 77%, with 43% from
spa t034 and 12% from
spa t011), where
spa t034 is dominant, and Spain (MRSA prevalence, 90%, with 11% from
spa t034 and 70% from
spa t011), where there is a clear dominance of
spa t011 (
30). Sieber et al. (
26) found that CC398 isolates from Danish pigs clustered into three different lineages. Our analysis, which included selected strains from this study, showed that the Swiss strains, for the most part, form separate clusters, which may reflect the low number of imports in the Swiss pork industry. Interestingly, Swiss MRSA ST398-t011 from pigs forms a distinct cluster separate from that of farmers and veterinarians in contrast to porcine MRSA ST398-t034 (
Fig. 3). Hence, porcine MRSA ST398-t011 seems to be a successful colonizer of pigs but not humans. Antimicrobial pressure as a driver can be excluded, as these strains harbor, in contrast to porcine MRSA ST398-t034, a very low number of resistance genes in addition to the beta-lactam resistance genes. The underlying molecular features need further research.
An obvious divergence between the CC398 MRSA and the others in the examined sample is the difference in resistance and virulence gene distribution. While LA-MRSA strains, especially those of
spa t034, contain a higher number of resistance genes, HA/CA-MRSA strains harbor more virulence genes. At least two out of three genes of the immune evasion cluster (IEC) (
chp,
sac, and
scn) were detected in all HA/CA-MRSA strains but only in three LA-MRSA strains. Two of these were
spa t034 strains from farmers, while one was a
spa t011 strain from a horse. The IEC genes are thought to be a marker of human adaptation and are rarely present in LA-MRSA (
31,
32). CC398 MRSA strains are thought to have lost the IEC during their evolution from human MSSA and adaptation to pigs (
15). It is thus conceivable that by regaining these virulence genes, they might readapt to the human host and thus increase their pathogenic potential. Since the current study looked at human carriers, not infections, we cannot draw conclusions about the pathogenicity of our isolates. Horizontal gene transfer during cocolonization has been demonstrated between different CC398 strains in pigs under experimental conditions (
16) and might therefore also occur in the noses of humans. Isolates from horses have been described to be more frequent carriers of IEC genes (
33). Since we included only three horse isolates, we cannot give an estimate of the IEC prevalence in Swiss horse strains. However, since one of them was positive but none of the 12 pig isolates was, it might be more frequent in horse than in pig isolates.
A prevalence of 6.6% in veterinarians was higher than in a comparable study from 2012, where it was 3.8% (
24). Treating horses clearly emerged as a risk factor for LA-MRSA positivity in veterinarians, with an odds ratio of 6.2. This observation is corroborated by the phylogenetic analyses clustering all but one veterinary isolate with horse strains. Contact with horses is a known risk factor for MRSA transmission to veterinary personnel (
14,
34). Recent studies in other European countries also found ST398-t011 to be the dominant type in horses (
14,
33), while older studies found other types, mainly CC8 (
35). Abdelbary et al. (
36) looked into the phylogeny of horse-associated LA-MRSA strains from different countries and found a CC398 subclone associated with equine hospitals. They used denaturing high-pressure liquid chromatography for mutation discovery as opposed to our whole-genome sequencing; therefore, our results are not directly comparable. However, the equine subclone found in their study was also associated with
spa t011. The emergence of ST398-t011 in horses coincides with the rise of this type in pigs. Considering the cgMLST and cgSNP analyses, there appears to be no epidemiological link between the two, as the majority of pig isolates clearly formed a separate cluster.
In the present study, we also included some strains isolated from pork and poultry meat. The prevalence in pork was found to be extremely low (<1%), meaning that it is a very unlikely source for human infection. One of the two pork strains was not even LA-MRSA, which could indicate that the meat was contaminated by human handling. All three sequenced poultry strains were isolated from imported meat since there were no isolates from Swiss meat; this could explain their separation from the main cluster of t034 strains. Overall, there is no indication that either poultry meat or pork plays a major role in human colonization with MRSA in Switzerland.
LA-MRSA strains are also found in hospital patients, meaning they not only colonize patients but also cause infections. While the incidence in Europe is still low (Goerge et al. estimated <3 infections/100,000 inhabitants per year in the German and Danish populations), it can be higher in areas with high livestock density (
37). Among the total MRSA isolates from two Swiss hospitals in 2017, only about 1% were LA-MRSA (two isolates, both
spa t011) (
25), indicating that the incidence seems to be low. LA-MRSA strains might reacquire more human-associated virulence factors and thus evolve into more virulent strains. Since LA-MRSA strains already harbor more additional resistance genes than do other types, these infections are also more difficult to treat.
Moreover, MRSA prevalence and the role of cattle have to be considered. Due to the comparatively low prevalence and the focus on pigs, they were not included in the sequencing study. The prevalence in dairy cattle is not systematically monitored in Switzerland; however, it is probably very low, as MRSA strains are rarely isolated from milk or other clinical samples. In 2012, 200 bulk milk samples were analyzed for the occurrence of MRSA, and three MRSA isolates were found (1.5%) (
38). The prevalence in veal calves, though still comparatively low (8% versus 44% in pigs), is on the rise and might also become a factor in transmission to humans. The majority of veal calf isolates in 2017 were
spa t011 (58%) which makes them an unlikely source for the farmers in the present study. However, they should be included in future studies, especially if the prevalence continues to rise.
In conclusion, LA-MRSA is a serious emerging problem in the pig industry, indicating possible antimicrobial overuse. However, the recent rise in spa t011 strains from pigs does not directly translate to a higher prevalence in humans. This type appears to be less likely to colonize humans or acquire resistance genes. Further research is necessary to confirm this finding and elucidate the underlying cause of the successful spread of this clone.