Many bacterial infections in humans and animals are caused by bacteria residing in biofilms, complex communities of attached organisms embedded in an extracellular matrix. One of the key properties of microorganisms residing in a biofilm is decreased susceptibility towards antimicrobial agents. This decreased susceptibility, together with conventional mechanisms leading to antimicrobial resistance, makes biofilm-related infections increasingly difficult to treat and alternative antibiofilm strategies are urgently required. In this review, we present three such strategies to combat biofilm-related infections with the important human pathogen Staphylococcus aureus: (i) targeting the bacterial communication system with quorum sensing (QS) inhibitors, (ii) a ‘Trojan Horse’ strategy to disturb iron metabolism by using gallium-based therapeutics and (iii) the use of ‘non-antibiotics’ with antibiofilm activity identified through screening of repurposing libraries.
The term ‘biofilms’ was first introduced by Costerton et al.  in 1978: although bacteria are single-celled organisms, they can organize themselves in complex sessile multicellular consortia, known as biofilms. Biofilms are formed everywhere: in industrial settings, in natural environments and in the human body. Bacterial cells residing in biofilms are attached to each other or to a surface and they are embedded in a matrix of extracellular polymeric substances [2,3]. Biofilm cells are phenotypically different from their planktonic counterparts because growth rate and gene transcription are altered . Biofilm formation is a complex developmental process and adhesion, proliferation and detachment are the three major steps in biofilm formation . The process is dynamic and starts with planktonic bacteria that attach reversibly to a surface. Pili, flagella, receptors or other adhesive surface appendages make contact with biotic or abiotic surfaces. Adhesion is followed by the secretion of extracellular polymeric substances resulting in an irreversible attachment. Subsequently, cells proliferate resulting in the formation of microcolonies. The biofilm grows and cells differentiate, resulting in a mature biofilm with multi-layered cell clusters. Finally, cells disperse from the biofilm and can form biofilms elsewhere [2,3,6,7]. Quorum sensing (QS), or cell-to-cell signalling, plays an important role in the development of the biofilm [8,9].
It is generally accepted that sessile microorganisms, i.e. microorganisms residing in a biofilm show decreased susceptibility towards antimicrobial agents. This is due to a decreased penetration of antibiotics, a decreased growth rate of the biofilm cells and/or a decreased metabolism of bacterial cells in biofilms. In addition, the presence of highly specialized survivor cells (so-called persister cells) and the expression of biofilm-specific resistance genes (including specific efflux pumps) contribute to this tolerance [10–13]. Because of these biofilm-specific resistance and tolerance mechanisms, combined with the emergence of methicillin resistant, vancomycin intermediate resistance and vancomycin resistant Staphylococcus aureus strains (MRSA, VISA and VRSA respectively) [14,15], treatment of S. aureus infections has become increasingly difficult and alternative antibiofilm strategies are urgently required.
In the present review, we will highlight three such alternative strategies. A first alternative approach is targeting the bacterial communication system. QS is a process by which bacteria produce and detect signal molecules in a cell density dependent way, and as QS plays an important role in bacterial biofilm formation and resistance, QS inhibitors (QSIs) are promising antibiofilm agents, either alone or as potentiators of conventional antibiotics [16–18]. In the second section, we will discuss the use of gallium-based therapeutics to combat S. aureus biofilm infections. The similarity between gallium and iron allows using a ‘Trojan Horse’ strategy to disturb iron metabolism, making gallium containing compounds interesting novel antimicrobial agents [19,20]. Finally, we will highlight the potential of ‘repurposing’ approaches, in which libraries of known and approved drugs are screened to identify novel compounds with antibiofilm activity and/or potentiating activity towards antibiotics [21–23].
The approaches highlighted are by no means the only ones being explored to tackle the problem of microbial biofilms. Other approaches not discussed here include the use of bacteriophages, bacteriophage-derived enzymes, nanoparticles (including silver, lipid and polymer nanoparticles), surfactant-based carriers, matrix degrading enzymes like dispersin B, and the reader is referred to several recent articles on these topics for more information [24–28].
The use of quorum sensing inhibitors to tackle
S. aureus biofilms
As far back as 1998, Davies et al.  showed that QS in Pseudomonas aeruginosa is important for biofilm formation, as a QS defective mutant formed flat and undifferentiated biofilms, in contrast with the wild-type. In 2005, it was shown that if QS was blocked in P. aeruginosa (either by knocking out the relevant genes or by using QSI), biofilms formed by this organisms became more sensitive to tobramycin and H2O2 . It was later shown that QSI can ‘potentiate’ the activity of antibiotics against various bacterial biofilms, in different model systems [16,30].
Hamamelitannin (HAM) is a QSI able to potentiate the activity of vancomycin against S. aureus biofilms . HAM targets the S. aureus TraP QS system and its effect on biofilm susceptibility is (at least partially) due to an effect on the cell wall thickness and release of extracellular DNA (eDNA) . While treatment of S. aureus biofilms with vancomycin typically results in thickening of the cell wall and release of eDNA, these defence mechanisms are down-regulated upon addition of HAM. At the molecular level, this can be explained by the differential expression of genes involved in biosynthesis of peptidoglycan and peptidoglycan precursors (including genes involved in synthesis of L-lysine and glucosamine-6-phosphate) and regulators of autolysis, like lytS .
Despite its activity, HAM is not an ideal drug-candidate, as it demonstrates several undesirable properties, including a high number of hydroxyl functions leading to high polarity, an aromatic hydroxyl function making the molecule oxidation- and glucuronidation-sensitive and metabolically unstable ester linkers (Figure 1) . To obtain more active derivatives with more drug-like properties, an extensive structure–activity relationship study was set up and several compounds with high in vitro and in vivo activity were identified (Figure 1) [32,33]. Several of these highly active compounds showed excellent metabolic stability and lacked toxicity in MRC-5 lung fibroblast cells, making them prime candidates for testing in more advanced models .
Chemical structure of HAM (top) and two more active derivatives
S. aureus biofilm infections with gallium-based therapeutics
The underlying rationale of gallium-based therapeutics lies in disrupting bacterial iron metabolism. Iron plays a crucial role in cellular processes, including DNA synthesis, respiration, energy production, for the protection against reactive oxygen species (ROS) and for biofilm formation [20,34], which depends on iron levels higher than needed for vegetative growth . The most abundant iron source in the human body is haem (iron protoporphyrin IX) as part of haemoglobin inside erythrocytes (Figure 2) [36–38] and this is the preferred iron source for S. aureus that is equipped with various systems, such as the iron-regulated surface determinant system, to sequester iron/haem from the host [38,39]. Once inside bacteria, iron is liberated from haem to be utilized for cellular processes or stored to foster bacterial survival and virulence, while excess iron/haem and their toxic metabolites are removed via efflux pumps [20,39].
Structural similarities between haem and gallium–protoporphyrin that is able to mimic haem as iron source
Bacterial iron metabolism as therapeutic target
As virtually all pathogens rely on iron for growth and virulence, bacterial iron metabolism represents a target for medical intervention strategies. Gallium and iron are two very similar elements and gallium can be utilized as iron analogue . By following a ‘Trojan Horse’ strategy, gallium containing compounds exploit bacterial iron acquisition systems for internal uptake or penetrate through the bacterial cell wall [19,20]. On a cellular level, gallium competes with iron and interferes with its absorption, metabolism and activity, thereby disrupting vital iron-dependent processes . Most importantly gallium is unable to transfer electrons and cannot induce redox reactions, thereby impeding respiration, DNA synthesis and bacterial proliferation and generating ROS [41,42]. Development of resistance against gallium is unlikely because if bacteria down-regulate receptors to reduce the binding and uptake of gallium, the uptake of iron would also be reduced [43,44].
Gallium can be delivered as simple salt or complex/conjugate with other molecules. Substantial antimicrobial effects were shown against various pathogens in vitro and in vivo, however, the antibiofilm activity appears to be species and strain dependent [45–50].
Gallium nitrate [Ga(NO3)3] is a salt whose pharmacokinetic profile and low-to-moderate toxicity are already known in humans . Ga(NO3)3 is mainly excreted by the kidneys and can potentially cause nephrotoxicity; however, in a clinical phase I trial this was dependent on the dose, treatment duration and the way Ga(NO3)3 was delivered . It was shown that renal toxicity can be reduced by a low drug dosage and longer infusion times. Apart from that, Ga(NO3)3 was shown to exhibit antibiofilm activity against S. aureus . It was reported that it is antibacterial even against stationary phase bacteria, which are usually found in the centre of biofilms and which frequently show reduced susceptibility to antibiotics . Another salt, gallium maltolate was reported to exhibit antimicrobial activity in a mouse model for burn wound infections, showing higher efficacy than Ga(NO3)3 and preventing the systemic spreading of P. aeruginosa. In the same model, a substantially reduced wound colonization of S. aureus was also demonstrated . While the in vitro and in vivo antimicrobial activity of gallium salts are interesting to note, their administration is challenging. After oral intake, gallium forms poorly soluble precipitates in the gastrointestinal tract limiting the bioavailability and antimicrobial activity in humans . Complexes like gallium citrate offer improved stability . Several in vitro studies showed broad antibacterial and antibiofilm activity of gallium citrate against Gram-positive and Gram-negative bacteria, as well as low drug resistance [19,48], and it was shown to impair biofilm formation on wound dressings and soft tissue in vivo, thereby improving wound healing . Siderophore–gallium complexes are another strategy to introduce gallium to bacteria. Siderophores are small molecules released by bacteria under iron limitation to sequester iron outside bacteria for internal uptake. Staphyloferrin A, a siderophore produced by S. aureus, was synthesized and loaded with gallium. This complex, however, failed to exhibit antimicrobial effects against MRSA .
Other gallium complexes with antimicrobial properties include synthetic haem analogues [55–57]. Being most similar to haem, gallium–protoporphyrin IX (GaPP, Figure 2) demonstrated strong activity against MRSA [58–60]. S. aureus is known to favour haem as preferred nutrient source  and is equipped with more active haem uptake systems than siderophore-based systems, facilitating the use of haem analogues like GaPP . However, GaPP’s low water solubility and modest dose-depending toxicity pose challenges for its applicability as therapeutic drug [59,62]. Cytotoxicity studies in human bronchial epithelial cells and murine fibroblasts showed no toxicity of GaPP at concentrations of up to 500 and 1000 μg/ml respectively , and GaPP used in a mouse model also showed no toxicity . There is potential for the development of synthetic gallium complexes with improved aqueous solubility and potency, and for incorporation of gallium into smart drug-delivery systems that overcome solubility and toxicity concerns.
Combination therapies with gallium
A multi-pronged approach using compounds with different modes of action offers the potential of additive or synergistic antimicrobial effects, while possibly reducing the risk for emerging resistance compared with monotherapy. As an example, by first starving bacteria with the iron chelator deferiprone (Def) and simultaneously introducing the haem-analogue GaPP as iron mimic, a significant antibiofilm activity was shown against S. aureus and MRSA . In addition, certain gallium compounds have the ability to potentiate antibiotics and combining antimicrobial compounds with gallium may lead to promising antibiofilm strategies .
Repurposing as a novel approach to find compounds active against
S. aureus biofilms
What is drug repurposing?
The use of drugs (either drug candidates, abandoned drugs, approved drugs or withdrawn drugs) to treat a disease for which they were initially not developed for is called drug repurposing or repositioning [21–23]. The main advantages of drug repurposing over de novo drug development are reduced time and costs in the R&D process, as knowledge concerning safety and pharmacology are available for the repurposing candidates . For lack of a better name, drugs in use for non-bacteriological indications but with antibacterial activity are in literature often called ‘non-antibiotic drugs’ or ‘non-antibiotics’ . These compounds might possess a direct antibacterial activity and/or enhance the activity of existing antibiotics by increasing the susceptibility of the bacteria towards the antibiotics, e.g. by controlling efflux pumps. In addition, they might also affect the pathogenicity of bacteria (virulence inhibitors)  or interfere with the host resulting in an improved pathogen clearance [68,69]. Several drug classes (e.g. antihistamines, local anesthetics, anti-hypertensive drugs, tranquilizers, statins and anti-inflammatory drugs) are known to possess antibacterial activity, although they were not developed to treat bacterial infections [67,70–72]. In the next paragraphs, we will discuss the activity of some of these non-antibiotics against S. aureus biofilms.
Jacobs et al.  screened the Prestwick Chemical Library for antimicrobial agents active against planktonic S. aureus; to this end they developed an adenylate kinase assay that identifies compounds that disrupt cellular integrity. Following their initial screen, they evaluated the activity of one hit compound, i.e. the antihistamine terfenadine, for activity against biofilms formed by S. aureus UAMS1 (an osteomyelitis clinical isolate). Treatment with terfenadine at 10× MIC resulted in a 2.7-fold increase in adenylate kinase release, corresponding to a 1.1-log reduction in biofilm cell viability. However, treating S. aureus infected Galleria mellonella larvae with terfenadine did not increase the latter’s survival . More recently, 84 terfenadine-based analogues were synthesized and evaluated for activity towards S. aureus planktonic cells. Two compounds had lower MIC in comparison with terfenadine, also against other Gram-positive pathogens but their antibiofilm activity has not yet been evaluated .
Antibacterial and antibiofilm activity of niclosamide and analogues
In another comprehensive screening using the same Prestwick Chemical Library, activity was evaluated by measuring inhibition of growth of planktonic S. aureus TCH1516; the screen resulted in the identification of 104 hits, most of them belonging to the group of antimicrobials and antiseptics. However, 18 non-antibiotic drugs were also identified and 9 of these hit compounds were evaluated for activity against S. aureus biofilms . Three of these showed modest activity, including the anthelmintic niclosamide that caused a reduction in the number of culturable S. aureus cells of 1–2 log . Rajamuthiah et al.  investigated the activity of niclosamide and another salicylanilide anthelmintic drug, oxyclozanide against planktonic grown ESKAPE pathogens (Enterococcus faecium, S. aureus, Klebsiella pneumoniae, Acinetobacter baumanii, Pseudomonas aeruginosa and Enterobacter species) . MICs against multiple S. aureus isolates ranged between 0.0625 and 0.5 μg/ml for niclosamide and between 0.5 and 2 μg/ml for oxyclozanide. These two compounds were also found to prolong survival of C. elegans infected with S. aureus MW2, with a similar effect as treatment with vancomycin. A third salicylanilide anthelmintic drug closantel was identified as a hit compound in a screen of the Biomol 4 compound library (containing 640 FDA-approved drugs) using a C. elegans–S. aureus infection assay . In this assay, C. elegans was infected with S. aureus MW2 BAA-1707 and treated with the library compounds in 384-well MTPs for 5 days. Next, Sytox Orange, a dye that stains death larvae, was added to the wells and an automated microscope was used the next day to generate both transmitted light and fluorescent images enabling the calculation of the number of surviving C. elegans. Using this approach, almost all antibiotics present in the library were identified as hits, as well as ten anticancer drugs, an antiviral drug, an antifungal, an antiarthritic drug, a non-steroidal estrogen and closantel. Although closantel turned out to have low MICs against several antibiotic-resistant S. aureus strains, the number of S. aureus cells in the infection assay using C. elegans was not reduced upon closantel exposure. It is thus possible that closantel targets bacterial virulence rather than survival, and/or has a direct effect on C. elegans .
Antibacterial and antibiofilm activity of 5-fluorouracil and analogues
5-fluorouracil is an antimetabolite widely used in treatment of cancers . Activity against biofilms was reported in 1992, as sub-MIC levels of 5-fluorouracil diminished biofilm formation of S. epidermidis [78,79]. Later, carmofur (1-hexylcarbamoyl-5-fluorouracil) was identified as a hit in the primary screen against planktonic cells of S. aureus and showed antibiofilm activity in a secondary screen . Recently, 5-fluorouracil and 5-fluoro-2'-deoxyfluridine, another fluoropyrimidine, were identified in a screen against planktonic S. aureus USA300 and activity of 5-fluoro-2'-deoxyfluridine was confirmed in a septicemic MRSA mice infection model with concentrations much lower than the concentrations therapeutically used for cancer treatment, and thus with reduced toxicity . The usefulness of 5-fluorouracil has been demonstrated in human clinical trials in which central venous catheters externally coated with 5-fluorouracil scored better in preventing catheter colonization than the control catheters coated with silver sulfadiazine or chlorhexidine .
Statins have been described for antibacterial effects by several research groups . Simvastatin at 1/16× MIC up to 4× MIC (62.5 μg/ml) significantly inhibited biofilm formation and at 4× MIC it significantly reduced the number of CFU/ml in mature biofilms of S. aureus ATCC 29213 [81,82]. Simvastatin was found to be able to disrupt S. aureus and S. epidermidis biofilms and its potency was higher than that of linezolid or vancomycin: at 2× and 4× MIC of simvastatin, the biofilm mass (as measured by Crystal Violet staining) was reduced by 40%, while 64 and 128× MIC of linezolid or vancomycin reduced biofilm mass by only 10% .
Repurposing candidates with potentiator activity
The combination of antibiotics with non-antibiotic drugs as potentiators could be a valuable approach to overcome antibacterial drug resistance [84–86]. Combination therapy might result in a broader spectrum of drug activity, synergy, a more rapid effect and the use of reduced drug concentrations . In a screening with 1059 previously approved drugs against planktonic P. aeruginosa PAO1, E. coli BW25113 and S. aureus ATCC 29213 in the presence of minocycline, 6, 41 and 35 hits respectively, were identified. These hits were non-antibiotic drugs that synergized with minocycline but had never been used clinically to treat bacterial infections . Disulfiram was one of the hits against S. aureus: alone, disulfiram has only weak antibacterial activity but it improved the activity of minocycline in a synergistic way against several MRSA strains, including MRSA USA300.
Ooi et al.  evaluated the antibiofilm activity of 15 redox-active compounds that have been safely used in humans for several applications (healthcare, cosmetics and consumption) and of which antibacterial activity had been described previously, but not against S. aureus planktonic cells and biofilms. All compounds tested were active against planktonic cells (MICs between 0.25 and 128 mg/l) and seven compounds (i.e. AO 2246, bakuchiol, benzoyl peroxide, carnosic acid, celastrol, nordihydroguaiaretic acid and totarol) were able to eradicate established biofilms of S. aureus SH1000 at concentrations ≤256 mg/l . Moreover, celastrol and nordihydroguaiaretic acid synergized the activity of gentamicin against S. aureus SH1000 biofilms . As both compounds did not cause irritation in a human living skin equivalent model, they might be valuable potentiators to treat superficial skin infections caused by S. aureus biofilms .
We recently screened the NIH Clinical Collection 1&2 against S. aureus Mu50 biofilms, formed in 96-well MTP . The screening was performed in the presence of vancomycin and resulted in the identification of 25 hit compounds that potentiated the activity of the antibiotic. Among these hits, we identified the disinfectants triclosan and hexachlorophene, the antiviral drug efavirenz and the antifungal imidazole drugs miconazole, econazole and oxiconazole. Antifungal imidazoles are known for activity against planktonic S. aureus for decades and perform activity by membrane damage and binding to flavohaemoglobins [91,92]. In addition, two anthracyclines, three selective estrogen receptor modulators, flutamide, oxymetholone, amiodarone, carvedilol, honokiol, loxoprofen, MK 886, 5-nonyloxytryptamide and ethacrynic acid were identified as well in our screen. Antibacterial activity of some of these hits had been reported before, e.g. for honokiol [93,94], MK-886 and 5-nonyloxytryptamine , but for most of the hits there were no previous indications for potential use as an anti-infective. Also four antipsychotic phenothiazine drugs (fluphenazine, perphenazine, thioridazine and trifluoperazine) and the antidepressant sertraline were hits. Anti-staphylococcal activity had been reported before for these compounds, both in vitro  and in vivo [96,97], but activity against biofilm was not investigated before. We showed that thioridazine enhances the activity of tobramycin, flucloxacillin and linezolid against in vitro grown S. aureus Mu50 biofilms (>1 log additional reduction in CFU/biofilm). Unfortunately, we were unable to confirm this activity in a biofilm model for chronic wounds.
The growing awareness about the problems associated with antimicrobial resistance and the role biofilms play in infectious diseases has increased the interest in the development of novel approaches to treat infections. In the present review, we highlighted three such alternative approaches for treating biofilm-related S. aureus infections, i.e. QS inhibition, gallium-based therapeutics and use of repurposed drugs.
While these approaches often appear very promising in early studies, there is a long road ahead of us before these compounds will make it to the market. For example, although gallium compounds show antimicrobial and antibiofilm activity against various bacteria, they have not found their way into clinical practice. There are two FDA approved gallium formulations on the market that are used as diagnostic agents in cancer therapy. The formulation Ganite (a gallium nitrate-citrate injection for cancer-related hypercalcaemia, FDA approved from 2003 to 2014) showed antimicrobial activity in vitro, however, this effect appeared species and strain dependent. Concentrations higher than the recommended dose would be required for a broad antibiofilm effect, raising toxicity concerns . In addition, it is worth stressing that promising results in in vitro tests are no guarantee for success in more advanced biofilm models and/or in vivo studies as the work with repurposing libraries has shown [86,98].
Despite these pitfalls, at least some of the approaches outlined above have considerable potential for the treatment of various biofilm-related infections. For example, HAM and a more potent derivative showed high potentiating activity in a mouse model for S. aureus mastitis [31,32]. By using an in vitro wound biofilm model, we could demonstrate that it is feasible to load wound care dressings with cyclodextrin–HAM complexes to increase the effect of antibiotics , opening the road towards further pharmaceutical developments. Finally, an exciting option is combining several of the novel approaches outlined above. For example, we have recently observed that the combination of Def, GaPP and HAM exceeded the antibiofilm activity of the individual compounds and the antibiotic control against S. aureus (Richter and Coenye, unpublished data), indicating that such combinations may lead to increased antibiofilm activity.
Biofilm formation contributes to the growing problem of reduced susceptibility to antimicrobial agents; this reduced susceptibility is multifactorial and is partially mediated by quorum sensing (QS).
There is an urgent need for alternative approaches to treat biofilm-related infections. Examples of such alternative approaches that have been explored to combat S. aureus biofilm infections are QS inhibition, gallium-based therapeutics and identification of potentiators from repurposing libraries.
Hamamelitannin and analogs inhibit QS in S. aureus and by doing so increase the susceptibility of S. aureus biofilms to antibiotics.
Gallium compounds exhibit antibiofilm activity by exploiting bacterial iron acquisition systems and disrupting cellular processes vital for bacterial growth, survival and pathogenesis.
Repurposing libraries contain a fair number of ‘non-antibiotics’ with antimicrobial activity against S. aureus biofilms (including statins and 5-fluorouracil) as well as potentiators (including phenothiazines and imidazoles).
Despite difficulties of translating results from in vitro to in vivo, some of the approaches outlined (or combinations thereof) seem to hold promise for future treatment of S. aureus biofilm-related infections.
The authors declare that there are no competing interests associated with the manuscript.
colony forming unit
minimum inhibitory concentration
methicillin resistant Staphylococcus aureus
reactive oxygen species
vancomycin intermediate Staphylococcus aureus
vancomycin resistant Staphylococcus aureus