45S5 Bioglass® concentrations modulate the release of vancomycin hydrochloride from gelatin–starch films: evaluation of antibacterial and cytotoxic effects

The aim of this work was to evaluate the release profile of vancomycin hydrochloride (VC), as well as the degradation, in vitro antistaphylococcal effect and cytotoxicity in MG-63 osteoblast-like cells of gelatin–starch (GS) films added with different concentrations of microparticles of the bioactive glass 45S5 (m-BG). The biomaterials were obtained through the gel-casting method. Four different composites were prepared at four different weight percentages of m-BG: 0, 5, 10, and 15 %. Glutaraldehyde 0.25 wt% (GA) was used as the cross-linker. The composites were characterized by scanning electron microscopy and the in vitro degradation of the films was studied by measuring the water uptake and weight loss. The drug release kinetics was quantified spectrophotometrically. The inhibition zone test and the plate count method were used to evaluate the antibacterial activity of the samples. Three staphylococcus strains were evaluated: Staphylococcus aureus ATCC6538, S. aureus ATCC29213, and Staphylococcus epidermidis ATCC12228. Cytotoxicity effects were evaluated through the MTT assay. The addition of m-BG to GS films showed no effects on the amount of water uptake, but led to an increase in the weight loss over time, even with m-BG content. The release rate of VC was also affected by the increasing concentration of m-BG in the composite films. However, the antibacterial effects of the composites were not improved by this modulation. All composites strongly inhibited staphylococcal cells with similar strength. On the other hand, liquid extracts from the composites resulted in cytotoxic effects on MG-63 osteoblast-like cells due to the presence of GA, but not to the concentration of VC or m-BG.


Introduction
The pharmacokinetics, pharmacodynamics, nonspecific toxicity, and efficacy of drugs may be controlled by various drug delivery systems [1]. The controlled and sustained delivery of a certain drug ensures that an adequate concentration will reach the target organ over a long period of time and will thus be therapeutically efficient [2]. In wound healing and tissue engineering, the modulation of the release of a drug is of particular interest because one important factor that interrupts the healing is the bacterial burden in the wound [3,4]. Staphylococcus aureus and Staphylococcus epidermidis are important pathogens involved in wound infection [4].
Several drug delivery systems based on biopolymers have been studied to optimize drug release profiles because of their many interesting features, including good biocompatibility, biodegradability, and good mechanical properties [5,6]. Among biopolymers, gelatin stands out strongly due to several key advantages: it is biodegradable, biocompatible, and nonimmunogenic, and has hemostatic properties [7,8]. In addition, gelatin can be an exceptionally adaptable drug delivery carrier because drug loading and releasing kinetics from gelatin matrices can be independently tuned [8][9][10][11][12]. Nevertheless, due to its high brittleness and low intensity, gelatin is often used after modification by different methods such as grafting [15], crosslinking [13,14], and blending [16][17][18].
Starch-based polymers and composites have also been proposed as drug delivery systems [20,21] and have great potential to be applied in the biomedical field [19]. Biodegradable starch-based polymers present good compatibility, proper mechanical properties, and their degradation products are non-toxic [19]. Corn starch can be used together with gelatin as reinforcement to the polymeric matrix [22] and for the controlled release of drugs [23].
In the last years, some researchers have shown that the combination of bioactive glasses (BGs) with biopolymeric matrices allows modifications in the release profile of antibiotics, in particular vancomycin hydrochloride (VC) [24][25][26]. Unfortunately, many often, the biological response or antimicrobial effects have been overlooked. VC is used to treat infections caused by Gram-positive bacteria such as methicillinresistant S. aureus (MRSA) [27].
The addition of BGs to a biopolymeric matrix can lead to several interesting biological properties [28]. BGs are being increasingly considered in bone tissue engineering, due to their ability to bind strongly to the bone and promote bone growth after in vivo implantation [29,30]. In addition, BGs have a huge potential for applications in wound healing and soft tissue engineering [31]. In particular, 45S5 BG is a promising material for applications in soft tissue engineering. For example, it has been demonstrated that microparticles of 45S5 BG possess angiogenic effects [32][33][34], promote the expression of genes related to wound healing [35], and accelerate the recovery of skin wounds [36]. The aim of this work was to study the influence of increasing concentrations of 45S5 BG microparticles (m-BG) on the water uptake capacity, degradation behavior, and release profile of VC. We also investigated the in vitro antistaphylococcal effect and potential in vitro cytotoxicity on MG-63 human osteosarcoma cells of m-BG-containing composites based on gelatin-starch (GS) biopolymers.

Composite films
The composite films were obtained by preparing a 5 % gelatin solution in distilled water and heating it at 80°C for 10 min, then a 1 % starch solution in J Mater Sci distilled water and heating it to 85°C for 5 min, and then homogenizing the solutions. The ratio between gelatin and starch was 5:1. Glycerol was added at a concentration of 20 % in relation to the total volume of solution and used as plasticizer. Four concentrations of m-BG particles were used to obtain films with 0, 5, 10, and 15 % of m-BG respect to the polymer solution. Before the addition of m-BG, the microparticles were passivated in phosphate-buffered saline (PBS) for 12 h to avoid the sudden release of a high ion concentration from the BG. The solutions were cooled to 30°C with constant stirring to incorporate VC at a concentration of 1 mg mL -1 . Alternatively, composites without VC loading were also prepared as controls. Then 20 mL of the respective mixes were placed in 8.5-cm-diameter polystyrene petri dishes and the moldings were dried at 4°C. After drying, composites were cross-linked for 10 min in a GA solution (0.25 wt%). This concentration was chosen because it has been previously reported that 0.25 wt% of GA allows obtaining a crosslinking degree of approximately 85 % and preventing the release of gelatin into the buffer solution [37]. Finally, the composites were washed gently in distilled water and then dried at room temperature (25°C) for 24 h.

Morphological characterization
Scanning electron microscopy (SEM) was used to characterize surface of films. According to previous work [26], the films were fixed with a 2.5 % GA 0.1 M PBS solution overnight at 4°C. After that the materials were washed with distilled water, and then dehydrated by means of a graded series of ethanol solutions. The samples were examined using a JEOL JSM 6480 LV, Japan microscope.

In vitro degradation study
Water uptake assay The water uptake capacity was determined as described previously in Rivadeneira et al. [26]. The films were cut and weighed. The materials were incubated in 15 mL of water at 37°C. At regular intervals, samples were removed and the excess of water at surface was withdrawn with a filter paper. The weights of films were recorded until reaching equilibrium swelling. The water uptake capacity of the films was calculated according to the Eq. (1): Weight loss The degradation pattern was studied at 37°C in PBS through the weight loss profile of composites. Circleshaped composites (10 mm in diameter) were immersed in PBS for up to 28 days. After each period of time, films were washed twice with distilled water to remove possible material adsorbed on the surface, and then dried at 37°C. The PBS solution was replaced twice a week. Three samples were assessed for each group. The weight loss (WL) was calculated by means of Eq. (2), where W 0 is the initial weight and W 1 is the weight after soaking in PBS at a given time: In vitro VC release The in vitro VC release was quantified as previously described in Rivadeneira et al. [26]. The composites were cut in 5-mm-diameter discs (area = 0.2 cm 2 ) with a paper punch incubated in 1 mL of distilled water at 37°C. At predetermined intervals, the VC concentrations were determined by measuring the absorbance at a wavelength (k) = 280 nm on a UV-Vis Thermo Spectronic Helios Beta v.460 (Thermo Electron Corporation, Massachusetts, USA). Three samples in each condition were evaluated. The data are expressed as mean ± SD.

Antimicrobial efficacy
Staphylococcus aureus ATCC6538, S. aureus ATCC29213, and S. epidermidis ATCC12228 were cultured in Mueller-Hinton broth for 24 h at 37°C. Bacterial cell suspensions were adjusted to 6-7 log cfu mL -1 . Previous to antibacterial experiments, film discs were sterilized under UV light for 20 min each side. Antibacterial efficacy was evaluated through the following two methods.

Zone of inhibition assay
Film discs of 5-mm diameter were placed on Mueller-Hinton agar plates which had been J Mater Sci previously seeded with 100 lL of described bacterial suspension. Then the plates were incubated at 37°C for 24 h. Finally, the diameters of the inhibition zones were measured in mm. All tests were performed in triplicate and the means and SD were determined.

Viable counts
Viable counts were determined in SWF. The composite samples were incubated for 48 h at 37°C in 2 mL of cellular suspensions. Each staphylococcus suspension in presence of SWF only was used as controls. Samples were collected at 0, 24 and 48 h and spread on agar plates containing Mueller-Hinton. The viability of cells was quantified by counting in. All tests were performed in triplicate.
The results are expressed as log 10 cfu mL -1 ±SD.
The antibacterial activity was determined as the difference between the log number of bacteria in the control and that in the test composites. The antimicrobial activity was classified as low, moderate, or high according to the criterion of Gallant-Behn et al. [38].

MG-63 osteoblast-like cells
MG-63 osteoblast-like cells were cultured in DMEM medium supplemented with 10 % FBS, 100 lg mL -1 streptomycin, and 100 U mL -1 penicillin at 37°C in a 5 % CO 2 atmosphere. Cells were seeded in a 75-cm 2 flask, and after reaching 70-80 % of confluence, the cells were removed from the flask by trypsinization. To determined total cell counts and the number of viable cells, a Trypan blue stain was performed using a Neubauer hemocytometer. For the experiments, cells were grown in multiwell plates. After cells reached the desired confluence, the monolayers were washed with DMEM and then incubated under the different conditions of each experiment.

Composite liquid extracts
Before cell incubation, the films (0.20 cm 2 ) were sterilized by UV light for 20 min each side. After, the film discs were immersed in 1 mL of DMEM for 48 h at 37°C. Control samples were obtained by incubating under the same condition DMEM medium without the composites.

MTT assay
MTT assay was carried out as previously [39]. Briefly, 2.5 9 10 4 cells per well were seeded in a 96-multiwell dish and incubated overnight to allow adherence. Afterwards, the cells were exposed to liquid extracts for 48 h. After exposure, the medium was remove and cells were incubated with 0.5 mg mL -1 MTT under normal culture conditions for 3 h. Cells were lysed in DMSO (100 lL per well), and the absorbance was measured at 570 nm in a Microplate Reader (7530, Cambridge Technology, Inc., Karlstad, Sweden). Cell viability is shown as the percentage of the control value (assuming data obtained from untreated cells as 100 %).

Statistical analysis
Data were statistically analyzed by one-way analysis of variance, ANOVA (SPSS 15.0 statistical package software). Tukey's multiple comparison post-tests for intergroup analysis and p values of \0.05 were considered to be statistically significant.

Results
Composite films Figure 1 shows the surface morphologies of VC-loaded composites studied by SEM. GS films showed a dense, smooth, and uniform surface (Fig. 1a). The presence of m-BG particles led to the formation of irregular protrusions on the surfaces of the films (see arrows). An increase of the roughness of the surface with increasing concentration of m-BGs in the films can be qualitative assessed by inspection of the SEM micrographs ( Fig. 1b-d).
The addition of VC to the films did not affect the morphology of the films in comparison to non-releasing composites (not shown). The coating was approximately 65.08 ± 3.75 lm for GS-BG5 %, 82.1 ± 2.55 lm for GS-BG10 %, and 106.67 ± 5.77 lm for GS-BG15 %.

In vitro degradation
The values of water uptake capacity for each composite (in %) are shown in Fig. 2a. We found no statistically significant differences in the water J Mater Sci absorption capacity of the composites. The water uptake capacity of the films was of the order of 250-300 %. The swelling reached equilibrium after 15 min. Figure 2b shows the weight loss of films containing VC grown in PBS for 28 days. The results show that the mass loss increased gradually over time. After a week, the weight loss of the composites was similar. Nevertheless, after this time, the GS-BG films showed faster degradation than the GS films. GS-10 % BG and GS-15 % BG were completely degraded by day 20. On the other hand, after 28 days of degradation, GS had a maximum mass loss of 43 % and GS-5 % BG a maximum mass loss of 57 %. However, the difference was not statistically significant. Figure 3 shows the release behavior of VC from the films. A burst release was expected during the first hours due to the hydrophilic nature of VC. In general, the release profiles followed typical three phases: a rapid release phase during the first hours; a decrease in the release rate and a change in the slope of the curves; and a zero order release (constant release rate), which lasted until the end of the experiment. The release rate of VC from the composites was modified by the presence of 45S5 m-BG. The amount of VC released to the medium increased as the m-BG concentration increased. Most of the VC was released during the first 24 h. At this time, the amount of VC released was 0.147 ± 0.080 mg mL -1 for GS, 0.2078 ± 0.087 mg mL -1 for GSBG-5%, 0.2526 ± 0.060 mg mL -1 for GSBG-10%, and 0.2867 ± 0.020 mg mL -1 for GSBG-10%.

Effects of 45S5 m-BG on VC release
In the first stage, GS-BG films showed an abrupt initial burst release, while GS films showed a much lower initial release of VC. In the second stage, GS films exhibited a longer release period than GS-BG composites and showed a controlled release of VC.

Inhibition zone
The values of the inhibition zone (in mm) around the composites are shown in Table 1. The presence of the composites inhibited the growth of the three staphylococcus strains tested. S. aureus ATCC6538 showed the largest inhibition zone, whereas S. aureus ATCC29213 and S. epidermidis ATCC12228 showed inhibition zones of similar size. The addition of 45S5

Viable counts
The initial cell concentrations (expressed as log 10 cfu mL -1 ) were 5.52 ± 0.27 for S. aureus ATCC6538, 5.66 ± 0.28 for S. aureus ATCC29213, and 5.79 ± 0.14 for S. epidermidis ATCC1228 (Fig. 4). No statistical differences were observed in the initial inoculum size between the strains. The composites strongly inhibited the cell viability of the Staphylococcus spp. tested, with similar strength. After 24 h, cell viability was around 4.30 log 10 cfu mL -1 for S. aureus ATCC6538, 4.33 log 10 cfu mL -1 for S. aureus ATCC29213, and 4.50 log 10 cfu mL -1 for S. epidermidis. After 48 h, the cell viability of the three strains was below the detection limits. In terms of antimicrobial activity, the composites presented high antimicrobial activity at 48 h, since growth inhibitions were higher than 3-log reduction.

Discussion
Several researchers agree that BG systems including 45S5 composition can modify the release profile of a drug [24][25][26]. Unfortunately, the biological consequences of this effect have been often overlooked.
Here, for the first time, we studied the release profile of a drug (VC), and the water uptake capacity, degradation, in vitro antistaphylococcal effect, and cytotoxicity (on MG-63 osteoblast-like cells) of GS films with increasing concentration of 45S5 m-BG. When chemical cross-linked biopolymer films are developed for drug delivery applications, a very important parameter should be considered is the water uptake capacity. This parameter will affect not only the film morphology and structure [40] but also the drug release rate in diffusion-controlled systems [41].
In the present study, the water uptake capacity of the polymer was not affected by the presence of b Figure 4 Viable counts of Staphylococcus aureus ATCC6538 (a), S. aureus ATCC29213 (b), and S. epidermidis ATCC1228 (c) in the presence of films. Asterisk indicates a statistically significant difference at p \ 0.05 between VC-releasing films and control cell suspensions at time 0. m-BG in the polymeric matrix. Although this is in accordance with our previous results [26] and with results of other authors [42], other studies have reported opposite findings [43][44][45]. These discrepancies can be explained by the nature of the biopolymeric matrix or the BG particle size. More specifically, when 45S5 m-BG nanoparticles are included in hydrophobic polymers, the water uptake capacity is improved [43,44]. On the other hand, 45S5 BG nanoparticles seem to enhance the water uptake capacity, probably due to the higher time of exposure and larger surface area [45]. Another parameter that should be considered is the degradation behavior of biomaterials in a physiological environment, since this parameter plays an important role in the regeneration of new cells [46]. After implantation, biomaterials interact with the fluids of the tissue and a degradation process starts. The in vitro degradation rate of composite materials is determined as a function of the hydrolysis time in PBS in normal physiological conditions (pH 7.4 at 37°C). The weight loss increases over the incubation period and with the increasing concentrations of 45S5 m-BG [44,47,48]. Nevertheless, some studies have shown that the addition of BGs reduces the degradation rate of biomaterials [49][50][51][52]. The proposed explanations for the slower degradation lies on that the BG particles induce a rapid exchange of protons in the water for the alkali in the BG, providing a pH buffering effect that neutralizes the acidic degradation of the by-products produced during hydrolysis, thus preventing the autocatalytic effect [43,44]. Another possible effect is that the degradation rate decreases because the incorporation of inorganic filler into the polymer matrix decreases the porosity of composites [51]. A decrease in porosity leads to a decrease in the surface area exposed to the medium [52]. In this study, the presence of BG particles could have increased the surface area of the composites and could have accelerated the degradation of GS-BG samples. Indeed, as mentioned in the materials and methods section, 45S5 m-BG samples were previously incubated in PBS before being added to the polymeric matrix to eliminate the increase in pH that could induce cell damage or that could degrade the antibiotic [53]. In that way, the aforementioned pH buffering effect is likely not taking place in the present composites. Finally, it has been proposed that the increased weight loss in the presence of BG could be related to the dissolution of the BG particles, as previously reported for other composites [47,54,55].
Contextualizing these results for a potential application of the composites in wound dressing engineering, some considerations can be made. In wound dressing engineering, the choice of a polymer or polymer combination depends on the fact that degradation rate matches the typical wound healing time frame for proper tissue influx. Many factors affect wound healing but, in normal conditions for skin wounds, the expected healing time frame is about 4 weeks [56]. Here, we found that, in general, films containing 45S5 m-BG degraded too fast in relation with the wound healing time frame. Nevertheless, the degradation of biomaterials can be modulated and improved by different techniques like blending [57], incorporation of enzymatic inhibitors [58], or crosslinking [59].
The drug release rate of VC was found to be affected by the presence of m-BG. This modulation depended, in general, on the m-BG concentration. This result is in agreement with previous research data [52,60]. A possible explanation for this behavior arises from chemical and physical considerations. Chemically, the drug and the biopolymeric matrix may have a greater bounding affinity [52]. Physically, BG particles in the polymeric matrix leave no free space for the antibiotic, promoting the fast release of the drug [52,61]. This can also be explained by the increase in the surface area of the composite by the addition of m-BG, as previously discussed [62].
Both the water uptake capacity and the degradation rate also play important roles in the release kinetics of a drug [60]. If water uptake is higher, the matrix would be more openly spaced and thus easier for drugs to diffuse out. Material degradation also offers relatively large open channels, thus accelerating drug diffusion [60]. Since the addition of m-BG accelerates the degradation of composite films, this effect could explain the results found here.
The treatment of the agar plates with drug-loaded composite films led to obvious inhibition zones. The inhibition zones were not affected by the m-BG concentration. On the other hand, there were no inhibition zones in films not releasing VC, which indicates the lack of bactericidal activity in the absence of the antibiotic. This result is relevant since previous research has reported antibacterial effects of 45S5 BG per se [63][64][65][66]. Here, the antibacterial effects depended on the release of VC. The plate count method J Mater Sci showed that the composites strongly inhibited the growth of staphylococcal cells in SWF. The exposure of the three staphylococcal strains tested to composites after 48 h resulted in a reduction of the total viable count. Interestingly, the composite exhibited similar antibacterial strength. This means that the greater concentration of VC released by the GS-BG composites (Fig. 3) does not indicate an improvement in their antibacterial effects. This phenomenon, where the drug concentration is not related to the antibacterial strength, is in agreement with our previous results [26,39] and with those of other authors with related biomaterials [67].
In the context of wound regeneration, the release of a drug to a wound depends on the condition of the wound [68]. Often, a burst release is considered as a negative consequence of long-term controlled release systems [69]. Nevertheless, high initial delivery rates or rapid release may sometimes be desirable [69] and should be considered in the immediate performance of dressings [68]. A high burst release followed by a decreased release over several days may allow a robust response and eliminate large numbers of bacteria, and can therefore be suitable in cases of developing infection. In a wound, it is commonly accepted that a concentration of bacteria above 10 5 cfu mL -1 determines an infection [70,71]. In this context, the presence of m-BG on films led to an abrupt initial burst of VC release, which is suitable to control the large number of bacteria present in infected wounds.
To complete the results of bacterial inhibition, the composites should be non-toxic to mammalian cells. Biomaterials may be toxic to cells due to the dressing material itself, the material processing or the incorporation of antimicrobials [68]. Here, the results obtained by the MTT assay (Fig. 5) showed that the composites exerted some cytotoxicity. Nevertheless, this cytotoxicity was related to the presence of GA but not to the concentration of m-BG or to the amount of VC released to the medium. This finding supports the idea of BGs as promising materials for soft tissue engineering and wound healing [31], since the presence of BGs can impart superior functionalities to a biopolymeric film than the biopolymeric matrix by itself.
As mentioned above, some of the limitations of biomaterials, such as the lack of adequate mechanical properties, can be overcome by crosslinking [59]. According to other authors, the cytotoxicity of GA depends on the concentration used, and up to 8 % GA has been found to be non-cytotoxic [72]. In this work, 0.25 % GA resulted in cytotoxicity to human osteoblast-like cells. Nevertheless, this kind of limitation can be overcome using other crosslinking agents like genipin, which has been demonstrated to be much less cytotoxic [71,73].

Conclusions
The present research demonstrated that the presence of 45S5 m-BG in gelatin-starch films can modulate a number of physicochemical parameters of relevance in drug delivery system engineering. In particular, the degradation and VC release rates can be affected by increasing concentrations of 45S5 m-BG, but the water uptake capacity of GS films does not depend on the BG content. These modifications did not lead to changes in the antibacterial effects of the composites or to cytotoxicity effects. Future research should focus on finding a proper cross-linker agent for gelatin and investigating the biological response of composites in relation to wound healing.

Compliance with ethical standards
Conflict of interest The authors declare that they have no competing interests.