2.4.1. Formulation Optimisation by 26−<sup>2</sup> Fractional Factorial Design

Based on the preliminary experiments and literature study, six independent variables (factors), namely surfactant type (X1), drug amount (X2), molar ratio of CH to surfactant (X3), DCP amount (X4), hydration medium volume (X5) and hydration time (X6) were selected to be evaluated for their effect on drug entrapment efficiency (EE), which was the dependable variable (response). The six factors were examined on two levels: low and high, which were represented by transform codes of -1 and +1, respectively. A one-quarter two-level six-factor (26−<sup>2</sup> ) fractional factorial design comprising 16 runs as highlighted in the design display table was constructed by Design-Expert® 7.0 (Stat-Ease, Minneapolis, MN, USA). The factors and levels employed in the design are listed in Table 1. In order to estimate the experimental error and check the response curvature, duplicates were added at two centre points (one for each surfactant type), totally giving 20 runs. The batches were produced in random order. Data analysis was performed by using Design-Expert® 7.0 statistical software. The main effect of variables and interactions were determined according to the Equation (1) listed below:

$$E\_{\overline{X}} = \overline{\overline{\mathcal{Y}\_{(+1)}}} - \overline{\overline{\mathcal{Y}\_{(-1)}}} \tag{1}$$

**Table 1.** 2 (6−2) screening design, providing values and coded units with centre points.


Contribution was used to determine which factors were larger contributors than others and it is calculated as:

$$\text{Contribration}\_{\chi\_i}(\%) = \frac{\text{SS}\_{\text{xi}}}{\text{SS}\_{total}} \times 100$$

where *SSxi* is the sum of square of factor X*<sup>i</sup>* ; *SStotal* is the total sum of square. The data was tested for significance by analysis of variance (ANOVA) with a level of significance of 5% (*p* = 0.05).

#### 2.4.2. Optimisation of Entrapment Efficiency by Central Composite Design (CCD)

The key variables that were identified to have significant effects on the EE were subjected to the optimisation step. A CCD was used to determine the optimum conditions and to investigate how sensitive the response was to the changes in the settings of the independent variables. The CCD, as described previously, consists of a full factorial design with centre and star points, which generates enough information to fit a second-order polynomial model. The two influential factors, namely drug amount (X1) and ratio of CH to surfactant (X2) were chosen as independent variables and EE was assessed as dependent variable, the other factors were fixed based on the findings obtained from the screening design. A total of 13 experiments were performed, including five replicates on the centre point which improved the assessment of the response surface curvature and simplified the estimation of the model error. The levels of the factors in EGCG-niosome optimisation are shown in Table 2.


**Table 2.** Optimisation of epigallocatechin gallate (EGCG)-niosome by central composite design.

Data analysis was performed by using Design-Expert® 7.0 statistical software. The data were tested for significance by analysis of variance (ANOVA) with a level of significance of 5% (*p* = 0.05). The second-order Equation (2) generated is described as:

$$y = \beta\_0 + \beta\_1 \mathbf{X}\_1 + \beta\_2 \mathbf{X}\_2 + \beta\_{11} \mathbf{X}\_1^2 + \beta\_{22} \mathbf{X}\_2^2 + \beta\_{12} \mathbf{X}\_1 \mathbf{X}\_2 \tag{2}$$

where *y* stands for the predicted response (dependent variable), *β<sup>o</sup>* is the intercept; *β*<sup>1</sup> − *β*<sup>22</sup> are the regression coefficients; X<sup>1</sup> and X<sup>2</sup> stand for the main effect of the two factors; X1X<sup>2</sup> is the interactions between the main effects; and X 2 <sup>1</sup> X 2 2 are quadratic terms of the independent variables that are used to simulate the curvature of the designed space.

Checkpoint analyses were carried out to establish the reliability of the CCD regression model in describing composition parameters' effect on entrapment efficiency. The optimum point was chosen according to the prediction based on the second-order equation. Predicted and experimental values were compared to determine the correlation extent between the actual and predicted responses.

#### *2.5. Characterisation Studies*

#### 2.5.1. Particle Size, Size Distribution and Zeta Potential Analysis

The mean particle size and polydispersity index (PDI) of the optimised EGCG niosome was determined by dynamic light scattering using the photon correlation spectroscopy (PCS) technique using a Zetasizer (Malvern instruments, Malvern, UK). A dilute suspension of the niosomes was prepared with Milli Q water. The size measurement was performed in triplicate at 25 ◦C.

The zeta potential (ζ), is an indicator of particle surface charge, which may arise from the adsorption of a charged species and/or from ionization of groups that at the surface of the formed particles. It determines particles' stability in dispersion. To measure zeta potential of the optimised EGCG niosome, they were dispersed into Milli Q water (pH 7) and measured in triplicate using the Malvern Zetasizer.

#### 2.5.2. Entrapment Efficiency (EE%)

To separate the free and entrapped drugs, ultracentrifugation was used. In summary, the niosomal dispersion was centrifuged for 1 h at 4 ◦C at 41,000 rpm using a WX80 centrifuge (Beckman Sorvall, Waltham, MA, USA). The amount of EGCG in the supernatant was measured with HPLC (Agilent LC1100, Santa Clara, CA, USA) after particle separation by centrifugation. Then the niosome pellets were gently rinsed with PBS and then dissolved in a methanol solution containing 10% Triton™X-100 solution and then sonicated in a water bath sonicator for 10 min. The resulting liquid was filtered and then subjected to

concentration determination by HPLC. The following Equation (3) was used to calculate the entrapment efficiency:

$$\text{Entropyment efficiency} = \frac{(\text{Total drug} - \text{Free drug})}{\text{Total drug added}} \times 100\% \tag{3}$$

#### 2.5.3. Morphology by Scanning Electron Microscopy (SEM)

SEM (XL30S FEG, Philips, Eindhoven, Netherlands) was used to study the optimised niosomes' morphology. The niosome dispersion was diluted 20 times with Milli Q water before being dried on the grid. Gold and palladium sputter coating was applied before morphological evaluation at 25 kV.

## 2.5.4. Differential Scanning Calorimetry (DSC) and Fourier Transform Infra-Red Spectroscopy (FTIR)

IR spectroscopy and DSC were used to investigate the drug's interaction and entrapment in the vesicular structure. DSC (TA Instruments, New Castle, DE, USA, Q2000+ RCS40) was used to analyse the thermal properties of the optimised niosomes. Span60®, cholesterol, pure drug, and a physical mixture of these components and the lyophilised niosomes were placed into T-zero aluminum pans and hermetically sealed. The temperature rises to 200 ◦C at a rate of 10 ◦C/min from a starting temperature of 20 ◦C for all experimental runs.

FTIR spectra of the individual and mixture of formulation components and lyophilized niosomes were determined using a Bruker FTIP tensor 37 (Bruker Optics, Billerica, MA, USA) at a 4 cm−<sup>1</sup> resolution between 500 and 4000 cm−<sup>1</sup> .

#### *2.6. In Vitro Drug Release*

In vitro release of drug-loaded EGCG-niosomes was studied using a Franz diffusion apparatus (FDC-6, Logan Instrument Corp, Somerset, NJ, USA). EGCG solution containing the equivalent quantity of EGCG as the drug loaded niosomes was used as a control. EGCG-niosomes and drug solution were added to the donor compartment of the Franz diffusion cell, with a cellulose membrane (MW 12,000–14,000, Membra-Cel ®, Viskase, Lombard, IL, USA) sandwiched between the donor and receptor chambers. The receptor chamber was filled with PBS (pH 5.5) and the temperature was maintained at 37 ± 1 ◦C. Aliquots (400 µL) were withdrawn at pre-determined time points (15 min, 30 min, 1 h, 2 h, 3 h, 4 h, 6 h, 8 h, 12 h and 24 h) and replaced with fresh PBS (400 µL). The samples were centrifuged at 13,000 rpm for 30 min, and the supernatant was filtered (0.22 µm) and analysed with the HPLC method.

To determine the release mechanism, the release data were fitted into the Korsmeyer– Peppas model:

$$\mathcal{Q}\_t = \mathcal{K}\_k t^n$$

where *Q<sup>t</sup>* is the cumulative drug released at time *t*, *k<sup>k</sup>* is a kinetic constant characteristic of the drug/polymer system, and *n* is an exponent describing the release mechanism.

#### *2.7. Ex Vivo Skin Permeation and Deposition Studies*

#### 2.7.1. Skin Permeation Studies

The full-thickness skin samples were kindly donated by patients who underwent elective skin reduction surgeries at Middlemore hospital, Auckland. This project has been approved by the University of Auckland's Human Participants Ethics Committee (approval number: 010990). The skin samples were stored at −20 ◦C immediately after the surgeries and used within one month.

The ex vivo skin permeation and deposition studies were carried out using Franz diffusion apparatus (FDC-6, Logan Instrument Corp, Somerset, NJ, USA). EGCG-niosomes and drug solution were added to the donor compartment of the Franz diffusion cell, with a piece of full-thickness human skin sandwiched between the donor and receptor chambers

(effective diffusional area: 1.77 cm<sup>2</sup> ). The receptor chambers were filled with PBS (pH 5.5) and the temperature was maintained at 37 ± 1 ◦C. Prior to the experiments, the integrity of the skin samples was verified by a Millicell-ERS equipment (Millipore, Burlington, MA, USA) to determine the electrical resistance (ER) across the skin. The skin samples that had an ER value above the cut-off value of 27.4 kΩ·cm<sup>2</sup> were used in the study and equilibrated in PBS for 4 h before the study. For the permeation test, EGCG-niosome suspension and EGCG solution (containing an equivalent amount of EGCG as the drug loaded niosomes) were added to the donor compartments. Aliquots (400 µL) were withdrawn at 12 h and 24 h and replaced with fresh PBS (400 µL). The samples were centrifuged at 13,000 rpm for 30 min, and the supernatant was filtered (0.22 µm) and the drug concentration was determined by HPLC.

## 2.7.2. Drug Deposition in the Skin

The skin tissues were removed from the Franz diffusion cells after 12 h and 24 h of the deposition study, and the surface of the skin tissue was thoroughly cleaned with methanol and then placed on a tissue cutting board. The SC layer of the skin was removed by a tape-stripping method with Scotch® Magic tapes (3M, Maplewood, MN, USA) [37]. A tape and a 2 kg weight were placed on each skin sample for 10 s, then peeled with forceps, and 15 strippings were applied consecutively to remove the SC. Subsequently, the skin was cut into smaller pieces and 60 mg of skin tissue was placed in a MACS™ tube (Mitenyi Biotec Inc, Cambridge, MA, USA) with methanol, then dissociated using a dissociator (Mitenyi Biotec Inc, Cambridge, MA, USA). Protein was precipitated by adding TCA, then centrifuged at 13,000 rpm for 30 min, and the supernatants were filtered and analysed by HPLC.

#### 2.7.3. Visualisation of Skin Penetration and Deposition

FITC was added to the hydration medium to prepare FITC-labelled niosomes. Fullthickness human skin was placed between the donor and the receptor chamber of the Franz cells, and FITC labelled niosomes were added to the donor compartment. The skin samples were removed after 12 h and thoroughly cleaned. They were frozen and directly embedded in wax and then cut into sections with a microtome. The tissue sections were fixed on glass slides and then observed using a fluorescence microscope (DMIL LED, Leica, Wetzlar, Germany), and the images were captured.
