Development of film-forming gel containing nanoparticles for transdermal drug delivery

HPMC 4000 was dispersed slowly in pre-heated water (60 °C) under continuous stirring to form a cloudy swelling solution. Then, the HPMC solution was cooled at −4 °C for 10 min. Simultaneously, zein, as a film-forming polymer, was dissolved in ethanol solution (90% v/v) at ambient temperature until a transparent solution appeared. The zein/HPMC FFH was obtained by blending zein and HPMC as film-forming and hydrogel agent, respectively, followed by gentle stirring for 2 h. After 2 h, a plasticizer or a mixture of plasticizers was added to the zein/HPMC blend and stirred thoroughly at room temperature. The final samples were stored in tubes and sealed tightly until further use. The composition of FFHs with varying concentrations of polymers and plasticizers is illustrated in table 1.

2.2.2.1. In vitro testing

2.2.2.2. Ex vivo testing

CUR-GNPs were prepared by sonoprecipitation (table 2). First, 200 mg of gelatine was dissolved in 20 ml of water at 40 °C. Meanwhile, CUR and POX 407 were dissolved in 5 ml of methanol at room temperature until a transparent solution was obtained (4 mg ml⁻¹ of CUR in the solvent phase). Then, the CUR/POX 407 mixture was dropped quickly into the dispersion phase under moderate magnetic stirring at 750 rpm. Then, the samples were immediately emulsified using an ultrasonic liquid processor (Qsonica, USA) with different sonication times and amplitudes. All samples were maintained in a water bath at 20 °C ± 1 °C during the sonication process to circumvent thermal variations. After ultrasonication, the samples were stirred at 30 °C during the dropwise addition of GTA 5% (v/v) solution to crosslink the particles. The methanol was completely evaporated after 2 h. The mixture was then centrifuged at 1000 rpm for 30 min to gradually remove small quantities of aggregated particles at the bottom of the tube, and the supernatant was lyophilized at −50 °C for 24 h to collect CUR-GNP powder.

A pre-determined quantity of lyophilized CUR-GNPs (equivalent to 1 mg of CUR) was dispersed slowly in 1 g of FFH under magnetic stirring at 250 rpm for 24 h at ambient temperature to ensure that all of the nanoparticles were dispersed in the gel, forming a homogeneous mixture.

The particle size, polydispersity index (PDI), and zeta potential of the CUR-GNPs were analysed using a Zetasizer Nano Series system (Malvern Instrument Limited, UK). First, 20 μl of the CUR-GNP solution was diluted with 20 ml of methanol followed by vortexing prior to analysis. To determine the particle size of the CUR-GNPs in the FFH, 5 mg of FFN was dispersed in 10 ml of a complex solvent containing ethanol and distilled water corresponding to the ratio of the FFH preparation under stirring for 1 h. Measurements were performed in triplicate at 25 °C.

2.2.5.1. Entrapment efficiency (EE) and drug loading (DL)

CUR-GNPs (equivalent to 1 mg of CUR) were dispersed in 25 ml of distilled water followed by sonication for 30 min at 50 °C to extract CUR completely. Then, absolute ethanol was added to the mixture under shaking and cooled to room temperature. The final volume was adjusted accurately to obtain 50 ml. Then, 100 μl of the solution was diluted with 900 μl of methanol for HPLC analysis (as described in section 2.2.9). Measurements were performed in triplicate for each sample, and the quantity of CUR entrapped in the form of CUR-GNPs was calculated using the following equations:

$$\begin{eqnarray*}&&\begin{array}{l}{\rm{EE}}=\displaystyle \frac{{\rm{The}}\,{\rm{mass}}\,{\rm{of}}\,{\rm{drug}}\,{\rm{in}}\,{\rm{nanoparticles}}}{{\rm{Total}}\,{\rm{mass}}\,{\rm{of}}\,{\rm{drug}}\,{\rm{used}}\,{\rm{in}}\,{\rm{preparation}}\,{\rm{of}}\,{\rm{nanoparticles}}}\\ \,\,\times \,100 \% \end{array}\end{eqnarray*}$$

$$\begin{eqnarray*}&&{\rm{DL}}=\displaystyle \frac{{\rm{The}}\,{\rm{mass}}\,{\rm{of}}\,{\rm{drug}}\,{\rm{in}}\,{\rm{nanoparticles}}}{{\rm{The}}\,{\rm{mass}}\,{\rm{of}}\,{\rm{nanoparticles}}}\times 100 \% .\end{eqnarray*}$$

The in vitro drug release of CUR-GNPs from the FFH was determined using a Franz diffusion cell (area = 3.8 cm², receptor chamber volume = 20 ml) with a cellulose diffusion membrane (Ø = 25 mm, 0.45 μm, Sartorius, Germany). The membrane was activated by drenching in isopropyl myristate (IPM) for 24 h before testing. Three Franz diffusion cells were equilibrated for 15 min in a water bath at 37 °C ± 0.5 °C before the sample was applied. The receptor medium, consisting of 20 ml of ethanol 20% (v/v), was continuously stirred at 500 rpm and 37 °C [13, 14]. The FFN (10 mg cm⁻²) was evenly spread on the surface of the membrane. Then, 500 μl of the receptor medium was collected after 1, 2, 4, 6, 8, 12, 18, and 24 h and replenished by an equivalent amount of fresh medium. For HPLC analysis, 100 μl of the solution was diluted with 900 μl of methanol.

The quantity of CUR was determined using an Ultimate 3000 HPLC system (Thermo Fisher Scientific, Inc., USA). HPLC analysis was performed using a reverse-phase column (150 × 4.6 mm, C18). The mobile phase consisted of methanol and water at a ratio of 80:20 (v/v), with a flow rate of 1.2 ml min⁻¹. The run time and UV/Vis detector were set to 5 min and a wavelength of 425 nm.

The crystallinity of the dry films, pure CUR and carriers was determined using a Powder x-ray diffractometer (D2 Phaser, Bruker, Germany) with Cu radiation at 30 kV and 100 mA. All of the samples were scanned across diffraction angles (2θ) ranging from 10° to 50° with a step size of 0.02° and a rate of 2θ/s.

An FTIR spectrophotometer (Vertex 70, Bruker, USA) was employed to obtain the spectra of dry films, pure CUR and carriers. One milligram of the sample was dispersed in 200 mg of dry potassium bromide (KBr). The compressed mixture of 1 mg of sample and 200 mg of KBr was scanned at wavelengths from 500 to 4000 cm⁻¹ with a resolution of 4 cm⁻¹.

All experiments were conducted three times. The data are presented as the mean ± SD and were analysed by ANOVA (SigmaPlot version 11). Statistically, P < 0.05 was considered to indicate a significant difference.

Ultrasonication was employed to reduce the particle size by varying the intensity of the sonic waves to create mechanical stress in the samples, resulting in a uniform and stable dispersion. The profile in figure 2(A) shows that the particle size was less than 300 nm and that the particles had a moderate zeta potential of 30–35 mV. Increasing the amplitude from 20 to 30 resulted in the delivery of more energy to break down the samples; hence, the mean particle size also decreased from 279.9 nm to 242.1 nm. However, the application of the high amplitude of 40 resulted in a slightly increased particle size. Moreover, there was no significant difference in the average particle size between N2 and N3 (P > 0.05). This result could be explained by the high intensity of the sonic waves corresponding to the high amplitude, leading to the destruction of the nanoparticles; thus, the drug may have dissociated from the nanoparticles and accumulated in the anti-solvent phase to form larger particles. Therefore, an amplitude of 30 was chosen as a suitable ultrasonic condition for particle size reduction.

The results shown in figure 2(B) indicate that an increase or decrease in the concentration of the POX 407 emulsifier led to an increase in the mean particle size. Decreasing the emulsifier concentration to 8 mg ml⁻¹ resulted in a less stable system and the formation of aggregates. Meanwhile, a high concentration of POX 407 resulted in the solubilization of CUR or gelatine in the precipitation process, which further reduced the stability of the system. Thus, an emulsifier concentration of 10 mg ml⁻¹ was applied in further experiments.

Zoom In Zoom Out Reset image size

On the other hand, the gelatine concentration could only be increased to a limited extent due to the high viscosity of the anti-solvent phase, which further reduced the uniformity of the dispersion of the solvent phase. Moreover, increasing the gelatine concentration could lead to high concentration of free gelatine in the solution after nanoprecipitation. Consequently, the aggregation of this free gelatine resulted in the formation of larger particles through crosslinking, and it became more difficult to obtain a uniform dispersion of CUR-GNPs incorporated in the FFH, as gelatine is practically insoluble in most organic solvents and the FFH system contains ethanol. Hence, the concentration of gelatine in N4 should be reduced. However, decreasing the gelatine concentration of 7.5 mg ml⁻¹ in N4 (figure 2(C)) resulted in a significantly increased particle size and decreased zeta potential compared to those of N2. It could be reasoned that a low gelatine concentration is not sufficient to encapsulate CUR/POX 407 inside nanoparticles.

Figure 2(D) depicts the particle size of N2, N7 and N8 resulting from ultrasonication at an amplitude of 30 for 10, 20, and 30 min, respectively. Compared with N2, the increased sonication time of N7 resulted in a significant increase in the zeta potential (P < 0.05). Moreover, there was no significant difference in particle size between N2 and N7 (P > 0.05); however, the encapsulation efficiency and drug loading of N2 were lower than those of N7 (details provided in section 3.2.2). In contrast to N7, when the sonication time was prolonged to 30 min for N8, the particle size was significantly increased to greater than 650 nm, and the zeta potential was reduced due to thermal variation resulting from the long period of sonication.

GTA 5% (v/v) solution was employed as a crosslinker to prevent the aggregation of and stabilize the CUR-GNPs. Different volumes of GTA were added dropwise. As shown in figure 2(E), the increased volume of GTA in N9 drastically increased the average particle size by approximately twofold compared to that of N7. Reasonably, greater concentrations of GTA resulted in the crosslinking of multiple groups and an increased degree of reticulation, thereby leading to an increased particle size. The use of less than 30 μl of GTA resulted in a large extent of aggregation because the quantity of crosslinker was insufficient to harden the nanoparticles (data not shown).

Based on the characterization of the CUR-GNPs, N1, N2 and N7 were selected for further investigation of the EE and DL (table 3) due to the smaller average particle size and the moderate zeta potential. The results show no significant difference in entrapment efficiency or drug loading between N1 and N2 (P > 0.05). Meanwhile, although there was no significant difference in particle size between N2 and N7, N7 performed significantly better, with the highest EE and DL of all other formulations (P < 0.05). Specifically, approximately 50% of the CUR was encapsulated in the nanoparticles, which were the smallest (233.4 nm) and showed good stability with the highest zeta potential (35.9 mV). Therefore, N7 was chosen as an optimal CUR-GNP model for dispersion in the FFH.