Survey of Staphylococcus aureus carriage by free‐living red deer (Cervus elaphus): Evidence of human and domestic animal lineages

Abstract Staphylococcus aureus is a pathogen that can affect multiple host species. Evidence of transmission between humans and animals and among different animal species has been reported in recent years. In this study, we investigated 284 free‐living red deer (Cervus elaphus) in the Central Italian Alps to assess the prevalence and molecular characteristics of S. aureus in nasal and intestinal samples in relation to host features and environmental factors. A prevalence of 90%, 26.2% and 10.7% of S. aureus was detected in nasal rectal swabs and faeces, respectively. Calves had a higher probability of being S. aureus intestinal carriers than adults, especially in females when considering faecal samples. Clonal complex (CC) 425 was the most prevalent lineage (61.5%). This is a lineage known to be widespread in both domestic and free‐living animals. It was followed by CC2671 (15.4%) and CC350 (6.4%). A high rate of the phage‐borne virulence factor lukM/lukF‐P83 was detected in CC425 and CC350. Further lineages, which are known to occur in both humans and animals, were detected sporadically in red deer faeces only, that is, CC7, CC9, CC121 and CC707, harbouring the genes of the penicillinase operon and a gene for macrolide resistance (CC9 and CC121). Methicillin resistance genes mecA and mecC were not found. Our results suggest that free‐living red deer may be reservoir for S. aureus in Alpine habitats.

various mechanisms of resistance and, for instance, methicillinresistant S. aureus (MRSA) strains were isolated few years after introduction of this compound (Jevons, 1961). In livestock, S. aureus is a major causal agent of mastitis, a disease that greatly affects the economics of dairy cattle and small ruminant industries (Bergonier et al., 2003;Ruegg, 2017). Staphylococcus aureus transmission events between humans and livestock have been described and the massive use of antibiotics in both, human and veterinary medicine, might provide a selective pressure favouring antibiotic resistance (Haag et al., 2019;Richardson et al., 2018). The role of other host species is still less clearly defined. A wide collection of strains of S. aureus from European wildlife, including cervids, has recently been investigated, highlighting a great genetic diversity and rather low antimicrobial resistance rates (Monecke et al., 2016). The occurrence of MRSA in wildlife without exposure to antibiotics has been recently reviewed (Heaton et al., 2020: Silva et al., 2020, showing that most MRSA isolated from wild animals are mecC positive. The dissemination of this mechanism of resistance (García-Álvarez et al. et al., 2011;Shore et al., 2011) in freeliving animals and river waters (Porrero et al., 2014a) suggests that the natural environment may be a reservoir of mecC gene, and indeed, a cocolonisation of hedgehogs with S. aureus and dermatophytes that produce penicillin provides a selective pressure favouring mecC in the wild (Dube et al., 2021;Larsen et al., 2022;Smith & Marples, 1965).
Red deer (Cervus elaphus) populations are currently increasing in density and distribution across Europe (Lovari et al., 2018), also owing to the ecological plasticity that allows this species to inhabit both natural environments such as forests or pastures, as well as agricultural landscapes. The interest to acquire baseline information on untreated free-living animals and monitor the possible effect of the influx of human-and livestock-associated S. aureus/MRSA on wildlife has been recognized (Heaton et al., 2020;Monecke et al., 2016).
The aim of this study was focused on an Alpine population of freeliving red deer to systematically survey the S. aureus population structure to evaluate the role of red deer in S. aureus epidemiology, in relation to host features and environmental factors.

Study population and area
The red deer study population inhabits the Central Italian Alps in the Stelvio National Park. A large increase in red deer density occurred over the past decades (up to some 31 individuals/km 2 in wintering areas), leading the Park to start a culling program in 2011, with the aim to reduce the red deer density as well as its impact on agricultural activities, forest regeneration and biodiversity. Between 2014 and 2017, annual surveys resulted in an estimation of the deer population within the Park of approximately 1800 individual animals, with a mean density of 9.5 individuals/km 2 in autumn and 27.4 ind./km 2 in winter (Corlatti et al., 2016;Pedrotti et al., 2017).
Other wild ungulates in the area include roe deer (Capreolus capreolus), chamois (Rupicapra rupicapra) and ibex (Capra ibex). Furthermore, herds of cattle and small domestic ruminants share alpine pastures with red deer in summer.
The study area corresponds to the red deer wintering site within the Province of Sondrio and ranges between 1200 and 2400 m a.s.l.
The study area was functionally divided into three macroareas with different levels of anthropization (low, medium, high) (Formenti et al., 2015;Lauzi et al., 2021). Information on land use were retrieved from CORINE land cover maps (https://land.copernicus.eu/pan-european/ corine-land-cover) and processed with QGIS (https://qgis.org/it/site/) to define the proportion between the surface occupied by human settlements/farming activities and the whole surface of the culling units.
Human settlements and agricultural landscapes around small villages constitute 7% of the whole surface in the low anthropized macroarea (772 ha), whereas they represent 32% of the total surface in the high anthropized macroarea (707 ha). A total of 1200 ha are considered medium anthropized macroarea. Observations on radio-tracked animals suggest limited deer movements among macroareas in autumn and winter.

Sampling and data collection
Red deer were annually culled by hunters in winter, under the supervision of the Park Authority. The culling plan was legally authorized by ISPRA, the Italian Ministry of Environment (Prot. 48585/T-A25-Ispra).
The present study was carried out in two consecutive seasons (2016-2017 and 2017-2018). Within a few hours after culling, animals were brought to the control centre of the Park, where individual information including age, sex, body mass and culling area were registered.
Animals were classified as calves (<1-year old), yearlings (1-year old) and adults (≥ 2years old). Nasal and rectal swabs were collected using Amies Agar Gel Transport Swab (Oxoid, Fisher Scientific Italia, Rodano, Italy). Swabs were taken with care to avoid touching the skin and were put directly into the transport tube to avoid cross-contamination. Faeces were taken, using disposable gloves, directly from the rectum and put into a sterile container. In the first season, nasal swabs, rectal swabs and faeces were sampled, whereas in the second season, only nasal swabs and faeces were collected. Specimens were stored at the check point at −20 • C. Subsequently, frozen samples were transferred to the laboratory monthly for further processing. A preliminary analysis, processing two replicates of samples after 1-2 days and after 28 days of storage at -20 • C showed the same S. aureus detection rate (data not shown).

Staphylococcus aureus isolation
Nasal swabs were incubated in 5 ml of Mueller-Hinton broth with 6.5% NaCl at 37 • C for 24 h. Pre-enrichment cultures were subsequently plated on Baird-Parker with rabbit plasma fibrinogen (Oxoid Baird-Parker RPF Agar Base, Fisher Scientific Italia, Rodano, Italy) and incubated for 48 h at 37 • C. Faeces were suspended in sterile H 2 O (≅ 1 g/sample in 4 ml), diluted (1:10) in Mueller-Hinton broth with 6.5% NaCl and then processed as previously described for swabs. One to four grey/black colonies with an opaque halo per sample were isolated on Blood Agar plates (Microbiol Diagnostici, Cagliari, Italy). Bacterial colonies were identified as S. aureus by colony morphology, haemolysis pattern and coagulase reaction, and the identification was confirmed by amplification of the nuc gene (Baron et al., 2004).

DNA microarray analysis
For each animal, one nasal S. aureus colony and two rectal or faecal colonies were analyzed in the first sampling season and two faecal colonies in the second sampling season. Specifically, S. aureus isolates were selected from animals that were positive at both, nasal and intestinal sites by rectal swab or faeces and from faecal carriers, irrespective of nasal status. Staphylococcus aureus isolates were characterized using the StaphyType DNA microarray based assay (Alere Technologies, Jena, Germany). This assay detects S. aureus target sequences, including species and antimicrobial resistance markers, allowing isolates to be assigned to MLST sequence types (STs) and/or clonal complexes (CCs), and staphylococcal cassette chromosome mec (SCCmec).
Details of these procedures as well as a list of target sequences have been previously described in Monecke et al. (2008).

Statistical analysis
A Paerson's chi-square test was used to assess the relationship between presence of S. aureus and independent variables such as type of sample, season of collection, sex, age and anthropization level. A p-value < .05 was considered as statistically significant, whereas tendency was considered in the presence of p-value >.05 but <.1. Statistical analysis was performed using Epitools-Epidemiological calculators available at: http://epitools.ausvet.com.au (Sergeant, 2018 where all terms were additive due to paucity of data. Next, stepwise selection was performed to find a more parsimonious model. We inspected the adequacy of each final model through quantile residual diagnostics (Dunn & Smyth, 2018), assessed the Tjur's coefficient of discrimination (Tjur, 2009) and calculated the predictive accuracy by means of the area under the receiver operating characteristic (AUC-ROC) curve (Fawcett, 2006). Given the small sample size, robust estimates of confidence intervals were obtained by bootstrap using 1000 iterations and the adjusted percentile method.

Staphylococcus aureus prevalence
All isolates were tested for methicillin resistance genes mecA and mecC but all were negative. CC7 and CC707 harboured the genes of the penicillinase operon (blaZ/I/R); CC9 and CC121 were multiresistant, carrying also a gene for macrolide resistance (ermA).
Strain typing showed in 15 out of 20 animals investigated a presence of the same CC at both sampling sites (Table 3). In detail, 12 animals harboured CC425, two CC2671 and one CC350. Concerning the two S. aureus isolates analyzed per sampled site, the same CC was always observed for both isolates from rectal mucosa, whereas in colonies from faeces a different CC was identified in four out of 14 cases.   Figure 1). The probability of presence of S. aureus in the faeces was best explained by the interactive effect of sex and age class (pseudo-R 2 = 5%, AUC-ROC = 0.69). Although the interaction was nonsignificant, the effect size was non-negligible, suggesting that the mean probability of presence of S. aureus from calves to adults decreased more rapidly in females (from 20% to 4%) than in males (from 19% to 13%) (Table 4, Figure 2). The variation in the probability of the presence of 'human-livestock type' S. aureus in the faeces and rectal swabs was also explained by a 'null' model, with a mean probability of presence of 27% (Table 4). All the selected models adequately described the data, as suggested by the unsystematic distribution of quantile residu-

DISCUSSION
Our study allowed us to assess the prevalence and genetic diversity of S. aureus in a population of free-living red deer in the Alps, identifying the frequency of carriers, genetic lineages and resistance genes at nasal and intestinal sites also in relation to host and environmental factors.
The prevalence of S. aureus was particularly high in nares, with a carriage rate higher than those previously reported in free-living red deer, that is, 90% in this study versus values between 19.2% in Spain (Porrero et al., 2014b) and 49% in Germany (Meyer et al., 2014). The S. aureus isolation protocol applied in this study, with enrichment and selective steps not or partially reported in the above investigations, may have contributed to an increase in the detection rates. The high population density of red deer, reaching a winter density of nearly 30 ind./km 2 in the study area (Corlatti et al., 2016), however, should be also considered, as it is likely to increase the transmission of S. aureus by direct contact, by ingestion of contaminated forages on pasture or by inhalation of infected droplets (Kozajda et al., 2019). It is difficult to compare results of S. aureus carriage with respect to density of red deer population because densities are subjected to annual fluctuations and different management strategies may occur in Europe, including reintroduction programs (Gnat et al., 2015).
In the present study, the rate of intestinal carriers is significantly lower than the one of nasal carriers both from rectal swabs (26.2%) and faeces (10.7%). A similar trend was reported in pigs (Khanna et al., 2008) and humans, where the average detection rate of intestinal carriage is 20% in healthy individuals and patients (Acton et al., 2009). In the present study, the significantly lower rate of intestinal S. aureus prevalence was also confirmed in the subset of red deer simultaneously sampled by nasal and rectal swabs and faeces.
The literature on S. aureus detection at intestinal site in red deer is scant, although that issue might be of relevance with regard to transmission. To the best of our knowledge, there are only two studies addressing this issue and they detected S. aureus in 0.67 % of faecal samples collected in the free-living population from Hungary, Slovakia and Poland (Gnat et al., 2015) and in 33.3% from Spain (García et al., 2020).
Noteworthy is the higher probability of calves being S. aureus intestinal carriers, compared to adults, with regard to both rectal mucosa and faeces. Even though some caution is required in the interpretation of the statistical results, owing to the limited sample size, this agerelated pattern is not surprising, as colonization of the intestinal tract with S. aureus in newborns and young children is observed at a high frequency also in humans, suggesting the occurrence of a mechanism of mother-to-child transmission (Acton et al., 2009). Calf behaviour with direct contact with mother, suckling and closeness among individuals likely explains the higher prevalence observed compared to adults.
Moreover, lactation in red deer is long termed and it could last until 1-year old m n/a n/a n/a CC7 CC7/CC2671 933 06/02/2018 1-year old f n/a n/a n/a CC7 CC425 975 08/02/2018 1-year old f n/a n/a n/a CC425 CC425 977 08/02/2018 1-year old m n/a n/a n/a CC425 CC7 Note: n/a: sample not available; +: S. aureus positive, microarray not performed.
12 months of age in barren females (Clutton-Brock et al., 1982) and occasional allosuckling, that is, suckling from a non-maternal hind, has been reported in the wild (Bartoš et al., 2001). Although we did not perform analysis on milk samples, long-term suckling and skin contact may have contributed to an increase in the probability of infection in calf.
The potential sex-related variation in the probability of presence of S. aureus in faeces from calves to adults, greater in females than in males, however, still remains unclear, though it might be related to a higher susceptibility to S. aureus of adult males compared to adult females, possibly due to the great energy expenditure of the former during rutting season.
Within the subset of S. aureus isolates analyzed by microarray, CC425 was the most prevalent lineage in all types of samples, reaching similar rates in the nares and rectal mucosa, 75% and 79.3%, respectively, and 33.3% in faeces. This lineage was previously identified in red deer in Austria, Germany and Spain (Porrero et al., 2014b;Monecke et al., 2016) and it is known to be widespread both in domestic and wild animals (Monecke et al., 2016). Differently from previous isolates, more than half of CC425 detected in our study carried the genes for lukM/lukF-P83, a leukocidin similar to the PVL (Kaneko & Kamio, 2004) and associated with bovine mastitis (Herron-Olson et al., 2007;Schlotter et al., 2012). Cross-infection might pose a risk to domestic animals, even though some caution is required in the interpretation of this finding. Indeed, the lukM/lukF-P83 positive lineages that have been reported in domestic cattle are others (CC151, CC479) than the ones in deer (Hoekstra et al., 2018).

TA B L E 4
Model estimates for the effects of different predictors selected to explain variation in (i) the probability of presence of S. aureus detected from nasal swabs; (ii) the probability of presence of S. aureus detected from rectal swabs; (iii) the probability of presence of S. aureus detected from faeces; (iv) the probability of presence of 'human-livestock type' S. aureus detected from faeces and rectal swabs. The  (García-Alvarez et al., 2011). This, however, was not the case in our isolates.
In the present study, CC2671 and CC350 were also identified in the nares, as well as in the rectum and faeces, but at much lower extent compared to CC425. CC2671 has sporadically been observed in red deer in Spain (Porrero et al., 2014b), and there are two entries of isolates from Spanish red deer in the MLST database (https://pubmlst.org/bigsdb?page = info&db = pubmlst_saureus_isolat es&id = 5037 and 5038). CC350 is rarely reported in humans and cattle (Monecke et al., 2016). Strain typing suggested involvement of the same strains for both nasal and rectal mucosae. Genes lukM/lukF-P83 were also found in CC350 isolates, another group in which they have, to the best of our knowledge, not yet been observed or where they are rare or where they might be restricted to host-specific lineages from animals that not yet have systematically been studied. The high rate of the phage-borne virulence factor lukM/lukF-P83 in red deer isolates (36 out of 79 isolates tested) suggests that it might play a pathogenetic role in S. aureus from red deer. It is beyond the scope of the present study to prove that experimentally, but further studies should aim at its role in infection or disease in red deer. It should be considered that the finding of the same CC or specific virulence factors in free-living animals as in humans and livestock is not evidence of an interspecies transmission event.
Further lineages were identified in faecal samples and included some that are known to infect humans and animals. In particular, CC7 has been reported in humans and cattle (Conceição et al., 2017;Cuny et al., 2019;Monecke et al., 2016). CC9 is recognized as a livestock-associated lineage in domestic pigs and farm personal (Mroczkowska et al., 2017;Ye et al., 2016), CC121 has been reported both in humans, in whom it is widespread and common, and rabbits (Kurt et al., 2013;Vancraeynest et al., 2006) and CC707 sporadically in humans (Egyir et al., 2014) and animals. The above-mentioned lineages were never reported in cervids, with the exception of CC707 that once was identified in reindeer (Rangifer tarandus) in Sweden (Monecke et al., 2016). CC2328 is too rare and poorly known so no assessment on geographic distribution or natural hosts can be pro-  (Luzzago et al., 2014) and might also represent livestock-associated lineages in Italy (Feltrin et al., 2016;Romanò et al., 2020).
Notably, no MRSA isolates were detected, supporting previous findings on free living red deer in Europe that reported absence or sporadic occurrence of MRSA (Monecke et al., 2016;Porrero et al., 2013Porrero et al., , 2014b. Nevertheless, CC7, CC9, CC121 and CC707 lineages harboured penicillin resistance genes and CC9 and CC121 also a macrolide resistance gene (ermA). These findings highlight that resistance genes are restricted to S. aureus CCs that are known to infect and or colonize both humans and domestic animals supporting the hypothesis of red deer exposure to an environment contaminated by livestock and human activities. Similarly, certain virulence genes, that is, sak, chp and scn, that are usually associated with isolates from humans, rather than with those from ungulates, might indicate exposure to human strains of S. aureus (Rohmer & Wolz, 2021).
Overall, our findings confirmed that nasal samples are the best option to detect S. aureus carriers in red deer compared to rectal mucosa and faeces, in accordance with previous results based on nares and skin (Porrero et al., 2014b). Nevertheless, although the nasal site is usually sampled to isolate S. aureus both from humans and animals due to the higher rate of detection compared to other sampling sites, S. aureus higher clonal complex diversity in faecal specimens versus nasal and rectal mucosa has to be considered in sampling dealing with S. aureus genetic investigation. Moreover, from an epidemiological standpoint, faeces represent an important route of dissemination of bacteria into the environment and analysis may be extended to other pathogens. Noteworthy, the present red deer population has also been reported as the carrier of potentially zoonotic Shiga toxin-producing Escherichia coli (Lauzi et al., 2021).
Our results suggest that free-living red deer may be a reservoir for S. aureus in Alpine environment. The S. aureus population mainly comprised lineages already known to occur in domestic or wild ruminants, although there were lineages on which no sufficient data are available.
In general, more typing, including whole-genome sequencing, on cervid isolates should be performed in order to identify host-specific lineages or host-specific factors and to understand which lineages might be endemic or involved in localized outbreaks. Human and other animal lineages, that harboured penicillin and macrolide resistance genes, were sporadic but may represent a source for environmental contamination. Untreated wild host populations, that do not experience selective pressure for resistance, could act as a reservoir for antimicrobial susceptible pathogens and may contribute to lower frequencies of the resistant pathogen (Park et al., 2015) on pastures and agricultural landscapes.