Microneedle-Assisted Percutaneous Delivery of a Tetramethylpyrazine-Loaded Microemulsion

This study examined the efficacy of the percutaneous delivery of a tetramethylpyrazine-loaded microemulsion (TMP-ME) on skin pretreated with microneedles (MN). The TMP-ME formulation was optimized in vitro with skin permeation experiments, using a uniform experimental design, guided by a pseudo-ternary phase diagram, in which the TMP skin permeation level and mean particle size were indices. The effects of MN pretreatment on skin permeation by TMP-ME were assessed using in vitro skin permeation, in vivo skin microdialysis, and pharmacokinetic studies in rats. The influence of MN pretreatment on the skin barrier function was evaluated by measuring the electrical resistance of rat skin before and after MN insertion. In the optimal formulation of TMP-ME, the weight percentages of Maisine® 35-1 (oil phase), Labrasol® (surfactant), and Transcutol® P (co-surfactant) were 7%, 30% and 10%, respectively, with 1.5% TMP loading. In the in vitro skin permeation study, MN-assisted TMP-ME exhibited a two-fold increase in a 24-h cumulative TMP permeation compared with TMP-ME alone (p < 0.05). In the skin microdialysis study, TMP in MN-assisted TMP-ME exhibited a 1.25-fold increase in Cmax, a 0.93-fold decrease in Tmax, and a 0.88-fold increase in AUC0–t (p < 0.05). Similarly, in the pharmacokinetic study, TMP in MN-assisted TMP-ME exhibited a 2.11-fold increase in Cmax, a 0.67-fold decrease in Tmax, and a 1.07-fold increase in AUC0–t (p < 0.05). The percutaneous electrical resistance of rat skin before and after MN insertion was 850 ± 50 Ω/cm2 and 283 ± 104 Ω/cm2 respectively, indicating that MN dramatically compromises the skin barrier. These results suggest that MN assistance increases the skin permeation rate and the extent of percutaneous absorption of TMP-ME, and that the mechanism may involve the reversible barrier perturbation effect. The rate and extent of percutaneous absorption of TMP-ME can be significantly enhanced by MN assistance, possibly because MN causes a reversible barrier perturbation effect on skin.


Introduction
The percutaneous route is attractive for drug delivery because it facilitates a stable level of drug in the plasma, avoids the first pass effect, and has a high patient compliance. However, transdermal delivery is currently limited to potent compounds that have low molecular weights and favorable n-Octanol/water partition coefficients, due to the formidable barrier function of the stratum corneum (SC). To address this challenge, methods for enhancing passive and active skin penetration are being

Optimization of Formulation for TMP-ME
The pseudo-ternary phase diagram for Maisine ® 35-1, Labrasol ® , Transcutol ® P, and water is shown in Figure 1. To optimize the formulation, nine TMP-ME formulations were compared using a uniform experimental design guided by the pseudo-ternary phase diagram. The accumulation of TMP in skin in vitro is shown as a function of time in Figure 2 for each formulation, and TMP accumulation values and mean particle sizes are listed in Table 1.
Using these data, the mathematical functions relating TMP accumulation (Y 1 ) or mean particle size (Y 2 ) for TMP-ME, and the weight percentages of the oil phase (X 1 ) and surfactant and co-surfactant (X 2 ) components in TMP were modeled using multiple linear regression. The resulting equations are shown below: Y 1 = 2163.81X 1 − 128.75X 2 − 6.95X 1 X 2 − 125.12X 1 2 + 1.75X 2 2 − 2815.02 Y 2 = −33.137X 1 − 14.626X 2 + 0.389X 1 X 2 + 1.888X 1 2 + 0.137X 2 2 + 470.191 The TMP-ME formulation was optimized using these functions. In the predicted optimum formulation, the weight percentages of the oil phase, surfactant, and co-surfactant were 7%, 30%, and 10%. The predicted accumulation of TMP over 24 h and the mean particle size were 1904.77 µg/cm 2 and 73.82 nm. TMP-ME was then prepared using these parameters and analyzed as described earlier.
The measured accumulation of TMP and the mean particle size were 2471.02 ± 41.57 µg/cm 2 and 79.04 ± 0.68 nm, with the relative standard deviation (RSD) between predicted and measured values ranging from 28.11% to 32.11%, and 6.07% to 7.91%, respectively. These results suggest that the characteristics of TMP-ME can be reliably predicted using the established functions. The relatively high RSD observed between the predicted and measured values for TMP accumulation may be due to the intrinsically large experimental error in the in vitro experiment. Using these data, the mathematical functions relating TMP accumulation (Y1) or mean particle size (Y2) for TMP-ME, and the weight percentages of the oil phase (X1) and surfactant and co-surfactant (X2) components in TMP were modeled using multiple linear regression. The resulting equations are shown below: The TMP-ME formulation was optimized using these functions. In the predicted optimum formulation, the weight percentages of the oil phase, surfactant, and co-surfactant were 7%, 30%, and 10%. The predicted accumulation of TMP over 24 h and the mean particle size were 1904.77 µg/cm 2 and 73.82 nm. TMP-ME was then prepared using these parameters and analyzed as described earlier. The measured accumulation of TMP and the mean particle size were 2471.02 ± 41.57 µg/cm 2 and 79.04 ± 0.68 nm, with the relative standard deviation (RSD) between predicted and measured values ranging from 28.11% to 32.11%, and 6.07% to 7.91%, respectively. These results suggest that the characteristics of TMP-ME can be reliably predicted using the established functions. The relatively high RSD observed between the predicted and measured values for TMP accumulation may be due to the intrinsically large experimental error in the in vitro experiment.    Using these data, the mathematical functions relating TMP accumulation (Y1) or mean particle size (Y2) for TMP-ME, and the weight percentages of the oil phase (X1) and surfactant and co-surfactant (X2) components in TMP were modeled using multiple linear regression. The resulting equations are shown below: The TMP-ME formulation was optimized using these functions. In the predicted optimum formulation, the weight percentages of the oil phase, surfactant, and co-surfactant were 7%, 30%, and 10%. The predicted accumulation of TMP over 24 h and the mean particle size were 1904.77 µg/cm 2 and 73.82 nm. TMP-ME was then prepared using these parameters and analyzed as described earlier. The measured accumulation of TMP and the mean particle size were 2471.02 ± 41.57 µg/cm 2 and 79.04 ± 0.68 nm, with the relative standard deviation (RSD) between predicted and measured values ranging from 28.11% to 32.11%, and 6.07% to 7.91%, respectively. These results suggest that the characteristics of TMP-ME can be reliably predicted using the established functions. The relatively high RSD observed between the predicted and measured values for TMP accumulation may be due to the intrinsically large experimental error in the in vitro experiment.

In Vitro Skin Permeation of MN-Assisted TMP-ME
The cumulative permeation of TMP in skin in vitro is shown in Figure 3 for MN-assisted TMP-ME and TMP-ME alone. Compared with TMP-ME alone, a two-fold increase in cumulative permeation was observed for MN-assisted TMP-ME (2471.02 ± 41.57 µg/cm 2 vs. 4963.61 ± 254.70 µg/cm 2 , p < 0.05). These results suggest that MN assistance can dramatically facilitate the penetration of skin by TMP-ME in vitro, consistent with the results obtained previously [28].

In Vitro Skin Permeation of MN-Assisted TMP-ME
The cumulative permeation of TMP in skin in vitro is shown in Figure 3 for MN-assisted TMP-ME and TMP-ME alone. Compared with TMP-ME alone, a two-fold increase in cumulative permeation was observed for MN-assisted TMP-ME (2471.02 ± 41.57 µg/cm 2 vs. 4963.61 ± 254.70 µg/cm 2 , p < 0.05). These results suggest that MN assistance can dramatically facilitate the penetration of skin by TMP-ME in vitro, consistent with the results obtained previously [28].

In Vivo Skin Microdialysis of MN-Assisted TMP-ME
To examine the influence of MN treatment on the ability of TMP-ME to penetrate skin, microdialysis experiments were conducted in vivo using rats. TMP levels in the recovered dialysate were measured over time for rats treated using MN-assisted TMP-ME or TMP-ME alone ( Figure 4). The pharmacokinetic parameters for both treatments are listed in Table 2. Compared to rats treated with TMP-ME alone, a 1.25-fold increase in Cmax, a 0.93-fold decrease in Tmax, and a 0.88-fold increase in AUC0-t were observed for TMP in rats treated with MN-assisted TMP-ME (p < 0.05). These results suggest that the rate and extent of percutaneous TMP absorption can be enhanced significantly by the application of MN-assisted TMP-ME, which is consistent with previous reports [29].

In Vivo Skin Microdialysis of MN-Assisted TMP-ME
To examine the influence of MN treatment on the ability of TMP-ME to penetrate skin, microdialysis experiments were conducted in vivo using rats. TMP levels in the recovered dialysate were measured over time for rats treated using MN-assisted TMP-ME or TMP-ME alone ( Figure 4). The pharmacokinetic parameters for both treatments are listed in Table 2. Compared to rats treated with TMP-ME alone, a 1.25-fold increase in C max , a 0.93-fold decrease in T max , and a 0.88-fold increase in AUC 0-t were observed for TMP in rats treated with MN-assisted TMP-ME (p < 0.05). These results suggest that the rate and extent of percutaneous TMP absorption can be enhanced significantly by the application of MN-assisted TMP-ME, which is consistent with previous reports [29].

In Vitro Skin Permeation of MN-Assisted TMP-ME
The cumulative permeation of TMP in skin in vitro is shown in Figure 3 for MN-assisted TMP-ME and TMP-ME alone. Compared with TMP-ME alone, a two-fold increase in cumulative permeation was observed for MN-assisted TMP-ME (2471.02 ± 41.57 µg/cm 2 vs. 4963.61 ± 254.70 µg/cm 2 , p < 0.05). These results suggest that MN assistance can dramatically facilitate the penetration of skin by TMP-ME in vitro, consistent with the results obtained previously [28].

In Vivo Skin Microdialysis of MN-Assisted TMP-ME
To examine the influence of MN treatment on the ability of TMP-ME to penetrate skin, microdialysis experiments were conducted in vivo using rats. TMP levels in the recovered dialysate were measured over time for rats treated using MN-assisted TMP-ME or TMP-ME alone ( Figure 4). The pharmacokinetic parameters for both treatments are listed in Table 2. Compared to rats treated with TMP-ME alone, a 1.25-fold increase in Cmax, a 0.93-fold decrease in Tmax, and a 0.88-fold increase in AUC0-t were observed for TMP in rats treated with MN-assisted TMP-ME (p < 0.05). These results suggest that the rate and extent of percutaneous TMP absorption can be enhanced significantly by the application of MN-assisted TMP-ME, which is consistent with previous reports [29].

Pharmacokinetic Study of MN-Assisted TMP-ME
TMP levels in plasma for rats treated using MN-assisted TMP-ME and TMP-ME alone are shown in Figure 5. The corresponding pharmacokinetic parameters are listed in Table 3. Compared with TMP-ME alone, a 2.11-fold increase in C max , a 0.67-fold decrease in T max , and a 1.07-fold increase in AUC 0-t were observed for TMP in rats treated with MN-assisted TMP-ME (p < 0.05). These results suggest that the rate and extent of percutaneous absorption of TMP-ME can be enhanced dramatically buy MN assistance, consistent with our skin microdialysis experiments and previous reports [30].

Pharmacokinetic Study of MN-Assisted TMP-ME
TMP levels in plasma for rats treated using MN-assisted TMP-ME and TMP-ME alone are shown in Figure 5. The corresponding pharmacokinetic parameters are listed in Table 3. Compared with TMP-ME alone, a 2.11-fold increase in Cmax, a 0.67-fold decrease in Tmax, and a 1.07-fold increase in AUC0-t were observed for TMP in rats treated with MN-assisted TMP-ME (p < 0.05). These results suggest that the rate and extent of percutaneous absorption of TMP-ME can be enhanced dramatically buy MN assistance, consistent with our skin microdialysis experiments and previous reports [30].

Influence of MN Treatment on Transcutaneous Electrical Resistance
The transcutaneous electrical resistance of rat skin was measured in vitro before and after insertion of MN. The resistances were 850 ± 50 Ω/cm 2 and 283 ± 104 Ω/cm 2 . respectively, suggesting that the skin barrier is severely compromised by treatment with MN. Our data are consistent with previous reports about the reversible reduction of skin integrity by MN [31,32].  4.56 ± 0.17 * 6.03 ± 0.33 Note: * represents p < 0.05, compared with TMP-ME.

Influence of MN Treatment on Transcutaneous Electrical Resistance
The transcutaneous electrical resistance of rat skin was measured in vitro before and after insertion of MN. The resistances were 850 ± 50 Ω/cm 2 and 283 ± 104 Ω/cm 2 . respectively, suggesting that the skin barrier is severely compromised by treatment with MN. Our data are consistent with previous reports about the reversible reduction of skin integrity by MN [31,32].

Formulation Optimization of TMP-ME
The pseudo-ternary phase diagram for the oil phase, surfactant, and co-surfactant of TMP-ME was constructed using the water titration method as reported previously [33]. Briefly, the ratios of the weight of the oil phase (Maisine ® 35-1) relative to the weight of the mixture of surfactant (Labrasol ® ) and co-surfactant (Transcutol ® P) were 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8 and 1:9, with the K m (the weight ratio of surfactant to co-surfactant) fixed at 3:1 (Table 4). Distilled water was then added drop-wise to these mixtures, and the weight ratios of oil phase, surfactant, co-surfactant, and water were recorded when the solutions changed from turbid from clear, and vice versa. Notes: X 1 represents the weight content of Maisine35-1 (oil phase) in the TMP-ME formulation; X 2 represents the weight content of Labrasol and Transcutol P (surfactant and co-surfactant); the loading of TMP is 1.5% (w/w).
The formulation for TMP-ME was optimized using a uniform experimental design and the pseudo-ternary phase diagram. The weight content of Maisine ® 35-1 (X 1 ), and Labrasol ® and Transcutol ® P (X 2 ) were used as factors. The indices were the quantity of TMP that permeated the skin over a 24 h period (Y 1 ), and the mean particle size of TMP-ME (Y 2 ).
The in vitro skin penetration behavior of the TMP-MEs was evaluated using a Franz-type diffusion cell. The effective penetration area was 1.7 cm 2 and the receptor compartment volume was 16 cm 3 . Freshly-excised abdominal skin from male Sprague-Dawley (SD) rats was used as a permeation barrier. TMP-ME (1 mL) was added to the donor compartment. Thirty percent ethanol and 20% PEG-400 in normal saline (v/v) was added to the receptor compartment. At predetermined times, 1 mL of receptor medium was withdrawn, and replaced with the same volume of blank medium at 32 • C. The level of TMP in the receptor medium was determined using HPLC, and the total quantity that permeated the skin over a 24 h period was calculated as: In this expression, C n is the TMP concentration in the receptor medium at each sampling time, C i is the drug concentration of the sample, V and V i are the volumes of the receptor medium and the sample, and S is the effective diffusion area.
Mean particle sizes for the TMP-MEs were measured using a Malvern Nano ZS90 Zetasizer (Malvern Instruments Ltd., Worcestershire, UK).
The mathematical functions relating indices and factors were established using a multiple linear regression method. On the basis of these functions, formulations for which the 24 h TMP permeation level exceeded 1900 µg/cm 2 and the TMP-ME mean particle size was less than 75 nm were re-screened. X 1 was varied from 5% to 10% (in 0.1% increments), and X 2 was varied from 40% to 60% (in 1% increments).
In the optimized formulation of TMP-ME, the weight percentages of Maisine ® 35-1, Labrasol ® , and Transcutol ® P were 7%, 30%, and 10%, respectively, and the loading of TMP was 1.5%. Triplicate TMP-ME samples were prepared using the optimum formulation. Mean particle size and TMP skin permeation over a 24 h period were measured and compared to the predicted values.

In Vitro Skin Permeation of MN-Assisted TMP-ME
The influence of MNs on the ability of TMP-ME to permeate skin was initially investigated using in vitro skin permeation experiments. For both TMP-ME alone and MN-assisted TMP-ME, experimental conditions and methods for TMP quantitation in the receptor medium were identical to those described above, except that the rat skin was pretreated with MN by manual insertion for 2 min in the MN-assisted TMP-ME analysis.

In Vivo Skin Microdialysis of MN-Assisted TMP-ME
Male SD rats were randomly assigned to either TMP-ME or MN-assisted TMP-ME groups (n = 5 for each). The skin microdialysis experiments for the TMP-ME group were conducted following methods described previously [34]. For the MN-assisted TMP-ME group, the protocol was identical, except that the rat skin was pretreated with MNs by manual insertion for 2 min. For both groups, rats were intraperitoneally anesthetized with urethane (1.3 g/Kg), and their abdominal hair was clipped. Microdialysis tubing (20 mm long, 200 µm inner diameter, 13,000 Da cut-off molecular weight) was inserted into the skin and perfused with 30% ethanol in phosphate-buffered saline (PBS) at 0.22 mL/h using a WZ-50C6 Micro Infusion pump (Smiths Medical, Norwell, MA, USA). an hour and a half after perfusion, 2 mL of TMP-ME was applied to the rat skin, and dialysate samples were collected from the tubes every 0.5 h for 8 h. TMP levels in the dialysate samples were determined using UPLC-MS/MS (Thermo Fisher Scientific Inc., Waltham, MA, USA).

Pharmacokinetics of MN-Assisted TMP-ME
The effect of MNs on the pharmacokinetic behavior of TMP-ME was assessed in male SD rats. The study design and treatments for the TMP-ME and MN-assisted TMP-ME groups were the same as in the skin microdialysis study. At 0.1, 0.5, 1, 2, 4, 6, 8, 12 and 24 h after the application of TMP-ME, 0.3 mL of blood was withdrawn from each rat. Plasma samples were immediately separated by centrifugation at 3000 rpm for 5 min. Plasma TMP levels were determined using UPLC-MS/MS. The non-compartmental pharmacokinetic parameters for TMP in both groups were calculated using BAPP2.0 (Bioavailability Program Package 2.0, Center for Metabolism and Pharmacokinetics, China Pharmaceutical University, Nanjing, China).

Influence of MN Treatment on Transcutaneous Electrical Resistance
The influence of MN treatment on skin barrier function was evaluated using transcutaneous electrical resistance [32]. Freshly-excised rat skin was mounted on a Franz-type diffusion cell, and PBS was used as the medium in both the donor and receptor compartments. Transcutaneous electrical resistance before and after MN insertion was measured in triplicate using Ag/AgCl electrodes.

Conclusions
The rate and extent of percutaneous absorption by TMP-ME can be significantly enhanced by MN assistance, which may be related to the reversible barrier perturbation effect of MN on skin.