Tailoring Drug Release Profiles From Electrospun Alginate Fibers Through Integration of Mesoporous Silica Nanoparticles

ABSTRACT

Electrospun hydrogels often lack sufficient mechanical integrity for biomedical applications, while nanoparticle‐reinforced systems are commonly associated with synthetic matrices or complex fabrication methods. Here, electrospun methacrylated alginate (AM) fibrous hydrogels reinforced with mesoporous silica nanoparticles (MSNs) are introduced as a simple, biocompatible platform for localized drug delivery. The mats are UV‐cross‐linked in ethanol to maintain fibrous morphology and allow chlorhexidine (CHX) loading from 0.2 to 5 mg ml^{− 1}. Incorporation of 1 wt.% MSNs doubles compressive and tensile strength while reducing swelling by 25%. Fourier transform infrared spectroscopy and thermogravimetric analysis confirm CHX adsorption in mats with and without MSNs. Although both systems exhibit similar initial CHX loading, MSN‐containing hydrogels show prolonged antibacterial activity and higher cytotoxicity, indicating sustained CHX retention and release. Ultraviolet–visible spectroscopy demonstrates reduced CHX release at higher loading in MSN‐reinforced mats, unlike MSN‐free mats, suggesting strong CHX‐MSN‐AM interactions and improved retention within MSN mesopores and the AM matrix. These findings show that MSN incorporation tunes both mechanical performance and drug release behavior, supporting localized antimicrobial delivery for wound dressings and implant coatings.

This graphical abstract presents electrospun alginate methacrylate/mesoporous silica nanoparticle fibers loaded with chlorhexidine and photo‐cross‐linked. By linking formulation, gel fraction, swelling, mechanics, drug release, antimicrobial activity and cell compatibility, it highlights how mesoporous silica nanoparticle loading modulates performance, inviting deeper insight into multifunctional wound‐relevant biomaterial design with promising therapeutic potential.

1. Introduction

Despite considerable progress in the use of biomedical devices as drug delivery platforms, challenges such as suboptimal dosage, local toxicity, lack of specificity, and the development of drug resistance remain. These issues often arise from insufficient spatiotemporal control over drug concentration at the target site according to the treatment needs, whether through an initial burst or sustained exposure over time [1].

Electrospinning has emerged as a promising strategy to address these challenges by enabling controlled, localized delivery from nanofibrous matrices [2, 3]. The technique produces nanofibrous networks with high surface area and tunable porosity, allowing regulated drug loading and release, thereby mitigating limitations such as burst release or rapid depletion [3, 4].

It relies on a relatively simple, low‐cost setup that is industrially scalable and compatible with biopolymers and functional nanofillers, offering precise control over fiber diameter and morphology through adjustable parameters, such as polymer selection, solvent system, voltage, and flow rate [2, 3].

Accordingly, electrospinning enables application‐driven design of nanofibrous matrices. These include tissue‐engineering scaffolds that prioritize porosity and cell infiltration, wound‐care dressings emphasizing exudate management and localized antimicrobial release, and drug‐delivery platforms targeting controlled release kinetics and stable loading of therapeutics [3, 5, 6].

Electrospun scaffolds are particularly relevant for wound healing due to structural similarity to the extracellular matrix (ECM), high porosity, and ability to support nutrient and metabolite exchange [6]. Additionally, the precursor solution can be functionally modified by incorporating inorganic nanoparticles, nanotubes, or therapeutic agents that impart mechanical reinforcement and bioactive functionality [5].

When these components are uniformly dispersed and stably integrated within the fiber matrix, controlled release can be achieved through interactions such as adsorption, entrapment, or diffusion barriers [4]. In the absence of such retention mechanisms, embedded drugs may exhibit uncontrolled release, highlighting the need for composite or cross‐linked architectures to extend delivery time [3, 4].

Among the natural polymers suitable for electrospinning, sodium alginate (SA) stands out as a versatile material for diverse biomedical uses owing to its inherent biocompatibility, mild ionic gelation, and permeable hydrogel networks [7, 8]. SA is a linear anionic polysaccharide derived from brown algae, composed of β‐D‐mannuronic (M) and α‐L‐guluronic (G) acid residues, covalently linked by 1–4 glycosidic bonds [2].

The M/G ratio and cross‐linking conditions allow tuning of mechanical response, swelling and diffusional properties, supporting applications ranging from localized drug delivery and antimicrobial dressings to cell encapsulation, tissue‐engineering and bioink formulations [9, 10, 11].

However, alginate‐based electrospun meshes typically have the disadvantages of water solubility and low mechanical strength. To address this, photo‐chemical cross‐linking via methacrylation has attracted interest as a fast, efficient, solvent‐free, and easily controllable method. Methacrylated alginate (AM) can be synthesized by modifying reactive hydroxyl (C2–C3) and carboxyl (C6) groups using agents such as methacrylic anhydride, glycidyl methacrylate, or 2‐aminoethyl methacrylate. Importantly, photo‐cross‐linkable alginate networks prepared with low doses of photo‐initiator have demonstrated biocompatibility [2, 5].

One disadvantage of directly blending bioactive molecules into electrospun fibers is the risk of activity loss during processing or preferential localization near the fiber surface due to phase separation during jet thinning and rapid solvent evaporation [2, 4, 12]. This behavior, governed by the physicochemical properties of the drug, can result in undesired burst release and reduced therapeutic efficacy [2]. To overcome these limitations, mesoporous silica nanoparticles (MSNs) can be incorporated into the fiber matrix to enable sustained and tunable release profiles, while also increasing the loading capacity and molecular stability [1, 6]. MSNs are commonly synthesized using a modified Stöber process in which surfactant removal via calcination generates a mesoporous structure within the silica framework. This mesoporosity substantially enhances the drug‐loading capacity of MSNs by increasing surface area and pore volume.

The physicochemical properties of MSNs, like particle size and pore size, are governed by several synthesis parameters, such as the pH of the synthesis solution and the type and concentration of the silica precursor and surfactant, which influence hydrolysis rates, nucleation, and corresponding grain growth reactions [13, 14]. These structural features, i.e., high surface area and uniform, tunable mesopores, enable efficient drug loading and controlled release. The surface chemistry of MSNs also allows for functionalization, useful in stimuli‐responsive or targeted drug delivery strategies [15].

However, MSNs used alone often exhibit burst release due to desorption of drugs from the external surface, which can result in cytotoxic effects for the host cells [16]. Embedding MSNs within a polymeric matrix provides an effective strategy to further moderate release kinetics by limiting the initial drug release, which is particularly advantageous for wound‐dressing applications [17, 18]. Such hybrid systems may also introduce additional mechanisms in the drug loading and release behavior, offering more refined control over release profiles compared to polymer‐only electrospun systems [19].

Composite electrospun fibers incorporating nanoparticles within polymer matrices have shown promise in improving the mechanical integrity and biological performance of scaffolds [20]. In the present study, electrospun fibers made of natural‐based AM were reinforced with a comparatively low content of MSNs (1 wt.%) (AM_MSN) to enhance mechanical stability and modulate drug release.

The fibers were ethanol UV‐cross‐linked to preserve morphology and post‐loaded with chlorhexidine (CHX) to minimize activity loss, while maintaining similar loading conditions. The resulting materials were systematically characterized in terms of fiber morphology, swelling capacity, mechanical strength, and drug loading and release performance.

CHX, a broad‐spectrum antiseptic widely used in wound care, dentistry, and preoperative skin disinfection, was selected as a proof‐of‐concept model drug to demonstrate the functionality of the platform [21]. More broadly, target applications include localized antimicrobial delivery in skin or oral wound dressings, implant coatings or infection‐resistant scaffolds, where confined and sustained release is desirable [22, 23, 24, 25, 26].

To the best of our knowledge, this work represents the first report of electrospun AM fibers reinforced with MSNs. CHX release kinetics, antibacterial activity, and cytocompatibility were evaluated to assess the potential of the platform for localized antimicrobial delivery applications.

2. Results and Discussion

2.1. Visualization of Composite Electrospun Fibers

Once the electrospinning parameters were determined, stable overnight processing was performed for AM and AM_MSN. The obtained mats were characterized for thickness, and then they were visualized by electron microscopy (Figure 1) for subsequent determination of fiber diameter.

FIGURE 1.

Scanning electron micrographs of cross‐linked electrospun mats composed of AM: (A, B) without MSNs and (C, D) with MSNs. Green arrows denote the electrospun fibers, while yellow arrows highlight agglomerated MSNs in contact with the fibers.

The measured mat thickness was 0.227 ± 0.025 mm for AM and 0.297 ± 0.068 mm for AM_MSN. Although the same solution volume was processed, AM_MSN was thicker; however, the difference was not statistically significant (unpaired t‐test, p = 0.103) and should be interpreted as a trend. This observed difference can be attributed to adjustments in the electrospinning parameters for each composition. These adjustments were necessary to achieve a stable electrospinning process consistent with reported effects of MSN incorporation on solution properties, such as increased viscosity and conductivity, which can alter the critical voltage (minimum electric field to initiate a stable electrospinning jet), whipping, deposition rate, and fiber stacking behavior [12, 27, 28].

The increased needle‐collector distance may have allowed more complete solvent evaporation, which also contributes to differences in mat structure [27]. No statistically significant difference in average fiber diameter was detected between AM (212 ± 18 nm) and AM_MSN (183 ± 17 nm). In AM_MSN mats, discrete silica agglomerates were observed distributed across the surface and in contact with fibers, with some appearing partially embedded (Figure 1D, yellow arrows). The synthesized MSNs were nanoscale (diameter ∼50 nm; as reported in our previous study); however, agglomerates formed during preparation of the electrospinning solution when particles were added directly to the water‐based AM phase, producing macro‐aggregates that hindered stable electrospinning in preliminary trials [29]. This was mitigated by pre‐dispersing MSNs in the ethanol fraction by probe sonication (ultrasonic processor) prior to incorporating into the aqueous polymer, which reduced large aggregates. A minority of clusters remained after sonication, more homogeneously distributed across the mats. While internal incorporation of MSNs is not required for this system, the morphology suggests that partial entrapment may occur during electrospinning. The similarity in scale between the nanoparticles (∼50 nm) and the fiber diameter (∼200 nm) allows the MSNs to be embedded within the solidifying jet as solvent evaporates, as similarly reported for silica‐polymer composite nanofibers [20, 30]. Loosely attached MSNs are removed during the overnight ethanol bath used for CHX loading, since silica nanoparticles remain well dispersed in ethanol whereas they aggregate in aqueous alginate environments [31, 32, 33]. The application environment is hydrated; consequently, the observed clusters are expected to remain immobilized within the matrix, supporting localized drug delivery and minimizing the potential for free‐particle release.

2.2. Hydration Behavior and Mechanical Properties

2.2.1. Swelling Capacity and Gel Fraction

Swelling experiments after 24 h incubation (shown in Figure 2) were conducted to assess fluid uptake under two conditions: distilled water, to evaluate maximum absorption and calculate the gel fraction, and phosphate buffer saline (PBS) to emulate a physiological environment. In both cases, the incorporation of MSNs caused a significant reduction in swelling. Two‐way ANOVA confirmed significant effects of both material and medium, as well as a strong interaction between both factors (p < 0.001), indicating that the influence of MSN incorporation on swelling depends on the surrounding medium. Specifically, MSN incorporation appeared to reduce the sensitivity of the hydrogel network to the ionic environment, resulting in a smaller difference in swelling behavior between distilled water and PBS for AM_MSN compared to AM (Table S1).

FIGURE 2.

Results for swelling, gel fraction, and mechanical response of electrospun mats are illustrated with (A) swelling degree after 24 h in distilled water (dark blue) and PBS (light blue) (left y‐axis) with gel fraction (ochre, right y‐axis) for both compositions, n = 4. Asterisks denote significance (unpaired t‐test): ^* p < 0.001; ^**** p ≤ 0.0001. (B) Compression and (C) tensile mean stress–strain curves for AM (blue) and AM_MSN (green); n = 3.

This is consistent with previous findings where nanoparticle incorporation reduces water uptake and alters network architecture, as reported in clay‐nanocomposite hydrogels showing that such fillers limit swelling by occupying free volume and increasing matrix stiffness [34, 35]. PBS consistently yielded lower values than distilled water due to ionic screening of the negatively charged alginate network, which reduces electrostatic repulsion between polymer chains and lowers the osmotic driving force for water uptake associated with its higher ionic strength and salt content, as documented in the literature [36, 37]. The reduction was more pronounced in the AM sample than in AM_MSN, possibly because the presence of MSNs introduces additional physical constraints and reduces polymer chain mobility, thereby dampening the network's responsiveness [38, 39]. These findings highlight the importance of considering both the composition of the swelling medium and the presence of nanoparticles when evaluating hydrogel performance. In this study, PBS is the more physiologically relevant condition of the two, although more complex media could more closely replicate in vivo environments [40].

Hydrogel gel fractions, shown on the right axis of Figure 2A, are obtained from the ratio of the dry weight of electrospun mats after photo‐cross‐linking and after incubation for 24 h in ultrapure water, relative to the initial dry weight.

Gel fraction is a property that reflects the proportion of the polymer network that is covalently cross‐linked and remains insoluble after incubation in water, providing an indirect measure of network formation and structural integrity [41]. Upon incorporation of MSNs, statistical analysis revealed a significant increase in gel fraction from 0.71 ± 0.03 to 0.89 ± 0.05. This suggests a more complete cross‐linking reaction, potentially influenced by interaction between MSNs and the polymer matrix that may alter network formation. A higher gel fraction indicates an increased extent of network formation, which may result in a more compact and mechanically robust hydrogel. This hypothesis is further supported by the mechanical testing results discussed in the following section. While nanoparticle incorporation has been reported to enhance mechanical stability and reinforce hydrogel composites, it can also contribute to reduced water uptake [41].

2.3. Measurement of Tensile and Compressive Strength

The capacity to absorb and retain physiological fluid is crucial for electrospun nanofiber composites designed for use in drug delivery. However, mechanical robustness under physiological loading is equally critical, particularly in biomedical contexts where materials are exposed to deformation, compression, and tensile stress [39].

To evaluate mechanical performance, uniaxial compressive and tensile tests were performed and displayed on both AM and AM_MSN hydrogel mats. The obtained stress–strain curves are presented in Figure 2B,C, and the resulting parameters: compressive and Young's modulus (MPa), maximum stress (MPa), and maximum strain (%) were calculated and reported in Table 1. The compressive modulus increased significantly from −7.27 ± 0.36 MPa in AM mats to −9.59 ± 0.36 MPa in AM_MSN samples (p ≤ 0.0001). Although both formulations remained within the same order of magnitude, the increase of more than 2 MPa represented a meaningful gain for hydrogel mechanics. Such reinforcement is not only relevant at the cellular scale, where stiffness guides spreading and migration, but also at the macroscopic level, as observed during handling of the scaffolds in the laboratory [42, 43]. The MSN‐reinforced mats were more robust and easier to manipulate, which adds to the reported advantages of electrospun architectures over bulk hydrogels in terms of stability and usability [44, 45, 46]. Compressive strength also increased from 0.35 ± 0.01 MPa to 0.42 ± 0.036 MPa (p < 0.05). Tensile modulus more than doubled, from 0.35 ± 0.04 MPa to 0.81 ± 0.07 MPa (p < 0.0001), and tensile strength improved as well (0.26 ± 0.02 MPa vs. 0.47 ± 0.07 MPa, respectively, p < 0.01). Together, these results confirm that MSNs incorporation enhanced mechanical performance under both compression and tension (n ≥ 3, α = 0.05).

TABLE 1.

Mechanical parameters AM mats without and with MSNs obtained in compression and tension using a dynamic mechanical analyzer and an Instron tensile tester, respectively.

	Compressive parameters			Tensile parameters
	Modulus [MPa]	Max stress [MPa]	Max strain [%]	Modulus [MPa]	Max stress [MPa]	Max strain [%]
AM	−7.27 ± 0.36	0.349 ± 0.00001	−76.0 ± 1.3	0.347 ± 0.044	0.258 ± 0.015	39.3 ± 1.0
AM_MSN	−9.59 ± 0.36	0.420 ± 0.036	−54.7 ± 1.7	0.813 ± 0.072	0.466 ± 0.070	35.5 ± 0.3

Accordingly, while stiffness and strength increased, maximum strain decreased for AM_MSN mats in both compressive (−76.0 ± 1.3% to −54.7 ± 1.7%, p ≤ 0.000001) and tensile (39.3 ± 1.0% to 35.5 ± 0.3%, p < 0.001) loading, suggesting a reduction in ductility. This trade‐off is well documented for nanoparticle‐reinforced hydrogel systems, where added fillers restrict polymer chain mobility by acting as physical cross‐linking sites, thereby enhancing rigidity while promoting stress concentration and reduced extensibility [38, 47]. The tensile modulus more than doubled (from 0.35 to 0.81 MPa), whereas the compressive modulus increased by ∼30%. This modest reinforcement raised the deformation and tearing thresholds, providing a more robust, resilient and easier‐to‐handle network, an advantageous characteristic for maintaining scaffold integrity and enabling controlled release applications [43].

A related study investigated the mechanical reinforcement of AM microfibers through dual cross‐linking via ionic gelation with CaCl₂ and UV‐induced covalent bonding. This significantly reduced swelling in PBS as cross‐linking density increased and improved material stability, as evidenced by decreased weight loss [48]. In contrast, the present system only relies on UV cross‐linking of AM (3 wt.%) yet shows comparable swelling values despite the absence of additional ionic crosslinks, suggesting that UV cross‐linking alone is sufficient to generate a stable network at this concentration. Moreover, the incorporation of MSNs resulted in reduced material loss, pointing toward enhanced structural stability. Although MSNs do not participate directly in covalent UV cross‐linking, silica‐polymer hydrogen bonding and polymer adsorption at the MSN surface are known to restrict chain mobility and introduce physical junctions. These interactions can decrease mesh size and increase modulus, thereby effectively increasing the apparent cross‐link density [38, 39, 49]. A similar effect has been reported previously in MSN‐chitosan nanofiber composites, where MSN incorporation reduced liquid uptake and increased structural integrity, as demonstrated by a lower weight loss and enhanced mechanical performance [50].

Reinforcement mechanisms have been widely explored in other nanocomposite hydrogel systems processed by electrospinning or freeze‐drying. For example, a composite gelatin‐SA hydrogel reinforced with carboxyl‐functionalized graphene nanosheets showed enhanced cross‐linking and reduced pore size as the graphene content increased up to 2 wt.%, whereas higher loadings led to a decline in performance [51]. At the optimal filler concentration, compressive strength rose from 25 to 380 kPa and tensile strength from 11 to 150 kPa, with the elastic modulus increasing from 0.03 to 1.2 MPa. Concurrently, elongation at break decreased from 14% to 5%, which is in line with the property of a stiffer material. In a different approach, SA‐based hydrogels have achieved mechanical reinforcement through ionic cross‐linking, such as calcium chloride treatment in a semi‐interpenetrating double network with poly(acrylamide). Although this mechanism differs from the UV‐based cross‐linking employed in the present work, it highlights the broader tunability of alginate networks for mechanical performance, with tensile strengths up to ∼730 kPa through network reinforcement [52].

Of particular relevance to the present system, a study on poly(vinyl alcohol) (PVA)–hydroxyapatite composite hydrogels demonstrated a similar compression profile. Simulation demonstrated that their hydrogels presented a nonlinear, exponential compression behavior attributed to a 3D network structure formed via hydrogen bonding between PVA and hydroxyapatite, acting as inorganic reinforcement. As the PVA content was tripled (up to 6 wt.%), the compressive strength increased from 0.17 to 0.43 MPa, while the modulus improved from 0.07 up to 2.9 MPa, the samples were compressed to 40% of their original size, with a compressive strain ratio chosen at 60%. The viscoelasticity observed is relevant for soft tissue applications and was explained by the rearrangement of flexible PVA chains under load, facilitated by microcrystalline and amorphous regions [53]. Similarly, the electrospun AM hydrogels reinforced with MSNs exhibited a matching exponential compression profile, confirming the viscoelastic character of the network.

Finally, a study on layer‐by‐layer electrospun PVA‐poly(acrylic acid)‐graphene oxide membranes engineered to mimic cartilage collagen fibrils reported an increased compressive modulus of 20.4 ± 2.5 MPa, validating the potential of fibrous stacking and synthetic nanofillers to enhance mechanical strength [54]. While such advanced multilayer systems with synthetic components can achieve high levels of functionality, the MSN‐reinforced electrospun hydrogels presented here offer a simpler and more biocompatible alternative based on natural polymers and UV cross‐linking. Under the conditions tested, mechanical characterization places these composites within the range reported for other fibrous nanocomposite systems. Importantly, these results were obtained using a straightforward electrospinning process and nanoparticle dispersion by probe sonication. Collectively, these attributes support the use of these fibrous hydrogels as mechanically robust, application‐relevant platforms for the delivery of bioactive agents.

2.4. Drug Loading Capacity and Release Profile

Chlorhexidine (CHX) loading and release were first tested with MSNs alone before studying AM and AM_MSN mats. Thermogravimetric analysis (TGA) and nitrogen physisorption results indicated good loading efficiency, Fourier transform infrared spectroscopy (FT‐IR) highlighted the presence of CHX, and ultraviolet–visible spectroscopy (UV–vis) was used to measure the CHX release concentration. These results are illustrated in Figure 3.

FIGURE 3.

CHX loading and release behavior from MSNs as studied using (A) TGA data of pure MSNs, pure CHX, and MSNs loaded with 0.2, 0.5, 1.0, and 5.0 wt.% CHX; (B) Nitrogen physisorption analysis showing pore size distribution and corresponding nitrogen adsorption–desorption isotherms before and after CHX loading (inset). (C) FT‐IR spectra of pure CHX and MSNs at different CHX loading concentrations and (D) cumulative CHX release over 7 days from MSNs loaded with 0.2, 0.5, 1.0, and 5.0 wt.% CHX.

TGA measurements were done to analyze the loading efficiency of different concentrations of CHX – 0.2, 0.5, 1.0, and 5.0 wt.% in the MSNs, and the results are illustrated in Figure 3A. These CHX concentrations were tested since 5 wt.% CHX is the concentration required to make a saturated ethanol‐CHX solution in order to have the highest loading capacity, and 0.2 wt.% is the CHX concentration present in commercial mouthwashes [55]. 0.5 and 1 wt.% CHX loading concentrations were also tested to study the CHX loading and release behavior. TGA profiles in Figure 3A for MSN loaded with CHX indicated major weight loss around 185°C and 360°C to 480°C, which corresponded well with the pure CHX melting and further decomposition temperatures [56, 57]. CHX was loaded into the MSNs in amorphous form which can be confirmed with the absence of the sharp endothermic peak at around 185°C for CHX‐MSN, but present in pure CHX (result in Figure S1). An increasing weight loss was observed ranging from around 6% to 40% with increasing CHX loading concentration inside the MSNs [49].

Nitrogen physisorption isotherms in Figure 3B exhibit a type IV isotherm with an H1‐type hysteresis loop according to IUPAC classification, indicating the presence of mesopores in which adsorbate filling occurs via capillary condensation, as previously observed [58]. The steady increase in adsorbed nitrogen between P P_o ⁻¹ = 0.2–0.8 indicates a uniform mesopore size distribution, while the pronounced nitrogen uptake at higher relative pressures corresponds to the presence of interparticle voids formed by the packing of the MSNs. The pore size distribution derived from the desorption branch confirms this, showing a sharp peak at 2 nm corresponding to intrinsic mesopores and a broader distribution between 10 and 30 nm associated with interparticle spacing. The TEM micrograph (in Figure S2) further corroborates the presence of isotropically distributed 2 nm mesopores, consistent with previous studies [59]. After 5 wt.% CHX loading, the slope of the sorption isotherms reduces, indicating partial filling of the mesopores with CHX and the reduction in maximum adsorbed volume is consistent with CHX occupying the interparticle voids [60]. The pore volume contribution around 2 nm becomes nearly negligible, demonstrating that these mesopores are effectively filled after CHX loading. The diminished intensity of the broad 10–30 nm feature, together with its slight shift toward smaller pore sizes, indicates reduced effective interparticle spacing due to CHX adsorption along the void walls. This interpretation is supported by the quantitative Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) analyses. As summarized in Table 2, upon 5 wt.% CHX loading, the BET surface area decreases by approximately 62%, from 847 to 319 m² g⁻¹, while the BJH pore volume decreases by about 46%, from 2.23 to 1.20 cm³ g⁻¹. These reductions confirm a substantial loss of accessible porosity upon CHX loading, consistent with CHX being present within the MSN pore network.

TABLE 2.

Total surface area and pore volume of MSNs before and after loading with 5 wt.% CHX, as determined by nitrogen physisorption and derived using the BET and BJH models, respectively.

	MSN	MSN‐C_5.0
Surface area (m2 g−1)	847.6 ± 1.8	318.8 ± 1.4
Pore volume (cm3 g−1)	2.23	1.20

FT‐IR results shown in Figure 3C complemented nitrogen physisorption and TGA measurements to confirm the presence of CHX after loading inside the MSNs. Characteristic CHX transmittance bands were observed around 2934 and 2859 cm⁻¹, corresponding to CH₂C‐H antisymmetric and symmetric stretching, respectively. Additional bands assigned to aromatic C═C stretching appeared at 1414, 1492, 1534 cm⁻¹, confirmed the presence of CHX in MSNs [61]. These aromatic features are highlighted with an orange rectangle. The characteristic Si‐O‐Si transmittance peak of MSN was observed at around 800 cm⁻¹, along with a broad band centered at 1025 cm⁻¹, highlighted with a violet rectangle [61, 62].

The release behavior of CHX from MSNs was evaluated over a seven‐day period to simulate the functional lifetime of a general wound dressing. The cumulative release profile is presented in Figure 3D, and the corresponding daily release concentrations are provided in Figures S3 and S4. All formulations exhibited an initial burst release phase followed by a transition to a slower, diffusion‐dominated regime. For MSN‐C_5.0, the burst phase extended over the first four days, with daily release amounts of approximately 5334, 4783, 1659 and 579 µg ml^{− 1} on days 0–3, respectively, followed by a marked decrease to 128 µg ml^{− 1} on day 4. Thereafter, the release rate stabilized, reaching approximately 67 µg ml^{− 1} by day 7. MSN‐C_1.0 exhibited a three‐day burst release, with daily release values of approximately 3115, 980, and 500 µg ml^{− 1}, followed by 202 µg ml^{− 1} on day 3. The release then transitioned to a more controlled regime, with a concentration of 44 µg ml^{− 1} measured on day 7. In the case of MSN‐C_0.5, a burst release occurred over the first two days, yielding approximately 2318 and 583 µg ml^{− 1}, respectively, followed by a gradual decline to 316 µg ml^{− 1} on day 2 and a sustained release phase, thereafter, reaching 30 µg ml^{− 1} by day 7. MSN‐C_0.2 showed only a brief burst release within the first day, releasing 1769 and 331 µg ml^{− 1} on days 0 and 1, respectively, followed by a steady release phase from day 2 onward, culminating at 21 µg ml^{− 1} on day 7. The progressive shortening and attenuation of the burst phase with decreasing CHX loading reflects differences in drug distribution within the MSNs architecture. At higher CHX loading concentrations, a larger fraction of CHX is likely adsorbed on the external particle surface in addition to being confined within the mesopores, enabling rapid early release. At lower loadings, CHX is predominantly localized near mesopore entrances and external surface sites, resulting in a more compressed burst and earlier transition to diffusion‐limited release.

To quantify these trends and enable an objective comparison of all MSN‐C formulations, the cumulative release data were fitted using the Weibull model, an empirical function that captures the statistical distribution of release events across a heterogeneous population of diffusion pathways. Its suitability for porous carriers arises from its ability to represent systems where drug molecules occupy sites of varying accessibility and pore depth, without assuming a single dominant transport mechanism. Its superior descriptive performance relative to conventional kinetic models has been widely reported across diverse drug delivery platforms [63, 64].

The detailed parameters obtained from the Weibull model fitting are provided in Table S2. The Weibull model demonstrated consistently strong agreement between observed and predicted values (adjusted R² = 0.982–0.997, RMSE = 7.2–152.8 µg ml⁻¹). Q_max increased proportionally with drug loading, from 2.695 µg ml⁻¹ for MSN‐C_0.2 to 13.276 µg ml⁻¹ for MSN‐C_5.0, confirming efficient and dose‐dependent CHX encapsulation. The scale parameter α did not follow a simple monotonic trend across the series. While MSN‐C_5.0 and MSN‐C_0.2 exhibited comparatively moderate and short characteristic timescales (α = 6.13 h and 0.55 h, respectively), MSN‐C_1.0 and MSN‐C_0.5 showed disproportionately high values (33.78 and 14.90 h, respectively). These elevated α values indicate that for these intermediate loadings the true release plateau was not fully reached within the 7‐day window, reducing the robustness of α estimation and suggesting that a substantial fraction of CHX remained confined in deeper mesopore regions. The shape parameter β was <1 for all formulations and declined progressively with decreasing drug loading, from 0.376 for MSN‐C_5.0 to 0.115 for MSN‐C_0.2. This behavior is indicative of Fickian diffusion with stretched‐exponential profile, reflecting a pronounced early burst attributed to the rapid mobilization of surface‐bound CHX, followed by a strong deceleration as diffusion originates from progressively deeper and less accessible mesopore sites [60, 63]. The decline in β at lower loadings reflects a growing dominance of near‐surface drug populations, which intensifies the burst character and compresses most of the release into the earliest phase of the profile.

Gram‐positive staphylococci are predominant skin commensals and a frequent cause of wound‐associated infections [65]. Reported minimum inhibitory concentration (MIC) values for CHX are generally situated at the low end of the spectrum between 1 and 10 µg ml⁻¹. In particular, Staphylococcus epidermidis, which was included in the present study, has been reported to display MIC values in the range of 2–8 µg ml⁻¹ depending on strain and experimental conditions [66, 67, 68, 69]. The MIC of CHX against various oral microorganisms has been reported to range between 2 and 80 µg ml^{− 1} [70]. The daily release from MSN‐C_0.2 fell within this MIC range from day 2 onward, while MSN‐C_0.5, MSN‐C_1.0, and MSN‐C_5.0 reached this range on days 3, 4, and 5, respectively. Consequently, all formulations achieved CHX levels consistent with the reported MIC for Staphylococcus epidermidis, indicating that even the lowest loading provides sustained antimicrobial‐relevant release.

The above results highlight the release behavior of individual MSNs; however, the target application in this research required their integration into macroscopic materials. Therefore, the subsequent experiments focus on CHX release from MSN‐containing mats, allowing assessment of how embedding the particles into a fibrous matrix influences release kinetics. The next important step was the CHX loading in AM and AM_MSN mats. CHX loading was done after synthesis of the mats to ensure similar loading conditions and prevent activity loss during the electrospinning. FT‐IR results showed MSN Si‐O‐Si bands for AM_MSN in Figure 4A with increased transmittance intensity compared to AM in the violet‐highlighted regions, indicating successful incorporation of MSNs. Further characteristic peaks for the cross‐linked AM backbone were identified in both AM and AM_MSN at 1600 cm⁻¹ (COO⁻), 1344 cm⁻¹ (C‐H), 1100 cm⁻¹ (O‐C‐O), and 840 cm⁻¹ (C‐H) [71]. The presence of CHX in AM and AM_MSN is further supported by the C═C stretching region around 1500 cm⁻¹ and highlighted in the orange rectangle [61]. Derivative weight change curves from the TGA of CHX‐loaded AM and AM_MSN mats are illustrated in Figure 4C,D. The first peak around 50°C corresponds to the evaporation of residual ethanol used as the cross‐linking medium and the CHX loading solvent [72]. The AM fibers exhibit two main degradation steps at around 250°C and 400°C, associated with cleavage of glycosidic (C‐O‐C) and carboxylate (COO⁻) groups and decomposition of methacrylate bonds, respectively [73]. The AM_MSN mats exhibit the same two degradation events, confirming that the thermal decomposition pathway of the AM matrix is preserved. However, the relative intensity of the second degradation peak is lower in AM_MSN mats compared to AM, due to the presence of thermally stable MSNs, which reduce the overall weight loss of the composite. A shoulder around 200°C, also visible in the weight derivative curve of pure CHX (Figure S5), indicates the presence of CHX within both fiber systems. This feature is more prominent in AM_MSN mats than in AM mats, suggesting higher CHX uptake in the composite. This may be attributed to competitive CHX adsorption between the AM matrix and the MSNs during immersion in the CHX solution. The mesoporous structure of MSNs, characterized by high surface area and large pore volume, and their negative surface charge enhance their affinity for CHX. The zeta potential measurements (Figure 4B) support this interpretation: AM, MSN, and the composite AM_MSN all exhibit negative surface charges, whereas CHX exhibits a positive charge due to protonated amine groups (═NH₂ ⁺) under physiological conditions [74, 75, 76, 77, 78]. The negative charge of AM originates from deprotonated carboxylate groups (–COO⁻), while the negative charge of MSNs arises from deprotonated silanol groups (≡Si‐O⁻). These electrostatic interactions promote CHX adsorption onto both AM and MSN surfaces, but the substantially larger accessible surface area of the MSNs results in greater CHX uptake in the AM_MSN mats [56]. Consequently, the incorporation of MSNs significantly enhances the CHX loading capacity of the AM_MSN composite mats.

FIGURE 4.

(A) FT‐IR spectra highlighting the successful modification of AM after incorporation of MSN in AM_MSN and 5 wt.% CHX in AM‐C_5.0 and AM_MSN‐C_5.0. (B) Zeta potential results for AM, MSN, CHX, and AM_MSN in a neutral environment mimicking the CHX release environment. TGA results indicating derivative weight percentage for different concentrations of CHX loading in (C) AM and (D) AM_MSN.

The cumulative CHX release from AM and AM_MSN mats was evaluated over a period of seven days, with results from Day 1 onward presented in Figure 5 and Day 0 release concentrations provided in Figures S6 and S7. This timeframe was selected to reflect the typical replacement interval of wound dressings.

FIGURE 5.

CHX cumulative release profiles for (A) AM and (B) AM‐MSN mats loaded with different CHX concentrations over a period of 7 days.

For AM mats, CHX release was concentration‐dependent, with higher CHX loadings yielding proportionally greater release (Figure 5A), confirming that CHX adsorption onto the AM fibers scales with the initial loading concentration. AM‐C_5.0 released approximately 8, 8, 5, 3, 2, and 4 µg ml^{− 1} from Days 1 to 6, with the release saturating at around 1.5 µg ml^{− 1} on Day 7. AM‐C_1.0 released approximately 3, 2.6, 3, 2, 2, and 1 µg ml^{− 1} during Days 1–6, with a final concentration of 0.2 µg ml^{− 1} on Day 7. Both formulations displayed higher initial release rates, indicative of the desorption of loosely adsorbed CHX molecules, followed by slower release governed by electrostatic interactions between the negatively charged AM fibers and the positively charged CHX molecules [79], which strengthened at higher CHX loadings. For AM‐C_0.5, the CHX release concentrations were approximately 2.6, 1, 2, 0.5, 0.3, and 0.1 µg ml^{− 1} from Days 1 to 6, with the release saturating around 0.3 µg ml^{− 1} on Day 7. AM‐C_0.2 exhibited the lowest CHX release, recording approximately 0.3 and 0.2 µg ml^{− 1} on Days 1 and 2, respectively, followed by negligible release thereafter. These results indicate that higher CHX loadings promoted an initial burst release, whereas lower loadings favored a more controlled and sustained release profile. To quantitatively compare these loading‐dependent behaviors and extract kinetic descriptors, the cumulative release data were fitted using the Weibull model, which provided good agreement across all AM‐C formulations (adjusted R² = 0.948–0.988, RMSE = 0.009–0.685 µg ml⁻¹). The detailed parameters obtained from the Weibull model fitting are provided in Table S2. Q_max scaled proportionally with CHX loading concentration, from 0.657 µg ml⁻¹ for AM‐C_0.2 to 44.433 µg ml⁻¹ for AM‐C_5.0, confirming dose‐dependent encapsulation within the AM matrix. The scale parameter α increased with CHX loading (31.14 h for AM‐C_0.2 to 118.66 h for AM‐C_5.0), reflecting slow CHX diffusion through the densely cross‐linked AM network. The incomplete plateau attainment for AM‐C_5.0 and AM‐C_1.0 within the 155.5 h window (71.5% and 82.0% of Q_max, respectively) further indicates that release from higher‐loaded fibers extends considerably beyond the observation period. The shape parameter β was close to or above unity across all formulations (0.847–1.165), indicating release behavior that transitions from near first‐order to mildly sigmoidal kinetics. These values suggest the absence of a pronounced burst release, with drug liberation occurring in a gradual manner followed by a smooth deceleration over time. Such release characteristics are consistent with diffusion and swelling‐controlled transport mechanisms typically observed in hydrogel or polymeric matrix systems [63, 64]. This contrasts sharply with the stretched‐exponential behavior observed for MSN‐C formulations (β = 0.115–0.376), where the isotropic mesopore architecture produced a pronounced burst character. The quantitative shift in β between the two systems thus provides mechanistic evidence that the AM matrix governs CHX release in AM‐C formulations, suppressing burst release and enabling a more gradual and sustained delivery profile. Regarding therapeutic relevance in dental applications, cumulative CHX release from AM mats was within the reported MIC range of 2–80 µg ml^{− 1} from Day 1 onward for all formulations except AM‐C_0.2 [70]. These findings demonstrate that AM mats can serve as tunable carriers for localized drug delivery, with release profiles modulated by CHX loading level and fiber–drug electrostatic interactions [4, 20].

The CHX release behavior from AM_MSN composites displayed distinct trends (Figure 5B), reflecting the combined influence of both MSNs and AM fibers on drug transport. This is consistent with findings in previous studies reporting that the drug release from hybrid drug delivery systems is governed by synergistic interactions between nanoparticles and the surrounding polymer matrix [58]. AM_MSN‐C_0.5 and AM_MSN‐C_0.2 displayed gradually increasing cumulative release profiles that did not reach a definitive plateau within the 7‐day experimental window, rising steadily from 3.12 to 5.96 and 3.84 to 7.24 µg ml^{− 1} respectively. In contrast, AM_MSN‐C_5.0 showed a markedly attenuated release profile, with cumulative concentrations ranging only from 1.07 to 1.12 µg ml^{− 1} across the entire 7 day period, indicating strong retention of CHX due to deeper mesopore occupancy combined with the diffusional resistance imposed by the surrounding AM network. AM_MSN‐C_1.0 exhibited an atypical stepwise release pattern, remaining essentially constant at 0.029 µg ml^{− 1} for the first 3 days before rising to 0.290 µg ml^{− 1} and subsequently to 0.456 µg ml^{− 1}. This behavior likely reflects an initial lag phase during which water penetration and AM swelling were required before CHX could begin diffusing out of the composite. The comparatively higher release observed for AM_MSN‐C_0.2 cannot be attributed solely to weaker electrostatic interactions at lower drug loadings. Differences in drug distribution likely also play an important role. At low loading, CHX is more likely located in easily accessible regions, such as the external surfaces of MSNs and AM fibers or near MSN pore openings, facilitating faster release. At higher loading, a larger fraction of CHX occupies deeper mesopores or interacts more strongly with both the MSN surface and alginate matrix, resulting in slower release.

Because the AM_MSN composites exhibited non‐monotonic, lag‐phase, and continuously rising release profiles that do not satisfy the assumptions of the Weibull function, model fitting would not yield physically meaningful kinetic parameters. For this reason, Weibull model fitting was not conducted for the AM_MSN‐C series. However, these trends indicate that distinct retention and transport mechanisms operate within the composites, pointing toward a more complex interplay between CHX, MSNs and the AM network, which requires a closer mechanistic examination of the interactions governing CHX binding and release. CHX, being positively charged, interacts electrostatically with two negatively charged components: the carboxylate groups of AM and the silanol groups on MSN pore surfaces (schematic in Figure 6) [79]. CHX confined within narrow mesopores forms multiple contact points with silanol groups, resulting in stronger binding than CHX associated with the AM surface alone. Consequently, CHX must first detach from the MSN pores and then diffuse through the alginate matrix, where additional electrostatic interactions further slow its movement. This sequential desorption‐diffusion process explains the markedly lower cumulative release observed for AM_MSN composites compared to AM‐C mats, where only AM‐CHX interactions are present.

FIGURE 6.

Schematic illustrating the different components of the AM_MSN‐C composite mat and their distribution potentially impacting the CHX drug release behavior, along with the molecular structure with their respective positive and negative ions depicting the interaction between the three different components.

Overall, incorporating MSNs into the AM matrix suppresses the burst CHX release characteristic of standalone MSNs by introducing an additional diffusion barrier through the alginate network. At the same time, the combined effects of electrostatic interactions and mesopore‐based entrapment enhance CHX retention, particularly at higher drug loadings, while promoting more sustained release at lower loadings. These findings indicate that CHX release is governed not only by molecular‐scale interactions between CHX, AM, and MSN surfaces but also by the spatial organization of MSNs within the AM fibers. Figure 1C,D shows that MSNs are partially agglomerated within the AM matrix, which may further influence CHX transport and release behavior. However, CHX release experiments performed on MSNs alone demonstrate highly reproducible CHX release profiles despite particle agglomeration, indicating that aggregates behave as consistent diffusion domains rather than introducing variability in release kinetics. Their primary influence is therefore likely related to spatial heterogeneity within the fibers. MSN‐rich regions within the fibers may lead to localized differences in CHX concentration and diffusion pathways: aggregates embedded deeper within the fiber are expected to release CHX more slowly because of longer diffusion distances through the alginate matrix, whereas those located near the fiber surface may contribute to relatively faster local release. Although these microscale effects average out at the macroscopic level, improving MSN dispersion within the AM matrix remains an important direction for future optimization.

For CHX‐loaded AM_MSN mats, cumulative CHX release within the MIC range was achieved from Day 1 for both AM_MSN‐C_0.5 and AM_MSN‐C_0.2. Among these, AM_MSN‐C_0.2 represents a more favorable formulation due to its lower CHX loading concentration while still maintaining therapeutically relevant release levels. Since wound dressings require different changing frequencies, which can range from daily to a couple of weeks as per the application, the AM and AM_MSN systems in this research together provide a versatile set of release profiles suitable for different clinical scenarios. While AM mats offer higher short‐term release at higher loading concentrations, AM_MSN mats provide extended CHX release at lower CHX loadings and improved mechanical properties. FT‐IR spectra (Figure 7) confirmed the presence of CHX in both AM_MSN and AM mats after the 7‐day CHX release study. Distinct CHX signals, particularly those associated with C═C stretching vibrations, remained clearly detectable for AM_MSN‐C_5.0 and AM_MSN‐C_1.0, confirming CHX retention despite the negligible release observed for these formulations compared to other AM and AM_MSN mats. These findings underscore the ability of the AM_MSN system to provide a more controlled and predictable CHX release, highlighting its potential as a platform for advanced wound‐dressing applications.

FIGURE 7.

Transmission FT‐IR spectroscopy results before and after 7 days of CHX release at different loading concentrations from AM and AM_MSN. The presence of CHX can still be observed in the mats after the drug release period. (A–D) represent 5.0, 1.0, 0.5, and 0.2 wt.% CHX loading respectively in both AM and AM_MSN mats before and after 7 days of CHX release.

2.5. Cytocompatibility of Chlorhexidine‐Loaded Hydrogels

The cytocompatibility of the system was investigated to ensure its suitability for biomedical applications. UV‐cross‐linked methacrylated alginate mats with and without MSN and CHX were evaluated by culturing the cells in direct contact with the hydrogels for 1, 3, 5, and 7 days and measuring the metabolic activity of skin fibroblasts with a commercial MTT assay.

Initially, fibroblasts cultured in the presence of unloaded AM and AM_MSN mats without pre‐drying exhibited reduced viability on day 1 (0.70 ± 0.12% and 0.79 ± 0.09%, respectively), indicating a mild cytotoxic effect. Interestingly, cell viability gradually improved over time, reaching non‐toxic levels by day seven of culture (0.97 ± 0.07% and 0.98 ± 0.1%, respectively). This suggested that the initial toxicity was transient and possibly linked to residual ethanol retained in the mats during photo‐cross‐linking, which may have leached out over time in culture.

Supporting this hypothesis, the pre‐dried AM and AM_MSN mats and pure MSN samples (that were not exposed to ethanol) displayed no detectable cytotoxicity from day 1 onward. Following this, an additional overnight drying step at 37°C was introduced before cell exposure. After this modification, a statistically significant increase in cell viability was observed in all time points of culture compared to the non‐dried samples, except on day 7, where viability was already within the non‐toxic range in both conditions (Figure 8A). Two‐way ANOVA confirmed significant effects of predrying and incubation time (p < 0.001), as well as a significant interaction between both factors (p < 0.001), indicating that the impact of predrying on cell viability evolves over time (Table S3).

FIGURE 8.

Survival rate of NHDF‐Ad cells cultured in direct contact with (A) unloaded AM and AM_MSN mats, with and without the pre‐drying step, as well as an equivalent solution of pure MSN nanoparticles (MSN) and CHX‐loaded mats at CHX concentrations of 0.0, 0.2, 0.5, 1.0, and 5.0 wt.% labeled accordingly as AM‐C_wt.% (B) and AM_MSN‐C_wt.% (C), over culture periods of 1, 3, 5 and 7 days (n = 3). Cell viability was assessed using the MTT assay, with results expressed as a fraction relative to the untreated control (1.0 = 100% viability). Asterisk notation indicates levels of statistical significance as follows: p > 0.05 (ns), p < 0.05 (^*), p < 0.01 (^**), p < 0.001 (^***), and p ≤ 0.0001 (^****).

Four concentrations of CHX were tested: 0.2, 0.5, 1.0, and 5.0 wt.%. As illustrated in Figure 8B,C, cell viability decreased significantly with increasing drug concentration. Two‐way ANOVA revealed that cell viability was strongly influenced by CHX concentration and incubation time (p < 0.001), with a smaller but significant effect of material (p < 0.001). Significant interactions between CHX and time (p < 0.001) and between material and CHX (p < 0.001) indicate that the cytotoxic effect of CHX is both time‐ and material‐dependent (Table S4).

The overall trend was similar for both AM and AM_MSN mats, except at 0.2 wt.% CHX, where cells exposed to AM mats demonstrated partial recovery of viability, whereas those exposed to AM_MSN‐C_0.2 mats did not. This suggests that in AM mats, at the lowest concentration, CHX may have been largely removed through media changes, allowing for improved cell recovery during culture. In contrast, the cells exposed to AM_MSN mats did not recover, potentially due to nanoparticle‐associated retention, resulting in sustained cytotoxic exposure and reduced viability. This outcome reflects the dose‐dependent cytotoxicity of CHX [80], rather than a limitation of the carrier system itself. Indeed, the mesoporous structure of MSN enabled a sustained release consistent with scientific literature on MSN‐based delivery systems for biomedical use [81, 82].

Although cell viability values were comparable between AM and AM_MSN mats at equivalent CHX loads, the mechanism of action was probably different. Based on the release study, CHX remained electrostatically bound to AM and MSN, which prevented its release. The presence of high unreleased CHX present on the AM_MSN surface likely contributed to lower cell viability at higher CHX loading concentrations. In contrast, the cytotoxicity in AM mats can be attributed to the progressive release of CHX over time, as observed in the release study. For comparison, the equivalent dose of pure CHX applied directly to the cells (without a hydrogel) followed the toxicity trend of loaded AM mats in all cases except for 0.2 wt.% CHX, where cells were able to recover significantly more than when loaded in AM. The findings highlight the importance of drug selection: while CHX was employed here as a proof‐of‐concept antimicrobial, translation of this platform would require alternative agents with improved therapeutic indices for wound healing applications.

Viability was simultaneously assessed using a Live/Dead assay, a fluorescent staining in which live cells are labeled green (calcein AM dye, CA‐AM) and dead cells fluoresce red (ethidium homodimer 1, EthD‐1), as presented in Figure 9. Observations via inverted microscopy and confocal laser scanning microscopy corroborated the MTT assay results, visually confirming the improvement in cell viability over time when there was no pre‐drying step. Notably, the mats that underwent the overnight drying step significantly promoted cell viability and healthier fibroblast morphology compared to non‐dried ones, restating the conclusion that residual solvent was the likely cause of initial cytotoxicity. Moreover, high cell viability was observed when fibroblasts were exposed to pure MSN nanoparticles at concentrations equivalent to those present in the AM_MSN hydrogel mats, suggesting their biocompatibility and confirming that the observed effects were not due to the nanoparticles themselves. The apparent gaps observed in Day 7 Live/Dead fluorescence images were not visible prior to staining. By this stage (effective culture day 8, given that the cells are seeded one day before hydrogel application), fibroblast confluency was high, which may have contributed to localized detachment or disruption during removal of the hydrogel discs and subsequent staining. Calcein AM staining has been reported to induce cell clustering or localized detachment under conditions of high confluency or stress in some cell types [83, 84, 85]. These observations suggest that the visible discontinuities are artifacts of staining and handling rather than indicators of cytotoxicity.

FIGURE 9.

Cell viability at 1‐, 3‐, 5‐, and 7‐days following exposure to AM and AM_MSN hydrogels, with and without the pre‐drying step, as well as an equivalent solution of pure MSN, under controlled conditions (37°C, 90% RH, 5% CO₂). Confocal laser scanning microscopy images of the LIVE/DEAD assay on human dermal fibroblasts (NHDF‐Ad). Testing conducted in accordance with the ISO 10993–5:2009 standard.

Additionally, cells exposed to CHX‐loaded hydrogels exhibited increased viability as CHX concentration decreased (Figures 10 and 11), a consistent trend between the MTT data and the LIVE/DEAD assay. Similarly, at the lowest drug loading (0.2 wt.%), AM‐loaded mats supported higher viability and more favorable cell morphology, highlighting the impact of lower drug loading on biocompatibility [62]. Notably, at this lowest concentration, AM‐loaded mats outperformed AM_MSN‐loaded mats, further supporting the sustained release effect of MSN incorporation, which prolonged cellular exposure to CHX and reduced viability accordingly.

FIGURE 10.

Cell viability at 1‐, 3‐, 5‐, and 7‐days following exposure to AM hydrogels, with decreasing amount of CHX (5.0, 1.0, 0.5, and 0.2 wt.%, denoted as ‐C_wt.%), under controlled conditions (37°C, 90% RH, 5% CO₂). Confocal laser scanning microscopy images of the LIVE/DEAD assay on human dermal fibroblasts (NHDF‐Ad). Testing conducted in accordance with the ISO 10993–5:2009 standard.

FIGURE 11.

Cell viability at 1‐, 3‐, 5‐, and 7‐days following exposure to AM_MSN hydrogels, with decreasing amount of CHX (5.0, 1.0, 0.5, and 0.2 wt.%, denoted as ‐C_wt.%), under controlled conditions (37°C, 90% RH, 5% CO₂). Confocal laser scanning microscopy images of the LIVE/DEAD assay on human dermal fibroblasts (NHDF‐Ad). Testing conducted in accordance with the ISO 10993–5:2009 standard.

Finally, to complement the visual observations, the images were processed using ImageJ for a semi‐quantitative assessment in triplicate (Figure 12A). This analysis quantitatively confirmed the improved biocompatibility upon the implementation of the drying step. Regarding the quantification of the cells exposed to CHX‐loaded hydrogels, a cumulative graph was generated (Figure 12B,C) to display the ratio of live to dead cells at the different drug concentrations. It can be seen that mats loaded with 5.0 wt.% CHX consistently presented a high ratio of dead cells (exceeding 50%) in both formulations. Particularly, the MSN‐containing mats displayed a generally higher proportion of dead cells across all concentrations Figure 12C. This suggests that, while AM mats allowed a faster washout of CHX during media changes, which resulted in partial cell recovery, the presence of MSN in the mats prolonged drug retention and release, leading to extended cytotoxic exposure even at lower loadings.

FIGURE 12.

(A) Quantification of living NHDF‐Ad cells cultured in direct contact with unloaded AM and AM_MSN mats, with and without the pre‐drying step, as well as with an equivalent solution of pure MSN. (B) Viability ratio (live/dead cell percentage) for NHDF‐Ad cells exposed to CHX‐loaded AM and (C) AM_MSN mats at concentrations of 0.2, 0.5, 1.0, and 5.0 wt.%, denoted as ‐C_wt.%, over culture periods of 1, 3, 5, and 7 days (n = 3). Cell count was performed using ImageJ on the confocal laser scanning microscopy images of the LIVE/DEAD assay.

Interestingly, a partial recovery of cell viability was observed between days 5 and 7 at intermediate CHX concentrations (0.5 and 1 wt.%). This could be due to a reduction in effective drug exposure over time, as the release rate stabilized and media refreshing diluted the available CHX. Similar delayed recovery effects have been reported in other CHX‐loaded fiber systems, where cells were able to adapt or repopulate after the initial burst phase of exposure [86]. The underlying mechanisms, whether associated with cellular responses, proliferation of surviving cells, or matrix‐associated modulation, require further investigation.

The cytocompatibility results presented in this study are consistent with previous reports on CHX‐loaded electrospun systems. For instance, a study evaluated CHX‐loaded PCL electrospun fibers and reported dose‐dependent cytotoxicity in fibroblasts, with a significant reduction in cell viability at loadings above wt.% [87]. In another study involving loaded PCL, researchers reported acceptable keratinocyte viability at CHX concentrations of 0.5 and 1 wt.%. Differences in biocompatibility across studies may be explained by the hydrophobic nature of PCL, which limits drug loading and release, highlighting the role of matrix properties in modulating release kinetics and cellular response. Given the differences in matrix composition and the initially unknown drug retention profile in the developed system, a broad range of concentrations (0.2‐5.0 wt.%) was explored to capture the spectrum of cytotoxic responses. This enabled the characterization of the release behavior and cellular impact more robustly.

Regarding the incorporation of MSNs, the findings align with previous studies demonstrating that MSNs support cell viability in different polymeric scaffolds, including PCL, PVA, alginate, gelatin, and chitosan, among others. Depending on surface modification, particle size, and dosage, MSN consistently shows low to negligible cytotoxicity on various cell types for biomedical applications [88, 89].

2.6. Antimicrobial Activity

2.6.1. Agar Diffusion Pre‐Screening

CHX loading, release kinetics, and cytotoxicity of the AM and AM_MSN electrospun mats, their antimicrobial efficacy was evaluated against Staphylococcus epidermidis using a disc diffusion assay to explore potential use in skin wound healing. After 18 h of incubation, both mats exhibit a power‐law response in their zones of inhibition, where increasing drug concentrations led to only marginal enhancements of antimicrobial activity (Figure 13). Although the AM mats demonstrated slightly larger inhibition zones compared to AM_MSN mats, this difference was not statistically significant (α = 0.05). Two‐way ANOVA confirmed that antimicrobial activity was significantly influenced by CHX concentration (p < 0.001), while no significant effect of material or interaction between factors was observed (Table S5). Interpreting the size of the resulting inhibition zones requires careful consideration, as several factors can influence drug diffusion through the agar medium. These include the size of the sample disc, the concentration of the released compound, the type and concentration of the agar, and the pH and thickness of the medium [90, 91, 92].

FIGURE 13.

Correlation of loaded CHX into the applied discs (0.0, 0.2, 0.5, 1.0, and 5.0 wt.%) to the zone of inhibition measured from each (n = 3).

2.6.2. Biofilm Formation

To further explore the potential of the CHX‐loaded mats for applications in the oral cavity, additional antimicrobial testing was carried out using the clinically relevant concentration of 0.2% CHX. A semi‐complex multispecies biofilm model was employed to better mimic the oral microbial environment typically targeted by CHX in the prevention and management of biofilm‐associated oral diseases. This model includes a diverse range of bacterial species, Gram‐positive and Gram‐negative, pathogenic and commensal, as well as aerobic and anaerobic organisms, allowing assessment of the system's efficacy across different bacterial types within the same community. The efficacy of 0.2% CHX drug delivery systems (DDS) was evaluated against both biofilm formation on hydroxyapatite disc (HAD) surfaces and planktonic bacterial growth. These effects were compared to those of non‐loaded systems (NDDS) and untreated control biofilms (CB). Over a 7‐day period, biofilm development was assessed by measuring the total viable bacterial load, alongside optical density readings of the supernatant growth media at 600 nm (OD₆₀₀).

As shown in Figure 14A, biofilms grown in the presence of DDS (AM‐C_0.2 and AM_MSN‐C_0.2) exhibited significantly reduced biofilm formation compared to control biofilms (CB) as early as day 1 (p < 0.05). Although DDS‐treated biofilms consistently showed lower viable microbial loads than the CB and the NDDS groups throughout the 7‐day period, the differences were not statistically significant beyond day 1. Notably, on days 5 and 7, AM_MSN‐C_0.2‐treated biofilms exhibited a more pronounced reduction in viable bacterial load compared to AM‐C_0.2, suggesting a more sustained release of CHX from the MSN‐based system, as shown by the data in Figure 3D. These findings from the prolonged antimicrobial activity were further supported by SEM images of the biofilms (Figure 15). SEM imaging revealed more biofilm accumulation on HAD surfaces exposed to NDDS compared to those exposed to DDS. Notably, surfaces treated with AM_MSN‐C_0.2 showed less biofilm accumulation than those treated with AM‐C_0.2 after extended incubation periods.

FIGURE 14.

(A) Total viable bacterial load in biofilms formed on hydroxyapatite discs (HADs) exposed to drug‐loaded systems (DDS: AM‐C0.2 and AM_MSN‐C0.2), non‐loaded systems (NDDS: AM and AM_MSN), or unexposed controls (CB) over 7 days. (B) Optical density at 600 nm (OD₆₀₀) of the supernatant from wells containing DDS, NDDS, or CB over 7 days. (C) Biofilm composition and species relative abundances of biofilms over 7 days. Data points represent the mean of three replicates; error bars indicate the standard error. * denotes statistically significant differences compared to the CB group (p < 0.05), and # denotes statistically significant differences compared to the NDDS group (p < 0.05).

FIGURE 15.

Scanning electron micrographs showing biofilm formation on hydroxyapatite discs (HADs) cultured in media containing AM‐C_0.2 and AM_MSN‐C_0.2 mats loaded with 0.2 wt.% CHX, compared to their unloaded counterparts (AM and AM_MSN), at days 1, 3, 5, and 7 with a 4 µm scale bar.

The inhibitory effect of the CHX was also evident in the supernatant media. As shown in Figure 14B, the optical density at 600 nm (OD₆₀₀) was lower in media containing AM‐C_0.2 and AM_MSN‐C_0.2 compared to their corresponding non‐loaded counterparts (AM and AM_MSN). Although the OD₆₀₀ values remained reduced in the DDS groups, the extent of reduction gradually diminished over time, indicating a decreasing antimicrobial effect in the planktonic phase as CHX release tapered. The reduction in OD₆₀₀ was statistically significant (p < 0.05) only up to day 3 for AM_MSN‐C_0.2 and on day 1 for AM‐C_0.2. The prolonged and significant antimicrobial effect of AM_MSN‐C_0.2 on planktonic bacteria, evident in OD600 values up to day 3, compared to its effect on biofilm bacterial counts, which was significant only up to day 1, can be attributed to the resilient nature of biofilm structures that require higher concentrations of antimicrobial agents for effective disruption [93].

The tested systems exhibited differential inhibitory effects on the microbial community. As shown in Figure 14C, biofilms exposed to the NDDS displayed a composition and relative species abundance profile comparable to that of the CB. In contrast, biofilms exposed to the DDS showed notable changes in composition and species distribution, particularly during the first 3 days. After day 3, all biofilms became increasingly dominated by the A. actinomycetemcomitans species. Changes in the biofilm microbial composition and species abundances are well‐documented with biofilm exposure to CHX. In addition, A. actinomycetemcomitans has been reported for its resilience to CHX.

These findings align with recent studies on CHX‐loaded electrospun mats. For example, electrospun polycaprolactone (PCL) nanofiber membranes loaded with CHX (0.3 and 1.2 wt.%) showed significant antibacterial activity against oral pathogens, with effective inhibition observed at low drug concentrations [94]. Similarly, cellulose acetate‐based electrospun mats loaded with CHX (0.05 – 0.15 wt.%) demonstrated sustained antimicrobial effects against both oral Gram‐negative (P. intermedia, P. gingivalis) and Gram‐positive bacteria (E. faecalis) with improved antibacterial activity at increasing CHX concentrations [87]. The comparable antimicrobial performance of both mats suggests that either system can effectively deliver CHX to inhibit bacterial growth. This outcome highlights the efficacy of prolonged CHX release in reducing biofilm development over time. The more pronounced reduction observed with AM_MSN‐C_0.2, particularly at later time points, suggests that mesoporous carriers may enhance the prolonged antimicrobial effect. This supports the potential application of such DDS in managing infectious diseases where localized, long‐term antimicrobial action is required(1‐4 weeks), such as in periodontitis, peri‐implantitis, or post‐surgical wound care [95, 96]. By maintaining antibacterial activity in both biofilm and planktonic phases, these systems could offer a promising alternative to conventional rinses or systemic antibiotics, reducing the need for frequent application and minimizing systemic side effects, including tooth staining, taste alteration, and mucosal irritation [80, 95, 97, 98]. However, certain bacterial species in the diverse microbial communities may be selectively enriched if they are less susceptible to the antimicrobials released from the system, as observed in this study with A. actinomycetemcomitans [99].

3. Conclusions

In this study, methacrylated alginate‐based electrospun hydrogels were developed and characterized, both with and without mesoporous silica nanoparticle (MSNs) reinforcement, for the controlled release of chlorhexidine (CHX). It was demonstrated that MSNs incorporation significantly enhanced mechanical strength with a preserved viscoelastic behavior, while significantly decreasing the swelling capacity, closely resembling profiles reported in the literature.

The CHX loading and release studies demonstrated that both the drug loading concentration and the electrostatic interactions among the components of the composite mats might play a critical role in governing CHX release behavior. For CHX‐loaded AM mats, the release was primarily concentration‐dependent, with higher CHX loadings resulting in increased drug release. In contrast, CHX‐loaded AM_MSN mats exhibited a distinct release profile governed by electrostatic interactions between the negatively charged AM and MSN matrices and the positively charged CHX molecules, as well as by the depth of CHX localization within the MSN mesopores, which varied according to the CHX loading concentration. At higher CHX loadings (1 and 5 wt.%), CHX molecules were preferentially confined deeper within the MSN mesoporous channels, where multiple electrostatic interactions with silanol groups generated stronger binding affinities and consequently suppressed CHX release, indicating effective drug retention within the matrix, as further corroborated by FT‐IR analysis following the release study. Conversely, at lower CHX loadings (0.2 and 0.5 wt.%), CHX molecules were more likely to be localized near the MSN pore entrances, reducing the extent of multipoint electrostatic interactions and facilitating enhanced CHX diffusion and release from the matrix.

These findings indicate that, in the presence of MSNs, the relationship between drug loading and release is no longer directly proportional but governed by competing retention and diffusion mechanisms. This behavior contrasts with conventional polymer‐only systems and highlights the potential of hybrid architectures to decouple loading from release kinetics, thereby enabling more precise tuning of drug delivery.

Antimicrobial agar diffusion assay confirmed a dose‐dependent efficacy, with both formulations showing comparable inhibition profiles. Biofilm formation and OD₆₀₀ measurements demonstrated a prolonged significant antimicrobial effect for the DDS incorporating embedded MSN.

Cytocompatibility assessment with MTT and Live/Dead assays revealed the importance of post‐processing, in particular ethanol removal via drying, to avoid initial toxicity. Both hydrogels showed excellent biocompatibility. In the presence of CHX, some biocompatibility was detected at the lowest concentration, while the prolonged release from AM_MSN mats slightly increased cytotoxic effects. These observations align with existing literature on CHX‐loaded electrospun mats and MSN‐based drug delivery, confirming the need to balance therapeutic efficacy and cellular compatibility.

Overall, the present work is a promising natural‐based alternative for drug delivery applications, combining tunable mechanical and drug‐releasing properties, effective antimicrobial performance, and biocompatibility through processing and formulation.

4. Experimental Section

4.1. MSN Synthesis

MSNs were synthesized using cetyltrimethylammonium bromide (CTAB, 99%), ethylene glycol (99%), aqueous ammonia (25%), tetraethyl orthosilicate (TEOS, 98%), 3‐aminopropyltriethoxysilane (APTES, 99%), and absolute ethanol (99.9%), all of which were procured from Sigma–Aldrich. The synthesis procedure was conducted following a previously described method, with minor modifications [100]. Briefly, CTAB (7.3 mmol) was dissolved in a mixture of ultra‐pure water (milliQ (MQ), 18.2 µs cm⁻¹, Merck Millipore), ethylene glycol, and aqueous ammonia in a 414:69:10 volume ratio. The solution was stirred vigorously at 50°C for 30 min. Subsequently, TEOS and APTES were introduced into the reaction mixture, yielding a final volume ratio of ultra‐pure water: ethylene glycol:ammonia:TEOS:APTES of 114:19:2.8:5.4:1 (initial ultra‐pure water amount 60 mL). The mixture was then heated to 80°C and stirred for an additional 2 h. After cooling to room temperature, the suspension was stored in a refrigerator. The precipitate was collected via centrifugation (Rotina 35R, Hettich) at 10 000 rpm for 15 min and washed three times using a 1:1 (v v⁻¹) mixture of ultra‐pure water and absolute ethanol and freeze‐dried (Alpha 2–4 LCSbasic, Martin Christ) at −90°C for 48 h at 0.01 mbar pressure. Finally, the dried product was ground and calcined in air at 550°C for 3 h with a controlled heating rate of 1°C min⁻¹ (RHF 1200, Carbolite Gero) to yield the final MSN powder.

4.2. Synthesis of Methacrylated Alginate

Methacrylation was performed as described earlier [101, 102] by adding four equivalents of methacrylic anhydride (MAAH, Sigma–Aldrich) dropwise to a 2 mol% SA solution in demineralized water. During the reaction, methacrylic acid formed and the pH was maintained at 8 using a 5 m NaOH solution. The mixture was stirred for 24 h at room temperature, followed by dialysis (Spectra/Por 4 ‐MWCO 12–14 kD) against distilled water, with two changes per day for 3 days. The modified alginate was then lyophilized. The degree of methacrylation (DS) was determined using ¹H‐NMR with D₂O as solvent [103].

4.3. Characterization of Methacrylated Alginate

The AM used in the present study was synthesized and characterized in‐house. The DS was a parameter obtained by ¹H‐NMR spectroscopy that influences the swelling and mechanical capacity of the cross‐linked material [104]. It was defined as the number of methacrylate moieties introduced relative to the hydroxyl functionalities of alginate. This was quantified using (Equation 1):

$$\mathrm{DS}=\frac{1}{2}\frac{\frac{{\mathrm{I}}_{5.73\mathrm{ppm}}+I_{6.16\mathrm{ppm}}}{2}}{{\mathrm{I}}_{4.58\mathrm{ppm}}+I_{4.97\mathrm{ppm}}}\times 100\%$$ (1)

where the intensities of the methacrylate alkene peaks present at 5.73 and 6.16 ppm were correlated to alginate's characteristic anomeric carbohydrate signals at 4.58 and 4.97 ppm, respective to mannuronic, M‐ and glucuronic, G‐, residues. Given that every unit has two available hydroxyl groups, the value was divided by two. The achieved degree of substitution for alginate was 18.5% per OH‐moiety present originally. This was consistent with previous studies employing similar reaction conditions for methacrylation of alginate [105, 106]. The corresponding ¹H‐NMR spectrum is provided in Figure S8.

4.4. Electrospinning for Fiber Processing

Two polymer solutions were prepared for electrospinning: the first solution (AM) contained 3 wt.% methacrylated alginate, 1 wt.% polyethylene oxide (1 Mo.), and 0.5 wt.% Pluronic F‐127 in a 90:10 (v v⁻¹) water‐absolute ethanol mixture. The second solution (AM_MSN) was identical but included MSN particles at a 1:3 wt.% ratio to alginate and was ultrasonicated using a gun‐tip within the ethanol phase. Both solutions were thoroughly mixed using a speed mixer and sonicated for 10 min at room temperature. Electrospinning was carried out using a custom‐built device developed by the research group at the Department of Materials Engineering, KU Leuven, operated inside a glove box to ensure a controlled environment. The relative humidity (RH) was kept below 15%, and the temperature was maintained at 25°C. A total of 10 mL per solution was electrospun. For AM mats, the needle‐to‐collector distance was set at 26 cm, the polymer flow rate at 0.65 mL h⁻¹ and the applied voltage at 12.5 kV. For AM_MSN mats, these parameters were adjusted to 28.5 cm, 0.45 mL h⁻¹ and 15 kV, respectively.

4.5. Cross‐Linking Protocol

The electrospun mats were cross‐linked using 2 wt.% Irgacure 2959 solution in absolute ethanol. Circular discs (5 cm diameter) were cut from each mat and placed between two parallel glass plates lined with Teflon foil, secured with screws to prevent ethanol evaporation and ensure uniform cross‐linking. The cross‐linking was done for 4 h under 365 nm UV‐A radiation (UV‐A LED Gen 2 Emitter LZ1‐00UV00), with a radiant flux of 800 mW. Afterward, the mats were washed with absolute ethanol to remove any photoinitiator residues and underwent a final overnight wash before drying overnight in an incubator at 37°C.

Initial conditions using 1 wt.% Irgacure and 1 h UV exposure resulted in partial dissolution of the mats, particularly for AM, indicating insufficient cross‐linking. Therefore, the protocol was optimized to improve network stability and reproducibility of gel fraction values. This adjustment was likely necessary due to the fiber structure, limited UV penetration, and use of ethanol as the cross‐linking solvent, which may restrict initiator diffusion into alginate during cross‐linking.

4.6. Electron Microscopy

Scanning electron microscopy (SEM, Nova NanoSEM 450, FEI) was used to analyze the morphological characteristics of the fiber mats with and without MSN and after bacterial biofilm growth. Prior to imaging, the samples were coated with a 5 nm‐thick platinum/palladium layer using a Q150TS sputter coater (Quorum, Kent, UK) to minimize charging effects. Transmission electron microscopy (TEM) examination was done on a JEOL ARM200F Cs‐corrected scanning transmission electron microscope (S/TEM) operated at 200 kV, using high‐angle annular dark field (HAADF) STEM imaging.

4.7. Fiber Diameter and Mat Thickness Analysis

For each group (AM, AM_MSN), at least two independent mats were prepared. From each mat, five randomly selected SEM micrographs (consistent magnification, central field; edges avoided) were acquired. Images were calibrated using the SEM scale bar and analyzed in ImageJ; ≥10 fibers per image were measured perpendicular to the fiber axis (junctions excluded). Diameter was reported as mean ± SD. Mat thickness was measured using a Micro Laser Scanner from Acacia Technology, and the measurements were performed in the Gocator Emulator software, LMI Technologies.

4.8. Nuclear Magnetic Resonance

¹H‐NMR spectra of AM were recorded at room temperature on a 400 MHz spectrometer equipped with a 5 mm four‐nucleus PFG probe using water suppression. Samples were prepared at 5 mg ml⁻¹ in D₂O (5 mg in 1 mL). Acquisition conditions as previously described [103].

4.9. Nitrogen Physisorption

The total pore volume and specific surface area of MSN and CHX‐loaded MSN were characterized through nitrogen physisorption analysis (Autosorb‐iQ instrument, Quantachrome Instruments). Prior to measurement, the samples were subjected to vacuum degassing at 40°C. The specific surface area was determined using the Brunauer–Emmett–Teller (BET) isotherm, while the total pore volume and pore size distribution were calculated from the desorption isotherm using the Barrett, Joyner, and Halenda (BJH) method.

4.10. Gel Fraction and Swelling

Dried electrospun mats were punched into 3 mm diameter discs and weighed before swelling (wd₀) in distilled water for 24 h. The samples were then re‐dried at 40°C and reweighed (w_d). The gel fraction (GF), representing the proportion of cross‐linked polymer, was calculated using Equation (2)

$$GF=\frac{{\mathrm{W}}_{\mathrm{d}}}{{\mathrm{W}}_{\mathrm{d}0}}$$ (2)

To determine the swelling degree (SD), pre‐weighed dry samples were immersed for 24 h at room temperature in distilled water and phosphate‐buffered saline (PBS, Sigma–Aldrich Belgium, pH 7.4), prepared according to the manufacturer's instructions. After swelling, excess solution was removed, and the hydrogels were weighed (w_s). The swelling degree (SD) was calculated using Equation (3)

$$\mathrm{SD}\left(\frac{\mathrm{g}\mathrm{water}}{\mathrm{g}\mathrm{hydrogel}}\right)=\frac{{\mathrm{W}}_{\mathrm{s}}-{\mathrm{W}}_{\mathrm{d}}}{{\mathrm{W}}_{\mathrm{d}}}\times \frac{1}{\mathrm{SG}}$$ (3)

where SG represents the specific gravity of the applied solution (PBS_avg ≈ 1.02–1.03 gcm⁻³ at 20–25°C) [107]. All experiments were performed in quadruplicate.

4.11. Mechanical Tests

Mechanical characterization was performed on hydrogels pre‐hydrated in distilled water for ∼3 h at room temperature on the day of testing, then gently blotted and tested for compression and tensile properties at room temperature. Stress–strain curves were generated to determine the modulus, strain‐at‐break, and stress‐at‐break. The moduli were derived from the linear elastic region of the curve, identified by the constant portion in the first derivative plot. For compression testing, a Dynamic Mechanical Analyzer (DMA) Q800 (Universal Instruments, USA) was used with circular specimens (6 mm diameter, n = 4) clamped in a compression fixture. The test was performed at a constant strain rate of 0.5 N min⁻¹ until sample failure (cracking followed by fracture), with a preload force of 5 mN and an upper force limit of 10 N. For tensile testing, a low‐force Instron 5943 tester with hydraulic clamping was used. Three dog‐bone‐shaped specimens (length = 15 mm, width = 3 mm, gauge thickness = 1 mm) were tested per composition. To prevent slippage, double‐sided tape covered with a paper towel was applied to the clamps, and hydraulic clamping pressure of 0.2 bar was used. The test was performed at a constant strain rate of 5 mm min⁻¹ until failure, with an upper force limit of 10 N. The load was zeroed before each test to ensure accuracy.

4.12. Zeta Potential Measurements

Electrophoretic light scattering (ELS, Litesizer 500, Anton Paar GMBH) was used to conduct particle zeta potential measurements for MSN, CHX, AM, and AM_MSN. All the samples were individually suspended in ultra‐pure water and vortexed for 2 mins to ensure adequate dispersion. 1 mL of the suspension was then transferred into a disposable folded‐capillary cuvette (Omega cuvette, Anton Paar GMBH), and triplicate measurements were conducted.

4.13. Drug Loading and Release Testing

CHX stock solutions were prepared by dissolving CHX in absolute ethanol in concentrations of 0.2, 0.5, 1.0, and 5.0 wt.%. These solutions were used to load the MSNs as well as AM and AM_MSN mats. For MSN loading, 11 mg ml⁻¹ MSNs were dispersed in the respective CHX solutions and agitated on a shaking platform for 24 h. The MSN‐CHX suspensions were centrifuged and the recovered precipitates were dried overnight at 60°C. For mat loading, the CHX solutions were added to AM and AM_MSN mats placed in Petri‐dishes and incubated on a shaking platform for 24 h to ensure homogeneous loading. Excess CHX solution was removed and the mats were dried at 60°C. CHX release studies were conducted by immersing MSN‐C, AM‐C and AM_MSN‐C samples in 500 µL of ultra‐pure water as the release medium, which was replenished at each time point. CHX was quantified using a ultraviolet–visible spectrophotometer (Varioskan, VWR, Leuven, Belgium) at 233 nm wavelength. A calibration curve was established over a concentration range of 6.255 ⁻ to 31.275 µg ml⁻¹ (R² of 0.9988). Weibull model (Equation 4) was used to fit the cumulative CHX release.

$$\mathrm{Q}\left(\mathrm{t}\right)={\mathrm{Q}}_{\max }\left(1-{\mathrm{e}}^{-{\left(\frac{\mathrm{t}}{\alpha }\right)}^{\beta }}\right)$$ (4)

where Q_max represents the asymptotic maximum drug release concentration, α defines the characteristic release timescale, and β governs the profile curvature and release mechanism.

4.14. Antimicrobial Testing

4.14.1. Agar Diffusion

Tryptic soy broth (TSB) and tryptic soy agar (TSA) were used for Staphylococcus epidermidis growth, a Gram‐positive bacterium commonly found on human skin. TSB was made by dissolving per liter: 17 g tryptone, 3 g soy peptone, 5 g NaCl, 2.5 g K₂HPO₄ and 2.5 g glucose. TSA was prepared similarly, with the addition of 15 g agar. The media were sterilized, and TSA was poured into petri dishes to solidify after cooling. S. epidermidis (1 million bacteria per lenticule, Sigma–Aldrich WDCM 000036) was inoculated in 10 mL TSB and incubated for 24 h. The inoculum was diluted to an OD₆₀₀ value of 0.5, and 100 µL was spread per TSA plate divided into three sections designated to a triplicate of each experimental condition: negative control (AM and AM_MSN hydrogels), CHX‐loaded hydrogels (0.2, 0.5, 1.0, and 5.0 wt.%), and positive control (Whatman filter paper soaked in the different CHX concentrations). Plates were incubated at 37°C for 18 h. After incubation, the zone of inhibition was measured with a digital caliper to assess antimicrobial activity.

4.14.2. Biofilm Formation

Fourteen oral bacterial species (Table 3) were cultured on blood agar (Oxoid) supplemented with 5 mg l⁻¹ hemin (Sigma), 1 mg l⁻¹ menadione (Calbiochem‐Novabiochem) and 5% sterile horse blood (E&O Laboratories). Overnight liquid cultures were prepared in brain heart infusion (BHI) broth (Difco) and incubated under either aerobic conditions (5% CO₂) or anaerobic conditions (80% N₂, 10% H₂, 10% CO₂), as appropriate for each species. The 14 bacterial species were allowed to grow together in a Biostat‐B Twin bioreactor using modified BHI medium (BHI‐2). The bioreactor‐derived multispecies communities were then used to grow biofilms on hydroxyapatite discs (HADs) under microaerophilic conditions over a 7‐day period [108]. Electrospun mats were cut into 14 mm diameter circles and placed at the bottom of 24‐well plates, with HADs vertically suspended in the wells using custom‐designed holders, allowing biofilm development on their surfaces. The wells were filled with 2 mL culture medium, enabling the electrospun mats, if loaded with CHX, to release the antimicrobial agent continuously throughout the 7‐day incubation. To ensure consistent CHX exposure, the culture medium was refreshed on days 1, 3, 5, and 7. HADs were retrieved and supernatants were collected on days 1, 3, 5, and 7 for assessment. Biofilm development was monitored over time by measuring the total viable bacterial load and the optical density (OD₆₀₀) of the supernatant (as indication of planktonic bacterial growth) was followed using a Varioskan LUX Multimode Microplate Reader. Biofilms were collected by incubating the discs in trypsin for 45 min at 37°C with shaking at 250 rpm. Viability‐based quantitative PCR (v‐qPCR) was performed using propidium monoazide (PMAxx), as previously described [50]. Three independent replicates were performed for the experiment.

TABLE 3.

Overview of bacterial strains used, grouped by pathogenic or commensal classification. For each strain, the incubation conditions and broth culture medium are indicated.

	Bacterial strain	Incubation conditions	Broth cultures
Periodontal pathogens	Prevotella intermedia ATCC 25611 Porphyromonas gingivalis ATCC 33277 Fusobacterium nucleatum DSM 20482 Aggregatibacter actinomycetemcomitans ATCC 43718	Anaerobic 5% CO2	BHIC BHI
Cariogenic species	Streptococcus mutans ATCC 20523 Streptococcus sobrinus ATCC 20742	5% CO2
Anaerobic oral commensals	Actinomyces viscosus DSM 43327 Actinomyces naeslundii ATCC 51655 Veillonella parvula DSM 2008	Anaerobic	BHI
Oral commensal streptococci	Streptococcus gordonii ATCC 49818 Streptococcus oralis DSM 20627 Streptococcus sanguinis LM 14657 Streptococcus mitis DSM 12643 Streptococcus salivarius TOVE‐R	5% CO2	BHI

Additionally, one HAD from each experimental condition was collected at each time point for SEM. Biofilms were fixed with 2.5% glutaraldehyde in 0.1 m sodium cacodylate buffer (pH 7.4) for 30 min, followed by washing in PBS. Samples were then dehydrated through a graded ethanol series, air‐dried at room temperature, and coated with a platinum/palladium (80/20) layer for imaging [51].

4.15. Cytocompatibility Testing

NHDF‐Ad cells were seeded in 48‐well plates at a density of 4 × 10⁴ cells cm⁻² and allowed to attach overnight. Circular samples (3 mm diameter, n = 3) of AM and AM_MSN mats at concentrations (0.0, 0.2, 0.5, 1.0 and 5.0 wt.%) were sterilized by UV irradiation at 254 nm by running the preset sterilization program of the biosafety cabinet on each side of the mats. The following day, the hydrogels were pre‐treated by rinsing three times with 500 µL sterile‐filtered distilled water and placed on top of the cells. The set‐up was incubated at 37°C, 5% CO₂ and 90% RH for 7 days, with medium refreshed every 2–3 days. The medium was composed of DMEM GlutaMAX with 1 µm pyruvate, supplemented with 1% antibiotic/antimycotic solution and 10% FBS. Cell viability was assessed on days 1, 3, 5, and 7 using both the MTT assay and the LIVE/DEAD Viability/Cytotoxicity Kit (L‐3224, Invitrogen). In the MTT assay, cells were washed with PBS, treated with 100 µL of MTT solution (yellow tetrazolium) for 3 h at 37°C, and formazan crystals were solubilized in 150 µL DMSO. Absorbance of a technical triplicate at 570 nm was measured with the SpectraMax ABSPlus microplate reader (Molecular Devices, USA). For LIVE/DEAD staining, cells were washed, stained with calcein acetoxymethyl ester (calcein AM ex/em 494/517 nm, live cells, green) and ethidium homodimer‐1 (ex/em 528/617 nm, dead cells, red) for 30 min and imaged using a Nikon Eclipse Ti2 inverted fluorescence microscope [109]. Images were processed using ImageJ for semi‐quantitative cell counting to confirm the MTT assay results. Three replicates were tested per condition on each day point. Controls included live and dead controls for both assays, a negative control (loaded hydrogels incubated in medium without cells to exclude false‐positive cytotoxicity), and pure doses of MSN and CHX at concentrations equivalent to their hydrogel incorporation, which served as positive controls.

4.16. Statistical Testing

Statistical analysis was performed using Microsoft Excel and JASP. For experiments involving more than one independent factor (e.g., MTT assays and swelling studies), two‐way analysis of variance (ANOVA) was conducted to assess the effects of the different variables and their interaction.

Following ANOVA, targeted pairwise comparisons between specific formulations were performed using Student's t‐tests (two‐tailed), after assessment of variance homogeneity using a two‐sample F‐test. Depending on the outcome of the F‐test, the appropriate t‐test, assuming equal or unequal variances, was applied.

For experiments involving only a single comparison between two groups (e.g., gel fraction and mechanical properties), statistical analysis was performed directly using Student's t‐tests.

A significance threshold of p < 0.05 was applied throughout [110, 111].

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: adhm71396‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.