This review of the human immune response to Staphylococcus aureus covers large amounts of new information that have become available since our last relatively recent review (Miller et al., 2020). In the previous review, evidence for increased numbers of human infections was linked to defects in innate immunity of keratinocytes, mucosal epithelial cells, and phagocytes (including neutrophils, monocytes/macrophages and dendritic cells). These genetic defects emphasized the role of these cells in preventing S. aureus infections. Also considered is how staphylococcal exotoxins (e.g. superantigens and pore-forming toxins) impair immune functions by lysing or dysregulating innate cells, further highlighting the crucial role of innate immunity in staphylococcal defense. Toxin-neutralizing antibodies have been hypothesized to improve outcomes in invasive S. aureus infections by mitigating such virulence and promoting immune function. This concept derived in-part from considerations of the cytokine response to invasive infections, especially in S. aureus bacteremia (SAB), and is further explored in the current review.

Most previous vaccine trials have been aimed at improving opsonophagocytic activity or promoting Th17 mediated adaptive immunity. New insights discussed herein provide perspectives on the challenges of these approaches. Moreover, previous vaccine trials did not fully appreciate the complexity of S. aureus-host immune interactions, thereby contributing to their limited efficacy. The lack of clear determinants of protective immunity against S. aureus infections, and direct or surrogate biomarkers thereof, also greatly present challenges to successful vaccine development vs. this pathogen.

This review examines a substantial body of recent data on human genetic variation and the crucial role of host-bacterial metabolic crosstalk in contributing to outcomes of S. aureus infections. The new information encompasses several areas, including: 1) the adaptive evolution of S. aureus within the host; 2) how S. aureus evades immune efficacy within the host; 3) how human genetic variability can be linked to outcomes of infections; 4) how the interplay between host and pathogen metabolism shapes the immune response to intracellular S. aureus, including the evolution of S. aureus metabolic genes over time within the host; 5) how vitamins enhance the innate immune response to S. aureus; and 6) how anti-toxin antibody levels and anti-glucosaminidase levels may be biomarkers for outcomes of S. aureus infections.

S. aureus has co-evolved with mammalian hosts for millennia, which has produced multiple mechanisms for evasion of the host immune response. Indeed, rates of invasive infections and deaths from S. aureus are similar if not higher compared to other severe respiratory pathogens such as N. meningitidis, H. influenzae, and S. pneumoniae for which vaccines have been developed (Proctor, 2021; Collaborators, 2022). [Incidence of invasive infections/100,000 human cases/year: S. pneumoniae = 15; H. influenzae = 23; S. aureus = 32]. Further complicating vaccine development is the kinetic relationship of in vivo host-staphylococcal interactions (discussed below). Moreover, over forty immune evasion molecules produced by S. aureus that disarm the immune system (e.g., complement inactivation, antibody subversion, TLR2 inhibition) have been described (de Jong et al., 2019), and these will not be covered in this review. Rather, this review will focus on key concepts regarding the evolution of S. aureus within the host.

During nasal colonization, S. aureus strains that caused invasive infections evolve within the host. Such genotypic variation is termed adaptive mutation. In a study by Young et al. (2017), genomes from 1163 strains were examined from 105 patients wherein 72% of the infections emerged from the nasal colonizing strain. There was a 2.8 to 3.6-fold enrichment of adaptive mutations in genes including rsp and agr. These gene products regulate the expression of surface proteins and toxins, which stimulate host-derived anti-microbial peptides. The adaptive mutations in pathogenesis-associated genes occurred in the strains that became invasive (3.1-fold enriched) as compared to those that only resided in the nose. Young et al. noted that “whole-genome sequencing case studies add weight to the idea that within-host evolution plays an important role in infection”. Enrichment of adaptive mutations was not observed in asymptomatic carriers, nor within unrelated bacteria.” This relationship indicates that some patients have selection pressures that place them at risk of invasive disease due to strains of S. aureus that may be uniquely advantaged in a given human host. In another study, when colonizing strains were compared to standard laboratory strains of S. aureus, colonizing strains showed a delay in inducing tlr2 expression in cultured human nasal epithelial cells (NECs) (Quinn and Cole, 2007). Moreover, human β-defensin-2 and human β-defensin-3 were negligibly induced in NECs by nasal carrier strain in comparison to the non-carrier strain of S. aureus. These data suggest that carrier strains are able to suppress the normal TLR-2 pathway, thereby reducing innate immune responses and imposing risks of more serious invasive disease as well as decreasing the barrier immunity given by cationic antimicrobial peptides. Finally, 15 SAB MRSA isolates from a single patient with persistent bacteremia were fully sequenced and shown to harbor a total of 37 sequence polymorphisms (Gao et al., 2015). This study identified how organisms evolved during treatment. Of note, the chromosome expanded to include a 20 kb tandem repeats encompassing mprF and parC, as well as transient mutations in genes such as relA and rplC, which can play a role in resistance to endogenous host defense peptides and peptide-based antibiotics and contribute to small colony variant (SCV) formation. A more recent study by Giulieri et al. (2022) examined the evolution of 2590 genomes in 396 S. aureus strains during infection in vivo. Results demonstrated a distinctive evolutionary pattern within the invasive S. aureus populations (early and late adapted) compared to those colonizing the nares. Late infection-adapted strains were enriched for variants predicted to interfere with intergenic regulatory regions through point mutations or insertions. However, several genes with the most significant enrichment (agrA, agrC, stp1, and sucA) were recurrently mutated across multiple host in vivo evolutionary scenarios, implying a global role in S. aureus adaptation during colonization and invasion. The mutations in succinate metabolism are consistent with adaptations seen in persistence (discussed below). Adaptation within the invasive population was strongly driven by antibiotics, e.g., enhanced MprF production following daptomycin pressure. This is a small sampling of the genes identified and the reader is referred to this paper. Together, understanding how these factors promote colonization vs. infection will allow novel strategies for vaccines or immunotherapeutics to prevent invasion or infection.

S. aureus employs many methods to adapt to the host environment including adaptive mutations in essential genes involved in pathogenesis and covert chromosomal rearrangements that promote invasive and/or persistent infections. S. aureus has surface proteins that bind fibronectin: fibronectin binding protein A (FnbpA) and FnbpB. Early studies correlated the numbers of fibronectin receptors expressed on S. aureus with overall invasive disease (Proctor et al., 1984). In these studies, 22 isolates from patients with invasive infections were compared to 19 isolates in patients with non-invasive infections. The differences were highly significant (p < 0.00025). Moreover, fibronectin was shown to enhance uptake of S. aureus by cultured endothelial cells (Hamill et al., 1986). Further studies refined these concepts. Two polymorphisms in S. aureus FnbpA were identified that promoted pathogenesis. One polymorphism increased adhesion (VDREED) to host fibronectin (Lower et al., 2011), which resulted in enhanced binding to endovascular devices, while another polymorphism (GIDRVED) enhanced uptake into host cells (Xiong et al., 2015). Isolates that had higher binding affinities were more likely to cause device-related infections in patients (Lower et al., 2011; Hos et al., 2015). Additional work examined five strains isolated from patients with persistent S. aureus bacteremia versus five strains with resolving S. aureus bacteremia (SAB) (Xiong et al., 2015). Isolates with increased affinity for host fibronectin were better able to initiate endovascular infections whereas other variants showed reduced affinity but increased invasion of host cells. Thus, binding to fibronectin does not always mean more intracellular uptake, as it is cell-type specific (Niemann et al., 2021). For endothelial cells, this appears to be true, but for osteoblasts large amounts of fibronectin can reduce uptake, via formation of fibrillar networks that inhibit uptake.

Additionally, staphylococcal adaptation for persistence also includes phenotype switching mechanisms that occur as a result of chromosomal rearrangements without any apparent genetic mutations (Guérillot et al., 2019). The inversion that was described involved half of the chromosome. This inversion occurred via recombination between methylase genes, and the activation of a conserved prophage harboring the Immune Evasion Cluster (IEC). IEC contains three genes (chi, scn, sak), which promote evasion of neutrophils and has been associated with persistent infections. Moreover, this inversion led to slow growth, enhanced host cell invasion, low hemolysin and pigment production, and reduced agr expression, phenotypes which are typical of SCVs. Cui et al. (2012) have also identified large symmetrical inversion (1.26 Mb) impacting two inverted copies of the S. aureus pathogenicity island (SaPl) that also resulted in the SCV phenotype, but this inversion does not appear as often as the one described by Guérillot et al. (2019). Thus, there is increasing recognition of S. aureus strain-specific characteristics that temporospatially exploit host cells or tissues that may differ considerably from human to human.

Colonized humans seem to have better outcomes in invasive infection than those who are not colonized (Wertheim et al., 2004). Nevertheless, the fact that the large majority of invasive infections are caused by the strain of S. aureus colonizing the nares and can evolve over time supports a need to revise conventional thinking on “commensalism” (Proctor, 2021). S. aureus is increasingly being referred to as a human pathobiont, especially given its ability to undergo adaptive mutations in response to host-mediated stresses and antibiotic exposure. S. aureus strains undergoing adaptive changes within the nose may make them more likely to invade the host, cause more severe disease and result in persistent infections. Hence, nasal colonization has serious impacts which imply a non-commensal interplay between the bacterium and the host.

Previously, we had reviewed data suggesting that staphylococcal toxins may suppress the immune response, especially in the context of the T cell—phagocyte axis (Miller et al., 2020). This concept has been further explored wherein each step in the interactions between antigen-presenting cells and T cells can be impacted by toxins (Teymournejad and Montgomery, 2021). Examples of toxin interference include LukAB killing of dendritic cells (Berends et al., 2019), LukED killing of dendritic and T cells when binding to CCR5 (Alonzo et al., 2013), alpha-toxin (α-hemolysin or Hla) inducing naïve T cell apoptosis (Nygaard et al., 2012), and TSST-1 interfering with differentiation of naïve T cells is given as a single example of the many examples of enterotoxins and superantigens altering T cell development. Or course, there is also the broader dysregulation of cytokine response caused by these toxins (Kametani et al., 2008). Strong evidence for these events impacting host responses to S. aureus has been presented in a prospective study of 76 children (31 had invasive infections and 45 non-invasive infections) wherein T cell and antibody responses were measured over time (Li et al., 2022). One specific rationale behind this study was the observation that S. aureus infections recurred in 50% of neonates within 12 months of birth. Major findings included globally impaired T-cell function in infected babies, with 5- to 12-fold reduction in IL-17A and IFNγ in response to S. aureus, while anti-LukE, -LukS-PV (but not Hla) increased following infections in those less than 5 years old. This raises questions regarding whether children with suppressed T-cells get S. aureus infections, or if S. aureus toxins suppress T-cells, or both?

Further evidence for S. aureus impairment of T cell-mediated immunity was provided by Sanchez et al. (Sanchez et al., 2017), in which humans failed to develop protective immunity after S. aureus bacteremia (SAB). The investigators proposed that the mechanism underlying this result was O-acetylation of peptidoglycan (PG), limiting Th17 response via pattern recognition receptors (PRRs). If so, it would follow that such immune subversion strategies would be strain-dependent. However, there are numerous other mechanisms by which S. aureus may subvert or deviate immune responses that would otherwise be protective against recurrent infection.

Recent evidence suggests that skin-resident T cells are involved in host defense against S. aureus skin infections (Clegg et al., 2021). For instance, healthy human skin is typically replete with S. aureus-specific tissue-resident memory CD4⁺ T cells that produce IL-17A, IL-22, IFNγ, and TNFα; in response to heat-killed S. aureus (Hendriks et al., 2021). Such responses indicate that adaptive immunity does develop against S. aureus in the skin.

However, whether these responses are protective in the presence of S. aureus toxins or can be targeted by vaccines for durable protection against recurrent infections warrants further investigation.

While T cell immunity has been emphasized as important for protection against invasive S. aureus infections (Armentrout et al., 2020), mispolarized, profusive or protracted T cell responses have also been implicated in adverse responses (Proctor, 2012). For example, the V710 vaccine (recombinant IsdB) stimulated protection via Th17 pathways that were implicated in adverse outcomes in clinical trials (Joshi et al., 2012). The data monitoring committee stopped the V710 vaccine trial because of such hyperimmune responses in subjects who had become infected with invasive S. aureus infections. A follow-up study to examine possible explanations for these results showed that patients who died after receiving V710 vaccine had constitutively low IL-2 levels before vaccination. Clearly, the vaccine disclosed this immune dysregulation correlate, substantiating the broader hypothesis that a protective but restrained immune response is necessary for vaccine safety and efficacy. Because IL-2 is needed for regulatory T cell (Treg) development, it is conceivable that patients with low IL-2 levels were deficient for generation of Treg cells that control systemic inflammatory response. As with most S. aureus immune responses, the Treg response is complex. Hence, more work is needed to sort out their exact role in patients with low IL-2 (Gaborit et al., 2020; Gao et al., 2022). Similarly, IL-17-producing T cells that are associated with hyperactive and pathogenic immune responses are induced in the presence of commensals such as staphylococcal species specifically upon the use of immune checkpoint inhibitors (Hu et al., 2022). This raised concerns over whether all commensal-specific immune responses are pathogenic in the context of anti-tumor therapy and how cancer patients would fare on Th17 stimulating vaccines. An additional observation comes from the V710 clinical trial. Infections in the vein graft donor site were significantly reduced in those subjects receiving the vaccine (Fowler et al., 2013). It is known that Th17 plays a very important role in skin infections (O'Brien and McLoughlin, 2019), hence this suggests that vaccination outcomes may be very location dependent. This concept has been termed contextual immunity and represents distinct temporospatial signatures of molecular and cellular response to infection (Yeaman et al., 2014; Chan et al., 2017). Thus, while T cell immunity is likely critical in protective efficacy, a balance between T cell subsets and their polarization paradigms that is optimal for protective immunity and minimizes off-target effects is necessary for safe and effective vaccines (Proctor, 2012; Yeaman and Hennessey, 2017).

Beyond the above observations, new information continues to emerge regarding the ways through which S. aureus appears to subvert the human immune system. For example, Uebele et al. (Uebele et al., 2020) reported that S. aureus protein A induces Treg numbers through interactions with antigen presenting cells. Askarian et al. (Askarian et al., 2018) reviewed the mechanisms through which S. aureus modulates innate immune responses via toll-like (TLR), NOD-like (NLR) and C-type lectin (CLR) receptor interactions. These are pattern recognition receptors that bind S. aureus lipoproteins and peptidoglycan, which results in transcriptional upregulation of inflammatory pathways in macrophages. This interaction provides an early activation system of innate immunity that results in bacterial clearance. However, S. aureus has developed ways to evade these systems. For example, S. aureus deficient in lipoproteins fail to activate TLR2 pathways. Moreover, S. aureus is able to inhibit heterodimer formation, use structural mimicry of the TIR domain to prevent activation of TLR2, and activate inhibitory receptor pathways [reviewed in (Askarian et al., 2018)]. Similarly, 12% of S. aureus produce a protein, TirS, which down-regulates NF-κB pathways that are critical for production of pro-inflammatory cytokines via NOD2 (Askarian et al., 2018). Staphylococcal evasion factors for CLR recognition have not been described to date.

Recent reports from Medie et al. (2019) and Chang et al. (2021) provide compelling evidence that poor outcomes in vancomycin treatment of S. aureus bacteremia (SAB; e.g. severity, persistence, etc.) are associated with host epigenetic signatures driving immune polarization. Interestingly, anti-IsdB antibodies serve as a negative biomarker as they correlate with poor outcome (IsdB is also highly immunogenic). Hence, it is conceivable or even probable that S. aureus may induce antibody subclass switching for evasion or exploitation of immune responses. Collectively, such findings suggest that prior experience (colonization, infection, invasion) with S. aureus may shape subsequent immune memory and protective efficacy. However, an outstanding question relating to these findings focuses on how S. aureus has evolved to attenuate or exploit host immune responses to its advantage.

Our previous review detailed the imbalances in cytokine release in those patients that did well versus those with a poor outcome in SAB (Miller et al., 2020). High IL-10 (anti-inflammatory cytokine) and low IL-1 and IL-6 (critically important cytokines early in the immune response) are associated with worse outcome. A more detailed examination of the cytokine response to S. aureus infections can be found in our previous review (Miller et al., 2020), and newer data on human genetic susceptibility is considered here. Some of the genetic variability is related to methylation patterns of the human genome that relate to the control of these cytokines.

Staphylococcal infection involves adaptation to numerous host selective pressures within specific niches. The metabolic microenvironment directs activity of both the host immune response as well as that of the bacteria that must compete for favored substrates and micronutrients. Many of the so-called S. aureus virulence factors are gene products that are regulated by the abundance of specific metabolites. Such regulatory systems enable bacteria that have disseminated to various tissues to rapidly adapt, often using metabolic pathways that are distinct from those expressed during bloodstream infection. The concept that central metabolism serves as a signaling pathway in S. aureus (Somerville and Proctor, 2009) is now being substantiated by multiple studies reviewed below.

Isolates that have undergone in vivo adaption often harbor loss of function mutations in the master virulence regulator agr (Traber et al., 2008; Suligoy et al., 2018), yet these strains are often better able to persist [reviewed in (Tuchscherr et al., 2020)]. Other adaptive mutations (gain or loss of function) occur in metabolic genes (McAdam et al., 2011; Acker et al., 2019; Gabryszewski et al., 2019; Giulieri et al., 2022), highlighting the crucial role of S. aureus metabolic flexibility in bacterial survival and persistence. For example, in glucose-replete conditions, the catabolite control protein A (CcpA) enables S. aureus to utilize the preferred carbon source glucose over secondary carbon sources such as amino acids (Seidl et al., 2009). However, when glucose is depleted, as is often the case in infected tissues such as the lung, as host and bacterial cells compete, catabolite repression is alleviated to enable rapid bacterial consumption of available substrates. In this section, we will discuss the metabolic crosstalk between bacterial and innate immune cells and how the dynamics of this interaction influences the infection outcome, often to the detriment of the host.

Upon bacterial stimulation, immune cells are activated and undergo profound metabolic changes that dictate their immune function, a process termed metabolic reprograming (reviewed in (O'Neill et al., 2016) ( Figure 1 ). This includes an increase in glycolysis, which is accompanied by the accumulation of the tricarboxylic acid cycle (TCA) metabolites including succinate and itaconate. Succinate stabilizes the host transcription factor, hypoxia-inducible factor-1α (HIF1-α), promoting the production of the pro-inflammatory cytokine IL-1β. In contrast, itaconate restores homeostatic balance following succinate-driven inflammation via its anti-inflammatory and anti-oxidative roles (Lampropoulou et al., 2016; Qin et al., 2019; Hooftman et al., 2020). There have been multiple demonstrations that metabolically active S. aureus stimulates macrophage and keratinocyte glycolysis and HIF1-α activation (Wickersham et al., 2017; Acker et al., 2019; Wong Fok Lung et al., 2020). This was shown to be dependent on the ability of S. aureus itself to perform glycolysis, highlighting host-bacterial cell competition for glucose. Immunometabolites are increasingly appreciated as drivers of bacterial adaptation and thus have an important role in pathogenesis. Moreover, a better understanding of how S. aureus adapts to specific microenvironments may provide novel, tissue-targeted strategies for staphylococcal eradication.

Among the key mysteries needing to be solved to accelerate the development of safe and effective vaccines targeting S. aureus is the identification of specific determinant(s) of protective immune efficacy. As can be surmised from the above sections, no single cellular subset or molecular class is solely responsible for protective immunity to this elusive pathogen. Furthermore, neither adaptive cellular immune mechanisms nor antibodies are sufficient for protection against in patients who face recurring S. aureus infection.

Vaccines demonstrating success in experimental models have provided hints regarding biomarkers for protection that remain to be fully explored in humans. Anti-toxin antibody levels against Hla, TSST-1, PVL, and SEA correlate with less severe disease including pneumonia, skin infections, osteomyelitis, and pyomyositis in the majority of studies (Miller et al., 2020). However, a direct quantitative correlation cannot always be made between presence of antibody and the infection outcome. For example, antibody subclasses IgG1 and IgG3 in humans have greater opsonophagocytic potential and likely contribute to more effective cell-mediated immune clearance of the pathogen than do IgG2 and IgG4. The opposite may be true for the neutralizing capacity of S. aureus exotoxins. Affinity maturation of antibody subclasses further complicates the specificity of antibody response. Of course, B cell production of antibody, particularly IgG, is governed by T cell recognition and crosstalk with B cells identifying cognate epitopes. Thus, the net protective efficacy of immune response to S. aureus infection or mitigation of severity likely results from a complex intersection of host and pathogen factors, including host genetic make-up, history of colonization or infection (immune memory), as well as S. aureus strain-specific combinations of immune evasion factors, antigenic epitopes, exotoxin signature and expression of these virulence factors. Moreover, neither the host immune response nor the adaptive pathogen virulence is static; rather, there is considerable dynamic interplay of host and pathogen in distinct temporospatial contexts.

Of the SA toxins, pore-forming toxins (PFTs) and superantigens (SAgs) have the most profound effect on the immune system. PFTs disrupt the skin barrier by killing keratinocytes (Soong et al., 2012), the first line of defense against SA i.e., polymorphonuclear neutrophils (PMNs), macrophages, and to lesser extent T cells (LukED) (Vandenesch et al., 2012; Aman and Adhikari, 2014). Uncontrolled lysis of PMNs leads to release of cytotoxic and proinflammatory molecules with pivotal consequences to host health and can also promote the dissemination of bacteria contained within phagosomes. Toxins are the only class of vaccine targets backed by strong human epidemiological and clinical data (Humphreys et al., 1989; Lina et al., 1999; Gillet et al., 2002; Barretti et al., 2009; Jacobsson et al., 2010; Rasigade et al., 2011; Adhikari et al., 2012; Bachert and Zhang, 2012; Fritz et al., 2013; Tong et al., 2015). Presence of pvl gene is associated with abscesses and furuncles (OR=10.5 (95% CI, 7.4 to 14.9) (Shallcross et al., 2013). The role of Hla in SSTI has been widely reported (Berube and Bubeck Wardenburg, 2013; Chua et al., 2014) (Inoshima et al., 2012). SAg-induced cytokine storm causes toxic shock (McCormick et al., 2001) and SAg expression has been related to septicemia (Humphreys et al., 1989; Jacobsson et al., 2010). PFTs and SAgs are implicated in SA skin colonization and exacerbation of atopic dermatitis (AD) that predisposes patients to recurrent SA SSTIs (Humphreys et al., 1989; Jacobsson et al., 2010; Brauweiler et al., 2013; Brauweiler et al., 2014). Most AD isolates produce SAgs (McFadden et al., 1993). SAgs are likely a major factor in the Th2 inflammatory response in AD patients (McFadden et al., 1993) and facilitate epithelial presentation of allergens to Th2 cells (Ardern-Jones et al., 2007; Forbes-Blom et al., 2012). Adhesion molecules that bind to SA in the skin like fibronectin and laminin are upregulated via the Th2 cytokine IL-4 and suppress IFNγ or TNFα induced release of antimicrobial peptides by keratinocytes (Cho et al., 2001). Thus, neutralization of PFTs and SAgs in the skin can change the immunological environment to unfavorable conditions for colonization and infection.

A large number of human clinical and epidemiological studies provide evidence for a protective role of antibodies against toxins in S. aureus disease. Lower levels of antitoxin antibodies in patients correlates with increased probability of complications (Jacobsson et al., 2010; Adhikari et al., 2012; Fritz et al., 2013). Skin infections with bacteria that produce PVL cause a higher frequency of abscess and need for incision and drainage (Gubbay et al., 2008; Jahamy et al., 2008; Munckhof et al., 2008; Kaltsas et al., 2011). Higher level of antibodies to Hla correlate with reduced rate of recurrent skin infection (Fritz et al., 2013). Hla production also correlates with poor resolution of staphylococcal peritonitis in continuous ambulatory peritoneal dialysis (CAPD) patients (Barretti et al., 2009). Children with the lowest antibody levels against Hla and the PVL component LukF are more likely to develop invasive disease (Fritz et al., 2013). Menstrual TSS patients that do not seroconvert are more likely to experience recurring bouts (Kansal et al., 2007). Humphreys et al. characterized 52 strains from septic patients and 27 strains from healthy SA nasal carriers. Of the septic isolates, 63% produced SEA, SEB or SEC, whereas 11% of the nasal isolates produced a SAg (Humphreys et al., 1989). Azuma et al. detected SAgs in the plasma of 31% of septic patients without septic shock and 41% of those with septic shock (Azuma et al., 2004). It is important to note that to-date studies have primarily evaluated total antibody titers (mostly total IgG) rather than toxin neutralizing titers. Functional antibody titers are more likely to be informative biomarkers than total IgG.

A series of recent studies in animal models highlight the importance of staphylococcal protein A (SpA) as a key virulence factor. Protein A (SpA) is expressed on the surface of SA and as secreted protein, and best known for its interaction with the Fc domain of IgG, preventing Ig hexamer formation required for C1q binding and subsequent bacterial clearance (Sharma-Kuinkel et al., 2019), and reducing interaction of IgG with the neonatal Fc receptor (FcRn) at pH6 shortening antibody half-life (Stapleton et al., 2011). Furthermore, SpA binds indiscriminately to the Fab portion of VH3 lineage antibodies and causes polyclonal expansion of these clones (Goodyear and Silverman, 2003). These SpA-mediated immune evasion mechanisms are believed to be critical for SA pathogenesis based on a series of animal studies showing that anti-SpA neutralizing antibodies as well as vaccination against SpA broadens the antibody response to other SA antigens and protects from lethal disease (Thammavongsa et al., 2015). A recent study showed that anti-IsdB antibodies in IsdB-immunized mice form complexes with SpA, haemoglobin and haptoglobin on the surface of SA (Nishitani et al., 2020). The complex is phagocytosed along with SA by a CD163 dependent mechanism into macrophages. Macrophages then act as trojan horse disseminating bacteria systemically leading to sepsis, suggesting a role for SpA in the failure of IsdB vaccine in the Merck clinical trial. However, clear evidence for the role of SpA in human disease remains to be explored.

Finally, a biomarker for good outcome has been identified that is not associated with toxin neutralization. Low levels of anti-glucosaminidase (Gmd) levels in patients with S. aureus osteomyelitis had a 2.7-fold increased chance of poor outcome. Gmd is the lytic portion of AltA (major autolysin) (Kates et al., 2020). Anti-Gmd antibodies that neutralize autolysin activity improve outcomes of invasive S. aureus infection (Sherchand et al., 2022).

There have been multiple attempts at producing a S. aureus vaccine that will prevent serious infections or mitigate their severity. Despite earnest efforts, none of these has yet to demonstrate safety and efficacy in human clinical trials. This result is in no small measure due to the fact that the complexities of both the virulence strategies of S. aureus and the human immune response are incompletely understood. In many respects, constitutive colonization with intermittent infection distinguishes this microorganism from many other bacterial pathogens. This host-pathogen relationship of course relates to the long evolutionary time span over which S. aureus has resided within or cultivated niches such as the nares of mammalian hosts. We are now learning that S. aureus enters human cells to subvert host defenses, and changes the basic metabolism of the host cells, which in turn has repercussions on the nature of the host immune response elicited. This is a two-way street as inflammatory immune cells, which sense the presence of intracellular S. aureus via its metabolic products or via competition for available substrates, also impose a strong selective pressure on the bacteria. This immunometabolic pressure is exemplified by the direct effects of metabolites such as itaconate on the bacteria that drive staphylococcal adaptation for survival. Of note, these host-staphylococcal metabolic cross-talks emphasize the importance of using live bacteria in infection models as opposed to bacterial proxies such as pathogen-associated molecular patterns (PAMPs). These complex interactions are just beginning to be understood.

Detailed information about the human genetic susceptibility to S. aureus infection is in its infancy, but we can already see that some humans seem predisposed for worsened outcomes, persistence or recurrent infection. For example, prior to recombinant IsdB vaccine administration, patients that developed a systemic inflammatory response to S. aureus invasion had pre-existing low IL-2 levels. This example raises key questions regarding potential dysregulation of Treg responses, which may fail to limit adverse inflammatory responses. There is also a chicken-versus-egg mystery remaining to be solved regarding host-pathogen relationships if we are to develop safe and effective vaccines for this pathogen. Surprisingly, a large proportion of children who become infected with S. aureus demonstrated impaired T-cell function. Further, patients with SNPs in CD207 of Langerhans cells or NLRP3 of phagocytes may be more prone to skin infections. Such correlates point to potential heritable as well as acquired dysfunction in protective immunity to S. aureus.

Because S. aureus and mammals have lived together for so long, preventing invasive disease presents unique and complex challenges for vaccine development that have yet to be solved. More detailed understanding of protective determinants of immunity to S. aureus and new experimental models in which to test candidate vaccines may provide new insights to overcome current hurdles. Even though these represent a very high bar, the continuous development of cutting-edge techniques and the potential leveraging of integrated omics data as a more holistic approach bring us closer to our goal. Alternatively, because S. aureus infections represent myriad forms of disease reflecting distinct temporospatial contexts, it may prove necessary to create vaccines aimed at protecting specific tissues such as skin or in persons with individual genetic make-up—truly precision meets personalized medicine.

In looking to the future, we may be able to pair conventional vaccine targets with novel host-directed strategies with the aim of decreasing the selective pressure on the bacteria while promoting immune clearance ( Figure 3 ). The field is currently on the threshold of testing multivalent anti-toxin vaccines with the hopes that recurrent skin infections can be reduced. Future vaccine attempts should also consider addressing potential immune imbalances such as low IL-2 levels/reduced Tregs to avoid the production of over exuberant immune responses in subgroups of vaccinated patients. Also, consideration will need to be given to the enhancement of M1 polarization of macrophages whose activities are down-regulated by the metabolic tricks used by S. aureus to disarm the immune response. Compounds such as oligomycin (OXPHOS inhibitor), which showed promise following nanoparticle delivery in a mouse model of S. aureus prosthetic joint infection, might be considered as adjuvants to the natural immune response. However, a deeper understanding of immunometabolic interactions within specific tissues is required. Because multiple studies have reported that increased IL-10 levels correlate with poor outcomes, perhaps anti-IL-2 therapies may improve outcomes. Finally, the novel anti-Gmd antibodies, while opsonic, are also directed at the fission plane, resulting in mega-clusters of S. aureus that are unable to disseminate. While anti-Gmd antibody levels have not yet been validated as a biomarker in human clinical trials, it serves as a potential example biomarker candidate of successful outcome and therefore a vaccine directed at Gmd may hold promise to reduce mortality in S. aureus infections.

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.