Beyond CC398: Characterisation of Other Tetracycline and Methicillin-Resistant Staphylococcus aureus Genetic Lineages Circulating in Spanish Hospitals

Tetracycline resistance (TetR) has been evidenced as a good phenotypic marker for detection of livestock-associated methicillin-resistant Staphylococcus aureus (LA-MRSA) isolates of the clonal complex CC398. The aim of this study was to characterise a collection of 95 TetR-MRSA isolates, not belonging to the lineage CC398, that were obtained in a previous multicentre study, to detect other MRSA clonal complexes that could be associated with this phenotypic TetR marker. The TetR-MRSA isolates were recovered from 20 Spanish hospitals during 2016 and they were characterised to determine their antimicrobial resistance and virulence phenotypes/genotypes as well as the presence of the immune evasion cluster (IEC). A high proportion of isolates belonging to the CC1 lineage (46%) were observed, as well as to the CC5, CC8 and CC45 lineages (11% each one). Thirty-two different spa-types were identified, being predominantly CC1-t127 (40%) and CC45-t1081 (11%). The IEC system (with the gene scn as marker) was present in 73% of isolates and 16% produced the Panton Valentine leucocidin (PVL). A high proportion of MRSA-CC1 isolates were scn-negative (38.6%) and 52.9% of them were blaZ-negative. A multidrug resistance (MDR) phenotype was identified in 86% of MRSA isolates. The knowledge of other TetR-MRSA genetic lineages, in addition to CC398, is highly relevant, since most of them were MDR and some of them presented important virulence factors. Strains potentially associated with livestock (as the subpopulation CC1-t127-scn-negative) or with humans (as the CC45 lineage or the subpopulation CC1-scn-positive) have been found in this study. The use of tetracycline-resistance for detection, not only of CC398 but also of other LA-MRSA lineages should be tracked in the future.


Introduction
Staphylococcus aureus is a common bacterium in human and animal regular microbiota, particularly present in the nose and on the skin. However, S. aureus can be an important opportunistic pathogen capable of causing anything from mild skin to life-threatening infections. S. aureus is also able to acquire multiple antimicrobial resistance mechanisms, and standing out amongst them, methicillin-resistant S. aureus (MRSA) isolates because of its resistance to almost all β-lactams [1].
Different clonal lineages have been initially proposed to be directly related to the environment in which they are found. In this way, we classically found hospital-associated (HA) or community-acquired (CA) MRSA. Nevertheless, nowadays this line of distinction within groups is becoming more and more blurred.
Newer livestock-associated (LA) MRSA lineages, genetically different from human isolates and closely related to farm animals, have been found more regularly in the healthcare system [2]. During the last decade, livestock have been considered to be a great reservoir of MRSA with high potential of zoonotic transmission to humans [3]. It has been observed that the spread of LA-MRSA from animals to humans occurs frequently and that livestock workers present a high risk for LA-MRSA colonisation and subsequent infection [4]. The abusive use of antimicrobials like tetracycline in food production animals leads to the selective pressure of tetracycline-resistant (Tet R ) LA-MRSA isolates [5]. The most common LA-MRSA lineage in Europe belongs to the clonal complex CC398. Tet R is one of its most outstanding features, a few atypical tetracycline-susceptible LA-MRSA strains have been detected [6], and this phenotype is commonly used for its detection [7]. LA-MRSA CC398 is related to pig farming and contact with these animals seems to be the driving force for human transmission. However, more and more often the frequent appearance of LA-MRSA cases in people without livestock contact is being reported [7].
Other LA-MRSA genetic lineages have been also related to animals and they are gaining relevance in the public health landscape [8][9][10][11]. While some of them have been also associated with pig farming (CC1, CC9 or CC97), others have been found in cattle (CC130, CC425), horses (CC8) and/or poultry (CC5) [12]. Nevertheless, frequently many of these clonal lineages have been detected in different animal hosts [12] and transmission among different livestock animals has been suggested [13].
In a previous study, a collection of 232 Tet R -MRSA isolates was obtained during a six-month period from 20 Spanish hospitals located in regions with different pig farming densities, in order to seek for potential CC398 isolates. Almost 60% of the Tet R -MRSA isolates corresponded to the lineage CC398 (n = 137 isolates), and a significant correlation was evidenced between the rate of MRSA CC398 detected in a hospital and the pig farming density in the adjoining region [14]. The non-CC398 isolates of this collection (n = 95) were not further characterised, and the unique data known were that 46% of them corresponded to the clonal complex CC1 [14]. The Sequence Type (ST) 1 belonging to CC1 lineage, producer of the Panton-Valentine leucocidin (PVL), is well recognised as a CA-MRSA in human infections, also known as USA400, and common in the USA [15]. In Europe, some cases of this PVL-positive clone have been reported, for instance in Italy [16] or Denmark [17]. Nevertheless, a different CA-MRSA ST1 subpopulation (presenting the spa-type t127) has been described in the United Kingdom, highlighting its lack of PVL production [18,19]. The spa-type t127 has been associated with LA-MRSA isolates and, when recovered from pigs, the isolates are generally IEC negative [20]. However, some studies have reported the presence of MRSA spa-type t127 in isolates of farmed animal origin (especially of bovine) with similar characteristics to the human clade, such as the presence of the genes scn, sak, sea or sat [20,21]. Several reports have highlighted the presence of MRSA-CC1-t127 isolates colonising pigs and also colonising or causing infections in people in contact with these farm animals [9,11,22].
Considering the results obtained in the previous study, in which Tet R was evidenced as a good phenotypic marker for LA-MRSA detection of the lineage CC398, the possibility of finding other LA-MRSA clonal lineages was analysed. The aim of the present study was to carry out the complete characterisation of the referred collection of the 95 non-CC398 Tet R -MRSA isolates by analysing their antimicrobial resistance and virulence phenotypes and genotypes, as well as the presence of the IEC system and their corresponding clonal lineages.

Sample Origin
According to the origin of MRSA isolates, 80% were obtained from clinical samples and the remaining 20% from epidemiological surveillance (ES) samples. Within clinical isolates, 64.5% were obtained from samples of skin and soft tissue infections (SSTI), 18.4% from respiratory tract infections (RTI), 5.3% from surgical site infections (SSI), 6.5% from urinary tract infections (UTI) and 5.3% from blood. Table 2 shows the distribution of the different sample origins according to the spa-types detected in the study.

Presence of the IEC and PVL Toxin
The IEC system (scn-positive) was present in 69/95 (73%) MRSA isolates analysed (Tables 1 and 2). The CC1, CC5, CC8 and CC30 lineages contained both IEC-positive and IEC-negative isolates. However, all CC45, CC80, CC88 and CC121 lineages included only IEC-positive isolates ( Figure 1A). Considering only the 38 isolates with the prevalent CC1-t127 lineage, 39% were scn-negative and, of them, 60% were recovered from SSTI (compared to scn-positive isolates where 34% were IPPB) ( Table 3). There was a statistically significant difference (p = 0.022) between the proportion of scn-negative isolates considering CC1 isolates and the rest of isolates included in other CCs ( Figure 2A).
Regarding macrolides and lincosamides, 72% of MRSA isolates showed resistance ( Figure 3A), with three different phenotypes detected: (a) 19 MRSA isolates (20%) showed resistance to erythromycin but not to clindamycin (ERY R -CLI S ); (b) 16 isolates (17%) exhibited the constitutive erythromycin-clindamycin resistance (ERY R -cCLI R ) phenotype; and (c) 33 isolates (35%) presented the inducible erythromycin-clindamycin resistance (ERY R -iCLI R ) phenotype. The phenotype ERY S -CLI R , very frequent among MRSA-CC398 isolates [14], was not detected among the isolates of this study. The resistance genes related to macrolide-lincosamide resistance detected in this study are shown in Table 4.   Isolates with ERY R -iCLI R phenotype carried more frequently the ermC gene (91%) compared with ERY R -cCLI R isolates (75%), but no statistically significant differences between both groups were found (p > 0.05). On the other hand, ERY R -cCLI R isolates presented a higher rate of ermB gene (44%) in comparison with ERY R -iCLI R group (6%), and, in this case, the statistical analysis showed a significant difference (p < 0.05) ( Figure 2C). The ermT and linB/vgaA genes were not detected among our ERY R and CLI R isolates, respectively.
Considering the aminoglycosides, 32% of isolates showed resistance to this group. More specifically, 14% of analysed isolates showed resistance to tobramycin alone (GEN S -TOB R ) and 18% of isolates did it in combination with gentamicin (GEN R -TOB R ) ( Figure 3A). The resistance gene found associated with the GEN S -TOB R phenotype was in all cases ant(4 )-Ia, and for GEN R -TOB R phenotype the aac(6 )-Ie-aph(2")-Ia gene ( Figure 3B).

Analysis of a Possible Relationship between MRSA CC1 and Pig Farming Density of Adjoining Region
Spearman correlations between pig densities (estimated in pigs/km 2 per each region in which hospitals are located) and different proportions of MRSA CC1 or MRSA t127 or scn-negative MRSA (respect to total number of MRSA, S. aureus, Tet R -MRSA or MRSA non-CC398) were analysed. For each pair of variables studied, a non-statistically significant association was detected (p > 0.05). Considering the population densities of each region (inhabitants/km 2 ) and the proportions of MRSA CC1 analysed before, the results were similar, with no statistically significant association (p > 0.05).

Discussion
Tetracycline has been widely used for preventing and treating farm-animal infections, contributing with its use, and potential abuse, to the increase of antimicrobial resistance worldwide. Tet R was evidenced as a good phenotypic marker for LA-MRSA detection in previous studies [7,14] In the present study, the Tet R -MRSA non-CC398 isolates obtained using this phenotypic marker were exhaustively characterised. The most prevalent clonal complex, among Tet R -MRSA non-CC398, was CC1 (46%) and the spa-type most frequently found was t127 (38 isolates). Thus, CC1-t127 isolates were identified in 14 out of the 17 participant hospitals. Considering the total of Tet R -MRSA isolates (including CC398, characterised in a previous study), CC1 was found in 19% of them [14]. CC1 isolates were identified among Tet R MRSA in other hospitals in the same range as those in our study (12-19% of Tet R -MRSA) [7,23].
Some clones related to MRSA CC1, such as the well-known PVL producer USA400, have been broadly described [24]. Among our CC1 isolates, only one CC1-t127 was PVLpositive. Several PVL-negative CC1 clones have also been identified; the Western Australia (WA) MRSA-1 clone was first discovered in Australia in the 1980s, and, since then, these isolates have been detected in several European countries [24]. This PVL-negative clone, like USA400, is not usually associated with multidrug resistance profiles [25]; however, it typically carries blaZ and ermC antimicrobial resistance genes, and other virulence genes, including scn [26]. More recently, another PVL-negative MRSA-CC1 clone (different from the WA-MRSA-1) was identified in humans in Ireland with an alarming MDR profile and mupirocin resistance as a common feature [27]; this clone has been referred to as the European CC1-MRSA-IV clone [28]. These three MRSA CC1 clones share the feature that all of them present with SCCmec type IV, despite the fact that other MRSA CC1 clones can present SCCmec type V [24], mostly among swine-related isolates [20].
In our study, we found 11 isolates (10 characterised as t127 and one as t1381) that could match the description of the WA-MRSA-1 clone (SCCmec IV, PVL-negative, ST1, blaZ-positive, ermC-positive and scn-positive), and one additional t127 isolate with the same features but a PVL producer. These potential WA-MRSA-1 isolates were mupirocinsusceptible and showed resistance to fewer families of antimicrobial agents in relation to the other three isolates (two t127 and one t2207) with features of the MDR-MRSA-CC1 clone (SCCmec IV, PVL-negative, and resistance to six-seven different groups of antimicrobial agents, including mupirocin).
Even though no correlation between pig density and the cases of MRSA CC1 or MRSA t127 or scn-negative MRSA in hospitals was found in our study (p > 0.05), we cannot discard that this result could be biased by the small number of samples collected. Further studies should be performed, and tetracycline-susceptible isolates should be considered as well.
On the other hand, CC45-t1081 (the second spa-type most prevalent in our study) corresponds with the clone known as the Berlin clone [29][30][31][32]. This clone is usually associated with a health care and nosocomial transmission (especially linked to nursing homes) [30,31]. Remarkably, tetracycline resistance was identified in more than 20% of isolates belonging to the Berlin clone in one study carried out in several European countries [33] and in most MRSA CC45 isolates (71%) obtained in one study in Poland [34]. In these studies, this phenotype was remarked for the microorganism's capacity to acquire multiple antimicrobial resistance determinants. This clone was also worryingly associated with toxigenic isolates [34].
Moreover, two of our t17235 isolates (new spa-type detected in this study) corresponded to ST2625. This ST was responsible for an outbreak in a Pediatric Intensive Care Unit in Italy in 2016 [35], and our t17235 isolates presented exactly the same resistance profile as those detected in the Italian hospital: tetM, tetL, mecA, blaZ and aac(6 )-Ie-aph(2")-Ia genes. Another clonal lineage detected was CC121 in a PVL-producer isolate with a new spa-type (t17234). This CC is becoming an emergent and hypervirulent clone, with more than 90% of CC121 (ST121) isolates carrying the PVL virulence factor [36,37].
The important virulence factor PVL was found in 16% of isolates, and they belonged to very different clonal lineages (CC1, CC8, CC30, CC80, CC88, and CC121). Interestingly, PVL was identified in three IEC-negative isolates. This virulence gene is not frequently found in LA-MRSA isolates, but unfortunately its presence in animal isolates cannot be discounted [9].
Regarding the origin of MRSA-Tet R non-CC398 isolates obtained in this study, it is important to remark that few of them were recovered from invasive infections or from urinary or surgical infections (4.2% each); as a matter of fact, most of them were from SSTI, respiratory tract infections or epidemiological samples (87.5%); this fact suggests that the dissemination routes for these tetracycline-resistant clonal lineages could be respiratory (inhalation) or contact transmission, as has also been indicated for MRSA-CC398 isolates [14]. These routes of transmission facilitate the colonisation of carriers that eventually could develop into cutaneous or respiratory infections.

Selection of Isolates
A collection of 95 Tet R -MRSA non-CC398 isolates obtained in a previous study [14] was analysed in the present work. Isolates were recovered from clinical and epidemiological samples during a six-month period (2016) in a multicentre study of 20 participant Spanish hospitals (full names of hospitals in Table 1). In particular, 17 out of the 20 hospitals contributed with Tet R -MRSA non-CC398 isolates.

Molecular Typing
The molecular characterisation of spa-types (S. aureus protein A) was performed by PCR and sequencing for all 95 Tet R -MRSA non-CC398 isolates, considering that 38 of them were typed in the previous study as t127-CC1 [18]. The spa gene sequences were analysed with Ridom ® StaphType software (version 2.2.1). The sequence types (STs) of selected isolates and those with a new spa-type were determined by multilocus sequence typing (MLST) as described in https://pubmlst.org/organisms/staphylococcus-aureus (accessed on 1 April 2021). For the remaining isolates, the clonal complex (CC) was assumed according to their spa-types. Determination of the SCCmec type was performed by PCR for isolates belonging to lineage CC1-t127 [41].

Detection of the Immune Evasion Cluster System (IEC) and the Panton-Valentine Leucocidin (PVL) Genes
The presence of the IEC system was analysed through the detection of the scn gene, present in all types of IEC and used as a marker [42]. For the detection of lukF/lukS genes, which encode the PVL toxin, all isolates were subjected to PCR technique [43].

Antimicrobial Susceptibility Testing
Susceptibility to 13 antimicrobial agents, in addition to β-lactams and tetracycline, was performed using automatic methods and/or disk diffusion tests for all isolates. The antimicrobials tested were as follows: penicillin, oxacillin/cefoxitin, tetracycline, erythromycin, clindamycin, ciprofloxacin, trimethoprim/sulfamethoxazole, vancomycin, teicoplanin, linezolid, daptomycin, fusidic acid, mupirocin, gentamicin, tobramycin and chloramphenicol. Breakpoints used for interpretation were those according to the Clinical and Laboratory Standards Institute and/or the European Committee on Antimicrobial Susceptibility Testing, depending on hospitals. When a MRSA isolate was resistant to, at least, one agent of three or more different antimicrobial families, a multidrug resistance (MDR) phenotype was associated.

Statistical Analysis
Comparison of rates of interest were completed by using Pearson's Chi-square tests or Fisher's exact tests when some sample sizes were small (below five isolates). For estimating the association between the Tet R -MRSA-CC1, Tet R -MRSA-t127 or Tet R -MRSA-noCC398scn-negative occurrence in hospitals and the pig density (pigs/km 2 ) in adjoining regions, Spearman rank correlations were performed.
All statistical analyses were made with the RStudio program (version 1.1.453) (p < 0.05 was considered statistically significant).

Conclusions
Tetracycline resistance is a good marker for the detection of MRSA CC398 isolates. Nevertheless, other clonal lineages related to livestock animals could also be detected using this marker; in this sense, scn-negative MRSA isolates belonging to CC1, CC5, CC8 and CC30 lineages, of potential animal origin, were frequently detected among non-CC398 Tet R -MRSA isolates (27.4%). The knowledge of other Tet R -MRSA genetic lineages is highly relevant, since most of them were MDR and some of them presented important virulence factors. Moreover, other human CC1 subpopulations (scn-positive) as well as the CC45 lineage (associated with facility centres/hospitals) were also detected among Tet R -MRSA isolates. The use of tetracycline resistance for detection, not only of CC398, but of other LA-MRSA lineages and other relevant human lineages should be tracked in the future.