Functional Fragments of AIMP1-Derived Peptide (AdP) and Optimized Hydrosol for Their Topical Deposition by Box-Behnken Design

Aminoacyl-tRNA synthetase complex-interacting multifunctional protein 1 (AIMP1)-derived peptide (AdP) has been developed as a cosmeceutical ingredient for skin anti-aging given its fibroblast-activating (FA) and melanocyte-inhibiting (MI) functions. However, a suitable strategy for the topical delivery of AdP was required due to its low-permeable properties. In this study, FA and MI domains of AdP (FA-AdP and MI-AdP, respectively) were determined by functional domain mapping, where the activities of several fragments of AdP on fibroblast and melanocyte were tested, and a hydrosol-based topical delivery system for these AdP fragments was prepared. The excipient composition of the hydrosol was optimized to maximize the viscosity and drying rate by using Box-Behnken design. The artificial skin deposition of FA-AdP-loaded hydrosol was evaluated using Keshary-Chien diffusion cells equipped with Strat-M membrane (STM). The quantification of the fluorescent dye-tagged FA-AdP in STM was carried out by near-infrared fluorescence imaging. The optimized hydrosol showed 127-fold higher peptide deposition in STM than free FA-AdP (p < 0.05). This work suggests that FA- and MI-AdP are active-domains for anti-wrinkle and whitening activities, respectively, and the hydrosol could be used as a promising cosmetic formulation for the delivery of AdPs to the skin.


Cytotoxicity and Immunogenicity Tests of FA-and MI-AdP
Since safety issue is important in the cosmeceutical filed, cytotoxicity and immunogenicity tests were performed on FA-AdP and MI-AdP (i.e., Fr3 and Fr4, respectively). To guarantee the safety of these peptides, a high amount of FA-AdP or MI-AdP was treated to achieve 100 × the effective concentration. In keratinocytes, FA-AdP and MI-AdP showed no cytotoxicity up to 100 µg/mL (Figure 2a). For immunogenicity tests, FA-AdP and MI-AdP were treated to monocytes. As shown in Figure 2b, the TNF-α level of peptide-treated groups exhibited no significant difference compared with that of the control group (Figure 2b).

Cytotoxicity and Immunogenicity Tests of FA-and MI-AdP
Since safety issue is important in the cosmeceutical filed, cytotoxicity and immunogenicity tests were performed on FA-AdP and MI-AdP (i.e., Fr3 and Fr4, respectively). To guarantee the safety of these peptides, a high amount of FA-AdP or MI-AdP was treated to achieve 100 × the effective concentration. In keratinocytes, FA-AdP and MI-AdP showed no cytotoxicity up to 100 μg/mL (Figure 2a). For immunogenicity tests, FA-AdP and MI-AdP were treated to monocytes. As shown in Figure 2b, the TNF-α level of peptide-treated groups exhibited no significant difference compared with that of the control group (Figure 2b). Non-cytotoxicity and non-immunogenicity of FA-AdP and MI-AdP: (a) To test cytotoxicity of FA-AdP and MI-AdP, keratinocytes (HaCaT cells) were incubated with doxorubicin (Dox, 10 μM), AdP (100 μg/mL), FA-AdP (10 or 100 μg/mL), and MI-AdP (10 or 100 μg/mL). After 24 h incubation, cytotoxicity was determined by CCK-8 assay; (b) for the immunogenicity test, monocytes (THP-1 cells) were used. After 12 h incubation with LPS (50 ng/mL), AdP (100 μg/mL), FA-AdP (10 or 100 μg/mL), and MI-AdP (10 or 100 μg/mL), the TNF-α level was measured in cultured media by ELISA. Data are presented as the mean ± SD (n = 3). ** p < 0.01, significantly different from the others.

Viscosity of AdP Hydrosol
To demonstrate the improved topical delivery of AdP through stratum corneum (SC), FA-AdP-loaded hydrosol formulations consisting of CMC, ethanol, and glycerin were prepared and evaluated. The viscosity values of the AdP hydrosols ranged from 116.1-267.1 cps (Table 1).

Viscosity of AdP Hydrosol
To demonstrate the improved topical delivery of AdP through stratum corneum (SC), FA-AdP-loaded hydrosol formulations consisting of CMC, ethanol, and glycerin were prepared and evaluated. The viscosity values of the AdP hydrosols ranged from 116.1-267.1 cps (Table 1). Table 1.
Experimental combinations of independent variables and observed responses in Box-Behnken design.

Independent Variables
Dependent Variables The reduced mathematical model equation for viscosity (y 1 ) in uncoded units is as below: where each term exhibited a significant effect on the viscosity of the hydrosol (p < 0.05; Table 2). The viscosity values increased as the concentrations of CMC (x 1 ), ethanol (x 2 ), and glycerin (x 3 ) changed from low to high values. The coefficients of x 1 , x 2 , x 3 , and x 2 ·x 3 in the coded regression equation were 28.25, 32.91, 30.16, and 19.53, respectively, indicating the linear terms have more substantial impacts on the viscosity than the interaction term. The actual, adjusted, and predicted coefficient of determination (R 2 ) values were 0.9049, 0.8668, and 0.7364, respectively. The similarity among these values suggests that the model was not over-fitted. The response surface plots showing the effect of the independent variables on the viscosity of the hydrosol are presented in Figure 3. As expected from the equation mentioned above, the concentrations of CMC, ethanol, and glycerin exhibited positive effects on viscosity. among these values suggests that the model was not over-fitted. The response surface plots showing the effect of the independent variables on the viscosity of the hydrosol are presented in Figure 3. As expected from the equation mentioned above, the concentrations of CMC, ethanol, and glycerin exhibited positive effects on viscosity.

Drying Rate of AdP Hydrosol
The weight loss of the hydrosol on drying is shown in Figure 4a, and the drying rate constant (k) values were successfully calculated by fitting the first-order model to these data (R 2 > 0.951; Table 1). The k-values ranged from 0.0083-0.0242 min −1 , and the reduced model equation for k (y2) in coded units is as below: y2 = 0.0142 + 0.0062·x2 − 0.00114·x3 + 0.00202·x2·x2 − 0.00118·x1·x2 (2) which consists only of the terms having a significant effect on the k-values (p < 0.05; Table 2). The positive coefficients of x2 and x2·x2 terms imply that the k-values increased as the concentration of ethanol raised, whereas the negative coefficient of x3 term clearly showed the inhibitory effect of the glycerin on drying. The actual, adjusted, and predicted R 2 values of this model were 0.9839, 0.9775, and 0.9659, respectively, supporting that the model can explain the k-values more than 98% without being over-fitted. The response surface plots exhibiting the relationship between the independent variables and the k-values of the hydrosol are presented in Figure 4b-d. Noteworthy is that although the interaction term, x1·x2, has the negative coefficient value, the overall effect of the ethanol concentration on the k-values was positive under the experimental conditions.

Drying Rate of AdP Hydrosol
The weight loss of the hydrosol on drying is shown in Figure 4a, and the drying rate constant (k) values were successfully calculated by fitting the first-order model to these data (R 2 > 0.951; Table 1). The k-values ranged from 0.0083-0.0242 min −1 , and the reduced model equation for k (y 2 ) in coded units is as below: which consists only of the terms having a significant effect on the k-values (p < 0.05; Table 2). The positive coefficients of x 2 and x 2 ·x 2 terms imply that the k-values increased as the concentration of ethanol raised, whereas the negative coefficient of x 3 term clearly showed the inhibitory effect of the glycerin on drying. The actual, adjusted, and predicted R 2 values of this model were 0.9839, 0.9775, and 0.9659, respectively, supporting that the model can explain the k-values more than 98% without being over-fitted. The response surface plots exhibiting the relationship between the independent variables and the k-values of the hydrosol are presented in Figure 4b-d. Noteworthy is that although the interaction term, x 1 ·x 2 , has the negative coefficient value, the overall effect of the ethanol concentration on the k-values was positive under the experimental conditions.
Molecules 2019, 24, x FOR PEER REVIEW 6 of 13 among these values suggests that the model was not over-fitted. The response surface plots showing the effect of the independent variables on the viscosity of the hydrosol are presented in Figure 3. As expected from the equation mentioned above, the concentrations of CMC, ethanol, and glycerin exhibited positive effects on viscosity.

Drying Rate of AdP Hydrosol
The weight loss of the hydrosol on drying is shown in Figure 4a, and the drying rate constant (k) values were successfully calculated by fitting the first-order model to these data (R 2 > 0.951; Table 1). The k-values ranged from 0.0083-0.0242 min −1 , and the reduced model equation for k (y2) in coded units is as below: which consists only of the terms having a significant effect on the k-values (p < 0.05; Table 2). The positive coefficients of x2 and x2·x2 terms imply that the k-values increased as the concentration of ethanol raised, whereas the negative coefficient of x3 term clearly showed the inhibitory effect of the glycerin on drying. The actual, adjusted, and predicted R 2 values of this model were 0.9839, 0.9775, and 0.9659, respectively, supporting that the model can explain the k-values more than 98% without being over-fitted. The response surface plots exhibiting the relationship between the independent variables and the k-values of the hydrosol are presented in Figure 4b-d. Noteworthy is that although the interaction term, x1·x2, has the negative coefficient value, the overall effect of the ethanol concentration on the k-values was positive under the experimental conditions.

Optimization of AdP Hydrosol Based on Composite Desirability
The optimization plots exhibiting the desirability functions are shown in Figure 5. The maximum composite desirability (D) of 0.8912 was obtained at the CMC, ethanol, and glycerin The old values of CMC, ethanol, and glycerin were set at 0.2%, 30%, and 10%, respectively.

Optimization of AdP Hydrosol Based on Composite Desirability
The optimization plots exhibiting the desirability functions are shown in Figure 5. The maximum composite desirability (D) of 0.8912 was obtained at the CMC, ethanol, and glycerin concentrations of 0.229% (w/v), 50% (v/v), and 20% (v/v), respectively. The D-value close to 1 suggests that both the viscosity and k achieved the favorable (i.e., maximized) results as a whole. The individual desirability values for viscosity (d 1 ) and k (d 2 ) were 1.0000 and 0.79429, respectively, implying the above composition of hydrosol is more effective in maximizing viscosity than k. Of note, although the concentrations of CMC and glycerin exhibited opposite effects on viscosity and k, only the CMC concentration term displayed maximum point in the D-function. The ethanol concentration term showed only a positive effect under our experimental condition. Figure 4. The weight loss of the hydrosol on drying (a). The best-fit straight line (first-order kinetics) for each run is shown together. Response surface plots exhibiting the effect of (b) CMC (x1) and ethanol (x2); (c) CMC (x1) and glycerin (x3); and (d) ethanol (x2) and glycerin (x3) on drying rate constant (k; y2). The old values of CMC, ethanol, and glycerin were set at 0.2%, 30%, and 10%, respectively.

Optimization of AdP Hydrosol Based on Composite Desirability
The optimization plots exhibiting the desirability functions are shown in Figure 5. The maximum composite desirability (D) of 0.8912 was obtained at the CMC, ethanol, and glycerin concentrations of 0.229% (w/v), 50% (v/v), and 20% (v/v), respectively. The D-value close to 1 suggests that both the viscosity and k achieved the favorable (i.e., maximized) results as a whole. The individual desirability values for viscosity (d1) and k (d2) were 1.0000 and 0.79429, respectively, implying the above composition of hydrosol is more effective in maximizing viscosity than k. Of note, although the concentrations of CMC and glycerin exhibited opposite effects on viscosity and k, only the CMC concentration term displayed maximum point in the D-function. The ethanol concentration term showed only a positive effect under our experimental condition.

Artificial Skin Deposition Study
The Strat-M membrane (STM; artificial skin) deposition of the fluorescent dye-tagged FA-AdP hydrosol was monitored by near-infrared fluorescence (NIR) imaging ( Figure 6). Based on the results of the optimization studies, the labeled FA-AdP loaded hydrosol containing CMC, ethanol,

Artificial Skin Deposition Study
The Strat-M membrane (STM; artificial skin) deposition of the fluorescent dye-tagged FA-AdP hydrosol was monitored by near-infrared fluorescence (NIR) imaging ( Figure 6). Based on the results of the optimization studies, the labeled FA-AdP loaded hydrosol containing CMC, ethanol, and glycerin at concentrations of 0.229% (w/v), 50% (v/v), and 20% (v/v) was selected as the final formulation. At 8 h incubation, the mean fluorescence intensity of the hydrosol treated STM was 127-fold higher than that of the free FA-AdP (i.e., the solution in double-deionized water) treated STM (p < 0.05). However, the coefficient of variation (CV) values were comparable between the two groups (10.6% and 14.8% for FA-AdP-loaded hydrosol and free FA-AdP, respectively), supporting the reproducibility of the dramatic enhancing effect of the developed hydrosol. and glycerin at concentrations of 0.229% (w/v), 50% (v/v), and 20% (v/v) was selected as the final formulation. At 8 h incubation, the mean fluorescence intensity of the hydrosol treated STM was 127-fold higher than that of the free FA-AdP (i.e., the solution in double-deionized water) treated STM (p < 0.05). However, the coefficient of variation (CV) values were comparable between the two groups (10.6% and 14.8% for FA-AdP-loaded hydrosol and free FA-AdP, respectively), supporting the reproducibility of the dramatic enhancing effect of the developed hydrosol.

Discussion
AdP was dissected to determine the functional domain. Fibroblast-activating domain of AdP (FA-AdP, sh-oligopeptide 84) and melanocyte-inhibiting domain of AdP (MI-AdP, sh-oligopeptide 91 SP) were proved as 14-28 and 18-32 residues of AIMP1, respectively. Despite the 66% homology between their sequences, these two peptides showed different functions in skin-associated cells, which could be explained by the difference in secondary/tertiary structures or interaction partners. Characteristic three-dimensional structures enable even the high homology peptides to have distinct biological activities, providing different interaction interfaces [35]. Meanwhile, FA-AdP and MI-AdP are 15-amino acid (aa) length peptides, which are much shorter than AdP (41-aa length peptide). In general, truncated peptides have advantages such as the feasibility of simple and cost-effective synthesis and enhancement in skin penetration, suggesting FA-AdP and MI-AdP are attractive cosmeceutical ingredients.
However, according to the previous study, these peptides need to be delivered to the dermis-where the fibroblasts reside in-to exert skin anti-aging activity [13]. Thus, hydrosol-type formulations were developed to address this delivery issue. Ethanol and glycerol were employed as the main components of the truncated AdP-loaded hydrosol. Ethanol is a commonly used penetration enhancer in the cosmetic industry, which can improve the peptide delivery to the skin by multiple mechanisms. Ethanol can alter the physical state of the skin tissue or even can extract some of the lipids composing the SC layer, resulting in the improvement of peptide flux through the skin [36]. Moreover, ethanol may increase the peptide concentration after administration due to its rapid loss by evaporation or skin absorption. The consequent supersaturation of peptide may provide a higher skin delivery potential [26]. Ethanol can also directly drag the peptide into the skin during its penetration [37]. Although ethanol concentrations up to 80% can be used without skin irritation concerns [38], the precipitation of CMC at high concentrations of ethanol limited its maximum content in the hydrosol to 50% in this study. Glycerin can also improve the peptide

Discussion
AdP was dissected to determine the functional domain. Fibroblast-activating domain of AdP (FA-AdP, sh-oligopeptide 84) and melanocyte-inhibiting domain of AdP (MI-AdP, sh-oligopeptide 91 SP) were proved as 14-28 and 18-32 residues of AIMP1, respectively. Despite the 66% homology between their sequences, these two peptides showed different functions in skin-associated cells, which could be explained by the difference in secondary/tertiary structures or interaction partners. Characteristic three-dimensional structures enable even the high homology peptides to have distinct biological activities, providing different interaction interfaces [35]. Meanwhile, FA-AdP and MI-AdP are 15-amino acid (aa) length peptides, which are much shorter than AdP (41-aa length peptide). In general, truncated peptides have advantages such as the feasibility of simple and cost-effective synthesis and enhancement in skin penetration, suggesting FA-AdP and MI-AdP are attractive cosmeceutical ingredients.
However, according to the previous study, these peptides need to be delivered to the dermis-where the fibroblasts reside in-to exert skin anti-aging activity [13]. Thus, hydrosol-type formulations were developed to address this delivery issue. Ethanol and glycerol were employed as the main components of the truncated AdP-loaded hydrosol. Ethanol is a commonly used penetration enhancer in the cosmetic industry, which can improve the peptide delivery to the skin by multiple mechanisms. Ethanol can alter the physical state of the skin tissue or even can extract some of the lipids composing the SC layer, resulting in the improvement of peptide flux through the skin [36]. Moreover, ethanol may increase the peptide concentration after administration due to its rapid loss by evaporation or skin absorption. The consequent supersaturation of peptide may provide a higher skin delivery potential [26]. Ethanol can also directly drag the peptide into the skin during its penetration [37]. Although ethanol concentrations up to 80% can be used without skin irritation concerns [38], the precipitation of CMC at high concentrations of ethanol limited its maximum content in the hydrosol to 50% in this study. Glycerin can also improve the peptide delivery efficiency, of which underlying mechanism is the hydration of the SC layer [39]. The skin hydration can augment the percutaneous absorption of the hydrophilic compounds, including peptides [26].
Moreover, the viscosity and k-values may also affect the peptide delivery efficiency, as well as the convenience of application [40,41]. More specifically, higher viscosity is closely related to the longer skin retention of the hydrosol, and faster drying rate guarantees the rapid development of the peptide concentration gradient between the hydrosol and the skin [40,41]. According to Fick's first law of diffusion, both of these properties need to be maximized to obtain the highest flux of peptide. Thus, the viscosity and k-values were regarded as critical quality attributes (CQAs) in this study. To achieve adequate viscosity, CMC was added to the hydrosol as a thickening agent. However, as CMC precipitates in the presence of a high concentration of ethanol, the CMC was not added at more than 0.3%. Interestingly, even under this low concentration, CMC exhibited a negative effect on the drying rate in our preliminary study (data not shown), suggesting the necessity of a systematic approach to achieving its optimized content. Therefore, BBD was adopted to evaluate the relationship between the compositions and CQAs of the hydrosol more thoroughly. Based on this relationship, the optimized composition that shows the maximized D-value was obtained.
The optimized hydrosol was evaluated for its AdPs delivery efficiency to the skin. However, the employment of conventional study models that use animal skin is discouraged when testing cosmeceuticals [42]. As an alternative of animal skin, STM, a synthetic membrane, was adopted, which shares a similar structure with the human skin. A tight top layer coated with lipids resembles the SC. Underneath the top layer, the two layers of polyethersulfone (PES) are supported by the polyolefin non-woven matrix. These structures mimic the dermis and subcutaneous tissue, respectively [43]. After the artificial skin deposition study, the STM samples were evaluated using the NIR imaging instrument to quantify the FA-AdP deposition. NIR imaging provides the information on the total extent of peptide deposition on STM almost instantaneously. Moreover, as this technique does not require destructive sample pretreatments, such as homogenization, extraction, and concentration, the fluorescence intensity of the intact samples can be visualized with spatial information. The observed dramatic enhancement of FA-AdP deposition in the artificial skin suggests that the developed hydrosol could be used as a promising topical AdPs delivery system for skin anti-aging.

Collagen ELISA
Foreskin fibroblast cells (HFF-1; 5 × 10 4 cells/well) were seeded at a 24-well culture plate. Before peptide treatment, cells were pre-incubated with serum-free DMEM media for 6 h. Cells were treated with AdP, hEGF (Sigma-Aldrich, St. Louis, MO, USA) and fragments of AdP for 24 h. After 24 h treatment, media were harvested for collagen ELISA assay. Procollagen type I in the culture media was determined by Procollagen type I ELISA kit (Takara, Kusatsu, Japan) in accordance with the manufacturer instructions. Procollagen type I amount in the media was normalized by the Procollagen type I amount of control group.

Proliferation Assay
Foreskin fibroblast cells (HFF-1; 3 × 10 3 cells/well) were seeded into 96-well plate and cultivated. Cells were replaced with serum-free media and then treated with AdP, hEGF, and fragments of AdP for 48 h. Cell proliferation was determined CCK-8 assay (Dojindo, Kumamoto, Japan) in accordance with the manufacturer instructions. Proliferation was normalized by the proliferation of control group.

Cytotoxicity and Immunogenicity Tests
The safety of FA-AdP and MI-AdP was verified in terms of cytotoxicity and immunogenicity. To test the cytotoxicity of the peptides, keratinocytes (HaCaT; 3 × 10 3 cells/well) were seeded into 96-well plate and cultivated. Cells were incubated with doxorubicin (Dox, 10 µM), AdP (100 µg/mL), FA-AdP (10 or 100 µg/mL), or MI-AdP (10 or 100 µg/mL). After 24 h incubation, cytotoxicity was determined by CCK-8 assay (Dojindo) according to the manufacturer's manual. The immunogenicity of the peptides was evaluated by performing TNF-α ELISA. THP-1 cells (5 × 10 4 cells/well) were seeded on a 24-well plate. Cells were pre-incubated with 1% FBS containing media for 6 h. After pre-incubation, cells were treated with 10 ng/mL LPS, AdP, and fragments of AdP. After 12 h treatment, conditioned media were harvested for TNF-α ELISA assay. TNF-α amount in conditioned media was assessed by human TNF-α ELISA kit (BD Bioscience, San Jose, CA, USA) in accordance with the manufacturer instructions. TNF-α amount in conditioned media was normalized by TNF-α amount of control group.

Preparation of FA-AdP-Loaded Hydrosol
FA-AdP hydrosol formulations were prepared using CMC, glycerin, and ethanol. CMC was dissolved in hot double-deionized water (DDW; 80 • C) by continuous stirring at 700 rpm for 3 h to prepare 3% (w/v) working solution. Glycerin and ethanol were mixed with the CMC solution using homogenizer (Ultra-Turrax T-25, IKA Works GmbH & Co. KG, Staufen, Germany) at 14,000 rpm for 5 min. FA-AdP dissolved in DDW was added to the mixture to produce the final formulation with FA-AdP concentration of 20 µM.

Experimental Design for Hydrosol Optimization
The Box-Behnken design (BBD) was employed to optimize the physicochemical properties of FA-AdP hydrosol. The concentrations of CMC (x 1 ), ethanol (x 2 ), and glycerin (x 3 ) were chosen as independent variables. The effects of these factors on the viscosity (y 1 ) and drying rate constant (y 2 ) of hydrosol were evaluated at three levels. Including three replicates in the center, 15 experiments were performed. The experimental combinations are presented in Table 1. The best-fit quadratic models were generated using Minitab ® 18.1 (Minitab Inc., State College, PA, USA) to investigate linear, quadratic, and interaction effects of the three factors. Only the variables with a p-value < 0.05 were considered significant and included in the model. Based on the results of randomized experimental runs, the 3-dimensional surface plots were obtained. The optimization plots showing desirability functions were generated, and the combination of independent variables exhibiting the maximum composite desirability (D) was estimated. The D-value was obtained as the geometric mean of the individual desirability (d) using below formula: where d 1 and d 2 represent the individual desirability of viscosity and drying rate constant of the hydrosol, respectively, and no weighting factor was applied.

Measurement of Viscosity and Drying Rate
The viscosity of the FA-AdP hydrosol was measured by using DV-II+Pro viscometer equipped with an LV-1 spindle (Ametek Brookfield, Middleborough, MA, USA). The measurement was made between 50-60% of the instrument torque scale at 25 • C. To evaluate the drying rate, aliquots (200 µL) of the developed formulations were loaded in dishes (surface area: 63.6 mm 2 ) of a known weight and incubated in a dry oven (32 • C). Their weight changes were monitored at 0, 15, 30, 45, and 60 min. The drying rate constant (k; min −1 ) was calculated by using the first-order correlation between the weight change on drying and elapsed time. The regression lines were plotted according to the following equation: where A 0 and A t represent the weights of the hydrosol at time 0 and t (min), respectively.

Evaluation of Artificial Skin Deposition by Near-Infrared Fluorescence Imaging
The artificial skin deposition study was conducted using NIR imaging. FA-AdP (4 mg) was dissolved in 0.1 M sodium bicarbonate buffer (1 mL, pH 8.3) and slowly mixed with Alexa Fluor™ 647 NHS ester (50 µg). The mixture was incubated at room temperature for 2 h. Fluorescence-labeled FA-AdP was purified by using a PD-10 desalting column (GE Healthcare, Chicago, IL, USA), and its concentration was measured by bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Waltham, MS, USA) according to the manufacturer's protocol.
Fluorescence-labeled FA-AdP-loaded hydrosol was prepared according to the same method in Section 4.6. The composition of the hydrosol showing the highest composite desirability was selected as the optimized formulation for the skin delivery study. The concentration of fluorescence-labeled FA-AdP in the hydrosol was 20 µM. An aliquot (1 mL) of the labeled FA-AdP hydrosol or labeled FA-AdP dissolved in DDW (20 µM, control group) was applied to the upper side of STM clamped between the donor and receptor cells of Keshary-Chien diffusion cell (FCDV-15 Diffusion Cell Drive Console; Labfine, Anyang, Korea). The diffusion cell area was 1.76 cm 2 , and the experiment was performed at 32 • C under dark condition. After 8 h of incubation, the STM was collected and washed with an excess amount of DDW for 5 times. The NIR intensity of the artificial skin samples was measured by VISQUE ® InVivo Smart (Vieworks, Anyang, Korea).

Statistics
The statistical analysis (unpaired two-tailed Student's t-test) was performed using SPSS statistics software (Version 21.0; IBM Corp, NY, USA). All results are presented as the means ± standard deviation (SD), and the p-value of less than 0.05 was considered statistically significant.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.