To aid in the initial screening of many excipients for further formulation exploitation, the dissolvability of LFs in different excipients was evaluated according to the beforehand reported instrument (Goindi et al., 2015). Dissolvability study for LFs were done separately using different oils, surfactants, and cosurfactants. The oils used for the study were dimethicone, olive oil, and Isopropyl myristate. The emulsifiers tested were tween 80, tween 20, laureth-9 and PEG-40. The co-emulsifiers were Glycerin, propylene glycol, and butylene glycol. All the excipients inspected are listed in the cosmetic raw materials list. Briefly, excess LFs were placed in 2 ml test tubes containing 1 g of its vehicle and kept in a thermostated water bath shaker for 48 h at 37 ± 1 °C to dissolve LFs in situ on the vehicle. After 48 h, the mix was centrifuged (Scilogex, United States) at 10,000 rpm for 15 min. Supernatants were taken and displayed with sufficient methanol. Afterward, A portion (20 µl) of the dilution was directed to HPLC to assess vehicle loading capacity. Means that provided LFs maximum solubility were selected for experiments.

In order to test the compatibility between the distinct emulsifiers and oils, different weight ratios of oil and emulsifiers, namely 1:9, 2:8, 3:7, 4:6, 5:5, 6:4, 7:3, gently stir with a magnetic stirrer for 10 min. After that, the mixing aliquot (1 ml)and 50 ml of water (37 C) were gently stirred with a magnetic stirrer (Liu et al., 2020). Properties such as self-emulsification time and appearance were visually assessed and evaluated in light of the grading system, while emulsification tendency and progression of emulsion droplets were observed. (shown in Table 1).

In this research, PTPD was exploited using Origin 9.0 software (OriginLab, USA) to determine the emulsifying region of LFs microemulsion formulations and shrink the coverage of emulsifier, co-emulsifier and oil in the following step. Pseudo-ternary phase diagrams of LFs microemulsion including various proportions of oils, emulsifier/co-emulsifier were generated using water titration method with little modification as described in previous studies (Franklyne et al., 2021). The weight ratio of oil to the system is shown in W (oil) and the weight ratio of emulsifier to co-emulsifier in Km. To choose the optimal co-emulsifier, the weight ratio of emulsifier and co-emulsifier mixture (Km = 2) to oil was 1: 9, 2: 8, 3: 7, 4: 6, 5: 5, 6: 4. 7: 3, 8: 2 and 9: 1. In order to select the optimum ratio of Km and oil dosage range, the ratios of emulsifier to co-emulsifier mixtures (Km) were as follows 6:1,5:1,4:1,3:1,2:1,1:1,1:2 and 1:3 while the different weights of oil W (oil) to Km were mixed at ratios of 1:9, 2:8, 3:7, 4:6, 5:5, 6:4, 7:3, 8:2 and 9:1. The general weight (1.0 g) of the mixture of emulsifiers with each oil and co-emulsifier was titrated with distilled water while gently stirring with a magnetic stirrer at room temperature until the solution became clear, transparent and fluid. Then, Pseudo ternary state plots were plotted by recording the mass fractions of each of the oils, the mixture of emulsifiers with emulsifiers, and distilled water.

Statistically, a central composite design (CCD) was applied for the optimization process of formulation (Design expert software, version 12, Stat-Ease, Inc. Minneapolis-United States) (Mandpe et al., 2021). In this experiment, CCD was used to investigate two variables: the weight percentage of the W (oil) and the range of the Km. Five test levels (−α; −1; 0; +1; +α) where α = 2^n/4; and n is the number of parameters corresponding to a central point. The specific factor level settings are shown in Table 2. The selected responses were dissolvability (Y₁), droplet size (Y₂) and polydispersity index (PDI) (Y₃). With these properties, the CCD recommended 13 formulation runs and the corresponding values of the dependent variables were entered into the CCD software through experiments. Finally, limits were applied to the responses to obtain an optimized formula.

The droplets size and PDI of LFs microemulsion pellets were evaluated by dynamic light Malvern Zetasizer (Malvern Instruments Ltd., Malvern, UK) at 25°C. All measurements were performed in triplicates (Wan et al., 2017).

The viscosity of the developed formulation was determined by the NDJ-8S viscometer (Shanghai JiTai Electronics Technology Co., Ltd, China). Preparations were cup in a container and temperature was kept at 35 ± 1°C in a water bath. Spindle no. 64 was used to apply a cutting speed of 60 rpm to the prepared formulations (Akilu et al., 2019).

The pH of optimized LFs microemulsion was examined by using an already calibrated pH-meter (PHS-3E pH Tutor, Shanghai Rex Xinjing Instrument Co., Ltd., China) (Poh et al., 2019).

The morphological feature of TAC-TPGS-ME was defined by electron transfer of a microscope (TEM, JEM1400, Tokyo, Japan). Briefly, the LFs microemulsion was immersed in the Formvar coated grid for 30 s retention. More ME was wiped off with a filter paper, then 100 µl of 2% aqueous phosphatic acid solution was dripped onto the grid and reserved for 1–2 min for negative staining, then continued with a thorough flush. grid with a filter paper. The copper grid was pre-dried with indoor air.

The electrical conductivity was surveyed by a digital conductivity meter (model DDS-307, Shanghai Rex Xinjing Instrument Co., Ltd., China). Measurements were made on microemulsion samples that were successively diluted with water to 90 wt%. After immersing the electrode until it reached equilibrium and the reading was stable (Fanun, 2008). The temperature was maintained at 25 ± 0.1°C and kept by the bath temperature.

For the determination of DC 10 g of LFs self-microemulsion was prepared. In this formulation, 1.50 g of LFs was ensured. A known amount (0.1 g) of this preparation was removed, diluted appropriately with methanol and filtered through a 0.22 Pm nylon filter. Analysis of a sample of the diluted preparation (20 µl) was performed by HPLC. DC was determined according to the formula: D C ( m g ⋅ g − 1 ) =   d r u g   l o a d i n g t o t a l   a m o u n t   o f   f o r m u l a t i o n The EE of the formulated LFs microemulsion was ascertained based on earlier work which employed centrifugal ultrafiltration technique (Ghadiri et al., 2012). The diluted LFs self-microemulsion aqueous solution (3.0 mg·ml⁻¹) was placed in centrifuge tubes before centrifugation at 10,000 rpm for 30 min to guarantee complete removal of the unencapsulated LFs microemulsion. The dilution of encapsulated and unencapsulated LFs microemulsions was performed with acetonitrile (HPLC grade), respectively, to a final volume of 10 ml. Finally, a sample of the diluted formulation (20 µl) was analyzed by HPLC. The following equations were used to calculate EE: E E ( % ) = t o t a l   a m o u n t   o f   d r u g   a d d e d   − a m o u n t   o f   d r u g   i n   t h e   s u p e r n a t a n t t o t a l   a m o u n t   o f   d r u g   a d d e d × 100

The stability study of the optimized LFs microemulsion was conducted at 5 ± 2°C, 25 ± 2°C and 40 ± 2°C for 6 months. It was filled in collapsible painted aluminum tubes and stored for a certain period of time. Samples were taken and analyzed for droplet size, drug content, and pH. Other transforms in parameters such as transparency and color change were inspected visually. The physical stability of the LFs microemulsion was evaluated by centrifuging the optimized formulation at 10,000 rpm for 25 min (Amuti et al., 2021).

Female ICR mice weighing 18–22 g were selected and killed using anesthetic ether. The abdominal hair was carefully trimmed with an electric razor and the rat abdominal skin was exfoliated. Any fat and residues were eliminated from the rat skin. Experiment with fresh mice skin. All procedures were authorized by the Committee of Animal Ethicals in the Faculty of Pharmacy, Jiangsu Ocean University.

Franz diffusion cells (model TK-12D, Shanghai Yuyan Instrument Co., Ltd., China) were used in the in vitro percutaneous absorption researches. Treated skin samples were placed on a vertical glass diffusion cell with the stratum corneum facing up. The chamber provides a 1.767 cm² diffusion surface with a receptor chamber capacity of 7 ml. To ensure sink conditions, 50% (v/v) ethanol in phosphate buffered saline (pH 7.4) and 50% (v/v) ethanol in normal saline was used as the receptor fluid. The ethanol receptor was used to provide sink conditions in skin permeation researches with lipophilic drugs including LFs (Mishra et al., 2018). In addition, due to the more complex ingredients in traditional Chinese medicine, studies have used higher concentrations of ethanol to ensure that the effective ingredients can be absorbed through the skin (El Khayat et al., 2018; Zhou G et al., 2020). The high-purity LFs used in the experiment are extremely difficult to dissolve in water, and have low transdermal solubility when the concentration of phosphate buffered saline or normal saline is high. Therefore, 50% (v/v) ethanol in phosphate buffered saline (pH 7.4) and 50% (v/v) ethanol in normal saline were used as the receiving fluid in this experiment. Diffusion cells were placed in a thermostatic water bath. The receptor fluid was stirred at 100 rpm and kept at 37 C ± 0.5°C. A test formulation (1.5 ml) was encased into the donor chamber before sealing it with aluminum foil. Samples (2 ml) were taken at the indicated times (0.5, 1, 2, 3, 4, 5, 6, 8 and 10 h) and immediately replenished with an equal volume of the latest receptor fluid after each specimen. All determinations were made in triplicate and results are presented as mean ± SD. The drug content of each specimen was evaluated by HPLC.

Cumulative amount of drug (Q_n, μg·cm⁻²) released to the receptor from LFs microemulsions for time (t, h) was calculated using the following equation (Zewail et al., 2021). Q n = C n × V 0 + ∑ i = 1 n − 1 C i × V i A Where C n = Drug concentration of the sample at each sampling.

C i   = Drug concentration of the ith sample,

The skin sample after the percutaneous absorption study of 10 h was removed and cut off the effective surface area of the skin, and rinsed thoroughly with acquiring fluid until all residues of the preparation were removed. The skin was then completely cut into pieces and placed in a bottle having 2 ml of methanol. The bottle was kept in an ultrasonic bath for 15 min. The resulting methanol was pipetted into a microcentrifuge tube and centrifuged for 15 min at 10,000 rpm and 25 °C. The amount of LFs in the supernatant was determined by HPLC. All formulations were tested in triplicate. The amount of LFs in skin retention (D, μg·cm⁻²) was calculated by the following equation. D = Q s P

Q_s = LFs amount in stratum corneum,

As higher drug solubility typically contributes to higher drug loading efficiency, it is imperative to test the solubility of LFs in various oils, emulsifiers and co-emulsifiers, which may increase oral bioavailability (Shewaiter et al., 2021). The dissolvability of LFs in various vehicles was presented in Table 3. According to the measurement results, in order to increase the loading of LFs in the microemulsion, the experiment preliminarily determined that olive oil (5.63 ± 0.12 mg·g⁻¹) and Isopropyl myristate (6.21 ± 0.20 mg·g⁻¹) were used as the oil phase; laureth-9 (124.34 ± 1.18 mg·g⁻¹) and tween 20 (119.42 ± 0.94 mg·g⁻¹) were used as emulsifiers; propylene glycol and butylene glycol acts as co-emulsifiers.

The compatibility effect between oil and emulsifier is one of the important factors affecting the quality of microemulsion. When the oil and emulsifier have a good compatibility effect, the stability and emulsifying effect of the microemulsion can be improved; on the contrary, the microemulsion may not achieve the ideal emulsifying effect, it is easy to flocculate during the placement process and the stability is poor (Villanueva-Martínez et al., 2020). The experimental results are shown in Table 4. It can be seen from the table that the emulsification effect of IPM is better than Olive oil in the compatibility of different oils and emulsifiers. Among the emulsifiers, laureth-9 has the best emulsification ability. The highest A rating is obtained when IPM is compatible with laureth-9. Therefore, IPM and laureth-9 were identified as the oil and emulsifier in the formulation, respectively.

The creation of phase diagrams facilitates the determination of the concentration scope of the components for the range of presence of the microemulsion (Park et al., 2021). pseudo-ternary phase diagrams were created by water titration means at room temperature (Buyuktimkin, 2020). The pseudo-ternary phase diagrams of ME, composed of IPM as oil, laureth-9 as emulsifiers, propylene glycol and butylene glycol as co-emulsifiers, were investigated at Km of 2. The effect of different co-emulsifiers on the micro emulsifying efficiency was shown in Figure 1. Compared to butylene glycol, larger microemulsion region was inspected when propylene glycol was used as the co-emulsifiers. This reveals that compared with butylene glycol, propylene glycol has better self-emulsification efficiency. Hence, propylene glycol was chosen as an excellent co-emulsifier for microemulsion.

Km is an important factor in the preparation of the microemulsion. The pseudo-ternary phase diagrams are shown in Figure 2, IPM as oil phase and various Km (mixture of laureth-9 and propylene glycol at a ratio of 6:1, 5:1, 4:1, 3:1, 2:1, 1:1, 1:2 and 1:3), and distilled water as the water phase. It can be seen from these phase diagrams that Km = 4 has the biggest microemulsion area (shaded area). The microe mulsifying zone tended to expand with increases Km; when Km > 4, the area of the microemulsion area begins to decrease, obvious viscosity increasing during the stirring process, and the emulsification time is prolonged; Km < 1, microemulsion cannot be formed, so there is no image. According to the results of the phase diagram and comprehensive consideration of the emulsification effect, the experiment limits the range of Km to one to five.

The dosage range of IPM can also be drawn in this process. The pseudo-ternary phase diagrams are described in Figure 3. As can be seen from the figure, when W (oil) is lower than 0.1, it is a coarse emulsion; when W (oil) is higher than 0.2, microemulsion cannot be formed; when the content of W (oil) is within the range of 0.1–0.2, uniform and transparent microemulsions can be emulsified in this area. Therefore, this oil weight ratio ranging from 10 to 20% was chosen for the next test.

Optimization of the formulations was accomplished by using design expert with two distinct variables (X₁-W (oil) and X₂-Km) and three responses (Y₁-Dissolvability, Y₂-Globule size and Y3-PDI) for diverse formulations. The model has 13 trial runs as described in Table 5. The minimum and maximum Dissolvability of developed formulations ranged from 82.47 (mg·g⁻¹) to 126.49 (mg·g⁻¹). The experimental upper and lower values for droplet size and PDI were found to be 13.85–29.75nm and 0.038 to 0.214 respectively. Regression calculations generated by the software helped determine the type of fit of the polynomial model to the chosen responses. The most suitable polynomial model for Dissolvability (Y₁), droplet size (Y₂) and PDI (Y₃) were found to be quadratic (p-value < 0.005).

As shown in Table 6, the high R ² of 0.9496 (Dissolvability),0.9911 (Droplet size) and 0.9764 (PDI) deduced that the model fitted well. The diversity among Predicted R ² (0.8216-Dissolvability, 0.9376-Droplet size and 0.9595-PDI) and the Adjusted R ² (0.9137-Dissolvability, 0.9848-Droplet size and 0.8386-PDI) was <0.2, which indicated that the predicted R ² is compatible with the reasonably adjusted R2. Usually, the sensitivity of the Adeq is used to measure the signal-to-noise ratio at >4, while the ratio of 13.1774 (Dissolvability), 36.8464 (Droplet size) and 22.4309 (PDI) indicated an adequate signal and excellent model competence. Overall, these results verify the adequacy of the model terms to stand for the trial data and forecast the three parameters analysed (Ge et al., 2014). The following sets of formulas provide polynomial model fitting equations for all three responses.

S o l u b i l i t y   ( Y 1 ) = 124.29 − 11.85 × X 1 + 10.75 × X 2 − 3.70 × X 1 × X 2 − 11.99 × X 12 − 9.00 × X 22 ( r = 0.9496 ,   P = 0.0002 ) Coded equation (1);

Usually, Response surface and contour maps can intuitively reflect the factors and response values as well as the interaction between each factor (Gull et al., 2020). Figure 4 Shows a 3D graph of the response surface and contour based on change (W-oil and Km) and the response (Dissolvability, Droplet size and PDI). It can be obtained from Figure 4, when Km is constant, the solubilization effect of microemulsion on LFs increases slowly at first and then decreases slowly with the increase of W (oil). The Droplet size decreases first and then increases with the increase of W (oil). PDI decreases first and then increases slowly with the increase of W (oil); When W (oil) is constant, the solubilization effect of microemulsion on LFs increases first and then decreases slowly with the increase of Km. PDI decreases first and then increases with the increase of Km. The Droplet size always decreases with the increase of Km.

The optimal conditions for the preparation of LFs microemulsion were Km = 3.73 and W (oil) = 14.76% under the conditions of minimum droplet size, PDI and maximum dissolvability. The optimal formulation of LFs microemulsion was composed of 14.76% IPM and 67.22% laureth-9 and 18.02% propylene-glycol.

The predicted formulation was the oil phase taking the percentage of 14.76% and the Km at a ratio of 3.73. The predicted dissolvability, particle size and PDI were 128.29 ± 0.00 mg·g⁻¹, 12.74 ± 0.00 nm and 0.046 ± 0.00, respectively. In order to validate the predicted results, the predicted and observed responses for the optimized incorporation of the LFs microemulsion were compared and the result is presented in Table 7. The observed dissolvability, particle size and PDI are 126.76 ± 1.74 mg·g⁻¹, 12.76 ± 0.48 nm and 0.047 ± 0.03, respectively. More importantly, there is no significant diversity among the predicted value of each indicator and the measured value, indicating that the experimental design of the study is appropriate.

According to the results of the laboratory, the dissolvability of the optimal microemulsion formulation for LFs was 126.76 mg·g⁻¹. It was found in experiments that when the dosage of the drug was more than 150 mg·g⁻¹, the microemulsion diluted with water would precipitate out after 24 h. The reason might be that when the dissolvability was too large, the microemulsion system could not form a stable interface film and LFs crystals precipitated out. Therefore, the dosage of LFs in the prescription was determined to be 150 mg·g⁻¹.

The droplet size and PDI of the LFs microemulsion estimated from the software were 12.80 nm and 0.050, respectively. Experimental mean droplet size and PDI values were found to be 12.68 ± 0.12 nm and 0.049 ± 0.005, respectively. A representative droplet size distribution plot is shown in Figure 5. The viscosity and pH of the optimized formulation were recorded as 3.05 ± 0.35 cPs and 7.8 ± 0.25, respectively. pH was close to normal skin, implying its relative compatibility with the skin.

TEM micrograph of the optimized preparations was obtained. The granule morphology of the LFs microemulsion is shown in Figure 6. The TEM image showed that the nanoparticles were nearly spherical and had homogeneous droplet size. The homogeneous spherical droplets presented in the photomicrograph show a nanosized in the recommended range for ME systems and a homogeneous distribution without agglomeration.

Conductivity measurement can supply a useful message of microstructural transitions occurring in microemulsion systems (Bharatiya et al., 2016). Figure 5 shows the change curve of microemulsion conductivity with the aqueous phase. After the aqueous phase dilution reaches 30 wt% aqueous phase, a precipitous and linear raise of conductivity is noticed, which is due to the emergence of percolation behavior (Guo et al., 2019). This phenomenon means that the addition of the aqueous phase causes swelling of the W/O micelles. The viscous collision between water droplets leads to droplet melting, in which conductive channels are well developed. However, when aqueous phase increases up to 60 wt%, conductivity causes the individual droplets to develop into a network of conductive channels and B.C. the structure emerges (Hathout and Nasr, 2013). After that, with the addition of the aqueous phase, the conductivity gradually decreased and the structure changed to an O/W microemulsion. Thus, the conductivity measurement results (as shown in Figure 7) describe the continuous microstructural transitions in LFs microemulsions from W/O micelles to B.C. type (between 60 wt% ∼ 75 wt% aqueous phases) after dilution of the aqueous phase. finally, to O/W structure (more than 75 wt% aqueous phases).

In fact, the determination of EE and DC is an important factor in evaluating the physical properties of liposomes, solid lipid nanoparticles and micro emulsification of nanoemulsions (Shewaiter et al., 2021). Therefore, EE and DC are frequently used as formulation screening indicators, as a satisfactory, clear and stable formulation is usually characterized by high EE and DC.

Herein, the calculated EE and DC of LFs microemulsion were 97.28 ± 0.13% and 122.67 ± 0.40 mg·g⁻¹, respectively. As the water content increases, the drug content will decrease. In comparison with other microemulsions (Kaur et al., 2017), The EE and DC predicted in this research was higher, indicating that LFs are mostly incorporated into the microemulsion. Thus, the LFs microemulsion prepared in this research met the target of obtaining a satisfactory drug delivery system with higher EE and DC.

The stability study of the optimized LFs microemulsion showed that after 6 months of storage at different temperatures, the microemulsion was transparent without discoloration (Figure 8). The drug content in stored formulations of the optimized LFs microemulsion changed slightly after testing. Organoleptic features such as droplet size and physical appearance were also noticed and no significant transforms were found. Centrifuge examinations showed excellent physical stability of all preparations. The O/W structure of the optimized LFs microemulsion with an observed pH range of 7.5–8.5 provided a favorable condition to prevent drug hydrolysis.

Evaluation of transdermal absorption is one of the main experimental steps in evaluating transdermal drug delivery systems (Patel et al., 2020). The permeation data is expressed as a cumulative amount of drug permeated as a function of time. After the drug penetration reaches a steady state, the slope of the straight line portion is the penetration rate of the drug ( J s s ) (You et al., 2019). The permeability coefficient ( K p ) was calculated using the following equations: K p = J s s C 0 Where C 0 is the initial concentration of the drug. Use time T (h) as the abscissa, and 10 h average cumulative permeation Q n (μg·cm⁻²) as the ordinate to draw the cumulative permeation-time curve, as shown in Figure 9. The total permeation and flux of LFs microemulsion from 50% (v/v) ethanol in normal saline as the receiving fluids were 644.95 ± 6.73 μg cm⁻² and 69.40 μg cm⁻² h⁻¹, respectively. The total permeation and flux of LFs microemulsion from 50% (v/v) ethanol in phosphate buffered saline as the receiving fluids were 519.43 ± 4.89 μg cm⁻² and 57.39 μg cm⁻² h⁻¹, respectively (Table 8). In addition, we also compared the transdermal penetration of free LFs at equal concentrations. As expected, the results of the skin permeability of the LFs suspension (15 mg/g deionized water) were very poor. Under the same condition of receiving liquid, the LFs suspension has not been detected to penetrate the skin after 10 h. This is due to the low water solubility of LFs. Therefore, the prepared microemulsion improves the skin penetration rate of LFs. The prepared LFs microemulsion was the first time to carry LFs through the skin.

The skin retention of LFs microemulsion was 4.25 and 9.98 μg for different receptor fluids (Figure 10). After 10 h of in vitro administration, the skin retention of LFs microemulsion in phosphate buffer saline with 50% (v/v) ethanol was twice of normal saline with 50% (v/v) ethanol. The reason for this may be that the normal saline increases the permeability rate of the drug, which reduces the drug retention in the skin (Goindi et al., 2015). There were residues of LFs microemulsion in the two different medium conditions, which further indicated that LFs microemulsion could stay in the skin, and could prolong the time of active ingredients in the skin, and had a better effect.