Poly(vinyl alcohol)/Pullulan Composite Hydrogels as a Potential Platform for Wound Dressing Applications

Hydrogels are 3D networks with an excellent ability to retain a high amount of water or biological fluids, representing suitable candidates for wound dressing applications. They can provide a protective barrier and a moist environment, facilitating wound treatment. The present paper focuses on physical hydrogels obtained from poly(vinyl alcohol) (PVA) and pullulan (PULL) mixtures in different weight ratios by using the freezing/thawing method. Hybrid hydrogels of similar polymer compositions were prepared in the presence of 0.5% Laponite® RD. The influence of polysaccharide and clay addition on the properties of PVA hydrogels was investigated. Scanning electron microscopy showed evidence of the inner porous structure. The viscoelastic properties were investigated in different shear conditions and revealed the influence of the hydrogel composition on the network strength. The swelling behavior was followed in physiological saline solutions at 37 °C and pH = 7.4. For all samples, a quasi-Fickian diffusion mechanism was found. The delivery of neomycin sulfate was studied in similar conditions as for the swelling tests (0.15 M NaCl solutions; 37 °C; pH = 7.4) and different kinetic models were used to determine the release mechanism. The Peppas–Sahlin approach described very well the in vitro drug release mechanism from the polymeric hydrogels in the absence of clay. However, the hybrid polymer/clay hydrogels showed the best fit with the Korsmeyer–Peppas model. According to the present study, the porous membranes containing 40–60% PULL (in absence of clay) are suitable for the release of therapeutic agents at wound sites in physiological conditions.


Introduction
Hydrogels represent water-insoluble networks with a high absorption capacity of water and biological fluids. A synergistic combination of various polymers, clays, proteins, and peptides generates composite hydrogels with different functions and improved properties. During recent years, multicomponent networks were obtained using synthetic [1,2] and natural [3,4] polymers and they were extensively investigated as potential materials for various biomedical applications [1][2][3][4][5]. The viscoelastic properties of hydrogels can be tuned during their preparation, resulting in characteristics similar to those of soft tissues. Generally, the hydrogels can be obtained using physical, chemical, or combined methods [5,6]. The physical hydrogels are promising drug delivery systems, with their high internal void fractions facilitating a high loading capacity and an effective control of the release of therapeutic agents [7,8].
Polysaccharides are found abundantly in nature, being widely used for preparing hydrogels [4,5]. Among them, pullulan (PULL) proved to be a versatile water-soluble natural polymer, produced by fungi (such as Aureobasidium pullulans), which forms viscous solutions, stable in time, over a wide range of temperatures and pH [9]. Due to its valuable properties (hydrophilicity, biocompatibility, high bioadhesion, biodegradability, non-toxicity, edibility, ability to form films that are oil-resistant and impermeable to oxygen, etc. [10]), it was accepted by the FDA and European Union as a safe biopolymer and has Table 1. The composition of mixtures used for preparing hydrogels and the loss tangent obtained for the investigated samples after they were subjected to freezing/thawing cycles (strain of 1%, oscillation frequency of 1 rad/s, temperature of 37 • C). The values of the swelling exponent (n s , Equation (1)) are also given.  Figure 1 shows SEM images for six selected hydrogels obtained in the absence of clay. A porous three-dimensional structure was observed for samples 1-8, while sample 9, with a low PVA content, appeared brittle due to the small number of crosslinking points. The network strength for this composition can be improved by subjecting the samples to a higher number of freezing/thawing cycles. According to rheological and SEM analysis, three cycles were selected for the preparation of the samples, ensuring the formation of the interconnected porous networks for polymer and hybrid hydrogels. Sample 10 did not form a network structure, regardless of the number of applied freezing/thawing cycles. The clay acts as a nanofiller; its addition to the polymer network can improve formation of the interconnected porous networks for polymer and hybrid hydrogels. Sample 10 did not form a network structure, regardless of the number of applied freezing/thawing cycles. The clay acts as a nanofiller; its addition to the polymer network can improve the rheological, thermal, and barrier properties [38,39], contributing to an extended wall formation, especially for high PVA content ( Figure 2).  (Table 1)   formation of the interconnected porous networks for polymer and hybrid hydrogels. Sample 10 did not form a network structure, regardless of the number of applied freezing/thawing cycles. The clay acts as a nanofiller; its addition to the polymer network can improve the rheological, thermal, and barrier properties [38,39], contributing to an extended wall formation, especially for high PVA content ( Figure 2).  The analysis of SEM images shows that both types of hydrogels present a highly porous structure with interconnected pores. During freezing/thawing cycles, a phase Gels 2023, 9, 580 5 of 16 separation accompanies the formation of junction points in the network; thus, the long PULL chains contribute to the pore connectivity and consolidation of the pore wall. Thus, for PULL contents between 40% and 60% (samples 5-7), the pore sizes are between 20 and 50 µm and they slowly decrease for higher PVA contents (more junction points are formed during freezing/thawing process). Generally, this porous structure provides a better drug and nutrient diffusion and metabolite exchange.

Viscoelastic Behavior of Hydrogels
Rheology is one of the most useful techniques to quantify the gel properties in correlation with their applications. Small-amplitude oscillatory shear investigations reveal the viscoelastic response and give information of the network structure. Figure 3 shows the experimental data obtained in frequency sweep tests for PULL/PVA hydrogel with 10% PULL in the presence and absence of clay. G' (elastic) and G" (viscous) moduli are nearly independent of oscillation frequency, with G' > G", and the loss tangent, tan δ = G"/G', is around 0.1, indicating a network structure. For this composition, the clay addition slowly increases all viscoelastic parameters.
The analysis of SEM images shows that both types of hydrogels present a hig porous structure with interconnected pores. During freezing/thawing cycles, a ph separation accompanies the formation of junction points in the network; thus, the l PULL chains contribute to the pore connectivity and consolidation of the pore wall. T for PULL contents between 40% and 60% (samples 5-7), the pore sizes are betwee and 50 μm and they slowly decrease for higher PVA contents (more junction points formed during freezing/thawing process). Generally, this porous structure provid better drug and nutrient diffusion and metabolite exchange.

Viscoelastic Behavior of Hydrogels
Rheology is one of the most useful techniques to quantify the gel properties in relation with their applications. Small-amplitude oscillatory shear investigations re the viscoelastic response and give information of the network structure. Figure 3 shows the experimental data obtained in frequency sweep tests PULL/PVA hydrogel with 10% PULL in the presence and absence of clay. G' (elastic) G" (viscous) moduli are nearly independent of oscillation frequency, with G' > G", the loss tangent, tanδ = G"/G', is around 0.1, indicating a network structure. For composition, the clay addition slowly increases all viscoelastic parameters.   Figure 4a presents the G' dependence on the oscillation frequency for polymeric hydrogels with increasing PULL content, and Figure 4b shows the behavior of hybrid polymer/clay hydrogels at 37 • C.
A synergistic effect is observed for low PULL content (sample 2), when intermolecular interactions determine an increase in the network strength. By subjecting the PVAentangled solution to successive freezing/thawing cycles, crystalline junction zones appear, acting as physical crosslinking points and generating the network structure [23]. The hydrogel formation is attributed to intermolecular hydrogen bonds achieved between the −OH groups of PVA chains, on the one hand, and between PVA and PULL macromolecules, on the other hand. Furthermore, the chain entanglements contribute to the network strength. Further PULL addition gradually decreases the G' value due to the decrease in the number of junctions between PVA macromolecules. Sample 10, containing only PULL chains, behaves as a viscous fluid (no physical crosslinking after applying even a high number of freezing/thawing cycles). This demonstrates that PVA is responsible for 3D network formation. In the presence of clay (Figure 4b), the PVA-PVA interactions are slowly perturbed, influencing the network formation. Thus, the loss tangent (tan δ) values increase for PULL contents between 20% and 60% ( Figure 5). When the polymer mixture is mostly composed of PULL macromolecules, the clay addition contributes to the formation of junction points between PULL chains during mixing, and the weak network is preserved after applying freezing/thawing cycles, determining an increase in G' value (tan δ decreases). A synergistic effect is observed for low PULL content (sample 2), when intermo ular interactions determine an increase in the network strength. By subjecting PVA-entangled solution to successive freezing/thawing cycles, crystalline junction zo appear, acting as physical crosslinking points and generating the network structure The hydrogel formation is attributed to intermolecular hydrogen bonds achieved tween the −OH groups of PVA chains, on the one hand, and between PVA and PU macromolecules, on the other hand. Furthermore, the chain entanglements contribut the network strength. Further PULL addition gradually decreases the G' value due to decrease in the number of junctions between PVA macromolecules. Sample 10, cont ing only PULL chains, behaves as a viscous fluid (no physical crosslinking after apply even a high number of freezing/thawing cycles). This demonstrates that PVA is resp sible for 3D network formation.
In the presence of clay (Figure 4b), the PVA-PVA interactions are slowly pertur The rheological behavior of entangled polymer solutions is not significantly affected by the presence of Laponite ® RD, because the chain entanglements hinder the formation of the platelet structure [40]. However, the addition of clay to polysaccharide-based systems determines qualitative changes in the flow behavior and mechanical spectra [41,42]. It was also shown that the structural state of Laponite ® RD dispersions depends on clay concentration and ionic strength [38,43]. The manifestation of attractive and repulsive interactions between particles can lead to arrested states (either attractive gel or repulsive glass), generating a variety of rheological behaviors, from quasi-Newtonian to viscoelastic and shear thinning fluids [43]. The rheological behavior of entangled polymer solutions is not signifi by the presence of Laponite ® RD, because the chain entanglements hinder of the platelet structure [40]. However, the addition of clay to polysacchar tems determines qualitative changes in the flow behavior and mechanical s It was also shown that the structural state of Laponite ® RD dispersions de concentration and ionic strength [38,43]. The manifestation of attractive interactions between particles can lead to arrested states (either attractive g glass), generating a variety of rheological behaviors, from quasi-Newtonia tic and shear thinning fluids [43].

Swelling Behavior
The swelling of hydrogels is carried out in NaCl solution at pH = 7.4. Figure 6 shows the swelling curves for hydrogels obtained in the absence and presence of clay. The first observation concerns the effect of clay: all polymer/Laponit samples present a smaller swelling degree as compared with similar polymer samples without clay. At equilibrium, a maximum swelling is obtained for sample 3 containing 25% PULL in the absence of clay, about 700% (Figure 6a), whereas PVA/clay hydrogel shows the maximum swelling level (around 600%, Figure 6b). A burst effect is registered for samples 9 and 9C, and the samples quickly disintegrate, with the network structure being unstable in the presence of the solvent. The clay determines the decrease in swelling degree for samples 1 to 6 (F then, when PULL chains are predominant (samples 7, 8, and 9), no significant diff are registered, in the limit of the experimental errors. The following relationship to determine the nature of the solvent diffusion through the hydrogel pores: is the total water uptake at time t; t m and  m are the amounts o absorbed by the network at a given time, t, and at equilibrium, respectively; k i acteristic rate constant depending on the network structure; ns is a transport nu dicating the solvent transport mode through the hydrogel pores (the swelling could be controlled by diffusion and/or relaxation).
Equation (1) is applied to the early swelling stage, and from the slope of li pendences of ) (t F vs. t, the diffusion exponent ( s n ) is obtained. For all PVA-co hydrogels, n < 0.5 (Table 1), suggesting a quasi-Fickian diffusion, when the init The clay determines the decrease in swelling degree for samples 1 to 6 ( Figure 7); then, when PULL chains are predominant (samples 7, 8, and 9), no significant differences are registered, in the limit of the experimental errors. The following relationship is used to determine the nature of the solvent diffusion through the hydrogel pores: where F(t) is the total water uptake at time t; m t and m ∞ are the amounts of solvent absorbed by the network at a given time, t, and at equilibrium, respectively; k is a characteristic rate constant depending on the network structure; n s is a transport number indicating the solvent transport mode through the hydrogel pores (the swelling process could be controlled by diffusion and/or relaxation). Previously, by using water as a swelling environment, the following s n va were obtained: 0.18 for pure PVA hydrogel and 0.21 for PVA/ Laponite ® RD hybrid drogel [38]. Furthermore, nanofibers of PULL, PVA, and montmorillonite, for through the exfoliation of clay layers, are superabsorbent in distilled water (143.42 and the absorbency becomes much lower in 0.15 M NaCl solution (39.75 g/g) [44].

Neomycin Delivery
Neomycin is an antibiotic with bactericidal and bacteriostatic action, clinically te against infections with Trypanosoma cruzi and Gram-positive bacteria (such as Stap coccus aureus), and very efficient against most Gram-negative bacterial strains (Pr vulgaris, Escherichia coli, Aerobacter aerogenes, P. aeruginosa), without toxic effects [45,4 Figure 8 presents the kinetics of neomycin release for some hydrogel sample lected on the basis of their swelling behavior shown in Figures 6 and 7. The drug rel begins when the hydrogel samples are introduced into the release medium and the s molecules of water and ions diffuse into the hydrogel. The mobility of polymer segm between two crosslinking points increases when the hydrogel swells, leading to a ri drug mobility and determining its delivery in the environmental liquid. The maxim efficiency is obtained by the addition of 40-60% PULL to PVA hydrogels, whereas clay improves the drug delivery for hydrogels with a high content of PULL.
The neomycin delivery from the hydrogels is discussed by using two models veloped by Peppas and coworkers [47,48]: Korsmeyer-Peppas [47]:

Mt/M∞ = k t n
Peppas-Sahlin [48]: where Mt/M∞ is the fraction of released drug at time t. The release kinetic constant k f Equation (2) depends on the characteristics of drugs [49]. Equation (2) is valid for the drug release stages up to 60% of the maximum level. The release exponent n from E tion (2) is correlated with the release mechanism of the drug from the hydrogel Fickian or non-Fickian diffusion). The constants k1 and k2 from Equation (3)   Equation (1) is applied to the early swelling stage, and from the slope of linear dependences of F(t) vs. t, the diffusion exponent (n s ) is obtained. For all PVA-containing hydrogels, n s < 0.5 (Table 1), suggesting a quasi-Fickian diffusion, when the initial diffusion of solvent molecules into the macroporous network occurs rapidly, and it is followed by a slow diffusion until the equilibrium state is reached. The deviations from Fickian diffusion are attributed to the interactions between the solvent molecules and the network constituents. It is found that the presence of clay or PULL in the PVA network has no influence on the diffusion mechanism.
Previously, by using water as a swelling environment, the following n s values were obtained: 0.18 for pure PVA hydrogel and 0.21 for PVA/Laponite ® RD hybrid hydrogel [38]. Furthermore, nanofibers of PULL, PVA, and montmorillonite, formed through the exfoliation of clay layers, are superabsorbent in distilled water (143.42 g/g), and the absorbency becomes much lower in 0.15 M NaCl solution (39.75 g/g) [44].

Neomycin Delivery
Neomycin is an antibiotic with bactericidal and bacteriostatic action, clinically tested against infections with Trypanosoma cruzi and Gram-positive bacteria (such as Staphylococcus aureus), and very efficient against most Gram-negative bacterial strains (Proteus vulgaris, Escherichia coli, Aerobacter aerogenes, P. aeruginosa), without toxic effects [45,46]. Figure 8 presents the kinetics of neomycin release for some hydrogel samples selected on the basis of their swelling behavior shown in Figures 6 and 7. The drug release begins when the hydrogel samples are introduced into the release medium and the small molecules of water and ions diffuse into the hydrogel. The mobility of polymer segments between two crosslinking points increases when the hydrogel swells, leading to a rise in drug mobility and determining its delivery in the environmental liquid. The maximum efficiency is obtained by the addition of 40-60% PULL to PVA hydrogels, whereas the clay improves the drug delivery for hydrogels with a high content of PULL.
PVA/PULL hydrogels in the presence or absence of clay, n < 0.45 was obtained (Ta suggesting a pseudo-Fickian diffusion with slow drug release [47,50,51]. According to the present data, the Peppas-Sahlin approach (Equation (3)) [ more suitable to describe the in vitro drug release mechanism of neomycin from polymeric hydrogels in the absence of clay (lower AIC values). The Korsmeyer-P model [47] better describes the drug release from the PVA/PULL/clay hydrogels.
The release of neomycin, a water-soluble drug included in the matrix durin hydrogel's preparation, is governed by the porous structure and the neomycin in tions with the network and the release environment. Each component of the netwo a specific role, with a direct effect on rheological properties, swelling behavior, and release rate. PVA ensures the network formation after three freezing/thawing cycle to the intermolecular hydrogen bonds in the crystalline regions acting as junction p in the network. When PVA macromolecules represent the majority, more crystalli gions are formed by applying successive freezing/thawing cycles, and this could detrimental parameter for drug release. The network is less extended by swelli salted solutions and the amorphous regions have a limited diffusivity with consequ on the drug release kinetics. The occurrence of the burst effect observed for sample be correlated with a highly porous macrostructure of the PVA hydrogel (influenc the crosslinking degree), drug diffusion through the pores in balance with the hyd The neomycin delivery from the hydrogels is discussed by using two models developed by Peppas and coworkers [47,48]: Korsmeyer-Peppas [47]: Peppas-Sahlin [48]: where M t /M ∞ is the fraction of released drug at time t. The release kinetic constant k from Equation (2) depends on the characteristics of drugs [49]. Equation (2) is valid for the first drug release stages up to 60% of the maximum level. The release exponent n from Equation (2) is correlated with the release mechanism of the drug from the hydrogel (i.e., Fickian or non-Fickian diffusion). The constants k 1 and k 2 from Equation (3) also illustrate the Fickian or non-Fickian contributions to diffusion during the drug delivery [48]. For PVA/PULL hydrogels in the presence or absence of clay, n < 0.45 was obtained ( Table 2), suggesting a pseudo-Fickian diffusion with slow drug release [47,50,51]. According to the present data, the Peppas-Sahlin approach (Equation (3)) [48] is more suitable to describe the in vitro drug release mechanism of neomycin from the polymeric hydrogels in the absence of clay (lower AIC values). The Korsmeyer-Peppas model [47] better describes the drug release from the PVA/PULL/clay hydrogels.
The release of neomycin, a water-soluble drug included in the matrix during the hydrogel's preparation, is governed by the porous structure and the neomycin interactions with the network and the release environment. Each component of the network has a specific role, with a direct effect on rheological properties, swelling behavior, and drug release rate. PVA ensures the network formation after three freezing/thawing cycles due to the intermolecular hydrogen bonds in the crystalline regions acting as junction points in the network. When PVA macromolecules represent the majority, more crystalline regions are formed by applying successive freezing/thawing cycles, and this could be a detrimental parameter for drug release. The network is less extended by swelling in salted solutions and the amorphous regions have a limited diffusivity with consequences on the drug release kinetics. The occurrence of the burst effect observed for sample 1 can be correlated with a highly porous macrostructure of the PVA hydrogel (influenced by the crosslinking degree), drug diffusion through the pores in balance with the hydrogel swelling when the release begins, osmotic pressure in the diffusion membrane [52][53][54], or high drug loading (when some molecules are adsorbed or weakly bound to the hydrogel surface [38,55]). In pharmaceutical formulations, the burst effect should be avoided because it decreases the therapeutic response and the drug is lost in an uncontrolled and unpredictable manner [52][53][54]. For the physical crosslinked hydrogels, the use of a minimum PVA concentration (when the macromolecular chains are in entangled state [23]) is recommended for hydrogel preparation. For chemically crosslinked PVA hydrogels, the burst effect can be minimized by surface extraction or surface preferential crosslinking [52].
Due to the high hydrophilicity of PVA, its hydrogels have a weak ability to load the drugs and to avoid the initial burst effect [54]. In our study, the clay or polysaccharide addition improves the delivery during the first stage. The pullulan chains ensure the connectivity between different PVA crystalline and amorphous regions. By increasing the content of polysaccharide, the amorphous part of the hydrogel is extended and the overall mobility and diffusivity become significantly higher, contributing to a higher release rate. In addition, the very favorable interactions established between Na + ions from the solvent and the monomeric units of PULL [56] improve the neomycin release. However, for polymer networks in the presence of clay, the increase in the ionic strength of the environment determines a diminution of the swelling degree and drug delivery ability from the hybrid gel matrix.

Conclusions
Hydrogels containing a synthetic polymer (PVA) and a polysaccharide (PULL) in different ratios in the presence and absence of clay are prepared and their main structural and rheological characteristics are investigated. The swelling in physiological conditions (0.15 M NaCl solution, 37 • C) reveals a higher swelling degree in the absence of clay, favoring the drug release.
The models developed by Peppas and coworkers provide an accurate description of neomycin delivery. The Peppas-Sahlin approach describes the in vitro drug release mechanism from the PVA/PULL hydrogels in the absence of clay, whereas the Korsmeyer-Peppas model is more suitable for PVA/PULL clay networks. The introduction of PULL into the PVA physical network improves the hydrogel biocompatibility, but also determines the increase in the mobility and diffusivity in the amorphous zone, favoring the drug release. On the other hand, the clay addition is not beneficial, because as the ionic strength of the environment increases, the swelling is diminished and less drug is released from the network.
The PVA/PULL physical hydrogels could be considered effective carriers for antibiotics delivery, suitable for wound dressing applications. These porous hydrogels provide a wet environment in physiological conditions, and by absorbing the wound exudates, they reduce the risk of infection.

Materials
Poly(vinyl alcohol) (PVA) with a molecular weight of 1.3 × 10 5 g/mol (99% hydrolyzed) was purchased from Sigma-Aldrich (Taufkirchen, Germany) and used as received. Pullulan (PULL) sample of 3 × 10 5 g/mol was purchased from TCI Europe N.V. An aminoglycoside antibiotic, neomycin trisulfate salt hydrate (denoted as neomycin), was purchased from Sigma-Aldrich (Taufkirchen, Germany). It is water-soluble (50 mg/mL) and enhances the cationic-lipid-mediated transfection of reporter plasmids and oligonucleotides [57]. The chemical structure of polymers and neomycin is given in Figure 9. Laponite ® RD with the chemical formula Na

Preparation of the Composite Hydrogels
PVA and PULL were dissolved in Millipore water for obtaining stock solutions of 5% wt. concentration (pH = 7.4). PVA solution was prepared under magnetic stirring at a temperature of about 90 °C and the homogeneous solution was kept for a few hours at room temperature. PULL was dissolved in Millipore water by shaking with a rolling mixer at room temperature, then the solution was stored in a refrigerator. PULL/PVA

Preparation of the Composite Hydrogels
PVA and PULL were dissolved in Millipore water for obtaining stock solutions of 5% wt. concentration (pH = 7.4). PVA solution was prepared under magnetic stirring at a temperature of about 90 • C and the homogeneous solution was kept for a few hours at room temperature. PULL was dissolved in Millipore water by shaking with a rolling mixer at room temperature, then the solution was stored in a refrigerator. PULL/PVA solutions of different compositions (Table 1) were prepared and homogenized under magnetic stirring. Solutions of similar compositions were mixed with clay under high-speed stirring, then ultrasonicated and used immediately to prevent the formation of aggregates.
The freezing of freshly prepared mixtures was performed in liquid nitrogen to avoid phase separation. The thawing step was performed at room temperature for 10 h. The gel formation was checked by rheological measurements and it was established that the application of 3 freezing/thawing cycles was optimum to obtain networks with a porous structure for drug delivery applications.

Scanning Electron Microscopy Studies
The morphology of hydrogels was examined on a cross-section of freeze-dried samples by using the Verios G4 UC Scanning Electron Microscope (Thermo Scientific, Brno, Czech Republic), operating at 10 kV in high-vacuum mode with a backscatter electron detector, ABS (Angular Backscattered Detector). The samples were coated with a 6 nm platinum layer using a Leica EM ACE200 Sputter coater prior examination to improve electrical conductivity and prevent charge buildup. The SEM images were analyzed at various magnifications. Image J software was used to determine the average pore sizes from the SEM micrographs and each value was obtained as an average of at least 80 dimensions.

Rheological Investigations
Rheological measurements were carried out with an MCR 302 Anton-Paar rheometer (Graz, Austria) by using a plane-plane geometry (the upper plate of 50 mm, gap of 500 µm) and a Peltier device for temperature control. A preliminary rheological study was performed for solutions of PVA/PULL mixtures of various concentrations in order to select the concentration for which the polymer chains are in an entangled state. Thus, for For the present study, the chosen polymer concentration was of 5% (wt.).
The viscoelastic behavior of hydrogels with various compositions was investigated at 37 • C in frequency sweep tests, in the linear range of viscoelasticity (established in amplitude sweep tests). The viscoelastic moduli (G' and G") and the loss tangent (tanδ) were determined for oscillation frequencies (ω) between 0.1 rad/s and 100 rad/s (at constant deformation, γ, of 1%). The loss tangent was determined as G"/G' ratio and gives information about the degree of viscoelasticity of the sample.

Swelling Behavior
Saline solutions with concentrations similar to physiological serum (0.9 g/dL aqueous solution of NaCl) were used for swelling tests. The hydrogel swelling was studied at 37 • C and pH = 7.4. The dry hydrogel (weight m o ) was immersed in saline solution and the weight of the swollen sample was determined (m t ) at different times t. The swelling degree (S), as the ratio between the weight of the liquid in the swollen state at a given time (m t ) and the weight of the dried sample (m o ), was calculated using the following equation:

In Vitro Drug Release Study
Samples with different compositions were loaded with drug (0.5% wt. neomycin sulfate) in the solution state. Then, they were subjected to freezing/thawing following the sample procedure mentioned above. In this way, the entire neomycin sulfate molecules Gels 2023, 9, 580 14 of 16 were entrapped in the network. The in vitro drug release behavior was investigated by immersing each hydrogel sample in a closed bottle with 10 mL of saline solution (0.9 g/dL aqueous solution of NaCl, pH = 7.4) at 37 • C in a thermostatic chamber. At different time intervals, 1 mL of the release medium was withdrawn and replaced by an equal volume of fresh, pre-heated solvent, maintaining a constant volume. The amount of drug in the release medium was determined by UV spectrophotometry, using a standard calibration curve (R 2 = 0.9924). The absorbance was recorded at 202 nm using a Libra UV-Vis spectrophotometer (Biochrom, Cambridge, UK).
The drug delivery results were presented as the mean of three values with standard deviation (±SD). The experimental data were analyzed using OriginPro 8.5 software to generate the best regression fits. The minimum values of the residual sum of the squares (RSS) were considered. The Akaike Information Criterion (AIC) [58] was used for the statistical analysis, being independent of the number of parameters (p) introduced by the model: AIC = N ln(SSR) + 2 p where N represents the number of experimental data samples. The smallest AIC values were used to consider the best-fitting approach for describing the drug release mechanism [7,59].