Quorum sensing (QS) is a bacterial communication system which allows neighbouring cells to send and receive signal molecules, called autoinducers, in a density-dependent manner. QS has been shown to play a pivotal role in biofilm regulation for many species (Brackman and Coenye, 2015). Targeting QS with inhibitors (QSIs) has been a significant innovation in the antibiofilm field, even though the role of QSIs in biofilm formation and dispersal is not always fully understood for different Gram-positive and Gram-negative bacteria (Brackman and Coenye, 2015). QS inhibition can occur by inhibiting autoinducer synthesis, degradation of signalling molecules, interfering with signal binding, and inhibition of the signal transduction cascade, which results in dysregulated biofilm signalling and subsequently, dispersal or inhibition of the biofilm (Brackman and Coenye, 2015; Jiang et al., 2019). Most QSIs are derived from proteins (autoinducers, transcription factors and regulators) that mediate QS in the bacterial target, acting as competitive inhibitors of these systems.

Several studies have been conducted using QSIs in combination with antibiotics to either inhibit and/or eradicate biofilms. One of the earliest studies, conducted by Brackman et al., focused on the efficacies of baicalin hydrate, cinnamaldehyde, and hamamelitannin (structures shown in Figure 2 ) in combination with the antibiotics tobramycin, clindamycin, and vancomycin (Brackman et al., 2011). These compounds were tested against established (24-hour) Pseudomonas aeruginosa and Staphylococcus aureus biofilms in Mueller-Hinton agar (Brackman et al., 2011). Initial in vitro testing showed that individual treatment with inhibitor or antibiotic alone was largely ineffective (<10% reduction in biofilm bacterial numbers) against P. aeruginosa ATCC 9027 and S. aureus CS1 and Mu50. Tobramycin alone reduced only P. aeruginosa PAO1 biofilm bacterial numbers by 45% (Brackman et al., 2011). When biofilms were treated with a combination of QSI and antibiotic, viable bacteria for all strains showed a 68-90% reduction ( Table 1 ), except S. aureus Mu50, which were only reduced by 6% when treated with clindamycin and a QSI (Brackman et al., 2011). This strain has been shown to be resistant to clindamycin planktonically, which may explain reduced efficacy also against its biofilms (Cui et al., 2009).

These QSI-antibiotic combinations were also evaluated in vivo against Burkholderia cenocepacia and Burkholderia multivorans, in a Caenorhabditis elegans and Galleria mellonella survival model and in a mouse lung infection model (Brackman et al., 2011). QSIs alone exacted a strain-dependent effect on the survival of G. mellonella, with a minimal protective effect observed against the Burkholderia spp. (0-30% survival) but very high protection in larvae infected with S. aureus (70-100% survival) (Brackman et al., 2011). In C. elegans, none of the QSIs alone exhibited any protective effect (<10% survival) (Brackman et al., 2011). Standalone antibiotic treatment in C. elegans showed increased survival compared to the inhibitor only, and this was also observed for G. mellonella (Brackman et al., 2011). Combination treatment exhibited the most protective effect, with most combinations resulting in 80-100% survival when infected with either P. aeruginosa and S. aureus. These outcomes were further confirmed in the mouse model, which showed that the combination of tobramycin (30 mg/kg) and baicalin hydrate (2 mg/kg) reduced pulmonary bacterial numbers by 99.9% (Brackman et al., 2011).

The same authors extended their investigation of hamamelitannin, by testing it in combination with several antibiotics (vancomycin, cefazolin, cefalonium, cephalexin, cefoxitin, daptomycin, linezolid, tobramycin, fusidic acid) against established S. aureus biofilms (Brackman et al., 2016). Here they measured the efficacy of single treatment (antibiotic only) and compared it to combination treatment (antibiotic and hamamelitannin). For all antibiotics, combination treatment was at least equally as effective but, in most cases, resulted in greatly increased efficacy (40-70% additional biofilm eradication compared to antibiotic treatment alone) (Brackman et al., 2016). The most effective combinations were hamamelitannin with cefazolin, cefoxitin, tobramycin or fusidic acid, which resulted in ≥90% eradication of the biofilm (Brackman et al., 2016). Similarly, Vermote et al. derived a hamamelitannin analogue (compound-38) that was reported to have a 20-fold lower median minimum bactericidal (MBC₅₀) value against S. aureus Mu50 planktonic cells (Vermote et al., 2016). Mirroring the initial hamamelitannin study (Brackman et al., 2011), all combinations were at least as, or more effective than their standalone counterpart (Vermote et al., 2016). The most effective combinations were compound-38 with vancomycin, cefazolin, daptomycin or tobramycin (≥90% eradication, Table 1 ) (Vermote et al., 2016). Combined, these studies suggest that hamamelitannin is most effective against biofilm bacteria when administered with the antibiotics cefazolin and tobramycin (Brackman et al., 2016; Vermote et al., 2016).

In a different study, Brackman et al. incorporated hamamelitannin into a delivery system to improve its efficacy against S. aureus biofilms (Brackman et al., 2016). They achieved this by incorporating the inhibitor into a cyclodextrin complex which could release the QSI and an antibiotic at a controlled rate (Brackman et al., 2016). Utilising these complexes as a delivery system increased efficacy in vitro, and the use of the delivery system with hamamelitannin alone reduced biofilm CFU by 1 log (Brackman et al., 2016). However, when the system was used in combination with both vancomycin and hamamelitannin, biofilm CFUs were reduced by 5.5 logs ( Table 1 ) (Brackman et al., 2016). Hamamelitannin thus appears to demonstrate a broad-spectrum of biofilm-eradication potentiating activity when administered as a co-treatment with antibiotics. Furthermore, it synergised well with antibiotics from different classes. These properties make hamamelitannin one of the more promising QSIs reviewed here, warranting further exploration into its synergistic capabilities. Given that hamamelitannin has already been tested in vivo in mice, this warrants further investigation into its potential toxicity and side effects, before moving into clinical development. As successful analogues with similar efficacies have also been developed (Vermote et al., 2016), this dispersal agent has potential to be modified chemically as well.

Other promising QSIs have also been tested in combination with antibiotics against biofilms. Some of these include furanone C-30 and C11 (structures shown in Figure 2 ), ajoene (garlic extract), 3-amino-7-chloro-2-nonylquinazolin-4(3H)-one (ACNQ) (PqsR receptor inhibitor), derivatives of cinnamic acid, horseradish extract and alkylquinolone-based inhibitors. Christensen et al. investigated combinations of tobramycin and the inhibitors furanone C-30, ajoene, and horseradish extract against P. aeruginosa biofilm infection in vivo using a mouse intraperitoneal implant model (Christensen et al., 2012). Tobramycin alone failed to reduce bacterial numbers by more than 1 log, but combination treatment reduced bacteria by up to 3 logs (Christensen et al., 2012). Treatment with inhibitor alone however, also showed similar efficacy, suggesting that the interaction between the two compounds was additive rather than synergistic. A derivative of a signalling molecule in P. aeruginosa QS, C11, was also tested in combination with the antibiotics ciprofloxacin, tobramycin, ceftazidime, and colistin against forming P. aeruginosa biofilms (Furiga et al., 2015). Treatment with individual agents reduced biofilm surface area (CFU/cm²) by 1-2 logs, while combination treatment with antibiotics and C11 resulted in a 4-6-log reduction ( Table 1 ) (Furiga et al., 2015). The only combination which was not synergistic was ceftazidime and C11, reducing biofilm surface area only by 1-2 logs (Furiga et al., 2015).

Building upon this work, Singh et al. used an inhibitor similar to C11 which also targets QS in Pseudomonas spp (Singh et al., 2019). Using engineered polymeric nanoparticles as a delivery system, they co-delivered the inhibitor ACNQ (4 µg/mL) and ciprofloxacin (60 µg/mL) to treat established P. aeruginosa biofilms (Singh et al., 2019). Individual treatments with the inhibitor and antibiotic were somewhat effective (10% and 50% eradication, respectively), however, combination treatment significantly enhanced eradication of established biofilms (80% eradication; Table 1 ) (Singh et al., 2019). Using an alternative delivery system, Ho et al. tested the alkylquinolone QSI [1] (20 µM) in combination with tobramycin (25 µg/mL) to treat P. aeruginosa biofilms (24-hours) (Ho et al., 2020). Here they employed a squalenyl hydrogen sulfate nanoparticle delivery system to deliver both compounds at the same time (Ho et al., 2020). Treatment with the antibiotic alone resulted in a 4-5-log reduction in biofilm CFU compared to the untreated control, regardless of the mode of delivery (Ho et al., 2020). Combination treatment with the delivery system achieved complete eradication (>6-log reduction), whereas combination treatment without the delivery system resulted in 3-4-log less eradication (Ho et al., 2020). Thus, the design of the delivery system is important for combination treatment with QSIs. The dual nature of a combination treatment suggests that for it to be successful, both compounds must be in the same environment at the same time to work effectively, as depicted in Figure 1B . As standalone treatments do not need to interact with any other compounds, this may be why the delivery system had no impact on their efficacy, as reported in this study. This should be taken into consideration when designing future co-treatment strategies.

Another studied QSI is baicalin ( Figure 2 ), a flavonoid isolated from the roots of Scutellaria baicalensis (Chinese skullcap) (Slachmuylders et al., 2018). Using this compound, Slachmuylders et al. examined biofilm eradication in combination with tobramycin, gentamicin, kanamycin or neomycin (Slachmuylders et al., 2018). Tobramycin was the most effective against several strains of Burkholderia spp. when combined with baicalin hydrate (75-95% increase in biofilm eradication against 5 out of 9 strains tested, compared to antibiotic treatment alone) (Slachmuylders et al., 2018). Similarly, Luo et al. tested baicalin in combination with levofloxacin, amikacin, and ceftazidime against P. aeruginosa biofilms (Luo et al., 2017). Amikacin and baicalin were the most effective combination for biofilm inhibition in vitro, while in vivo (mouse peritoneal implant model), ceftazidime and baicalin were the most effective combination for reducing bacterial numbers (Luo et al., 2017).

Another series of QS inhibitors shown to be active against S. aureus are analogues of the RNA-III inhibiting peptide (RIP) (Cirioni et al., 2013; Simonetti et al., 2013; Simonetti et al., 2016). Mouse studies using the QSIs FS3 and FS8 with daptomycin and tigecycline, respectively, demonstrated additive and synergistic efficacy against S. aureus biofilms (Cirioni et al., 2013; Simonetti et al., 2013). In these studies, Cirioni et al. and Simonetti et al. implanted grafts (with or without inhibitor) and injected S. aureus into the graft site while the antibiotic was administered intraperitoneally (Cirioni et al., 2013; Simonetti et al., 2013; Simonetti et al., 2016). Treatment with the QSI or antibiotic alone resulted in a 2-3-log reduction in bacterial numbers after 7 days of infection, compared to the untreated control group (Cirioni et al., 2013; Simonetti et al., 2013). Treatment with both compounds resulted in an additive effect for FS3 and daptomycin (4-log reduction) and a synergistic effect for FS8 and tigecycline (5-log reduction, Table 1 ) (Cirioni et al., 2013; Simonetti et al., 2013). Following on from this, FS10 was also tested in combination with tigecycline and showed similar results in the same mouse infection model against both methicillin-susceptible S. aureus (MSSA) and MRSA (Simonetti et al., 2016). Groups treated with FS10 alone showed minimal CFU reduction after 7 days post implantation (1-2-log reduction), while groups treated with tigecycline only showed higher log reduction in bacterial numbers (3.5-4 logs, Table 1 ) (Simonetti et al., 2016). Combination treatment was the most efficacious against both MSSA and MRSA, reducing bacterial numbers by 5.5-6 logs (Simonetti et al., 2016). Considering that the inhibitor and antibiotic are administered separately, administration of both at the same site could potentially improve efficacy.

Lastly, cinnamic acid, a metabolite of Cinnamomum cassia (Chinese cinnamon), has recently been shown to inhibit quorum sensing in bacteria (Vasconcelos et al., 2018). In a study by Cheng et al., two synthesised cinnamic acid derivatives, 4-dimethylaminocinnamic acid (DCA) and 4-methoxycinnamic acid (MCA) were tested in combination with tobramycin against Chromobacterium violaceum biofilms (Cheng et al., 2020). Individual treatments were largely ineffective at eradicating established biofilms (≤1-log reduction in biofilm CFU counts), though combination treatment was highly successful and resulted in a 5-log CFU reduction compared to the untreated control, see Table 1 (Cheng et al., 2020). It is also noteworthy that the core structures of these compounds are amenable to synthetic modification, which makes them ideal candidates for the development of more potent derivatives. In conjunction with this, additional in vivo/ex vivo studies should be conducted to test the toxicity of newly modified inhibitors while also assessing discrepancies between in vitro vs. in vivo effects.

Overall, quorum sensing inhibitors are a promising class of biofilm disruptors. Most QSI studies to date demonstrate that while they are not effective at eradicating biofilms on their own, when used together with antibiotics they can effectively eradicate established biofilms both in vitro and in vivo. Continued research into QSIs is sorely needed, as many of the mechanisms inhibited by them are poorly understood (Brackman and Coenye, 2015). Additionally, further manipulation of the structures of these inhibitors would open many avenues of therapeutic development, considering their amenability to synthetic modification.

Nitric oxide (NO) is a free radical colourless gas (at room temperature), and a well-established signalling molecule in eukaryotic organisms (Barraud et al., 2015). One of the earliest uses of NO as a biofilm dispersal agent was documented by Barraud et al. (2006) and since then its potential as a biofilm dispersal agent has been widely documented. While NO concentrations at the mM range are antibacterial (Barraud et al., 2015), inhalation of NO gas at such high concentrations is also acutely toxic to the respiratory tract of humans (Weinberger et al., 2001). However, NO-mediated biofilm dispersal occurs at concentrations sublethal for bacteria (nM range) (Barraud et al., 2006). Thus, the use of NO as an effective biofilm eradication strategy requires supplementation with an antimicrobial agent. NO signals dispersal of bacterial biofilms by interacting with enzymes that affect intracellular concentrations of bis-(3′-5′)-cyclic dimeric guanosine monophosphate (c-di-GMP) (Williams and Boon, 2019). The molecular mechanism was recently characterised in some species of bacteria, including P. aeruginosa, Nitrosomonas europaea, and Shewanella oneidensis (Hossain and Boon, 2017; Hossain et al., 2017; Nisbett et al., 2019). Interaction of NO with NO-sensitive enzymes stimulates the activity of phosphodiesterases (PDEs), which degrade c-di-GMP and signal biofilm dispersal (McDougald et al., 2011).

Several studies have tested NO for its ability to potentiate antibiotics in vitro (Barraud et al., 2009; Nguyen et al., 2016; Liu et al., 2020; Soren et al., 2020). NO is a difficult molecule to work with due to its gaseous form at room temperature and high reactivity (Barraud et al., 2015). Hence, most studies have used NO donors, such as sodium nitroprusside (SNP) and (Z)-1-[N-Methyl-N-[6-(N-methylammoniohexyl)amino]]diazen-1-ium-1,2-diolate (MAHMA NONOate) (Barraud et al., 2009; Barnes et al., 2013; Marvasi et al., 2015) (structure shown in Figure 3 ), which albeit are also known to be inherently unstable molecules (Wang et al., 2002). In order to address this, several NO delivery systems have been reported to release NO in a controlled manner at the target site (Lu et al., 2013; Nguyen et al., 2016; Ren et al., 2016; Liu et al., 2020). Examples of these include diazeniumdiolate (NONOate), micro/nanoparticles, and amphiphilic poly(amidoamine) (PAMAM) dendrimers (Lu et al., 2013; Nguyen et al., 2016; Soren et al., 2020). These systems improve handling and dose mediation, prevent or limit off-target effects, and increase the efficacy of biofilm dispersal by NO while minimising cytotoxicity to the host (Davies et al., 2001; Findlay et al., 2004; Poh et al., 2017; Yang et al., 2018; Verderosa et al., 2019c). In this section, we review studies that have combined NO with antibiotics for either dispersing or eradicating clinically relevant biofilms, while excluding studies which have only tested NO alone in vitro.

In one of the earlier in vitro studies, Barraud et al. investigated the potentiation of tetracycline by NO against biofilms from a variety of different bacterial species, using the NO-donor sodium nitroprusside (SNP) (Barraud et al., 2009). Pre-treating Vibrio cholerae biofilms with NO and then adding tetracycline reduced the biofilm surface area by 90% compared to untreated biofilms (Barraud et al., 2009). NO alone reduced the biofilm surface area by 67%, while tetracycline alone only afforded a 21% biofilm reduction (Barraud et al., 2009). The study also investigated the ability of NO to potentiate chlorine used for water treatment. Similar to tetracycline, NO treatment of V. cholerae biofilms followed by chlorine resulted in significant biofilm reduction (85-90%) compared to untreated controls (Barraud et al., 2009). The authors concluded that NO could enhance the activity of tetracycline and chlorine against biofilms when administered sequentially (NO treatment followed by tetracycline/chlorine treatment), however co-treatment with NO and tetracycline was not investigated, nor was the viability of the dispersed cells.

While sequential treatment of biofilms with a dispersal agent followed by an antibiotic is informative and may find use in industrial settings, such as water treatment, this approach would not be suited to the clinical setting. Dispersing a clinical biofilm prior to antimicrobial intervention can lead to systemic bacterial dissemination, which can result in spread of infection, septicaemia, and bacteraemia (Fleming and Rumbaugh, 2018; Müsken et al., 2018). For this reason, subsequent biofilm studies have focused on NO co-treatment with antibiotics.

NO co-administration with antibiotics has been facilitated by delivery systems such as those developed by Soren et al. which use the prodrug diethylamin-cephalosporin-30diazeniumdiolate (DEA-C3D) to investigate the delivery of NO in combination with tobramycin to eradicate P. aeruginosa biofilms (Barraud et al., 2012; Soren et al., 2020). DEA-C3D belongs to a class of NO-donor prodrugs known as cephalosporin-30-diazeniumdiolates (C3Ds), which contain the phenacetyl side chain of the first-generation cephalosporin cefaloram and the NO donor diazeniumdiolate (Soren et al., 2020). These compounds deliver NO by local release upon contact with bacterial β-lactamase enzymes (Soren et al., 2020). When used to deliver NO for the treatment of P. aeruginosa biofilms, NO alone reduced biofilm biomass by 50%, while treatment with the antibiotic tobramycin alone had no effect on biofilm biomass. However, NO and tobramycin co-delivery reduced biofilm biomass by 65%. While the potentiation of tobramycin was relatively modest, it should be noted that the concentration of antibiotic used in the study was sublethal (Soren et al., 2020), suggesting that efficacy could be likely increased by using tobramycin at higher concentrations. A similar potentiation trend was also observed for colistin. Despite colistin treatment alone being extremely effective against P. aeruginosa biofilms (90% biomass reduction), co-treatment with DEA-C3D resulted in almost complete biofilm eradication (98% reduction, Table 1 ) (Soren et al., 2020). The design of this NO delivery strategy is certainly innovative; however, its application is limited to β-lactamase-producing bacteria, and furthermore, it is unfortunate that the antibacterial properties of the delivery system (β-lactam) are lost upon NO release, which was an intentional design feature (Soren et al., 2020). The authors comment that the ultimate goal of this delivery system is to deliver NO to a targeted site, containing β-lactamase-producing bacteria, to reduce tissue damage from NO’s reactivity (Soren et al., 2020). In addition, as the study determined biofilm eradication by visualising remaining cells, dispersed cells were not studied. This makes it difficult to evaluate the efficacy of this treatment in the context of a “disperse and eradicate” strategy, as it is important to determine whether dispersed cells remained viable or were killed.

Another study targeting P. aeruginosa biofilms used a diazeniumdiolate nanoparticle delivery system in combination with the antibiotic gentamicin (Nguyen et al., 2016). Here, Nguyen and co-workers designed and produced a polymeric nanoparticle delivery system that simultaneously released NO and gentamicin at a target site by placing a gentamicin-NONOate complex (generated by reacting gentamicin and NO gas) at the core of the nanoparticle (Nguyen et al., 2016). Co-treatment with NO and gentamicin reduced biofilm viability by 90%, whereas treatment with the individual compounds only reduced viability by ≤30%, ( Table 1 ) (Nguyen et al., 2016). These results are promising and should the compound prove to be non-toxic to mammalian cells, a property that was not examined, the strategy would certainly merit in vivo analysis. Similarly, Liu et al. engineered a NO delivery system using a derivative of a naturally occurring compound, chitosan (Liu et al., 2020). The donor, chitosan-graft-poly(amidoamine) dendrimer (CS-PAMAM), was used to deliver methicillin and NO to methicillin-resistant Staphylococcus aureus (MRSA) biofilms (Liu et al., 2020). They observed that despite the strains being methicillin resistant, co-treatment with methicillin increased biofilm eradication a further 10% than with NO treatment alone (Liu et al., 2020). This study is unique as the target pathogen has acquired resistance to the antibiotic yet co-treatment with NO had an effect, which we speculate suggests that NO was likely dispersing cells from the biofilm outer layer and allowing higher levels of methicillin to access and kill underlying biofilm-residing cells.

The above studies have showcased the promising potential of NO as a biofilm dispersal agent with and without antibiotics in vitro. However, studies on NO co-treatments of clinical biofilms using in vivo models (cystic fibrosis, urinary tract infections, chronic wounds, etc. (Vestby et al., 2020)] are currently lacking. Nevertheless several in vivo studies on NO have provide valuable insight into its potential use as an infection treatment and or control strategy (Webert et al., 2000; Miller et al., 2012; Miller et al., 2013).

Inhaled NO has been studied as a standalone treatment for pneumonia in rats and has more recently progressed to phase I human trials (Webert et al., 2000; Miller et al., 2012; Miller et al., 2013; Deppisch et al., 2016). A rat model of P. aeruginosa pneumonia was initially used to test the efficacy of inhaled NO (Webert et al., 2000; Miller et al., 2012; Miller et al., 2013). Rats were injected intratracheally with 10⁸ colony forming units (CFU) of P. aeruginosa and then monitored over time for any physiological changes and for bacterial carriage in the lungs (Webert et al., 2000). The study successfully demonstrated that NO could reduce the pulmonary bacterial load after 24 hours of NO inhalation at a concentration of 40 ppm (Webert et al., 2000). A follow up study by the same group demonstrated that an even higher concentration of NO (160 ppm) given intermittently was also effective at lowering viable P. aeruginosa counts in the lungs of infected rats (Miller et al., 2013). This confirmed previous in vitro studies, which had shown that at a concentration of 160ppm, NO is bactericidal (Schairer et al., 2012). The success of these preclinical animal studies paved the way for a small Phase I study on 10 healthy adult volunteers, which confirmed the same dosing regimen of 160 ppm for 30 minutes was safe (Miller et al., 2012). A follow-up pilot clinical study treated eight cystic fibrosis patients suffering from long term (>6 months) bacterial and/or fungal lung infection(s) with gaseous NO for 30 minutes, three times daily, at a concentration of 160 ppm for two periods of five days over two weeks (Deppisch et al., 2016). Patients were followed over the course of seven months and post-NO treatment were all found to have an average log-reduction of 3.6 and 3.0, respectively, of bacteria and fungi (regardless of species) in their sputum samples (Deppisch et al., 2016). In addition to this, ESBL-producing Escherichia coli or Aspergillus spp. biofilms were undetectable after treatment. This study is one of the first of its kind and is unique from other NO dispersal studies in that it did not require the use of a NO delivery system. Additionally, these studies demonstrate the safety of NO administration even at bactericidal concentrations, which are higher than the concentrations shown to be required to disperse bacterial biofilms in vitro. Future clinical studies will greatly enhance the development of new respiratory therapeutics with potential benefits for other respiratory diseases.

The development of antibacterial and antibiofilm therapies based on NO is often limited by its high reactivity, instability, and inherent human toxicity at high concentrations (Barraud et al., 2015; Verderosa et al., 2019c). Thus, alternatives that mimic the behaviour of NO, but lack its limitations, have recently been explored (de la Fuente-Núñez et al., 2013; O'Loughlin et al., 2013; Rajasekaran et al., 2019).

Nitroxides are structurally similar to NO as both contain an unpaired electron delocalised over the nitrogen-oxygen bond. However, unlike NO, nitroxides are generally air-stable crystalline solids at room temperature, which makes their handling and delivery far more convenient than NO. Furthermore, nitroxides have been recently shown to exhibit similar antibiofilm properties to NO in vitro (de la Fuente-Núñez et al., 2013; Verderosa et al., 2016). The antibiofilm properties of nitroxides were first explored by de la Fuente Núñez et al., who showed that nitroxides not only mimicked the biofilm dispersal activity of NO but also prevented biofilm formation in vitro (de la Fuente-Núñez et al., 2013). In a follow up study by the same group, nitroxides were shown to potentiate the action of ciprofloxacin against both P. aeruginosa and E. coli biofilms (Reffuveille et al., 2015). Here, they investigated the ability of the nitroxide 4-carboxy-2,2,6,6tetramethylpiperidine 1-oxyl (CTEMPO) (structure shown in Figure 4 ) to potentiate the activity of ciprofloxacin when administered as a co-treatment (Reffuveille et al., 2015). CTEMPO (20 µM) alone dispersed P. aeruginosa and enterohemorrhagic E. coli (EHEC) biofilms resulting in significantly reduced biomass (60% and 71%, respectively, Table 1 ) (Reffuveille et al., 2015). At the same concentration, CTEMPO also potentiated ciprofloxacin activity against P. aeruginosa and EHEC biofilms, eradicating 99.3% and 93%, respectively (Reffuveille et al., 2015). This occurred at a ciprofloxacin concentration of just 320 ng/mL for P. aeruginosa and 20 ng/mL for EHEC when ciprofloxacin treatment alone had no or minimal effect on biofilm biovolume (Reffuveille et al., 2015). Dispersed cells from CTEMPO treated P. aeruginosa biofilms were also collected and quantified at 3-, 6- and 24-hours post treatment. Between the 3 and 6 hour treatment timepoints, there was minimal change in bacterial numbers, however after 24 hours, there was a minimum 5-fold change in dispersed bacterial counts (Reffuveille et al., 2015). This suggests that after 24 hours of CTEMPO treatment, the biofilm has been sufficiently dispersed that many cells are becoming dislodged and flowing out of the system.

Verderosa et al. built upon this work by synthetically linking the nitroxide to the ciprofloxacin moiety to create a variety of ciprofloxacin-nitroxide hybrids (Verderosa et al., 2016; Verderosa et al., 2017; Verderosa et al., 2019a; Verderosa et al., 2019b). These hybrid compounds were effective in dispersing and eradicating both Gram-negative (P. aeruginosa, uropathogenic E. coli (UPEC)) as well as Gram-positive (S. aureus) biofilms in vitro (Verderosa et al., 2017; Verderosa et al., 2019a; Verderosa et al., 2019b). P. aeruginosa biofilms were 94% eradicated with ciprofloxacin-nitroxide hybrid-27 (structure shown in Figure 4 ) treatment at 20 µM (Verderosa et al., 2017). Treatment with the hybrid was on par with the synergy reported previously against P. aeruginosa biofilms using CTEMPO and ciprofloxacin co-treatment (Reffuveille et al., 2015). However, hybrid compounds provide some intrinsic advantages over co-treatments, such as i) more favourable pharmacokinetic and pharmacodynamic properties and ii) insurance that the dual action of dispersal and eradication are contained in the one compound, which is especially important for in vivo studies (Verderosa et al., 2017; Fleming and Rumbaugh, 2018). In a follow-up study by Verderosa et al., co-treatment and treatment with hybrid compounds was tested against S. aureus biofilms (Verderosa et al., 2019a). Co-treatment with nitroxide (8 µM for CTEMPO) and ciprofloxacin (256 µM) was effective at eradicating biofilms, affording a 16-fold improvement ( Table 1 ) in the minimum biofilm eradication concentration (MBEC) (99.9% eradication) compared to ciprofloxacin treatment alone (Verderosa et al., 2019a). Interestingly, treatment with the ciprofloxacin-nitroxide hybrid-27 at 64 µM was 4-fold more potent than co-treatment and 64-fold more potent than ciprofloxacin alone, suggesting that hybrid compounds (dispersal agent linked to antibiotic) may provide distinct advantages over co-treatments against some species (Verderosa et al., 2019a). In a subsequent study, the authors developed a synthetic strategy for altering the ratio of nitroxide to antibiotic producing a dinitroxide-ciprofloxacin hybrid (CDN-11, structure shown in Figure 4 ) with potent activity against UPEC biofilms in vitro (Verderosa et al., 2019b). CDN-11 was effective in eradicating biofilm residing UPEC, with a 99% reduction in bacterial numbers (vs. untreated group) at 12.5 µM ( Table 1 ) (Verderosa et al., 2019b). The use of nitroxides in the eradication of biofilms is certainly encouraging, especially considering how well they can potentiate the activity of ciprofloxacin (either as a co-treatment or in hybrid compounds). However, their ability to potentiate the activity of other classes of antibiotics remains to be explored as does their antibiofilm activity in vivo. The lack of human cell toxicity reported for nitroxides and nitroxide functionalised antibiotics (Sadowska-Bartosz et al., 2015; Verderosa et al., 2017) however supports their further development as antibiofilm antimicrobials.

Clearly, NO and nitroxides are effective biofilm dispersal agents and based on in vitro and in vivo studies, they both hold genuine promise as treatment strategies for biofilm related infections. As co-treatment agents they have also showed synergy with several antibiotics and proved to be effective at eradicating established biofilms. While the combined use of NO, nitroxides and antibiotics is still in early preclinical testing, its development towards clinical applications is likely to progress. Continued assessment against other biofilm-forming organisms (in vitro and in vivo) is needed to ascertain how broad-spectrum these combinations can be. NO has already been tested against several well established biofilm-forming species, including the fungal pathogen Candida albicans, and proven efficacious for biofilm dispersal (Barraud et al., 2009). Further work expanding the testing of such combination treatments towards several other species would aid both their development as therapeutics and understanding of the mechanism behind NO’s biofilm dispersing ability, which is likely to differ between microbes.

Antimicrobial peptides (AMPs) are small cationic and amphipathic molecules of 12-50 amino acids that have the ability to disrupt bacterial membranes (Grassi et al., 2017). AMPs exhibit a broad-spectrum of activity and have a high potential to target metabolically dormant cells in biofilms (Grassi et al., 2017). AMPs are often studied in combination with antibiotics and other antibiofilm compounds, and many of these studies have been briefly described by Grassi et al. in 2017 (Grassi et al., 2017). Here, we report the efficacy afforded through combination treatment compared to standalone and also include studies that have been published since then.

An early study on AMP-antibiotic combinations assessed the synergistic activity of the synthetic AMP G10KHc (amino acid sequence shown in Table 2 ) in combination with tobramycin for the eradication of P. aeruginosa biofilms (Eckert et al., 2006). Treatment with G10KHc or tobramycin alone (each at 100 µg/mL) for 4 and 24 hours had little effect on established biofilms (Eckert et al., 2006). However, when used together at the same concentrations, a 4-log reduction in bacterial numbers ( Table 1 ) was observed after 4 hours of co-treatment, with no viable CFU recovered from the biofilms after 24 hours (Eckert et al., 2006). The same group further investigated how G10KHc potentiated tobramycin activity by staining cells with propidium iodide (PI) and observing whether the dye was present in treated cells, as it cannot enter cells with intact membranes (Krishan, 1975; Eckert et al., 2006). Bacterial cells treated with G10KHc were found to fluoresce, indicating that the peptide disrupted the cell membrane, subsequently allowing tobramycin to enter and kill cells (Eckert et al., 2006).

Amino acids can exist in nature as a D-isomer or L-isomer, where the orientation of the alpha carbon in the molecule determines its chirality. The L-isomer is most common in ribosomal peptide synthesis, but recently D-amino acids have also been found in mammals as regulators of neurogenesis and brain receptor function, and also components of some bacterial membranes (Cava et al., 2011). These have recently emerged as a class of potential biofilm dispersal agents given their role as regulators of biofilm dispersal (Cava et al., 2011). So far, the mode of action for D-amino acids has only been delineated in Bacillus subtilis biofilm dispersal, where the bacteria were reported to release D-leucine, -methionine, -tyrosine, and -tryptophan, which disrupted the amyloid fibers linking biofilm cells together at nanomolar concentrations (Kolodkin-Gal et al., 2010).

While their exact mode of action has yet to be elucidated in other species, a few studies have been conducted using amino acids and antibiotics together as a biofilm treatment strategy. In one of these studies, Sanchez et al. investigated the ability of D-amino acids and antibiotics to inhibit P. aeruginosa and S. aureus biofilm formation (Sanchez et al., 2014). Here, an equimolar mixture of D-amino acids (methionine, phenylalanine, and tryptophan) in combination with rifampicin were shown to reduce the minimum biofilm inhibition concentration (MBIC) by 4-fold and a >2-log reduction in viable bacteria was reported with treated S. aureus biofilms (Sanchez et al., 2014). Similar results were observed for P. aeruginosa, with a >2-log reduction in viable bacteria when treated with 64 µg/mL rifampicin, and 5 mM of the D-amino acid mixture (Sanchez et al., 2014). Another study investigating amino acids in combination with antibiotics, conducted by Warraich et al., reported that compared to standalone treatment, lower concentrations of ciprofloxacin and D-amino acids were synergistic (<40 mM amino acids, <90.54 µM ciprofloxacin), and resulted in almost 97% inhibition of biofilm formation and 97.6% dispersal (D-isomer structures shown in Figure 7 ). However, biofilm eradication activity was not tested. The study also investigated the ability of L-amino acids to inhibit biofilm formation, and found that L-isomers of aspartic acid and glutamic acid were equally as effective as the D-isomers at 40mM (Warraich et al., 2020). Only two other studies have reported similar findings with L-amino acids (aspartate and glutamate) (Tong et al., 2014; Yang et al., 2015), challenging the dogma that L-isomers do not have antibiofilm activity. Conflicting evidence reports that L-amino acids can even encourage biofilm formation in some species (Hochbaum et al., 2011; Velmourougane and Prasanna, 2017).

For D-amino acids the evidence is more clear-cut, with several studies reporting some antibiofilm activity (inhibition and dispersal) and in a range of settings, from the industrial to development of drug delivery systems (Si et al., 2014; Wei et al., 2015; Li et al., 2016; Zilm et al., 2017). In these studies, D-amino acids (d-leucine, -methionine, -tyrosine, and -tryptophan) alone and in combination with other compounds effectively inhibited Desulfovibrio vulgaris biofilms and polymicrobial biofilms (Si et al., 2014; Wei et al., 2015; Li et al., 2016; Zilm et al., 2017). Li et al. and Si et al. investigated disruption of established D. vulgaris biofilms that had been grown on carbon steel coupons (Li et al., 2016) or multi-species aggregates collected from an activated sludge reactor (Si et al., 2014) that had grown for longer than 10 days and up to 6 months (Si et al., 2014; Li et al., 2016). Both studies reported that combinations of D-amino acids with other compounds (hydroxymethyl phosphonium sulfate (THPS) and norspermidine) disrupted biofilm surface attachment (Si et al., 2014; Li et al., 2016). Titanium oxide nanoparticles engineered to release D-amino acids upon stimulation with UV light were also shown effective in dispersing B. subtilis biofilms (Wei et al., 2015), and D-amino acids were also found to inhibit E. faecalis biofilm formation in vitro (Zilm et al., 2017). Some conflicting evidence exists around the capacity of D-amino acids to disperse P. aeruginosa and S. aureus biofilms (Sarkar and Pires, 2015; Kao et al., 2017). Kao et al. investigated the ability of D-amino acids alanine, leucine, methionine, tryptophan, and tyrosine (10 μM, 1 mM, and 10mM) to inhibit biofilm formation of P. aeruginosa strains PAO1 and PA14. Here they concluded that D-amino acids slow biofilm growth but do not prevent its formation (Kao et al., 2017), which Sarkar et al. also report for S. aureus biofilms treated with D-tryptophan and D-tyrosine at 1 and 5 mM (Sarkar and Pires, 2015).

Several questions remain unanswered about the antibiofilm properties of amino acids and thus their potential to be employed in future antibiofilm strategies needs more fundamental research. This should aim to resolve the role of both D- and L- isomers in biofilm inhibition and dispersal and explore the specificity of their activity in key biofilm-producing species, such as P. aeruginosa and S. aureus. Their mode of action is likely to vary between species, and so, future research on this group of potential antibiofilm agents should be prioritized.