Materials

Imiquimod (IMQ, ≥ 95%) was purchased from Cayman Chemical (Michigan, USA), Tween 80, vinylpyrrolidone (VP, ≥ 98%), polyethylene glycol diacrylate Mw = 700 g/mol (PEGDA), lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP, ≥ 95%), methanol (≥ 99.9%), acetonitrile (≥ 99.9%), gentamicin sulphate (≥ 95%), Nile Red (≥ 97%), isopropyl alcohol (≥ 98%), ammonium acetate (≥ 98%) and phosphate-buffered saline (PBS; 0.01 M phosphate buffer, 0.0027 M potassium chloride and 0.137 M sodium chloride, pH 7.4) tablets were obtained from Merck KGaA (Darmstadt, Germany). Water was deionized, distilled, and filtered through a Millipore Q purification system.

Preparation of nanocrystals

NCs (5 wt% in final formulation) were prepared according to Petrová etl al [16]. via small-scale wet stirred medium milling. In short, 30 mg IMQ with 7.5 mg Tween®80 and 1 ml deionized water were placed in a 25 ml amber glass vial containing the milling medium (5 g of 0.4–0.5 mm diameter zirconium oxide milling beads and a 8 × 20 mm PTFE cross stirrer). The milling was performed at room temperature at constant 700 rpm, for 24 h. After the milling period, the suspension was diluted by 1 ml of deionized water and rehomogenized at 700 rpm for 1 min in order to achieve good homogeneity of the suspension. After brief homogenization, the vial was tilted and the suspension was carefully drained using a PTFE pipette so as not to drain milling beads along with the sample. The suspension was immediately analysed via dynamic light scattering using the NANO-Flex fiber-optic instrument (Microtrac Retsch GmbH, Haan, Germany). No filtering steps were included, since satisfactory particle size distribution was obtained. The nanosuspension was subsequently freeze dried (1 h − 40 °C → 3 h − 35 °C → 5 h 35 °C, 26 Pa) using AdVantage 2.0 benchtop freeze dryer (SP Scientific, Warminster, USA). This procedure provided the freeze dried powder with residual moisture ≤ 2% and was based on the standard procedure for previously used nanoformulations [17]. The intactness of freeze-dried NC was checked by transmission electron microscopy (TEM, Jeol JEM-1010, Jeol Ltd., Tokyo, Japan).

DLP printing and characterisation of microneedles

MNs were designed by TinkerCAD (Autodesk, San Francisco, USA), an online computer-aided design software. The patch with MNs was designed as a square of 15 × 15 × 2 mm (Fig. 1). Each patch contained 36 conical MNs which were of 1 mm in diameter and 1.5 mm in height after printing (the dimensions of final patch are described in Chap. 3.1). The design was converted to.stl format and uploaded to LumenX (Cellink, Göteborg, Sweden) 3D printer. The resin used for the printing consisted of VP, PEGDA, photoiniciator– LAP (1 wt%), and distilled water as the solvent. The printing resolution was 50 μm and intensity of UV light was 22.5 mW/cm2 per layer.

Fig. 1

The patch CAD design (obtained by TinkerCAD software) used for the DLP printing of IMQ-loaded microneedle patches

After the printing process, the patch was gently removed, washed in 70% ethanol, and put in the UV-chamber (Asiga, Alexandria, Australia) for 1 min. Finally, the patch was dried at room temperature for 2 days, when the moisture in the patch stabilized. In the case of IMQ loaded patches, the freeze-dried IMQ NCs were added to resin and stirred for 15 min (500 rpm). The patches were visualized by optical microscopy Leica EZ4D (Leica Microsystems, Milton Keynes, UK) at room temperature.

Fourier-transform infrared spectroscopy (FTIR)

FTIR was used to determine the monomer conversion in MNs patches and to confirm the IMQ incorporation in the matrix and its distribution though matrix. All results were reported as average from 3 independent measurements. Spectra were recorded using a Nicolet iZ10 (Thermo Scientific, Watham, MA, USA) equipped with a single reflective Miracle ZnSe crystal (PIKE technologies, Madison, WI, USA) at constant clamping force. The spectra were collected at a resolution of 2 cm− 1 by co-addition of 64 scans and evaluated using the OMNIC™ software. Spectra were normalized in Origin PRO (OriginLab, Wellesley Hills, USA) to minimize the effect of variations.

Differential scanning calorimetry (DSC)

DSC was used to determine the thermotropic behaviour of MN patches and confirm the IMQ incorporation. All results were reported as average from 3 independent measurements.

The thermograms were measured using the Netzsch DSC 214 Polyma with autosampler (NETZSCH GmbH & Co.Holding KG, Selb, Germany). Samples (5–7 mg) were accurately weighed into aluminium pans and crimped with a lid. The measurements were performed under a heating range of 20 °C/min and temperature range of − 10– 400 °C. Nitrogen was used as a purge gas at flow rate 40 ml/min. As the standard, the empty aluminium pan was used. The melting points were determined as the position of the main phase transition peak in the thermogram.

Insertion studies in porcine skin

Fresh neonatal pig skin was kindly donated from a local slaughterhouse following all local regulations. The skin was cleaned in distilled water and stored at − 20 °C. The skin was defrosted at room temperature and placed on the microscope glass and fixed with tape on the sides. The patch was pressed in the skin with thumb pressure (~ 20 N) for 1 min. Subsequently, the patch was removed from the skin and a solution of Nile Red in isopropyl alcohol (1 wt%) was applied on the skin to visualize the holes created by MNs. The skin surface was observed by optical microscopy, with images taken before and after MNs application. The image analysis was applied for the determination of holes sized created by MNs. The experiments were performed in triplicates.

Insertion studies in Parafilm® layers

Insertion studies in Parafilm® layers were performed according to published methods [3, 18]. 10 layers of Parafilm® were cut in the 5 × 5 cm squares and stacked on the top of each other to create a ~ 1 mm high layer as a model of the skin. A TA.XTPlus Texture Analyser (Stable Micro Systems, Surrey, UK) was used to insert the MN in the membrane by a cylindrical probe. The probe was approaching the model with the patch attached (MN pointing into the skin model) at a speed of 1.19 mm/s. Subsequently, the 32 N force was applied for 30 s. Finally, the probe automatically picked up at the same speed. Skin model with inserted MN patch was removed and the created holes were observed under optical microscope. Based on the number of created holes, the effectivity of insertion was evaluated. The study was performed in triplicates.

Swelling study

Swelling test was conducted using the part of the patch with 6 MNs. PEGDA and VP polymerise into hydrophilic polymers. The rate of swelling is important for the IMQ release. Each patch was accurately weighted and placed in the 10 ml of buffers (PBS buffer, pH = 7.5; acetate buffer, pH = 5). At different time points (e.g., 0.5, 1, 2, 4, 6, 24 and 48 h), the patch was carefully taken off, dried with paper tissue, weighted, and placed back into the vial with buffer. The temperature of experiment was set to 32 °C. The study was performed at least in triplicates for each buffer.

In vitro release study

For the in vitro study, only part with 6 MNs were cut out and used. The patches were placed in 15 ml vial in 15 ml of PBS buffer of pH 7.4 or in acetate buffer of pH = 5.5 (20 mM), which is the pH of the skin surface [19, 20]. The temperature was set to 32 °C (skin temperature). Samples (0.5 ml) of the dissolution media were taken at 0.5, 1, 2, 4, 6, 24 and 48 h. Then, the patch was extracted in 3 ml of methanol/acetate buffer (20 mM, pH = 4) in ratio 7:3 to determine the total residual quantity of IMQ in patch. All samples were immediately analysed by high-performance liquid chromatography (HPLC, see Chap. 2.6) and the curve of cumulative release was determined. The study was performed at least in triplicates for each buffer.

For better understanding the drug release kinetic, several mathematical models were applied [21, 22]. The coefficients were determined by Solver in Microsoft Excel.

Ex vivo permeation studies

The study was performed in Franz diffusion cells and porcine skin was used as the tested barrier. The acceptor part was filled with PBS buffer with gentamicin (50 mg/ml, microbial protection). The skin (ca. 1 cm2) was placed between the donor and acceptor part. For equilibration, the cells were tempered at 32 ± 0.5 °C in a water bath for 1 h. Then, patches were pressed in the skin on the donor side and the MN patch was covered with Parafilm®. After 48 h, 300 µl of samples were taken from each cell and analysed immediately by HPLC. Then, the patch and skin were gently removed. The skin was extracted in 3 ml of the extraction medium (methanol/acetate buffer (20 mM, pH = 4) in ratio 7:3). Finally, the extracts were filtered through 0.22 μm filters and analysed by HPLC (see 2.7). Each formulation was tested in 8 skin samples.

High-performance liquid chromatography (HPLC)

IMQ concentration was analysed by UV-HPLC Agilent Infinity 1220 LC system (Agilent Technologies Inc., Santa Clara, USA) using a flow rate of 1 ml/min, sample injection volume of 20 µl and detection wavelength of 242 nm. The mobile phase was acetonitrile/acetate buffer (pH = 4, 20 mM) in a 3:7 ratio (v/v). A Kinetex® column 150 × 4.6 mm, 5 μm, RP C18, 100 Å (Phenomenex, Torrance, USA) equipped with a guard column was used as the stationary phase. The retention time of IMQ was 3.0 min. Data were evaluated in Chemstation software 02.09 (18). A calibration curve was created from standards and the exact IMQ concentration was calculated. The method was previously validated in house for linearity, accuracy and precision.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software (GraphPad Software, Boston, USA). The Grubbs’ test was used to identify the outlier values. Unpaired t-test was used for determination of p value with confidence level 95%. The data are presented as the mean values ± SEM of minimal three independent measurements.