Transdermal Insulin Delivery Using Ionic Liquid-Mediated Nanovesicles for Diabetes Treatment

Abstract

Transdermal insulin delivery is a promising method for diabetes management, providing the potential for controlled, sustained release and prolonged insulin effectiveness. However, the large molecular weight of insulin hinders its passive absorption through the stratum corneum (SC) of the skin, and high doses of insulin are required, which limits the commercial viability. We developed ethosome (ET) and trans-ethosome (TET) nanovesicle formulations containing a biocompatible lipid-based ionic liquid, [EDMPC][Lin], dissolved in 35% ethanol. TET formulations were obtained by adding isopropyl myristate (IPM), Tween-80, or Span-20 as surfactants to ET formulations. Dynamic light scattering, ζ-potential, transmission electron microscopy, and confocal laser scanning microscopy studies revealed that the nanovesicles had a stable particle size. The formulations remained stable at 4 °C for more than 3 months. ET and TET formulations containing IPM (TET1) significantly (p < 0.0001) enhanced the transdermal penetration of FITC-tagged insulin (FITC-Ins) in both mouse and pig skin, compared with that of the control FITC-Ins solution and other TET formulations, by altering the molecular structure of the SC layer. These nanovesicles were found to be biocompatible and nonirritants (cell viability >80%) in the in vitro and in vivo studies on three-dimensional (3D) artificial human skin and a diabetic mouse model, respectively. The ET and TET1 formulations were applied to the skin of diabetic mice at an insulin dosage of 30 IU/kg. The nanovesicle formulations significantly reduced blood glucose levels (BGLs) compared with the initial high BGL value (>150 mg/dL). The nanovesicle-treated mice maintained low BGLs for over 15 h, as opposed to only 2 h in the injection group. The ET and TET1 formulations reduced the BGLs by 62 and 34%, respectively, of the initial value. These ET and TET1 formulations have a high potential for use in commercial transdermal insulin patches, enhancing comfort and adherence in diabetes treatment.

1. Introduction

Insulin therapy is essential for managing blood glucose levels (BGLs) in patients with type 1 diabetes mellitus and those with advanced type 2 diabetes mellitus.¹ Subcutaneous injections remain the most common method for insulin delivery.² However, there are several problems with subcutaneous injection, including needle anxiety, pain, risk of infection, poor adherence, and peripheral hyperinsulinemia.³ Various studies have shown that needle anxiety is prevalent among both adults (∼22%) and children (∼68%).⁴ To address such limitations, various technologies have been developed for alternative insulin delivery methods that are noninvasive, safe, and painless. These methods include oral, buccal, nasal, peritoneal, and transdermal systems.⁵ Inhaled insulin was the first noninvasive method to receive approval, although the product was subsequently withdrawn from the market because of low sales and patient inconvenience.⁵ The oral route provides a patient-friendly option for insulin delivery, but several challenges still exist, including the gastrointestinal (GI) enzyme inactivation and low bioavailability of insulin.⁶ Intranasal delivery provides benefits over the oral and inhalation routes of insulin administration by bypassing GI enzymes and avoiding lung complications. However, this method still has challenges, including permeability issues and mucociliary clearance, which result in variable absorption.⁷ In addition, the buccal route is limited by the low and inconsistent absorption of insulin.⁸

Transdermal insulin delivery can overcome the issues related to needles and injections and bypasses the first-pass metabolism, and the large surface area of the skin provides a convenient method for insulin administration.^9,10 Recently, transdermal drug delivery systems (TDDS) have become the most extensively investigated noninvasive method of drug delivery through the skin, offering promising treatment options for a wide range of diseases, including skin cancer, cardiovascular diseases, diabetes, pain management, and hormonal therapies.^11,12 Transdermal insulin delivery is being investigated for managing diabetes, particularly for patients who require frequent insulin administration.⁹ However, insulin, a hydrophilic peptide hormone with a molecular weight of approximately 5800 Da, has a strong affinity for water, which restricts its capacity to diffuse through the lipid-rich stratum corneum (SC), the outer barrier of the skin.^9,10 The transdermal delivery of insulin requires the development of advanced strategies to efficiently transport insulin molecules across the skin, utilizing either physical or chemical methods.¹³ Physical approaches to deliver insulin include electrical facilitation techniques, such as iontophoresis, which uses electrical currents to drive insulin through the skin and enhance the permeability.^14,15 Additionally, mechanical force-triggered methods, such as the application of pressure or vibration, can assist in overcoming the skin barrier. Microneedles can also be used to create microchannels that enable insulin to bypass the outer skin barrier.¹⁶ In chemical approaches, permeation enhancers, such as ionic liquids (ILs), surfactants, fatty acids, esters, amino acids, and peptides, are crucial for facilitating transdermal insulin delivery by disrupting the lipid structure of the SC, thereby increasing its permeability.^17,18 ILs, which are molten organic salts that are liquid at ambient temperatures and are nonvolatile, nonflammable, and thermally stable, have gained attention in drug delivery systems because of their tunable synthesis from various cations and anions.¹⁹⁻²² ILs are extremely useful for transdermal drug delivery (TDD) because their distinctive physicochemical properties make ILs highly effective permeation enhancers.^20,23 For example, Mitragotri et al. have utilized a choline geranate (CAGE) IL to increase the transdermal permeability of insulin in vitro.²⁴

ILs have been incorporated into various delivery systems, including solid-in-oil nanodispersions,²⁵ microemulsions,^10,12 and ethosomes (ETs),²⁶ to enhance the stability and bioavailability of the encapsulated drugs. These systems also facilitate the solubilization and transport of hydrophilic and lipophilic compounds in the skin, promoting efficient TDD.^10,26,27 Microemulsions-mediated formulations have been reported by our group that improved the transdermal permeation of insulin in BALB/c diabetic mice.¹⁰ However, this complex ternary system has limitations, as it requires dissolving insulin through sonication and high temperatures in a choline propionate ([Chl][C3]) IL, which increases the risk of peptide degradation during formulation preparation.²⁸ Additionally, the quantity of insulin needed for this method is 5 times greater than that used in standard injections, making it costly and commercially challenging to implement.¹⁰ Recently, we have developed biodegradable lipid-based IL (LBIL)-mediated ETs, for the effective TDD of insulin, which significantly increased skin penetration compared with conventional 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)-mediated ETs.²⁶ The biodegradable LBILs have also been previously utilized in the TDD of various drugs, including cancer drugs,²⁹ antisense oligonucleotides,²⁵ and antigens,²⁹ demonstrating their versatility and effectiveness. In the present study, we developed an improved ET formulation incorporating a safe amount of LBIL, using a simplified preparation method designed to preserve insulin functionality. Additionally, a modified version of the ET formulation, known as trans-ethosome (TET), was developed, incorporating different edge activators, such as isopropyl myristate (IPM), Tween-80 (T80), and Span-20 (S20).³⁰ ET and TET preparation methods were simple, involving water dispersion in an ethanol solution of LBIL under stirring, without the need of any size reduction methods.³¹ Ethanol is known to enhance the quality of vesicles by allowing the solubilization of lipophilic compounds and by opening pores in the SC layer, thereby promoting the penetration of the loaded compounds.³² Edge activators were incorporated into TETs to further increase the permeability of the vesicles and make them more malleable.³³ For example, in vitro permeation of quercetin was found to be higher with TETs compared to ETs.³⁴ Numerous in vitro and in vivo studies have highlighted the transdermal effectiveness of ET and TET by evaluating their therapeutic potential and efficiency in delivering loaded compounds through the skin.^26,35,36

This study aimed to address the challenges of transdermal insulin delivery using ET and TET nanovesicles, both in vitro and in vivo, while ensuring scalability for potential commercial applications.¹⁰ The improved ET and new TET formulations showed enhanced in vivo performance using less insulin compared with previous studies. Both formulations were evaluated through various in vitro and in vivo experiments and demonstrated long-term stability, biocompatibility, and increased transdermal efficacy.

2. Materials and Methods

2.1. Materials

IPM, T80, S20, DMPC, and ethyl trifluoromethyl sulfonate were purchased from Tokyo Chemical Industry Co., Ltd., (Tokyo, Japan). 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT), human recombinant insulin (Lot. CAM195, potency ≥27.5 U/mg), and fluorescein isothiocyanate isomer I (FITC-I) were obtained from Wako Pure Chemical Industries Ltd. (Osaka, Japan). Ethanol (99.5%) was purchased from Kishida Chemical Co., Ltd. (Osaka, Japan). Linoleic acid was purchased from Sigma-Aldrich Chemical Co., (St. Louis, MO). PD-10 Sephadex G-25 M columns were purchased from Sankei Chemical Co., Ltd. (Kagoshima, Japan). Dulbecco’s phosphate-buffered saline (D-PBS) and trypsin/ethylenediaminetetraacetic acid (EDTA) (0.25% trypsin/1 mM ethylenediaminetetraacetic acid) were purchased from Nacalai Tesque (Kyoto, Japan). Isopropyl alcohol (IPA) and acetonitrile (ACN) were obtained from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan).

The human insulin enzyme-linked immunosorbent assay (ELISA) kit and three-dimensional reconstructed human epidermal model-24 (LabCyte EPI-24) were purchased from Japan Tissue Engineering Co., Ltd. (Aichi, Japan). Mouse skin sections (Hos: HR-1) were purchased from Hoshino Laboratory Animals (Ibaraki, Japan) and stored at −80 °C. Frozen skin samples, excised from a female Yucatan micropig (YMP), were purchased from Charles River Japan, Inc. (Tokyo, Japan) and stored in the frozen state at −80 °C. Five-week-old male BKS.Cg-m+/+ Lepr^db mice were purchased from Kyudo (Saga, Japan) and raised under standardized conditions. All animal experiments were approved by the Ethics Committee for Animal Experiments of Kyushu University (Approval No. A24-173-0) and carried out following the guidelines set by the Science Council of Japan.

All of the other chemicals and solvents were of analytical grade and used without any further purification.

2.2. Preparation of ET and TET Nanovesicles

ET vesicles were prepared using a DMPC-based LBIL consisting of the phospholipid derivative 1,2-dimyristoyl-sn-glycero-3-ethyl-phosphatidylcholine (EDMPC) as the cation and unsaturated linoleic acid (Lin) that contains two double bonds (C18:2) as the anion. The LBIL was synthesized following previously reported protocols.²⁶ ETs were also prepared following the previously reported “‘cold method”’ with minor modifications to the procedure (Scheme 1).²⁶ A stock solution was prepared by dissolving a specific amount of LBIL in ethanol. Edge activators, including IPM, T80, or S20, were added to the LBIL stock solution for TET preparation. For the ET and TET formulations, insulin was first dissolved in 0.01 M HCl (for every 20 mg of insulin, 1 mL of 0.01 M HCl was added), and then Milli-Q water was added to create a drug solution. The solution was neutralized with 0.01 M NaOH and mixed gently with the LBIL solution. Water was added dropwise with stirring at 750 rpm to create the formulations. The mixture was stirred for approximately 3 h with the speed gradually reduced from 750 to 250 rpm over this time. The formulations were then stirred at 4 °C for 2 h and subsequently stored at 4 °C in a refrigerator. The compositions of the ET and TET formulations are detailed in Tables 1 and S1.

Scheme 1. Schematic Representation of the Preparation of ET and TET Formulations.

Table 1. Composition of LBIL ET and TET1 Nanovesicles.

formulations	LBIL (mg/mL)	ethanol (μL/mL)	water (μL/mL)	IPM (μL/mL)	drug (mg/mL)
ET	6	350	650		1
TET1	6	350	630	20	1
ET (in vivo)	1.5	350	650		0.25
TET1 (in vivo)	1.5	350	645	5	0.25

2.3. Characterization of ET and TET1 Formulations

The particle size and ζ-potential (ZP) distribution of the formulations were determined using a Zetasizer Nano-ZSP (Malvern Panalytical, Malvern, U.K.) through dynamic light scattering (DLS) measurements. These measurements were used to calculate the average hydrodynamic radius (R_H) and ZP of the particles. Disposable cuvettes were positioned at an angle of 173°, and the temperature was maintained at 25 °C ± 0.1 °C during the measurements. Ten measurements were taken for each sample, and the mean values were calculated using Zetasizer 7.03 software, which provided Z-averaged values of the R_H and polydispersity index (PDI). Circular dichroism (CD) spectroscopy was employed to assess the stability of the insulin secondary structure in 35% ethanol and in the formulations. A J-1500 CD spectrometer (JASCO, Tokyo, Japan) was used, with the parameters shown in Table 2.

Table 2. Parameters for CD measurements^a.

cell	path length (mm)	wavelength range (nm)	integral count	temperature (°C)
quartz cell	0.1	190–250	5	25

^aThe obtained CD spectra were subjected to an adaptive smoothing method and then converted to the absolute value of CD.

2.4. Morphology Analysis of ET and TET1 Nanovesicles

The size and shape of ET and TET1 were observed using a JEOL JEM-2010 transmission electron microscope (TEM). Aliquots (5 μL) were placed on a carbon-film-coated copper grid and allowed to air-dry for 2 min, and excess material was removed with a filter paper. Staining was performed by applying 2.5 μL of a 2% uranyl acetate solution, followed by incubation for 5 min. Subsequently, the TEM grid was vacuum-dried in a desiccator and TEM images were captured at an accelerating voltage of 120 kV.

The formulations were also examined by using a confocal laser scanning microscope (CLSM). A 5 μL sample of each formulation was mounted between a glass slide and a cover glass for fixation. The observations were made using a Carl Zeiss LSM900 microscope, Oberkochen, Germany, at 63× magnification. LSM software from Carl Zeiss was used to process the images acquired by the microscope.

2.5. Stability Study of Insulin-Encapsulated Nanovesicles

The physical stability of the nanovesicles was determined at 4 °C over different time intervals through visual inspection, particle size distribution analysis, and ZP measurements. These stability assessments of drug-loaded nanovesicles were performed on day 0, after 1, 2, and 3 months.

2.6. In Vitro Skin Permeation Study and CLSM Observations

The skin penetration ability of the ET and TET1 formulations through YMP skin was investigated using Franz diffusion cells (FDC) following a previously described procedure.³⁷ First, the extra fat layer of the skin was removed by using a stainless steel scalpel. The skin was then cut into 1.75 cm × 1.75 cm pieces by using the same scalpel. The receiver phase and donor compartment were prepared as previously described,³⁷ given in detail in the Supporting Information. The FITC-Ins formulations (150 μL and 1 mg/mL) were applied onto the SC layer of the YMP skin within the donor compartment. The FDCs were then incubated at 32.5 °C for 24 h with continuous stirring at 500 rpm. After 24 h, the samples from the receiver phase were collected using needles to measure the content of FITC-Ins transdermally delivered via different formulations by using a multiple microplate reader at 485–535 nm. To quantify the delivery within the skin, the skin was washed with Milli-Q water and 20% ethanol and cut into pieces to extract FITC-Ins from the skin. The extraction was carried out in a mixture (v/v) of 50% PBS, 25% MeOH, and 25% ACN overnight with continuous shaking; then, the fluorescence intensity was measured.

To investigate the deepness of the skin penetration, clean YMP skin was placed in an embedding medium, optimal cutting temperature compound (Sakura Fine Tech, Tokyo, Japan), for frozen tissue specimens at −80 °C. Then, the solid immobilized skin tissue was sliced into 20 μm cross sections with the help of a cryostat microtome (Leica CM186OUV, Leica Biosystems, Wetzlar, Germany) and immediately transferred to a glass slide. Images of the skin sections were obtained by using a CLSM (LSM900, Carl Zeiss, Jena, Germany).

2.7. FTIR Analysis of SC Structural Modification Induced by ET and TET1 Formulations

The changes in the SC layer at a molecular level were observed by Fourier transform infrared (FTIR) spectroscopy after treating the skin with ET, TET1, and 35% ethanol. The experiment was performed following previously established protocols.¹² First, the fat layer of the YMP skin was removed with a scalpel. Then, the skin was segmented and kept in a heat block at 60 °C for 10 min to isolate the epidermis. The epidermal layer was then submerged in a solution containing 0.25% trypsin and 1 mM EDTA and incubated for 24 h at room temperature. Then, the SC layer was collected from the epidermis and washed multiple times with water. Subsequently, the SC layer was dried for 24 h at room temperature and segmented into pieces of approximately 1 cm² in size. The segmented pieces were treated with ET and TET1 formulations at 32.5 °C for 24 h. Then, the treated SC layers were washed with 20% EtOH and dried overnight. Finally, FTIR spectra of the treated SC sheets were acquired by using an Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) system (PerkinElmer, Waltham, MA) with a resolution of 1 cm^–1 for 16 scans. As a control, nontreated SC sheets were also prepared and the FTIR spectra obtained following a similar protocol.

2.8. In Vitro and In Vivo Skin Irritation Study

A 3D cultured human skin tissue model, LabCyte EPI-MODEL 24, was used to investigate the cytotoxicity and skin irritation of the ET and TET1 formulations. The procedure was similar to that of previous studies with slight modifications.^12,27 First, skin tissues were incubated for 24 h in 500 μL of assay medium in 24-well plates at 37 °C in a 5% CO₂ incubator. Then, 25 μL of the test samples were applied to the surface of each tissue and the tissues were incubated for 3 h. Then, the tissues were washed 5 times with phosphate-buffered saline (PBS). The tissues were subsequently transferred into new 24-well plates, with each well containing 500 μL of freshly prepared MTT medium at a concentration of 0.5 mg/mL. The tissues were then incubated for 3 h, after which each epidermal tissue fragment was transferred to microtubes containing 300 μL of IPA to extract the formazan. Absorbance was measured at 570 and 650 nm, and the number of viable cells was determined using the formula below. IPA was used as a negative control. The relative cell viability was calculated according to eq 1

1

The in vivo skin irritation caused by the ET and TET1 formulations was assessed from transepidermal water loss (TEWL). The unit of measurement was the weight of water per unit area (m²) per unit time (h), g/(m²·h). Before administration, diabetic BKS mice were anesthetized for 10 min to ensure sufficient sedation. The TEWL value of the dehaired mice was measured using a VAPOSCAN (Asch Co., Ltd.) before application of each formulation (n = 5). The TEWL value was measured 24 and 48 h after patch removal in the mice treated with the formulations. For comparison, the TEWL value was also measured at the same time points in the positive and negative control groups.

2.9. In Vivo Pharmacodynamic Response of Insulin-Loaded Formulations on BGLs and Evaluation of Pharmacokinetic Parameters in Diabetic Mice

BKS.Cg-m+/+ Lepr^db mice were received at 5 weeks of age and kept in an environment with sufficient water and food. After 5 weeks, the mice were dehaired with a trimmer and then treated with a depilatory cream 1 day before the experiment. Four groups of mice, with five mice in each group, were selected randomly. A comparison of the stabilizing effects on the BGL was conducted for the injection, no treatment, ET, and TET1 groups. For the injection group, an insulin/D-PBS solution was prepared by dissolving insulin in D-PBS (10 IU/kg). The backs of the mice were depilated the day before, and both BGLs and body weight were measured before the administration of the formulations. Mice with BGLs > 150 mg/dL were considered diabetic. The mice that had received insulin via injection were the positive control group and mice that were not treated with insulin were the negative control group. In the injection group, a dose of 10 IU/kg insulin/D-PBS solution was administered subcutaneously to the mice, while for the formulations, the dose was 30 IU/kg, 3 times the injection dose. The ET and TET1 formulations were applied to the dorsal area of the dehaired mice. The insulin solution was applied to the back of the mice and secured with a patch (2 × 2 cm²) and bandage. Measurements of BGLs were conducted at different time points from the time of administration (considered 0 h). During the monitoring period, the mice were supplied with a normal feed. The transdermal patches were removed after 24 h.

In vivo pharmacokinetic parameters were also determined in diabetic BKS Cg-m+/+ Lepr^db mice following a previously reported protocol.¹⁰ Around 100 μL of blood was collected at different time intervals in heparinized tubes. Plasma was separated by centrifuging the blood samples at 10,000g for 10 min at 4 °C. A human insulin ELISA kit was used to determine the insulin concentration in the plasma. This immunoassay represents a quantitative method with two insulin-specific monoclonal antibodies.

2.10. Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (Version 6.05). One-way ANOVA with Dunnett’s and Tukey’s multiple comparison test were used to determine the statistical significance of the data. A p-value < 0.05 was considered statistically significant. All values shown are the mean ± standard deviation (SD).

3. Results and Discussion

3.1. Selection and Composition of ET and TET Formulations

For effective transdermal insulin delivery, optimization of the formulations is crucial. In our recent study, an ET formulation prepared from a LBIL containing 35% ethanol was identified as an effective insulin carrier, demonstrating minimal cytotoxicity and good permeability.²⁶ In the present study, we incorporated different surfactants, IPM, T80, and S20, into the initial preparation step of the ET formulations (Scheme 1) and evaluated the particle size distribution, stability, and skin permeability of the resulting formulations. The preparation of ET and TET formulations is depicted in Scheme 1. Our goal was to identify the most effective TET formulation for use in an in vivo BGL-lowering study alongside the ET formulation.

It is well-known that during TDD, vesicles with a larger size (≥1000 nm) tend to remain on the SC.³⁸ In our recent study, we also observed that increased permeation of insulin could be achieved using vesicles with a reduced size distribution.²⁶ Therefore, the vesicle size is a critical factor in selecting carriers for TDD. On the basis of the particle size distribution and PDI values, three TET formulations, including IPM, Span-20, and a combination of IPM and Span-20 (highlighted in Table S1), were selected for further evaluation. Formulations with particle sizes close to or exceeding 1000 nm were not evaluated further (Table S1). ET and TET containing IPM (TET1) as an edge activator were selected for further characterization in the in vivo experiments because of their high stability and permeability (Figures S2 and S3). Both FITC-Ins and insulin were fully dissolved in all of the selected formulations.

DLS and TEM observations revealed that the particle size for insulin-containing ET was approximately 160 nm with a PDI value < 0.3, and for TET1, the particle size was approximately 300 nm with a PDI value < 0.15 (Figure 1A,B and Table 3). The ZP values were highly positive for both ET and TET1 formulations, which was indicative of vesicle stability and reflected the degree of repulsion between vesicles in the dispersions.³⁹ The positive ZP values for both formulations suggested a good vesicle stability. All of the DLS results are expressed as mean ± SD in Table 3.

Figure 1.

TEM images showing the physical appearance of ET (A) and TET1 (B) formulations. Scale bar = 50 nm. (C) CD spectra of insulin in ET and TET1 formulations compared with the insulin control solution in water.

Table 3. Size Distribution and ZP Values of ET and TET1 Formulations.

system	Z-avg (nm)	PDI	ZP
ET	162.1 ± 6.0	0.289 ± 0.061	56.4 ± 3.21
TET1	404.5 ± 9.8	0.128 ± 0.076	75.9 ± 1.94

Insulin has an α helix structure, and its interaction with receptors for bioactivity depends on the maintenance of the native structure.⁴⁰ The CD spectra of the insulin-loaded formulations were analyzed to assess the stability of the secondary structure of insulin in the formulations. Compared with the control insulin solution in PBS, the ET and TET1 formulations exhibited a slight increase in negative ellipticity from wavelengths of 230–210 nm (Figure 1), consistent with previous findings for insulin in ethanol.⁴¹ In previous studies, the native structure of insulin was shown to be maintained in solutions with a slight increase in negative ellipticity in ethanol concentrations up to 40%.^41,42 Therefore, it is likely that the structure of insulin was also preserved in the ET and TET1 formulations.

3.2. Morphology Analysis of the Formulations

The physical appearance of the insulin-loaded TET1 formulations was more turbid compared to the ET formulations (Figure 1A,B). The physical appearance of the TET formulations loaded with FITC-Ins was also more turbid compared to the ET formulations (Figure S1). TEM observations revealed the presence of spherical particles with sizes consistent with the DLS results (Figure 1). The FITC-ins-containing formulations were visualized by CLSM to observe the drug distribution within the formulation (Figure 2). The imaging revealed a uniform fluorescence distribution across both the ET and TET1 formulations, suggesting a uniform distribution of the drug (Figure 2). However, the vesicles in the TET1 formulation showed a tendency to be dispersed more extensively and to form larger aggregates than those in the ET formulation (Figure 2).⁴³ This increased dispersion was because of the inclusion of the oil IPM, which enhanced the vesicle flexibility, allowing nanovesicle aggregation.⁴⁴ This structural flexibility of the TET1 vesicles might be the reason for the larger particle size distribution compared with the ET nanovesicles (Table 3).⁴⁵

Figure 2.

Observation of the distribution of FITC-Ins encapsulated in ETs and TET1s using CLSM visualization through a 63× lens, scale bars = 10 μm.

3.3. Stability Study of ET and TET1 Nanovesicles

Good stability of drug-encapsulated formulations is crucial for pharmaceutical applications. The stability of insulin-loaded ET and TET1 formulations was examined by visual examination during storage for 3 months at 4 °C. Throughout this period, no aggregation, phase separation, color change, or drug precipitation was observed (Figure S3). The particle size distribution observed through DLS showed no significant changes in the size distribution of the ET formulation with only a minimal decrease in the size over time (Table 4).

Table 4. Size Distribution, PDI, and ZP Values of ET and TET1 Formulations over 3 Months at 4°C.

particle size (nm)a
	0_day	1_month	2_months	3_months
ET	162.1 ± 6.0	162.7 ± 3.8	162.9 ± 9.5	155.3 ± 4.0
TET1	404.5 ± 9.8	375.4 ± 11.07	374.8 ± 14.04	304 ± 11.87

PDIa
	0_day	1_month	2_months	3_months
ET	0.289 ± 0.061	0.171 ± 0.056	0.197 ± 0.071	0.169 ± 0.059
TET1	0.128 ± 0.076	0.215 ± 0.045	0.422 ± 0.109	0.242 ± 0.063

ZPa
	0_day	1_month	2_months	3_months
ET	56.4 ± 3.21	54.4 ± 0.96	52.3 ± 1.4	50.1 ± 1.15
TET1	75.9 ± 1.94	72 ± 1.23	72.2 ± 3.87	67.7 ± 2.3

^an = 10; mean ± SD.

A decreasing trend was observed in the size distribution of the TET1 formulation, where the average particle size was 404.5 ± 9.8 nm on the day of preparation, decreasing to 304 ± 11.87 nm after 3 months (Table 4). The PDI values were maintained under 0.5 for both ET and TET1 formulations (Table 4). There were no significant changes in the ZP values for the ET and TET1 formulations, although a slight decreasing trend was observed (Table 4). ZP values are an indicator of vesicle stability, as these values reflect the degree of repulsion between vesicles in a dispersion. According to previous studies, greater positive ZP values indicate enhanced stability of the vesicles.³⁹

3.4. In Vitro Skin Permeation Study and CLSM Observations in Pig Skin

Insulin, being hydrophilic, typically has limited skin permeation through the hydrophobic barrier of the SC, resulting in minimal absorption when insulin is delivered in water or PBS. However, when insulin is combined with an IL and ethanol in a nanovesicle system, insulin permeation might be increased considerably, as indicated by the FTIR results for the SC layer treated with ET and TET1 formulations. Additionally, the presence of one or more π-bonds in the IL structure increases the hydrophobicity of the IL, contributing to a more appreciable reduction of the barrier function.⁴⁶ The double bonds in unsaturated fatty acids enable the lipid structures to bend or kink, allowing for improved drug permeation through the skin.

To evaluate the penetration and permeation efficacy, the ET and TET1 formulations were applied on YMP skin, and cross-sectional samples were observed using CLSM to assess the depth of drug permeation. Compared with the control, which showed minimal fluorescence, the skin treated with ET and TET1 formulations exhibited a high level of fluorescence throughout the skin layers after 24 h (Figure 3A). The drug content was also quantified after 24 h of administration. Both the ET and TET1 formulations significantly (p < 0.0001) increased drug penetration across the skin (Figure 3B). Furthermore, the drug permeation was significantly higher for the ET and TET1 nanovesicles compared with that of the control (p < 0.0001 for ET and p < 0.005 for TET1), with the ET formulations showing a notable increase in drug permeation compared with the TET1 formulations (Figure 3C). This increased penetration and permeation may be attributed to the smaller particle size distribution of the ET nanovesicles, as smaller vesicles are generally known to improve TDD.³⁹ Additionally, the simpler structure of ETs, consisting only of phospholipids and ethanol, likely facilitates more efficient interactions with skin components.⁴⁷ In contrast, TET1s contain an edge activator, which might control sharp penetration by increasing the vesicle size or altering the vesicle properties.⁴³ Visually, the ET formulations appeared as less turbid solutions compared to the TET1 formulations (Figure 1). Overall, both the ET and TET1 formulations effectively facilitated the penetration of FITC-Ins macromolecules through the skin.

Figure 3.

Skin permeability of FITC-Ins using ET and TET1. (A) CLSM images of skin cross sections of YMP skin treated with 1.0 mg/mL FITC-Ins in ET and TET1 formulations along with the control FITC-Ins solution for 24 h with a 20× lens. Scale bars: 200 μm. The total amount of FITC-Ins penetrated (B) and permeated (C) after 24 h; ns = nonsignificant; n = 3, mean ± SD; **p < 0.005, and ****p < 0.0001.

3.5. Effect of ET and TET1 on Skin Permeability

TDD is often hindered by the hydrophobic SC layer, a lipid matrix that forms a barrier to skin permeation.⁴⁸ The SC layer must be disrupted by altering the molecular structure to allow drugs to penetrate.¹³ These changes can be analyzed by measuring the vibrational shifts in cellular lipids using FTIR spectroscopy.^49,50 FTIR spectroscopy was performed on the SC layer after treatment with the drug-loaded ET and TET1 formulations (Figure 4). The SC layer was also treated with PBS and 35% ethanol as controls for comparison. The shifts in the spectral peaks were then compared with those of an untreated SC layer. The characteristic absorption peaks of lipids were observed in the untreated SC layer at 2920 cm^–1 (CH₂ asymmetric stretching) and 2850 cm^–1 (CH₂ symmetric stretching), as shown in Figure 4A,C. The ratio of the intensities of these peaks reflects the lateral interactions between acyl chains within the SC and indicates lipid fluidization.⁵¹ Keratin peaks were observed at 1645.18 and 1535.52 cm^–1, corresponding to the NH–C=O vibrations (Figure 4B,D). These peaks are indicative of an organized α-helical structure, suggesting that the keratin in the SC retained a stable and well-ordered α-helical configuration. All of the treated skin samples exhibited shifts in the peaks corresponding to both the lipid and keratin components of the SC (Figure 4).

Figure 4.

(A, B) FTIR spectra of the SC of YPS treated with different test samples. Peaks shifted in the cellular lipid (C) and keratin layer (D); n = 3, mean ± SD; ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, and ns, nonsignificant (Tukey’s posthoc test).

Both the ET and TET1 formulations produced significant peak shifts compared to PBS and 35% ethanol (Figure 4). Notable shifts observed with 35% ethanol suggested a synergistic enhancement effect from the combined components within the nanovesicle system.⁵² For the ET formulations, the CH₂ vibration peak shifted to 2923.72 cm^–1 (asymmetric) and 2953.88 cm^–1 (symmetric) and the NH–C=O peaks shifted to 1646.36 and 1546.94 cm^–1 (Figure 4). For TET1, the CH₂ vibration peak shifted to 2923.83 cm^–1 (asymmetric) and 2953.83 cm^–1 (symmetric), and the NH–C=O peaks shifted to 1647.39 and 1545.85 cm^–1 (Figure 4). These changes indicated that the nanovesicles altered the molecular organization of the SC layer.⁵³ The shift of the CH₂ symmetric stretching vibration peak to a higher wavenumber reflected the transformation of the lipid from an orthorhombic phase to a liquid-crystalline structure, which indicated an increased fluidity in the lipid matrix. Additionally, the shift of the NH–C=O vibration peaks to higher wavenumbers indicated that the keratin structure transitioned to a random coil configuration following treatment with the test samples (Figure 4B,D). The ionic nature of the IL likely contributed considerably to these effects on the SC.⁵⁴ Furthermore, the influence of ethanol can be attributed to an optimal balance of push–pull and blending effects on the skin. This balance reduced the skin’s barrier function, resulting in an increased permeability coefficient and improved TDD.⁵⁵

3.6. In Vitro Skin Irritation Study

The biocompatibility of the ET and TET1 formulations containing LBIL was evaluated by using artificial human epidermal tissue. PBS was used as a negative control, while 5% sodium dodecyl sulfate (SDS) was used as a positive control in the skin irritation study. A nontreated sample was also included as a negative control. The relative cell viabilities of the nontreated control and PBS were 99.95% ± 0.05% and 97.21% ± 0.05%, respectively, indicating that PBS had no negative impact on the cell viability (Figure 5A). The cell viability for the ET formulation was >90% and for the TET1 formulation, the viability was ∼90% (Figure 5A). In contrast, the cell viability was significantly decreased (<20%) with 5% SDS compared with the ET and TET1 formulations and the other control samples (Figure 5A).

Figure 5.

(A) Biocompatibility of ET and TET1 formulations compared with positive and negative controls. n = 3; mean ± SD; ns, not significant; ****p < 0.0001 and **p < 0.005, significant values were determined using Tukey’s posthoc test. (B) Mouse skin condition before and immediately after patch removal. (C) TEWL values for the mouse skin before application, after patch removal, and at 48 h after removal; n = 5, mean ± SD; ns, nonsignificant (compared with the TEWL value before application of formulation) using Dunnett’s multiple comparison test.

The high cell viability for both the ET and TET1 formulations indicated the safety of these formulations toward human epidermal tissues. Similarly, the cell viability of IPM, used as an edge activator in the TET1 formulation, was also ∼90% (Figure 5A). It has also been previously reported that LBIL, along with LBIL formulations containing oligonucleotides²⁵ and hormones⁵⁶ were nonirritating and biocompatible, which is consistent with our findings.

The TEWL value was measured to evaluate the potential for the ET and TET1 nanovesicles to cause skin irritation, as the TEWL value is as an indicator of the skin barrier function and can be used to detect any dysfunction.⁴⁸ Initial TEWL measurements were taken prior to the application of formulations in all of the mouse groups. After 24 h, the patches were removed, and the TEWL values were measured again to evaluate the water loss after application of the formulations. Although a slight increase in the TEWL value was noted, the changes were not statistically significant (Figure 5C). A follow-up measurement at 48 h indicated that the TEWL values had returned to the initial values, suggesting full restoration of the skin’s barrier function post application of the formulations (Figure 5C). TEWL measurements were also performed for both the positive (injection) and negative (control) groups, and no significant changes were observed in either group (Figure 5C).

3.7. In Vivo Response of Insulin-Loaded ET and TET1 Formulations toward Lowering BGLs in Diabetic Mice

The efficacy of the ET and TET1 formulations in lowering BGL was evaluated using diabetic BKS mice, with four groups of five mice each: (1) positive control group, which received subcutaneous insulin injections (10 IU/kg); (2) negative control group, which received no treatment; (3) ET formulation group, treated with insulin-loaded ET formulation (30 IU/kg); and (4) TET1 formulation group, treated with insulin-loaded TET1 formulation (30 IU/kg). BGLs were monitored regularly over 24 h. In the negative control group, no significant BGL-lowering effect was observed at any time, and the BGL levels remained steady throughout the observation period with no notable weight changes (Figure S5). In contrast, a rapid hypoglycemic effect was observed in the positive control group within 30 min of insulin injection, with a significant decrease in BGLs of 49.5% (Figure 6C). This reduction lasted approximately 2 h, followed by a sharp increase in BGLs after 3 h, with no further reduction observed (Figure 6B,C). No significant change in the body weight of the mice was noted in this group, although there was a slight change 24 h after injection, which was recovered within 1 day (Figure S5).

Figure 6.

(A) In vivo experimental design and the physical appearances of the mice after being dehaired, with the applied patch, and after patch removal. (B) Changes in diabetic mice BGLs over time after transdermal insulin delivery with ET and TET1 nanovesicles. (C) BGL-lowering potency (%) of different formulations compared with baseline levels. Mice with BGLs > 150 mg/dL were considered diabetic. Untreated mice were used as the negative control, while mice that were given a subcutaneous insulin injection (10 IU/kg) were used as positive controls. ET and TET1 nanovesicle formulations were applied at 30 IU/kg (B); n = 5, mean ± SD; ns, nonsignificant; ***p < 0.001 and ****p < 0.0001 (compared with 0 h BGL value using Dunnett’s multiple comparison test).

In the ET formulation group, the BGLs were decreased by 62% within 1 h (Figure 6C). This effect was maintained for up to 15 h before increasing to >150 mg/dL at 24 h (Figure 6B). Mice in this group showed a more pronounced and sustained hypoglycemic effect compared with those of the injection and TET1 groups, with only a slight temporary body weight loss observed (Figure S5). In the TET1 formulation group, the BGLs steadily declined by approximately 34%, reaching stable levels of 100–150 mg/dL for up to 15 h (Figure 6B,C). The TET1-treated mice did not experience hypoglycemia but exhibited a more gradual and sustained reduction in BGLs, with a slight and temporary decrease in body weight (Figures 6B and S5).

Both formulations effectively reduced BGLs, with ET showing a sharper hypoglycemic effect, while TET1 provided a steadier, more gradual reduction, consistent with the in vitro penetration study (Figures 3 and S4). At the end of 24 h, the patches were removed, revealing no visible skin damage. Although transdermal delivery using these formulations required 3 times the insulin dose compared with delivery via injection, stable BGLs were maintained for nearly 15 h, significantly longer (7 times) than delivery by injection (Figure 6).

In vivo pharmacokinetic parameters were also assessed in diabetic mice to evaluate the TID efficacy of the ET and TET1 nanovesicles. Insulin concentrations were measured at multiple time points (0, 1, 5, 10, 15, and 24 h), where 0 h represents the plasma insulin concentration before treatment (Figure 7C). A significant increase in insulin concentration was observed within 1 h of transdermal delivery for both formulations (Figure 7C). It has been gradually reduced over time and reached the previous values within 24 h (Figure 7C). The concentration gradually decreased over time, returning to baseline levels within 24 h (Figure 7C). The elevated insulin levels were sustained for 15 h, consistent with the pharmacodynamic response of the formulations (Figure 6).

Figure 7.

In vivo efficacy of insulin-loaded ET and TET1 nanovesicles for increasing the bioavailability of insulin in diabetic mice before and after transdermal delivery. Nontreated mice were considered as a negative control, and mice treated by subcutaneous injection were considered as a positive control. (A) Insulin concentration before and after insulin delivery over 24 h for injection, nontreatment, ET, and TET1 nanovesicles-treated mice. (B) Insulin concentration before and after subcutaneous insulin delivery over 3 h. (C) Insulin concentration before and after insulin delivery over 24 h for nontreatment, ET and TET1 nanovesicles-treated mice; n = 5, mean ± SD; ns, nonsignificant; **p < 0.005 and ****p < 0.0001 (compared with insulin concentration before treatment using Dunnett’s multiple comparison test).

In comparison, for mice treated with subcutaneous injection, insulin concentration increased significantly within 30 min, reaching its peak (over 400 μIU/mL) within 1 h (Figure 7A,B). In contrast, insulin concentrations remained relatively unchanged over the 24 h period in nontreated mice. These findings highlight the efficacy of the ET and TET1 formulations in providing controlled and sustained TID, avoiding the sharp peaks associated with subcutaneous injections.

4. Conclusions

In conclusion, we have successfully prepared insulin-loaded ET and TET1 formulations, which contained lower insulin dosages compared with those of previously reported formulations for transdermal insulin delivery. The ET and TET1 formulations showed long-term stability, increased skin permeability, biocompatibility, and a significant BGL-lowering capacity for a longer time compared to injection. DLS and TEM observations indicated that the formulations contained spherical particles with sizes of approximately 160 and 400 nm for the ET and TET1 nanovesicles, respectively. Compared with the control FITC-Ins solution, the ET and TET1 formulations increased the drug penetration and permeation significantly. These formulations facilitated insulin permeation by altering the molecular structure through the activation of a fluidizing effect on the lipid matrix and the transition of the keratin structure in the SC layer of the skin. In vitro and in vivo skin irritation studies indicated the safety of the formulation applied on artificial human tissue and mouse skin, respectively. Finally, an in vivo study using diabetic mice showed that the formulations had a significant BGL-lowering effect compared with insulin delivery via subcutaneous injection. Lower BGLs were also maintained for a significantly longer time of 15 h, which was 7 times longer than the effect of injection (2 h). The plasma insulin concentrations measured at different time points further validated these findings. The ET and TET1 formulations have strong potential as effective transdermal carriers for insulin and other peptide-based therapeutics. By providing an alternative to traditional injection methods, while maintaining both safety and efficacy, these formulations could revolutionize transdermal insulin delivery and also enable the transdermal administration of other peptide drugs usually administered by injection.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsbiomaterials.4c02000.

• Additional experimental details with results and discussion including composition, size distribution, and PDI values for TET formulations (Table S1); physical appearances of FITC-Ins containing formulations (Figure S1); physical appearances and particle size distribution of ET and TET1 formulations at day 0 and after 3 months (Figure S2); drug distribution of TET formulations with S20 and with both S20 and IPM (Figure S3); and weight variation in all of the mice during the experiment (Figure S4) (PDF)

Author Contributions

F.H.N. contributed to conceptualization, methodology, investigation, visualization, data analysis, writing, and editing. R.I. and L.Y. contributed to investigation and editing; Y.K., R.W., and N.K. contributed to resources, review, and editing; M.M. contributed to supervision, review, and editing; M.G. contributed to supervision, fund acquisition, conceptualization, review, and editing.

The authors declare no competing financial interest.