Bacteria Live Dead Assay
We performed an additional test after antibacterial assay with gelatin hydrogels to evaluate the amount of bacteria alive or dead on the surface of samples. We used the BacLight Redox Sensor CTC Vitality Kit (Molecular Probes). SYTO 24 green-fluorescent nucleic acid staining is used at 0.001 mM to counting all bacteria and CTC red staining is used at 50 mM to detect bacteria alive.
3. Results
3.1 Synthesis and characterization of PDA-PAR30 NPs
In this study, polydopamine nanoparticles (PDA NPs) were synthesized using a protocol developed by Chassepot et al.(Chassepot & Ball, 2014) In this work, the ability of proteins (either negatively or positively charged) to stabilize and control the size of dopamine aggregates formed in dopamine solutions upon oxidation in order to get functional nanoparticles was demonstrated. In the current work, we have used a chain-length controlled polycation, polyarginine with 30 residues (PAR30), known to have a strong antimicrobial activity.(Mutschler et al., 2016) Then we have synthesized PDA-PAR30 nanoparticles by adding a PAR30 solution (1 mg.mL-1) in contact with different concentrations of dopamine hydrochloride solution ranging from 0.2mg.mL-1 to 0.5 mg.mL-1 in alkaline Tris Buffer (pH=8.5) for 24 hours under stirring (Scheme 1). We first studied the influence of the dopamine concentration on the nanoparticle stability. After 2 months in a closed vial, only the formulation synthesized with 0.2mg/mL of dopamine hydrochloride (PDA-PAR30 0.2) was still stable without apparent sedimentation (Figure 1A). For the other formulations containing 0.5 and 0.3 mg.mL-1 of dopamine hydrochloride (noted as PDA-PAR30 0.5 and PDA-PAR30 0.3) a total sedimentation was observed after few days. The size of the aggregates was analysed using dynamic light scattering (Figure S1). Aggregates at the microscale were found for both formulations. The size of these aggregates was estimated at about 2 µm and 4 µm for the PDA-PAR30 0.5 and PDA-PAR30 0.3 formulations respectively. For the stable formulation, PDA-PAR30 0.2, Transmission Electron Microscope (TEM) analysis was performed to determine the size of the resulting nanoparticles (Figure 1B). Nanoparticles with a narrow size distribution of 3.9 ± 1.7 nm (mean diameter based on the measurement of 300 nanoparticles) were obtained. As a control, we also synthesized PDA NPs using the same protocol but without any decoration but these particles were not stable and formed aggregates between 1-10µm (Figure S1). The only way to obtain stable PDA based NPs was to use polypeptides or proteins to stabilize them. Thus, for the rest of the experimental work, we used the stable nanoparticles obtained with a solution of 0.2 mg. mL-1 of dopamine hydrochloride (PDA-PAR30 0.2 condition). To further probe and demonstrate the decoration of polydopamine NPs with chain-length controlled polyarginine, zeta potential measurements were performed. It is known that polydopamine is negatively charged above pH 5. Indeed, the measurement on PDA NPs gave a zeta potential value of -22.7mV (Figure 1C). By adding PAR30 to the PDA particles, we obtained zeta potential value of +30.3 mV. The surface charge of the NPs has switched from negative to positive values as PAR30 is a polycation. This demonstrates that the polycation was successfully deposited on the surface of PDA NPs.
In the next step, we have investigated the antimicrobial properties of PDA-PAR30 0.2 NPs to check if the PAR chains maintained their biological activity once immobilized at the surface of PDA NPs. Indeed, our previous studies demonstrated the need for the mobility of PAR molecules in the coatings to exert their antimicrobial activity and this mobility in the particles is not guaranteed.(Mutschler et al., 2016) We have tested different NPs dilutions ranging from 1:10 to 1:100 from the initial particle concentration and incubated them with S. aureus ( 8.105 CFU.mL-1) for 24 hours. As a control, we also synthesized unmodified PDA NPs and used them in the same range of dilution against S. aureus . The results are presented in Figure 2A. PDA-PAR30 0.2 NPs ranging from 1:10 to 1:50 in dilution completely inhibit the bacterial growth (100% growth inhibition) and then from the dilution 1 :100, we can again observe a significant bacterial growth (40% of growth inhibition for the dilution 1:100). Based on these results, we can deduce that the MIC (Minimum inhibitory concentration) of these PDA-PAR30 0.2 NPs is between 1:50 and 1:100. As it was difficult to estimate the true concentration of NPs, based on the absence of a precise knowledge about the dopamine oxidation-oligomerization reaction yield, we have performed this experiment on different batches of NPs coming from different synthesis done in identical conditions. The MIC was always estimated between 1:50 and 1:100 (Figure S2) which means that the synthesis is reproducible, and we can assume that we always obtain comparable amount of NPs of similar size in each synthesis. As a control, experiments performed on PDA NPs without the decoration of PAR30 show absence of antimicrobial activity whatever the dilution used. This means that the antimicrobial activity of the NPs can be attributed to the presence of PAR30 and that PAR30 chains maintain their biological activity after immobilization on PDA NPs.
After the evidence of the antimicrobial properties of the NPs, it was necessary to check their cytotoxicity. For this purpose, we used Balb 3T3 cells as a well-established model (mouse fibroblasts). PDA-PAR30 0.2 NPs at different dilutions in cell culture medium were put in contact with an almost confluent layer of 3T3 cells for 24 hours and the viability of these cells was then estimated with a MTT assay and compared to a control (normal medium without NPs). A material is defined as cytotoxic if a decrease of at least 30% in viability compared to the control is observed according to ISO/EN 109935. The obtained results are presented in Figure 2B and images of the cell morphology are given in Figure S3. As a control, we also investigated the cytotoxicity of Tris Buffer at a dilution 1 :10 which represents the highest concentration of Tris used in the experiments to dilute NPs in the cell culture medium. At this concentration of Tris Buffer, it was shown that the buffer was not cytotoxic. This indicates that for the rest of the experiment, if a cytotoxic effect is monitored, it can only be attributed to the presence of PDA-PAR nanoparticles. Finally, it was demonstrated that for PDA-PAR30 0.2 NPs, only the 1:10 dilution exhibits a cytotoxic behaviour with a decrease in viability of about 90%. Hence, we can conclude that PDA-PAR30 0.2 NPs particles are non-cytotoxic at dilutions higher than 1:10.
3.2 Design and characterization of antimicrobial Gelatin hydrogels loaded with PDA-PAR30 NPs (Gel-PDA-PAR30)
We have previously demonstrated that decoration of PDA NPs with PAR30 allows to confer antimicrobial properties. In the second part of the study, we have incorporated these PDA decorated NPs into gelatin hydrogels to develop an antimicrobial hydrogel that can be further used as an implant surface coating to fight infection after implantation while providing a degradable environment for remodelling. For this purpose, PDA-PAR30 0.2 NPs were incorporated into gelatin type A hydrogel without any dilution (Gel-PDA-PAR30 and the hydrogels were then crosslinked with microbial-Transglutaminase in order to reinforce their mechanical properties and stability. This hydrogel Gel-PDA-PAR301:1 is depicted in Figure 3A and compared to gelatin hydrogel without NPs (Gel). We observed a brown color for the hydrogel loaded with NPs which is the characteristic colour of PDA. The presence of NPs at the surface of the gelatin hydrogel was checked using SEM and it was found that the particles were homogeneously distributed. However, we can notice that once incorporated in the hydrogel, there is an increase of particle size from few nanometers (as found for solution dispersed NPs, Figure 1B) to almost 100 nm. This can be explained by the aggregation of the NPs during the hydrogel formation (Figure 3B). We then investigated the swelling properties of the composite hydrogels and compared them to pure gelatin hydrogels (Figure 3C). It has been found that the addition of NPs in the gelatin network had an effect on the swelling behaviour: a significant increase with a value of 500% was monitored, compared to pure gelatin hydrogel where a value of only about 300% was measured. This could be explained by Donnan effect due to the high positive charge density of the PAR covered NPs inducing a strong incorporation of counteranions and hence of hydration water.(Rathna, Mohan Rao, & Chatterji, 1996) Then, the release of PDA-PAR30 0.2 NPs from the gelatin hydrogel was studied using fluorescently labelled NPs with PAR30-FITC (Figure 3D). We can observe a burst release of the NPs during the first hours with most of the NPs released during the first 3 days. However, the hydrogel was still brown after 3 days of release (data not shown), which means that some of the NPs were still present in the gelatin network even if no further release was recorded.
Next, we investigated the rheological properties of the gelatin-PDA-PAR30 hydrogel composite to determine the influence of the addition of PDA-PAR30 NPs on the mechanical properties. We first analysed the shear viscosity in solution of the two formulations (Gel and Gel-PDA-PAR30) without the crosslinking agent, microbial Transglutaminase (Figure 4A). The shear viscosity of both formulations remained constant over a large range of shear rates (from 0.1 to 1000 s-1). Moreover, the addition of nanoparticles in the formulation did not affect the shear viscosity since the values of both Gel and Gel-PDA-PAR30 formulations were similar with values ranging from 0.1 to 0.06 Pa.s depending on the shear rate applied. After demonstrating that the addition of NPs in the formulation did not affect their viscosity, we crosslinked both formulations with microbial-transglutaminase and studied the viscoelastic properties of the resulting hydrogels. The two hydrogels were crosslinked for at least 16 hours in situ in the rheometer to ensure complete crosslinking of the network. Then we recorded the storage modulus of the two hydrogels as a function of frequency (Figure 4B). Over the frequency range tested (0.1 - 50 Hz), it was found that the storage modulus (G’) of the two formulations increased slightly as a function of frequency. In addition, we observed a higher storage modulus in the case of Gel-PDA-PAR30 hydrogel composite with values between 8 to almost 11 kPa, whereas for the gelatin hydrogel the values were between 5 and 7 kPa. Thus, the addition of NPs within PDA-PARhydrogel enhanced its viscoelastic properties. We can assume that the NPs act as a filler in the structure of the hydrogel.
After characterization of the mechanical properties of the hydrogels, we performed a stability study in different conditions (Figure 4C, D, E): enzymatic stability in the presence of collagenase, stability in acidic conditions (pH=5) as a model of infection and stability in physiological pH at 7.4 in PBS. All these experiments were conducted at 37°C and the mass of hydrogels were recorded as a function of time. For enzymatic stability, hydrogels were incubated in 1 mL of collagenase solution (1 mg.mL-1) at 37°C. In Figure 4C, it is shown that the two formulations, Gel or Gel-PDA-PAR30, were completely degraded after 2 hours with the same kinetics. For the stability at pH=5, hydrogels were incubated in 1mL of citric buffer and the experiment was performed for 7 days with an estimation of mass loss at days 1, 4 and 7 (Figure 4D). After the first day, the hydrogel containing PDA-PAR30NPs showed better stability (86% of remaining hydrogel versus 69 % for the pure gelatin hydrogel). But during the next days, the degradation was indistinguishable for both formulations. After 7 days, the difference between the two hydrogels was not significant (64% degradation for Gel-PDA-PAR30 vs 59% for Gel). The last stability experiment was also conducted for 7 days in PBS buffer at pH=7.4 (Figure 4E) and the results showed the same degradation profile between the two formulations with values of remaining hydrogel after 7 days of about 55% (56% for Gel-PDA-PAR30 vs 53% for Gel). With this set of experiments, we can conclude that the addition of PDA-PAR30 NPs in the gelatin structure only increases the viscoelastic properties of the hydrogel (increase of G’, Figure 4B) but did not change the stability of the resulting hydrogels compared to pure gelatin hydrogels.
Once the hydrogel composites were characterized, their antimicrobial activity starting from a dilution 1:1 (no dilution) and until 1 :50 was tested against S. aureus (Figure 5A). Hydrogels were prepared in PBS but as the NPs were prepared in Tris Buffer, the final hydrogel composite was prepared in a PBS/Tris mix. It was found that Gel-PDA-PAR30 hydrogels exhibit complete antimicrobial activity with a 100% growth inhibition after 24 hours in contact withS. aureus until a dilution of 1:4. For higher dilutions (1:10 and 1:50), the antimicrobial activity was lost because the NPs were probably too diluted in the hydrogel to maintain a biological activity. SEM analyses were also performed to compare the behaviour of Gel-PDA-PAR30 (1:4) vs Gel (Figure 5B). After 24 hours of contact with S. aureus , the bacteria fully colonized gelatin hydrogel and started to form a biofilm whereas with the NPs loaded hydrogel almost no bacteria were found. Zooming on the Gel-PDA-PAR30 surface allows to show the presence of particles at the surface of the hydrogel which confer the antimicrobial activity.
These results are also in agreement with the metabolic activity of bacteria, using CTC/Syto 24 staining where no viable bacteria were found on the hydrogel loaded with NPs (Figure S4). For final applications, the gel formulation with NPs diluted in a ration 1:4 seems to be optimal. Then, the cytotoxicity of the formulation Gel-PDA-PAR30. has been tested with Balb 3T3 cells using extraction method following ISO Standard 10993-5 recommendations. As the hydrogel composite was prepared in a PBS/Tris mix, we have first tested the cytotoxicity of pure gelatin hydrogels prepared in either Tris or PBS and then the cytotoxicity of hydrogel prepared in different ratios of PBS/Tris (Figure S5). It was found that none of the configurations tested exhibited cytotoxicity with viability up to 70% maintained in all cases. Then, we prepared Gel-PDA-PAR30 hydrogel composites starting from a dilution 1 :4 and until 1:50 (NPs were also diluted from 1:4 to 1:50 in the hydrogel) and studied their cytotoxicity. None of the formulations tested were cytotoxic and a viability up to 85% was maintained (Figure 6). The resulting cell morphology is given in Figure S6 and there was no evidence of anormal cell shape. The formulation 1:4 was tested towards Balb 3T3 cells and its viability was estimated to be around 90%. Hence we can conclude that this formulation was not cytotoxic. Thus, we have designed a powerful antimicrobial hydrogel ony based on natural and bioderived materials, i.e gelatin, polydopamine and polyarginine. An adequate formulation has been obtained both in term of antimicrobial activity with 100% bacteria growth inhibition and in term of biocompatibility.
4. Discussion
4.1 Polyarginine decorated PDA NPs are stable and antimicrobial
Polydopamine NPs-based chemistry is versatile and should open the road to develop bioactive biomaterials with the possibility to decorate nanoparticles with multiple bioactive agents enabling us the design of hydrogel composites with bioactive properties. As these particles have a wide range of inherent properties as a potential substrate for photothermal effect or as quenchers and catalysts, they can be used in many biomedical applications including cancer therapy and stimuli responsive drug delivery. Adding a layer of bioactive agents on this structure creates a tool that is highly modular and can have applications in incorporating theranostic delivery and biosensing for tissue engineering and regenerative medicine applications. In this study, our main goal was to create stable polydopamine based nanoparticles with antimicrobial properties and to do that we have used a known polypeptide antibiotics (polyarginine).
In our study, we have first demonstrated that the addition of a sufficient amount of PAR30 to a sufficiently diluted solution of dopamine, 0.2 mg.mL-1, in Tris buffer allows to produce small PDA-PAR and stable nanoparticles having an average diameter of 3.9 ± 1.7 nm and displaying a positive zeta potential of about +30.3 mV compared to pristine PDA nanoparticles having a zeta potential close to -23 mV in the same conditions (Figure 1D).
Polyarginine, a positively charged polymer, has been originally used as a highly efficient cell penetrating peptide for gene delivery applications. Beyond this cell penetrating properties, due to its polycationic nature, polyarginine is a widely used component to build polyelectrolyte multilayers. Recently, it has also been shown that PAR18 and also its D-enantiomer has neuroprotective properties.(Meloni et al., 2019) As mentioned before, polyarginine also has antimicrobial activity whose efficacy is related to the mobility of PAR chains within polyelectrolyte structures.(Dong et al., 2016) In solution, a wide range of polyarginine chain lengths have shown antimicrobial activity; however in film configurations where polyarginine was in interaction with polyanions (heparin, chondroitin sulfate, alginate, hyaluronic acid) only a narrow range of chain length and only films with HA were active. Thus, decoration of polyarginine on PDA particles might have hindered the antimicrobial activity of polyarginine due to the lack of mobility.
However, our experiments demonstrated that polyarginine chains on PDA surfaces sustain their antimicrobial capacity. Indeed, the PDA-PAR30 0.2 NPs display an antimicrobial activity against S. aureus up to a dilution of about 1:50 to 1:100 (Figure 2A) whereas the PDA NPs are not antimicrobial whatever their dilution; this demonstrates that the antimicrobial activity arises from the presence of PAR on its surface and also that the particle formation does not hinder the antimicrobial activity of PAR, excepting the cases of thin layer-by-layer coatings where the antimicrobial properties depend on the polyanion. This is due to the fact that, even though immobilised, PAR chains are at the surface of the particles thus they are free to electrostatically interact with bacterial membranes even though the interaction is partially shared with the interaction with the PDA surface. Our experimental results clearly demonstrated that such interactions cannot overshadow the antimicrobial properties of polyarginine.
4.2 PDA-PAR NPs incorporated hydrogels have improved viscoelastic properties and antimicrobial capacity
The need for antimicrobial coatings for non-degradable implants has been long recognized such with systems such as TYRXTMantimicrobial, degradable polymer pouch used for pacemakers in clinical applications. The surface of non-degradable implants, can create an immune-privileged zone that can facilitate the bacterial attachment and the biofilm formation. For example, the rate of infections with fixation devices were estimated between 2-5% The replacement of passive implants with engineered artificial tissues can in one way alleviate this problem as the porous nature of such scaffolds will enable the angiogenesis. Consequently, the implanted structure will integrate with the host immune system rather than being an isolated surface. But, on the other hand, materials such as natural polymers (as for instance collagen, gelatin, hyaluronic acid etc.,with certain exceptions like chitosan) used for scaffolds generation can also boost the attachment and proliferation of bacteria just as well that of the host tissue. Moreover, potential use of engineered tissues in non-sterile wounds would also necessitate an antimicrobial component. Previously, such precautions has been tried to be put in place by either doping hydroxyapatite with silver or copper to confer antimicrobial activity to it(Sahithi et al., 2010; Stanić et al., 2011) or incorporation of silver based nanoparticles in the scaffold formulations. (Saravanan et al., 2011) Specifically, with gelatin foams crosslinked with genipin, Yazdimamaghani et al. demonstrated antimicrobial activity againstE. coli and S. aureus by in situ formed silver nanoparticles.(Yazdimamaghani et al., 2014) However, it has been shown that silver nanoparticles are less cytotoxic compared to silver sulfadazine which at high concentrations can demonstrate hepatotoxicity in clinical settings.(Trop et al., 2006) Moreover, at the cellular level, it has been shown that silver exhibits mild cytotoxicity against mesenchymal stem cells, which are one of the major cell sources in tissue engineering.(Samberg, Loboa, Oldenburg, & Monteiro-Riviere, 2012) Thus, alternative antimicrobial solution would be beneficial, particularly in formats that would also contribute to the mechanical properties of the overall scaffold in tissue engineering. The nanoparticles developed in this study could potentially achieve these two effects concomitantly in the context of hydrogels.
In this study, to demonstrate such utility of PDA-PAR NPs, we have used gelatin hydrogels, which creates a highly amenable surface for bacterial attachment as demonstrated by our results. The incorporation of the particles in the gel structure resulted in an increase in G’ values which indicates a slightly stiffer hydrogel due to the nanoparticle based reinforcement. The incorporation of the particles in the gel structures could decrease their mobility and as a result their antimicrobial capacity; however, experiments showed that it was possible to maintain antimicrobial activity of the NPs at dilutions below the cytotoxic limits, the reason for this observation being the ability of the hydrogels to release the NPs as demonstrated by the release experiments. The relatively controlled manner of release of the NPs also points out using the hydrogels as a potential delivery system for such NPs.