The pathogenicity and virulence of the opportunistic pathogen Staphylococcus epidermidis

ABSTRACT

The pervasive presence of Staphylococcus epidermidis and other coagulase-negative staphylococci on the skin and mucous membranes has long underpinned a casual disregard for the infection risk that these organisms pose to vulnerable patients in healthcare settings. Prior to the recognition of biofilm as an important virulence determinant in S. epidermidis, isolation of this microorganism in diagnostic specimens was often overlooked as clinically insignificant with potential delays in diagnosis and onset of appropriate treatment, contributing to the establishment of chronic infection and increased morbidity or mortality. While impressive progress has been made in our understanding of biofilm mechanisms in this important opportunistic pathogen, research into other virulence determinants has lagged S. aureus. In this review, the broader virulence potential of S. epidermidis including biofilm, toxins, proteases, immune evasion strategies and antibiotic resistance mechanisms is surveyed, together with current and future approaches for improved therapeutic interventions.

KEYWORDS:

• Staphylococcus epidermidis
• infection
• virulence
• biofilm
• host–pathogen interactions
• therapeutic approaches

Introduction

Colonizing up to 80% of healthy humans,¹ the frequent presence of Staphylococcus epidermidis and other coagulase-negative staphylococci (CoNS) on human skin demonstrates both its resilience to environmental extremes and capacity for host immune modulation,^2,3 allowing it to evade detection by host defences. The skin can be a harsh environment for microorganisms, due to environmental factors such as variations in pH, temperature, and low moisture levels. Human epithelial surfaces also have multifaceted defence mechanisms mediated by the innate immune system, including production of a diverse array of antimicrobial peptides, sebum production, and the shedding of keratinocytes, all of which collectively inhibit colonization and infection by pathogenic organisms. Yet paradoxically, the skin still needs bacteria: formation of a healthy skin microflora can help prevent skin dysfunction, particularly well documented in the case of atopic dermatitis but also playing a significant role in stabilizing the microbial communities of healthy skin.^4,5 Commensal inhabitants must balance strategies to survive desiccation, acid stress, and temperature fluctuations against the risks of activating the host immune system.

Due to its ubiquity, isolation of S. epidermidis in diagnostic samples taken during clinical infections was often dismissed as accidental contamination,⁶ before its emergence as an important pathogen in infections associated with implanted medical devices (Figure 1). Moreover, this pathogen is increasingly recognized as the sole causative factor in serious infections ranging from bacteraemia,⁷ sepsis,^8,9 endocarditis, meningitis, and toxic shock syndrome,¹⁰ as well as superficial skin infections, ocular infections^11,12 and the aforementioned infections associated with indwelling medical devices including shunts,¹³ catheters¹⁴ and prosthetic joints.¹⁵

Figure 1. Medical interventions including surgical wounds and various implanted medical devices associated with increased risk of S. epidermidis and other CoNS opportunistic infections. Created with Biorender.com.

The diverse array of infection vectors, including medical interventions and the patient’s own microflora, serve to emphasize the importance of S. epidermidis, through its ubiquity, as a challenging pathogen to eradicate.¹⁶ Moreover, to eliminate this commensal from epithelial surfaces may do more harm than good in the majority of cases, and this opportunistic pathogen continues to pose a significant health risk to vulnerable patients.¹⁷

Any discussion about S. epidermidis as a pathogen is invariably qualified by the comment that it lacks the diverse array of virulence factors found in S. aureus. Nevertheless, although this is true, we argue that it is important to consider what aspects of S. epidermidis pathogenicity are being overlooked when this microorganism is primarily categorized as a “common commensal” and not a ubiquitous opportunistic pathogen.

This review summarizes the pathogenicity of S. epidermidis infections including the molecular basis of biofilm-related and non-biofilm virulence determinants, as well as current treatment strategies and future directions for research on this remarkably persistent pathogen.

Establishment of infection

As a frequent constituent of the human microbiome, S. epidermidis is generally considered to lead a commensal lifestyle on epithelial surfaces, such as the skin, nares, and the mucosa. However, S. epidermidis exploits breaches in the host’s cutaneous barriers, both through unintentional injury or surgical procedures in order to gain access to deeper tissues. Normally, the immune system is able to clear low-level infiltration by S. epidermidis cells, but immunocompromised individuals are susceptible to a spectrum of infections, from localized tissue inflammation to severe, systemic disease. Early establishment of infection involves bacterial adherence to surfaces, facilitated by microbial adhesins. This initial adhesion is a critical step, allowing the bacterium to evade clearance and establish a foothold within the host. Following successful adhesion, biofilm formation can begin (Figure 2).

Figure 2. Graphic illustration of the biofilm life cycle in Staphylococcus epidermidis including depictions of important mediators of attachment aggregation, maturation and dispersal stages of biofilm are depicted. The attachment phase is characterised by the deposition of planktonic cells onto a biotic or abiotic surface. This phase is mediated by interactions between host ligands present on the surface, including fibrinogen (Fg) and fibronectin (Fn). Host proteins are then bound by bacterial surface adhesins, such as SdrG and microbial surface components recognizing adhesive matrix molecules (MSCRAMMs), which allow the bacterial cells to remain attached. The accumulation phase results in microcolonies of bacterial cells that self-associate using electrostatic forces and protein adhesins such as accumulation associated protein (Aap). Negatively charged extracellular DNA (eDNA) also surrounds the cells, attracted to the positively charged cell surface and resulting in a “sticky” mesh. A mature biofilm forms when polysaccharide intracellular adhesin (PIA) is excreted by S. epidermidis, forming profuse multi-layered structures. The biofilm matrix shields the bacteria from immune system cells, such as neutrophils, and provides protection from harsh environmental conditions such as desiccation, pH stress and the presence of antibiotics. Channels in the biofilm allow perfusion of nutrients. Within the mature biofilm, a significant proportion of cells remain metabolically dormant, rendering them tolerant to antibiotics that target metabolic processes. The dispersal of biofilm in S. epidermidis is co-ordinated by the action of the agr quorum sensing system, which secretes an autoinducing peptide when cell numbers increase to a critical level. In response to this, transcription of the ica operon, which governs the production of PIA, is reduced and individual cells are released from the biofilm structure. These planktonic cells are then free to disseminate in the host again.

Common vectors of infection

Intravascular catheters, particularly central venous catheters and peripherally inserted central catheters, stand out as major sources of S. epidermidis infections. These devices provide an ideal substrate for bacterial adhesion and biofilm formation, leading to catheter-related bloodstream infections. More generally, CoNS are leading causative agents of bloodstream infections, representing up to 31% of reported cases.^18–21 Prosthetic devices, including joint implants and cardiac devices, also represent significant sources of infection, causing chronic and difficult-to-treat biofilm-related infections. Periprosthetic joint infection (PJI) isolates are significantly more likely to be resistant to one or more antibiotics when compared to commensal isolates, and 44% of PJI isolates are from ST2, a sequence type associated with hospital-acquired highly resistant strains.²² Infection of cardiac devices such as pacemakers,²³ as well as prosthetic heart valves^{24, 25} can necessitate removal of the device, or in refractive cases due to highly resistant strains, heart transplantation as salvage therapy.^{26, 26} In one study of bloodstream infections originating from cardiac assist devices, S. epidermidis was identified as the most common causative pathogen.²⁷ In orthopedic S. epidermidis infections, which represent up to 43% of orthopedic device-related infections, aminoglycoside resistance, and strong biofilm-forming capabilities are associated with worsened patient outcomes.²⁸

S. epidermidis is the causative agent in up to 60% of ventriculoperitoneal (VP) shunt infections,²⁹ which are particularly challenging to treat due to poor antimicrobial drug penetration into cerebrospinal fluid (CSF). Post-operative infections of implanted VP shunts, which can affect up to 16% of patients, higher in neonates^29–32 is accompanied by a mortality rate at 10%.²⁹

Up to 31% of surgical site infections following spinal surgery are caused by S. epidermidis, and whilst S. aureus is a more frequent cause of these spinal surgery infections, methicillin resistant S. epidermidis (MRSE) was significantly more abundant than methicillin resistant S. aureus (MRSA) in spinal infections, which can limit treatment options.³³ S. epidermidis was identified as the most common cause of latent surgical site infections after spinal fusion surgery, characterized by the slow development of clinical symptoms often mistaken for post-surgical back pain, which allowed the bacteria to become established at the surgical site and cause deeper infection, particularly inflammation of the vertebrae (spondylitis).³⁴ Intracranial infection with MRSE may be associated with poor prognosis, even following vancomycin treatment.³⁵ Isolates collected from infected sutures from a variety of surgical cases also implicated S. epidermidis as the most common pathogen.³⁶ Surgical tools can also become contaminated with skin flora from either the patient or from surgical personnel during surgery, with S. epidermidis being the most commonly isolated species from laparotomy instruments,³⁷ representing a vector of infection even when strict hygiene and aseptic technique are practiced.

Other vectors of implant-associated S. epidermidis infection include, but are not limited to, intrathecal pumps,³⁸ breast prostheses and implants,^39–41 intraocular lenses,^{42, 43} surgical mesh used for hernia repair,^44,45 and bone-anchored hearing aids.⁴⁶ Despite these risks, rigorous infection control measures, such as pre-operative decolonization, surgical site disinfection, and intra-operative infection control procedures can help prevent bacteria from breaching the skin during surgical procedures.^47–49

Chronic versus acute infection

S. epidermidis is typically responsible for chronic infections involving biofilm, rather than acute infections. However, examples of acute S. epidermidis infections include those associated with ocular implants^{43, 50} and knee arthroplasty.⁵¹ For the latter, poor response to treatment compared to non-staphylococcal infections can also lead to chronic infections.^52–55 Prevention, or timely diagnosis and treatment of S. epidermidis infections is therefore of particular importance. The release of bacterial cells from biofilm into the bloodstream is noted as the primary mechanism of CoNS-related sepsis.⁵⁶ Sepsis, a serious condition in which host immune response to bacteria in the bloodstream leads to systemic inflammation and potentially escalates to severe sepsis or septic shock, can in the case of S. epidermidis infection be viewed as an acute host response to a chronic bacterial infection.⁵⁷

Biofilm – the primary virulence factor

In a typical S. epidermidis infection scenario, biofilm formation begins with the transfer of bacteria from the skin onto a medical device during surgical implantation. The individual bacterial cells utilize their own surface protein adhesins to adhere to the implant, including interacting with host matrix proteins that coat the device. After adherent microcolonies are established, biofilm formation enters the accumulation phase, driven by the abundant production of polysaccharide intercellular adhesin (PIA)/poly-N-acetylglucosamine (PNAG), or protein adhesins that are covalently attached to the cell wall or associated with the cell surface. PIA forms a protective layer, enveloping the cells until mature biofilms begin to release planktonic cells, initiating a new cycle of dispersion. The role for different adhesins and biofilm mechanisms indicates that “biofilm” is not a uniform process, but a term encompassing various survival modes employed by S. epidermidis to adapt to dynamic and challenging host niches.

Biofilm formation

Several reviews and articles have comprehensively covered S. epidermidis biofilm formation, structure, regulation, and disassembly in detail.^52,58–64 Here, the role of the ica operon and the role of non-PIA biofilm factors such as extracellular DNA (eDNA) and key protein adhesins will be summarized to contextualize the interaction and importance of S. epidermidis virulence determinants with these biofilm constituents and regulators.

Role of ica operon in PIA-dependent biofilm

Any review on S. epidermidis virulence cannot overlook the essential role of the icaADBC operon-encoded PIA in biofilm formation by many S. epidermidis isolates. The capacity of S. epidermidis to form biofilm strongly correlates with the establishment of chronic infections, and the accompanying morbidity and mortality to the host. Whilst S. epidermidis is also capable of ica-independent biofilm formation, PIA-mediated biofilm presents a uniquely profuse and resilient state of existence for staphylococci. Early research on “slime-producing” isolates of S. epidermidis identified the ability of what had previously been considered a harmless skin-dwelling commensal organism to persist within the host, turning simple implantation of medical devices into a procedure carrying significant risk for opportunistic infection by a ubiquitous skin commensal microorganism.^65–72

The ica operon encodes a set of proteins responsible for PIA production, export, and function. The icaA and icaD genes encode transmembrane proteins that dimerize to form an N-acetylglucosaminyltransferase, responsible for PIA synthesis.⁷³ The IcaB enzyme deacetylates PIA, introducing a positive charge necessary for its adherence to the cell surface, with icaB mutants markedly impaired in adhesion and colonization in the host, as well as being more vulnerable to phagocytosis.⁷⁴ IcaC elongates the polysaccharide molecules, rendering them functional, whereas icaC mutants produce short, 20-residue oligomers in comparison to full-length PIA, which normally contains over 130 N-acetylglucosamine (GlcNAc) residues.⁷⁵ The icaR gene located immediately adjacent to the ica operon encodes a transcriptional repressor centrally involved in the regulatory network that controls icaADBC transcription.^76–78 Expression of icaR is in turn controlled by a complex network of proteins and transcription factors, including the alternative sigma factor σ^B and SarA, that combine to regulate the ica operon in response to metabolic, environmental, and host factors (Figure 3).⁷⁹ Furthermore, transcription of the icaADBC operon does not precisely align with PIA production, revealing the importance of post-transcriptional regulation in the biofilm phenotype.

Figure 3. Graphic illustration of major S. epidermidis regulators of growth, biofilm formation and virulence. The icaADBC locus encodes the enzymes responsible for synthesis, export and deacetylation of polysaccharide intercellular adhesin (PIA), which serves as a major mediator of biofilm. The divergently expressed icaR gene encodes the major transcriptional repressor of the ica operon. In response to environmental cues activated SrrB phosphorylates its cognate SrrA effector promoting pflA and ndrB expression and growth under low oxygen conditions. SrrA binds to both the icaA and icaR promoters to differentially regulate biofilm under oxic and microaerobic conditions. Activation of the SaeS kinase leads to phosphorylation of the SaeR response regulator increasing fibronectin (Fn) binding and cell-cell adhesion through upregulation of GehD lipase and the extracellular matrix binding protein (Embp). SaeR also upregulates activity of the serine endopeptidase Esp, negatively impacting S. aureus colonisation and promoting S. epidermidis immune evasion through proteolysis of complement proteins. ClpPX is an ATP-dependent protease and chaperone system that degrades the global regulator SarA, resulting in repression of the repressor icaR and activation of icaADBC and biofilm. Rbf downregulates expression of sarR, which encodes a repressor of the icaADBC operon. The agrD gene from the agrACDB operon encodes a pro-peptide that is exported and processed by AgrB to release the autoinducing peptide (AIP) that constitutes that major quorum sensing system in staphylococci. In response to AIP, the AgrC sensor kinase phosphorylates the AgrA response regulator, thereby leading to activation of the P2 and P3 promoters and increasing expression of the agrACDB and RNAIII loci, respectively. RNAIII is the second effector molecule of the Agr system, functioning as an antisense RNA to control translation of target genes. The cytolysin phenol soluble modulin g (PSMg, δ-haemolysin), encoded by the hld gene located within RNAIII plays a role in biofilm dispersal. The Agr system also negatively regulates expression of the clpP-encoded protease, which degrades Spx, a negative regulator of icaADBC. SarA positively influences Embp expression facilitating cell-cell adhesion. Processing of the major autolysin AtlE by SepA to generate active amidase (AM) and glucosaminidase (GL) autolytic enzymes is important for eDNA release during the early stage of biofilm development.

PIA-independent biofilm

Approximately 30% of S. epidermidis isolates from non-superficial infections lack the ica operon.⁸⁰ However, the absence of the ica operon does not preclude biofilm formation.⁸¹ PIA-independent or protein-mediated biofilm represents a second modality of biofilm formation. Protein-mediated biofilm relies on cell wall proteins that can bind to host ligands, as well as extracellular DNA (eDNA) released by autolysins that also enhances attachment and structure of biofilm. The spatial distribution of adhesins significantly influences biofilm architecture, for example, PIA encapsulates the cell, whereas protein adhesins localize to the membrane, resulting in different overall biofilm architecture which can give advantages in surviving in variant niches. PIA-producing isolates are more commonly isolated from sites exposed to shear force, e.g. the bloodstream and urinary tract, whereas PIA-independent biofilm in S. epidermidis represents the major mode of biofilm production in other prominent infection types, such as ocular infections.^{80, 82} PIA-dependent biofilm-forming strains are approximately twice as likely to cause prosthetic joint infections, compared strains capable of PIA-independent biofilm formation.⁸¹

Aap

The cell wall-anchored (CWA) proteins like accumulation-associated protein (Aap) is the best characterized mediator of PIA-independent biofilm in S. epidermidis. Aap has several domains: a signal peptide, an N-terminal A domain, comprising A-repeats and L-type lectin domains, followed by a B-repeat region containing G5-E repeats. The C-terminal end contains an LPXTG motif that is covalently anchored to the cell wall by sortase.⁸³ The A domain can be proteolytically cleaved by the metalloprotease SepA, exposing the B domain, which is then capable of self-association with other Aap B domains, to promote cell aggregation.⁸⁰

The A-domain of Aap binds to host surfaces through the lectin domain, essential for adherence to host glycans. Heterophilic sugar binding by the lectin region of the A domain is also involved in initial staphylococcal cell–cell contact early in biofilm formation, before homophilic B-repeat bonds allow for tighter binding. Homophilic interactions between adjacent cells involve the G5-E domains binding to one another. Notably, the B-repeat domains exhibit amyloid fibril formation, forming “rope-like” structures within the biofilm matrix, a process dependent on zinc equilibrium.^{80, 84} Single-molecule experiments have revealed the extraordinary mechanical strength of Aap homophilic bonds, with forces reaching up to 1,000 pN – close to the bond strength of covalent chemical bonds⁸⁵ reinforcing their role in forming highly adhesive and cohesive biofilms (Figure 2). Intriguingly, the cleavage patterns of Aap vary between strains, with a 2022 study by Wang et al. finding that strain CSF41498 has lower rates of Aap processing resulting in more intact Aap when compared to strain 1457, which has increased cleavage of Aap.⁸⁶

Extracellular matrix binding protein (EmbP)

Embp is a large, membrane bound protein involved in adhesion. Whilst EmbP plays a moderate role in adhesion to fibronectin (Fn) under low flow conditions, the “Velcro-like” binding that occurs between multiple repeats within EmbP and the repeats exposed in the fibrillated, deposited form of Fn was shown to be crucial to S. epidermidis binding to surfaces under high shear conditions, for example on an implanted intravenous catheter.⁸⁷ Embp expression is highly induced when adherent cells, such as those in a biofilm, are subjected to osmotic stress indicating that Empb may be more important for survival of S. epidermidis by promoting cell–cell adhesion under the high-salinity conditions found on the skin, rather than as a factor in invasion and adhesion during infection. However, given the role of Embp in binding to Fn, osmotic stress induced Embp expression should enhance adherence to the host extracellular matrix proteins and enhanced biofilm formation in the infection niche.⁸⁸

Biofilm associated protein (Bap)/bap homolog protein (Bhp)

The first report of a PIA-independent mechanism of biofilm formation in staphylococci identified Bap.⁸⁹ This 239-kDa sortase-anchored surface protein contributes to the initial attachment and accumulation phases of biofilm.⁹⁰ Unlike Bap in S. aureus, which is associated primarily with bovine mastitis isolates, Bap in S. epidermidis is commonly found in mastitis isolates from several veterinary species.⁹⁰ Bap consists of several repeated domains, referred to as Bap repeats, which are responsible for its adhesive properties and the formation of cell-to-cell interactions within the biofilm. Bhp is a protein with significant homology and similarity to Bap,⁹⁰ which is also found in non-animal derived isolates, including in MRSE isolates from clinical infections.⁹⁰ Bhp has an N-terminal export signal, followed by an A-domain and a region of tandem repeats that shares 45–57% identity with S. aureus Bap tandem repeats.⁹¹ Bhp expression is highest during early exponential phase and decreases in stationary phase.⁹¹ Serum reactivity to Bhp was not commonly detected in patients with S. epidermidis infections, suggesting that Bhp is only weakly immunoreactive or not highly expressed during infection, although other cell-wall proteins like SdrF exhibited a similar response.⁹¹ Found in both commensal and invasive strains,⁹² and in up to 15% of blood culture isolates,⁹³ the role of Bhp in S. epidermidis biofilm formation remains unclear, and whilst it shares considerable similarity with Bap, it is important not to over-extrapolate from the small amount of available literature on Bhp.

eDNA and cell lysis

Extracellular DNA (eDNA) is released from lysed bacterial cells, either through breakdown of dead cells or by the activity of autolysins.⁹⁴ eDNA has multiple properties that make it a valuable biofilm constituent. eDNA is associated with the initial attachment phase to abiotic surfaces.^95–97 Strands of eDNA create a scaffold that helps maintain the overall architecture and resilience of the biofilm (Figure 2).⁹⁸ eDNA also acts as a barrier against host immune defences by binding and sequestering antimicrobial peptides.^{99, 100}

S. epidermidis possesses two autolysins, the major autolysin AtlE and a secondary autolysin Aae, both having dual roles as both autolysins and adhesins. AtlE consists of a signal peptide and a propeptide, followed by amidase (AmiE) domain linked by three direct repeats (R1, R2, R3) to a glucosaminidase domain at the C-terminus. Proteolytic processing occurs by an unidentified protease, resulting in removal of the N-terminal signal peptide and the propeptide after secretion.^101,102 After further processing by the metalloprotease SepA,¹⁰¹ the amidase domain retains two of these repeats, and the glucosaminidase domain retains the third. Each repeat contains a glycine-tryptophan (GW) module. The amidase domain is anchored to the cell wall by the R1 and R2 repeats and is required to cleave peptidoglycan during cell division.^103,104 Evidence from the functionally interchangeable Atl protein in S. aureus indicates that presence of WTA on the mature cell wall but not the newly created septum discourages AtlE binding during division, encouraging enzymatic separation of daughter cells.^104,105

Binding and internalisation by host cells are mediated by the AmiE domain, which interacts with the host heat shock cognate protein Hsc70 directly and with α5β1 integrin via fibronectin bridging.^{106, 107} Vitronectin binding activity is seen with the full-length protein, and to a lesser extent with each of the catalytic domains.¹⁰² Deletion mutants of atl are defective in biofilm formation⁹⁴ and have decreased virulence in a rat catheter model.¹⁰⁸ eDNA is found in microcolonies during the initial adherence phase of biofilm accumulation. DNase I is capable of dispersing early-stage biofilms, but ineffective at removing mature biofilm.^{94, 109}

Aae is a more recently identified autolysin.¹¹⁰ The N-terminal domain of Aae consists of an LysM domain, implicated in peptidoglycan recognition and binding, made up of three direct repeats. Unlike other bacterial species where it is located at the C-terminal end, the N-terminal position of the LysM domain in staphylococci may allow Aae to associate with the cell wall in the absence of a specific cell wall anchor. Like AtlE, Aae also has adhesive properties and can bind fibrinogen, fibronectin and vitronectin in a dose-dependent manner.¹¹⁰

The link between PIA production and eDNA in biofilm matrix is complex, though the overall trend is positive, with strong PIA production being associated with increased levels of eDNA. PIA may play a role in retaining eDNA within the biofilm structure.⁹⁸ Thus, eDNA appears to play a role in both PIA-dependent and PIA-independent biofilm development.

Interestingly, several eDNA-binding proteins have been described in the matrix of S. aureus biofilms, many of which were previously not known to bind DNA,⁹⁶ forming positively charged surface anchors that stabilize the electrostatic net of eDNA in the matrix.⁹⁶ Several of these proteins have homologues in S. epidermidis, including CopL, which is associated with the COMER-like pathogenicity island found in the hospital-associated MRSE lineage ST2.¹¹¹ Furthermore, the glucosaminidase region of S. aureus Atl is also known to bind DNA.¹¹² This raises the intriguing possibility that there remain uncharacterized eDNA and protein interactions in S. epidermidis biofilms. Regulation of eDNA release is mediated through SarA, which indirectly negatively regulates AtlE by downregulating SepA, responsible for the proteolytic cleavage required for the bacteriolytic function of AtlE.¹⁰¹

A small regulatory RNA RsaE was recently identified that positively regulates both icaADBC expression and eDNA release. RsaE is expressed differentially throughout S. epidermidis biofilm, and interestingly is not present in all icaADBC-positive strains. Regulation of PIA production is mediated through RsaE interaction icaR mRNA, whilst the effect on autolysis occurs due to negative regulation the lrgA-encoded antiholin protein that prevents cell lysis by the holin CidA.¹¹³ In S. aureus, eDNA release by CidA contributes to the function of biofilm,¹¹⁴ suggesting that it may play a similar role in S. epidermidis.¹¹³ In S. epidermidis small colony variants (SCVs), imbalance between CidA and LrgA expression results in increased eDNA abundance in biofilm.¹¹⁵

Non-biofilm virulence determinants

Despite being commensals, S. epidermidis strains are capable of expressing several enzymes and toxins, which collectively appear to play a protective role underpinning the characteristic resilience of this organism.

Lipases

Lipase-mediated hydrolysis of lipids can serve as a source of nutrients for S. epidermidis. Interestingly, the glycerol ester hydrolase (geh) lipases of staphylococci have a secondary role as surface-anchored adhesins. S. epidermidis has two known geh lipases, GehD and GehC, under the control of the agr quorum sensing system.¹¹⁶ GehD is a collagen-binding MSCRAMM, which despite lacking an LPXTG motif is localized to the cell wall¹¹⁷ (Figure 3).

Superantigens and enterotoxins

Staphylococci produce multiple superantigens implicated in toxic shock syndrome (TSS) including TSS toxin 1 (TSST-1), and staphylococcal enterotoxins B and C.^{118, 119} Superantigens bind to Major Histocompatibility Complex class II (MHC II) receptors, precipitating interactions with T-cells and inciting a cascade of events leading to heightened cytokine production, marked inflammation, and the onset of TSS symptoms such as fever, hypotension, vomiting, and confusion.¹²⁰ Menstrual TSS is caused almost exclusively by TSST-1, whereas staphylococcal enterotoxin B and C as well as superantigens from group B, C, and G streptococci are also known to be involved in non-menstrual TSS.¹²¹ CoNS are generally considered to lack superantigens as they lack TSST-1, however, they may possess staphylococcal enterotoxins.¹⁰ Up to 95% of S. epidermidis isolates from blood cultures possess at least one staphylococcal enterotoxin.¹²²

Staphylococcal enterotoxins (SEs) exhibit a remarkable tolerance to both heat and pH stress, as well as protease degradation, maintaining their integrity in the harsh environment of the gastrointestinal (GI) tract¹²³. Staphylococcal food poisoning (SFP) from SEs is accompanied by a rapid onset of symptoms, including nausea, emesis, abdominal pain, diarrhea and GI tract inflammation and damage.¹²³ Whilst reports of disease caused by superantigen- and enterotoxin-producing S. epidermidis isolates are limited to sporadic cases in the literature, a small minority of strains do possess and express these virulence factors, which can be clinically important. An early case of SEA-producing S. epidermidis was implicated in a major outbreak of SFP in the 1960s.¹²⁴ A novel C-type enterotoxin (SEC) from S. epidermidis was shown to induce high levels of cytokine secretion, including IL-2, −4, −6, −8, −10, IFN-γ, TNF-α and GM-CSF.¹²⁵ This massive release of multiple cytokines in response to MHC II binding is characteristic of superantigenic activity.¹²⁰ Production of IL-6 by lymphocytes in response to SEC is notable as IL-6 is a “pro-sepsis” cytokine that contributes to the systemic inflammation seen in septic shock.¹²⁵ Interestingly, the S. epidermidis homologue of C-type enterotoxin induced higher secretion of IL-6 when compared to its S. aureus homologue.¹²⁵ SEC and another enterotoxin, SEL are often carried on the same S. epidermidis pathogenicity island¹²⁶ and oral administration of SEL and SEC in a mouse model was shown to cause oedema and cytotoxicity in the GI tract.¹²⁷

A recent case of S. epidermidis toxic shock syndrome was reported in which novel PCR assay was used to detect multiple staphylococcal enterotoxins including TSST-1 in blood plasma.¹⁰ Interestingly in this patient, S. epidermidis was only isolated from a urinary culture and not blood cultures.¹⁰ TSST-1 is also found infrequently in veterinary isolates of S. epidermidis, and its horizontal gene transfer (HGT) from CoNS to S. aureus has been widely studied.¹²⁸ Notably one paper described likely HGT of a pathogenicity island in the opposite direction: from S. aureus to two virulent strains of S. epidermidis, allowing for expression of an enterotoxin. This was previously considered to be controversial due to the presence of strong restriction barriers in S. epidermidis.¹²⁸ The presence of superantigens in S. epidermidis is uncommon, and these authors concluded that enterotoxin produced by these strains may have contributed to septic shock in the patients from which they were isolated.¹²⁸ Overall, while reports of S. epidermidis toxic shock and enterotoxigenicity remain rare in the literature, it nevertheless appears that individual S. epidermidis strains may have the capacity to produce both superantigens and enterotoxins.

Proteases

Serine protease (Esp)/glutamyl endopeptidase (GluSE)/SspA

The serine endopeptidase Esp is a secreted protease that belongs to the glutamyl endopeptidase I family, sharing about 50% identity with the better-characterized S. aureus V8 protease.¹²⁹ Esp is ubiquitously found in S. epidermidis strains, in which it is also known as SspA and GluSE.^{17, 130} Production of Esp is associated with adherent, biofilm-like growth rather than planktonic growth.¹³⁰ Esp contributes to the ability of S. epidermidis to compete with S. aureus¹³¹ by cleaving the B-domain of the S. aureus Aap protein required for the attachment phase of S. aureus biofilm formation, as well as degrading fibrinogen, fibrin.^{80, 132, 133} Elastin, fibronectin, and type I collagen,¹³⁴ which all play a role in S. aureus interaction with host cells.^{80, 132–134} Esp is also involved in evading host immune defences by degrading complement protein C5.^{132, 134} Finally, Esp is a significant virulence factor in ocular infections,¹³⁵ which is important given that S. epidermidis is a leading cause of keratitis, blepharitis, conjunctivitis, and endophthalmitis.¹³⁵

Metalloprotease (SepA)/aureolysin

As noted above SepA processes Aap, cleaving off the A-domain, and sometimes the lectin domain to leave the B-domain.¹³⁶ The regulator SarA controls expression of sepA, which is not observed for the other proteases (Figure 3).⁸⁰ A notable role for SepA is in promoting bacterial survival following neutrophil phagocytosis, contributing to immune evasion and maintenance of infection.¹³⁷ SepA has also been reported to degrade AMPs and process AtlE¹³² (Figure 3). SepA-mediated degradation of the anionic human AMP dermcidin is important for survival on the skin and also notable given that the standard staphylococcal defence mechanisms against CAMPs, such as altering the bacterial cell wall charge via D-alanylation of wall teichoic acids, are ineffective against anionic dermicidin.¹³⁸

Extracellular cysteine protease (EcpA)

EcpA is a secreted cysteine protease with gelatinase and collagenase activity.³⁵ EcpA is negatively regulated by the protease inhibitor EcpB, also known as staphostatin A, and positively regulated by agr system.^{132, 139} Secretion of EcpA damages the epithelial barrier by breaking down host barrier proteins, primarily desmoglein-1.³⁵ This breakdown in dermal integrity increases disease severity in atopic dermatitis.³⁵ EcpA is homologous to SspB in S. aureus which is known to degrade the LL-37 epithelial antimicrobial peptide.¹³² Cau et al. recently showed that S. epidermidis EcpA also degrades LL-37, and increased expression of IL-6, IL-8, TLSP, IL1-α and IL-1β in host keratinocytes, promoting inflammation.¹⁴⁰ Autoinducing peptides (AIPs) from other CoNS species, such as S. hominis, can decrease expression of EcpA and the presence of a healthy skin flora may therefore inhibit S. epidermidis-mediated skin damage in patients with AD.¹⁴⁰

Epidermidin leader peptide processing serine protease

The extracellular epidermin leader peptide processing serine protease (EpiP) cleaves the propeptide EpiA, generating the functional lantibiotic form of epidermin.¹⁴¹ Lantibiotics such as epidermin have broad antibacterial activity microorganisms, and consequently self-immunity proteins are required to prevent ill effects from the production of these molecules.¹⁴² Whilst the primary function of epidermin is in defence, EpiP is also involved in cleavage of host proteins such as collagen and casein,^{132, 143} resulting in breakdown of connective tissue. Furthermore, the Streptococcus pyogenes homolog of EpiP impairs neutrophil recruitment, raising the possibility that S. epidermidis EpiP has a similar function.¹⁴³

Cytolysins

The phenol soluble modulins (PSMs) are the major cytolysins in S. epidermidis^{144, 145} and comprise a family of small amphipathic, α-helical peptides that also serve as effectors of staphylococcal biofilm structuring and dispersal both in vitro and in vivo.^{144, 146, 147} Clinical PJI-associated biofilm forming strains are less likely to produce PSMs, potentially due to the dispersal effect PSMs have on established biofilm.¹⁴⁸

S. epidermidis produces 6 types of PSMs: PSMα, PSMβ1, PSMβ2, PSMγ (also known as δ-toxin), PSMδ, and PSMε, and may additionally possess PSM-mec.^{149, 150} PSMγ also known as δ-haemolysin is encoded by the hld gene located in the RNAIII locus of the agr operon.^{122, 148} α-PSMs are small, approximately 20–25 amino acids in length and are associated with cytolysis of immune cells and induction of neutrophil chemotaxis.¹⁵¹ β-PSMs, characterised by their larger (45 amino acid) size, disrupt the biofilm matrix by interfering with the non-covalent forces involved in cell-cell interactions.¹⁵² At lower concentrations of β-PSMs, this allows for the formation of channels, a process essential for the delivery of nutrients to deeper biofilm layers, whereas higher levels of the PSM result in dispersal of biofilm, allowing planktonic cells to be released.^{152, 153} The chromosomal locations for the genes for α and β PSMs are found separate from PSMγ/agr.¹⁴⁸ β-PSMs genes occur in tandem on the chromosome, with two identical genes next to each other.¹⁴⁸ The psm-mec gene is located in the SCCmec mobile genetic element.¹⁵⁴ Independent of the PSM-mec peptide, the psm-mec RNA acts as an independent regulator of RNAIII.¹⁵⁵ PSM-mec influences the virulence of S. epidermidis in sepsis and is highly inflammatory and induces neutrophil chemotaxis.¹⁵⁴

Beyond the PSMs, β-toxin, encoded by the hlb gene, is a sphingomyelinase, which in S. aureus can lyse erythrocytes¹⁵⁶ and kill proliferating human T-lymphocytes,¹⁵⁷ but in S. epidermidis has a protective role in promoting production of ceramides important for skin integrity.¹⁵⁸

Regulation of virulence

Quorum sensing

Staphylococcal quorum sensing is primarily mediated by autoinducing peptides (AIPs) that are secreted from the cell and recognised by their cognate receptors.¹⁵⁹ Quorum sensing regulates S. epidermidis virulence in a cell-density dependent manner, triggering changes in exotoxin secretion and adherence factors as well as regulating formation and dispersal of biofilm.^160–163

Accessory gene regulator (Agr) system

The Agr system, encoded by the agrABDC-RNAIII locus, is a quorum sensing system that plays a key role in regulating virulence in S. epidermidis and is regulated in a density-dependent manner¹⁶⁴ The agr system produces two effector molecules. The first, AgrA, is a cytoplasmic response regulator that is phosphorylated by AgrC, a membrane-bound histidine kinase, in response to the agrD-encoded autoinducing peptide (AIP) after a threshold concentration is reached.^{165, 166} The AgrD pro-peptide is processed and exported by AgrB, in tandem with the signal peptidase SpsB.^{116, 167} AgrA also controls expression of a secondary effector molecule, RNAIII.¹⁶³ RNAIII is a 510 nucleotide non-coding RNA molecule that controls protein production by acting as an antisense RNA, with the long 5’ untranslated region (UTR) of RNAIII when binding to the mRNA product of target genes, impairing translation.^168–171

In S. epidermidis, three different agr groups have been described, (I, II, III) based on polymorphisms identified within the agrC, agrB and agrD genes that result in different types of AIP production.^{116, 172, 173}

AgrA autoregulates its own expression, and also regulates RNAIII transcription¹⁶⁶ (Figure 3). Agr controls expression of many S. epidermidis virulence genes, including geh lipase and ecp cysteine protease.¹⁶⁰ An active agr system is strictly required for the expression of PSMs.^{144, 148, 174} PSM production regulated by the AgrA/AIP circuit occurs primarily in the late-log phase of growth, which coincides with maximal agr activity.^160,166

As noted above, the hld gene encoding for production of PSMγ (δ-haemolysin) is encoded within the transcript for RNAIII.¹⁷⁰

The cysteine protease EcpA, which triggers production of pro-inflammatory cytokines and disrupts the epithelial membrane is positively regulated by the agr system.¹⁴⁰ In the absence of a functional agr system, no activation of host TNF-α and HIV1-LTR is observed, and neutrophil chemotaxis was reduced.¹⁷⁵ The presence of the agr system is also associated with resistance to cathelicidin and β-defensin AMPs.¹⁷⁶

The agr system plays a central role in controlling the virulence of S. epidermidis under both biofilm and planktonic conditions. A biofilm contains a variety of niches with cells in different metabolic and physiological states, and so the ability of S. epideridis to tightly control the regulation of biofilm dispersal as well as the induction of inflammation in a host is critical for establishment and maintenance of disease (Figure 3).

LuxS/Autoinducer-2 quorum sensing

Autoinducer-2 (AI-2) is the signal sensed by a proposed cross-species quorum sensing system present in both Gram-positive and Gram-negative bacteria, originally identified in Vibrio harveyii.¹⁷⁷ AI-2 is derived from a precursor molecule produced by LuxS. The characterization of AI-2 signalling in staphylococci remains somewhat unresolved, with separate papers describing a role in both positive¹⁷⁸ and negative regulation¹⁷⁹ of biofilm. Furthermore, the positive regulation of biofilm was strain dependent with RP62A showing decreased biofilm in a luxS deletion mutant, whereas similar experiments with CSF41498 and clinical isolates revealed no impact on biofilm¹⁷⁸ Addition of exogenous AI-2 was shown to differentially regulate genes involved in glycometabolism and amino acid, nitrogen, and nucleotide metabolism.¹⁷⁸ PSM production was also increased in the presence of AI-2, independent of agr regulation.¹⁷⁸ Whilst its role remains to be fully elucidated, the available evidence suggests that when AI-2 quorum sensing is active in S. epidermidis, it plays a strain-dependent role in modulation of biofilm formation and metabolic regulation.

Global regulators

Alternative sigma factor σ^B

σ^B is centrally involved in the environmental regulation of icaADBC expression and biofilm production¹⁸⁰ and controls the transcription of >200 staphylococcal genes.¹⁸¹ Under normal growth conditions σ^B, which is encoded by the sigB gene in the rsbUVWsigB operon, is bound to the anti-σ factor, RsbW and consequently inactive.^{79, 182} The RsbU phosphatase is activated by environmental stress leading to dephosphorylation of RsbV, which binds to RsbW, thereby releasing σ^B to engage with the core RNA polymerase and up-regulate transcription of genes with σ^B-dependent promoters.^{79, 183} However, neither the icaR nor icaADBC promoters are σ^B-dependent.¹⁸⁴ Furthermore, σ^B controls ica operon expression by indirectly regulating expression of icaR.^{79, 180, 183} Thus stress-induced activation of σ^B leads to repression of icaR transcription, via an unidentified intermediary, and the consequent activation of the ica operon. σ^B also plays a role in the stability of mature S. epidermidis biofilm, contributing to persistence in a rat catheter infection model.^{185, 186}

CodY

The CodY global regulator in S. epidermidis acts as a sensor of nutrient availability to coordinate gene expression in response to nutrient fluctuations. The CodY regulon consists of hundreds of negatively regulated genes that contain CodY boxes, typically located within their promoter regions.¹⁸⁷ CodY binding interferes with RNA polymerase interaction with the promoter, downregulating expression. When nutrient levels are high, CodY is activated by its ligands, branched-chain amino acids (BCAAs) or GTP, and binds to the CodY box, resulting in transcriptional repression of target genes^{160, 188} In contrast, under conditions of nutrient limitation, dissociation of CodY from the CodY box leads to de-repression of target gene expression.

In staphylococci, CodY is particularly associated with negative regulation of virulence.¹⁸⁷ In S. aureus, CodY negatively regulates the ica operon, leading to reduced production of PIA and impaired biofilm formation.^{187, 189} The role of CodY in S. epidermidis biofilm formation has yet to be clearly elucidated, though a possible role in induction of cell dormancy during stationary phase and in biofilms has been shown, and a role in regulation of agr has been clearly demonstrated.^190–192

Other staphylococcal virulence factors regulated by CodY include secreted extracellular proteases like SspA and AtlE, as well as the geh lipase, the hla α-haemolysin and the surface protein SasG.¹⁹³ CodY also indirectly regulates the global regulator SarA by repressing the expression of rot, which encodes a repressor of sarA. Reduced levels of Rot relieve SarA repression, leading to increased expression of SarA-regulated virulence genes¹⁹⁴ (discussed in more detail below). In staphylococci, expression of CodY itself is negatively regulated by the agr system¹⁶⁰ and in turn, strongly represses expression of agr through an unknown intermediate, revealing the complexity of this network of interconnected virulence regulators. Whilst these targets have been primarily identified in S. aureus, S. epidermidis homologues (ica,¹⁹⁵ sspA,¹³² atlE,¹⁹⁶ geh,¹⁹⁶ sasG/aap,¹⁹⁷ sarA,¹⁹⁸ agr¹⁶⁵) may also be regulated by CodY.

Sar family regulators and Rbf

Sar family regulators are characterized as winged-helix family of transcriptional regulators with conserved DNA binding domains that typically bind to the promoter regions of target genes, up- or down-regulating their expression.¹⁹⁹

SarA is particularly associated with the control of biofilm and virulence gene expression including a central role in the positive regulation of PIA-mediated biofilm formation (Figure 3). IS256 insertional inactivation of sarA impaired biofilm formation in the icaADBC-positive strain RP62A, which was reversed by clean excision of the insertion sequence.¹⁸⁰ SarA can bind to the icaA promoter, activating transcription of ica and production of PIA,¹⁹⁸ which implied that SarA controls biofilm in an agr-independent manner.¹⁹⁸ The S. epidermidis sarA mutant also showed decreased extracellular protease activity.¹⁹⁸ SarA may also play a role in switching between polysaccharide- and protein adhesin-mediated biofilm formation.⁸⁰ A transposon inactivation of sarA increased biofilm production in ica-negative S. epidermidis 1585 due in part to upregulation of the sarA-repressed Embp biofilm protein and SepA metalloprotease (which processes AtlE and Aap)^{80, 101} (Figure 3). Regulation of sarA itself remains incompletely understood and is driven by three promoters designated SarP1, SarP2, and SarP3, of which SarP1 is σB-dependent. σB regulates ica operon transcription in S. epidermidis via indirect regulation of the icaR repressor.¹⁸⁴ SarA positively regulates the ica operon by directly binding to the ica promoter, independently of σB¹⁸⁴ (Figure 3).

SarZ is involved in both the initial attachment and accumulation phases of biofilm formation, as well as regulating the expression of serine protease and geh lipase genes.⁷⁷ Interestingly, deletion of the sarZ gene resulted in β-haemolysis, a phenotype not typically observed in S. epidermidis.²⁰⁰

SarX also positively regulates the expression of the ica operon though direct interaction with the ica promoter, resulting in increased production of PIA.^{201, 202} SarX also regulates the agr locus, but the effect on biofilm production is predominantly due to ica expression.²⁰¹

Rbf

The AraC/XylS-family transcriptional regulator rbf (Regulator of biofilm formation) gene is located immediately upstream of sarX and positively influences biofilm in both S. aureus and S. epidermidis.^{203, 204} Rbf does not interact with the ica operon promoter region of S. epidermidis demonstrating that it controls PIA-mediated biofilm indirectly.²⁰⁵ Electrophoretic mobility shift assays also showed that Rbf bound specifically to the sarR promoter but not to the promoters of the sarX, sarA, sarZ, spx, and srrA genes.²⁰⁵ SarR was subsequently shown to be a repressor of ica operon transcription revealing that Rbf controls biofilm by downregulating expression of the sarR repressor.^{202, 205} These findings further reveal the complexity of S. epidermidis PIA regulation as well as the importance and sophisticated fine tuning of this phenotype.

Two-component signal transduction systems

S. aureus exoprotein expression regulator (SaeRS)

SaeRS has been extensively studied in S. aureus,²⁰⁶ but the cognate S. epidermidis system has notable differences, including the absence of the SaeP and SaeQ accessory proteins involved in S. aureus response to neutrophils.^{207, 208} Virulence genes under the control of SaeSR include the geh lipase gene and the esp serine protease gene (Figure 3).²⁰⁷ In S. aureus, SaeS senses human neutrophil peptide 1 (HNP1), but not in S. epidermidis.²⁰⁶ SaeRS also contributes to inflammation in a mouse catheter infection model, with increased proliferation of polymorphonuclear neutrophils, which occurs independently of PSM production.²⁰⁷ Deletion of SaeR results in impaired nitrate utilization and decreased expression of genes involved in anaerobic growth, implicating the SaeRS system in mediating the switch between aerobic and anaerobic growth²⁰⁷. However, saeR deletion does not impact ica transcription levels, and it is suggested that the differential regulation of virulence genes in a saeR mutant may be indirect, due to the effects on metabolism²⁰⁹. Activation of SaeRS decreases autolysis by downregulating transcription of aae and atlE (Figure 3), which impacts cell survival and viability during biofilm and planktonic growth.²⁰⁹

Staphylococcal respiratory response protein (srrAB)

Another two-component system that regulates both survival and virulence of S. epidermidis is SrrAB (Figure 3). Unlike its cognate system in S. aureus, which is only expressed only under microaerobic conditions,²¹⁰ SrrAB in S. epidermidis is active under aerobic conditions and upregulated under low-oxygen conditions.²¹¹ SrrAB regulates both initial adherence and subsequent biofilm matrix production, as well as overall growth and fitness under both aerobic and microaerobic conditions.²¹¹ Specifically, under aerobic conditions, the phosphorylated response regulator SrrA, which binds to both the icaA and icaR promoters, positively regulates icaADBC expression while simultaneously downregulating icaR expression, resulting in increased PIA production and biofilm formation.²¹¹ Under microaerobic conditions, deletion of srrA results in a significant growth defect, and transcription of both icaR and icaADBC is downregulated.²¹¹ SrrA also positively regulates the qoxBACD-encoded respiratory chain terminal oxidase under oxygen-replete conditions. SrrA controls expression of pflBA and nrdD during microaerobic growth, also through direct interaction with the promoter regions of these genes.²¹¹ The ability of SrrAB to modulate growth and biofilm production in oxygen-replete and oxygen-limited environments highlights the capacity of S. epidermidis to adapt to a wide range of environmental conditions and is likely to contribute to its pathogenicity.

GraXSR/VraFG

The GraXSR (also referred to as ApsRSX) two-component system and neighboring VraFG efflux system are involved in resistance to cationic AMPs (CAMPs).²¹² GraS is a membrane-associated histidine kinase which works together with GraX and the response regulator GraR.²¹³ GraXSR is known to modulate the activity of VraFG, a permease and ATPase respectively, and all five proteins function together as a signal transduction complex.²¹³ The function of GraX remains unclear, but it may act to potentiate the signal transduction of the GraXRS/VraFG complex.²¹³ A nine amino acid extracellular loop of the GraS protein senses CAMPs.²¹³ One model suggests that in S. epidermidis, CAMP interaction with the extracellular sensor of GraS dislodges the guard loop of VraG from the membrane, leading to a change in VraG conformation that could positively modulate the activity of GraS.²¹³ Deletion of the guard loop of VraG also leads to knock-on effects such as increased transcription of the MprF gene, which lysinylates phosphatidylglycerol and helps repel CAMPs from the cell membrane.²¹³ The GraXRS system also affects regulation of the dlt operon, which is responsible for D-alanylation of wall teichoic acids, increasing the positive surface charge of the cell envelop and repelling CAMPs.²¹³ CAMPs are currently under investigation as an alternative treatment strategy to traditional antibiotics.²¹⁴ S. epidermidis, like other species, has intrinsic resistance to some CAMPs, in this case through the activity of GraXSR, and so understanding the mechanisms behind this recognition and repulsion of CAMPs may help develop synthetic CAMPs to overcome resistance.²¹⁵

Other regulators

ClpP protease and ATPase chaperones (ClpX, ClpC)

ClpP and ClpX constitute an ATP-dependent protease-chaperone system in S. epidermidis²¹⁶ that orchestrates controlled degradation of specific protein substrates, contributing to cellular homoeostasis.²¹⁷ Under cellular stress (e.g. antibiotics, temperature, oxidative stress), ClpP expression is increased to manage the accumulation of misfolded proteins, contributing to cellular fitness and survival.²¹⁷ ClpXP-mediated proteolytic regulation of stress response regulators enhances cellular resilience.²¹⁸ Paradoxically, deletion of ClpX or ClpP alone improves stationary-phase fitness because the survival of individual cells in stationary phase is enhanced in the absence of ClpXP-dependent degradation of misfolded proteins.²¹⁹ Deletion of ClpX in S. aureus also enhances survival at high temperatures due to the compensatory role of the heat shock-induced chaperone ClpC,²²⁰ whose role in S. epidermidis remains largely unexplored, although it is present in the genomes of many clinical and animal isolates.^{196, 221}

ClpP positively regulates biofilm formation, especially primary attachment and PIA production (Figure 3), impacting virulence in a rat IV-catheter infection model.²¹⁶ Mutations in spx, encoding an RNA polymerase binding protein conserved in the Bacillota, reversed growth and competence defects caused by ClpX or ClpP mutations in Bacillus.²²² Deletion of ClpP appears to have no direct effect on ica operon transcription. However, because ClpP controls the activity of Spx by proteolytic degradation (Figure 3),²²³ mutation of clpP is accompanied by downregulation of the ica operon, implicating an unidentified intermediate factor, possibly another ClpP protein target, that neutralizes the negative regulatory effect of Spx on ica transcription.^216,224 Downregulation of ClpP by a functional agr system is also consistent with agr-mediated negative regulation of biofilm formation, favoring release of planktonic cell for dissemination.^{216, 224} This flexible regulation of a key virulence factor highlights the capacity of S. epidermidis to respond to environmental and population density cues to ensure persistence within the host.

Phenotypic variation and IS256

IS256 belongs to the IS6 family of insertion sequences and is characterized by its relatively small size (1.2 kb) and the presence of terminal inverted repeats (IRs). These IRs facilitate its transposition within the bacterial genome, allowing IS256 to introduce genetic variability within S. epidermidis populations, including gene inactivation, up- or down-regulation of gene expression and chromosomal rearrangements.²²⁵ This genetic diversity can lead to variations in virulence factor expression and pathogenicity amongst isolates.

Compared to icaADBC⁺ commensal strains, clinical isolates with icaADBC are considerably more likely to carry IS256 (85% vs. 15%).²²⁶ IS256 is known to modulate the biofilm phenotype and can cause a switch from PIA-dependent to PIA-independent biofilm when inserted into the icaC locus,²²⁷ as well as decreasing ica operon expression and PIA production when inserted into the rsbU and sarA loci.¹⁸⁰ An important aspect of IS256 phenotypic switching is that it is reversible, as the insertion sequence can be subsequently excised from the gene, leaving an intact wild-type locus.²²⁸ This clean excision is a rare event, however, and appears to be independent of IS256 transposition, which normally leaves 8-bp target site duplications (TSDs).²²⁸

The presence of IS256 is considered a virulence marker,^{229, 230} as IS256 is found more frequently in clinical isolates (up to 80%) of S. epidermidis compared to community isolates (13%) from healthy volunteers.²³¹ Isolates from sequence type 2 (ST2) have a high rate of IS256 carriage, up to 100%.²³² ST2 is a widely distributed sequence type that is commonly found in clinical infections, especially catheter infections²³¹ and is prevalent in hospital environments.²³³ In prosthetic joint infections, up to 83% of isolates show the presence of IS256, compared to just 4% of commensal isolates.^{234, 235}

IS256 is frequently associated with the acquisition and dissemination of antibiotic resistance genes within S. epidermidis populations. Isolates containing IS256 are more likely to exhibit resistance to multiple antibiotics,²³⁶ including β-lactams and aminoglycosides.²²⁶ IS256 presence is also correlated with carriage of mecA.²³⁷ Whilst multiple copies of IS256 are typically distributed throughout the S. epidermidis genome, it is also found as a component of transposon Tn4001, which confers aminoglycoside resistance.²³⁸

A striking example of the clinical significance of genomic rearrangements induced by IS256 was reported in a case of recurrent meningitis, in which a switch from a biofilm-positive to biofilm-negative phenotype was associated with IS256 translocation and large chromosomal rearrangements.²²⁵ The ability of a single strain to rapidly adapt within the host gives it a clear fitness advantage, particularly when host immune factors increase selective pressure. Furthermore, several studies have suggested that IS256 carriage is a more predictive indicator of S. epidermidis virulence than the presence of icaADBC, supporting the idea that S. epidermidis virulence is also linked to its capacity for adaptation through establishment of genetically diverse subpopulations.^{226, 229, 234, 239}

Host-S. epidermidis interactions

S. epidermidis must modulate the innate immune response to facilitate its persistence. Moreover, the host response to S. epidermidis as a commensal organism and as a pathogen can vary significantly in response to the location and strain, as outlined in several recent reviews.^{3, 8, 16, 240}

Recognition and response

Identifying pathogens involves initial recognition of bacterial components by pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) and nucleotide-binding oligomerization domain (NOD)-like receptors on epithelial cells, macrophages, dendritic cells, and neutrophils^{241, 242}. Following recognition, signalling pathways are triggered, leading to production of pro-inflammatory cytokines, chemokines, and AMPs that induce chemotaxis of immune cells and result in phagocytosis and bacterial killing.²⁴³ S. epidermidis activates different PRRs depending on its cell wall components and secreted toxins, such as lipoteichoic acid (TLR2), peptidoglycan (TLR2 and NOD2), lipoproteins (TLR1/TLR2 and TLR2/TLR6), and PSMs (TLR2 and TLR6).^244–250

Notably, S. epidermidis can inhibit TLR signalling, as is seen on the skin surface where it dampens TLR2 response in keratinocytes and reduces TLR3-mediated inflammation following skin injury.²⁴⁵ Along with the inhibition of TLR2, S. epidermidis can also induce the expression of anti-inflammatory cytokines, such as IL-10 and TGF-β, and regulatory T cells that suppress the inflammatory response and promote tolerance to the commensal bacteria, discussed in further detail below.^251–253 Strains isolated from atopic dermatitis are more likely to cause a pro-inflammatory cytokine response in an in vitro model than strains from the skin of healthy patients, highlighting the complex role of S. epidermidis in epithelial immunity.²⁵⁴

Immune evasion

Neutrophils/polymorphonuclear leukocytes (PMNs) form an important part of host defence against S. epidermidis. Biofilm is particularly associated with downregulation of pro-inflammatory cytokines, allowing infection to persist undetected.²⁵⁵ Production of PIA shields S. epidermidis from PMN phagocytosis, although PIA-independent biofilm can also impede effective opsonization by antibodies and complement components.^{74, 256, 257} Evasion of PMNs by S. epidermidis is also associated with repulsion and proteolysis of AMPs by GraRS and SepA respectively, as well as biofilm-mediated shielding of cells from IgG deposition.^{137, 258} Broader evasion of the complement system also contributes to S. epidermidis survival and persistence in the host²⁵⁹. Phagocytic killing of staphylococci is mediated primarily by the C3b and C5a opsonins,²⁵⁹ and whilst S. epidermidis can activate the complement system, effective opsonization by C3b is impeded in a biofilm.²⁵⁸

S. epidermidis as a beneficial member of the skin microflora

Competitive exclusion

S. epidermidis agr signals can interfere with the S. aureus agr system, resulting in downregulation of virulence factors required for S. aureus to establish infection.^{165, 260} Agr type I and IV S. epidermidis strains from patients with atopic dermatitis exhibited inhibitory effects on the expression of S. aureus virulence factors in vitro, and reduced S. aureus colonization of the skin in a mouse model.²⁶¹

S. epidermidis produces several antimicrobial peptides that inhibit the growth of S. aureus, including bacteriocins. Bacteriocins are antimicrobial peptides synthesized by the ribosome and typically target species that are closely related to the producer,²⁶² in order to stave off competition for a particular niche. For example, 96% of nasal S. epidermidis isolates produce bacteriocins.²⁶³ As a result, many S. epidermidis bacteriocins are active against S. aureus as well as other S. epidermidis and CoNS strains. S. epidermidis is known to produce the bacteriocins pep5,²⁶⁴ epidermin,²⁶⁵ epilancin K7²⁶⁴ epilancin 15×,²⁶⁶ nukacin IVK45²⁶³ and epicidin 280.²⁶⁷ A skin isolate of S. epidermidis has been described that produces epidermicin NI01, a plasmid-encoded bacteriocin.²⁶⁸ Plasmid-encoded nukacin IVK45 shows sequence similarity to genes in other staphylococcal species, indicating horizontal gene transfer of bacteriocins between species.²⁶³

On healthy skin, S. epidermidis is found to inhibit colonization with harmful Cutibacterium acnes, including selective activity against C. acnes strains associated with acne.²⁶⁹ These strains of S. epidermidis are typically not from the sequence types associated with infection²⁶⁹. Co-culture experiments with C. acnes and a representative strain of S. epidermidis showed that EpiA (precursor peptide for epidermin, which is known to inhibit survival of C. acnes) and PSMβ expression increased upon co-culture with C. acnes.²⁶⁹

S. epidermidis can also interfere with the virulence of other staphylococcal species. Co-culture of S. epidermidis with S. aureus results in SaeRS-dependent downregulation of the S. aureus hla-encoded haemolysin,²⁷⁰ via a mechanism involving an unidentified non-proteinaceous factor present in culture supernatants.²⁷⁰ The secreted serine protease Esp, which as noted above processes Aap and degrades complement protein C5, also impairs S. aureus biofilm production²⁷¹ and can decolonise S. aureus from the nasal passages.¹³¹ Furthermore, Esp is also involved in nutrient acquisition via the breakdown of proteins into glutamate, which can be used by S. epidermidis as an energy source.¹³² Esp highlights the overlapping functions of S. epidermidis persistence/colonisation and virulence factors.⁸⁰

Immunomodulation

Gamma delta T-cells (GD T-cells) found in the epithelium are activated by the presence of S. epidermidis, leading to increased expression of p-2, an antimicrobial effector protein that kills S. aureus by forming pores in the bacterial membrane of internalised cells, as well as the TNF family transmembrane protein Fas Ligand and the pore-forming toxin granulysin.²⁷² S. epidermidis LTA can activate Toll-like receptor 2 (TLR2) on keratinocytes that have been triggered to induce inflammation via Toll-like receptor 3 (TLR3) in response to skin injury.²⁷² This interaction triggers an anti-inflammatory response through the NF-κB signalling pathway,²⁷³ reducing proinflammatory cytokines like TNF and IL-6 production, as well as inducing the production of AMPs.^{245, 274} Similarly, LTA can induce production of IL-6, IL-1β and TNF and elevate nitric oxide levels in macrophage cells lines while also stimulating an IgG response in patients with S. epidermidis device-associated infections.²⁷⁵ Commensal colonisation with S. epidermidis has been shown to induce a specific and co-ordinated dendritic cell response that enhances IL-17A⁺ CD8⁺ T-cell migration to the skin, which is associated with improved barrier function.²⁴⁷ These findings indicate that S. epidermidis-mediated modulation of the immune system is associated with both benefit and harm to patients, in keeping with its ability to switch from a protective member of the skin microflora to a pathogen. Indeed, the beneficial effects of S. epidermidis on keratinocyte and GD T-cell mediated-responses in the outermost epithelial cells of the skin contrasts with the deleterious impact on macrophage-mediated responses, which are more commonly detected in the underlying dermis following skin breach.²⁷⁶

Maintenance of skin barrier function and homeostasis

S. epidermidis plays a role in maintaining the skin’s hydration and preventing skin disorders like atopic dermatitis by regulating skin barrier function.²⁷⁷ Production of the sphingomyelinase Sph, a hydrolase that breaks down sphingolipids onto ceramides and phosphocholine, benefiting the organism, protecting the skin from dehydration and promoting healthy turnover of skin cells.^{3, 158} Using a mouse model, butyrate, which is a fatty acid metabolite produced by S. epidermidis, was shown to protect the epidermis from IL-6 production induced by damaging UVB light.²⁷⁸ S. epidermidis also produces trace amines from aromatic amino acids (AAAs) found in high concentrations in skin.²⁷⁹ Whilst S. epidermidis does not appear to utilise AAAs in its own metabolism, excretion of trace amines by SadA-mediated decarboxylation of AAAs accelerates wound healing in a mouse model, via a mechanism associated with internalisation of the bacteria into host cells (offering protection from the host immune system) and inhibition of epinephrine production by keratinocytes (which negatively impacts wound healing by inhibiting cell migration).²⁷⁹ However, it is important to note that whilst S. epidermidis may provide benefits for the host, the function of these interactions is clearly to ensure the survival of the bacteria. Persistence in the host may be benign until circumstances allow for the establishment of infection.

Diagnostics

Diagnostics for invasive S. epidermidis isolates responsible for infection is complicated by the ubiquitous presence of this organism on skin and mucous membranes. Efforts to evaluate the significance of a single positive blood culture of S. epidermidis as indicative of bacteraemia when concurrent with clinical symptoms have shown that whilst consecutive positive cultures is less prone to false positives, the incubation time between from culture initiation to culture positivity also predictive of bacteraemia and requires a shorter delay before diagnosis and treatment.^280–282 For patients with symptoms of catheter-related sepsis, the gold standard is isolation of the same strain from catheter tips and blood cultures.²⁸³ In the case of PJIs, such as elbow arthroplasty, up to 78% of joint aspirations fail to yield a positive culture, which may be due to difficulties in culturing bacteria from a biofilm, whilst the current gold standard of biopsy and culture is prone to false positives (7.5%) and may result in overtreatment.^284–286 Non-culture-based methods such as Raman spectroscopy are faster than traditional culture-based methods and show promising results for diagnosis of bone graft infections.²⁸⁷ Quantitative PCR methods have thus far been hindered by difficulty in finding suitable markers which are both actively transcribed during infection and conserved amongst isolates, whilst minimising the risks of false positives from non-invasive commensal strains.²⁸⁸ Multiplex real time PCR is capable of distinguishing S. epidermidis, MRSE, S. aureus and MRSA.²⁸⁹

Automated identification systems such as the bioMérieux Vitek 2, BectonDickinson Phoenix and Siemens Microscan assign a species level identification based on the results of biochemical tests.²⁹⁰ These automated systems reduce the time and workload required for bacterial identification, and can show high accuracy for S. epidermidis, though other CoNS species are not so easily identified.²⁹¹ However, unlike real time PCR methods, these systems still require preculture of the isolates.²⁹⁰ Matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF MS) systems facilitate rapid species level identification of CoNS.^292–295 Whilst most typically require growth of the isolate on agar plates prior to analysis, MALDI-TOF MS can be carried out directly on the clinical specimen, for example synovial fluid directly from PJIs, with better sensitivity than culture-based methods.^{296, 297} Attempts to use MALDI-TOF MS to differentiate MRSE from MSSE have been only partially successful, as the results are prone to false negatives.^{298, 299} A small study also showed that MALDI-TOF MS could potentially differentiate between biofilm-producing and non-biofilm producing strains.³⁰⁰

Current antimicrobial treatment strategies

In the treatment of S. epidermidis infections, antibiotic selection is predicated upon the location of infection, antibiotic susceptibility profile, and the presence of prosthetic devices.

For bloodstream infections and catheter-related infections, vancomycin remains the gold standard due to its efficacy against MRSE, with β-lactams such as nafcillin and oxacillin also considered based on the resistance profile of the pathogen.^{17, 301} For ocular infections, fluoroquinolones, such as moxifloxacin or levofloxacin, are frequently utilized owing to their good penetration into ocular tissues and broad-spectrum activity.³⁰² In the case of endophthalmitis, which can cause severe visual loss if time to treatment is delayed and of which S. epidermidis is a leading cause of infection, intravitreal injection of vancomycin can be used.³⁰³ Device-related infections, particularly those involving biofilm formation, often require a combination of surgical intervention and antibiotic therapy, with rifampicin combined with a glycopeptide (e.g. vancomycin) or a lipopeptide (e.g. daptomycin) recommended to eradicate biofilm-associated bacteria without promoting the emergence of rifampicin-resistant staphylococci.^{301, 304, 305}

Antimicrobial resistance

Emergence of methicillin-resistant S. epidermidis (MRSE)

The emergence of MRSE mirrors the emergence and spread of MRSA. Acquisition of mecA, reduces susceptibility to methicillin and a broad spectrum of β-lactam antibiotics. The mecA-encoded penicillin-binding protein 2a (PBP2a) is less susceptible to binding by all β-lactams, including β-lactamase-resistant derivatives such as methicillin, cloxacillin, flucloxacillin, and oxacillin.^{196, 306, 307} The 13 known variants of SCCmec, which carries the mecA gene, are differentiated based on the composition of their cassette chromosome recombinase (ccr) and mec gene complexes.^308–311 Types I-VI, including various subtypes have been described in S. epidermidis, with type IV generally being most frequent.^312–319 Through the action of the recombinase genes in SCCmec, the cassette can be excised from the genome and transferred between isolates and between species, and there is evidence that S. epidermidis can function as a reservoir for methicillin resistance that can be passed on to other species, including S. aureus.^320–322 The prevalence of MRSE has increased drastically in frequency over the past two decades and is more common in nosocomial isolates.³²³ In certain populations, such as neonates with bloodstream infections, methicillin resistance rates can reach 100%.^{324, 325}

Fluoroquinolone resistance

Infections caused by fluoroquinolone-resistant S. epidermidis, typically associated with point mutations in DNA gyrase (gyrA, gyrB) and topoisomerase IV (parC, parE) genes,³²⁶ narrow the range of therapeutic options. Moreover, fluoroquinolone resistance is increasing in prevalence, with 23% to 67% of S. epidermidis isolates being resistant to ciprofloxacin,^327–329 and up to 60%, 55%, and 27% of MRSE and MSSE intravitreal isolates exhibiting resistance to levofloxacin, moxifloxacin, and delafloxacin, respectively.³³⁰ Topical fluoroquinolone solutions used prophylactically for cataract surgery have been correlated with an increased prevalence of fluoroquinolone-resistant S. epidermidis strains, including potential within-host adaptation during a 14-day treatment timeframe,³³¹ although transmission of resistant strains from outside the host remains more common.³³²

Glycopeptide resistance

Whilst planktonically-growing strains of S. epidermidis are susceptible to glycopeptides like vancomycin and teicoplanin, biofilms exhibit high tolerance to these antibiotics.^333–335 Heteroresistance in subpopulations of laboratory grown strains to vancomycin was first recorded in a S. epidermidis human isolate in 1996.^{336, 337} Distinguishing homogeneous resistance from heteroresistance is challenging, but even heteroresistant strains can cause recurrent infection following treatment with vancomycin.^338–341 Heteroresistance in catheter-associated bloodstream infections is associated with poor response to glycopeptide treatment, resulting in persistent infection, prolonged hospitalization and death.³⁴² S. epidermidis is more commonly resistant to teicoplanin than vancomycin^{323, 343, 344} and newer generation lipoglycopeptides, including dalbavancin, are currently very effective against S. epidermidis in vivo and against planktonic and biofilm S. epidermidis cells in vitro, extending therapeutic options even after vancomycin treatment failure.^345–351 However, mutations in walK, the histidine kinase of the essential two-component signalling system WalKR involved in regulating cell-wall biosynthesis and turnover,³⁵² can lead to dalbavancin heteroresistance in patient isolates during treatment.³⁵³

Fusidic acid resistance

Resistance to fusidic acid is mediated by Fus family proteins (FusA, FusB, FusC, FusD, FusF) that interfere with antibiotic binding to the ribosome.^{354, 355} FusA, FusD, and FusF are responsible for intrinsic resistance in species including S. saprophyticus and S. cohnii. In S. epidermidis, the most reported Fus proteins are FusB and FusC.^{356, 357} FusB is often carried on mobile phage-related resistance islands that are inserted at att integrase sites in heterologous strains.^{358, 359} The capacity for inter-strain or inter-species transfer of fus genes raises concern, particularly as their carriage has a low fitness cost for staphylococci.^{355, 360}

Linezolid resistance

Resistance to the oxazolidinone drug linezolid can arise due to mutations in the 23S rRNA gene.³⁶¹ Furthermore, methylation of the 23s rRNA A₂₅₀₃ residue by the cfr-encoded methyltransferase leads to resistance to oxazolidinones, phenicols, lincosamides, pleuromutilins, and streptogramin A via decreased antibiotic binding.³⁶¹ Linezolid-resistant S. epidermidis has been implicated in several endemic outbreaks in clinical settings.^362–365 Unlike 23S rRNA mutations, cfr-mediated resistance is carried on a plasmid with a low fitness cost to the host cell,^366–370 although linezolid resistant isolates can possess both cfr plasmids and 23S rRNA mutations.³⁷¹ First reported in a S. sciuri bovine isolate in 2000 and human staphylococcal strains in 2008, cfr-bearing plasmids may also carry resistance genes for macrolide-lincosamide-streptogramin B antibiotics (MLS_B) and spectinomycin.^{366, 372, 373} Recent reports of MRSE carrying cfr-containing plasmids are particularly concerning, as all strains investigated belonged to ST2, a clonal lineage that is a frequent cause of nosocomial infections and is notable for its worldwide dissemination and multi-drug resistance characteristics.³⁷⁴ Furthermore, several S. aureus isolates from the same hospital contained highly similar plasmids, potentially because of transfer from S. epidermidis.³⁷⁴

Macrolide resistance

Resistance to macrolides associated with cross-resistance to lincosamides and streptogramin B is primarily associated with ermC.³⁷⁵ Erythromycin resistance can be found in approximately 60–65% of S. epidermidis isolates.^{376, 377} ermC is found in up to 85% of constitutively-MLS_B resistant and 100% of inducible – MLS_B resistant S. epidermidis strains.³⁷⁸ Whilst ErmA, ErmB, and ErmC mediate resistance by methylation of 23S rRNA, other mechanisms of resistance to macrolides include drug efflux (associated with the msr genes) and phosphorylation of the antibiotics (associated with the mph genes).³⁷⁸

Multi-drug resistance (MDR)

The prevalence of multi-resistant strains is increasing in clinical settings.^379–381 MRSE already precludes the use of one of the best-tolerated classes of antibiotics, but highly-methicillin resistant strains can also exhibit high rates of resistance to erythromycin (up to 80%), tetracycline (33%), clindamycin (33%), and gentamicin (27%).³⁸⁰ Up to 90% of MRSE isolates can be multi-drug resistant (MDR), posing a significant challenge to treatment.^{55, 382} Amongst ocular isolates, MDR was detected more frequently in infection-related isolates than amongst non-infection-causing isolates, and MDR isolates were significantly more likely to be biofilm producers.³⁸³

S. epidermidis as a reservoir of resistance genes

Horizontal gene transfer of resistance cassettes such as SCCmec highlights the adaptability of S. epidermidis in both acquiring and disseminating resistance determinants.^{360, 384–386} Whilst the precise mechanism of SCCmec transfer in vivo remains unclear, in vitro experiments indicate that of S. aureus can be induced to take up SCCmec by transformation via upregulation of competence genes by two-component systems such as SrrBA.³⁸⁷ In a particularly striking illustration of inter-species transfer of resistance, sequence analysis of three staphylococcal strains (MRSA, MSSA, and MRSE) isolated from the same patient indicated that in-patient transfer of SCCmec from the MRSE strain to MSSA had occurred.³²¹ Plasmid-mediated resistance to tetracycline, streptomycin, and erythromycin can also drive dissemination from S. epidermidis to other species.^{360, 386}

Prevention and treatment strategies

Prevention

Prophylactic measures used as part of medical procedures include fluoroquinolones, known for their efficacy against MRSE, during ocular surgery,³⁸⁸ vancomycin in prosthetic joint revision surgery,³⁸⁹ and clindamycin, vancomycin (teicoplanin in Europe), cephalosporins or vancomycin in combination with rifampicin to prevent PJIs.^{47, 390} Cephalosporins and vancomycin are routinely used for cardiac device implantation, with in vitro data supporting their use in combination with rifampicin.³⁹¹ The FDA has also approved the use of rifampicin/minocycline-impregnated mesh on cardiac implants as an additional preventive measure.³⁹² Notably, antimicrobial catheter lock solutions (CLSs), which are traditionally used post-infection, also have prophylactic potential.³⁹³ Anticoagulants such as sodium citrate used in CLSs can be antimicrobial, although their use as prophylactics to prevent infection have not been standardized.³⁹³ Nevertheless, sodium citrate was shown to be the most effective CLS against S. epidermidis, while taurolidine demonstrated efficacy against E. coli.^{392, 393} However, the potential benefit of using antibiotics in CLSs must be weighed against the risk of promoting resistance, which has been observed with fluoroquinolones in ocular procedures.³⁹⁴ Furthermore, the choice of antibiotic may change over time, as is seen with low susceptibility to clindamycin among S. epidermidis and MSSA PJI isolates when this antibiotic is used prophylactically.³⁹⁰ In this regard, antimicrobial stewardship and susceptibility testing of the frequently isolated strains that cause infection are imperative for responsible use of prophylactic antibiotics.

Antimicrobial coating of medical devices

An attractive approach to combatting infection of biomaterials exploits novel antimicrobial coatings to limit S. epidermidis colonization. Osteopontin, a bovine milk protein, dose-dependently blocks bacterial adhesion under shear flow, potentially targeting early biofilm formation adhesins.³⁹⁵ Polymer blends of polyvinyl chloride (PVC) and polystyrene-ethylene-butylene-styrene (SEBS), infused with the ionic liquid HdmimDMSIP, have activity against S. epidermidis and can be incorporated into devices such as catheter tips.³⁹⁶ Micro-fibrillated cellulose-based antimicrobial films applied to biomaterials are also efficacious against S. epidermidis.³⁹⁷ Cellulose-based hydrogels, incorporating crosslinking agents or ε-poly-L-lysine, also have potential as part of antimicrobial wound dressings.^{398, 399}

Colloid materials, characterized by their small particle size and large surface area, offer promising avenues for designing surfaces that incorporate antimicrobial agents. Silver nanoparticles in chitosan nanogel, combined with ampicillin, were effective against oxacillin-resistant S. epidermidis.⁴⁰⁰ Colloidal suspensions with titanium-peroxy gel layers provide non-antibiotic solutions for combating implant infections.⁴⁰¹

In dual-purpose biomaterial modification, coupling titanium with a cationic antimicrobial peptide exhibited significant bactericidal activity against S. epidermidis and improved osteoblast adherence.⁴⁰² Coating titanium with carboxymethyl chitosan and bone morphogenic protein 2 also inhibited S. epidermidis colonization.⁴⁰³ Surface treatments must be carefully chosen due to varying efficacy against bacterial species and potential cytotoxicity to host cells.^{400, 404–408} For titanium implants, hydrophilicity plays a crucial role in minimizing S. epidermidis attachment, challenging the assumption that hydrophilicity alone predicts implant material efficacy.⁴⁰⁹ In vivo studies showed that hydrothermally and electrophoretically modified titanium surfaces had superior bacterial repulsion properties.⁴¹⁰

Quorum sensing inhibition

The therapeutic potential of quorum sensing inhibition, or “quorum quenching,” which generally centers around competitive inhibition using biosimilar molecules that interfere with autoinduction, the use of enzymes that degrade autoinducing peptides (AIPs), or interference with AIP synthesis was the subject of a recent commentary.¹⁶³

Vaccines and antibody therapy

Vaccination against an important commensal like S. epidermidis presents significant challenges and even if achieved may be potentially detrimental. However, anti-biofilm vaccines are an interesting possibility. Here, the choice of biofilm antigen will be critical key given that staphylococci express different surface epitopes under varying environmental and physiological conditions.^{411, 412} The antigen will need to be highly conserved, if not amongst all strains, then at least amongst strains associated with the target infection type, and the antigenicity of the bacterial epitope also needs to be considered.⁴¹¹

Early work using whole chloroform-inactivated S. epidermidis cells and cell-free filtrate confirmed the immunogenicity of these vaccines in mice, where both vaccines provided protection from S. epidermidis intraperitoneal challenge but did not offer significant protection from S. aureus mortality.⁴¹³ Targeting SdrG, a fibrinogen-binding protein, with antibodies impaired S. epidermidis attachment to Fn-coated catheters and increased phagocytosis by macrophages.^{414, 415} Immunization with a subdomain of SdrG was also shown to induce antibody production and impaired infection by S. epidermidis O-47 in a mouse model.⁴¹⁶ Another cell wall protein, SesC, showed promise as a vaccine candidate in mouse and rat models of infection.⁴¹⁷ Other Ses proteins, including SesB, were also found to exhibit vaccine potential⁴¹⁸ and immunization with the B-domain repeat of Aap resulted in decreased inflammation, impaired bacterial colonization and a significant IgG antibody response.⁴¹⁹ The iron transport protein SitC has also been investigated as a vaccine epitope and more recently its S. aureus ortholog MntC, a manganese transporter, was shown to provide protection from both S. aureus and S. epidermidis infection in a murine infection model.^420–422

Pagibaximab, a monoclonal antibody therapy that targets LTA, showed promise in preventing neonatal sepsis in phase II clinical trials.⁴²³ However, phase III trials showed that whilst the treatment was safe and well-tolerated, prevention of sepsis was unsuccessful.⁴²⁴

Unfortunately to date vaccines and antibody therapies to combat S. epidermidis infection and biofilm formation have not progressed beyond animal trials. Vaccination against more serious S. epidermidis infections, such as neonatal sepsis, is complicated by the immature immune response of neonates and infants that can fail to provide long-lasting memory T- and B-cell responses, as well as poor activity of antigen-presenting cells and lower IgG response compared to adults.⁴²⁵ Nevertheless, vaccine research will continue to be an important part of efforts to find improved ways to control S. epidermidis and other infections^426–429

Concluding remarks

The pathogenic success of S. epidermidis is deeply rooted not only in its biofilm-forming capacity but also in its resilience and remarkable genomic plasticity, which facilitates the acquisition of virulence determinants and rapid adaption to shifting environmental conditions. Research into CoNS infections lags S. aureus, and whilst S. epidermidis is the most studied of the CoNS species, there are still substantial gaps in our understanding of its virulence determinants, resistance modalities, and host–pathogen interactions. Recognition of the importance of CoNS as pathogens is imperative to maintain and expand the global research base in this field. Furthermore, our understanding of CoNS virulence would be improved if S. aureus investigators extended their research, where appropriate, to include comparative studies with S. epidermidis or other CoNS.

The pathogenicity of S. epidermidis is unique in several ways: i. the persistence factors that make S. epidermidis such an effective colonizer of the harsh epithelial environment and how these are redeployed to cause disease, ii. the role of genomic flexibility in the success of S. epidermidis as a pathogen, and iii. the ever-increasing emergence of multidrug resistant strains. Regulation of the major virulence determinant in S. epidermidis, biofilm, is significantly different to S. aureus, and further research is needed to fully understand this in the context of the distinct lifestyle of this pathogen and its interactions with the host including opportunistic infections. To revisit the central question of this review: are we overlooking important disease determinants when S. epidermidis is overlooked as a “common commensal” and not a ubiquitous opportunistic pathogen? The research literature focused on S. epidermidis in recent decades reflects a gradual shift in how this organism is viewed, both as an important commensal for cutaneous health and a complex pathogen capable of causing significant harm to vulnerable individuals. Taking a holistic view continues to provide a more complete understanding of the lifestyle, host–microbe interactions and virulence of S. epidermidis.

All authors agree to be accountable for all aspects of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The figures in this review are openly available in Figshare (DOIs: 10.6084/m9.figshare.25663989 (Figure 1); 10.6084/m9.figshare.25092626 (Figure 2) and 10.6084/m9.figshare.25092623 (Figure 3).

Additional information

Funding

This work was supported by grants from the Irish Research Council [GOIPG/2019/2011 to Ó.B. and J.P.O’G.], Health Research Board [ILP-POR-2019-102 to J.P.O’G.], and Science Foundation Ireland [19/FFP/6441 to J.P.O’G. and 22/PATH-S/10804 to M.S.Z].

Notes on contributors

Órla Burke

Órla Burke Conception and design, writing – original draft, writing – review, editing, and critical revision for intellectual content; writing – approval of the final version for publication.

Merve S. Zeden

Merve S. Zeden Conception and design, writing – original draft, writing – review, editing, and critical revision for intellectual content; writing – approval of the final version for publication.

James P. O’Gara

James P. O’Gara Conception and design, writing – original draft, writing – review, editing, and critical revision for intellectual content; writing – approval of the final version for publication.