When antibiotics fail: a clinical and microbiological perspective on antibiotic tolerance and persistence of Staphylococcus aureus

Abstract

Staphylococcus aureus is a major human pathogen causing a vast array of infections with significant mortality. Its versatile physiology enables it to adapt to various environments. Specific physiological changes are thought to underlie the frequent failure of antimicrobial therapy despite susceptibility in standard microbiological assays. Bacteria capable of surviving high antibiotic concentrations despite having a genetically susceptible background are described as ‘antibiotic tolerant’. In this review, we put current knowledge on environmental triggers and molecular mechanisms of increased antibiotic survival of S. aureus into its clinical context. We discuss animal and clinical evidence of its significance and outline strategies to overcome infections with antibiotic-tolerant S. aureus.

Introduction

Staphylococcus aureus is one of the most versatile pathogens infecting humans. It harmlessly colonizes one-third of the human population in the nares or throat or on the skin,¹ but can also lead to a vast array of potentially lethal invasive infections. S. aureus can infect virtually any tissue of the body such as skin, muscles, bones and joints, lungs, kidneys, heart or even the eyes or brain.^2,3 Moreover, it readily grows on artificial surfaces of implanted foreign bodies such as orthopaedic devices, prosthetic heart valves or intravascular and urinary catheters. More than three decades ago, S. aureus was named the ‘persistent pathogen’⁴ due to its unusual recalcitrance towards antibiotic treatment. This has remained, to date, a major clinical challenge. S. aureus can persist for several days in the blood of up to one-third of patients despite correct antibiotic treatment;^5–9 this is by far more common than with other bacteria.^8,10 The associated mortality rate of 15%–50% leads to a calculated annual 2–10 deaths/100 000 population.¹¹ Finally, current antibiotics have limited capacity to eradicate S. aureus in abscesses or in implant infections without the support of surgery.^11–14 Physicians choose antibiotics according to a strain’s susceptibility measured as the MIC in the microbiology laboratory. When treatment nevertheless fails, this can have several clinical or pharmacological reasons (Table 1), many of which are preventable through optimized application of the antibiotics. However, evidence is emerging that the ability of S. aureus to become antibiotic tolerant represents a major microbiological reason for the high rate of treatment failure.

Earlier studies arbitrarily defined antibiotic tolerance as in vitro measurement of a ratio of MBC to MIC of ≥32. This definition, however, does not encompass the various populations and mechanisms of tolerance and its measurement is fraught with considerable variability.^15,16 Therefore, more recently, leading experts have provided clarification about the terminology defining the effect of antibiotics on bacterial growth and survival (for details, see Balaban et al.¹⁷). Accordingly, tolerance describes the bacterial ability to survive usually lethal concentrations of a bactericidal antibiotic without change of the MIC.^17,18 This can classically be visualized and measured in a slower killing rate in in vitro time–kill curves. However, so far, no new classification or measurable cut-offs have been introduced that may be applicable to clinical practice. From a population-level perspective, the term tolerance applies if an entire bacterial population shows increased survival whereas the term persistence is used if only a small fraction of a population (named persisters) exhibit a tolerant phenotype. Although the clinical relevance of this differentiation is not clear, we will use the term ‘increased antibiotic survival’^19,20 to encompass both tolerance and persistence.

In this review, we put mechanisms and triggers of increased antibiotic survival into the context of various types of infection and life forms of S. aureus. By taking a broad clinical perspective, we extend the scope of previous reviews on tolerance and persistence in general^17,18,21 as well as on S. aureus in particular.^22,23 With the goal of translating basic research to clinical practice, we provide an overview of highly discussed as well as rather neglected studies of in vitro, in vivo and clinical aspects of antibiotic-tolerant S. aureus. We discuss approaches of how to overcome infections with antibiotic-tolerant S. aureus and identify future research questions and open gaps in knowledge.

Environmental triggers of increased antibiotic survival of S. aureus

Development of tolerant phenotypes can be interpreted as a survival strategy of a bacterial population in quickly changing and potentially toxic environments.²⁴ Tolerant phenotypes can develop due to genetic traits,^25–29 can evolve stochastically^30,31 or can be triggered by specific environmental factors¹⁸ in both Gram-positive and Gram-negative bacteria. The majority of studies on increased antibiotic survival have focused on Gram-negative bacteria.³² Basic mechanisms of becoming tolerant are probably common for both Gram-positive and Gram-negative bacteria.^20,21,33,34 However, not only do the identified molecular targets leading to tolerant phenotypes differ but S. aureus also stands out through the variety of evolved life forms causing infections associated with characteristics of increased antibiotic survival. The scientific pioneer Bigger had already noticed by 1944 that growth-inhibiting conditions induce increased antibiotic survival in S. aureus.³⁵ Renewed research on the topic has now confirmed these observations and found several environmental factors such as low oxygen level,³⁶ low pH³⁷ or low amount of nutrients³⁸ as responsible for triggering tolerant phenotypes (Figure 1).

Most mechanistic in vitro studies have focused on the free-floating planktonic single-cell life form of S. aureus in liquid culture. High planktonic population density is typically strongly associated with development of increased antibiotic survival. This is due to the exhaustion of growth-sustaining factors and accumulation of metabolic waste products or bacterial signalling molecules.³⁹ However, it is not known which role such tolerant planktonic forms of S. aureus play in the clinical context. Single cells or small aggregates of S. aureus dividing in the interstitial space are most likely responsible for several acute infections such as pneumonia or skin and soft tissue infections. Short antibiotic regimens generally cure such infections¹² and recurrence is uncommon. Hence, even if tolerant planktonic subpopulations survive the exposure to antibiotic treatment in these acute infections, they presumably do not withstand killing by the immune system in the unprotected environment of the interstitial space.

S. aureus increases its survival in vivo through the formation of multicellular conglomerates. Indeed, aggregating into cell clusters seems to be an intrinsic growth pattern of S. aureus,⁴¹ a phenotype constantly observed since its first discovery in the 1880s and serving as inspiration for its microbiological name (Greek: σταφνλοσ meaning grape cluster).⁴² The tendency to cluster dramatically increases in the presence of fibrinogen/fibronectin, which is abundant in synovial fluid of joints (e.g. in septic arthritis) or in intravascular fibrin and platelet clots (e.g. in endocarditis).⁴³ These cell clusters can survive antibiotic exposures that would usually be bactericidal even without attachment to a surface,^43,44 requiring prolonged antibiotic treatment to cure the respective infections. Inside a tissue, staphylococcal cell clusters can evolve into abscesses by digesting and transforming the surrounding tissue into a protective capsule of fibrin and necrotic debris.⁴⁵ Antibiotics in the absence of surgery generally fail to eradicate S. aureus in abscesses.⁴⁶ This has been attributed to reduced antibiotic penetration and physicochemical or enzymatic inactivation.⁴⁷ However, since several antibiotics can achieve adequate concentrations inside an abscess,⁴⁷ tolerant S. aureus in the acidic and low-oxygen environment of an abscess may play an important role in their recalcitrance against antibiotics. Concordantly, tolerant subpopulations have been isolated from abscesses manifesting as non-stable small colony variants (SCVs) with extended lag periods before regrowth.⁴⁸

S. aureus can further increase its survival by forming multicellular biofilms encased by a self-produced matrix on organic or inorganic surfaces.^49,50 Prototypical examples of such biofilm infections include endocarditis or implant-related infections.^51,52 These are clinically relevant due to their ability to survive more than 100–1000-fold higher antibiotic concentrations compared with planktonic S. aureus.^53–56 In such cases, increased survival cannot be explained by lack of biofilm penetration since even large antibiotics such as vancomycin,^57,58 daptomycin⁵⁹ or telavancin⁶⁰ easily penetrate S. aureus biofilms. Gradients of oxygen and nutrients inside staphylococcal biofilms lead to spatial patterns of diverse physiological states with different levels of cellular respiration, DNA replication and protein synthesis⁶¹ and therefore potentially diverse levels of increased survival. Non-growing, tolerant fractions of biofilm bacteria are thought to be responsible for the prolonged survival of biofilms during exposure to bactericidal antibiotics requiring extended treatment and responsible for relapses once treatment ends.⁶² The number of tolerant cells in a biofilm increases with age and cell density,^63,64 supporting the clinical concept that after a certain time of biofilm maturation, prosthetic implants need to be surgically removed to achieve cure.^65,66 Moreover, biofilm-derived detached cell clusters exhibit a higher level of antibiotic survival than purely single-cell planktonic bacteria^67,68 and may play a major role in perpetuating and spreading an infection.

With its versatile physiology, S. aureus can also survive by invading professional and non-professional phagocytic cells.^69–73 This may lead to treatment failure⁷⁴ and even the spread of infection through haematogenous transport inside host cells.⁷⁵ Intracellular S. aureus has been documented in skin and soft tissue, bone and upper airway infections^72,76,77 and is probably involved in an even wider spectrum of chronic and recurrent infections.^58,75,78 The intracellular compartment is a protected niche since it shields S. aureus from the immune system and from the high extracellular antibiotic drug concentrations.^75,79,80 However, the specific metabolic state of intracellular S. aureus further reduces the lethality of cell-penetrating antibiotics.^81,82,S. aureus from intracellular locations can manifest on the laboratory growth plate as an SCV due to its slowed or delayed growth and hence eponymous reduced colony size.^83–85 Compared with normal phenotypes, SCVs not only show reduced metabolism but also reduced expression of regulator and virulence factors.⁸⁴ SCVs from clinical samples often present as so-called ‘non-stable’ or ‘dynamic’ SCVs, since they rapidly revert to normal colony sizes and metabolism after reculture on rich media.^86,87 SCVs have been found to be tolerant to aminoglycosides and cell-wall active antibiotics^88,89 and are associated with recalcitrant and relapsing infections.^90,91 Nutrient starvation, oxidative stress and low pH are characteristics of the encountered intracellular milieu and all induce the formation of SCVs.^85,92,93 Concordantly, alkalinization of the phagolysosome by the antimalarial drug chloroquine reduced survival in antibiotic treatment of intracellular S. aureus.⁹³ Finally, exposure to host defence mechanisms can stimulate SCVs and other tolerant phenotypes in S. aureus during an infection.^85,94

Taken together, S. aureus has developed various life forms to persist in different niches of the human body and thereby encounters or produces environmental triggers of increased survival, rendering an antibiotic treatment ineffective.

In vitro evidence for mechanisms of increased antibiotic survival of S. aureus

Understanding the mechanisms of antibiotic survival is essential in order to improve therapeutic strategies against S. aureus infections. Reduced cellular metabolism leading to decreased growth rate or dormancy seems to be a shared feature of most tolerant bacteria.^18,30,95,96 This may not just be a passive state but may involve active metabolic changes focusing on adaptation to a hostile environment.^24,97–99 Various molecular targets have been found to change the level of antibiotic survival in S. aureus (Table 2). Individual targets may only induce survival to some antibiotics tested whereas other antibiotics may stay effective (Table 2). However, separate studies have found some conflicting results, which likely reflect the complexity of mechanisms and the sensitivity of its regulation. Results may differ greatly depending on the genetic background of the individual strains and on the experimental setup.¹⁷ The growth phase strongly influences the level of antibiotic survival since slow- or non-growing S. aureus from high-density stationary-phase cultures are considerably more tolerant to antibiotics than when in exponential growth.^{30,37,100–104} A low intracellular level of ATP may represent a mechanism of antibiotic survival for S. aureus³⁰ as seen in other bacteria.¹⁰⁵ The reduced cellular fuel in the form of ATP slows down or halts cellular processes targeted by antibiotics and thereby protects the bacterial cell from their bactericidal action. Nevertheless, low ATP levels seem not to be a prerequisite, since increased antibiotic survival has been found in S. aureus with normal ATP levels.^106,107 Moreover, antibiotic survival may not always be associated with decreased growth according to a study that found a higher rate of cells tolerant to vancomycin in a more rapidly growing WT S. aureus compared with its mutant with defects in purine synthesis.¹⁰⁸

Deletion of the Clp proteolytic system [involved in degradation of several antitoxins in S. aureus toxin/antitoxin systems¹⁰⁹ and in regulation of the stress response^110,111 as well as the tricarboxylic acid (TCA) cycle¹¹²] was associated with increased rates of tolerant S. aureus.¹¹³ Similarly, deletion mutants of the S. aureus toxin/antitoxin system mazF/mazE exhibited decreased antibiotic survival after exposure to β-lactam antibiotics or vancomycin in planktonic as well as biofilm growth.^114,115 It should be cautioned that this effect on antibiotic survival could not be confirmed after simultaneously knocking out three major toxin–antitoxin systems including MazF/MazE.³⁰ Nevertheless, dysregulation of the MazF/MazE system can phenotypically lead to growth inhibition¹¹⁶ and SCVs⁸⁵ as an indirect clue to its association with antibiotic survival in S. aureus. Further, several studies found increased antibiotic survival of S. aureus associated with activation of the guanosine tetraphosphate ppGpp-dependent stringent response and vice versa.^117–119 This may not be true for all strains or other experimental setups according to one conflicting study.³⁰ Nonetheless, parallels to phenotypes of tolerant life forms again exist since higher levels of ppGpp have been associated with SCVs¹²⁰ and intraphagosomal survival.¹²¹ More than 40 years ago, the lack of autolysis was found to lead to increased survival with β-lactam antibiotics in S. aureus.¹²² More recent studies were able to associate deficiencies in CidABC and LrgA, both regulators of extracellular murein hydrolases, to increased survival through lack of autolysis.^123,124 Concordantly, WGS associated mutation in the murein hydrolase atl with in vitro-generated phenotypes tolerant to oxacillin.¹²⁵ Addition of glucose increased the activity of CidABC and LrgAB, thereby simultaneously decreasing the level of antibiotic survival.¹²⁶

Other studies have specifically investigated mechanisms of increased survival of S. aureus with daptomycin, a membrane-targeting bactericidal antibiotic known for its growth-independent killing capacity.¹⁰¹ Changes in the membrane-lipid composition in the stationary phase increased antibiotic survival to daptomycin treatment. This observation could be influenced by the lipids available in the growth medium.¹⁰⁴ Similarly, up-regulation of the dlt operon, encoding proteins responsible for d-alanylation of teichoic acids in the Gram-positive cell wall, increased daptomycin-specific antibiotic survival.^127,128 However, changes in available glucose can diminish the rate of daptomycin-tolerant S. aureus.¹²⁹ Recently, a genome-wide analysis identified several new genomic regions associated with daptomycin-tolerant S. aureus, again a strong hint for the plasticity and versatility of mechanisms of antibiotic survival in S. aureus.²⁹

Studies have found a higher level of antibiotic survival in mutants with defects in the quorum sensing accessory gene regulator (agr).^107,130 Dysfunctional agr may promote tolerant phenotypes through increased biofilm formation¹³¹ or through excreted factors interfering extracellularly with antibiotics such as gentamicin¹³⁰ or daptomycin.¹³² Such increased antibiotic survival is supported by clinical data linking agr mutants to persistent infections.^133–136

Antibiotics themselves may trigger antibiotic survival by ‘toxifying’ the environment of S. aureus. Indeed, evidence suggests that antibiotics can not only induce biofilm¹³⁷ and SCV^85,138 formation, but also trigger other mechanisms of antibiotic survival.^139,140 In the case of S. aureus, repeated in vitro exposure to antibiotics can lead to increased antibiotic survival, as shown for β-lactams,^25,118,141 aminoglycosides,¹⁴² vancomycin¹¹⁸ or daptomycin.¹²⁷ For vancomycin, the proposed mechanisms include the induction of the ppGpp-dependent stringent response¹¹⁸ and direct blocking of the access of murein hydrolases to their cell-wall substrates.^134,143 Moreover, pre-treatment with an antibiotic can result in increased survival not only for the applied antibiotic but also for other antibiotic classes.^144–146 This cross-effect can vary greatly depending on which antibiotics are chosen^{97,103,147,148} and even which sequence of antibiotics is selected.¹⁴⁴ It stretches from full cross-effect (as shown for tobramycin followed by ciprofloxacin) to no cross-effect (as shown for tobramycin and daptomycin).¹⁴⁴ Finally, the widely used topical antimicrobial triclosan can increase antibiotic survival up to 10 000-fold, a mechanism again requiring ppGpp synthesis.¹⁴⁹

Taken together, multiple mechanisms may mediate increased antibiotic survival of S. aureus. It remains to be seen whether they act in parallel or as redundant or upstream/downstream pathways. Nevertheless, the selective effect on individual antibiotics is a strong clue that different mechanisms of antibiotic survival are present. Clearly, a better understanding of the mechanisms of cellular death mediated by individual antibiotics should help decipher the various mechanisms underlying antibiotic survival.

In vivo experimental evidence for increased antibiotic survival of S. aureus

Although knowledge about in vitro phenomena of increased antibiotic survival is extensive, how this translates in vivo has not been ascertained. Moreover, there is a paucity of in vivo data on the outcome of infections with tolerant S. aureus since most available studies were performed more than 20 years ago. An early animal study by Chuard et al.¹⁵⁰ gave proof of the impact of the in vivo environment on antibiotic survival. Measuring the S. aureus MIC and MBC of various antibiotics before and after a 6 week untreated foreign-body infection revealed such a drastic rise of the MBC that even the highest antibiotic concentrations could no longer kill the recovered S. aureus, notably without changes to the MIC. Various other animal studies have shown both the failure of antibiotics to eradicate in vivo infections, possibly due to tolerant phenotypes,^55,151,152 and the increased in vivo survival of S. aureus already determined in vitro to be tolerant.^108,153,154 However, not all studies found a correlation of in vitro with in vivo antibiotic survival.^155,156 In summary, in vivo studies show the relevance of the host environment on antibiotic survival but also reveal the difficulty of translating in vitro measurements into in vivo differences of outcome.

Evidence of the clinical impact of increased antibiotic survival of S. aureus

Although bedside experience and the growing in vitro evidence argue for a major role of antibiotic-tolerant phenotypes of S. aureus in the clinical setting, the number of studies analysing antibiotic survival of S. aureus from patients is limited. The awareness of the clinical relevance of the antibiotic killing capacity, however, is not new and earlier studies have analysed the correlation between in vitro antibiotic survival and clinical outcome (Table 3). The MBC/MIC ratio, the most commonly used clinical measurement of tolerance, is fraught with considerable variability.^15,16 Nevertheless, up to now it has been the best available marker of its clinical relevance. Most studies focus on vancomycin and describe a rate of tolerant S. aureus from 6%–43% in clinical strains;^157–164 however, up to 63% oxacillin-tolerant S. aureus have been reported in patients with endocarditis.¹⁶⁵ Single case reports describe treatment failures with tolerant strains.^166–168 Larger series present a wide spectrum of clinical outcomes associated with ‘in vitro tolerant’ S. aureus. It ranges from no clinical difference¹⁶⁵ to longer duration of fever,¹⁶⁹ longer duration of bacteraemia^170,171 and increased treatment failure in composite endpoints^172–175 to an increased mortality rate.^176–178 The relevance and clinical impact of increased antibiotic survival may depend on the site of infection: one study showed, for example, a worse outcome with tolerant strains only in bacteraemic patients with endocarditis but not in non-endocarditis patients.¹⁷⁷ Most studies focus on increased survival with vancomycin, which may be a marker for treatment failure even when no vancomycin is used in the treatment.¹⁷⁵ Remarkably, Miyazaki et al.¹⁷⁸ were able to demonstrate a correlation of the level of bacterial glycopeptide survival (measured as reduction of cfu in kill curves) with the probability of survival of patients with MRSA bacteraemia. In conclusion, several studies over the last few decades found a correlation between increased antibiotic survival and clinical failure, but the differences in patient population and the lack of standardized methods of measuring antibiotic survival make a comparison between them difficult. Hence, there is an urgent clinical need to reliably measure and understand the dynamics and rate of increased antibiotic survival of S. aureus in patients.¹⁷⁹

Although expert consensus on the definition of tolerance and persistence in general has been achieved,¹⁷ there is so far no validated and reliable tool to quantify and most of all classify tolerance or persistence for clinical application and decision-making. An optimal test for measuring tolerance or persistence would need to be highly standardized, reproducible within and between laboratories and strains, rapid, not too labour-intensive and correlated with clinical outcome. First promising steps in this direction have been taken with proposed new metrics and methodological approaches.^{117,180–184} For example, the metric MDK99 (minimal duration necessary for killing 99% of a bacterial population) could be used to quantify tolerance.^17,180 However, standardization of pre-test growth conditions, starting inoculum, test media and conditions as well as the applicable antibiotic concentrations still need to be established and clinically relevant cut-offs determined. Moreover, the antibiotic concentration needs to reach saturation for killing in the MDK99 assay, which does not work for all antibiotics.^180,185 Finally, the proposed setup to measure MDK99 probably needs a robotic system to fit into the workflow and manpower of a routine diagnostic microbiology laboratory.¹⁸⁰

Approaches to treating antibiotic-tolerant S. aureus

With the increasing awareness of increased antibiotic survival, research has turned its focus on new ways of treating and eradicating tolerant S. aureus. Three main approaches have been explored so far to combat tolerant S. aureus (Figure 2): (i) disrupting growth- or metabolism-independent targets of increased antibiotic survival with anti-tolerance compounds; (ii) improving the activity of existing antibiotics (by changing methods and duration of administration); and (iii) resensitizing tolerant cells to existing antibiotics.¹⁸⁶

Strategies targeting mechanisms of increased antibiotic survival have produced compounds capable of killing tolerant S. aureus, the majority of which target the cell membrane, an optimal target in metabolically quiescent cells. Retinoids, a new class of antibiotics, disrupt the lipid bilayer of tolerant S. aureus.¹⁸⁷ Various membrane-active small molecules, such as NH125,^188–190 XF-73,¹⁹¹ HT-61¹⁹² or NCK10¹⁹³ have shown good activity against tolerant S. aureus. A broad clinical application may be limited by specific properties such as induction of significant haemolysis through NH125 or low bioavailability of XF-73 and HT-61 due to strong hydrophobicity. Nevertheless, XF-73 has successfully completed Phase I/II tests and HT-61 has passed Phase II trials and is currently undergoing Phase III trials.¹⁸⁶ Engineered cationic amphipathic peptides,¹⁹⁴ bacteriophage-derived or synthetic lysins,^195–197 the potassium ionophore boromycin,¹⁹⁸ surface-attached polyethylenimine¹⁹⁹ or even the anthelminthic biothionol²⁰⁰ can kill tolerant S. aureus through cell wall/membrane interaction.

Another promising new synthetic antibiotic is ADEP4, which has the unique feature of not blocking but activating its target. It binds intracellularly to the protease ClpP resulting in an ATP-independent uncontrolled self-digestion and death of the cell.^55,201,202 However, this substance must be used in combination with other antibiotics as its use is fraught with rapid development of escape mutants.⁵⁵ Similarly, a clinical drug-library screen identified the quinolone tosufloxacin as highly active against tolerant S. aureus—even more active than ADEP4 in one in vitro study.²⁰³ Novel compounds such as SPI009,²⁰⁴ halogenated indoles,²⁰⁵ phenazine-inspired small molecules²⁰⁶ or chelators of iron and zinc²⁰⁷ have an effect on tolerant S. aureus cells. Even the anticancer drugs mitomycin and cisplatin demonstrate bactericidal action on tolerant S. aureus through crosslinking DNA.^208,209 However, their clinical application in S. aureus infections seems limited due to their inherently unfavourable toxicity.

Enhancing existing antibiotics may be the most straightforward way of tackling tolerant S. aureus. For example, combinations of antibiotics may achieve a synergistic effect on tolerant S. aureus as shown in vitro for daptomycin and tobramycin.^103,144 However, that study applied very high doses of 100× MIC, which is not achievable systemically in patients due to toxicity issues, especially considering protein binding of 92%. Localized administration could, in theory, circumvent systemic toxicity in implant-associated infection.²¹⁰ Prolonging the duration of antibiotic treatment represents another approach. Concordantly, in vitro long-term antibiotic exposures of up to 24 days led to complete eradication of tolerant S. aureus populations.¹⁰³ Recently, a mathematical model successfully predicted the duration of treatment for eradication of infections with tolerant bacteria.²¹¹ These results foster the discussion of treatment duration in patients and may also explain why short-course antibiotic treatment (<14 days) is associated with more relapse even in uncomplicated S. aureus bacteraemia.²¹² As early as 1944, Bigger suggested an intermittent antibiotic treatment to let the dormant tolerant subpopulation awaken in antibiotic-free intervals, rendering them amenable to the next antibiotic killing.³⁵ The potential efficacy of this pulsed treatment was recently validated in vitro.²¹³ However, whether this effect can be applied clinically to improve the outcome in treatment of tolerant infections remains to be proven. Another way of improving the available antibiotics is to synthetically fuse them to cleavable moieties with cell-penetrating properties^79,214–217 or to change the carrier formulation (Figure 2).^218–220 This either increases the penetration into S. aureus per se, as shown for the fusion antibiotic pentobra, or ameliorates the penetration into mammalian cells to access the intracellular S. aureus population.

Finally, resensitizing tolerant S. aureus to existing antibiotics has become a focus of research. Changing bacterial metabolism by varying the composition of environmental metabolites can influence antibiotic killing of tolerant S. aureus. As an example, addition of fructose to biofilm cultures enabled growth-independent killing through aminoglycosides;⁹⁷ addition of glucose increased the effect of daptomycin,¹²⁹ vancomycin and rifampicin,¹²⁶ or of quinolones when combined with oxygen.³⁸ Mechanistically, this has been linked to metabolite-enabled generation of proton-motive force,⁹⁷ activation of cell respiration³⁸ or increased transcription of murein hydrolase.¹²⁶ It is worth noting that potentiation is generally antibiotic and metabolite specific,⁹⁷ supportive of the notion that various tolerance mechanisms are at play. Changing the environment is not limited to metabolic stimuli, as pH modification by addition of the basic amino acid l-arginine or exposing tolerant S. aureus to alkaline conditions increased the killing effect of aminoglycosides.^221,222 Further, interfering with chemical signalling by using supernatant from S. aureus medium^223,224 or fatty acid signalling molecules²²⁵ succeeded in resuscitating dormant S. aureus. These results of sensitizing by ‘environmental reprogramming’ are intriguing; however, the complex nutritional and metabolic interactions in an infected tissue could make it challenging to specifically trigger the desired response for synergistic therapeutic reasons. Finally, the effect of existing antibiotics on tolerant S. aureus may be augmented by adding phenol-soluble modulins,²²⁶ bacteriophages²²⁷ or rhamnolipids²²⁸ to the treatment.

Open questions and conclusions

Our knowledge of the determinants and significance of antibiotic-tolerant S. aureus in patients remains patchy at best. Does the same mechanism of increased antibiotic survival account for a late relapse after stopping antibiotic treatment⁶² and continuous infection as found in persistent bacteraemia?⁷ Do the various microenvironments in the infected milieu determine antibiotic survival more than the cellular stochasticity that is thought to underlie the persister phenotype? How relevant is increased antibiotic survival in non-bloodstream infections such as in implant infections? What role does the antibiotic selected have on development of tolerant phenotypes during treatment? What is the therapeutic role of so-called bacteriostatic antibiotics in infections with tolerant S. aureus? What influence does the immune system exert on S. aureus to become tolerant? Can we identify bacterial or host markers for tolerant infections? How long can tolerant S. aureus stay in the metabolically quiescent state in the human body? Do they revert stochastically or only by induction through environmental triggers? Do tolerant subpopulations actually harbour mutations, which rapidly revert back upon subculturing? Would it be an advantage to regularly screen all patients for tolerant S. aureus or only in those with treatment failure? Could we adapt treatment duration to the level of antibiotic survival or should we aim to individualize an effective combination therapy? To answer these and other questions, a reliable and reproducible tool of measuring antibiotic survival is critical.

In conclusion, S. aureus may display increased survival to antibiotic treatment through a plethora of inducers and mechanisms, reflecting its physiological versatility and infectious context. Increased antibiotic survival should be considered as a potential factor for the unacceptably high rate of treatment failure. New insights and promising approaches are appearing on the horizon.

However, clinical studies are needed to better understand the dynamics, mechanisms and relevance of tolerant S. aureus at the site of infection and to step ahead in the fight against one of mankind’s enduring foes.

Funding

R. K. was supported by the research funds of the University of Basel.

Transparency declarations

None to declare.

References