Staphylococcal adaptation to diverse physiologic niches: an overview of transcriptomic and phenotypic changes in different biological environments

Abstract

Host niches can differ strongly regarding, for example, oxygen tension, pH or nutrient availability. Staphylococcus aureus and other staphylococci are common colonizers of human epithelia as well as important human pathogens. The phenotypes that they show in different host environments, and the corresponding bacterial transcriptomes and proteomes, are currently under intense investigation. In this review, we examine the available literature describing staphylococcal phenotypes, such as expression of virulence factors, gross morphologic characteristics and growth patterns, in various physiological environments. Going forward, these studies will help researchers and clinicians to form an enhanced and more detailed picture of the interactions existing between the host and staphylococci as some of its most frequent colonizers and invaders.

The study of effects that the host environment has on bacteria is crucial to understanding pathogenesis and discovering new avenues for the treatment of infection. Staphylococcus aureus, a Gram-positive coccal species, permanently or transiently colonizes a large portion (up to 60%) of the human population and is thus generally considered part of normal human microbiota. However, it has the potential to cause many diseases, including chronic and severe infections such as persisting skin and soft tissue infections (SSTIs), infectious endocarditis, infections of prostheses, septic arthritis and osteomyelitis [1]. Environmental factors that can vary widely within the host, such as pH, nutrient availability or oxygen tension, can induce S. aureus to express a variety of different phenotypes [2–4]. Host environmental factors that contribute to the pathogenesis of staphylococcal disease are generally acknowledged to be of great importance, but are at the moment poorly characterized [1].

Presently, heavy emphasis is placed on researching genetic cues within staphylococci that are crucial to pathogenesis. For the purpose of fully understanding mechanisms of pathogenesis, researchers commonly examine phenotypes that result from the artificial perturbation of bacterial genes. In vitro investigation of molecular pathways using such strains is performed under a highly controlled environment and with laboratory media. Most in vivo studies, while more clinically relevant, typically evaluate outcomes of bacterial infection using gene deletion strains (e.g., the ability of an infection to thrive with or without expression of key virulence genes), or assess strains that are isolated from clinical samples, cultured in laboratory media and tested for the presence or absence of genes associated with virulence.

Further information can be derived from the bacterial genome composition; and certain clones of S. aureus show stronger propensities than others for causing specific infections, such as SSTIs [5]. However, since bacterial behavior depends strongly on the environment, gene presence or absence alone only gives limited information about host–pathogen interaction. In the past, genome-wide analysis of gene expression in samples obtained from clinical or experimental infections has been challenging due to difficulties in obtaining sufficient amounts of RNA at high purity [6]. However, recent technical advances have allowed researchers to take a closer look at the impact of the host environment on bacterial gene expression and in vivo phenotypic characteristics. Furthermore, proteomic analyses are increasingly being employed to decipher protein expression under different conditions and in vivo environments. In this review, we present and discuss current knowledge about the effect of different in vivo or simulated in vivo environments on the behavior, expression of virulence factors, growth, antibiotic susceptibilities and gross morphological characteristics of staphylococci, focusing on S. aureus as the most important human pathogen.

S. aureus virulence factors

Virulence is defined as the ability of a pathogen to reduce host fitness as well as the ability of an organism to establish an infection and cause disease in a host. S. aureus produces a vast array of virulence factors, some of which we briefly present in the following, which are involved with diverse virulence mechanisms, such as immune evasion, colonization, immune stimulation or lysis of host cells. Notably, researchers have put much emphasis on the search for genes necessary for staphylococcal virulence, but much less consideration has been afforded to elucidating the specific effects of the host environment on these virulence-associated genes.

S. aureus produces a series of toxins that subvert mechanisms of innate host defense [7], and many different toxins that can be classified as cytotoxins, pyrogenic (fever-producing) toxins, enterotoxins and exfoliative toxins [8]. Studies that were performed using gene deletion strains have shown that several staphylococcal toxins are instrumental in causing massive tissue damage. Among the most potent cytotoxins is the phenol-soluble modulin (PSM) family of peptides [9]. Expression of PSMs is controlled by the agr locus, which also controls many other toxins such as hemolysins (e.g., α-toxin) and leukotoxins [10,11]. Often, toxins are encoded on mobile genetic elements; and thus, in those cases, the analysis of gene presence alone may give important information on strain toxicity [8]. This is the case for many of the toxins that block mechanism of innate host defense as well as some cytotoxins (e.g., Panton–Valentine and other leukocidins) [12]. In contrast, several toxin genes such as those coding for PSMs or α-toxin are invariably present in S. aureus isolates. S. aureus isolates may differ strongly in the capacity to produce those latter toxins, which directly correlates with virulence in animal models [13,14].

Biofilm formation is known to be key to the pathogenesis of persistent staphylococcal infection, such as infective endocarditis and prosthetic joint infection. During biofilm-associated infection, the bacteria form sessile communities on surfaces or in aggregates with host tissue [15]. This phenotype contributes to immune evasion and recalcitrance to antibiotic treatment [16]. Genetic manipulation has frequently been used to decipher which genes are needed for staphylococcal biofilm formation [17]. An extracellular biofilm matrix forms to serve the dual purpose of containing the staphylococci and shielding the bacteria from immune cell surveillance, all while simultaneously permitting diffusion of nutrients into the biofilm. One very important biofilm extracellular matrix molecule is the exopolysaccharide polysaccharide intercellular adhesin, PIA, which is also known as poly-N-acetyl-glucosamine, or PNAG [18,19]. The PIA/PNAG biosynthetic operon ica comprises the icaA, icaB, icaC and icaD genes [20]. Together, IcaA, IcaB, IcaC and IcaD produce, modify and export PIA into the extracellular space, where it adheres to bacterial surfaces and provides scaffolding for the accumulation of other biofilm factors [21,22]. Biofilm structuring, dispersal and dissemination of biofilm-associated infection are dependent upon factors that are controlled by the Agr quorum-sensing system, most notably PSMs [23].

In many S. aureus strains, immune evasion is enhanced by a capsule, consisting primarily of hexosaminuronic acid polymers, that provides protection from phagocytosis. This capsule is formed by gene products of the cap gene locus, of which several variants exist, the most frequent of which produce capsules of serotypes 5 and 8 [24]. Capsule synthesis is a demonstrated virulence determinant in S. aureus infection [25], but it is absent from several clinically relevant strains such as USA300 [26]. It is subject to regulation by a variety of environmental factors and regulatory systems [24].

Surface proteins, such as accumulation-associated protein (Aap), extracellular matrix binding protein (Embp), protein A (SpA), the fibronectin-binding proteins (FnbpA and FnbpB) and other similar proteins, in addition to establishing initial contact with host matrix proteins, contribute to strengthening cell–cell adhesion [27]. The difference in staphylococcal gene expression between the biofilm and planktonic states is quite marked [28–30]. In an in vitro analysis of global gene expression in S. aureus biofilms, 48 genes were induced at least twofold in biofilm cultures compared with planktonic cultures. Additionally, 84 genes were repressed by a factor of two or more (e.g., spa and agr genes) [28]. This may be due to the change in the microenvironment that is known to exist once bacteria form biofilms; in mature biofilms, the pH decreases and anaerobic respiration increases. Similar observations have been made in other staphylococci, including S. epidermidis [30].

It is crucial to consider the varying conditions that may exist within each host niche in order to determine what effects the host niche may impact on the ability of staphylococci to effectively colonize and cause damage to the host.

Main regulatory pathways in staphylococci

S. aureus has a number of regulatory circuits, which allow control of virulence determinants such as adhesion proteins and toxins, in addition to metabolic targets, in response to the host environment. Among the most intensely studied are the quorum-sensing system Agr [31], the small transcriptional regulator proteins of the SarA family [32], the components of the alternative sigma factor SigB [33] and several two-component systems such as Arl and Srr [34,35]. Agr is tightly controlled by bacterial density. When sufficient cell density is reached, Agr is activated by a secreted autoinducing peptide (AIP), which through signal transduction via the AgrAC two-component system, the regulatory RNAIII and the repressor of toxins (Rot) DNA-binding protein generally induces production of toxins and exoenzymes and decreases production of several surface proteins [31]. SarA is the prototype of the Sar regulator family [32]. It controls many of the extracellular and cell wall associated virulence proteins and is known to serve as a direct activator of the agr operon [36]. The SigB regulatory pathway, which is also a regulator of SarA expression, modulates the expression of surface proteins such as clumping factors [37–39]. The Srr pathway comprises a 2-component system that is responsive to oxygen tension [40]. The ArlRS two-component system is involved in the downregulation of some virulence genes such as α-toxin and protein A, is important for cell growth, division and colonization of polymer surfaces [34,41]. All of these pathways are activated/repressed in a temporal and environment-dependent manner and are interconnected in a way to permit bacterial infection in the host.

Important environmental cues

Perhaps the most frequently considered environmental factors of the various host niches are oxygen tension, pH and nutrient starvation. S. aureus is facultative and responds to oxygen limitation through the induction of anaerobic respiration and direct changes in regulatory gene expression, mediated by the SrrAB (Staphylococcal respiratory response) pathway [42]. SrrAB is an upstream effector of ldh, the TCA cycle enzymes, agr and ica [35]. The switch from aerobic to anaerobic respiration is associated with profound changes in the transcription patterns of approximately 9% of all S. aureus genes [42]. Decreased oxygen tension activates SrrAB and directly influences SrrA binding to the promoter region of the agr P2 promoter, which controls cytolytic proteins and quorum sensing [43]. Conversely, SrrAB activation under decreased oxygen tension directly promotes the production of ica transcripts, leading to increased biofilm formation [44]. These effects, however, have only been shown in laboratory media, and further research is necessary to determine whether similar changes are observed in physiological fluids. In vivo, hypoxic conditions have been reported in abscesses, the urogenital tract and within synovial joints and the spine (disc) [42,45,46].

Altered gene expression can also be due to shifts in pH. The pH values of host environments vary widely. Most important for staphylococci is the shift from the slightly acid conditions on human skin as their premier habitat to higher pH values encountered during infection, such as a pH of approximately 7.4 in human blood. During in vitro growth of S. aureus, the accumulation of metabolites can cause a 2-unit decrease in pH, which causes over 400 transcripts to change significantly, independent of nutrient availability. Notably, agr-controlled transcripts, many of which of virulence-associated genes, are increased at lower pH, as present in the skin, urogenital tract, lung, mouth and within abscesses [47]. There are conflicting results about the effect of pH on biofilm formation. In one in vitro study, decreased pH (5.5) was reported to induce staphylococcal biofilm formation [47], whereas in another study, a decrease to pH 4.5 did not appear to induce biofilm/PIA formation, induce agr-controlled virulence factors or promote enterotoxin production [2]. Clearly, more comprehensive in vitro and in vivo studies will be necessary to better determine the impact of the pH on staphylococcal phenotypes.

Nutrient availability and osmolarity of the growth environment strongly influence staphylococcal gene expression. For example, high NaCl concentration (1 M) decreases the growth rate of S. aureus and the transcription of protein A (spa), α-toxin (hla) and toxic shock syndrome toxin 1 (tst)-encoding genes by at least tenfold, while only a moderate effect of high salt concentration on sarA and agr expression was reported [48] Similarly, varying concentrations of sucrose, from 0.2 to 1 M, reduced expression of spa, hla, tst and sarA, but had no effect of agr expression [48]. In contrast, increased concentrations of glucose repress agr expression in multiple strains of S. aureus [49]. High osmolyte levels have been reported to repress the expression of some agr-controlled exoproteins such as Eta and staphylococcal enterotoxins B and C [48,50–52]. Furthermore, the availability of metal ions such as Mg²⁺, Ca²⁺, Fe²⁺ and Zn²⁺ is crucial to the expression of several virulence genes, for example, tst [48,53,54]. Finally, analysis of in vitro and in vivo transcriptional changes under low-iron conditions showed an impact of Fe²⁺ concentration on agr-controlled transcripts [55].

In vivo, there are a variety of compounds, either produced endogenously (antimicrobial peptides [AMPs], reactive oxygen species, etc.), or introduced exogenously (e.g., antibiotics), that can alter the host environment and may cause a change in staphylococcal gene expression. In vitro experiments have been performed to determine the effects of these antibacterial agents on gene expression in staphylococci. AMPs form an important part of innate host defense on human epithelia [56]. Several studies have investigated the impact that AMPs have on gene expression of staphylococci [57–60]. Central to the transcriptional adaptations of staphylococci to the presence of the usually membrane-damaging AMPs is the increased expression of bacterial defense mechanisms, which include increased expression of the mprF, dlt and vraFG transcripts [58,59]. MprF is an enzyme that produces lysyl-phosphatidylglycerol, whose incorporation in the cytoplasmic membrane decreases interaction with the commonly cationic AMPs [61]. The dlt operon is responsible for the D-alanylation of teichoic acids, decreasing attraction of cationic AMPs to the bacterial surface [62]. Finally, the ABC transporter VraFG may be directly involved in AMP resistance by AMP expulsion [59], but it may also form an integral part of the staphylococcal AMP sensing mechanism [63], which is based on binding of AMPs to the ApsS (GraS) histidine kinase and the ApsRSX (GraRSX) three-component signal transduction system [58].

Another method that the host uses to combat infection is the production of reactive oxygen species. Macrophages and other lymphocytes use reactive oxygen species (such as hydrogen poeroxide and superoxides) to kill S. aureus. A microarray study performed to assess the effects of H₂0₂ on S. aureus gene expression showed a notable decrease in the expression of the ica locus, indicating that H₂0₂ production inhibits the ability of S. aureus to produce biofilm [64]. Finally, there are multiple studies that have shown an impact of subinhibitory antibiotic concentrations of antibiotics on the expression of virulence genes, such as hla, PSMs and toxic shock syndrome toxin (TSST-1) [65–67].

Host niches may influence host–pathogen interaction not only depending on their physico-chemical composition, but also depending on the interactions between host and pathogen proteomes. For example, tissue adherence and colonization are facilitated by the interaction of S. aureus with host matrix proteins, which are present in different amounts in the various host niches. Staphylococcus aureus expresses a family of surface proteins, called ‘microbial surface components recognizing adhesive matrix molecules’ (MSCRAMMs), which binds host matrix proteins such as collagen, fibronectin, fibrin and fibrinogen [27,68]. In addition to impacting attachment to host tissue, MSCRAMMs contribute to bacterial aggregation and biofilm formation. Different strains of S. aureus have varying expression of MSCRAMMS [69]. Notably, many MSCRAMMS are under, usually negative, regulation by Agr [11,69]; and thus, environmental cues that impact Agr expression will also impact the expression of MSCRAMMs.

Table 1 summarizes possible environmental cues in each host niche, as well as the corresponding reported S. aureus adaptations. The following is a summary of data collected from studies on S. aureus asymptomatic colonization or infections of various host niches.

Table 1. . Major Staphylococcus aureus adaptations in response to environmental cues .

Host niche	Potential environmental cue(s)	Ref.	Reported adaptation phenotype and factors up/down- regulated	Ref.
Skin and soft tissues	Varying humidity, pH, temperature, microbiota	[76,78,136]	Upregulated hlg, spa, hla, lukS-PV, lukE	[43,87–89]
			Downregulated agr, psmβ
Urinary tract	Varying pH	[47,137]	Slow growth	[130,131]
			Biofilm formation
Nares	Low pH	[101,102,138]	Dispersed phenotype rather than biofilm (?)	[3,101]
	Low nutrient availability
			Downregulation of agr
	AMPs
	Mucociliary interactions
	Competing microflora
Reproductive tract	Varying oxygen tension and pH	[139]	Biofilm formation	[133]
			Low production of virulence factors
Lungs	Antecedent viral infection	[104,140]	Biofilm formation, PIA expression	[105]
	Varying pH		Downregulation of agr
			Upregulation of hld, psm
Blood/vasculature	Presence of serum proteins	[141–145]	Aggregation	[43,107,111,112,117,146]
	Presence of red and white blood cells, platelets		Spread/dissemination
	Flow/shear stress		Differential expression of agr, psm (conflicting results)
			Increased production of surface proteins
Bone/joints	High mineral content (bone)	[46,147–150]	Aggregation/biofilm formation intracellular persistence (osteoblasts/chondrocytes)	[118,124,151]
	Fibrous tissues		Antibiotic recalcitrance
	High viscosity, varying oxygen tension and cellularity, protein composition (synovial fluid)		Decrease in agr, psm
			Increase in surface protein (FnbpA, ClfA) gene expression
CSF	Ultra-filtrate of serum with CSF-specific proteins (e.g., cystatin), low cellularity	[152]	N/A	N/A

Skin colonization & infection

Coagulase-negative staphylococci (CoNS) such as S. epidermidis are common colonizers of human skin [70,71]. Staphylococcus aureus permanently colonizes the nares and ano-genital areas of at least a third of the human population [72,73]. Furthermore, S. aureus is a common cause of SSTIs [74,75]. The milieus on human skin can differ strongly, not only from human to human, depending on lifestyle and hygiene, but also from one anatomic location to the next [76,77]. Temperature, humidity, pH, lipid composition and production of AMPs vary considerably over time and location [76,78]. Thus, it is difficult to simulate the conditions staphylococci encounter during SSTIs and even more so during asymptomatic skin colonization. Abscess models with subcutaneous or intradermal injection of bacteria may give a valuable representation of human SSTI, but there are virtually no models that mimic staphylococcal skin colonization with acceptable accuracy. Some researchers have characterized the interactions of S. aureus with skin through the use of skin equivalents. In skin equivalents, collagen and fibroblasts are combined to form a dermal matrix; and keratinocytes are subsequently seeded onto the matrix and permitted to differentiate, thus simulating a skin-like substrate [78].

While S. aureus is not primarily thought of as an intracellular organism, internalization of S. aureus by a variety of cell types is believed to contribute to persistence and the phenomenon of recurring infection [79,80]. In addition to surviving in macrophages [81], S. aureus is known to persist intracellularly in epithelial cells [82]; both mechanisms likely matter during SSTI. In one study that examined gene expression of S. aureus after internalization in human epithelial cells, a significant downregulation of agr in the internalized bacteria was shown [83]. Specifically, AgrA was shown to be downregulated by approximately tenfold by microarray analysis and 40-fold by qPCR validation as compared with noninternalized bacteria cultured in the presence of cells. In the same study, the virulence regulator sarS and the gene encoding superoxide dismutase (sodA), an enzyme detoxifying reactive oxygen species, were strongly induced by internalization in an in vitro simulation of epithelial cell colonization [83]. However, in another study, phagosomal escape and survival in epithelial cells were attributed to the Agr-controlled PSMs [84], which is in line with similar findings in macrophages and neutrophils [84,85], and suggest upregulation of Agr and PSMs after internalization. In a microarray study specifically addressing gene expression changes after neutrophil phagocytosis, genes involved in capsule synthesis, oxidative stress and virulence, were upregulated following ingestion. Agr showed slight down-, while the virulence regulator Sae showed slight upregulation [86]. Overall, gene expression changes after internalization appear to be dependent strongly on the experimental setup and will need to be verified in in vivo settings.

There is also an increasing number of in vivo studies examining gene expression in human or experimental SSTI. In a subcutaneous infection model in rabbits, Yarwood et al., using subgenomic DNA microarrays, found a 17-fold decrease in RNAIII levels, representative of Agr expression, as compared with expression in Todd–Hewitt medium [43]. Another study focused on virulence gene expression during human cutaneous S. aureus infection. Using material obtained directly from 40 clinical cutaneous abscesses, a relative upregulation of cytolytic toxin gene transcripts, such as hla, lukS-PV, lukE and hlgB, and a relative decrease in RNAIII, bacteriocin (bsa) and protein A (spa) transcripts, as compared with laboratory broth, was found using qRT-PCR [87]. Researchers also monitored transcriptomic changes by microarray in a rabbit model of skin infection with the community-associated MRSA strain USA300 and noted that virulence factors such as leukocidins, as well as extracellular matrix binding protein expression increased dramatically (e.g., between five- and eightfold for leukocidins, and 12-fold for fibronectin-binding protein) within 1 day of infection [88].

Recently, Date et al. showed that in a human cutaneous abscess, the expression levels of 262 genes were significantly altered. Using microarray analysis of three clinical samples, these authors reported that 10.1% of genes were upregulated, and 7.3% of genes were downregulated, as compared with a 4-h culture grown in tryptic soy broth [89]. While some toxins, such as the Panton–Valentine leukocidin (PVL) and γ-hemolysin, were among the most highly upregulated in these cutaneous abscesses, others such as PSMs, were significantly downregulated. Interestingly, the transcription profile in human cutaneous abscesses was remarkably similar to that in infected mouse kidneys [89]. These results may reflect the importance of certain toxins in SSTI, but they also show the limitations of an approach that only measures overall transcriptional changes of bacteria in abscesses, disregarding specific micro-environments and interaction with cells. PSMs, for example, are known to be upregulated after internalization in professional and nonprofessional phagocytes [84,85]. Thus, changes observed in the entire abscess material may not necessarily mirror well-established roles of virulence factors in the pathogenesis of SSTI that are specific to the interaction with a subset of cells.

Nasal & respiratory tract colonization & infection

The nares as the predominant niche of S. aureus colonization present a unique environment; and several staphylococcal factors influencing nasal colonization have been analyzed, such as teichoic acids and specific surface proteins [90,91]. Animal models of nasal colonization use cotton rats or mice [92,93]. Results from such models have shown that there is a distinct pattern of gene expression in the nasal ecological niche. For example, virulence factors controlled by the agr locus are weakly expressed, as assessed by qRT-PCR of samples collected from four persistently colonized individuals [94]. Nasal colonization with staphylococci, particularly S. aureus [95], has long been associated with subsequent infection (see below), although approximately 30% of the general population carries S. aureus asymptomatically [72]. Nevertheless, this association has driven decolonization efforts, particularly for the purpose of prophylaxis prior to many cardiothoracic and orthopedic procedures [96,97]. S. aureus in the nasal niche is therefore routinely eradicated with antibiotics and does not appear to be recalcitrant to antibiotics [98–100]. This, and the finding that swabs of the nares only contain low amounts of S. aureus, have led researchers to believe that S. aureus in the nares prefer a dispersed rather than a biofilm phenotype [101]. In one study, human nasal secretions were analyzed by metabolomics and found to contain nutrients in rather low amounts, which led to the creation of a growth medium for the in vitro simulation of the nasal environment, called synthetic nasal medium (SNM) [3]. Bacterial culture of S. aureus USA300 in laboratory medium and SNM showed that in SNM, there was no increase in transcript levels of RNAIII, psmβ or the clumping factor gene clfB, as measured by qRT-PCR [3]. These in vitro findings are in accordance with those of an in vivo study, which was performed using swabs taken directly from nasal S. aureus carriers (all asymptomatic) and showed by qRT-PCR that there was no activation of agr in the nasal niche [94].

Respiratory infections commonly originate from nasal colonization and traumatic injury or underlying conditions such as influenza or immune deficiencies. Certain clones are better suited to persist in the respiratory tract, indicating that the immune environment in the respiratory tract is greatly important to the pathogenesis of S. aureus respiratory tract infections and supporting the notion that a compromised immune system, as in the case of postviral pneumonia or in hosts with immunodeficiency syndromes such as chronic granulomatous disease (CGD) [102,103], permits S. aureus to take hold in the respiratory tract. In a postinfluenza model of S. aureus pneumonia, expression of α-toxin, a factor known to be crucial for S. aureus lung infection, was greatly decreased, and antecedent viral infections caused significant alterations in bacterial gene expression by qRT-PCR analysis [104]. In a murine study of early lung infection, the transcriptome of S. aureus at 6 h post infection was analyzed using microarrays. Principal component analysis revealed an overall downregulation of most virulence factors controlled by agr. Interestingly, however, transcript levels of all PSMs were greatly increased (between eight and 13-fold) [105].

Blood & vascular infection

Staphylococcal dissemination in the blood, known as bacteremia, is a life-threatening condition, and can result in sepsis and acute shock, as well as long-term colonization of the vasculature, including cardiac valves, as seen in endocarditis [106]. Endocarditis is largely associated with intravenous drug users, who introduce S. aureus directly into the blood stream through contaminated needles or poor sterilization of the injection site [107,108]. There is a shortage of data available concerning gene expression of S. aureus in human cardiac vegetations; however, one study has shown that there is reduced expression of hla in vivo [109].

A study of infective endocarditis in a diabetes mellitus model assessed the S. aureus MRSA strain COL transcriptome within endocardial vegetations compared with in vitro growth. Microarray analyses showed that 61 genes, mainly associated with metabolism, were upregulated. Among toxins and toxin-related genes, no significant difference was detected in the expression of agrA, hld or hla. However, there was a significant increase in the expression of the leukocidin lukF and lukS genes in cardiac vegetations compared with controls [110].

In human blood, gene expression is altered compared with laboratory medium. Some studies showed that agr is suppressed in human serum or whole blood [43,111], while also reporting that expression of the Agr-controlled γ-hemolysin gene was strongly upregulated when measured by USA300-specific oligonucleotide microarrays [111]. During S. aureus bacteremia in human patients, surface proteins (e.g., FnbpA and ClfA) were upregulated by more than twofold, and levels of PSM expression were low according to microarray analysis, which is in accordance with suppression of agr under those conditions [112]. Downregulation of Agr was assumed to represent a likely reason for the formation of S. aureus aggregates in blood, as this results in increased expression of several MSCRAMMs that bind, for example, fibrin and fibrinogen. However, it needs to be stressed that Agr control of MSCRAMMs can be strongly strain-dependent [69]. Based on results from other infection types and in vitro investigation, low PSM production likely further contributes to aggregation [113–115]. Notably, S. aureus bacteremia isolates often have a dysfunctional Agr system [116], which is in accordance with a beneficial role of Agr downregulation for S. aureus survival in human blood. Contrasting these studies, another study performed using calf serum reported upregulation of Agr and Agr-controlled virulence determinants including PSMs, as assessed by qRT-PCR [117]. Whether this contradiction is due to species-specific differences or the experimental setup is not clear.

Bone & joint infections

Osteomyelitis, septic arthritis and prosthetic joint infections are notoriously insidious and difficult to treat [118–120]. Staphylococci have been shown to persist in chondrocytes and osteoblasts [121], thereby evading antibiotics, making treatment without surgical invention nearly impossible. There is no information on gene expression in humans or animals during joint infections, but limited studies on virulence factor expression have been performed in an ex vivo synovial environment.

One study by Szafranska et al. determined the adaptive response of S. aureus during osteomyelitis. Using high-resolution transcriptomic analysis and mouse models of both chronic and acute osteomyelitis, these authors found that 444 genes were differentially expressed in S. aureus comparing in vivo acute osteomyelitis and in vitro control experiments. Among these, were several virulence factors such as PSMs and α-toxin. Interestingly, levels of agr expression were significantly lower under in vivo as compared with in vitro conditions. Matrix-binding genes such as ClfA and FnbpA were also found to be significantly upregulated in osteomyelitis (both acute and chronic) [122].

Bacteria in the synovial environment are especially recalcitrant to antibiotic treatment and difficult to detect [123], a situation that has recently been explained by an extraordinary strong propensity of staphylococci, as the most frequent causes of joint infections, to form aggregates in synovial fluid. In that environment, fibrin- and fibronectin-binding proteins on the surface of S. aureus contribute to the formation of large, fibrous agglomerations [124] (Figure 1A). Staphylococcus aureus undergoes immediate aggregation upon incubation in synovial fluid through the engagement of those MSCRAMMS with fibrous matrix components (Figure 1B) [124]. In general, aggregation behavior in synovial fluid holds many similarities to the behavior of S. aureus in blood. However, synovial fluid, a filtrate of serum with added proteins, causes a more pronounced biofilm phenotype as well as larger aggregate formation than serum [124]. In synovial fluid, there is significantly decreased expression of agr and PSMs (when assessed by qRT-PCR) compared with serum and laboratory medium, which is believed to be the cause for the exceptionally strong degree of S. aureus aggregation during joint infection [113]. Decreased expression of the biofilm-dispersive PSMs means that after initial aggregation has occurred, biofilm adhesion is increasing and disaggregation is inhibited [113] (Figure 2). This leads to the buildup of PIA/PNAG and likely other matrix components, such as teichoic acids and eDNA, within the dense matrix, and results in biofilm-like aggregations that are recalcitrant to antibiotics [113,124].

Figure 1. . Staphylococcus aureus in joint infections visualized by scanning electron microscopy.

(A) Biofilm isolated from a patient with orthopedic infection. (B) Rapid aggregation of S. aureus after 20 min incubation in synovial fluid.

Figure 2. . Aggregation of bacteria and formation of biofilm in response to host fibrin.

(A) Staphylococcus aureus binds to human matrix proteins present in traumatized joints (such as fibrin) using dedicated, surface-located binding proteins (ClfA, ClfB, FnbA, FnbB). (B) The low activity of Agr in fibrin-mediated aggregates causes low production of the surfactant PSMs, with the result of pronounced cell-to-cell attachment and retention of biofilm macromolecules on the cell surface. This ultimately leads to the formation of enormous, macroscopic cellular aggregates with high resistance to antibiotics.

Infections of the CNS

Staphylococcal infections of the CNS are less frequent than those of the skin, mucosa, and bone and joints. The few cases that are reported usually mention the formation of abscesses, disseminated disease or cavernous sinus thrombosis, and are believed to result from superficial skin infections, sinusitis or otitis media [125]. Postsurgically, infections are present in the form of shunt infections and ventricular drain infections. Fortunately, antibiotic penetration into the cerebrospinal fluid is sufficient for antimicrobial activity [126], which is why S. aureus infections of the CNS tend to respond positively to antibiotics, suggesting that biofilm formation is likely less pronounced in CSF than in other fluids, such as serum or synovial fluid. No data are available on staphylococcal gene expression during infection of the CNS.

Infections of the urinary & reproductive tract

Among the staphylococci, Staphylococcus saprophyticus is the pathogen most frequently involved with urinary tract infections (UTIs) [127]. Specific adhesins and biofilm formation play key roles for the pathogenesis of S. saprophyticus UTI, but how the environment in the urinary tract impacts gene expression of S. saprophyticus is unknown [128]. In contrast to S. saprophyticus, S. aureus is not as common a cause of UTI. S. aureus only represents <5% of outpatient and 16% of inpatient UTIs [129], and is primarily a cause of UTIs among patients with urinary tract catheterization. In accordance with the notion of a key role of biofilms during UTI, S. aureus was shown to form biofilms in the urinary catheter lumen [130]. Interestingly, urine has been reported to be inhibitory for S. aureus growth [131]. For ex vivo studies, pooled urine is sometimes used to simulate the urinary environment [132]. Nevertheless, no studies have been performed on the transcriptome of S. aureus or S. saprophyticus in urine.

S. aureus is one of the most prevalent species present in the male and female genital tracts. However, S. aureus overgrowth in the genital tract is normally inhibited and overpowering infection is rarely a concern, except for in the immunocompromised, or where there is the presence of a foreign body (tampons, intrauterine devices, etc.) [133]. Gene expression by S. aureus in the reproductive tract has not yet been studied, but biofilm formation by S. aureus on tampons and on menses components has been confirmed through fluorescent in situ hybridization [133]. Not much is known about the phenotype of staphylococci in the reproductive tract, but there is a link of S. aureus colonization to infertility [134]. Specifically, the expression of spermagglutinins by S. aureus in the vaginal tract causes aggregation and decreased motility of sperm leading to decreased fertilization [135]. Other than infertility and the commonly reported toxic shock syndrome due to staphylococcal overgrowth on foreign bodies, no other clinical manifestation of S. aureus in the vaginal tract has been reported. Likely, the presence of other bacteria, the pH and the relatively low oxygen tension keeps S. aureus in check. Similar to other host niches, much work remains to be conducted on staphylococcal gene expression in both the male and female reproductive tracts.

Conclusion

The host environment can have significant effects on staphylococcal gene expression and overall phenotype. The control of gene expression in response to environmental conditions is crucial for effective staphylococcal colonization and infection. Of particular importance is the ability to switch between planktonic and biofilm phenotypes in response to the host environment. However, our understanding of how staphylococci respond to specific micro-environments in the host is still very limited.

Future perspective

Staphylococcal virulence, growth, colonization and antibiotic susceptibility vary from niche to niche in the human host. These differences require tailoring of therapeutic approaches, or decolonization measures, to specific infection types and locations, taking into account the different patterns of gene expression and phenotypic changes that occur within each environment. Many of the studies cited in this review describe conflicting results. Inconsistencies may be a result of varying qualities and rigors of study, and must be resolved using state-of-the art methodologies in future studies. It is apparent that our knowledge about staphylococcal phenotypes in most colonization and infection scenarios is still very limited. Going forward, the improvement of techniques to determine transcriptomic and proteomic changes, such as by RNA sequencing or MS-based approaches to analyze protein composition, will play a key role in those endeavors. Furthermore, genetic and imaging tools have been recently developed that allow us to measure gene expression levels in in vivo settings that closely resemble the situations bacteria encounter during human infection and colonization. Thus, it is to be expected that there will soon be much more detailed insight into host niche-dependent differences in staphylococcal phenotypes and pathogenesis strategies, hopefully leading to a significant improvement of therapeutic approaches to combat staphylococcal infections.

EXECUTIVE SUMMARY.

Staphylococci in diverse physiologic niches

• Each physiologic niche (e.g., the nares, skin, joints, etc.) comprises specific environmental characteristics, which can strongly affect gene expression.

• Bacterial virulence in each niche can be heightened or diminished, depending on the environmental factors present.

• It is crucial to understand the effects of each physiologic niche on the ability of staphylococcal strains to cause infection, or maintain colonization.

Important environmental cues

• Gene expression can be altered by variations in pH, nutrient availability, oxygen pressure and other host factors. Many of those factors impact biofilm production and production of agr-controlled virulence factors.

• Staphylococcus aureus is more prone to biofilm formation where serum proteins and ECM components are abundant (e.g., in cardiac valves or synovial joints).

Staphylococcus aureus gene expression in various host niches

• Transcriptomic analyses of samples from in vivo infection or colonization are very limited.

• In vitro transcriptomic analyses closely emulating the in vivo environment are currently being developed (e.g., nasal media, ex vivo use of physiological fluids, use of physiological pH or oxygen tension).

• Future studies should expand comparisons of gene expression in various physiologic niches, using direct isolation from in vivo infections or ex vivo simulations adequately representing in vivo situations.