Chronic wounds correspond to those that do not progress through the healing process in an orderly and timely manner (Frykberg and Banks, 2015). The definition of time without complete or partial healing differs across countries, ranging from 4weeks to 3months (Järbrink et al., 2016). They include diabetic foot ulcers (DFU), venous leg ulcers, and pressure ulcers. Among them, DFU are one of the main complications in diabetic patients, with a lifetime prevalence varying between 15 and 25% (Armstrong et al., 2017). These wounds are frequently infected (more than 50%) and the infection that spreads to underlying soft tissues and bone structures is responsible for numerous lower-limb amputations (Senneville et al., 2020). Moreover, the consequences of DFU in terms of mortality, morbidity, and diabetes-related hospital admissions represent an important cost for society (Lipsky et al., 2020).

Several studies have shown that diabetic foot infections (DFI) are polymicrobial (Liu et al., 2020); however, Staphylococcus aureus is the most frequently isolated pathogen (Dunyach-Remy et al., 2016), mainly in Occidental countries. This species is found either alone or in combination, regardless of the depth of the infection. Staphylococcus aureus is both a commensal bacterium and an opportunistic human pathogen. Indeed, although approximately 30% of the human population is colonized with this species (Wertheim et al., 2005), it can also cause a wide range of clinical infections (e.g., bacteremia, endocarditis, skin and soft tissue, osteoarticular, pulmonary, and device-related infections; Tong et al., 2015). The numerous virulence factors and toxins produced by S. aureus are well characterized and responsible for the different clinical situations (Serra et al., 2015). One of the main problems in the management of DFI is the misuse of antibiotics. This is particularly frequent due to the mitigated clinical signs of ulcer infection. The unnecessary antimicrobial treatment has an impact on the bacterial microbiota and contributes to the selection and emergence of multidrug-resistant bacteria. Thus, the prevalence of methicillin-resistant S. aureus (MRSA) in DFI is high (Dunyach-Remy et al., 2016). Different risk factors for MRSA acquisition have been identified including prior hospitalization, living in a chronic care facility, previous antibiotic use, and previous amputation (Tentolouris et al., 2006; Ertugrul et al., 2012).

DFU colonization/infection by S. aureus is clearly described and S. aureus is known to hide and persist in skin cells and osteoblasts, thereby contributing to wound chronicity, yet very few data are available about the persistence rate of the species in this chronic infection. Longitudinal studies on persisting S. aureus strains in this context are lacking (Huitema et al., 2020). This contrasts with other chronic infections, the most well-known being lung infections in cystic fibrosis (CF) patients. In CF, longitudinal studies showed that S. aureus isolates belonging to the same genetic lineage are repeatedly recovered, indicating chronic infection with a single strain, and revealed adaptive modifications of the species during persistence under diverse biotic and abiotic selective pressures (Branger et al., 1996; Vu-Thien et al., 2010; McAdam et al., 2011).

This prompted us to investigate S. aureus persistence and adaptive ability during DFI. Indeed, during chronic DFU infection, S. aureus is also subjected to numerous selective pressures resulting from different antibiotic regimens, the host immune system, and the environment of the wound in uncontrolled diabetic patients and the presence of multiple microorganisms with colonizing and/or infecting potential, which interact with one another (Pouget et al., 2020). Here, we determined the rate and characteristics of S. aureus persistence in DFU based on the identification of the same lineage of S. aureus during a longitudinal study of DFI, and we performed a comparative genome analysis of sequential S. aureus strains isolated in the same patient to investigate the in vivo adaptative evolution.

Staphylococcus aureus represents the main pathogen isolated in DFI in Occidental countries (Dunyach-Remy et al., 2016). To date, no longitudinal study has evaluated the ability of this species to persist in DFU over long periods of time, yet the knowledge of this persistence is crucial to understand the bacterial adaptation in this specific environment. Among our patients, only 12 (25%) presented a persistent DFU colonization by a related strain during a period ≥4weeks with a median persistence of 12weeks. Only one patient presented successive S. aureus strains belonging to the same clonal lineage over time for an extended period exceeding 30weeks (patient 20). These results suggested that long-term persistence of S. aureus in DFI after initial colonization is a rare finding in DFI (2.1% of the patients in our study), a low implantation rate compared with other chronic conditions like lung colonization in CF patients (Kahl et al., 2003). We assume that the dynamics of S. aureus implantation in DFU differ from the dynamics of bacterial colonization and adaptation in the lower respiratory tract during CF due to distinct environmental conditions and infection management (Kahl et al., 2003; Vandenesch et al., 2012). Our results support that DFI caused by S. aureus is mainly recurrent due to iterative reinfection. This could be explained by the efficiency of the debridement performed in our center by a trained orthopedist and of the antibiotics. This observation is reinforced by the high diversity of S. aureus clonal lineages identified in DFI at inclusion and during the follow-up.

Among S. aureus clonal lineages in the chronic wounds of diabetic patients, MRSA was identified in 14.5% of the patients at inclusion. This observation could be explained by the inclusion criteria because only patients with a first episode of the DFU were included in our studies. MRSA persisted in wounds for short periods of time only (mean of 4weeks) and was mainly replaced by MSSA clones. More patients (9, 18.8%) had MRSA during their follow-up and one had the same CC8-MRSA-IV Lyon clone identified in three specimens sampled over a 7-week period (patient 11). This patient had a Grade 4 DFI and was hospitalized for 3weeks, possibly explaining this persistence. However, due to both their low representation and implantation rate in DFU, the potential of virulence of these MRSA strains remains contentious. Similarly, the fact that MRSA could be less adaptable to the DFU environment and/or more “sensitive” to the debridement due to a more superficial localization in the wound remains on open hypothesis.

Interestingly, the comparison of the distribution of virulence genes between clonal/persistent and non-clonally-related/sporadic strains isolated at different periods in the same patients showed a significant difference concerning tst and agrIII genes (Supplementary Table S2). The absence of the tst gene but also of other major toxinogenic factor-encoding genes (e.g., lukF/S-PV, etA, etB, edinA, and edinC) and the low prevalence of etD and edinB genes in persistent S. aureus suggests that these toxinogenic strains do not have the potential to adapt to the DFU environment. They have high virulence and invasiveness and the host immune response is more activated to eliminate these pathogens (Horn and Kielian, 2021). Recently, we also observed, in an in vitro model mimicking the DFI environment, that the strains harboring lukF/S-PV and edin genes were particularly affected by this condition (Pouget et al., 2020). A significant reduction of bacterial virulence can be observed, thus, the immune system could more easily eliminate the toxinogenic strains.

Comparative gene contents of S. aureus isolates obtained sequentially over several weeks revealed that the majority of strains (58.3%) showed no modification over time, but among the virulence factor-encoding genes that were lost in persistent isolates, two factors associated with mediation of immune evasion of S. aureus in humans (scn and sak) were lost by three strains. These factors can specifically modulate the human innate immune system, and are considered as mechanisms of S. aureus adaptation to the human host (Senghore et al., 2016). Interestingly, the median time to the loss of virulence genes was high [21.5weeks (range 9–52)]. In DFI, the intrahost adaptation is a continual process, where the decrease of virulence could be a key point to limit the host immune response and favor the chronicity of the lesions, as previously suggested (Pouget et al., 2020).

Concerning the resistant gene contents, we observed a similar evolution with four and three isolates having either loss or gain of resistance-encoding genes, respectively, and two isolates having both gain and loss of these genes (Table 2). In the cases of gene loss, qacA, tetM, and ermC were the most prevalent. These genes encode for resistance to antiseptics, tetracyclin, and macrolides, respectively. They are associated on mobile genetic elements and are easily acquired or lost by bacteria (Anitha et al., 2016). In the five patients (no. 11, 20, 32, 33, and 45) harboring these S. aureus showing evolution during DFI, none of the antibiotics or antiseptics used could explain the loss of these genes (data collected in the three previous studies and cross-referenced with the results of this study). In the cases of gene gain, the acquisition of β-lactamase-encoding genes was the most commonly observed. The prescription of amoxicillin/clavulanic acid during the management of these ulcers in patient 20 and 41 could have favored the resistance selection. Although, the blaZ gene was particularly frequent in our population and qacA present regardless of the period of the isolation of the S. aureus strains, tetM and ermC were rarely identified, but in all cases these genes were present in strains isolated at inclusion and lost in the follow-up strains, suggesting that they could be more susceptible than others to be lost during the adaptation of S. aureus to the DFU environment.

To gain insight into the adaptive ability of S. aureus in DFU, we compared the genomes of clonal S. aureus isolated from the same wound after 48weeks. No modification in the genome size was observed, suggesting that the bacterial adaptation in this hostile environment does not need the reduction of the genome, as previously observed for Escherichia coli (Lienard et al., 2021). However, some genome modifications were noted concerning important virulence factor-encoding genes. Indeed, four genes were affected by a stop codon affecting two main bacterial functions: the bacterial defense (bepA and oatA) and the two-component systems (TCS) SsrA-SsrB and VraS/VraR (ssrB and vraR). Beta-barrel assembly-enhancing protease (BepA) acts both as a chaperone and a metalloprotease. BepA maintains the integrity of the outer membrane both by promoting assembly of the outer membrane proteins (OMPs) and by degrading misfolded OMPs (Lütticke et al., 2012; Narita et al., 2013; Shahrizal et al., 2019). O-acetyltransferase (OatA) is an O-acetylating enzyme present in Gram-positive bacteria (Jones et al., 2020). O-acetylation of the peptidoglycan protects bacteria from the lytic activity of lysozyme, an enzyme integral for one pathway of innate immune response inside phagocytic cells (Bera et al., 2005). Different bacterial enzymes can modify the essential cell wall polymer peptidoglycan by O-acetylation, thus representing an important bacterial virulence factor. Moreover, S. aureus possess different members of the TCS, which is involved in the global regulation of multiple targets that determine virulence, stress tolerance, and persistence in these bacteria (Shankar et al., 2015). Among those, the TCS VraS/VraR has been shown to be involved in the control of the cell wall peptidoglycan biosynthesis. This system inhibits the host autophagic flux and delays the early stage of autophagosome formation, thereby promoting bacterial survival. It facilitates the ability of S. aureus to resist host polymorphonuclear leukocytes-mediated phagocytosis and killing, thus contributing to immune evasion (Belcheva and Golemi-Kotra, 2008). Mutations in this VraSR system are involved in increasing resistance to antibiotics (Galbusera et al., 2011; McCallum et al., 2011; Qureshi et al., 2014). Finally, SrrA/B is also a member of the TCS, which regulates a type III secretion system and is thus involved in the global regulation of staphylococcal virulence factor production. This TCS regulates different genes involved in bacterial survival under modified environmental conditions such as anaerobic condition, oxidative environment, cytochrome biosynthesis, and assembly, in biofilm formation and programmed cell death. Recently, SrrA/B has been identified to play a role in host-derived nitric oxide resistance (Yarwood et al., 2001; Tiwari et al., 2020). Altogether, these four mutations indicate that during long-term intrahost evolution, S. aureus decreased its potential to adapt to the stress environment and its ability to hijack the immune system through virulence attenuation. This could explain the low persistence of S. aureus in DFU and a “preference” of the bacteria for an intracellular life (Rasigade et al., 2016). However, we could not exclude the influence of the “environment” encountered by S. aureus in this chronic wound. If we previously highlighted that the diabetic foot medium modifies the virulence of S. aureus (Pouget et al., 2021), we also know that the competition between microorganisms in these polymicrobial infections directly impacts the invasiveness of S. aureus (Pouget et al., 2020). The importance of the bacterial adaptation could also be observed by the formation of small colony variants (Tuchscherr et al., 2020), as previously observed in isolates from CF patients (Schwerdt et al., 2018) and DFU (Cervantes-Garcia et al., 2015). All these multifactorial aspects influence the behavior of this pathogen and could explain the modulation of the S. aureus virulence. Further work on the transcriptomic profile of these persistent isolates must be developed to evaluate the regulation of the different pathways involved in this long-term evolution.

This study is the first longitudinal one in DFI; it highlights a low prevalence of long-term persisting S. aureus infection observed in 2% of the patients. This work provides the basis for the understanding of S. aureus colonization dynamics in chronic wounds.

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 in the article/Supplementary Material.

The studies involving human participants were reviewed and approved by CPP sud Mediterranée III. The patients/participants provided their written informed consent to participate in this study.

AS and J-PL: conceptualization, resources, and project administration. MH, CD-R, AY-M, BS, AS, and J-PL: methodology. MH and AY-M: software. J-PL, MH, HM, and AS: validation. J-PL, MH, CD-R, HM, and J-PL: formal analysis. J-PL, MH, CD-R, AB-D, and AS: investigation. AB-D, SS, NC, CD-R, VM, and J-PL: data curation. AS, HM, and J-PL: writing – original draft preparation. MH, CD-R, AB-D, SS, NC, AY-M, VM, and BS: writing – review and editing. HM, AS, CD-R, BS, and J-PL: visualization. AS and J-PL: supervision. AS, BS, and J-PL: funding acquisition. All authors contributed to the article and approved the submitted version.

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.