Versatility of targeted antibiotic-loaded gold nanoconstructs for the treatment of biofilm-associated bacterial infections

Abstract

Background: We previously demonstrated that a photoactivatable therapeutic approach employing antibiotic-loaded, antibody-conjugated, polydopamine (PDA)-coated gold nanocages (AuNCs) could be used for the synergistic killing of bacterial cells within a biofilm. The approach was validated with a focus on Staphylococcus aureus using an antibody specific for staphylococcal protein A (Spa) and an antibiotic (daptomycin) active against Gram-positive cocci including methicillin-resistant S. aureus (MRSA). However, an important aspect of this approach is its potential therapeutic versatility.

Methods: In this report, we evaluated this versatility by examining the efficacy of AuNC formulations generated with alternative antibodies and antibiotics targeting S. aureus and alternative combinations targeting the Gram-negative pathogen Pseudomonas aeruginosa.

Results: The results confirmed that daptomycin-loaded AuNCs conjugated to antibodies targeting two different S. aureus lipoproteins (SACOL0486 and SACOL0688) also effectively kill MRSA in the context of a biofilm. However, our results also demonstrate that antibiotic choice is critical. Specifically, ceftaroline and vancomycin-loaded AuNCs conjugated to anti-Spa antibodies were found to exhibit reduced efficacy relative to daptomycin-loaded AuNCs conjugated to the same antibody. In contrast, gentamicin-loaded AuNCs conjugated to an antibody targeting a conserved outer membrane protein were highly effective against P. aeruginosa biofilms.

Conclusions: These results confirm the therapeutic versatility of our approach. However, to the extent that its synergistic efficacy is dependent on the ability to achieve both a lethal photothermal effect and the thermally controlled release of a sufficient amount of antibiotic, they also demonstrate the importance of carefully designing appropriate antibody and antibiotic combinations to achieve the desired therapeutic synergy.

Keywords:

• Photothermal killing
• Staphylococcus aureus
• Pseudomonas aeruginosa
• biofilms
• gold nanocages

Introduction

Acquired antibiotic resistance is a growing problem that has dramatically limited the efficacy of conventional antibiotic therapy in the context of many bacterial pathogens. In fact, recent years have seen the emergence of pan-resistant bacterial strains, particularly among Gram-negative pathogens including Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii [1–3]. This problem is further compounded by the fact that many bacterial pathogens cause biofilm-associated infections, with the presence of the biofilm conferring a therapeutically relevant degree of intrinsic antibiotic resistance [4]. Taken together, these observations emphasise the urgent need for alternative methods to fight bacterial infections.

It is this need that prompted us to investigate a nanotechnology-based approach as an alternative means of combating bacterial infections. Such approaches have been widely applied in the context of cancer [5,6], but only recently have begun to be investigated in the context of infectious disease. Using 10–40 nm spherical gold nanoparticles conjugated to antibodies targeting staphylococcal protein A (Spa), we demonstrated that pulsed laser irradiation could be used to achieve photothermal (PT) effects, including the formation of microbubbles that were capable of killing Staphylococcus aureus grown under planktonic conditions in vitro [7]. Subsequent studies employing antibody-conjugated gold nanorods and nanotubes confirmed the utility of this approach in vivo in the clinical context of S. aureus sepsis [8]. However, in subsequent unpublished studies, we found that the efficacy of this approach was significantly reduced when assessed in vitro or in vivo in the context of an established biofilm-associated infection.

This limitation led us to explore the use of antibody-conjugated, antibiotic-loaded, polydopamine (PDA)-coated gold nanocages (AuNC) as a means of achieving highly targeted, laser-assisted PT effects and the simultaneous controlled release of antibiotics directly at the site of infection [9]. Using an anti-Spa (aSpa) antibody and an antibiotic (daptomycin) that is effective against Gram-positive cocci, we confirmed that these combined effects were therapeutically synergistic and capable of eradicating viable S. aureus cells, including methicillin-resistant S. aureus (MRSA), even when the targeted cells were present within an established biofilm. Thus, we believe this approach offers great therapeutic promise with the potential to overcome the intrinsic resistance of biofilm-associated infections.

Another advantage of this approach is its potential for therapeutic versatility [10]. More directly, using photoactivatable PDA-coated AuNCs (AuNC@PDA) as the central component, it is possible that different antibodies could be used to enhance the sensitivity and/or coverage of our approach for diverse strains of S. aureus. For instance, while essentially all strains of S. aureus produce Spa, they do so at widely variable levels [9]. To the extent that both the PT and antibiotic release effects are dependent on the localisation of a sufficient number of AuNCs to the bacterial cell surface, this leaves open the possibility that the therapeutic efficacy of our approach could be compromised with S. aureus strains that produce Spa at relatively low levels. This accounts for our inclusion of the USA300, MRSA strain LAC (Los Angeles County clone) and the USA200, methicillin-sensitive S. aureus (MSSA) strain UAMS-1 in our previous experiments, with the former producing Spa at significantly lower levels than the latter [9]. While we found that this did not diminish the efficacy of our approach, this does not preclude the possibility that reduced efficacy might be observed with other strains, nor does it preclude the possibility that, even if this is not the case, sensitivity could be enhanced through the use of alternative S. aureus-specific antibodies, perhaps in combination with each other. This is important because it could reduce the amount of laser irradiation required to achieve the desired therapeutic effects. A second important aspect of this potential versatility is that our approach could be used to target other bacterial pathogens depending only on the use of appropriate antibodies and antibiotics.

In this report, we explored these possibilities by examining the therapeutic efficacy of our nanotherapeutic approach using alternative antibodies and antibiotics targeting S. aureus as well as an antibody and antibiotic combination chosen specifically to target the gram-negative pathogen P. aeruginosa. The results confirmed the utility of two alternative antibodies that target proteins shown to be present on the surface of biofilm-associated S. aureus cells [11,12] but also revealed the need to carefully consider the choice of antibiotic. Perhaps most importantly, they also confirmed the efficacy of our nanotherapeutic approach in the context of biofilms formed by P. aeruginosa.

Materials and methods

Synthesis of PDA-coated AuNCs

AuNC@PDA were synthesised using a protocol modified from our previously published method [9]. Briefly, the AuNCs with LSPR peak at ∼750 nm were synthesised by a galvanic replacement reaction between Ag nanocubes and HAuCl₄ [13,14]. Then, 1 ml of 24 nM AuNCs were added to 49 ml of an aqueous solution of Trizma base (0.5 mM, 60.5 mg) preheated to 50 °C, followed by adding dopamine hydrochloride (2.9 mM, 55 mg) to initiate the self-polymerization of dopamine on the surface of AuNCs [15]. The reaction was carried out in a 200 ml Erlenmeyer flask with a disperser (IKA T 18 digital ULTRA-TURRAX) equipped with a dispersing element (IKA S 18 N-19 G) and allowed to proceed for 1 h to generate AuNC@PDA. After this reaction, the product was collected by centrifugation at 9000 relative centrifugal force (rcf) for 30 min, washed with H₂O twice, and recovered by centrifugation at 21 000 rcf for 10 min at 4 °C. The AuNC@PDA preparation was suspended in H₂O at a concentration of 2.5 nM for biophysical characterisation and conjugation to antibodies as detailed below.

Conjugation of antibodies to AuNC@PDA constructs

All antibodies employed in these experiments were affinity-purified preparations of IgG. As in our previous report, antibodies were conjugated to the surface of AuNC@PDA via Michael addition of the primary amines of the lysines and/or N-terminus of the antibody to the aromatic rings of the PDA coating [9]. Briefly, 1 nM AuNC@PDA was dispersed in 1 ml of 10 mM bicene buffer (pH = 8.5) before adding 0.1 nmol of the appropriate IgG antibody to the solution. The reaction was allowed to proceed at 4 °C for 1 h. The conjugates were collected and washed three times with phosphate buffered saline (PBS) followed by centrifugation at 21 000 rcf for 5 min at 4 °C. The conjugates were dispersed in PBS and stored at 4 °C.

Antibodies used in these studies included one targeting Spa (Sigma Aldrich, P3775, designated aSpa), another targeting an S. aureus lipoprotein (Lpp) of unknown function encoded by SACOL0486 that was previously shown to be expressed in increased amounts in an S. aureus biofilm [16] (antibody designated aLpp), a third targeting a manganese transporter (MntC) encoded by SACOL0688 and also shown to be expressed at increased levels in an S. aureus biofilm [12] (antibody designated aMntC), and a fourth targeting a conserved P. aeruginosa outer member protein (Abcam, ab35835, designated aPa).

Loading of antibiotics to antibody-conjugated AuNC@PDA constructs

Following antibody conjugation, antibiotics were loaded into antibody-conjugated AuNC@PDAs at pH = 5.5 to prepare antibody-conjugated, antibiotic-loaded AuNC@PDAs. Briefly, 1 nM antibody-conjugated AuNC@PDA was incubated with 2 mg/mL of antibiotic in 1 ml of 100 mM citrate buffer solution (pH = 5.5). The reaction was stirred overnight at 4 °C in the dark. The product was collected by centrifugation at 21 000 rcf for 10 min, purified by washing with citrate buffer solution twice, and collected by centrifugation at 8000 rcf for 10 min to remove free antibiotic. Antibody-conjugated, antibiotic-loaded AuNC@PDA were suspended in Dulbecco’s phosphate buffered saline (DPBS; pH = 7.4) at a concentration of 2 nM and stored at 4 °C until used.

Antibiotics incorporated into antibody-conjugated AuNC@PDA constructs included daptomycin, ceftaroline, vancomycin, and gentamicin. Specific combinations included daptomycin loaded into AuNC@PDA conjugates with each S. aureus antibody (AuNC@Dap/PDA-aSpa, AuNC@Dap/PDA-aLpp, and AuNC@Dap/PDA-aMntC), ceftaroline fosamil and vancomycin loaded into AuNC@PDA conjugated to aSpa (AuNC@Cef/PDA-aSpa and AuNC@Van/PDA-aSpa, respectively), and gentamicin loaded into AuNC@PDA conjugated to aPa (AuNC@Gen/PDA-aPa). As determined under standard planktonic growth conditions, the S. aureus strain used in these experiments (LAC) was determined to be susceptible to daptomycin, vancomycin, and ceftaroline with a minimum inhibitory concentration (MIC) of 0.5, 2.0 and 0.5 µg per ml, respectively (data not shown). The P. aeruginosa strain used (ATCC 27317) was also determined to be susceptible to gentamicin with an MIC of 1.5 µg per ml (data not shown).

Quantification of antibiotic release as a function of PT effects

The release of antibiotics from each of the antibiotic and antibody formulations described above was assessed in the absence of laser irradiation (dark release) and as a function of the time of laser irradiation (PT release). These studies were carried out in DPBS (pH = 7.4). Briefly, each construct was suspended in 200 µL DPBS at a concentration of 0.4 nM at room temperature. To assess dark release, samples were incubated for 24 h at 37 °C without near-infra-red (NIR) irradiation. To assess PT release, samples were irradiated using an 808-nm diode laser at a power density of 1 W/cm² in a clear 96-well plate. Samples were then taken immediately following irradiation.

All samples were filtered through a 0.22 µm syringe filter before determining antibiotic concentrations by ultra-performance liquid chromatography (UPLC, Waters Acquity). Elution was performed using a mobile phase consisting of a gradient (90:10 to 10:90) of H₂O and acetonitrile (1% trifluoroacetate) through a phenyl stationary phase (BEH phenyl, Acquity) at a flow rate of 0.2 ml/min with ultraviolet detection at 262 nm for daptomycin, 254 nm for ceftaroline fosamil, 205 nm for vancomycin, and 250 nm for gentamicin. The gradient of the mobile phase was as follows: from 0 to 2 min 90:10 of H₂O and acetonitrile; at 3 min 50:50 of H₂O and acetonitrile; at 5 min 40:60 of H₂O and acetonitrile; at 6 min 10:90 of H₂O and acetonitrile; and from 7 to 8 min 90:10 of H₂O and acetonitrile. Daptomycin was eluted at ∼3.5 min while ceftaroline fosamil, vacomycin, or gentamicin were eluted at ∼1.5 min. Peak integral for each antibiotic was linear over the concentrations tested.

Characterization of AuNC nanocontructs

Characterization of nanocontructs was done by transmission electron microscopy (TEM), analysis of UV–Vis extinction spectra, and analysis of temperature profiles as previously described [9]. To assess the temperature profiles associated with irradiation, a diode laser centred at 808 nm was used to irradiate 200 µL of a 0.4 nM suspension of nanoconstructs suspended in a well of 96-well plate at a power density of 1 W/cm². Temperature measurements were obtained using an infra-red sensor. The time-dependent temperature profile of the suspension was plotted as previously described [9]. Based on curve fitting analysis, the rate of energy absorption and the rate of heat dissipation were determined as previously described [17].

Assessment of antibody-mediated targeting efficiency

Localization of different AuNC@PDA formulations with S. aureus cells was done using photothermal microscopy (PTM). Specifically, the S. aureus strain UAMS-1 was transformed with a plasmid (pCM11) that allowed for constitutive expression of the gene encoding super green fluorescent protein (sGFP). This strain was grown in tryptic soy broth (TSB) overnight at 37 °C, harvested by centrifugation, and suspended in PBS at a concentration of 10⁷ colony-forming units (cfu) per mL. Four AuNC formulations (AuNC@PDA, AuNC@Dap/PDA-aSpa, AuNC@Dap/PDA-aLpp and AuNC@Dap/PDA-aMntC) were tested for labelling efficiency. Three spots per AuNC formulation were demarcated on microscope slides using an ImmEdge pen (Vector Laboratories, Burlingame, CA, USA) and 5 µL of the standardised bacterial culture added to each spot. After heat fixation as commonly employed in a conventional Gram stain protocol, 90–140 µL of each AuNC formulation was placed on each spot to achieve a ratio of approximately 2.7 × 10⁴ AuNC per bacterial cell. Samples were allowed to incubate for 1 h at room temperature. Spots were then washed with PBS, covered with a clear coverslip, and sealed using ProLong Diamond Antifade Mountant (Molecular Probes, Eugene, OR, USA).

Photothermal microscopy was performed as previously described [18]. Briefly, a custom built platform based on an inverted Olympus IX73 (Olympus America, Inc., Central Valley, PA, USA) using a 3-wavelength Wavelength Division Multiplexer (WDM or RGB Combiner, RGB46HF; Thorlabs, Newton, NJ, USA) was used to combine 488 nm (fluorescence excitation: IQ1C45 (488–60) laser diode; Power Technology, Little Rock, AR, USA), 532 nm (PT pump: LabSpec 532 nm DPSS Laser; Laserglow Technologies, LLS series, Toronto, Canada) and 635 nm (PT probe: LP637 SM Fiber-Pigtailed Laser Diode; Thorlabs, Newton, NJ, USA) laser beams into a single mode fibre. High resolution confocal fluorescence and PTM imaging were carried out simultaneously by steering laser beams using galvo-mirrors (GVSM002; Thorlabs, Newton, NJ, USA) across the sample. Probe beam intensity was collected using 40× objective located above the sample and measured by amplified photodiode (PDA10A; Thorlabs, Newton, NJ, USA). PT signal was measured using a digital lock-in amplifier (MFLI, 500 kHz, 60MSa/s, Zurich Instruments, Switzerland) and recorded using custom software developed on the LabView platform. Conventional fluorescent imaging was carried out using a CCD camera DP80 (Olympus America, Central Valley, PA, USA). PT and fluorescence images were merged and PT signal was quantified pixel-by-pixel using ImageJ software. Each AuNC sample was run in triplicate and 95% confidence intervals were obtained for labelling efficiency and signal intensity relative to a control of a sample of bacteria alone.

Antibacterial efficacy studies

The antibacterial efficacy of our various AuNC formulations was assessed as previously described [9]. Briefly, 14-gauge fluorinated ethylene propylene catheters (Braun, Melsungen, Germany) were cut into 0.5 cm segments, sterilised, and coated in human plasma [19]. Catheters were then placed in the wells of a 12-well microtiter plate containing 2 ml of TSB supplemented with glucose and sodium chloride (biofilm medium, BFM) [19]. Each well was then inoculated with the MRSA strain LAC or the P. aeruginosa strain ATCC 27317 at an optical density at 560 nm of 0.05. A biofilm was allowed to form on the catheters by incubation at 37 °C for 24 h. To treat catheters with our various AuNC formulations, catheters were rinsed in sterile PBS and transferred to the wells of a 48-well microtiter plate. Each well contained 500 μL of BFM with the test AuNC formulation at a concentration of 0.4 nM (1.2 × 10¹¹ AuNC/well). Before addition to the wells, all AuNC formulations were sonicated to disrupt aggregates (Bransonic 2800, Branson Ultrasonics, Danbury, CT, USA). For the assessment of formulations loaded with daptomycin, the BFM was supplemented with 2.5 mM CaCl₂ [20]. All plates were covered with a gas-permeable sealing membrane (Diversified Biotech, Dedham, MA, USA) to prevent evaporation during irradiation. As assessed using a power metre (Newport Power Meter Model 1918-R; Newport Corp., Irvine, CA, USA), this membrane was previously shown to reduce the laser power reaching the underlying catheters by ∼30% [9]. Colonized catheters were either not irradiated or irradiated for 2.5, 5, 7.5, or 10 min using a Diomed 25 clinical laser with a peak at 808 nm. Taking into account power loss through the membrane, the final power density used for efficacy studies was measured to be 0.8 W/cm². Following irradiation, a subset of catheters (n = 3) for each group was harvested immediately (0 h), rinsed in sterile PBS to remove non-adherent cells, and sonicated to disrupt the biofilm. Viable bacteria were subsequently quantified by serial dilution and plate counts. Reductions in the number of viable bacteria observed at this time point were considered indicative of the PT effect. The remaining catheters (n = 3) were allowed to incubate for an additional 24 h at 37 °C in the wells in which they were irradiated, after which catheters were removed, sonicated, and the number of viable bacteria determined. Reductions in the number of viable bacteria at this time point were considered indicative of the effect of antibiotic release.

Results

The central element of our nanotherapeutic approach is the AuNC@PDA. The PDA coating is included to minimise AuNC aggregation and enhance in vivo stability and biodistribution. In addition to the AuNC itself, the PDA coating also provides a reservoir for antibiotic loading as well as a surface amenable to antibody conjugation. By comparison to the AuNC@PDA formulation used in our previous studies [9], we modified the coating protocol for these studies (see Materials and Methods) in an effort to achieve more uniform coating without altering the absorption spectrum of our AuNC@PDA, both of which were subsequently confirmed. Specifically, nanocontructs prepared in this fashion were found to be composed of a 36.6 ± 4.3 nm AuNC core and 41.4 ± 3.0 nm PDA shell, and the UV–Vis spectrum confirmed an LSPR peak in the targeted NIR range at 830 nm (Figure 1). Both of these parameters were unchanged following laser irradiation (data not shown). The rate of energy absorption and the rate of heat dissipation were determined to be 0.24 °C/s and 0.0061 s⁻¹, respectively.

Figure 1. Structural characteristics of PDA-coated AuNCs (AuNC@PDA). TEM image of AuNC@PDA showing relative size of AuNC core and PDA coating (left). UV−Vis spectrum of AuNC@PDA in aqueous suspension (right).

Using a 48-well microtiter plate experimental format, colonised catheters exposed to daptomycin-loaded PDA-coated AuNCs conjugated to anti-Spa antibody (AuNC@Dap/PDA-aSpa) were subjected to laser irradiation (0.8 W/cm²) for varying periods of time (0, 2.5, 5.0, 7.5, and 10.0 min). These studies revealed a time-dependent heating profile that was consistent with the profile observed in our previous experiments [9]. Specifically, the temperature increased in a relatively linear manner through the first 5 min of irradiation before reaching a plateau of approximately 60 °C, with this heat being rapidly dissipated once laser irradiation was stopped (Figure 2).

Figure 2. Temperature profile as a function of laser irradiation. Temperature was recorded in samples consisting of a colonised catheter in 500 μL of a 0.4 nM AuNC@Dap/PDA-aSpa suspension in BFM. Temperature was recorded in the absence of irradiation (time 0) and after continuous irradiation for 10 min, at which point laser irradiation was stopped and the temperature recorded for an additional 10 min. The experiment was done in triplicate with the results shown as the average ± the standard error of the mean (SEM).

As with our previously published study [9], bacterial viability was assessed at two time points following irradiation to allow for the assessment of PT killing as defined immediately after irradiation (0 h) and antibiotic-mediated killing as defined after continued incubation (24 h). In these studies, the AuNC@Dap/PDA-aSpa formulation was found to be capable of clearing an established S. aureus biofilm (Figure 3). Specifically, colonised catheters exposed to AuNC@Dap/PDA-aSpa and irradiated for various lengths of time exhibited a dose-dependent PT effect, with PT killing increasing as a function of increased irradiation time. Additionally, as the PT effect and temperature increased with increasing duration of irradiation, the amount of antibiotic released from the AuNC@Dap/PDA-aSpa formulation also increased (Figure 4). This result was to be anticipated as it was previously demonstrated that antibiotic release is in fact dependent on heat released during the PT effect [9]. The fact that the PT-mediated effect did not result in complete clearance from all catheters at any time point allowed for demonstration of the synergism of subsequent antibiotic release. Specifically, all catheters irradiated for 7.5 and 10 min were not completely cleared of viable bacteria by the PT effect alone but were completely cleared when assessed 24 h following irradiation (Figure 3).

Figure 3. Killing of biofilm-associated S. aureus with AuNC@Dap/PDA-aSpa. Biofilms were allowed to form on catheters before being placed into 500 μl of biofilm medium (BFM) containing AuNC@Dap/PDA-aSpa at a final concentration of 0.4 nM. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Figure 4. Antibiotic release from AuNC@Dap/PDA-aSpa as a function of laser irradiation. A 200 µl suspension of AuNC@Dap/PDA-aSpa constructs at a concentration of 0.4 nM were either not irradiated (0 min) or exposed to laser irradiation for the indicated period of time. Numbers above each bar indicate the concentration (µg/ml) observed with each sample as determined by UPLC. Concentration observed in the absence of laser irradiation (0 min) indicate dark release. Concentrations observed after irradiation indicate PT release.

An important aspect of our nanotherapeutic approach is the versatility of design, making it possible to substitute certain components to optimise bactericidal effects. One of the components suitable for substitution is the antibody conjugated to the AuNC for bacterial targeting, theoretically allowing for increased coverage for different bacterial strains and/or enhanced PT effects. For this reason we next sought to demonstrate the ability to conjugate our AuNC@PDA to alternative antibodies targeted to S. aureus surface proteins. Two antibodies (aLpp and aMntC) were chosen for these studies based on the observation that both of the S. aureus targets of these antibodies are produced in increased amounts in a biofilm relative to planktonically grown cells [12,16,21,22]. These alternative constructs were designated AuNC@Dap/PDA-aLpp and AuNC@Dap/PDA-aMntC, respectively.

As with the aSpa antibody, both of these consisted of purified IgG, thus allowing us to employ the same conjugation methods employed in our earlier studies [9]. Therefore, the focus of these studies was not on the ability to conjugate AuNCs to these alternative antibodies but rather on their utility as potential targeting agents. An essential component of that assessment was the demonstration that conjugation to these antibodies did not significantly alter critical properties of our AuNC system. Specifically, the possibility existed that conjugation to alternative antibodies could alter the degree of targeting and, in turn, the PT profile from the conjugated AuNCs. To this end, we assessed the heating profile of AuNC@PDA conjugated to these alternative antibodies following exposure to identical irradiation parameters. The same heating profile was observed with all daptomycin-loaded, AuNC@PDA irrespective of the antibody used (Figure 5), thus confirming that the heating profile is a function of the time of irradiation and that it is not affected by the antibody component of the formulation.

Figure 5. Temperature profile as a function of AuNC@PDA formulation. Temperature was recorded in samples consisting of a colonised catheter in 500 μL of a 0.4 nM suspension of the indicated AuNC formulation in BFM. Temperature was recorded in the absence of irradiation (0 min) and after irradiation, with readings taken at 2.5 min intervals through 10 min. The experiment was repeated 6 times with each AuNC formulation. Results are shown as the average ± the standard error of the mean (SEM).

We next examined the relative targeting capacity of antibody-conjugated AuNCs relative to unconjugated AuNC@PDA. Using a derivative of the S. aureus strain UAMS-1 transformed with pCM11 allowed us to do this using confocal fluorescent microscopy and PTM. The results confirmed antibody conjugation greatly enhanced the localisation of AuNCs to bacterial cells (Figure 6). Quantitative analysis confirmed that AuNC@Dap/PDA-aSpa co-localized with 92.0 ± 5.9% of S. aureus cells compared to only 21.2 ± 10.9% of S. aureus cells that co-localized with AuNC@PDA. PTM also allowed for quantification of relative signal intensity from AuNCs co-localized to S. aureus cells as measured by arbitrary units (a.u.), and the relative signal intensity was found to be higher for AuNC@Dap/PDA-aSpa than AuNC@PDA (7.1 ± 3.6 a.u. vs. 0.8 ± 0.4 a.u., respectively). These results indicate that, by comparison to unconjugated AuNC@PDA, more S. aureus cells co-localized with antibody-conjugated AuNCs and that the number of antibody-conjugated AuNCs per cell was higher.

Figure 6. Localization of AuNC@PDA as a function of antibody conjugation. Images illustrate fluoresence (sGFP) and photothermal microscopy (PTM) images of S. aureus cells (green) incubated with AuNC@PDA (blue) with and without conjugation to aMntC, aLpp, or aSpa. Scale bars =10 µm.

Photothermal microscopy also allowed us to compare targeting efficiency of AuNCs conjugated to different antibodies (Figure 6). Specifically, when these experiments were repeated with AuNC@Dap/PDA-aLpp and AuNC@Dap/PDA-aMntC, we found that 53.2 ± 10.4% and 44.3 ± 8.1% of S. aureus cells co-localized with these AuNC formulations, respectively. Although these percentages are lower than those observed with AuNC@Dap/PDA-aSpa, it must be noted that these experiments were performed with S. aureus cells that were grown under planktonic conditions, and both of the antibodies were chosen because they target antigens that were previously shown to be produced in increased amounts in a biofilm [12,16,21,22]. Thus, these results confirm the necessity of antibody conjugation, but also demonstrate the need to carefully choose appropriate antibody targets.

We next sought to assess the relative therapeutic efficacy of AuNC formulations conjugated to these alternative antibodies in comparison to AuNC@Dap/PDA-aSpa. In the case of AuNC@Dap/PDA-aLpp, we observed a clear, time-dependent PT effect (Figure 7), although it did appear to be slightly reduced by comparison to AuNC@Dap/PDA-aSpa (Figure 3). This is perhaps due to decreased targeting of the AuNC@Dap/PDA-aLpp formulation as suggested by PTM analysis. Nevertheless, our AuNC@Dap/PDA-aLpp formulation clearly demonstrated a synergistic antibiotic-mediated effect with complete clearance of viable bacteria at 24 h from all catheters irradiated for 10 min (Figure 7).

Figure 7. Killing of biofilm-associated S. aureus with AuNC@Dap/PDA-aLpp. Biofilms were allowed to form on catheters before being placed into 500 μl of BFM containing 0.4 nM AuNC@Dap/PDA-aLpp. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Efficacy studies with our AuNC@Dap/PDA-aMntC formulation yielded more complicated results (Figure 8). As with AuNC@Dap/PDA-aSpa and AuNC@Dap/PDA-aLpp, irradiation of AuNC@Dap/PDA-aMntC resulted in PT effects that increased with increasing irradiation duration, although in this case the PT effect was reduced even by comparison to the results observed with AuNC@Dap/PDA-aLpp (Figure 7). Once again, this could be a function of reduced targeting efficiency as suggested by PTM analysis. Additionally, analysis at 24 h demonstrated a decrease in bacterial viability from all groups, including the group not exposed to irradiation. One possible explanation for this finding would be a lack of stability of the AuNC@Dap/PDA-aMntC formulation and release of antibiotic in the absence of irradiation (dark release). Therefore, we evaluated the level of dark release as well as the release of antibiotic from each of our formulations following irradiation for 10 min (PT release) (Figure 9). The results confirmed that some dark release was observed with all antibiotics tested, but that the amount of antibiotic released was also significantly increased after laser irradiation, thus suggesting some other factor may be responsible for this observation.

Figure 8. Killing of biofilm-associated S. aureus with AuNC@Dap/PDA-aMntC. Biofilms were allowed to form on catheters before being placed into 500 μl of BFM containing 0.4 nM AuNC@Dap/PDA-aMntC. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Figure 9. Antibiotic release as a function of laser irradiation. The amount of antibiotic released from the indicated AuNC@PDA formulations in the absence of laser irradiation (dark release) and after 10 min of laser irradiation (photothermal release) is shown for each AuNC formulation. Numbers above each bar indicate the concentration of antibiotic (µg/ml) observed with each formulation as determined by UPLC. Designations on the X-axis indicate the antibiotic and antibody incorporated into each formulation. Dap, daptomycin; Cef, ceftaroline; Van, vancomycin; Gen, gentamicin.

In an effort to further explore the versatility of our nanotherapeutic approach, we next sought to determine whether alternative antibiotics could be loaded into our AuNC@PDA formulations. As the S. aureus strain used in these studies was an MRSA strain, and in the interest of coverage for antibiotic-resistant strains, we chose antibiotics with activity against MRSA. Specifically, we chose to load ceftaroline or vancomycin into AuNC@PDA. Ceftaroline was also chosen because of previous studies demonstrating its relatively high level of activity within established S. aureus biofilms [20]. For studies testing alternative antibiotics, the antibiotic component was changed and the antibody held constant, with all formulations against S. aureus being conjugated to aSpa. In comparison to daptomycin-loaded AuNCs, higher concentrations of PT-mediated release of ceftaroline were achieved (Figure 9), possibly due in part to the relatively small size of ceftaroline compared to daptomycin (Figure 10). Efficacy studies with ceftaroline-loaded AuNCs (AuNC@Cef/PDA-aSpa) were performed as with other formulations and demonstrated similar but somewhat diminished PT-mediated effects compared to those observed with daptomycin-loaded AuNC formulations (Figure 11). The antibiotic-mediated effects as determined by bacterial viability after 24 h were also less effective in clearing catheters of viable bacteria compared to the daptomycin-loaded formulations.

Figure 10. Comparison of antibiotic structures. Molecular structure and relative molecular weight of each antibiotic incorporated into our AuNC@PDA nanoconstructs is shown for comparison.

Figure 11. Killing of biofilm-associated S. aureus with AuNC@Cef/PDA-aSpa. Biofilms were allowed to form on catheters before being placed into 500 µl of BFM containing 0.4 nM AuNC@Cef/PDA-aSpa. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

We also examined the efficacy of our AuNC@PDA loaded with vancomyin and conjugated to aSpa (AuNC@Van/PDA-aSpa). Importantly, despite the fact that we obtained a clear PT effect, we did not observe a synergistic antibiotic-mediated effect at 24 h irrespective of the time of irradiation (Figure 12). This suggests that limited amounts of active vancomycin were released. To address this, antibiotic release of vancomycin was assessed as with each of our other AuNC formulations (Figure 9). These studies demonstrated that low concentrations of vancomycin were released by comparison to other antibiotics even after laser irradiation. This is particularly true when the concentration of vancomycin present following irradiation (3.7 µg/ml) is viewed in the context of the breakpoint minimum inhibitory concentration (MIC) for vancomycin susceptibility (2.0 µg/ml). Moreover, previous studies from our group have demonstrated that concentrations this low have little to no effect on bacterial viability within the context of a biofilm [20]. Whether these relatively low concentrations are a function of a reduced capacity to load vancomycin or a reduced capacity to achieve its PT-mediated release is not clear, but it should be noted that vancomycin is also a relatively large antibiotic (Figure 10). By comparison to the other antibiotics tested, vancomycin also has a relatively large number of hydrophobic benzene rings, which could limit its loading and/or release capacity.

Figure 12. Killing of biofilm-associated S. aureus with AuNC@Van/PDA-aSpa. Biofilms were allowed to form on catheters before being placed into 500 µl of BFM containing 0.4 nM AuNC@Van/PDA-aSpa. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

In an attempt to further test the versatility of our system, we sought to test the ability of our nanotherapeutic approach to target and efficiently eliminate biofilms formed by a pathogen other than S. aureus. To this end, we chose the clinically-relevant, Gram-negative pathogen, P. aeruginosa. To demonstrate this versatility, both the antibiotic and the antibody from our initial formulation were substituted to use an antibiotic (gentamicin) shown to be active against the test strain (ATCC 27317) and an antibody specific for a conserved P. aeruginosa outer membrane protein (aPa). Photothermal release was determined for gentamicin-loaded, aPa-conjugated AuNCs (AuNC@Gen/PDA-aPa) and demonstrated relatively high concentrations compared to daptomycin-loaded formulations (Figure 9). As with ceftaroline, the increased loading capacity of gentamicin is possibly due in part to the smaller size of gentamicin relative to daptomycin and vancomycin (Figure 10).

Efficacy studies were also carried out with AuNC@Gen/PDA-aPa using catheters colonised with P. aeruginosa biofilms in the same manner as that described for S. aureus. The results demonstrated a PT-mediated effect that increased with the time of irradiation to the point that 10 min of irradiation resulted in the complete clearance of all catheters by the PT effect alone (Figure 13). Nevertheless, enhanced efficacy associated with antibiotic release was evident with complete clearance after 24 h of all catheters following irradiation for 7.5 min, a duration which was not sufficient to completely eradicate viable bacteria from all catheters owing to the PT effect alone.

Figure 13. Killing of biofilm-associated P. aeruginosa with AuNC@Gen/PDA-aPa. Biofilms were allowed to form on catheters before being placed into 500 µl of BFM containing 0.4 nM AuNC@Gen/PDA-aPa. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Finally, as a further test of specificity, we repeated these efficacy studies using the AuNC@Gen/PDA-aPa formulation with biofilms formed by S. aureus (Figure 14) and, conversely, the AuNC@Dap/PDA-aSpa formulation with biofilms formed by P. aeruginosa (Figure 15). In contrast to the results discussed above, none of the catheters were cleared in either of these experiments even after 10 min of irradiation.

Figure 14. Killing of biofilm-associated S. aureus with AuNC@Gen/PDA-aPa. S. aureus biofilms were allowed to form on catheters before being placed into 500 μl of BFM containing 0.4 nM AuNC@Gen/PDA-aPa. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Figure 15. Killing of biofilm-associated P. aeruginosa with AuNC@Dap/PDA-aSpa. P. aeruginosa biofilms were allowed to form on catheters before being placed into 500 μl of BFM containing 0.4 nM AuNC@Dap/PDA-aSpa. Control catheters were not irradiated (0 min), while test catheters were irradiated for the indicated period of time. The number of viable bacterial cells was then determined at 0 h (open circles, PT effect) and after an additional 24 h incubation (filled circles, antibiotic effect).

Discussion

We previously demonstrated that a novel therapeutic approach employing laser irradiation of PDA-coated AuNCs loaded with daptomycin and conjugated to an antibody specific for S. aureus protein A could be used to eradicate viable bacteria from an established S. aureus biofilm [9]. While these results were limited to in vitro studies, it is important to note that we have not been able to achieve a comparable effect under such conditions using any existing antibiotic alone [20]. Thus, we believe this approach offers tremendous therapeutic promise, particularly in the context of biofilm-associated infections. In addition, because it consists of component parts, some of which can be changed, it also potentially offers tremendous therapeutic versatility.

The purpose of this report was to experimentally examine this potential by evaluating the use of different antibodies and different antibiotics not only in the context of S. aureus but also in the context of the important Gram-negative pathogen P. aeruginosa, and we believe the results reveal a number of important considerations. First, we confirmed that it is possible to employ alternative antibodies to achieve effective localisation of AuNCs to the targeted pathogen. At the same time, from the perspective of therapeutic effect, not all antibodies behaved in an identical fashion. For instance, by comparison to all other formulations, we observed a greater degree of killing with our AuNC@Dap/PDA-aMntC formulation even in the absence of laser irradiation.

One possible explanation for this enhanced bactericidal activity would be increased dark release of antibiotic, but this seems unlikely based on the observation that a comparable level of dark release was observed with the AuNC@Dap/PDA-aLpp formulation. Moreover, the amount of antibiotic released in the absence of laser irradiation was <2 µg/ml, and our previous studies would suggest that this concentration is not effective in the context of an established S. aureus biofilm [20]. This suggests that the killing observed 24 h after exposure to the AuNC@Dap/PDA-aMntC formulation in the absence of laser irradiation may be due to some other factor. MntC is a manganese transporter and it is possible that binding of antibody to this target alters bacterial viability in and of itself. Indeed, a mutation in the corresponding gene (SACOL0688) has been shown to exhibit an in vitro growth defect in a biofilm (M. Shirtliff, unpublished observations). Examining this possibility will require further experimentation, but in the context of this report it is important to note that, should this prove to be the case, it would argue for the choice of this antibody as it would potentially provide a third level of bacterial killing unrelated to PT effects or the release of antibiotics.

We also observed variability with respect to the use of different antibiotics when tested in the context of S. aureus. Specifically, when evaluated in the context of the same aSpa antibody, daptomycin exhibited greater efficacy than ceftaroline, while ceftaroline exhibited greater efficacy than vancomycin. The results observed with vancomycin are not particularly surprising both because loading and PT release were found to be limited by comparison to other antibiotics and because our previous studies have demonstrated that vancomycin exhibits less efficacy in a biofilm than either daptomycin or ceftaroline [20]. The results observed with ceftaroline are more difficult to explain, particularly since our ceftaroline-loaded formulation exhibited a higher level of PT-mediated antibiotic release than any other antibiotic tested. One possible explanation is that ceftaroline is more sensitive to NIR irradiation and/or heat, thus resulting in a much lower concentration of active ceftaroline than what is measured by UPLC. Thus, these results demonstrate that, while it is possible to load alternative antibiotics into our PDA-coated AuNCs, they also emphasise the need to carefully consider the antibiotics that are most effective in this specific therapeutic context.

Despite these limitations, we were able to demonstrate using an AuNC formulation containing gentamicin and an antibody targeting a conserved outer membrane protein that our approach is effective in the context of other bacterial pathogens, in this case P. aeruginosa. This is perhaps the best indication of the therapeutic versatility of our approach in that it suggests that it could be effectively applied in the context of any bacterial pathogen depending only on identification of an appropriate antibiotic and antibody combination. This is further evidenced by the observation that the use of antibody–antibiotic combinations designed to target S. aureus had little impact on biofilms formed by P. aeruginosa and vice versa.

A potential concern with respect to in vivo applications is the possible tissue damage associated with temperatures like those we observed in these studies. This would presumably be minimised with effective targeting of AuNCs to bacterial cells, and perhaps by employing alternative means of laser irradiation (e.g., pulsed rather than continuous wave), but this is nevertheless an important issue that will ultimately have to be addressed. Thus, while many challenges remain with respect to adapting and optimising our approach for effective use in vivo, particularly in cases involving deep infections and a substantial tissue barrier to laser penetration, we nevertheless believe the results we report provide strong justification to pursue the experiments necessary to achieve this important goal. This is particularly true in an era of increasing antibiotic resistance and in the specific context of the unique clinical problem of biofilm-associated infections.

Conclusions

We demonstrate that PDA-coated AuNCs can be used as a central core element around which different antibodies and antibiotics can be incorporated to achieve the laser-assisted synergistic PT and antibiotic-mediated killing of diverse bacterial pathogens, including both gram-positive and gram-negative species, even in the intrinsically-resistant environment of an established biofilm.

Acknowledgements

The authors would like to acknowledge Dmitry A. Nedosekin for technical assistance with PTM experiments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was funded by a kind gift from Robert H. Schmidt, MD and the Texas Hip and Knee Centre and in part by a grant [R56-AI093126] from the National Institute of Allergy and Infectious Diseases. Support was also provided by Arkansas Biosciences Institute and the Centre for Advanced Surface Engineering from NSF EPSCoR OIA 1457888. T.W. was supported by the Arkansas Breast Cancer Research Program. Additional support was provided by core facilities supported by the Centre for Microbial Pathogenesis and Host Inflammatory Responses [P20-GM103450] and the Translational Research Institute [UL1TR000039].