A comparative study on stevioside- and rubusoside-assisted micellar solubilization of clofazimine for physicochemical and in vitro release properties

Abstract

This study compares the micellization and encapsulation abilities of stevioside (STE) and rubusoside (RUB) for enhancing the solubility, stability, and bioavailability of clofazimine (CFZ), a poorly water-soluble antibiotic. CFZ-loaded micelles were prepared and characterized by FT-IR, XRD, DSC, SEM, and NMR. RUB showed a lower critical micelle concentration and significantly improved CFZ solubility (up to 2.06 mg/mL) compared to STE. CFZ–RUB complexes offered greater protection against UV and oxidative degradation. In vitro release tests indicated sustained and enhanced CFZ release in simulated gastrointestinal fluids. CFZ–RUB also showed higher permeability across Caco-2 cell monolayers (62.5 %) and reduced cytotoxicity versus free CFZ. These results suggest that RUB's amphiphilic structure facilitates stable micelle formation and efficient drug loading. Overall, RUB presents a promising natural solubilizer and oral delivery vehicle for lipophilic bioactives in food and pharmaceutical applications, providing new insights into the structure–function roles of steviol glycosides.

Graphical abstract

Highlights

• •Rubusoside (RUB) formed clofazimine (CFZ) micelles more efficiently than stevioside.
• •RUB enhanced aqueous solubility of CFZ by up to 2400-fold.
• •CFZ–RUB micelles showed superior stability against UV and oxidative stress.
• •RUB micelles improved intestinal permeability and reduced CFZ cytotoxicity.
• •RUB offers a natural and effective oral carrier for poorly soluble bioactive.

1. Introduction

Clofazimine [CFZ, N,5-bis(4-chlorophenyl)-3-propan-2-yliminophenazin-2-amine] is an antimicrobial and anti-inflammatory agent used to treat mycobacterial infections, including leprosy and tuberculosis. CFZ is a lipophilic, low-molecular-weight antibiotic with low aqueous solubility (<0.001 mg/L) and limited bioavailability (Bodart et al., 2020). The absorption rate of orally administered CFZ in humans ranges from 45 % to 62 % (Nugraha et al., 2021). According to WHO guidelines for multidrug-resistant tuberculosis, 100 mg of CFZ is administered once daily regardless of patient weight (Bodart et al., 2020). Strategies to improve CFZ solubility include forming complexes with cyclodextrins or cucurbituril, modifying molecular structures through salt formation, or using lipid carriers, such as oil-wax-based microcrystalline suspensions, as another method (Bodart et al., 2020; Nugraha et al., 2021).

Approximately 70 % of new drug candidates exhibit low solubility, and the solubility of a drug is a critical factor affecting its dissolution rate (Ku and Dulin, 2012). Poor solubility leads to low bioavailability of orally administered drugs, limiting their effectiveness and absorption. This often necessitates increased dosages to reach therapeutic levels, which can induce gastrointestinal toxicity and reduce patient compliance (Kawabata et al., 2011; Kumar et al., 2013). Improving the bioavailability of poorly soluble drugs is a major challenge in oral drug design. Bioavailability is influenced by aqueous solubility, drug permeability, dissolution rate, first-pass metabolism, systemic metabolism, and susceptibility to efflux mechanisms (Le Garrec et al., 2004; Savjani et al., 2012). Low aqueous solubility, slow dissolution rates, low permeability, and poor absorption through biological membranes restrict pharmacological use. Various methods to enhance solubility include modifying molecular structures, reducing particle size, forming complexes, employing natural product-based nanotechnology, micelle solubilization, using prodrugs, or adjusting pH (Alshamrani et al., 2022; Kawabata et al., 2011).

Steviol glycosides (SGs) are natural sweeteners derived mainly from Stevia rebaudiana Bertoni and Rubus suavissimus S. Lee, which have long been used as safe natural sweeteners and folk medicines. Since their approval as food and drug additives by international regulatory authorities such as the FAO, WHO, and FDA, SGs have been widely applied in food, beverages, and pharmaceuticals due to their high sweetness intensity, low caloric value, and excellent safety profile (Lan et al., 2019; Nguyen et al., 2014; Singh et al., 2019). SGs share a steviol aglycone backbone but differ in the number and position of glucose residues, leading to distinct physicochemical and functional properties (Shibata et al., 1995). Among them, stevioside (STE) and rubusoside (RUB) are two representative components: STE contains three glucose units, while RUB contains two. Although STE is more abundant and economical, RUB exhibits superior amphiphilic characteristics that enable it to act as a natural solubilizer for poorly water-soluble bioactive compounds. The amphiphilic structure of RUB, comprising a hydrophobic steviol skeleton and hydrophilic glucose moieties, facilitates micelle formation and enhances the solubility of a wide range of lipophilic compounds, including paclitaxel, curcumin, resveratrol, and vitamins E and D₂ (Liu et al., 2015; Wan et al., 2013; Zhang et al., 2016; Zhou et al., 2021). In addition to its effective solubilizing ability, RUB possesses excellent biocompatibility, low toxicity, and multifunctional pharmacological activities such as anti-inflammatory, anti-hyperglycemic, and antioxidant effects (Guan et al., 2020; Wang et al., 2020). These attributes make RUB a particularly promising amphiphilic excipient, offering both formulation versatility and biological safety superior to other steviol glycosides such as STE.

Therefore, this study aims to investigate the micellization behavior of CFZ using two structurally distinct SGs/STE and RUB to evaluate their potential as natural solubilizing agents. By comparing the solubility enhancement effects of STE and RUB, which differ in their glycosylation patterns, this work seeks to elucidate how subtle structural variations influence micelle formation and drug encapsulation efficiency. The findings are expected to provide fundamental insights into the structure–function relationship of SGs in micellar drug delivery systems and contribute to the development of natural, biocompatible platforms that improve the solubility of poorly water-soluble drugs.

2. Materials and methods

2.1. Materials and chemicals

CFZ was purchased from Sigma-Aldrich (St. Louis, MO, USA). STE (purity: > 95 %) was purchased from ALFS Korea Co., Ltd. (Kyonggi-do, Korea). RUB was synthesized by beta-glucosidase from Lactobacillus plantarum GS100 using our laboratory (Ko et al., 2022). The solvents used in high-performance liquid chromatography (HPLC) were of HPLC grade and were acquired from Duksan General Science Co., Ltd. (Seoul, Korea).

2.2. Preparation of CFZ-STE and CFZ-RUB complexes

The CFZ-STE and CFZ-RUB complexes were prepared following previously reported methods with slight modifications (Hong and Kim, 2023). The CFZ-STE and CFZ-RUB complexes were prepared by adding an excess amount of CFZ to mixtures of ethanol and water (1:1, v/v) containing STE and RUB. The samples were subjected to ultrasonic treatment for 30 min, repeated twice, and stirred continuously at 25 °C for 48 h until equilibrium was reached. The resulting solution was filtered through a 0.45 μm pore size filter membrane and then freeze-dried. Freeze-dried CFZ-STE and CFZ-RUB were used as the samples in all experiments. The CFZ concentration was determined by HPLC analysis. The HPLC analysis condition was as follows: HPLC system (LC-20 CE; Shimadzu, Kyoto, Japan) equipped with a UV detector (SPD-40; Shimadzu); Agilent ZORBAX 300SB-C18 column (4.6 × 150 mm, Agilent, Palo Alto, CA, USA); mobile phase was the gradient conditions of water with 0.1 % trifluoroacetic acid (TFA) (65–2 %, v/v) and acetonitrile with 0.1 % TFA (35–98 %, v/v); flow rate, 1 mL/min; detection time, 30 min; column temperature, 45 °C; injection volume, 10 μL. The detection of CFZ was recorded at 280 nm.

2.3. Critical micelle concentration (CMC)

The CMC values of STE and RUB micelles were determined by fluorescence spectroscopy using pyrene as a molecular probe (Basu Ray et al., 2006). Fluorescence emission spectra of pyrene (6 × 10⁻⁷ M) were recorded at various concentrations of STE and RUB (0–10 mg/mL) using a fluorescence spectrophotometer with excitation wavelengths of 378 and 394 nm. The intensities of the third (I₃ at 394 nm) and first (I₁ at 378 nm) vibronic bands were measured. The intensity ratio (I₃/I₁) of pyrene emission was calculated to determine the CMC.

2.4. Apparent solubility test

The apparent solubility of CFZ in STE and RUB was evaluated. Aqueous solutions of STE and RUB (0–60 mg/mL) were prepared, and an excess amount of CFZ was added to phosphate buffer (pH 7.4) in 10 mL amber-capped tubes. The suspensions were sonicated twice for 30 min and then stirred at room temperature for 48 h to reach equilibrium. The samples were filtered through 0.45 μm membrane filters, diluted appropriately, and analyzed at 280 nm using an HPLC method (see Section 2.2).

2.5. Octanol-water distribution coefficient

The octanol–water distribution coefficient (log D) was determined using the classical shake-flask procedure (Andrés et al., 2015). Briefly, 5 mg of each compound (CFZ, STE, RUB, CFZ–STE, and CFZ–RUB) was dissolved in 1 mL of phosphate buffer (pH 7.4). An equal volume of 1-octanol was then added, and the resulting biphasic system was vortexed for 1 h at room temperature. After vortexing, the mixture was allowed to stand undisturbed for 2 h to facilitate phase separation. Aliquots of the upper (octanol) and lower (aqueous buffer) phases were carefully withdrawn with a pipette, and the concentrations of CFZ in each phase were quantified by HPLC according to the method described in Section 2.2. The log D values were calculated using the following Eq. (1):

$$\mathit{Log}D=\mathit{Log}\left(\frac{C_{\mathit{oct}}}{C_{\mathit{aq}}}\right)$$ (1)

C_oct: concentration in the octanol sample (corrected for dilution).

C_aq: concentration in the aqueous sample (corrected for dilution).

2.6. Physicochemical properties

2.6.1. Fourier-transform infrared (FT-IR) analysis

FT-IR spectra of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB were obtained using a PerkinElmer Spectrum 400 spectrophotometer (Waltham, MA, USA) over the range of 4000–400 cm⁻¹.

2.6.2. X-ray diffraction (XRD) analysis

XRD patterns of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB were recorded using an X'Pert PRO MPD diffractometer (PANalytical, Netherlands) equipped with Cu-Kα radiation (λ = 0.154 nm), operating at 50 kV and 50 mA. Diffraction patterns were collected over a 2θ range of 5–90° at a scanning rate of 1°/min.

2.6.3. Differential scanning calorimetry (DSC) analysis

DSC analysis was conducted to confirm the formation of the complexes. Approximately 2 mg of each sample (CFZ, STE, RUB, CFZ-STE, and CFZ-RUB) was sealed in an aluminum pan and heated from 25 °C to 300 °C at a rate of 10 °C/min under a nitrogen atmosphere using a DSC3 instrument (Mettler Toledo, GmbH, Greifensee, Switzerland).

2.6.4. Particle size distribution (PSD)

The particle sizes of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB were measured after 10-fold dilution in distilled water. Electrophoretic mobility was measured using dynamic light scattering (DLS) on a Zetasizer Advance Pro (Malvern Instruments Ltd., Malvern, UK). Samples were taken at 25 °C.

2.6.5. Stability test

Photodegradation experiments were carried out on solutions of CFZ, CFZ-STE, and CFZ-RUB using a UV lamp (F4T5BLB, Sankyo Denki Co., Ltd., Kanagawa, Japan). At each time point, 100 μL of sample was mixed with 400 μL of 99 % DMSO to release encapsulated CFZ. Samples were placed 10 cm from the UV source and irradiated with UV-A light (315–400 nm) for 6 h at 25 ± 2 °C. Absorbance was measured at 280 nm at hourly intervals using an HPLC method (see Section 2.2).

To evaluate oxidative stability, 1 mL of each sample (containing 10 mg of CFZ, CFZ-STE, or CFZ-RUB) was treated with 100 μM ferric chloride (FeCl₃) at room temperature for 6 h. Aliquots (100 μL) were collected at specified intervals, and the absorbance at 280 nm was measured to determine the remaining CFZ using HPLC. Free CFZ in 0.02 % (w/v) DMSO and CFZ-STE and CFZ-RUB in 0.1 % (w/v) deionized water were used as controls.

2.7. Scanning electron microscopy (SEM)

The surface morphology of each freeze-dried sample (CFZ, STE, RUB, CFZ-STE, and CFZ-RUB) was observed using a field emission scanning electron microscope (JSM-7800F, JEOL, Tokyo, Japan). Samples were mounted on aluminum stubs using double-sided conductive tape and coated with platinum under vacuum using a sputter coater (Q150T-ES, Quorum Technologies, East Sussex, UK) before imaging.

2.8. Nuclear magnetic resonance (NMR)

The ¹H and ¹³C NMR spectra of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB were recorded using an INOVA-500 NMR spectrometer (Varian, Walnut Creek, CA, USA) at 500 MHz and 25 °C. Each sample (2 mg) was dissolved in 0.2 mL of DMSO‑d₆ and transferred to an NMR tube for analysis in triplicate.

2.9. Drug release properties

2.9.1. Drug release study

The release profile of CFZ was evaluated using the dialysis bag method, as previously described by Popat et al. (2014), with slight modifications. CFZ-STE and CFZ-RUB, each equivalent to 1 mg of CFZ, were suspended in 1 mL of 0.1 M phosphate buffer (pH 7.4) containing 0.5 % sodium lauryl sulfate (SLS) and placed in a dialysis bag with a molecular weight cutoff of 10 kDa. The bag was then immersed in 9 mL of the same buffer at 37 °C under constant stirring. At predetermined time intervals, 100 μL of the external medium was withdrawn and replaced with an equal volume of fresh buffer to maintain sink conditions. The concentration of CFZ in the collected samples was determined by measuring the absorbance at 280 nm using an HPLC method (see Section 2.2) and calculated using a standard calibration curve.

2.9.2. In vitro simulated digestion release analysis

The in vitro digestion was performed using simulated saliva fluid (SSF), simulated gastric fluid (SGF), and simulated intestinal fluid (SIF), prepared according to previously reported protocols with minor modifications (Hong and Kim, 2023). Briefly, 50 mg of CFZ, CFZ-STE, and CFZ-RUB were placed in a 125 mL Erlenmeyer flask and stirred continuously at 37 °C in a static model to simulate gastrointestinal conditions. Sequential digestion was conducted as follows: 10 mL of SSF was added and mixed for 5 min, after which a 1 mL aliquot was collected. Subsequently, 10 mL of SGF was added, and aliquots were taken after 30 and 60 min of incubation. Finally, 10 mL of SIF was added, and 1 mL aliquots were collected at 2 and 4 h. All samples were centrifuged at 6238 ×g for 5 min and filtered through a 0.45 μm membrane. The supernatants were stored at −80 °C until analysis. The concentration of released CFZ was quantified by measuring the absorbance at 495 nm using an HPLC method (see Section 2.2).

2.10. Cell culture

2.10.1. In vitro cell viability studies

The cytotoxicity of unloaded CFZ, CFZ-STE, and CFZ-RUB formulations was evaluated using the MTT assay on Caco-2 cells. Caco-2 cells were seeded at a density of 5 × 10⁴ cells per well in 96-well plates. After allowing the cells to adhere, the culture medium was removed, and the cells were incubated with 100 μL of fresh culture medium containing free CFZ, CFZ-STE, or CFZ-RUB at concentrations ranging from 3.13 to 50 μg/mL. For the free CFZ, stock solutions were prepared in DMSO, and the appropriate dilutions were made to achieve a final concentration of ≤4 % (v/v) DMSO in each well. After 24 h of incubation, the supernatant was removed, and 100 μL of MTT solution (0.5 mg/mL) was added to each well. The plates were incubated for an additional 2 h at 37 °C. After incubation, the MTT solution was removed, and 100 μL of DMSO was added to dissolve the formazan crystals. The optical density was measured at 570 nm (OD₅₇₀) using a UV–vis microplate reader (HIDEX Sense 425–301; Hidex, Turku, Finland). Untreated cells were considered as the negative control, representing 100 % viability, while cells treated with 4 % DMSO served as the negative control for CFZ. The cell viability was calculated using the following formula:

$$\text{Cell viability}\left(\%\right)=\left({\mathrm{OD}}_{570}\text{of cells incubated with formulation}/{\mathrm{OD}}_{570}\text{of cells incubated only with culture medium}\right)\times 100$$ (1)

2.10.2. Permeability study on Caco-2 monolayers

The permeability of CFZ, CFZ-STE, and CFZ-RUB was evaluated using a CacoReady™ kit (ADMEcell, Emeryville, CA, USA), which consists of differentiated Caco-2 cell cultures (day 21 of differentiation) on polycarbonate microporous filters in HTS Transwell-24 plates (6.5 mm diameter, 0.33 cm² area, and 0.4 μm pore diameter). Prior to the experiments, the shipping medium (5.4 mM KCl, 0.44 mM KH₂PO₄, 0.137 mM NaCl, 0.33 mM Na₂HPO₄, 1.1 mM MgCl₂·6H₂O, 0.13 mM CaCl₂·2H₂O, and 5 mM d-glucose; pH 6.5) was replaced with standard Caco-2 cell culture medium (Dulbecco's modified Eagle medium [DMEM], Lonza Walkersville, MD, USA) according to the manufacturer's protocol. The transepithelial electrical resistance (TEER) was measured before each experiment to assess the integrity of the Caco-2 monolayers using an EVOM™ epithelial voltmeter with EndOhm electrode (World Precision Instruments, Sarasota, FL, USA). The TEER values of >1000 Ω·cm² confirmed that the monolayers had formed tight junctions. For the permeability study, 0.5 mL of culture medium containing 40 μg/mL of CFZ, CFZ-STE, or CFZ-RUB was added to the apical side of the Caco-2 monolayers. The formulations were incubated for 8 h, and 0.5 mL samples were withdrawn from the basolateral compartment at predetermined time intervals (1, 2, 3, 4, 6, and 8 h) and replaced with an equal volume of fresh culture medium. The CFZ content in the samples was analyzed using HPLC (see Section 2.2). The cumulative amount of CFZ transported across the monolayers was calculated, and the apparent permeability (P_app) was determined using the following equation:

$${\mathrm{P}}_{\mathrm{app}}=\frac{\mathrm{d}Q}{\mathrm{d}t}\times \frac{1}{A{C}_0}$$ (2)

where dQ/dt is the amount of complex/drug in the basolateral compartment as a function of time (mg/min), A is the monolayer area (cm²), and C₀ is the initial concentration of complex/drug in the apical compartment (mg/mL).

2.11. Statistical analysis

Data were expressed as the mean ± standard deviation (SD), and there were triplicate experiments. All statistical analyses were performed using SPSS version 23.0 (SPSS Inc., Chicago, IL, USA). One-way analysis of variance (ANOVA) was used to assess the significance of differences among groups, followed by Duncan's multiple range test. A p-value of less than 0.05 was considered statistically significant.

3. Results and discussion

3.1. CMC of STE and RUB

CMC, a key parameter reflecting the micellization characteristics of surfactants, is commonly used to assess the solubilization capacity of glycosides, including SGs (F. Wang et al., 2020). As shown in Fig. 1A, CFZ was dissolved by STE and RUB in a concentration-dependent manner. Furthermore, the CMCs of STE and RUB were determined to be 5.20 ± 0.04 and 4.88 ± 0.06 mmol/L, respectively (Fig. 1B). This CMC of STE is similar to that reported previously for STE, which was 5.22 mmol/L (Wang et al., 2020). These results indicated that STE required higher concentration to form micelles capable of dissolving an equivalent amount of substance compared to RUB, which exhibited a lower CMC, suggesting that RUB formed micelles more efficiently than STE for encapsulating solutes.

Fig. 1.

Dissolution rate of CFZ in STE and RUB solution (A), critical micelle concentration (B) of STE and RUB, and apparent solubility test (C) of CFZ-STE and CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

3.2. Apparent solubility of CFZ-STE and CFZ-RUB

Fig. 1C presents the apparent solubility of CFZ in the CFZ–STE and CFZ–RUB systems at 25 °C. The solubility of crystalline CFZ in water is extremely low, with a reported maximum of 0.210 μg/mL (Salem et al., 2003). Both STE and RUB have been shown to enhance the apparent solubility of poorly water-soluble active compounds (Luo et al., 2022). As the concentration of STE increased up to 50 mg/mL, the solubility of CFZ increased to 0.64 mg/mL. In contrast, when the RUB concentration increased to 60 mg/mL, the solubility of CFZ reached 2.06 mg/mL. This corresponds to an approximately 2400-fold increase in solubility with RUB. However, in the STE-CFZ system, the solubility of CFZ increased only up to a certain concentration and then decreased at the maximum tested concentration of 60 mg/mL. This phenomenon suggests that at high concentrations (>50 mg/mL), STE may form micelles that encapsulate excess STE molecules rather than CFZ, thereby reducing CFZ solubility. This observation is consistent with previous findings, which indicate that increased STE concentrations can reduce the solubility of target compounds (Luo et al., 2022; Nguyen et al., 2014). Meanwhile, RUB exhibited a different trend from STE: the solubility of CFZ continued to increase with RUB concentration, suggesting that RUB forms a more stable micellar structure than STE and binds more effectively to the hydrophobic regions of CFZ.

3.3. Log D value of CFZ-STE and CFZ-RUB

Water solubility and log D are key factors in determining the partitioning behavior of chemicals in pharmaceuticals (Miller et al., 1985). As summarized in Table 1, the measured log D values were 5.76 ± 1.34 for free CFZ, −1.51 ± 0.22 for STE, and − 1.45 ± 0.34 for RUB. Upon complexation, the log D values of CFZ–STE and CFZ–RUB were reduced to 4.99 ± 1.20 and 4.11 ± 0.88, respectively, suggesting a partial decrease in lipophilicity while maintaining a considerable hydrophobic character. These findings indicate that, despite the formation of hydrophilic complexes with steviol glycosides, CFZ retains substantial lipophilic characteristics within the complex matrix. Considering that CFZ is an extremely lipophilic drug, the moderate decrease in log D observed in the present study suggests that the complexation process did not completely suppress its inherent hydrophobicity.

Table 1.

Log D values of CFZ, STE, and RUB complex.

Compounds	Log D
CFZ	5.76 ± 1.34
STE	−1.51 ± 0.22
RUB	−1.45 ± 0.34
CFZ-STE	4.99 ± 1.20
CFZ-RUB	4.11 ± 0.88

Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

This phenomenon can be explained by several complementary factors. First, the amphiphilic steviol glycosides (e.g., stevioside and rubusoside) can self-assemble or associate to form hydrophobic microdomains or micellar-like interiors capable of solubilizing lipophilic guest molecules, thereby retaining CFZ in a relatively nonpolar microenvironment. Such self-assembly and natural surfactant-like behavior have been reported for rubusoside and other steviol glycosides and have been exploited to improve the aqueous solubility of poorly water-soluble compounds (Chen et al., 2020; Zhang et al., 2017). Second, the experimentally measured distribution coefficient represents the net partitioning of both complex-bound and unbound CFZ. Because a dynamic equilibrium can exist between free drug and complexed species, the observed log D effectively reflects an averaged distribution encompassing both states rather than exclusively the free molecule. Third, the inherent diterpenoid backbone of steviol glycosides confers residual affinity for organic phases; consequently, complex formation does not necessarily eliminate lipophilic character (Dahlgren and Lennernäs, 2019). Maintaining a moderate degree of lipophilicity in CFZ–STE and CFZ–RUB systems can be advantageous: sufficient lipophilicity often facilitates transmembrane diffusion, and the simultaneous preservation of some lipophilic character together with enhanced aqueous solubility may promote intestinal permeability and absorption. Thus, the steviol glycoside-based complexation approach provides a balanced physicochemical profile for CFZ, enabling improved solubilization without compromising permeability.

3.4. Physicochemical properties of CFZ-STE and CFZ-RUB

3.4.1. FT-IR analysis

Fig. 2A presents the FT-IR spectra of CFZ, STE, RUB, CFZ–STE, and CFZ–RUB. The FT-IR spectrum of CFZ exhibited characteristic peaks, including N—O stretching frequencies at 1500–1550 cm⁻¹, N—H bending at 1625 cm⁻¹, and weak N—H stretching at 2800–3000 cm⁻¹, consistent with previous reports (Chaves et al., 2018a, Chaves et al., 2018b; Hong and Kim, 2023). In contrast, the CFZ–STE and CFZ–RUB spectra lacked the prominent N—H stretching and bending peaks observed in pure CFZ, likely due to the encapsulation of CFZ within micelles formed by STE and RUB. This encapsulation may have obscured or suppressed the characteristic peaks of CFZ. The spectral similarities between CFZ-STE and CFZ-RUB further suggest good compatibility among the formulation components. Therefore, FT-IR analysis should be interpreted alongside other spectroscopic and thermal techniques for a more comprehensive characterization of the inclusion complexes.

Fig. 2.

FT-IR spectra (A), X-ray diffractograms (B), DSC curve (C), and particle size (D) of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

3.4.2. XRD analysis

The complexation of CFZ induces changes in the diffractogram, observable as variations in signal intensity or peak positions compared to those of the isolated compound or physical mixture, as shown in Fig. 2B. The diffractogram of pure CFZ corresponds to its polymorphic forms, with characteristic peaks at 9.2°, 13.4°, 19.8°, and 21.7°, as reported in the Cambridge Structural Database (Groom et al., 2016). STE exhibited distinct crystalline peaks, while RUB was confirmed to be in an amorphous state. In the micellar complexes, CFZ–STE and CFZ–RUB, the characteristic peaks of CFZ were absent. Notably, CFZ–STE also showed a reduction in the intensity of STE's major peaks, suggesting that complex formation with CFZ diminished the crystalline signal of STE. Similarly, the diffractogram of CFZ–RUB displayed an amorphous pattern resembling that of RUB; however, unlike RUB, in which some minor peaks were present, CFZ–RUB showed no distinct CFZ-related peaks. These observations indicate that CFZ was successfully incorporated into micelles formed by STE and RUB. This result is consistent with previous reports, where the formation of micelles by STE with poorly soluble substances led to reduced or undetectable diffraction peaks.

3.4.3. DSC analysis

DSC can be used to identify the formation of inclusion complexes. The DSC thermogram of STE exhibited a single endothermic peak centered at 100 °C, attributed to the release of water. In contrast, CFZ displayed a broad endothermic peak with a maximum at 226 °C (Fig. 2C), consistent with its boiling point of 212 °C. The DSC profiles of the CFZ–STE and CFZ–RUB complexes showed substantially altered thermal behavior, with a notable weakening of the endothermic peak at 100 °C and the complete disappearance of the characteristic CFZ peak. These results suggest that CFZ was successfully entrapped within the micellar structures of STE and RUB rather than existing as a simple physical mixture. The disappearance of the CFZ melting peak indicates that the active compound was molecularly dispersed within the micellar matrix. This observation is further supported by comparisons between the thermograms of the physical mixtures and those of the inclusion complexes (Chaves et al., 2018b; Hong and Kim, 2023; Rongala et al., 2024). Collectively, these findings demonstrate that the active ingredient was effectively protected within the micellar cavities of STE and RUB.

3.4.4. Particle size analysis

STE, RUB, CFZ–STE, and CFZ–RUB were completely reconstituted in water on the day of preparation; the corresponding average particle sizes were 1209.88 ± 260.24, 197.6 ± 0.65, 310.78 ± 0.88, and 1635.98 ± 144.25 nm, respectively (Fig. 2D). These results suggest that micelle formation by STE and RUB led to the entrapment of CFZ, which influenced particle size. The average particle sizes increased upon CFZ encapsulation for both STE and RUB, indicating that micelle formation with CFZ altered their size characteristics. Notably, the initial PSD of STE was broader than that of RUB; however, the distribution narrowed upon CFZ entrapment, suggesting that micellization can stabilize PSD. In contrast, RUB exhibited an increase in PSD upon CFZ incorporation, although it maintained a more uniform distribution compared to STE. This indicates that RUB forms more structurally consistent micelles and is more efficient at encapsulating CFZ. The observed increase in particle size upon encapsulation is consistent with previous studies, which reported similar size increases for STE, Reb A, and RUB when used to entrap various compounds (Zhang et al., 2021; Zhao et al., 2020).

3.4.5. Stability analysis

The stability of free CFZ, CFZ–STE, and CFZ–RUB was evaluated under UV-induced photodegradation and FeCl₃-induced oxidative degradation to assess the ability of STE and RUB to protect CFZ from environmental stresses. To investigate the stability of CFZ encapsulated in STE and RUB, the amounts of active CFZ remaining in CFZ–STE and CFZ–RUB were measured over 6 h of exposure (Fig. 3). As controls, free CFZ, CFZ–STE, and CFZ–RUB were subjected to identical environmental stress conditions. The results revealed that degradation of free CFZ under UV exposure was significant, with over 40 % degraded within the first hour and approximately 80 % degraded after 6 h. In contrast, both CFZ–STE and CFZ–RUB showed improved photostability compared to free CFZ. However, CFZ–STE still exhibited substantial degradation (∼60 %) after 6 h of UV exposure. Notably, CFZ–RUB demonstrated superior protection, with less than 50 % degradation over the same period, indicating that RUB was more effective than STE in shielding CFZ from photodegradation.

Fig. 3.

Photodegradation (A) and chemical oxidation (B) stability of CFZ, CFZ-STE, and CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

In Fig. 3B, FeCl₃ was used as an accelerating oxidant to evaluate the oxidative stability of CFZ in the inclusion complexes. Oxidative degradation was generally less severe than photodegradation. The micelle formation with STE and RUB contributed to improved stability of CFZ under oxidative stress, supporting the conclusion that successful encapsulation inhibited the degradation of CFZ by isolating it from external oxidants. These results are consistent with previous findings on the protective effect of cyclodextrin and cyclodextran complexes (Hong and Kim, 2023; Salem et al., 2003).

Overall, our results demonstrate that CFZ–STE and CFZ–RUB effectively protect CFZ from environmental stresses such as UV irradiation and oxidation, which are commonly encountered during manufacturing, processing, storage, and consumption. These findings suggest that both CFZ–STE and CFZ–RUB have potential as formulation strategies for developing stable and effective pharmaceutical products. However, further studies are necessary to optimize these formulations for therapeutic applications and to evaluate their safety and efficacy.

3.5. SEM analysis

Morphological differences between the individual components, physical mixtures, and inclusion complexes were examined using SEM (Fig. 4). The pure CFZ and STE (Fig. 4A and B) exhibited well-defined crystalline structures, with CFZ showing a standard morphology and STE displaying rectangular particles with distinct angles. In contrast, notable morphological changes were observed in the CFZ–STE and CFZ–RUB inclusion complexes (Fig. 4D and E). Specifically, the original crystalline structures of the pure compounds were no longer evident in the complexes. Instead, an amorphous powder with slightly reduced particle sizes and no visible separation of individual components was observed in all inclusion complex samples. These morphological alterations suggest a successful interaction between CFZ, STE, and RUB, consistent with the formation of inclusion complexes. The loss of the native crystalline structure and transformation into an amorphous form may contribute to the improved dissolution rate of the inclusion complexes. Overall, the changes in surface morphology support the hypothesis that CFZ was effectively encapsulated within the micellar matrices of STE and RUB (Chen et al., 2020; Yang et al., 2022; Zhang et al., 2016).

Fig. 4.

Scanning electron microscopy of morphological characteristics (×300). (A) CFZ, (B) STE, (C) RUB, (D) CFZ-STE, and (E) CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

3.6. NMR analysis

In the ¹H NMR spectrum of clofazimine dissolved in DMSO‑d₆, two distinct doublets corresponding to the methyl groups of the isopropyl moiety were observed at approximately δ_H 1.03 and δ_H 1.05, each integrating for 6 protons (Fig. 5A, CFZ). Additionally, aromatic and amine protons appeared in the expected downfield region. The ¹³C NMR spectrum exhibited well-resolved carbon peaks, including a prominent signal near δ_H 49 assignable to the isopropyl methine carbons (Fig. 5B, CFZ). These results are consistent with the previously reported ¹H and ¹³C NMR spectra of clofazimine (Li et al., 2016; Verbić et al., 2023).

Fig. 5.

¹H NMR (A) and ¹³C NMR (B) spectra of CFZ, STE, RUB, CFZ-STE, and CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

STE and RUB in DMSO- d₆ showed typical glycosidic proton and carbon signals. The ¹H NMR spectra displayed multiple resonances between δ_H 3.0–5.5, while the ¹³C NMR spectra revealed signals ranging from δ_C 60 to δ_C 105, corresponding to various sugar ring carbons. Minor chemical shift differences between STE and RUB were attributed to their structural variations in the aglycone regions (Lan et al., 2019; Wang et al., 2015). Upon complexation with CFZ, the resulting CFZ-STE and CFZ-RUB mixtures exhibited a remarkable loss of CFZ-related signals in both ¹H and ¹³C NMR spectra (Fig. 5A and B, CFZ-STE, CFZ-RUB). All proton and carbon peaks from CFZ, except for the isopropyl methyl groups, were either significantly broadened or entirely undetectable, indicating restricted molecular mobility within the complex. Notably, the isopropyl methyl doublets remained observable around δ_H 1.05–1.10 in the ¹H NMR spectra, while a single carbon signal near δ_C 56 was retained in the ¹³C NMR spectra. These observations suggest that CFZ becomes encapsulated or immobilized within the SGs matrix upon complex formation, leading to severe line broadening due to restricted rotational freedom. In contrast, the terminal isopropyl methyl groups likely remain surface-exposed and dynamically mobile, thus preserving their NMR visibility. Additionally, the glycosidic ¹H- and ¹³C- signals of both STE and RUB exhibited slight downfield chemical shifts (typically δ 0.1–0.3) upon complexation, supporting the occurrence of non-covalent interactions such as hydrogen bonding, hydrophobic contacts, and possible π–π interactions with the aromatic moieties of CFZ.

3.7. Drug release properties

3.7.1. Drug release analysis

The drug release behavior of CFZ–STE and CFZ–RUB composites was evaluated using a dialysis bag method, as shown in Fig. 6A. The results demonstrated that CFZ–RUB exhibited the highest drug release rate, reaching up to 75 %, while CFZ–STE showed a release of up to 55 %. These findings indicate that RUB is more effective than STE in encapsulating and delivering CFZ. Although STE possesses a higher encapsulation capacity, its lower solubilization ability appears to limit the drug release efficiency.

Fig. 6.

Drug release properties (A) and in vitro simulated gastrointestinal drug release properties (B) of CFZ, CFZ-STE, and CFZ-RUB. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

Both CFZ–STE and CFZ–RUB showed a time-dependent increase in drug release, and their release profiles were comparable to those typically observed from cyclodextrins. This suggests that the solubilization capacity of the micellar aggregates and their binding affinity to CFZ significantly influenced the drug release behavior (Chaves et al., 2018a; Zhang et al., 2022). Additionally, the enhanced drug release may be attributed to the amorphous form of CFZ within the inclusion complexes. Alterations in crystallinity can lead to less ordered molecular arrangements, thereby promoting faster and more efficient drug release (Luo et al., 2022). These results support the conclusion that RUB-based micelles serve as a highly efficient delivery system for encapsulated CFZ.

3.7.2. In vitro simulated gastrointestinal release analysis

Micellization is a widely utilized strategy to enhance the bioavailability of poorly water-soluble drugs. In the current study, the release profiles of free CFZ and CFZ encapsulated within STE and RUB micelles were evaluated in a simulated gastrointestinal environment to assess the impact of these carriers on drug release kinetics. As shown in Fig. 6B, the release behavior of CFZ–STE and CFZ–RUB was initially examined in SSF (pH 7.0), where micellized CFZ was released even in the early stages of exposure. The amount of released CFZ increased significantly in SGF (pH 2.0), suggesting that the micelles formed by STE and RUB were destabilized under acidic conditions. Furthermore, the release of CFZ from both CFZ–STE and CFZ–RUB gradually increased in SIF (pH 7.0) over time. This sustained release is likely due to the progressive weakening of the micellar structure, allowing for continued drug diffusion. Notably, up to 80 % of CFZ was released in the intestinal simulation, indicating a strong potential for effective intestinal absorption. These results are consistent with previous findings that highlight the influence of particle size on release behavior. For example, Zhang et al. (2022) reported that RA–idebenone micelles, which contain larger molecules, exhibited lower release rates. In the current study, RUB-based micelles, characterized by smaller particle sizes than CFZ–STE, facilitated greater CFZ release, reinforcing the idea that particle size plays a crucial role in release efficiency.

3.8. Cytotoxicity effects of CFZ-STE and CFZ-RUB on Caco-2 cells

CFZ-STE and CFZ-RUB were designed for oral delivery in leprosy therapy. Given the high hydrophobicity of CFZ, drug absorption in the gastrointestinal tract can be problematic due to its tendency to precipitate at physiological pH, potentially causing toxicity to intestinal tissues. To evaluate the impact of CFZ-STE and CFZ-RUB on intestinal cell viability, Caco-2 cells were selected as a representative in vitro model of the intestinal epithelium. Cell viability assays were conducted using free CFZ, CFZ-STE, and CFZ-RUB (Fig. 7A). At high concentrations of free CFZ (25 and 50 μg/mL), a significant decrease in Caco-2 cell viability was observed (p < 0.0001 and p < 0.001, respectively). In contrast, exposure to CFZ-STE and CFZ-RUB nanoparticles-maintained cell viability above 90 %, with no statistically significant differences compared to untreated controls. However, the incorporation of CFZ into the STE and RUB systems resulted in a concentration-dependent reduction in cell viability, indicating that the observed cytotoxicity was attributable to the drug itself rather than the carriers. At the highest tested concentrations, CFZ-STE and CFZ-RUB resulted in cell viabilities of approximately 64 % and 73 %, respectively. Exposure to free CFZ yielded an IC₅₀ value of 23.3 μg/mL for Caco-2 cells, confirming the known high cytotoxicity of CFZ (Chaves et al., 2018b). This aligns with previous reports, which indicate that long-term oral administration of CFZ can lead to drug bioaccumulation and associated toxicity (Arbiser and Moschella, 1995; Li et al., 2016). Importantly, in this study, micellization of CFZ with STE and RUB helped mitigate drug-induced cytotoxicity for up to 24 h. These results suggest that encapsulating CFZ in CFZ-STE and CFZ-RUB may prevent in vivo drug crystallization and the formation of insoluble precipitates, as approximately 70 % of CFZ remained in its amorphous form. Therefore, these formulations show potential not only for enhancing the oral delivery of CFZ but also for reducing adverse effects by protecting intestinal cells from drug-induced toxicity.

Fig. 7.

Cell viability (A) and permeability (B) properties of CFZ, CFZ-STE, and CFZ-RUB. Data are presented as mean ± SD of triplicate samples in a representative experiment of three independent experiments. Abbreviations: CFZ; clofazimine, STE; stevioside, RUB; rubusoside.

3.9. Permeability effects of CFZ-STE and CFZ-RUB on Caco-2 cells monolayers

The in vitro bioassay using Caco-2 cell monolayers is widely recognized as a reliable model for predicting drug permeability, and it has shown a good correlation with in vivo absorption in humans (Chaves et al., 2018b). Therefore, a Caco-2 monolayer assay was conducted to evaluate the permeability of CFZ loaded in STE and RUB. The tested concentration was selected based on the results of the cytotoxicity assays, with 40 μg/mL of CFZ (equivalent to 200 μg/mL in polymer concentration) used in the nanoparticles-CFZ formulations. Cell monolayer integrity was monitored through TEER measurements throughout the experiment, and no cellular damage affecting TEER values was observed (data not shown). Initial TEER values were greater than 400 Ω·cm², as recommended for absorption studies (Chaves et al., 2018b). The addition of the samples studied did not alter these values over the 8 h experiment. The permeation profiles of CFZ-STE and CFZ-RUB across Caco-2 monolayers are shown in Fig. 7B. After culturing the Caco-2 monolayer, approximately 20.14 μg of CFZ permeated into the basolateral chamber in the CFZ-STE group, representing about 50 % of the initial CFZ concentration (40 μg/mL). In comparison, CFZ-RUB resulted in a permeation of approximately 25.33 μg, representing about 63 % of the initial amount. And P_app of CFZ-STE, and CFZ-RUB were detected at 2.7 ± 0.3 × 10⁻⁶ cm/s and 4.1 ± 0.1 × 10⁻⁶ cm/s, respectively (data not shown). These results suggest that the differences in CFZ permeation across the Caco-2 monolayers are likely due to the size-dependent uptake of nanoparticles by intestinal cells (Chaves et al., 2018a, Chaves et al., 2018b). This phenomenon could enhance the bioavailability of CFZ upon oral administration. Based on these findings, it is evident that the current therapy regimen may need to be adjusted according to the pharmacokinetics of CFZ-STE and CFZ-RUB in order to achieve the desired therapeutic effect while minimizing the risk of undesired accumulation. The developed CFZ-STE and CFZ-RUB formulations, with their protective properties, offer a significant advancement in improving CFZ intestinal permeability, which is particularly important given the high toxicity of free CFZ.

4. Conclusion

This study demonstrated the superior performance of RUB over STE in forming stable micellar systems for encapsulating the hydrophobic drug CFZ. Compared to STE, RUB exhibited a lower CMC, a higher solubilizing capacity, and greater protection against environmental stresses, such as light and oxidative degradation. The CFZ-RUB complex also showed enhanced in vitro release behavior and improved permeability across intestinal epithelial models, along with reduced cytotoxicity, suggesting better compatibility with biological systems. These findings indicate that RUB effectively encapsulates CFZ in its amorphous form, facilitating both solubility and bioavailability enhancement. Overall, RUB proved to be a highly efficient natural micellization agent and delivery platform for poorly water-soluble drugs, supporting its application in functional food and pharmaceutical formulations requiring improved oral absorption and safety profiles.

CRediT authorship contribution statement

Seong-Jin Hong: Writing – original draft, Methodology, Conceptualization. Hye-Su An: Writing – original draft, Methodology. Bo-Ram Park: Methodology, Data curation. Ki-Nam Yoon: Visualization, Supervision. Hye-Jin Kang: Visualization, Supervision. Sung Jae Shin: Supervision. Young-Min Kim: Writing – review & editing, Writing – original draft, Supervision, Conceptualization.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Data availability

Data will be made available on request.