Osteomyelitis (OM) is an infective and inflammatory process of the bone, which can progress to osteonecrosis and bone destruction. Treatment of OM is challenging due to a variety of factors including the scarce penetration of antibiotics into the bone tissue, the biofilm-like presentation of the infection, and the increasing antibiotic resistance of bacterial pathogens. Despite the improvement in the diagnosing of OM, the incidence of OM has been increasing over the last few decades. This could be ascribed to the increase of certain risk factors such as diabetes and surgery procedures such as arthroplasty (Kurtz et al., 2012; Geraghty and LaPorta, 2019). The number of orthopaedic surgeries is constantly increasing. In the United States (US), the average orthopaedic surgical volume increased of 38% in the last 3 years (Shukla et al., 2021) and is estimated that by 2030, 572,000 hip revisions and 3.48 million knee revisions could be performed with an increase of 174 and 673%, respectively (Kurtz et al., 2012). In line with the increase of arthroplasty, the incidence of OM increased from 11.4 to 24.2 cases per 100,000 person-years in four decades in United States (Kremers et al., 2015). If a prosthetic-joint infection develops, revision surgery for removal of the implant is often required, which leads to prolonged hospitalization and a high risk of re-infection (Trampuz and Zimmerli, 2008).

S. aureus is the principal cause of OM, being responsible for 30–60% of cases (Karwowska et al., 1998; Grammatico et al., 2008; Byren et al., 2009; Howard-Jones and Isaacs, 2013). OM caused by S. aureus is associated with severe prognosis and persistent infections in approximately 40% of patients (Kremers et al., 2015). During bone infection, S. aureus attachment and colonization are facilitated by adhesins. The colonization of bone occurs through direct interaction of S. aureus with the bone cells or the extracellular matrix (ECM). After bone colonization, S. aureus can grow as biofilm and produce toxins, which facilitate persistence and dissemination of the infection in the host, respectively (Patti et al., 1994; Tuchscherr et al., 2019). The success of S. aureus in bone infections is related to a multitude of virulence factors involved in critical steps of the pathogenesis, including adhesion to surfaces, invasion of host tissues, evasion of the immune system, and biofilm formation (Ricciardi et al., 2018).

S. aureus strains from different sources carry various combinations of virulence genes (Montanaro et al., 1999; Peacock et al., 2002; Campoccia et al., 2008), and some studies were conducted to identify and characterize the most frequent S. aureus clones cause of OM (reviewed by Pimentel de Araujo et al., 2021). Pandemic lineages, including CC5, CC8, CC22, CC30, and CC45, were the most common in OM, and their distribution greatly differed among the countries reflecting the local epidemiology of S. aureus and the MSSA heterogeneity (Pimentel de Araujo et al., 2021). However, the characterization of S. aureus clones causing bone and joint infections in Italy is limited to only few studies (Campoccia et al., 2008; Montanaro et al., 2016).

It is unknown to date if the ability to cause OM of some lineages is correlated with the presence of certain combination of virulence factors, or it is only a reflection of the local epidemiology of clones. The purpose of this study was to identify the pattern of virulence determinants and the clonal distribution of S. aureus isolates from OM in Italy, by using whole-genome sequencing (WGS) for a detailed genetic characterization.

S. aureus is the most frequent pathogen causing OM, and a wide range of virulence factors involved in adhesion, host cells damage, and evasion of the immune system are likely to account for its success in OM. Given the multifactorial nature of S. aureus pathogenesis, this study was aimed at understanding whether S. aureus isolates from OM were characterized by (i) distinctive epidemiological traits and (ii) definite repertoire(s) of virulence-related genes. To this purpose, the clonal characteristics and the virulence determinants inferred from WGS data of 83 S. aureus isolates responsible for OM in Italy were investigated. The most frequent CCs were CC22, followed by CC5, CC8 and CC30, which are among the prevalent lineages in Italian nosocomial settings (Giufrè et al., 2017). The phylogenetic relationship based on cgMLST revealed a non-uniform distribution of the isolates that clustered together according to the ST-CC grouping, regardless of the geographic origin or methicillin resistance status. Close genetic relatedness between isolates from either the same or different hospitals was rarely observed, suggesting limited transmission of S. aureus strains causing OM within the same hospital and between hospitals in the same geographical area.

The S. aureus lineages identified in the present study are largely consistent with those responsible for bone and joint infections worldwide (Pimentel de Araujo et al., 2021), and essentially match the most frequent lineages causing both invasive and non-invasive infections in Italian hospitals (Gagliotti et al., 2012; Grundmann et al., 2014; Giufrè et al., 2017). Although ST8/CC8 is the most common clone involved in OM in several countries (Luedicke et al., 2010; Senneville et al., 2014; Gaviria-Agudelo et al., 2015), the most common lineages in Italy were CC22 and CC5. MSSA isolates showed polyclonality with CC30 being the most represented CC, as previously reported in OM isolates (Post et al., 2014; Valour et al., 2014; Montanaro et al., 2016). MRSA clones belonging to the CC1 and CC8 lineages carried a high number of resistance determinants. Conversely, isolates belonging to CC22, which is one of the most common MRSA lineages in hospital settings in Italy, carried a limited number of resistant genes.

Studies attempting to characterize the virulence of OM strains identified some genetic traits that could be associated with a greater propensity to cause bone infections, such as the cna and bbp adhesins genes (Post et al., 2014; Montanaro et al., 2016). In our collection cna and bbp gene were present in 48.2 and 77.1% of the isolates, respectively. In particular, isolates belonging to CC22, the prevalent clone in our collection, showed a wide pattern of adhesin genes, including those coding for collagen adhesin (Cna), bone sialoprotein-binding protein (BbP), and fibronectin binding protein A (FnbA). However, it is worth noticing that each lineage expressed a different combination of adhesin genes, likely reflecting functional redundancy. Previous studies showed how these adhesin genes were frequently associated with invasive infections including bone infections (Palmqvist et al., 2005; Rieg et al., 2013; Foster et al., 2014; Post et al., 2014). Cna has previously been identified in few S. aureus clones (Peacock et al., 2002) and has been shown to play an important role in bone tropism in a murine model of haematogenous OM and in the development of septic arthritis (Patti et al., 1994; Elasri et al., 2002). Similarly, BbP is involved in bone tropism and in the initial adhesion to osteoblasts due to its ability to bind bone sialoprotein and fibrogen (Tung et al., 2000; Testoni et al., 2011). FnbA and FnbB have frequently been detected in clinical S. aureus isolates (Peacock et al., 2002). FnbA mediates the internalization of S. aureus into osteoblasts and adhesion to implant materials enhancing the severity of bone infection and its chronicity (Ahmed et al., 2001; Palmqvist et al., 2005; Testoni et al., 2011; Gries et al., 2020).

After the initial adhesion of the bacteria to bone tissue, the subsequent infection step is biofilm formation. The locus icaACD involved in the production of the polysaccharide intercellular adhesin (PIA), which promotes the aggregation of the bacterial cells in biofilm (Heilmann et al., 1996), was detected in over 95% of the S. aureus isolates examined in our study. Indeed, the presence of the ica locus is very frequent among strains from biofilm-related infection (90%) and less frequent in isolates from carriers (43%; VictoriaMartín-López et al., 2002). Capsular genes were detected in all the isolates, and cap5 was the most frequent irrespective of the CC. Other important virulence factors, such as the PVL toxin genes, which are a hallmark of community-acquired MRSA (Otto, 2013), were rarely detected, being present in only two CC30 strains. This is probably because OM patients in our study were adults, while PVL genes were more frequently detected in S. aureus from paediatric OM (Kechrid et al., 2011; Bouras et al., 2018), and only occasionally in isolates from adults (Senneville et al., 2014; Valour et al., 2014). Other leucocidin genes, such as lukX and lukY, and haemolysin genes aur, hla, hlII and hlgA/B/C, were ubiquitous among the strains. The enterotoxins cluster seg-sei-sem-sen-seo, which has been documented to increase the commensal fitness of S. aureus (Nowrouzian et al., 2015), was detected in the most frequent lineages CC5, CC22, and CC30. The expression of several virulence genes such as alpha-toxin (hla), gamma-hemolysin (hlg), leucocidins (lukX and lukY), and adhesins is under the control of the agr system (Cheung et al., 2011; Jenul and Horswill, 2019), which has been directly associated with the pathogenesis of OM (Gillaspy et al., 1995). The agr system was detected in nearly 90% of the isolates. In this study, the higher frequency of isolates carrying agr I and agrII, compared with agrIII and agrIV, is in accordance with previous reports with other collections of isolates from bone infections (Montanaro et al., 2010; Kawamura et al., 2011).

Studies on the association between S. aureus genetic traits and clinical manifestations are challenging and often generated contradictory results. Some studies support that certain S. aureus lineages harboring specific sets of virulence genes are more successful than others in causing invasive disease (Rasmussen et al., 2013; Tasse et al., 2018). However, the limit of these studies is the consensus repertoire of virulence genes that are shared within the lineage. Thus, the association of virulence factors and disease can be biased by an uneven distribution of CC between the group of isolates investigated (hitchhiker effect; Lindsay et al., 2006). Strains causing OM show substantial heterogeneity of virulence factors as it is for isolates causing colonization. A recent study showed that the commensal nasal isolates shared the same CC and genetic determinants with isolates from joint infection, suggesting that commensal S. aureus clones can cause bone and joint infections (Wildeman et al., 2020). Similarly, there is no evidence that a particular lineage or a single virulence factor or a combination of factors were distinctive of isolates from bone and implant infections (Luedicke et al., 2010) or from invasive infections (van Belkum et al., 2009). Although belonging to several lineages and being characterized by heterogeneous virulence profiles, strains isolated from carriers can become invasive under certain circumstances (Mehraj et al., 2016; Deinhardt-Emmer et al., 2018). Given that S. aureus clones causing OM originate from commensal clones which are characterized by an uneven distribution of virulence genes, it is not surprising that S. aureus strains from our collection show a broadly diversified pattern of virulence-related traits.

In summary, this is the first Italian study providing a genome-level characterization of a large collection of S. aureus isolates from bone infections. The prevalence of the different CCs in OM follows the epidemiology of S. aureus infections in Italy. It is conceivable that the prevalence of certain S. aureus lineages in OM is due to their more frequent circulation among patients rather than to a particular pattern of associated virulence factors. Understanding the virulence of S. aureus and the consequent infection tropism(s) continues to be a challenging topic for the scientific community. A better understanding of this aspect could be useful for the development of prevention and treatment strategies for OM.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ncbi.nlm.nih.gov/, PRJNA784720.

The studies involving human participants were reviewed and approved by the Ethics committee of Istituto Superiore di Sanità no. 0013802 of 18 April 2019. Patients were required to sign an informed consent that included the acceptance of the collection and analysis of clinical and microbiological data for epidemiologic and scientific purposes. Patients’ data were anonymized.

FPA, AP, and PV conceived the study. AP, PV, MM, and MDG contributed to the design of the study. SA, DL, TC, RG, and ER provided the isolates and patients’ data. FPA performed the experiments and drafted the manuscript. FPA, MP, and MDG performed the WGS. FPA and MP analysed the data. FPA, MP, AP, PV, MDG, and MM revised the manuscript. All authors read and approved the final version of the manuscript.

This work was supported in part by the Excellence Departments grant (art. 1, commi 314–337 Legge 232/2016) to the Department of Science, Roma Tre University, and grant PRIN 2017 (Prot. 20177J5Y3P) to PV, both from the Italian Ministry of Education, University and Research (MIUR), and in part by the Italian Ministry of Health, Centro Controllo Malattie (CCM), 2019 project “Sostegno alla Sorveglianza delle Infezioni correlate all’assistenza anche a supporto del PNCAR.”

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.