Development and Evaluation of Curcumin-Loaded Mucoadhesive Buccal Films Using Green Deep Eutectic Solvents via Design of Experiments

Abstract

Background/Objectives: This study aimed to develop and formulation-design curcumin-loaded buccal films using a green deep eutectic solvent (DES) to improve drug solubility and support localized mucosal delivery, with the help of a Design of Experiments (DoE) approach. Methods: Various DESs with different components and molar ratios were prepared, characterized, and the optimal DES was selected. Curcumin-loaded buccal films were prepared by solvent casting, employing the optimal DES as the solvent system. A three-factor and five-level central composite design was applied to systematically investigate the amounts of HPMC K100, Kollicoat^® IR, and DES in buccal films, and mucoadhesive strength, elongation (%), swelling index, swelling time, and thickness were identified as critical responses. Based on these responses, characterization, in vitro release, ex vivo permeation, and in vitro biological activity studies were performed on the model-predicted optimal formulations. Results: Choline chloride: propylene glycol (1:4 molar ratio) was selected as the optimal DES due to its viscosity, pH, organoleptic properties, and the highest curcumin solubility (11.90 ± 0.15 mg/mL). While DES increased curcumin solubility, buccal films were successfully obtained. Six formulations identified through model-based DoE optimization were advanced to detailed experimental evaluation. The drug release exhibited sustained curcumin release profiles over 24 h, and ex vivo studies showed <2% mucosal permeation. The lead formulations demonstrated in vitro cell compatibility, anti-inflammatory and antioxidant activities, and enhanced wound-healing-related bioactivity. Conclusions: Curcumin-loaded buccal films containing DES and developed using a DoE-guided formulation strategy successfully demonstrated the acceptable formulation properties and screening-level in vitro biological activities, offering a promising formulation platform for localized buccal delivery applications.

1. Introduction

Mucoadhesive buccal films adhere to the oral mucosa. They can deliver therapeutic agents in a sustained, site-specific manner, making them useful for lesions such as aphthous ulcers and oral lichen planus. Conventional topical formulations, most commonly mouth rinses and gels, are rapidly washed away by saliva, and systemic administration often fails to achieve sufficient drug levels at the affected mucosal surface [1,2]. Buccal films, by remaining in intimate contact with the tissue, allow direct release of the drug at the lesion site and can therefore improve both patient comfort and local therapeutic performance [3,4], while also providing ease of removal due to salivary clearance from the buccal mucosa [5].

Curcumin, the main bioactive component of Curcuma longa, possesses well-documented anti-inflammatory, antioxidant, and antimicrobial activities, making it a compelling candidate for managing oral mucosal diseases [6]. Despite its therapeutic potential, its clinical utility is limited by extremely poor aqueous solubility and rapid metabolic degradation, which restricts drug availability when administered through conventional dosage forms [7]. To address these limitations, numerous formulation approaches have been developed in recent years. For instance, chitosan-based mucoadhesive films designed for the topical management of oral cancer have demonstrated promising mechanical strength, adhesion, and enhanced ex vivo permeation [8]. Other strategies, such as incorporating amorphous curcumin–chitosan nanoplexes into hydroxypropyl starch matrices, have improved drug loading and provided superior disintegration and release characteristics compared with unmodified starch or HPMC films [9]. Electrospun nanofibrous oral films optimized using central composite design have achieved rapid dissolution and markedly improved apparent solubility [10], while zein/β-cyclodextrin nanoparticles have shown prolonged mucoadhesion and sustained curcumin release profiles [11]. A comprehensive review conducted in 2025 discussed innovative formulations such as solid lipid nanoparticle-loaded wafers, hydrogels, and electrospun fibers developed to solve curcumin’s solubility issues and significantly increase its therapeutic effect [12]. In addition to these systems, a limited number of studies have specifically reported curcumin-loaded buccal film formulations. Fast-dissolving polymeric buccal films have been developed using conventional polymers and solid-dispersion strategies to enhance dissolution and systemic absorption [13,14], while more recent approaches have combined nanoemulsion technology with mucoadhesive polymeric films to improve mucosal permeation and therapeutic efficacy in oral mucositis models [15].

Deep eutectic solvents (DESs) are a promising class of green solvents formed by combining hydrogen bond acceptors (e.g., choline chloride) and donors (e.g., lactic acid) in defined molar ratios (e.g., 1:1 to 1:5), and are capable of dissolving hydrophobic drugs such as curcumin [16]. DESs help avoid the use of volatile organic solvents, offer adjustable viscosity and pH, and have demonstrated low mucosal irritation and high mucosal safety in preliminary studies [16,17,18]. In addition to their role as green solubilizing media, DESs may also act as formulation-enabling components by facilitating homogeneous drug distribution within polymeric matrices and supporting film formation and processability during film preparation [19]. Despite these advantages, their application in curcumin-loaded buccal films remains unexplored. Although nanoparticles, nanoemulsions, and cyclodextrin complexes [9,10,11] have been used for buccal administration of curcumin, these approaches generally involve complex preparation steps and formulation difficulties. In contrast, DES-based systems offer a simpler and more scalable strategy by allowing the direct incorporation of poorly soluble drugs into buccal film matrices without the need for multistep processing [20,21], and provide an alternative option for drug formulations used in buccal applications, such as buccal films. Within this framework, the present study can be positioned relative to previously reported curcumin delivery systems. While earlier curcumin delivery systems have primarily focused on enhancing solubility or permeation—often aiming to increase systemic bioavailability—these approaches generally rely on complex carrier structures, multistep manufacturing processes, and in many cases the use of organic solvents. In contrast, the present work targets prolonged local residence and sustained release at the buccal mucosa, which is more appropriate for localized delivery to the oral mucosa. Moreover, by enabling direct dissolution of curcumin within the film matrix through a DES-based formulation strategy, the system avoids the need for organic solvents and complex carrier architectures, offering a simpler, safer, and more scalable platform. This green and patient-friendly formulation approach supports sustainable pharmaceutical development while providing an efficient and scalable preclinical formulation platform for localized buccal drug delivery.

Optimizing multiple formulation factors, such as the polymer ratio and DES content, can be efficiently achieved using the Design of Experiments (DoE). DoE systematically identifies key interactions with fewer experiments than traditional trial-and-error methods. Although DoE has been applied to buccal systems [22,23,24], its application in curcumin buccal film development remains rare. Recent studies have confirmed the efficiency of DoE in systematic formulation design of buccal films: Hassan et al. (2023) [22] applied a quality-by-design approach to develop mucoadhesive films with well-defined mechanical and mucoadhesive properties; Suksaeree et al. (2021) [23] used a factorial design to tailor the polymer composition for nicotine films, and Elkanayati et al. (2024) [24] implemented DoE in hot-melt extruded buccal films to control release kinetics. However, despite these advances, no research has integrated DoE with DESs in the design of curcumin-loaded buccal films. Therefore, using DoE to optimize polymer-solvent interactions in DES-containing formulations will provide a sustainable and systematic approach.

In this study, curcumin-loaded mucoadhesive buccal films containing DES as the solvent were developed, and a DoE strategy was applied to optimize the film components. Comprehensive evaluations, including physicochemical characterization, mechanical testing, mucoadhesion assessment, in vitro release profiling, ex vivo permeation studies, and in vitro biological activity, were conducted to demonstrate that the proposed formulation can be processed simply, support localized curcumin delivery at the buccal site, and exhibit functional potential for localized oral mucosal applications.

2. Materials and Methods

2.1. Materials

Curcumin (≥98% purity) was purchased from Thermo Scientific (Waltham, MA, USA). Betaine was obtained from Macklin Biochemical Co., Ltd. (Shanghai, China), and choline chloride was obtained from Ambeed (Buffalo Grove, IL, USA). Propylene glycol was supplied by ÖZLAB (Istanbul, Turkey), and glycerol by ZAG Kimya (Istanbul, Türkiye). Hydroxypropyl methylcellulose (HPMC K100, 100 mPa·s in a 2% aqueous solution at 20 °C; Colorcon, Poughkeepsie, NY, USA) was kindly gifted by Colorcon, and Kollicoat^® IR was kindly provided by BASF SE (Ludwigshafen, Germany). Acetonitrile (HPLC grade) was purchased from Carlo Erba Reagents Srl. (Cornaredo, Italy). Lactic acid and glacial acetic acid (≥99%), lipopolysaccharide (L4391), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), indomethacin (100 μM), Dulbecco’s Modified Eagle Medium (DMEM), isopropanol, 2,2-diphenyl-1-picrylhydrazyl (DPPH), and agar were also obtained from Sigma-Aldrich (St. Louis, MO, USA). pH 6.8 and 7.4 phosphate-buffered tablets, Tween 80, and calcium chloride were obtained from Merck (Darmstadt, Germany). Sodium nitrite was obtained from Fluka Chemika (Buchs, Switzerland).

2.2. HPLC Method and Validation

Curcumin was quantified using an Agilent 1100 HPLC system (Agilent Technologies, Santa Clara, CA, USA) equipped with a C18 column (250 × 4.6 mm, 5 µm; Aisimo Corporation, Tokyo, Japan) maintained at 33 °C. The mobile phase, acetonitrile and 4.2 M acetic acid (95:5, v/v), was delivered at 2.0 mL/min, with 20 µL injections and UV detection at 425 nm [25]. The analytical method was validated in accordance with the ICH Q2 (R1) guidelines [26] for specificity, linearity, accuracy, precision, limit of detection (LOD), limit of quantification (LOQ), and robustness of the method. The specificity was confirmed by the absence of interfering peaks at the curcumin retention time. The method showed linearity over the range of 0.5–15 µg/mL (r² = 0.9999). Accuracy and precision studies demonstrated recovery within 95–101% and relative standard deviations (RSD) below 1%. The LOD and LOQ were 0.035 and 0.106 µg/mL, respectively. Stability testing confirmed that curcumin solutions remained within 96.9–101.2% of the initial value for up to 48 h, indicating good robustness. This validated method was applied to determine curcumin in drug content uniformity, in vitro release, and ex vivo permeation studies.

2.3. Deep Eutectic Solvent Preparation and Characterization

Deep eutectic solvents were prepared by stirring the components at 80 °C on a magnetic stirrer (MR Hei-Standard, Heidolph, Germany) until a clear and homogeneous liquid was obtained [27,28,29]. Betaine and choline chloride were used as hydrogen bond acceptors (HBA), and propylene glycol, glycerol, and lactic acid were used as hydrogen bond donors (HBD) in molar ratios of 1:1, 1:2, 1:3, 1:4, and 1:5 (

Table 1. Deep eutectic solvent mixtures and their molar ratios.

HBA	HBD	Code	Molar Ratio
Choline chloride	Propylene glycol	ChCl:PG	1:1, 1:2, 1:3, 1:4, 1:5
Choline chloride	Glycerol	ChCl:Gly	1:1, 1:2, 1:3, 1:4, 1:5
Choline chloride	Lactic acid	ChCl:LA	1:1, 1:2, 1:3, 1:4, 1:5
Betaine	Propylene glycol	Beta:PG	1:1, 1:2, 1:3, 1:4, 1:5
Betaine	Glycerol	Beta:Gly	1:1, 1:2, 1:3, 1:4, 1:5
Betaine	Lactic acid	Beta:LA	1:1, 1:2, 1:3, 1:4, 1:5

Beta: Betaine, ChCl: Choline chloride, Gly: Glycerol, LA: Lactic acid, PG: Propylene glycol.

). The pH of the DESs was determined using a digital pH meter (HI83141, Hanna Instruments, Smithfield, RI, USA). The viscosity of the DESs was determined at 25 °C using a Brookfield-type digital viscometer (DV2T RV model, Brookfield Engineering Laboratories, Middleboro, MA, USA). The SC4-27 type spindle was submerged in the DES, and the viscosity was measured at 200 rpm. The analyses were conducted in triplicate, and mixtures that appeared turbid or showed crystallization were excluded from the study design.

The solubility of curcumin in the prepared DESs was evaluated using HPLC. To prepare saturated solutions, an excess amount of curcumin was introduced into each DES at room temperature and mixed thoroughly using a vortex mixer. The resulting suspensions were centrifuged (Sigma 3-18 KS, Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany) to remove undissolved solids, after which the clear supernatant was carefully collected. This fraction was passed through a 0.45 µm membrane filter and subsequently analyzed by HPLC to quantify curcumin. The solubility in the DES identified as optimal was also evaluated at 40 °C by adding an excess amount of curcumin and determining the dissolved concentration using the same validated HPLC method. For photostability studies, the saturated curcumin–DES solution was stored under typical laboratory lighting conditions for one month, and its curcumin content was re-examined using HPLC [30].

2.4. Preparation of Curcumin-Loaded Buccal Films

Curcumin-loaded buccal films were prepared using the solvent-casting technique under a low-light environment. Curcumin was first dissolved in the optimized DES system (ChCl:PG, 1:4), selected for its strong solubilizing ability. HPMC K100 and Kollicoat^® IR served as the film-forming and mucoadhesive polymers, respectively [31]. The polymers were dissolved in distilled water at room temperature under continuous stirring with the magnetic stirrer. In a separate step, curcumin was dissolved in the DES at 40 °C to ensure complete solubilization. The polymer dispersion was then slowly incorporated into the curcumin–DES solution to obtain a homogeneous casting mixture. Once combined, the pH of the mixture was adjusted with 0.1 N acetic acid. The final formulation was poured into a glass Petri dish and dried at 25 °C in an oven for approximately two weeks to allow gradual solvent evaporation and film formation. All mixing and casting operations were conducted under low-light conditions to protect the curcumin from photodegradation. After drying, the films were carefully removed from the Petri dishes, cut into 1.5 × 1.5 cm² pieces, and stored in a desiccator for further studies.

The buccal films were formulated to contain 1 mg of curcumin per 1.5 × 1.5 cm² film piece. The other component percentages of the buccal films were determined using the DoE approach. A total of 15 formulations were prepared based on the design points, which consisted of eight factorial, six axial, and one center point (the center point was repeated six times). The detailed compositions of these formulations are presented in

Table 2. Composition and design points of buccal films with DoE.

Design Points	HPMC K100 (%)	Kollicoat® IR (%)	DES (g)
Factorial	2	0.260	3
Factorial	5	0.260	3
Factorial	2	1	3
Factorial	5	1	3
Factorial	2	0.260	3.500
Factorial	5	0.260	3.500
Factorial	2	1	3.500
Factorial	5	1	3.500
Axial	0.977	0.630	3.250
Axial	6.023	0.630	3.250
Axial	3.500	0.007	3.250
Axial	3.500	1.252	3.250
Axial	3.500	0.630	2.829
Axial	3.500	0.630	3.670
Center	3.500	0.630	3.250

.

2.5. Design of Experiment for Buccal Films

The optimization of curcumin-loaded buccal films was performed using a three-factor-at-five-level design in Design-Expert^® software (13.0.1.0, Stat-Ease Inc., Minneapolis, MN, USA) with a central composite design (CCD) [31,32]. The effects of HPMC K100 amount (X₁), Kolliocat IR amount (X₂), and DES amount (X₃) on thickness (Y₁), bioadhesion (Y₂), percentage elongation (Y₃), swelling percentage (Y₄), and swelling time (Y₅) were optimized using a three-factor, five-level CCD.

Table 3. Independent variables, levels, coded responses, and desired outcomes of the CCD.

Independent Parameters	Coded	Levels
		−α	−1	0	+1	+α
HPMC K100 (%)	X1	0.977	2.000	3.500	5.000	6.023
Kollicoat® IR (%)	X2	0.008	0.260	0.630	1.000	1.252
DES (g)	X3	2.829	3.000	3.250	3.500	3.670
Response		Goals
Thickness (mm)	Y1	minimize
Bioadhesion (mJ·cm−2)	Y2	maximize
Percentage elongation (%)	Y3	in range
Swelling percentage (%)	Y4	maximize
Swelling time (min)	Y5	maximize

presents the levels of the independent variables and the coded responses, while the experimental methods used to determine these response parameters are described in detail in Section 2.6.

2.6. Characterization of the Buccal Films

2.6.1. Thickness and Weight Uniformity

The thickness of the 1.5 × 1.5 cm² films was measured five times using a digital micrometer (Insize Co., Ltd., Suzhou New District, China), and the mean value was recorded. Then, five films (1.5 × 1.5 cm²) were individually weighed using an analytical balance (Ohaus Corporation, Parsippany, NJ, USA). The mean weight and standard deviation (SD) were calculated to determine weight uniformity [9].

2.6.2. pH

Each buccal film (1.5 × 1.5 cm²) was immersed in 5 mL of pH 6.8 phosphate-buffered solution (PBS) and allowed to equilibrate for 2 h at room temperature. The pH of the buccal films was measured using the pH meter [31]. Measurements were performed in triplicate for each formulation, and the results were expressed as mean ± SD.

2.6.3. Percent Elongation at Break and Tensile Strength

The tensile strength and elongation of the 3 × 1 cm² films were determined using a texture analyzer equipped with a 5 kg load cell (TA.XT.Plus C Stable Micro Systems, Haslemere, Surrey, UK). The samples were clamped between 0.5 cm probes and measured, and a force was applied until the film ruptured. The test parameters included a pre-test speed of 0.5 mm/s and a test speed of 0.5 mm/s. The studies were conducted four times. The percent elongation and tensile strength were determined using the following equations (Equations (1) and (2)) [33].

% Elongation= (Increase in length (mm)/Initial length (mm)) × 100

(1)

Tensile strength (N·cm⁻²) = Force required to break (N)/Area of the film (cm⁻²)

(2)

2.6.4. Mucoadhesion Studies

Ex vivo mucoadhesion studies were performed to evaluate the mucoadhesive properties of the buccal films using the texture analyzer, following the methodologies described in previous studies [34,35]. Freshly excised bovine buccal mucosa obtained from a local slaughterhouse served as the model tissue. The underlying fat and loose connective tissues were carefully removed to obtain a mucosal thickness of approximately 1 mm. The prepared tissue samples were stored at −20 °C and thawed at room temperature prior to the experiments. For the adhesion test, the thawed mucosal tissue was fixed onto a plexiglass holder and moistened with 75 μL of pH 6.8 PBS to simulate the oral environment. A circular section of the buccal film was attached to a 10 mm cylindrical probe using double-sided adhesive tape. The probe was lowered at a pretest speed of 0.5 mm/s until it contacted the mucosal surface. A contact force of 1 N was applied for a residence time of 120 s to ensure adequate adhesion. Subsequently, the probe was withdrawn at a post-test speed of 0.5 mm/s. The force required to detach the buccal film from the tissue was recorded as a function of distance. The maximum adhesive force was determined directly from the peak of the force–distance plot. The work of adhesion (mJ·cm⁻²) was calculated using the area under the curve (AUC), according to Equation (3):

Work of adhesion (mJ·cm⁻²) = AUC_1–2/πr²

(3)

where AUC represents the area under the force-distance profile, and πr² denotes the surface area of the buccal film in contact with the mucosa. All measurements were performed at four time points (n = 4), and the results are expressed as means with SDs.

2.6.5. Swelling Studies

The swelling behavior of the buccal films was evaluated using the pH 6.8 PBS. The buccal films were weighed and placed in Petri dishes with a wire mesh, as previously described in our earlier study [33]. Ten milliliters of pH 6.8 PBS at 37 ± 0.5 °C was added to ensure complete wetting of all surfaces. The films were placed in an oven and allowed to hydrate at 37 ± 0.5 °C. At predetermined intervals, each sample was removed, and excess surface moisture was carefully blotted with cellulose filter paper to avoid altering the internal hydration level. The films were then weighed, and the swelling percentage was calculated from the change in mass. Measurements continued until the samples either reached their maximum swelling capacity or began to lose structural integrity. All formulations were evaluated in triplicate.

Swelling time was defined as the point at which the film began to disintegrate during the swelling test and was recorded in min, following the criteria described in previous studies [33,36]. The swelling percentage was calculated using Equation (4):

%Swelling = ((M₂ − M₁)/M₁) × 100

(4)

where M₁ is the initial weight of the film, and M₂ is the weight of the swollen film at a given time point.

2.6.6. Moisture Loss Percentage

Moisture loss was assessed using anhydrous calcium chloride as the desiccant. Buccal films were cut into predetermined sizes, and their initial weights (M₁) were recorded. The samples were then placed in a desiccator containing 100 g of anhydrous calcium chloride and kept at room temperature for 72 h. After the designated period, the films were removed and reweighed (M₂) [33]. All measurements were performed in triplicate, and the results are reported as mean ± SD. The percentage of moisture loss was calculated according to Equation (5):

%Moisture Loss = ((M₁ − M₂)/M₁) × 100

(5)

2.6.7. FT-IR Analysis

Fourier-transform infrared (FT-IR) spectroscopy was employed to assess the potential molecular interactions among curcumin, HPMC K100, Kollicoat^® IR, and the selected DES in the lead buccal film formulations. The active ingredient, blank, and drug-loaded optimal formulations were analyzed on an FT-IR spectrometer (PerkinElmer Inc., Waltham, MA, USA) using a UATR accessory with a diamond crystal, and spectra were recorded. Each sample was scanned over the 4000–400 cm⁻¹ range at a resolution of 4 cm⁻¹ with eight accumulations per measurement. The resulting spectra were averaged and plotted as wavenumber versus transmittance using Spectrum™ software (Version 10.03.09) [37].

2.7. In Vitro Drug Release Studies

The in vitro release of curcumin from the buccal films was evaluated using the USP apparatus V (paddle over disc) method [38]. The study was conducted in 500 mL of pH 6.8 PBS containing 1% (w/v) Tween 80 to maintain sink conditions. The dissolution medium was maintained at 37 ± 0.5 °C, and the paddle rotation speed was set to 50 rpm. To protect the curcumin from light, the experiment was conducted in a dark environment, and the dissolution tanks were covered with light-blocking material. Samples were collected at predetermined time intervals of 5, 10, 20, 30, 45, 60, and 90 min, as well as 2, 3, 4, 5, 6, 7, 8, 10, 12, and 24 h. At each time point, an aliquot was withdrawn and immediately replaced with an equal volume of fresh dissolution medium to maintain the constant volume. The collected samples were filtered through a 0.45 µm membrane filter, and the curcumin concentration was determined by HPLC. To evaluate the release profile of the buccal film, zero-order, first-order, Higuchi, and Korsmeyer–Peppas kinetic models were applied [33].

2.8. Short-Time Stability Tests

Mucoadhesive buccal films (1.5 × 1.5 cm²) were hermetically sealed in aluminum foil and airtight pouches to exclude light and air, and then stored for one month under the following conditions: in stability chambers at 25 ± 2 °C/60 ± 5% RH and 40 ± 2 °C/75 ± 5% RH, and in a 4 ± 1 °C refrigerator [39]. On the other hand, the films were stored in transparent pouches under controlled light exposure. The buccal films were analyzed at the initial time point (t = 0) and after one month of storage. For curcumin content analysis, each film was transferred into a volumetric flask containing 100 mL of mobile phase (acetonitrile:4.2 M acetic acid, 95:5 v/v) and stirred magnetically at room temperature for 2 h under light-protected conditions. The solutions were then filtered and quantified using HPLC. Moreover, the pH of the films was measured using the pH meter, as described in the method above. All assays were performed in triplicate and reported as mean ± SD [9]. To isolate the matrix effect, DES-containing films were evaluated using DES-free control films (curcumin dispersed in water and cast under the same conditions) under each of the four storage conditions. This design permitted a direct, controlled comparison of the effect of the DES matrix on drug-content retention across one month of long-term, accelerated, refrigerated, and photostable storage [40]. In addition, immediately after combining the DES–curcumin solution with the aqueous polymer solution, aliquots were withdrawn at predefined time points (t = 0, 2, 24, and 48 h) to monitor the curcumin content in the liquid matrix prior to film casting using the validated HPLC method. To verify that the DES itself does not cause curcumin degradation, we confirmed the stability of curcumin content in the pre-casting polymer mixture over 48 h, so any later loss can be attributed to drying and storage rather than the DES.

2.9. Ex Vivo Permeation Study

Ex vivo permeation studies were performed on the lead formulations (F34 and F62) according to a previous study with minor modifications [34]. The ideal buccal film was evaluated using a Franz diffusion cell system with bovine buccal tissue as the biological membrane. The buccal tissue, excised to a thickness of 1 mm and a surface area of 4 cm², was placed between the donor and receptor compartments of Franz diffusion cells with an effective diffusion area of 1.77 cm² and securely clamped in place. The donor compartment was filled with 0.5 mL of pH 6.8 PBS, while the receptor compartment contained 12 mL of a mixture of phosphate-buffered saline (pH 7.4) and ethanol (90:10, v/v) containing 1% (w/v) Tween 80 to maintain sink conditions. The buccal films, cut into 0.5 × 0.5 cm² sections, were carefully placed on the mucosal surface before starting the permeation study. The system was maintained at 37 ± 0.5 °C under continuous stirring at 300 rpm throughout the study. At predetermined time points (30 min, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 24 h), 0.5 mL aliquots were withdrawn from the receptor compartment and filtered through a 0.2 μm membrane filter. An equal volume of fresh receptor medium was added to maintain the constant volume. The study was conducted in three replicates, and the samples were analyzed by HPLC to determine cumulative curcumin permeation.

2.10. Cell Culture Studies

2.10.1. Cytotoxicity

The toxicity and cell migration of curcumin-loaded (F34 and F62) and blank buccal films were assessed using the L929 mouse fibroblast line according to the ISO 10993-5 protocol [41]. L929 cells were seeded and incubated for 24 h (37 °C, 5% CO₂). The cells were subsequently exposed to film extracts (0.5–1 mg/mL). Cytotoxicity was determined via the MTT assay, which measures mitochondrial function. After exposure, the supernatants were removed, and 0.5 mg/mL MTT solution was added for 2 h of incubation at 37 °C. The medium was then discarded, and the resulting formazan crystals were dissolved in 100 μL of isopropanol. Absorbance was measured at 570 nm using a Varioskan^TM Lux Spectrum Microplate Reader (Thermo Scientific, Waltham, MA, USA). The MTT assay was also performed on RAW 264.7 murine macrophage cells to identify the non-toxic concentrations. The cells were seeded at a density of 5 × 10⁴ cells/well and incubated for 24 h. The cells were then exposed to various film concentrations (0.5–1 mg/mL) for 2 h, followed by an additional 22 h of incubation with 1 μg/mL of lipopolysaccharide (LPS). Cellular viability was determined using the formula provided in Equation (6) [42].

Relative Cell Viability % = 100 × OD570_sample/OD570_NC

(6)

where OD570S represents the average measured optical density of the sample, whereas OD570NC represents the average measured optical density of the negative control.

2.10.2. Nitric Oxide Level

To assess inflammation, the non-toxic concentration of nitric oxide in the samples was quantified indirectly by measuring its stable metabolite, nitrite, with the Griess reagent. After 24 h of treatment, the RAW264.7 cell culture supernatants were collected. The nitrite concentration was determined spectrophotometrically at 532 nm. Quantification was achieved using a sodium nitrite standard curve, and the results were expressed in μM. Indomethacin (100 μM) served as a positive control for nitric oxide inhibition, consistent with a previous study [43].

2.10.3. Direct Contact Assay

The in vitro cytocompatibility of the films was evaluated using L929 healthy mouse fibroblast cells as per the requirements of ISO 10993-5. To compare the cytotoxicity profiles of the films, cells were plated in 24-well plates and incubated at 37 °C with 5% CO₂ for 24 h to achieve confluency. Film samples, along with positive (PC) and negative (NC) controls, were applied to 1.9 cm² filter paper (pore size 0.45 μm) for extraction. Following a 24 h exposure period, cellular viability was determined using the MTT assay and calculated according to Equation (6).

2.10.4. Indirect Contact Assay

L929 cells were seeded into 24-well microplates and incubated for 24 h (37 °C, 5% CO₂). For the indirect contact assay, the supernatant was discarded, and a 50% 2× DMEM and 50% agar mixture (v/v) was added to the wells to create a barrier. The film extractions were then spread onto a filter paper (0.45 μm pore size, 1.9 cm² area) and carefully placed on the surface of the DMEM–agar mixture using forceps. After 24 h of further incubation, MTT solution (0.5 mg/mL) was added. Following a 2 h incubation period at 37 °C, formazan was dissolved using 300 μL of isopropanol. The resulting purple chromophore, formed by the isopropanol migrating under the agar, was measured spectrophotometrically at 570 nm using the microplate reader. Cellular viability was calculated as defined in Equation (6).

2.10.5. DPPH Radical-Scavenging Activity

The in vitro antioxidant capacity of the F34 and F62 curcumin-loaded and blank buccal films was assessed using the DPPH (2,2-diphenyl-1-picrylhydrazyl) radical-scavenging assay with the methodology described by Reis et al. [42]. Briefly, 50 µL of each sample was mixed with 250 µL of freshly prepared 0.1 mM DPPH solution in methanol. The mixture was then incubated in the dark for 50 min at room temperature. The absorbance was measured at 517 nm using the spectrophotometer. Butylated hydroxytoluene (BHT) was utilized as the reference substance (positive control), while phosphate-buffered saline was used as the negative control. The radical scavenging activity for the control and all samples was calculated using Equation (7).

DPPH radical-scavenging activity (%) = (Abs_control − Abs_sample)/Abs_control × 100

(7)

Abs_control is the absorbance value of the control group (phosphate-buffered saline), and Abs_sample is the absorbance of the reaction mixture containing the films or BHT. All measurements were conducted in triplicate (n = 3), and the results were expressed as %. The statistical significance of the differences between the formulations was analyzed using the Mann–Whitney U test.

2.10.6. Relative Wound Healing Capacity

The wound-healing activity of the formulations was assessed at the previously determined non-cytotoxic concentrations. To create a simulated wound, pre-incubated L929 cells were uniformly scratched using a cell-scratcher. Images of the scratched areas were captured immediately (0 h) and after 24 h. The resulting wound width was quantified using the ImageJ analysis software (version 1.54g, National Institutes of Health, Bethesda, MD, USA), and the wound closure percentage was calculated using Equation (8) [43].

$$\begin{gathered} \text{Wound healing (\%) = (Wound area} {\mathrm{t}}_0\text{) − (Wound area} {\mathrm{t}}_{\mathrm{final}}\text{)/(Wound area} {\mathrm{t}}_{\mathrm{final}}\text{) × 100} \\ {\mathrm{t}}_0\text{: The moment when the wound model was created and the first image was taken} \\ {\mathrm{t}}_{\mathrm{final}}\text{: The moment when the negative control group was completely closed} \end{gathered}$$

(8)

2.11. Statistical Analysis

All quantitative results are expressed as the mean ± SD of at least three independent experiments. Design of Experiments (DoE) was conducted using Design-Expert^® v22 (Stat-Ease, Minneapolis, MN, USA). A three-factor, five-level Central Composite Design (CCD) with six center-point replicates generated 20 runs. The significance of differences between blank and loaded films was analyzed using the Mann–Whitney U test (p < 0.05 was considered significant). One-way ANOVA, Tukey’s test, and Mann–Whitney U test were performed using GraphPad Prism software (version 10.1.1, GraphPad Software, San Diego, CA, USA).

3. Results and Discussion

This study is framed as a formulation-focused, exploratory study rather than a translational or clinically predictive therapeutic investigation. The primary aim of the study is not to demonstrate clinical efficacy or define therapeutic tissue concentration thresholds, but to show the feasibility of integrating green DESs with mucoadhesive buccal film systems and a structured DoE approach to create a reproducible and systematic formulation development framework for localized buccal applications. In this context, the study is structured as a platform-level formulation study aiming to reveal the physicochemical suitability, process compatibility, and functional performance characteristics for topical buccal treatment, rather than predicting clinical outcomes. Biological tests and ex vivo models were used not as predictors of in vivo therapeutic outcomes, but as supportive screening tools for evaluating local biocompatibility and functional biological activity. The novelty of this study lies in developing a scalable, green, and systematically optimized formulation platform for the localized buccal transport of curcumin, a low-solubility compound.

3.1. Deep Eutectic Solvent Preparation and Characterization

DES systems comprising various HBA-HBD components at different molar ratios were prepared to increase curcumin solubility and identify the most suitable DES system based on preformulation screening criteria for buccal films. In this study design, betaine and choline chloride, which are most frequently used as HBA in pharmaceutical DES systems in the literature and are considered to have high biocompatibility and pharmaceutical safety, were preferred. These naturally occurring endogenous molecules are widely reported in DES-based drug formulations due to both their low toxicity profiles and their effectiveness in dissolving hydrophobic drugs [44,45], and were evaluated as suitable candidates for buccal film formulation development. Following the preparation, pH and viscosity measurements were performed for the DES systems that formed a clear and homogenous appearance, and curcumin solubility was evaluated.

The results of the characterization of betaine-based DES are shown in

Table 4. Viscosity, pH, and curcumin solubility of betaine-based DESs.

DES	Ratio	Viscosity (mPa·s ± SD)	pH (±SD)	Solubility of Curcumin (mg/mL ± SD)
Beta:PG	1:4	145.0 ± 1.3	8.38 ± 0.22	0.51 ± 0.01
	1:5	116.0 ± 0.2	7.71 ± 0.03	0.60 ± 0.01
Beta:Gly	1:2	2506.2 ± 1.1	8.26 ± 0.19	- *
	1:3	1823.0 ± 0.0	7.67 ± 0.07	- *
	1:4	1523.1 ± 21.4	7.32 ± 0.20	- *
	1:5	1248.0 ± 10.0	7.09 ± 0.31	- *
Beta:LA	1:2	242.5 ± 0.1	3.49 ± 0.03	0.10 ± 0.01
	1:3	127.5 ± 0.8	3.01 ± 0.01	0.19 ± 0.01
	1:4	90.8 ± 0.7	2.69 ± 0.01	0.17 ± 0.01
	1:5	70.6 ± 0.9	2.50 ± 0.04	0.14 ± 0.01

* No measurement was taken.

. Beta:PG mixtures at molar ratios of 1:1 to 1:5 were screened; only the 1:4 (145.0 ± 1.3 mPa·s, pH: 8.38 ± 0.22) and 1:5 (116.3 ± 0.2 mPa·s, pH: 7.71 ± 0.03) ratios formed clear, homogeneous liquids at 25 °C. The solubility of the curcumin was found to be 0.51 and 0.60 mg/mL for the Beta:PG systems at 1:4 and 1:5 molar ratios, respectively.

The Beta:Gly and Beta:LA systems at a 1:1 molar ratio did not form a transparent and homogeneous solution system. All other Beta:Gly systems exhibited extremely high viscosities (1248–2506 mPa·s) and could not be measured for solubility due to their high viscosities, and the pH range was found to be 7.09–8.26. The other Beta:LA DES systems, although clear, had a pH of 2.50–3.49 and an offensive odor. The solubility of the curcumin in the Beta:LA systems was found to range from 0.10 to 0.19 mg/mL.

In the present study, consistent with previous studies, betaine-based systems were found to have a low solubility capacity for curcumin [46]. It was observed that the Beta-Gly systems did not provide a suitable solubility environment for curcumin because of their very high viscosity; therefore, curcumin solubility could not be evaluated. Although curcumin solubility was achieved in Beta:LA systems, it was found to be quite low because the acidic nature of these systems was not optimal for curcumin dissolution. Beta:PG systems offer higher solubility than compared to other betaine systems due to their low viscosity and high pH values [47]; however, they did not provide sufficient solubility for buccal films, and the solubility values were still well below 1 mg/mL. Consequently, none of the betaine-based DESs were selected for buccal films because of their excessive viscosity, irritant pH, and poor organoleptic properties.

The pH, viscosity, and solubility of curcumin in the ChCl-based DESs are listed in

Table 5. Viscosity, pH, and curcumin solubility of choline chloride-based DESs.

DES	Ratio	Viscosity (mPa·s ± SD)	pH (±SD)	Solubility of Curcumin (mg/mL ± SD)
ChCl:LA	1:2	46.25 ± 0.01	0.61 ± 0.01	0.25 ± 0.01
	1:3	41.88 ± 0.62	0.44 ± 0.01	0.39 ± 0.01
	1:4	34.38 ± 0.63	0.31 ± 0.01	0.47 ± 0.01
	1:5	30.53 ± 0.60	0.19 ± 0.02	0.22 ± 0.01
ChCl:PG	1:2	56 ± 1.00	4.88 ± 0.10	5.57 ± 0.01
	1:3	53.75 ± 0.01	5.56 ± 0.05	3.54 ± 0.01
	1:4 *	48.13 ± 0.63	5.70 ± 0.04	8.54 ± 0.01
	1:5	41.88 ± 0.62	5.57 ± 0.03	7.59 ± 0.01
ChCl:Gly	1:2	244.40 ± 0.60	5.59 ± 0.04	- **
	1:3	243.15 ± 3.15	5.53 ± 0.09	- **
	1:4	371.30 ± 0.10	5.42 ± 0.03	- **
	1:5	344.40 ± 0.60	5.47 ± 0.05	- **

* Selected as the optimal DES. ** No measurement was taken.

. It was observed that ChCl-based DESs formed stable and clear liquids at ratios ≥ 1:2. The ChCl:LA systems exhibited viscosity values ranging from 30.53 to 46.25 mPa·s and pH values between 0.19 and 0.61. However, these systems are characterized by a pungent odor. In addition, due to highly acidic pH, the ChCl:LA systems exhibited limited curcumin solubility, ranging from 0.22 to 0.47 mg/mL, even though favorable low viscosity values were observed. Ch:Gly systems were found to have higher pH values (approximately 5.50); however, due to the high viscosity (243.15–344.40 mPa·s), the solubility study could not be performed.

In contrast, ChCl:PG systems exhibited excellent characteristics, including low viscosity (41.88–56.00 mPa·s), a moderately acidic pH range suitable for formulation processing and curcumin stability (4.88–5.70), high curcumin solubility (3.54–8.54 mg/mL), and no offensive odor. Among the ChCl-based systems, the ChCl:PG 1:4 system showed the highest curcumin solubility (8.54 ± 0.01 mg/mL) combined with moderate viscosity (48.13 ± 0.63 mPa·s) and favorable pH (5.70 ± 0.04) at 25 °C. Therefore, this system was selected as the most suitable DES for curcumin dissolution and buccal film formulation development.

Although the solubility obtained at 25 °C with the ChCl:PG (1:4) system was higher than previously reported values [48,49], it was deemed insufficient for buccal film formulation to load the desired curcumin dose. Therefore, to obtain a higher solubility, the solubility process was carried out at 40 °C, selected based on literature reports indicating curcumin degradation at temperatures above 70 °C [50], resulting in a 39% increase (11.90 ± 0.15 mg/mL). The increase in solubility at 40 °C is most likely associated with enhanced molecular mobility and stronger curcumin–DES interactions at elevated temperatures [48]. To further evaluate the stability of curcumin within the optimized DES, a photostability study was performed under light. After one month of exposure, approximately 92.23% of the curcumin remained intact, demonstrating that the DES provides substantial protection against light-induced degradation [17]. This protective effect appears to arise from the combined influence of extensive hydrogen-bonding networks, restricted molecular diffusion due to the viscosity of the medium, and the inherently low water activity within the DES matrix, all of which collectively limit degradation pathways [17,30].

A broader comparison of all DES formulations prepared in this study showed that curcumin solubility is governed by a multifactorial interplay between pH, viscosity, and the chemical composition and molar ratios of the DES constituents. Systems characterized by either very high viscosity or excessively low pH generally exhibited lower solubility, likely because high viscosity impedes molecular diffusion, whereas low pH suppresses the deprotonation of phenolic functional groups on curcumin, reducing its solvation capacity [47].

By contrast, the ChCl:PG systems consistently produced superior solubility as seen in our previous study [44]. This performance can be attributed to the ability of chloride ions and the hydroxyl groups of propylene glycol to form strong hydrogen bonds with curcumin, supported by additional van der Waals interactions that stabilize the dissolved drug. The moderate viscosity of this DES further contributes by maintaining a balance between structural cohesiveness and sufficient molecular mobility, allowing curcumin to dissolve at comparatively high concentrations [16]. In contrast to traditional solubilization methods for curcumin, which frequently depend on harsh thermal processing, multistep preparation procedures, organic solvents, or surfactants that may irritate the mucosa, the optimized DES system enables curcumin to dissolve directly without the need for auxiliary excipients. This feature makes the system particularly suitable for the buccal film formulation process. The single-step dissolution process substantially reduces the complexity of film manufacturing by allowing the DES to be incorporated directly into the polymer matrix, thereby eliminating the requirement for separate pre-dissolution steps. Beyond its application in buccal films, this DES-based approach provides a broadly adaptable platform for formulating other hydrophobic drugs, especially in cases where high drug loading and gentle processing conditions are essential.

Overall, the water solubility of curcumin has been reported in the literature at the µg/mL level [51], and these values are considered practically insoluble for pharmaceutical applications. In contrast, the ChCl:PG (1:4) DES used in this study exhibited excellent solubility capacity for curcumin at both 25 °C (8.54 ± 0.01 mg/mL) and 40 °C (11.90 ± 0.15 mg/mL), resulting in a thousand-fold increase in solubility compared to literature-reported values. Furthermore, this DES system facilitated a simplified and scalable production by enabling direct integration of curcumin into the polymer matrix. The combination of high solubility, suitable viscosity, formulation-compatible pH, and inherent photoprotective effects underscores its promise as a green solvent system for developing drug-loaded buccal films. Moreover, these characteristics suggest its potential utility in a wider range of dosage forms where maintaining drug stability and minimizing processing complexity are crucial.

3.2. Preparation of Curcumin-Loaded Buccal Films and Preformulation Studies

Ideally, buccal films should exhibit mechanical flexibility, sufficient durability for handling, and a homogeneous structure to ensure dose uniformity [52]. To ensure these critical quality attributes and determine the component ratios specified in the methodology, the relationship between polymer concentration and film-forming capacity was evaluated based on macroscopic observations and physical handling properties in preformulation studies of the buccal films. HPMC K100, used as the main matrix-forming agent, produced formulations with poor film-forming ability at concentrations below 0.9% (w/w). Thus, the resulting structures were not sufficiently thick or durable to separate from the casting surface. However, at levels above 6%, there was insufficient water for the dissolution of the polymer, and due to the high viscosity, the casting process became difficult. Therefore, buccal films were designed with a total polymer concentration of approximately 1–6%. Additionally, the high water content of buccal films with low polymer concentrations caused destabilization of the hydro-sensitive curcumin–DES system [48], and precipitation of curcumin. In contrast, at optimum polymer levels (3% and above), it was observed that the viscous structure of the obtained polymer network acted as a barrier, and the curcumin–DES system in the matrix was distributed homogeneously without forming precipitates. Additionally, adjusting the pH of the casting mixture with the acetic acid to the range of 5–6 contributed to the buccal physiological compatibility and chemical stability of curcumin [50].

The amount of DES used in the formulations emerged as a critical factor during optimization, as it significantly affected the films’ mechanical flexibility, drying profile, and overall drug-loading capacity. Notably, the films maintained adequate flexibility and structural integrity even in the absence of an external plasticizer, suggesting a plasticizer-like functional contribution of the DES rather than a definitive plasticizing role to the polymer matrix [18,21], consistent with the known plasticizing properties of its constituent components (e.g., propylene glycol, glycerol) [53]. However, when the DES proportion exceeded approximately 10% of the total formulation, complete solvent removal during drying became difficult. Under these conditions, the films tended to remain soft and tacky, necessitating extended drying periods to achieve acceptable handling characteristics. In preliminary studies, the drying method was found to be a decisive parameter for the morphological integrity and dose homogeneity of the films. Film integrity and dose homogeneity could not be achieved in films produced at high temperatures and rapid drying. The water-holding capacity of the DES [18] naturally prolongs drying time. Therefore, a controlled and slow drying strategy at 25 °C was adopted. Although the drying phase extended over 14–20 days, depending on the DES-to-water ratio, the gradual evaporation proved essential for achieving high-quality films. The extended drying period enabled the polymer chains to arrange themselves into low-energy conformations, ultimately producing smooth, optically clear, and mechanically robust buccal films (Figure 1).

The limited amount of DES that could be used in the formulation, combined with curcumin’s solubility (8.5 mg/mL) at room temperature, made it difficult to load a high dose of curcumin onto buccal films. To address the limitation in curcumin loading, a two-pronged strategy was implemented. (i) Thermal adjustment: Increasing the dissolution temperature to 40 °C enhanced molecular mobility and improved curcumin solubility to 11.9 mg/mL [48]. To avoid thermal degradation, the temperature was not raised further, ensuring an appropriate balance between chemical stability and solubility [50]. (ii) Surface-area optimization: Because the amount of DES that could be incorporated into the formulation was limited, the buccal film area was designed with a surface area of 2.25 cm² (1.5 × 1.5 cm²). Based on literature recommendations, each film was designed to contain a 1 mg dose of curcumin [9]. The selected dose was used as a technically practical formulation level to enable reproducible film preparation and system characterization, rather than as a clinically optimized therapeutic dose. Following these optimization steps, curcumin-loaded mucoadhesive buccal films were prepared using a green DES system in combination with the conventional solvent-casting technique. Controlled drying conditions enabled the formation of uniform films with adequate mechanical integrity and structural homogeneity, without the use of external plasticizers or additional plasticizing excipients. The process allowed reproducible film production and provided a consistent formulation basis for subsequent physicochemical and functional characterization.

3.3. Design of Experiment for Buccal Films

The formulation of curcumin-loaded buccal films was systematically optimized using a three-factor, five-level central composite design in Design-Expert^®. Using HPMC K100 (X₁), Kollicoat^® IR (X₂), and DES (X₃) as variables, predictive quadratic models were generated for five critical quality attributes: mucoadhesive strength, percent elongation, swelling index, swelling time, and thickness. For all responses, statistically significant models with non-significant lack of fit values (p > 0.05) and adequate precision ratios above the recommended threshold (>4) were obtained, demonstrating an adequate fit to the experimental data, although predictive capabilities (Predicted R²) varied among parameters. The corresponding quadratic polynomial equations describing the relationships between the formulation variables and the evaluated responses are presented in

Table 6. Quadratic polynomial equations of the DoE models for the evaluated responses (coded factor levels).

Response	Code	Coefficient Equations
Thickness (mm)	Y1	0.6353 + 0.0772X1 + 0.0461X2 + 0.0489X3
Bioadhesion (mJ·cm−2)	Y2	0.0089 − 0.0006X1 − 0.0020X2 − 0.0007X3 + 0.0013X1X2 − 0.0005X1X3 − 0.0005X2X3 + 0.0014X12 + 0.0011X22 + 0.0014X32
Percentage elongation (%)	Y3	84.80 + 29.14X1 + 0.7408X2 − 5.58X3 + 2.76X1X2 − 5.53X1X3 + 0.1489X2X3 − 3.63X12 + 6.73X22 + 5.59X32
Swelling percentage (%)	Y4	152.80 + 15.53X1 − 9.51X2 − 13.55X3 − 4.14X1X2 + 0.6169X1X3 + 6.48X2X3 + 1.66X12 − 9.17X22 − 12.15X32
Swelling time (min)	Y5	10.08 + 5.28X1 + 0.7322X2 + 0.7322X3 + 1.25X1X2 + 1.25X1X3 − 1.25X2X3 + 2.14X12 + 3.02X22 − 0.5143X32

. The statistical adequacy, model fit parameters, and predictive performance indicators (R², adjusted R², predicted R², Adeq Precision, F-value, and p-value) for each response are summarized in

Table 7. Statistical model summary and adequacy parameters for response variables in DoE optimization.

Response	Model	R2	Adjusted R2	Predicted R2	Adeq Precision	F-Value	p-Value
Thickness	Linear	0.7344	0.6847	0.5607	13.5449	14.75	<0.0001
Bioadhesion	Quadratic	0.7313	0.4895	−1.0352	5.0458	3.02	0.0498
Percentage elongation	Quadratic	0.9288	0.8648	0.4589	15.4159	14.50	0.0001
Swelling percentage	Quadratic	0.7812	0.5842	−0.6623	7.1681	3.97	0.0213
Swelling time	Quadratic	0.8784	0.7690	0.0503	8.5624	8.03	0.0016

. Response surface plots allowed us to determine how polymers and DES content affected the performance and responses of buccal films. For responses exhibiting low or negative predicted R² values, the CCD models were used solely for exploratory trend visualization and formulation space mapping, rather than for predictive optimization or mechanistic interpretation.

In addition, following model construction and fitting, multi-response numerical optimization was performed using a desirability function–based approach, in which a global desirability score was calculated to optimize all response variables simultaneously. Based on this optimization process, the model generated multiple predicted optimal formulation points, and six of these model-predicted optimal formulations with the highest global desirability scores were selected for experimental preparation, validation, and further characterization. The tests on these six formulations were then repeated to assess reproducibility and consistency of formulation performance. The components of the model-predicted DoE-optimal formulations selected for further studies are listed in

Table 8. Composition of the six DoE optimized buccal film formulations.

Code	HPMC K100 (%)	Kollicoat® IR (%)	DES (ChCl:PG 1:4) (g)	Water (g)	Curcumin (mg)
F1	5	0.260	3.012	q.s. 33	28.260
F34	2	0.260	3.006	q.s. 33	28.260
F62	5	0.260	3.474	q.s. 33	28.260
F73	4.986	0.938	3.058	q.s. 33	28.260
F79	3.562	0.260	3.000	q.s. 33	28.260
F84	2.016	0.260	3.422	q.s. 33	28.260

q.s.: quantum sufficient.

. The influence of formulation variables on the key quality attributes of curcumin-loaded buccal films was evaluated using the DoE. As shown in Figure 2, the response surface plots demonstrate the combined effects of the HPMC K100, Kollicoat^® IR ratio, and DES content on the film characteristics, and are used for exploratory formulation space mapping and trend visualization, rather than predictive optimization.

3.3.1. Mucoadhesive Strength

The quadratic model describing mucoadhesive strength was statistically significant (p < 0.05) with an adequate precision of 5.046 (>4), indicating a sufficient signal-to-noise ratio. A non-significant lack of fit (p = 0.4895) confirmed that the model adequately represented the experimental data. Changes in the concentration of Kollicoat^® IR (X₂) showed a negative linear effect (p = 0.0107), indicating that increasing the concentration of Kollicoat^® IR was associated with reduced mucoadhesive strength. Both HPMC K100 (X₁) and the DES content (X₃) produced significant quadratic effects, suggesting that mucoadhesion values were higher at intermediate levels of these components, whereas very low or very high concentrations were associated with lower performance. The interaction terms were not statistically significant, indicating that each factor influenced mucoadhesion independently. Although the model accounted for a considerable portion of the experimental variability (R² = 0.7313), the negative predicted R² value (−1.0352) indicates limited predictive reliability, consistent with experimental variability and sample size limitations. The corresponding surface plots depicting these effects are shown in Figure 2A.

From a formulation standpoint, achieving an appropriate degree of mucoadhesion is critical to ensuring that the film remains in contact with the buccal mucosa long enough to deliver therapeutic benefit, while avoiding excessive adhesion that could lead to discomfort or mucosal irritation [34]. The variability observed among measurements may be partly due to the sensitivity of mucoadhesion testing to external factors, such as the hydration level of the mucosal surface and natural differences in bovine tissue substrates. Within the exploratory DoE framework, the observed trends suggest that formulations with lower Kollicoat^® IR content and intermediate levels of HPMC K100 and DES are associated with more favorable mucoadhesion profiles, without implying predictive optimization or mechanistic causality.

3.3.2. Percent Elongation

The quadratic model developed for elongation was statistically significant (p < 0.05), supported by a high coefficient of determination (R² = 0.9288) and an adequate precision value of 15.416, which together indicated a strong signal-to-noise ratio. The lack of fit test was non-significant (p = 0.8648), confirming that the model reliably described the experimental data. Among the individual factors, HPMC K100 (X₁) showed a pronounced positive linear effect (p < 0.0001), demonstrating that higher concentrations substantially increased film flexibility, reflecting changes in the polymer matrix structure and hydration behavior. Kollicoat^® IR (X₂) did not exert a significant linear effect; however, its quadratic term was significant (p = 0.0317), indicating that flexibility values were higher at intermediate levels, while very low or high concentrations were associated with reduced performance. DES (X₃) exhibited a combination of a negative linear effect with a positive quadratic contribution, suggesting that moderate DES levels were associated with improved mechanical behavior. None of the interaction terms were significant, implying that the factors influenced elongation largely independently. Although the model accounted for a substantial proportion of variability (adjusted R² = 0.8648), the predicted R² (0.4589) indicated only moderate predictive capability, highlighting the need for confirmatory studies (Figure 2B). From a formulation perspective, the observed trends indicate that increasing HPMC K100 while maintaining Kollicoat^® IR and DES at intermediate levels is associated with improved flexibility and mechanical performance, within the exploratory DoE framework.

3.3.3. Swelling Index

The quadratic model generated for the swelling index was statistically significant (p = 0.0213), and the adequate precision value of 7.168 (>4) indicated a satisfactory signal-to-noise ratio. The lack of fit test was non-significant (p = 0.5842), confirming that the model successfully reflected the experimental data. HPMC K100 (X₁) exerted a significant positive linear effect (p = 0.0079), indicating that higher levels of this polymer were associated with increased water uptake and swelling. In contrast, DES (X₃) showed both significant negative linear (p = 0.0163) and quadratic effects, suggesting that higher DES concentrations were associated with reduced swelling capacity. Kollicoat^® IR (X₂) displayed a negative trend in both linear and quadratic terms, although these effects were not statistically significant. This tendency may reflect its high aqueous solubility [31], which is associated with altered hydration behavior and reduced swelling capacity [54]. Interaction terms were not significant, suggesting that each factor influenced swelling behavior independently. While the adjusted R² value (0.5842) indicated a moderate degree of model fit, the negative predicted R² (−0.6623) reflected limited predictive reliability, emphasizing the need for additional confirmatory trials (Figure 2C). From a formulation standpoint, within the exploratory DoE framework, the observed trends indicate that increasing the proportion of HPMC K100 while maintaining DES and Kollicoat^® IR at intermediate levels is associated with more balanced swelling and mechanical stability, without implying predictive optimization.

3.3.4. Swelling Time

The quadratic model describing swelling time was statistically significant (p = 0.0016), and the adequate precision value of 8.562 indicated a strong signal-to-noise ratio. The non-significant lack of fit result (p = 0.7690) further supported that the model appropriately represented the experimental data. HPMC K100 (X₁) exhibited a marked positive linear effect (p < 0.0001), demonstrating that higher polymer levels prolonged the swelling duration before disintegration, reflecting changes in matrix hydration and structural integrity [55]. Both HPMC K100 and Kollicoat^® IR exhibited non-linear effects, with intermediate concentrations yielding the longest swelling times, whereas very low or very high levels resulted in shorter durations. Neither Kollicoat^® IR (X₂) nor DES (X₃) showed significant linear effects, and none of the interaction terms reached statistical significance, indicating that the factors acted largely independently. Although the adjusted R² (0.7690) suggested that the model had good descriptive capability, the low predicted R² value (0.0503) indicated limited predictive strength, highlighting the need for additional confirmatory experiments (Figure 2D). From a formulation standpoint, within the exploratory DoE framework, increasing HPMC K100 while maintaining Kollicoat^® IR at intermediate concentrations is associated with extended swelling time and surface residence, without implying predictive optimization or mechanistic causality.

3.3.5. Thickness

The DoE analysis of film thickness identified the linear model as the most appropriate fit, supported by a highly significant model p-value (<0.0001), a strong Adequate Precision value (13.54), and good agreement between the adjusted (0.6847) and predicted (0.5607) R² values. Together, these indicators reflect good descriptive model performance and internal consistency. All three formulation variables, HPMC K100 (X₁), Kollicoat^® IR (X₂), and DES (X₃), had statistically significant effects (p = 0.0001, 0.0085, and 0.0058, respectively), demonstrating that increasing the concentration of any of these components produced thicker films. Examination of the coefficients showed that HPMC K100 exerted the strongest influence on thickness (+0.0772), followed by DES (+0.0489) and Kollicoat^® IR (+0.0461). The non-significant lack of fit value (p = 0.6847) confirmed that the model appropriately represented the experimental data. Overall, the combination of significant model statistics, consistent model behavior, and well-defined factor effects indicates that film thickness is systematically influenced by compositional variation, with polymer content serving as the primary determinant, within the exploratory DoE framework.

3.3.6. Consistency Assessment of DoE Responses

To verify the reproducibility and consistency of the DoE models, the formulations were re-prepared and tested for all five critical quality attributes under identical experimental conditions.

Table 9. Predicted vs. Observed DoE Responses for the Buccal Film Formulations.

Code	Mucoadhesion (mJ·cm−2) Pred/Obs/Dev (%)	Elongation (%) Pred/Obs/Dev (%)	Swelling Index (%) Pred/Obs/Dev (%)	Swelling Time (min) Pred/Obs/Dev (%)	Thickness (mm) Pred/Obs/Dev (%)
F1	0.014/0.013/−7.1	129.3/87.2/−32.5	181.9/181.3/−0.3	14.99/15/0.1	0.620/0.546/−11.9
F34	0.016/0.015/−6.3	66.3/82.5/+16.2	143.7/157.4/9.5	9.27/10/7.8	0.464/0.511/10.1
F62	0.012/0.012/0.0	107.9/127.5/18.2	147.3/142.8/−3.2	21.02/20/−4.9	0.710/0.847/19.3
F73	0.012/0.012/0.0	129.3/102.4/−20.9	150.7/144.7/−4.0	20.13/20/−0.6	0.713/0.669/−6.2
F79	0.014/0.015/7.1	103.4/95.5/−8.3	161.8/161.9/0.0	9.99/10/1.0	0.543/0.586/7.9
F84	0.016/0.015/−6.3	63.6/53.8/−9.8	115.3/118.6/2.9	10.75/10/−6.9	0.546/0.527/−3.5

Pred = DoE prediction; Obs = Experimental observation; Dev = (Obs − Pred)/Pred × 100.

summarizes the predicted values, experimental observations, and percentage deviations (Dev) calculated as “((Obs−Pred)/Pred) × 100”.

Overall, the DoE models demonstrated consistent descriptive performance, with most responses falling within the predefined acceptance criteria. Mucoadhesion displayed excellent agreement between predicted and observed values (−7.1% to +7.1%), indicating limited experimental variability under standardized conditions. Both the swelling index (−4.0% to +9.5%) and swelling time (−6.9% to +7.8%) similarly showed close alignment with model predictions, suggesting that these measurements were relatively stable across repeated experiments. Thickness deviations ranged from −3.5% to +19.3%, a level of variability consistent with previously reported studies in which prediction errors below 20% are considered acceptable [23]. The only response exceeding this threshold was elongation for formulation F1 (−32.5%). Such variability is well documented in the characterization of thin polymeric films, as elongation is highly sensitive to subtle differences in processing variables commonly encountered during solvent-casting [52,56]. Since the deviations observed for all other formulations remained within ±20%, the model was considered adequate for descriptive formulation mapping of mechanical behavior, rather than for predictive optimization.

Collectively, these findings confirm that the DoE models established in this study are experimentally reproducible and suitable for exploratory formulation space analysis, with prediction errors comparable to those reported in similar formulation-optimization work [31,36,57,58]. From a formulation development perspective, the magnitude of variability, particularly in elongation, underscores the need for careful control and optimization of experimental conditions to ensure reproducibility in future studies and large-scale manufacturing.

3.4. Characterization of Optimal Buccal Films

Comprehensive characterization tests were performed on the model-predicted optimal buccal film formulations obtained through DoE-based multi-response optimization. These films were evaluated for physical uniformity (thickness and weight variation), pH, mechanical strength (percent elongation and tensile strength), mucoadhesive properties, hydration behavior (swelling index and time), moisture loss, and molecular interactions (FT-IR) to assess their physicochemical suitability and formulation performance. The characterization results of the formulations are presented in

Table 10. Characterization results for the model-predicted optimal buccal films.

Code	Thickness (mm ± SD)	Weight (g ± SD)	pH (±SD)	Elongation (% ± SD)	Tensile Strength (N·cm−2 ± SD)	Mucoadhesion (mJ·cm−2 ± SD)	Swelling Index (% ± SD)	Swelling Time (min ± SD)	Moisture Loss (% ± SD)
F1	0.55 ± 0.03	0.17 ± 0.02	6.71 ± 0.03	87.23 ± 4.83	0.11 ± 0.01	0.013 ± 0.001	181.34 ± 25.83	15 ± 0	24.05 ± 0.61
F34	0.51 ± 0.05	0.17 ± 0.01	6.60 ± 0.01	82.54 ± 3.19	0.04 ± 0.00	0.015 ± 0.002	157.36 ± 6.38	10 ± 0	32.94 ± 0.37
F62	0.85 ± 0.09	0.20 ± 0.00	6.59 ± 0.01	139.84 ± 21.54	0.10 ± 0.01	0.012 ± 0.002	142.85 ± 23.53	20 ± 0	25.63 ± 0.67
F73	0.67 ± 0.04	0.22 ± 0.01	6.55 ± 0.01	102.39 ± 9.72	0.15 ± 0.01	0.012 ± 0.002	144.68 ± 35.92	20 ± 0	21.29 ± 0.47
F79	0.59 ± 0.01	0.18 ± 0.01	6.60 ± 0.02	82.44 ± 32.92	0.17 ± 0.08	0.015 ± 0.001	161.86 ± 17.74	10 ± 0	21.29 ± 0.67
F84	0.53 ± 0.02	0.19 ± 0.01	6.56 ± 0.02	53.79 ± 5.00	0.02 ± 0.00	0.015 ± 0.001	118.55 ± 50.00	10 ± 0	28.79 ± 1.50

.

3.4.1. Thickness and Weight Uniformity

The thickness of the buccal films varied between 0.51 ± 0.05 mm and 0.85 ± 0.09 mm, respectively. The weight of the buccal films varied between 0.17 ± 0.02 g and 0.22 ± 0.01 g (Table 10). The high thickness observed in formulation F62 was associated with the higher polymer concentration in the casting solution and reduced solvent evaporation rate, leading to the formation of a thicker film. By contrast, formulations containing lower polymer levels (F34 and F84) yielded noticeably thinner films. As expected, increasing the proportions of polymer and DES resulted in higher film weights, consistent with the corresponding increases in thickness. Importantly, all films remained below 1 mm in thickness, a range considered comfortable for buccal application and unlikely to irritate. Slightly thicker formulations may even offer practical advantages, such as improved mechanical robustness and easier handling during placement [59].

3.4.2. pH

The pH values of the films (Table 10) were close to that of the buccal mucosa (approximately pH 6.8), indicating good physiological compatibility and a low likelihood of irritation. All pH values remained within the acceptable range for buccal use. Maintaining this range is essential for ensuring patient comfort and supporting compliance during administration [31].

3.4.3. Percent Elongation at Break and Tensile Strength

Flexibility is an essential mechanical requirement for buccal films, as it reduces the likelihood of cracking during placement and allows the film to conform to mucosal movements without losing structural integrity [60]. In this study, elongation at break (%) ranged from 53.79 ± 5.00% for F84 to 139.84 ± 21.54% for F62, demonstrating that all formulations possessed sufficient flexibility for routine handling and application (Table 10). According to Tukey’s post hoc analysis, F62 exhibited significantly greater elongation than F1, F34, and F84 (p < 0.05), while F73 also showed higher elongation compared with F84 (p < 0.05). These trends indicate that formulations containing adequate amounts of HPMC K100 and moderate-to-high DES levels (approximately 3.0–3.4 g) generate more compliant films, whereas those with limited DES content tend to be more brittle. The DoE findings support this interpretation: HPMC K100 (X₁) produced a strong positive linear effect on elongation, Kollicoat^® IR (X₂) demonstrated a significant quadratic influence with optimal flexibility at intermediate levels, and DES (X₃) exhibited a negative linear but positive quadratic trend, indicating improved flexibility at moderate DES levels without implying mechanistic causality.

Tensile strength values ranged from 0.02 N·cm⁻² (F84) to 0.17 N·cm⁻² (F79), with a statistically significant difference among formulations (p = 0.0005). Tukey’s post hoc analysis showed that formulation F79 exhibited significantly higher tensile strength than F34 and F84 (p < 0.01), while F73 also demonstrated greater strength than F34 and F84 (p < 0.05). In contrast, no significant differences were observed between F62 and the lower-strength formulations (F1, F34, and F84) (p > 0.05). The enhanced tensile strength observed in F79 and F73 was associated with the specific compositional balance of HPMC K100, Kollicoat^® IR, and DES, which supports stronger polymer chain interactions and improved film cohesiveness.

Overall, the mechanical performance results indicate that formulations exhibiting higher tensile strength, particularly F79, achieved a balanced mechanical profile between rigidity and flexibility, offering sufficient stability during storage and application without compromising patient comfort. These findings further support the notion that DES may function in a plasticizer-like manner at moderate concentrations by increasing polymer chain mobility and reducing intermolecular forces. However, when used at excessively high levels, DES may weaken film structure, underscoring the importance of optimizing its proportion to achieve the desired mechanical properties [61].

3.4.4. Mucoadhesion Studies

The mucoadhesive strength of the films ranged from 0.012 to 0.015 mJ·cm⁻² (Table 10, Figure 3), values sufficient to maintain film attachment to the buccal surface under standardized experimental conditions and resist displacement during routine oral movements [3]. One of the highest adhesion values (0.015 mJ·cm⁻²) was observed for formulation F79, which contained a moderate amount of HPMC K100 (3.562%), a low proportion of Kollicoat^® IR (0.26%), and a low-to-medium DES level (3.0 g). This composition is consistent with the formulation trends identified within the DoE framework.

In contrast, F62 and F73 exhibited the lowest adhesion values (0.012 mJ·cm⁻²). Although these formulations contained relatively high levels of HPMC K100 (5%), the elevated concentration of Kollicoat^® IR in F73 (0.938%) was associated with reduced mucoadhesive performance.

Consistent with the DoE findings, increasing the proportion of Kollicoat^® IR (X₂) led to a reduction in mucoadhesive strength, showing a negative association with adhesion performance. In contrast, both HPMC K100 (X₁) and DES (X₃) exhibited significant quadratic effects, with the greatest adhesion occurring at intermediate concentrations, while very low or very high levels were less favorable. The absence of significant interaction terms indicates that the factors influenced mucoadhesion largely independently. One-way ANOVA revealed no statistically significant differences among the formulations (p = 0.3542), and the model’s explanatory power was modest (R² = 0.3855). The lack of significant differences between groups (p > 0.05) suggests that the small variations observed between predicted and experimental mucoadhesion values were due to normal experimental variability rather than systematic errors. Accordingly, the DoE model for this response is considered suitable for descriptive trend analysis and exploratory formulation mapping, rather than predictive optimization.

3.4.5. Swelling Studies

The swelling index of the buccal films ranged from 118.55 ± 50.00% to 181.34 ± 25.83% (Table 10), demonstrating their ability to absorb moisture and form a hydrated gel layer associated with matrix hydration and structural integrity [57]. The highest swelling was recorded for F1 (181.34%), which contained the greatest proportion of HPMC K100 (5.0%), low Kollicoat^® IR (0.26%), and a moderate DES level (3.012 g). This observation is consistent with the formulation trends identified within the DoE framework, where higher polymer content is associated with increased water uptake. In contrast, F84 exhibited the lowest swelling (118.55%), corresponding to its lower HPMC K100 level (2.016%) and relatively higher DES content (3.422 g). The association between higher DES content and reduced swelling is also consistent with the observed DoE trends. In accordance with the model results, HPMC K100 (X₁) showed a significant positive linear effect, increasing swelling through enhanced hydrophilicity and gel-layer formation. DES (X₃), on the other hand, demonstrated significant negative linear and quadratic effects, suggesting that excessive DES levels impede hydration by modifying polymer–water interactions. Kollicoat^® IR (X₂) displayed a negative but non-significant trend, which may reflect its tendency to disrupt matrix cohesion during hydration. None of the interaction terms reached significance, indicating that the factors acted independently. One-way ANOVA showed no significant differences among the formulations (p = 0.3396), and the R² value (0.3677) indicated moderate explanatory power. Overall, these results support the use of the DoE model for descriptive trend analysis and exploratory formulation mapping and suggest that minor differences between predicted and observed swelling values arise from inherent variability in swelling measurements rather than systematic model inaccuracy.

3.4.6. Moisture Loss Percentage

The moisture loss of the buccal films differed notably across formulations, ranging from 21.29% to 32.94% (Table 10). Formulation F34 displayed the highest moisture loss. From a formulation perspective, films containing higher total polymer proportions and lower DES levels (such as F73) exhibited better resistance to moisture evaporation. In contrast, formulations with lower total polymer concentrations (e.g., F34, F84) demonstrated higher moisture loss. From these formulations perspective, increasing DES content appeared to be associated with a reduction in moisture loss. These findings highlight the importance of optimizing the polymer–DES ratio to achieve an appropriate balance between mechanical robustness and moisture retention. Such a balance is essential not only for maintaining structural stability during storage but also for supporting patient comfort and acceptability upon administration [61].

3.4.7. FT-IR Analysis

FT-IR analysis was performed exclusively on the blank and curcumin-loaded forms of the selected optimal formulations (F34 and F62), which were later identified as lead formulations based on the release study results. FT-IR spectra were recorded on a % transmittance scale over 4000–400 cm⁻¹; band positions are reported as transmittance minima (i.e., absorbance maxima) (Figure 4). The spectrum of the neat drug displayed the expected markers for curcumin: a broad O–H envelope centered at 3506 cm⁻¹, a conjugated C=O/C=C band at 1623 cm⁻¹, aromatic C=C bands at 1610 and 1495 cm⁻¹, an enolic/aryl C–O band at 1273 cm⁻¹, and a methoxy C–O–C band at 1025 cm⁻¹ [50]. The FT-IR spectrum of the blank films showed characteristic bands for the mixture composed of polymer components. The broad band centered at ~3340–3360 cm⁻¹ corresponds to O–H stretching vibrations, indicating extensive intermolecular hydrogen bonding between the cellulosic backbone of HPMC and the PVA/PEG units of Kollicoat^® IR [31]. The peak observed at ~1650 cm⁻¹ is attributed to the bending vibration of water molecules absorbed in the hydrophilic polymer matrix. This peak is considered a characteristic feature of such hydrocolloid systems. In the fingerprint region, the strong band at ~1060 cm⁻¹ arises from the C–O–C stretching vibrations of HPMC and the ether linkages in the Kollicoat^® IR copolymer structure [31]. No unexpected peaks were detected in the 1700–1750 cm⁻¹ region, confirming the chemical stability of the matrices.

In the drug-loaded film F34, curcumin markers were retained but overlapped with matrix bands and shifted systematically: O–H moved from 3506 to ~3342 cm⁻¹ (downshift and broadening), the conjugated C=O/C=C band appeared at 1650 cm⁻¹ (Δν = +27 cm⁻¹ vs. neat drug; also present at 1650 cm⁻¹ in the blank), the enolic/aryl C–O band at 1290 cm⁻¹ (Δν = +17 cm⁻¹), and the methoxy C–O–C at 1040 cm⁻¹ (Δν = +15 cm⁻¹), while the aromatic C=C bands remained at 1610 and 1495 cm⁻¹ [50]. The coincidence of the 1650, 1290, and 1040 cm⁻¹ peaks with those of the F34-curcumin-free spectrum indicates that the apparent shifts arise primarily from spectral overlap with the hydroxyl-rich polymer matrix rather than the formation of new functional groups. No additional bands were detected in the 1700–1750 cm⁻¹ range, arguing against oxidative or esterification-type transformations during processing [31,62].

The same pattern was observed for formulation F62, O–H centered at ~3348 cm⁻¹ (blank F62-curcumin-free ~3357 cm⁻¹), the conjugated C=O/C=C band at 1650 cm⁻¹ (Δν = +27 cm⁻¹; also present in the blank), enolic/aryl C–O at 1290 cm⁻¹ (Δν = +17 cm⁻¹), and methoxy C–O–C at 1040 cm⁻¹ (Δν = +15 cm⁻¹), with aromatic C=C bands preserved at 1610 and 1495 cm⁻¹. Prominent matrix bands in the fingerprint region were again centered near ~1060–1090 and ~1040 cm⁻¹. Similarly to F34, the coincidence of band positions between F62 and F62-curcumin-free, together with the absence of new peaks in the 1700–1750 cm⁻¹ region, indicates that curcumin was incorporated without detectable chemical degradation [31,62].

Taken together, the persistence of all curcumin marker bands, the matrix-matching positions in the loaded versus blank films, and the pronounced broadening/downshift of the hydroxyl envelope support a noncovalent mode of interaction dominated by hydrogen bonding and dipolar contacts between curcumin and the polymeric film environment. These interactions rationalize the small systematic band shifts (particularly at 1650, 1290, and 1040 cm⁻¹) and are consistent with the physical incorporation of the drug rather than chemical modification [31,50,62].

3.5. In Vitro Drug Release Studies

The in vitro release of curcumin from the DoE optimized buccal films (F1, F34, F62, F73, F79, and F84) was evaluated using a paddle over disk apparatus (USP apparatus V). Over 24 h, cumulative release followed the rank order: F34 (91.06%) > F73 (88.99%) ≈ F62 (88.95%) > F1 (72.29%) > F84 (71.51%) > F79 (68.66%) (Figure 5). Specifically, F84 reached equilibrium much earlier than the other films, with curcumin release plateauing markedly after approximately 6–7 h. Formulations F1, F79, and F84 exhibited comparatively lower total release values at 24 h (around 70%). In contrast, formulations F34 and F62 demonstrated a more controlled and sustained release profile, characterized by a gradual increase over the first ~12 h and reaching approximately 90% curcumin release at 24 h. This release behavior is consistent with a sustained-release profile under the tested in vitro conditions. Therefore, considering the total amount released, the temporal evolution of release rates, and the overall continuity and reliability of the release profiles, kinetic modeling was carried out exclusively on formulations F34 and F62.

All buccal film formulations exhibited a biphasic release profile characterized by an initial release phase followed by a significantly slower second phase as the system approached saturation. Since F34 and F62 showed sustained and reproducible release profiles, kinetic modeling was conducted for these two formulations only. The second phase of the release profile (12–24 h) for both films showed minimal additional drug release, representing a near-plateau region with a significantly reduced concentration gradient (~6% for F34 and ~12% for F62). For this reason, kinetic analyses (zero-order, first-order, Higuchi, and Korsmeyer-Peppas) were applied only for the initial release phase (0–12 h). According to the release kinetics, high correlation coefficients (R² > 0.99) indicated a good fit to the experimental data (

Table 11. Kinetic modeling results of the lead buccal films.

Code	Zero-Order (R2)	First-Order (R2)	Higuchi (R2)	Korsmeyer-Peppas
				n	R2
F34	0.9384	0.7800	0.9806	0.6413	0.9601
F62	0.9654	0.8026	0.9654	0.5907	0.9780

). Higuchi kinetics showed the highest agreement in formulations F34 (R² = 0.9806) and F62 (R² = 0.9654), suggesting that curcumin release is predominantly governed by diffusion processes under the tested conditions. The n values obtained in the Korsmeyer–Peppas model (F34: 0.6413; F62: 0.5907) point to a hybrid mechanism, classified as anomalous diffusion, where polymer matrix relaxation contributes to the process in addition to Fickian diffusion [63,64]. In contrast, first-order kinetics showed low agreement (F34: 0.7800; F62: 0.8026), indicating that the release cannot be explained by a concentration-dependent dissolution process. The high agreement of zero-order kinetics in both formulations (F34: R² = 0.9384; F62: R² = 0.9654) is consistent with a sustained-release profile during the first 12 h. Taken together, all data indicate that the release behavior of formulations F34 and F62 is governed by a diffusion-dominant mechanism [65]. Overall, kinetic evaluation has shown that the F34 and F62 formulations provide a balanced release profile, making these formulations appropriate candidates for further formulation and preclinical investigations, and therefore suitable to be carried forward as lead formulations in subsequent studies.

3.6. Short-Term Stability Studies

Short-term stability testing was performed on the two formulations identified as lead formulations in the release study (F34 and F62). Films (1.5 × 1.5 cm²) were sealed in airtight resealable pouches and stored for one month under four conditions: 25 °C, 40 °C, 4 °C, and controlled light exposure. After one month, the residual curcumin content and pH were measured in triplicate. The results are summarized in

Table 12. Drug content of the buccal films after one month under different conditions.

Condition	Time (Day)	Drug Content (% ± SD)
		F34	F34 DES-Free	F62	F62 DES-Free
25 ± 2 °C and 60% RH	0	85.00 ± 4.00	90.93 ± 4.26	90.64 ± 2.27	79.55 ± 2.89
	30	69.90 ± 1.50	62.68 ± 14.60	73.87 ± 12.49	83.27 ± 4.15
40 ± 2 °C and 75% RH	0	85.00 ± 4.00	90.93 ± 4.26	90.64 ± 2.27	79.55 ± 2.89
	30	18.19 ± 2.56	62.44 ± 1.32	20.03 ± 1.56	69.72 ± 0.13
4 ± 1 °C	0	85.00 ± 4.00	90.93 ± 4.26	90.64 ± 2.27	79.55 ± 2.89
	30	78.04 ± 2.98	102.45 ± 11.60	84.60 ± 15.55	78.77 ± 1.33
Light exposure	0	85.00 ± 4.00	90.93 ± 4.26	90.64 ± 2.27	79.55 ± 2.89
	30	17.33 ± 2.63	70.30 ± 8.05	29.83 ± 4.88	82.31 ± 11.42

RH: relative humidity.

.

The pH values of formulations F34 and F62 remained stable under different storage conditions for one month, indicating good physicochemical stability of the formulations. At room temperature (25 °C), both formulations showed minimal variation, maintaining pH values between 6.61 ± 0.01 and 6.64 ± 0.02 for formulation F34 and between 6.62 ± 0.05 and 6.63 ± 0.04 for formulation F62. Similar consistency was observed at refrigerated (4 °C) and elevated (40 °C) temperatures, as well as under light exposure. All pH values remained within the physiological range of the buccal mucosa, suggesting that the films are unlikely to cause mucosal irritation. These results confirm that the incorporation of DES and polymer components did not adversely affect the pH stability of buccal films during storage.

After the DES-curcumin solution was combined with the aqueous polymer phase, ≥90% of the curcumin remained in the polymer mixture for 0–48 h prior to casting, indicating that the curcumin in this solution did not experience any DES-induced reduction. The pH remained narrowly within ~6.5–6.7 across conditions (Table 11), which is the curcumin’s most stable and mucosally compatible pH [50], arguing against pH-driven degradation. After the solvent casting and drying process, the initial (t = 0) assay values were already close to 90% (F34: 85.00 ± 4.00%; F62: 90.64 ± 2.27%), consistent with loss during prolonged ambient drying and/or partial analytical under-recovery from the solid matrix. Over 1 month, the stability was strongly dependent on the conditions. Refrigerated storage resulted in modest declines (F34: −8.2%; F62: −6.7%). At 25 °C/60% RH, losses increased (F34: −17.8%; F62: −18.5). In contrast, 40 °C/75% RH and controlled light exposure resulted in a pronounced loss of curcumin in the DES films.

DES-free control films have demonstrated higher apparent chemical stability under certain stressful conditions (e.g., 40 °C and light). However, they have not exhibited fundamental characteristics of buccal film properties such as poor content homogeneity and insufficient flexibility. Functionally, films prepared with DES are more advantageous due to film flexibility and drug solubility that cannot be achieved in DES-free control films. DES systems with hygroscopic properties can cause stabilization problems by facilitating oxidation and hydrolytic pathways during prolonged solvent casting and drying, followed by storage [18]. The data suggest that DES does not trigger immediate curcumin loss (stable pre-casting mixture), while decreases after film formation are driven by process and storage exposures.

For industrial-scale production or preclinical process optimization, future studies aim to prioritize stability-focused processes and formulation controls. Developing production processes under dark and low-humidity conditions, shortening the solvent casting and drying steps, using packaging that protects against moisture, oxygen, and light, and using substances such as antioxidants should be important steps in the development of DES-containing curcumin-loaded buccal films [18].

In summary, films prepared with DES are beneficial in terms of mechanical properties and curcumin solubility, but exposure to moisture and oxygen during prolonged processing increases the risk of instability. Future implementation of the above controls should preserve the advantages of DES and yield chemically stable buccal films.

3.7. Ex Vivo Permeation Study

Based on the permeation study findings, over a 24 h period, the cumulative amount of curcumin that permeated through the tissue remained below 2% of the applied dose for both formulations, with most of the drug retained within the buccal membrane (Figure 6), as evaluated for the lead formulations (F34 and F62). This limited passage is consistent with the high lipophilicity of curcumin (Log P ≈ 3.0) [50]. Although its lipophilic character facilitates penetration into the mucosal tissue, it also restricts complete transmembrane transport [66]. As a result, curcumin released from the films tends to accumulate within the buccal tissue rather than permeate through it.

In a study, Ferreira and colleagues (2019) [67] reported that the nano-based buccal film formulations they developed exhibited extremely limited permeation of curcumin through porcine oral mucosa, with no detectable levels in the receptor compartment after 24 h; retention within the mucosa was reported as 6.99%. In another study, Tolentino et al. (2024) [8] also reported that, in mucosal permeation studies of the chitosan-based buccal films they developed, curcumin levels in the receptor compartment remained below the limit of quantification, indicating that the drug was localized within the mucosal tissue. These observations are consistent with our findings and reflect a similar retention-dominant permeation pattern reported across different buccal film systems.

Such retention is therapeutically desirable for conditions like aphthous ulcers and oral lichen planus [4,6], where localized drug action is preferred. Sustained localization within the mucosa supports higher drug concentrations at the site of inflammation while reducing systemic exposure, thereby minimizing potential toxicity and avoiding extensive first-pass hepatic metabolism [66]. Accordingly, the low permeation flux observed for F34 and F62 supports their suitability as locally acting mucoadhesive delivery systems, suggesting their potential as preliminary therapeutic platforms rather than providing direct clinical outcomes.

3.8. Cell Culture Studies

3.8.1. Cytotoxicity and Nitric Oxide Level

The effects of 24 h exposure to varying doses of blank and curcumin-loaded films on both L929 and RAW 264.7 cell lines were evaluated. As illustrated in Figure 7, the L929 cell viability profiles, determined by the MTT assay, showed dose-dependent changes for all samples. Importantly, neither the blank nor the loaded films induced cytotoxicity in L929 cells, with relative cell viability remaining above 70% across all tested concentrations.

The cytotoxicity and anti-inflammatory activity of curcumin-loaded (F34 and F62) and blank buccal films were systematically evaluated using RAW 264.7 macrophage cell lines. The results are shown in Figure 8A; cell viability in all treatment groups remained above 70% relative to the untreated control across the tested concentrations, fulfilling the cytotoxicity threshold recommended by ISO 10993-5 [41] for biocompatible formulations. These results indicate that both the blank and curcumin-loaded films are non-cytotoxic within the studied concentration range and are therefore suitable for further biological assessments. There was no significant difference between blank and loaded films (p > 0.05).

Regarding anti-inflammatory efficacy, significant differences were observed between the tested groups in terms of nitric oxide production (Figure 8B). According to the results, at a dose of 1 mg/mL, F34 curcumin-loaded film extract produced a stronger inhibition of nitrite accumulation (56%) than the reference nonsteroidal anti-inflammatory drugs indomethacin and L-NAME. This effect was substantially greater than that of 1 mg/mL F62 curcumin-loaded film extract, where the inhibition was approximately 47%. In particular, the curcumin-loaded films demonstrated an ability to reduce LPS-induced nitrite levels. The incorporation of curcumin significantly enhanced the anti-inflammatory effect compared to that of the blank film extractions. For both loaded films, the higher concentration showed a highly significant reduction in nitrite levels (p **** < 0.001), confirming the dose-dependent anti-inflammatory effect of the released curcumin.

3.8.2. Direct & Indirect Contact Assay

Figure 9A,B demonstrated a significant reduction in cell viability in the positive control group compared to the negative control group (p **** < 0.001). Critically, all extractions, F34 and F62 curcumin-loaded and blank buccal films, maintained a relative cellular viability above the 70% cytocompatibility threshold. Ultimately, all formulations exhibited strong cytocompatibility profiles, with cellular viability exceeding 80% in the skin cytocompatibility test, both indirectly and through direct contact.

3.8.3. DPPH Radical-Scavenging Activity

The radical scavenging activity of F34 and F62 curcumin-loaded and blank buccal film extracts at a dose of 1 mg/mL is presented in

Table 13. Radical scavenging activity of film extracts.

Samples	Parts	RSA (%)	n
F34	Blank	65	3
	Loaded	78	3
F62	Blank	63	3
	Loaded	72	3

RSA: Radical scavenging activity. DPPH radical scavenging activity assays were performed in triplicate (n = 3).

. The data demonstrated that curcumin incorporation was associated with an increased radical-scavenging capacity in both film formulations (F34 and F62). This observation indicates that curcumin retains its functional antioxidant activity following incorporation into the film matrix and subsequent extraction, consistent with its known physicochemical properties. Furthermore, F34 curcumin-loaded extraction exhibited a higher total antioxidant activity than F62 curcumin-loaded extraction. These findings represent preliminary functional screening observations and serve as supportive indicators of retained antioxidant functionality, rather than quantitatively validated therapeutic efficacy.

3.8.4. Relative Wound-Healing Capacity

The proliferative and migratory effects of the buccal film extractions on L929 cells were quantified using a wound-healing assay over 24 h (Figure 10A,B). Compared to the control group, which exhibited a relative wound-healing capacity (RWHC) of 74%, all tested concentrations of both blank and curcumin-loaded films (F34 and F62) showed increased cell migration and closure under the tested experimental conditions. Specifically, the F34 curcumin-loaded film at 1 mg/mL demonstrated the highest efficacy, yielding an RWHC of 83%, representing a higher closure value compared to the control. Furthermore, a dose-dependent increase in wound closure was evident, with the highest RWHC values consistently achieved at the maximum tested concentrations. An increasing trend in wound closure was observed with increasing formulation concentration within the tested range; however, this was not intended as a clinical dose–response analysis. Additionally, the curcumin-loaded film extractions were observed to support cell migration behavior in vitro, confirming their positive effect on the L929 cell migration, which is necessary for tissue repair.

The local applicability of curcumin for oral wound healing is also supported by the literature. In a clinical study conducted after periodontal surgery, mucoadhesive films containing curcumin significantly reduced postoperative pain and inflammation and ensured a complication-free healing process in all patients [68]. Guo et al. (2025) [69] developed a curcumin-loaded polyurethane-based adhesive oral patch and reported that this system showed a significant wound-healing effect in oral ulcers due to its antioxidant and antibacterial properties. The study demonstrated that the curcumin-loaded patch adhered to the tissue for a long time, accelerated mucosal epithelialization, and significantly supported ulcer healing. These findings demonstrate that curcumin’s anti-inflammatory and tissue-regenerating properties can accelerate wound healing in the oral mucosa through local action, which is consistent with the in vitro wound-healing performance observed in our curcumin-loaded buccal films.

4. Conclusions

This study successfully developed mucoadhesive buccal films containing curcumin by employing a ChCl:PG (1:4) DES system in combination with a DoE-guided formulation design approach. The selected DES not only markedly improved curcumin solubility but also enabled its direct incorporation into the polymer matrix without the need for surfactants or additional complex processing steps. Following model-based numerical optimization within the DoE framework, six model-predicted optimal formulations with the desirability approach were prepared and evaluated experimentally, and formulations F34 and F62 were identified as lead candidates based on their favorable mucoadhesive strength, mechanical flexibility, controlled swelling behavior, and sustained in vitro release characteristics. Although mucosal permeation of curcumin remained low, consistent with its physicochemical profile and the objective of achieving localized therapy, both films exhibited effective mucoadhesion and prolonged attachment in ex vivo bovine buccal tissue, supporting their suitability for localized buccal delivery. Additionally, it was observed that buccal films could offer functional potential in localized applications due to their anti-inflammatory, antioxidant, and wound-healing effects demonstrated in cell culture studies. Overall, integrating DES technology as a green solvent with DoE-guided formulation design provides a promising pathway for developing efficient delivery systems for poorly water-soluble compounds such as curcumin. The buccal films developed in this study show strong potential for further preclinical investigation as localized buccal delivery platforms. The present work is limited by the absence of in vivo validation and statistically powered biological efficacy studies. Accordingly, future investigations involving in vivo evaluation, dose–response analysis, and integrated translational modeling will be required to establish clinical relevance and therapeutic applicability.