Nanoparticle loaded hydrogel for the light-activated release and photothermal enhancement of antimicrobial peptides

Rising concerns over multidrug resistant bacteria have necessitated an expansion to the current antimicrobial arsenal and forced the development of novel delivery strategies that enhance the efficacy of existing treatments. Antimicrobial peptides (AMPs) are a promising antibiotic alternative that physically disrupts the membrane of bacteria resulting in rapid bactericidal activity, however clinical translation of AMPs has been hindered by their susceptibility to protease degradation. Through the co-loading of liposomes encapsulating a model AMP, IRIKIRIK-CONH 2 (IK8), and gold nanorods (AuNRs) into a poly(ethylene) glycol (PEG) hydrogel we have demonstrated the ability to protect encapsulated materials from proteolysis and provide the first instance of triggered release of AMPs. Laser irradiation at 860 nm, at 2.1 W cm 2 2 , for 10 mins led to the photothermal triggered release of IK8, resulting in bactericidal activity against Gram-negative Pseudonomas aeruginosa and Gram-positive Staphylococcus aureus . Furthermore, by increasing the laser intensity to 2.4 W cm -2 we have shown the thermal enhancement of AMP activity. The photothermal triggered release, and enhancement of AMP efficacy, was demonstrated to treat two rounds of fresh S. aureus , indicating the therapeutic gel has the potential for multiple rounds of treatment. Taken together, this novel therapeutic hydrogel system demonstrates stimuli-responsive release of AMPs with photothermal enhanced antimicrobial efficacy in order to treat pathogenic bacteria. transition The thermally induced release is close to the phase transition T m to the coexistence of both gel and fluid producing grain boundaries that have an increased permeability of hydrophilic molecules. The liposomal release of the MIC of IK8 showed a similar level of bactericidal activity as the gel containing the MIC of free IK8 against both bacterial types, ~5.5- and ~6.7-log reductions in CFU ml -1 against S. aureus and P. aeruginosa respectively. This indicates that the liposomal encapsulation and subsequent triggered release did not affect the peptide’s antimicrobial efficacy. When taken together with the liposomal protease protection, these results indicate that this therapeutic gel offers a potential means of maintaining AMPs in a protease rich wound environment until required, with no decrease in bactericidal efficacy upon triggered release. By utilizing a photothermal trigger, we are also able to enhance the bactericidal activity of the therapeutic gel by heating the bacteria to higher temperatures. Enhanced bacterial kill was observed when treating both bacteria with IK8 and laser irradiation at 2.4 W cm -2 (60°C) for 10 min, leading to a 7.8-log reduction in CFU ml -1 of S. aureus and complete bacteria killing of P. aeruginosa . This is similar to previous observations demonstrating thermal enhancement of antibiotics against planktonic S. aureus and P. aeruginosa Furthermore, s imilar studies of the effects of for 10 mins on no induced to peripheral upon (heating to 60°C), at which point there was a 2.6-log decrease in S. aureus and a 4.4-log decrease in P. aeruginosa . Further increase in the laser intensity to 2.8 W cm -2 (heating to 65°C) showed complete antimicrobial killing of both bacterial species in all samples containing the NRs. All bactericidal activity was attributed to the photothermal heating, with no bacteria death observed under irradiation at 2.8 W cm -2 in the absence of AuNRs. These results show that the IK8 liposome and AuNR loaded hydrogel has the potential for use as a broad-spectrum antimicrobial treatment that, through the regulation of the applied laser intensity, can not only trigger the release of AMPs, but amplify their antimicrobial effects. By harnessing the photothermal enhancement, the chances of providing non-lethal treatment are reduced and as such this treatment should provide a decreased likelihood of bacterial AMP resistance.


Introduction
Societal misuse and overconsumption of antibiotics has led to the emergence of multidrugresistant bacteria that are set to pose serious global medical, social and economic issues. 1 Such is the threat that bacteria have evolved resistance to "last resort" antibiotics such as carbapenem, leaving critically few options for treatment. [2][3] Naturally derived and synthetic antimicrobial peptides (AMPs) have been identified as a promising antibiotic alternative. AMPs act through the 0 W cm -2 1.8 W cm -2 2.1 W cm -2 2.4 W cm -2 2.8 W cm -2 Laser intensity physical disruption of the bacterial membrane, promoting cytoplasmic leakage, and resulting in rapid bacteria death. Further, the high metabolic cost associated with membrane repair reduces the risk of the emergence of drug resistance. [4][5][6] However, the clinical translation of AMPs has suffered from the peptide instability and susceptibility to protease degradation in vivo. [7][8] As such, the development of protective delivery systems is seen as a key strategy for increasing the chances of AMP implementation. [9][10][11] However, such systems have primarily focused on the sustained release of AMPs, neglecting stimuli-responsive alternatives. [12][13][14][15] Recently, systems utilizing the triggered delivery of antibiotics have drawn widespread attention due to their capacity for spatially and temporally controlled delivery of lethal antibiotic dosages. [16][17][18][19] The ability to promote site-specific delivery in response to environmental cues also minimizes systemic loss into peripheral tissues ensuring that local microflora are not subjected to sub-lethal doses that may induce resistance. 4 With the increased treatment efficacy and reduced risk of resistance development, we believe that the implementation of triggered delivery of AMPs could provide the next step toward clinical translation.
In recent years, hydrogels have become increasingly studied as drug delivery vehicles due to the diversity of gel characteristics (e.g. crosslinking density, mesh size, charge, polymer density) that can be controlled to tailor the retention rate of impregnated materials. [20][21] Furthermore, the doping of hydrogels with stimuli-responsive materials has become increasingly popular due to the potential for increasing the accuracy of the location and rate of drug release. [22][23][24] Traditionally, the fabrication of stimuli-responsive materials has utilized the chemical modification of existing biomaterials, however synthesis of these materials can be complex and difficult to scale-up, leading to increased cost of production and premature leakage of cargo with decreased functionality. [25][26] Such issues can in principle be circumvented by the incorporation of optically active nanoparticles that, under near-infrared irradiation (NIR), exhibit photothermal heating that can be used to trigger drug release, [27][28][29][30] as well as provide thermally induced antimicrobial effects. [31][32][33][34] The combination of these treatments has been demonstrated to produce synergistic antimicrobial effects. Meeker et al. (2016) developed a daptomycin encapsulating gold nanocage with conjugated staphylococcal targeting proteins, that provided triggered antibiotic and photothermal co-therapy that demonstrated complete bacteria killing of both planktonic suspensions and biofilm models of S. aureus. 35 Treatments utilizing the same system without laser irradiation or in the absence of the daptomycin, did not provide complete bacteria eradication.
Liposomes are the most widely researched nanoscale antibiotic delivery system, [36][37] primarily due to their ability to increase the biocompatibility, bioavailability and safety profiles of encapsulated antimicrobial materials. 38 Conventional liposomes predominantly consist of phospholipids, providing an inherent temperature dependent gel-liquid phase transition that increases the membrane permeability. As such liposomal photothermal delivery has been widely demonstrated using dye molecules and anti-cancer chemotherapeutics, 28 Here, we describe a hydrogel/liposome system for the photothermal triggered release of antimicrobial agents (Figure 1), in which a poly(ethylene glycol) (PEG) based gel containing AMP-loaded liposomes and lipid coated NRs is used for the controlled treatment of bacteria. We have previously shown that the β-sheet forming AMP, IK8, possess potent broad-spectrum antimicrobial activity against various antibiotic-susceptible and antibiotic-resistant microorganisms. 43 However, IK8 was found to be rapidly degraded by proteases such as trypsin and proteinase K, leading to a significant loss in antimicrobial activity. 44   and 24 h. 10 µl of the liposomes were extracted and 10 % v/v of DMSO was added to induce peptide leakage before the sample was diluted to 300 µl. The solution was then run through an Agilent 1290 series high-performance liquid chromatography (HPLC) system with a 4.6 x 250 mm Insertil ODS-SP column, through which the sample was flowed with an acetonitrile gradient (acetonitrile/water ratio at 0.01 mins 13:87, at 10 mins 100/0) at a rate of 2 µl min -1 . A peak was observed at 220 nm after 2.56 mins corresponding to the IK8. The concentration of IK8 encapsulated was determined by integrating the area beneath the peak and comparing this to a predetermined concentration curve ( Figure S1).

Thermal stability
The thermal release of the liposomes was investigated through the release of the self-quenching dye calcein. Liposomes were fabricated through the hydration of lipid films (DSPC/cholesterol/DSPE-mPEG2k, 65/30/5 mol% respectively) using a solution of 0.1 M calcein dissolved in PBS. The resulting liposomes were homogenized using heated-extrusion before unencapsulated dye was removed by passing the liposome suspension through a Sephadex® G-50 gel chromatography column. The first 0.5 ml of sample to pass through the column was collected. The sample was then diluted 2000 times before 200 µl was added to a 96well plate, for measurement of the sample fluorescence (Exc/Em 496/515 nm). Thermocouples were then added to the wells and the plate was placed in an incubator preheated to 55°C. Once the temperature of the solution had been maintained at 55°C for 5 min the fluorescence was again measured. 1% Triton-X100 was added to the control wells to find the maximum fluorescence, and data normalization.

Liposome leakage
To assess the liposome leakage, 100 µl of calcein loaded liposomes diluted 1000 times was added to a 96-well plate along with 100 µl of MHB II, or MHB II S. aureus suspension (initial concentration 1 x 10 6 colony forming units (CFU) ml -1 ) and incubated at 37°C. The sample fluorescence was measured immediately after preparation and once daily to identify passive leakage of calcein. 1% Triton-X100 was added to the control wells at day 0 and to all wells at the end of the test to identify to determine maximum fluorescence.

Colloidal stability of IK8 Liposomes
Size and polydispersity measurements were performed in triplicate using dynamic light scattering (DLS, Zetasizer Nano ZS) at 25°C. IK8-liposomes diluted 100 times in PBS were sized using a 4mW, 633nm laser with a measurement angle of 173°. In between measurements the liposomes were stored at 4°C.

Liposomal protection of encapsulated AMP from protease degradation
The efficacy of liposomal encapsulation in protecting the peptide against potential proteolytic degradation was tested by adding 10 µl of the proteolytic enzyme trypsin (50 µg ml -1 ) to 1 ml of 0.1 mg ml -1 free IK8 and IK8-encapsulated liposomes containing the equivalent quantity of peptide. After an hour 150 µl of each sample was removed and the trypsin was inactivated through the addition of 10 µl of 40% w/w trichloroacetic acid. 45 This process was repeated each hour for 5 h. The concentration of IK8 remaining in the samples was quantified using HPLC, liposomally encapsulated IK8 was released by adding 5% v/v DMSO before being added to the HPLC.

Lipid-coating of AuNRs
Binary surfactant centrimonium bromide (CTAB) and sodium oleate (NaOL) stabilized AuNRs were synthesized as described by Roach et al. (2018), these were then lipid coated to increase colloidal stability. The lipid replacement of the CTAB-NaOL bilayer was performed by initially forming a 10 ml suspension of liposomes via tip sonicating DSPC/DSPE-PEG2000 (95/5 mol%, 10 mg ml -1 ) for 2 h before addition to a 10 ml AuNR solution containing 60 µg ml -1 of Au. The liposome-NR suspension was then sonicated for 24 h before pelleting using centrifugation (9,000 x g for 30 minutes), removal of the supernatant and re-dispersal in a fresh liposome solution.
This process was repeated three times, with the AuNR pellet dispersed into deionized water following the final centrifugation. The dimensions of the AuNRs were determined through transmission electron microscopy (TEM) imaging using a Tecnai G2 Spirit TWIN/BioTWIN with an acceleration voltage of 120 kV. Characterization of the AuNR absorption was measured using a PerkinElmer Lambda 35 spectrophotometer and the Au density was determined using atomic absorption spectroscopy Varian 240 fs ( Figure S2d).

Fabrication of a hydrogel containing IK8-loaded liposomes and AuNRs
Briefly, 4APM was dissolved in 20 µl of 10mM sodium citrate buffer (pH = 6), before addition of lipid coated AuNRs (48 µg ml -1 of gold) and liposomes encapsulating 32 µg ml -1 of peptide, the suspension was topped up to 40 µl using the citrate buffer. The PEGSH was dissolved into 10 µl of 10 mM citrate buffer and added to a well in a 96-well plate. The 4APM solution, containing the AuNRs and IK8-liposomes, was rapidly added to the PEGSH solution and vortexed to ensure thorough mixing.

Liposome and AuNR retention within the gel
The retention of liposomes and AuNRs within the hydrogel was determined by incubating the gels with buffer and measuring the leakage of the particles daily. 50 µl PEG gels containing both liposomes and AuNRs were fabricated into opaque 96-well plates and 150 µl of citrate buffer was added to the gels after gelation. To quantify the amount of AuNRs lost from the gel the supernatant absorbance at 860 nm, the longitudinal absorbance wavelength of the AuNRs, was measured. The liposome retention was assessed by including the dye Texas Red within the liposome bilayer, thus providing them with a measurable fluorescence (EX/EM, 561 nm/594 nm). The Texas Red was included by the addition of 0.5 wt% Texas Red to the lipid mixture, prior to the formation of the thin lipid film, the rest of the fabrication protocol was unchanged.
The absorbance and fluorescence of the citrate buffer was measured before addition to the gels, and the absorbance and fluorescent values were obtained for the gels on day 0. After 24 hours the supernatant was added to a clean well and the absorbance and fluorescence were both measured, after which the solutions were added back to the gels. The liposome and AuNR release was determined using the equation where X is the fluorescence when of the liposomes or the absorbance of the AuNRs (860 nm).

Rheometry
The mechanical properties of the IK8-liposome and AuNR loaded gels was assessed using an Anton Paar modular compact rheometer 302. The storage and loss moduli (G' and G'' respectively) of the gels was obtained through applying a constant strain of 1% to the gel and by applying a 1% strain at frequencies of 0.1-100 Hz. A 500 µl gel was fabricated on the bottom plate of the rheometer that was preheated to 37°C. Immediately after gelation the 25 mm diameter top plate was lowered to 1 mm above the bottom plate before initiating the measurement. Silicon oil was applied to the periphery of the rheometer plates to restrict evaporation from the hydrogel.

Hydrogel swelling ratios
The swelling ratio is the fractional increase in the weight of the hydrogel due to the absorption of water. PEG gels of 2.5, 5 and 10 wt% were fabricated into the eppendorfs, before freeze-drying.
The gels were then weighed before 1 ml of Milli-Q was added. After 10, 30, 60, 120, 180 and 240 mins the excess Milli-Q was removed and the gel was weighed again. The swelling ratio was then calculated using: Where W 1 is the weight of the swollen gel and W 2 is the initial weight of the gel before hydration.

Cytotoxicity testing
To determine the cytotoxicity of the hydrogel containing IK8-loaded liposomes and AuNRs, 100 µl of fibroblast growth media containing 5 × 10 3 human dermal fibroblast cells was added to wells of a 96-well plate before incubation at 37°C for 24 h. The fibroblast growth media was replaced with the 90 µl of fresh media and a 10 µl 5 wt% PEG hydrogel containing liposomes encapsulating 32 µg ml -1 of IK8 and 48 µg ml -1 of AuNRs, fabricated as previously described (Section 2.6), and incubated for a further 24 h The media was then replaced with fresh media containing 10% v/v of the cell viability reagent WST-1, and incubated for a further 2 h. An equal volume of the media containing 10% v/v WST-1 was added to wells containing no cells to account for the background absorbance by the WST-1 reagent. The media was transferred to wells that did not contain cells and the absorbance was measured at 440 and 660 nm. The relative viability was calculated using the equation given in Figure 2.  CFU ml -1 ) before irradiation for 10 min using the same laser intensities in both treatment events.
The samples were then incubated for 18 h before being plated onto agar plates for determination of colony counts as described in section 2.8.

Statistical analysis
All statistical analysis was performed using two-tailed student's t-testing. Results where considered as statistically significant when P < 0.05, all significant results are denoted with asterisks with the probability range denoted in the corresponding figure caption.

Liposomal formulation and IK8 release
To determine the efficacy of the liposomal AMP reservoirs, the different key aspects of thermal,  against S. aureus (32 µg ml -1 , Figure S4), whilst providing low sample-to-sample variability, compared to film-hydration using higher IK8 concentrations. The mass of encapsulated IK8 increased linearly with the concentration of IK8 used to hydrate the lipid film, however, the proportion of the initial IK8 added to the lipid film that is encapsulated decreases causing a drop in the encapsulation efficiency from 24 ± 4% to 13 ± 2%. The resulting liposomes showed excellent colloidal stability with little change in size upon incubation in PBS at 4°C, over a three- week observation period (Figure 4b). Post fabrication, the liposome diameter was 365 ± 36 nm with a coefficient of variation (CV) of 0.12 ± 0.05, whereas after 3 weeks the liposome diameter increased slightly to 385 ± 42 nm (CV = 0.17 ± 0.04). The final stage of testing was to assess the efficacy of encapsulation against enzymatic degradation. Despite the great potential of AMPs as antibiotic alternatives, clinical translation has been hindered by proteolytic instability in vivo. 49 By loading AMPs within delivery vehicles, it is possible to protect the AMP from protease degradation before reaching the infection site, greatly enhancing treatment efficacy. 50 To test the liposomal protection of encapsulated IK8 against degradation we incubated free and liposomal IK8 in a trypsin solution for 5 h. Our data a) b) shows that 81 ± 6% of liposome encapsulated IK8 remained intact, over double that of the free IK8 in solution, 36 ± 2% ( Figure 5). The lipid bilayer encapsulating the IK8 therefore provides a protective barrier that restricts access of proteases, thus potentially allowing treatment infection environments that are often rich in such proteolytic enzymes, such as an open wound. 51

Hydrogel formation
IK8-liposome / AuNR loaded hydrogels were prepared by mixing the nanoparticles with 4APM as the monomer, before initiating gelation using a PEGSH solution. This gelation occurs through a Michael-type reaction between the maleimide and thiol groups of the two PEG molecules, providing a gel with tunable gelation rates and mechanical properties. [52][53][54] The PEG hydrogel was also chosen due to its high permeability by small hydrophilic molecules, meaning the gel should provide minimal interference with the AMP release kinetics. 55 By controlling the pH and the buffer concentration the gelation time could be varied up to two minutes. A 5 wt% gel with a gelation time of 20 seconds was deemed appropriate, as this produced a stiff gel and allowed for a reasonable mixing time. Gels greater than 5 wt% were stiff and difficult to mix, whereas the 2.5 and 1 wt% gels were very ductile almost fluidic, as shown by the small discrepancy between the storage and loss moduli ( Figure S5a). The 5 wt% gel also was also advantageous in terms of it's swelling in an aqueous environment. The swelling ratio of the gels was found to decrease as the wt% increased, after 4 hours the 2. aggregation of the AuNRs under irradiation, at these laser intensities. Subsequently, 64 ± 6% of liposome encapsulated calcein was release when the liposomes were loaded into a 5 wt% gel and irradiated at 2.4 W cm -2 , heating the sample to 55°C, an increase upon the calcein released from liposomes free in suspension, 56 ± 2% (Figure 6d). This increase may be due to the confinement of the liposomes and AuNRs within the 50 µl volume of the gel, rather than when they are free in solution and are able to diffuse throughout the 200 µl total volume of liquid. This would increase the likelihood of the liposomes and thermally radiating AuNRs being in a close proximity to one another making the liposomes more susceptible to thermal destabilization.

Cytotoxicity of the therapeutic gel
Considering the cytotoxicity of any dressing used on a wound susceptible to infection is of vital importance. Hindering the regenerative process through increasing the toxicity to mammalian cells leaves the wound vulnerable to infection and increases the risks of a wound becoming chronic. 57 As such, the cytocompatibility of the therapeutic gel was assessed using HDF cells; which play a critical role in the formation of granulation tissue. 58    The ability to utilize the hydrogel as a drug depot to provide repeated triggered release of antimicrobial agents at lethal dosages and in combination with photothermal ablation of bacteria would provide significant advantages in wound management. To demonstrate this, gels were fabricated with AuNRs and an increased number of liposomes such that they contained 2.5 times the MIC of IK8. The gels were inoculated with the first round of S. aureus before irradiation at 2.1 or 2.4 W cm -2 for 5 mins, resulting in release of 43 ± 8 % of encapsulated materials ( Figure   S9). The bacteria were then incubated with the gel for 18 h, before removal and quantification of CFU

Conclusion
We have demonstrated a AuNR / AMP liposome loaded hydrogel, that can effectively treat pathogenic bacteria through the photothermal stimulated release of AMPs, which is, to our knowledge, the first instance of triggered release of AMPs in response to exogenous stimuli.
Irradiation at 2.1 W cm -2 for a period of 10 min caused heating to 55°C, triggering the release of

Corresponding Author
* Email: s.d.evans@leeds.ac.uk; z.y.ong@leeds.ac.uk Present Addresses †If an author's address is different than the one given in the affiliation line, this information may be included here.

Author Contributions
The manuscript was written with contributions from all authors. All authors have given approval to the final version of the manuscript.

Notes
The authors declare no competing financial interest.

ACKNOWLEDGMENTS
The authors would like to thank the following for financial support: EPSRC (1819417)