Potential of Topical Microemulsion Serum Formulations to Enhance In Vitro and Clinical Anti-Skin Wrinkle Benets of Cordyceps Militaris Extracts

This research is the rst to investigate the anti-skin wrinkle properties of Cordyceps mili-taris extracts both in vitro and in vivo. Anti-skin wrinkle activities of C. militaris were investigated by means of matrix metalloproteinase-1 (MMP-1), elastase, and hyaluronidase inhibitions. Microemulsions and topical serum formulation containing C. militaris extract were developed. The anti-skin wrinkle ecacy and irritation properties of the topical serum formulations were clinically investigated in human volunteers. Cordycepin was identied as a major component of C. miliaris extract that was responsible for MMP-1, elastase, and hyaluronidase inhibition. The C. miliaris water extract possessed the most potent inhibition on MMP-1 (77.9 ± 5.3%) and elastase (84.4 ± 4.0%). Interestingly, CW was as a potent MMP-1 and elastase inhibitor as oleanolic acid and EGCG. CW was incorporated into the microemulsion with the smallest internal droplet size (146.1 ± 1.5 nm) and further developed as a topical serum formulation. The resulting serum formulation effectively enhanced skin moisture (42.2 ± 14.2%), increased the skin elasticity (39.9 ± 7.3%), and induced no skin irritation in 30 human volunteers. The effectiveness on the skin was detected after 1 week of the applications. Therefore, it was suggested as an effective anti-skin wrinkle formulation. emollient (4.7 ± 0.5/5.0), skin absorption (4.6 ± 0.6/5.0), skin texture (4.6 ± 0.7/5.0), skin slippery (4.6 ± 0.8/5.0), no skin irritation (4.9 ± 0.3/5.0), skin elasticity enhancement (4.3 ± 0.8/5.0), winkle reduction (4.2 ± 0.9/5.0), and non- sticky (4.3 ± 1.1/5.0). CA: fractionated ethyl acetate extract; CC: crude ethanolic extract; CE: fractionated ethanolic extract; CH: fractionated hexane extract, CW: water extract; ECM: extracellular matrix; HaCaT: human keratinocyte; IC 50 : The half maximal inhibitory concentration; MMP-1: Matrix metalloproteinase-1.


Introduction
Human skin is an organ directly contact with the external environments, which are primary reasons for skin damage and leading to skin-winkles [1]. Skin ageing is naturally occurred by various factors, such as genetics, metabolic process, environments, mechanical stress, ultraviolet (UV) irritation, pollution, etc. [2].
UV irradiation or photo ageing is the one of the important factors which promote the production of matrix metalloproteinase (MMP), which leads to the decrease or degradation of collagen bers [3]. Recently, plenty of cosmetic products claim to improve skin problems, especially anti-ageing and reduce wrinkles.
The active ingredients for anti-ageing can be obtained from both natural and synthetic process. However, some chemicals could cause serious side effects, such as retinoic acid which has been reported to cause serious skin irritations [3]. Therefore, tretinoin (all-trans retinoic acid) is currently not allowed in Europe (cosmetic regulation 1223/09) and strictly regulated by Food and Drug Administration (FDA) [4]. trypan blue were supplied from Invitrogen™ (Grand Island, NY, USA). GlutaMAX™-I supplement was supplied from Thermo Fisher Scienti c, Inc. (Grand Island, NY, USA). HPLC-grade methanol was supplied from Labscan (Dublin, Ireland).

Extraction of C. militaris
Crude ethanolic extract (CC), fractionated hexane extracts (CH), fractionated ethyl acetate extracts (CA), fractionated ethanolic extracts (CE), and water extract (CW) of C. militaris were prepared by according to the method previously described by Marsup et al. [6]. All organic solvents were completely eliminated by rotary evaporator (N-1001S-W, Eyela, Tokyo, Japan), whereas, the DI water was eliminated by a freeze dryer (Beta 2-8 LD plus, Christ, Osterode, Germany). All extracts were kept in a well-closed container in a refrigerator until further experiments.
In vitro determination of anti-wrinkle activities of C. militaris extracts Determination of matrix metalloproteinase-1 (MMP-1) inhibition C. militaris extracts and cordycepin were investigated for their inhibitory activity against MMP-1 by using spectrophotometric method according to the previously described method of Chaiyana et al. (2019) which was slightly modi ed from Thring et al. (2009) [10,11]. The nal concentration of cordycepin and C. militaris extracts in the tested system was 0.1 mg/ml. Blank of sample, which was a sample solution, was performed to reduce an interference from colored samples. Blank of control, which was the native solvent, was performed to reduce the interference from each solvent. The MMP-1 inhibition was calculated using the following equation; MMP-1 inhibition (%) = [1-(A/B)] ×100, where A is a different absorbance between sample solution with MMP-1 and the sample solution without MMP-1 and B is a different absorbance between MMP-1 solution and its native solvent. Oleanolic acid was used as a positive control. The experiments were analyzed in triplicate.
Elastase enzyme inhibition C. militaris extracts and cordycepin were investigated for their inhibitory activity against elastase by using spectrophotometric method according to the previously described method of Chaiyana et al. (2019) which was slightly modi ed from Thring et al. (2009) [10,11]. The nal concentration of cordycepin and C. militaris extracts in the tested system was 0.1 mg/ml. Blank of sample, which was a sample solution, was performed to reduce an interference from colored samples. Blank of control, which was the native solvent, was performed to reduce the interference from each solvent. The inhibition of elastase was calculated using the following equation; Elastase inhibition (%) = [1-(A/B)] ×100, where A is a different absorbance between sample solution with elastase and the sample solution without elastase and B is a different absorbance between elastase solution and its native solvent. EGCG was used as a positive control. The experiments were analyzed in triplicate.

Hyaluronidase inhibition
Page 5/25 C. militaris extracts and cordycepin were investigated for their hyaluronidase inhibitory activity by using spectrophotometric method according to the previously described method of Nema et al. (2011) [12] with slight modi cations. The nal concentration of cordycepin and C. militaris extracts in the tested system was 0.1 mg/ml. Blank of sample, which was a sample solution, was performed to reduce an interference from colored samples. Blank of control, which was the native solvent, was performed to reduce the interference from each solvent. The inhibition of elastase was calculated using the following equation; Hyaluronidase inhibition (%) = [1-(B/A)] ×100; where A is a different absorbance between a sample solution with hyaluronidase, hyaluronic acid, and BSA and a sample solution without these reagents, whereas, B is a different absorbance between hyaluronidase solution and its native solvent. EGCG was used as a positive Chemical characterization of CW The chemical characterization of CW, which was the most potent anti-wrinkle extract, were performed using high performance liquid chromatography (HPLC) regarding to our previous study [6]. Cordycepin was used as a standard compound. The experiments were analyzed in triplicate.

Development of microemulsion
Pseudoternary phase diagram construction Pseudoternary phase diagrams were prepared by using water titration method [13]. The samples were de ned as microemulsions when they were transparent liquid with low viscosity. Various factors affecting microemulsion formation were investigated, including oil types, surfactant types, co-surfactant types, and ratio of surfactant to co-surfactant (Smix).
To investigate the effect of oil type, various types of oils, including jojoba oil, argan oil, apricot kernel oil, perilla oil, corn oil, canola oil, and sugar squalane were used as the oil phase in microemulsion system when other components are xed Tween® 85 was used as a surfactant, butylene glycol was use as a cosurfactant, and the Smix ratio was 1:1. To investigate the effect of surfactant type, various types of nonsurfactants, including Tween® 20, Tween® 80, Tween® 85, PEG 40-caster castor oil, Plantacare® 1200 UP, and Triton™ X-114 were used as the surfactant in microemulsion system when sugar squalane was used as an oil, butylene glycol was use as a co-surfactant, and the Smix ratio was 1:1. To investigate the effect of co-surfactant type, various co-surfactant types, including glycerin, propylene glycol, and butylene glycol were used as the co-surfactant in microemulsion system when sugar squalane was used as an oil, Tween® 85 was use as a surfactant, and the Smix ratio was1:1. To investigate the effect of surfactant to co-surfactant ratio, various surfactant to co-surfactant ratio, including 1:2, 2:1. 3:1, 4:1, and 5:1 were investigated when sugar squalane was used as an oil, Tween® 85 was used as a surfactant, and butylene glycol was used as a co-surfactant.
Microemulsions from the pseudoternary phase diagram which give the largest area of microemulsion region, were prepared.

Characterizations of microemulsions
Organoleptic inspections were used to assess the external presence of microemulsions. The zeta sizer (Zetasizer®, Malvern Instruments Ltd., Malvern, UK) was used to investigate the internal droplet size, polydispersity index (PDI), and zeta potential. A pH meter was used to measure the pH of the formulation.

Stability of microemulsions
The stability of microemulsions was investigated after 8 heating-cooling cycles in which the microemulsions were stored at 45°C for 24 h and then at 4°C for 24 h. The physical appearance, particle size, PDI, and zeta potential were then investigated as previously stated.
Preparation of microemulsion containing C. militaris extract C. militaris extract with the most potent anti-wrinkle activities was selected for incorporating into the best microemulsion system. Isotropic appearance, small internal droplet size, narrow size distribution, stable after storage, and appropriate concentration of surfactant used in the formulation were among the selection criteria for the best microemulsion. To incorporate the C. militaris extract, it was rst dissolved in DI water and then combined with the other ingredients until homogeneous. The microemulsions containing C. militaris extract were characterized and investigated for their stability in heating-cooling conditions as described above.
In vitro cytotoxicity effect determination on HaCaT cells Before the preparation of topical serum formulations for evaluation in human volunteers, the cytotoxicity effects of on HaCaT cells of microemulsion containing C. militaris extract were investigated in a comparison with their blank formulations. Additionally, aqueous solution of CW was also evaluated for the cytotoxicity effect. In brief, the HaCaT cells (Cell Lines Service, Eppelheim, Germany) were grown and the cell viability was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay regarding to our previous study [6].

Preparation of topical serum formulation from microemulsion containing C. militaris extract
The microemulsions containing C. militaris extract were incorporated into the serum formulation, the ingredients of which are mentioned in Table 1. The topical serum formulation from microemulsion containing C. militaris extract were prepared regarding good manufacturing practices (GMP), characterized and investigated for the stability in heating-cooling conditions as described above. Microbial enumeration test of non-sterile serum formulation Methods for microbial limit numbers and speci ed microorganisms in non-sterile cosmetic products were assessed in accordance with the procedure and criteria laid down in standard requirements of the United States Pharmacopeia 32-National Formulary 27 [14] and the European Pharmacopoeia (EP) [15]. Total viable aerobic microbial count (TAMC) and total yeasts and molds count (TYMC) were performed by plating serum sample on soybean casein digest agar (SCDA) and Sabouraud dextrose agar (SDA), respectively. The speci ed harmful pathogens, such as Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, Clostridium sp. and Candida albicans were determined dependent on the method validation under the guidelines. Culture plates were incubated and the results are displayed as colony forming units per mL of serum (CFU/mL).

In vivo irritation test in human using human patch test
Before the in vivo skin irritation study, the ethic for human research was approved by the Human research ethics committees, Faculty of Pharmacy, Chiang Mai University. The ethics approval number was 04/2563. The skin irritation was investigated in 30 healthy volunteers by using human patch test with some modi cations [16,17]. The samples were applied on the inner lower arm skin for 4 h. During the study, the skin irritation was observed at 15, 30, 60, 120, 180, and 240 min. After that the patch was taken off and the skin irritation was further assessed at 24, 48, and 72 h.

E ciency evaluation in human volunteers
Before the e ciency evaluation, the ethic for human research was approved by the Human research ethics committees, Faculty of Pharmacy, Chiang Mai University. The ethics approval number was 04/2563. The e cacy of microemulsion containing C. militaris extract were investigated in 30 healthy volunteers. Two products, including topical serum formulation from microemulsion containing C. militaris extract (CW serum) and blank serum formulation (Blank) were applied on the inner forearm skin of the left and right arm, respectively. The blank serum formulation was served as placebo. During the experimental period of 14 days, the volunteers can use all other products in normal daily life. Before and after applying the products on the skin twice daily for 14 days, the skin was evaluated for moisture and elasticity by using portable skin test equipment (M-6602, Shenzhen Fuhengtong Technology, Shenzhen, China). The exact area of the skin, where skin elasticity and moisturization were evaluated, was marked by using a plastic sheet and punching a hole with an area of 3×3 cm placed on the inner forearm area, setting the edge on the volunteers' wrist. Skin moisturization, which was recorded as a function of electrical conductance, and skin elasticity, which was recorded as a function of skin stretching, were recorded and the changes in these parameters were calculated using the equations; Skin elasticity change (%) = A/B×100, where A is a different skin elasticity between before and after application of the formulations and B is a skin elasticity before the formulation application. Additionally, Skin moisture change (%) = A/B×100, where A is a different skin moisture between before and after application of the formulations and B is a skin moisture before the formulation application.
Additionally, Likert-scale questionnaires were also used to evaluate the e ciency of the products in terms of moisturizing, elasticity, and winkle reducing effects.

Investigation of satisfaction in human volunteers
Before the satisfaction evaluation, the ethic for human research was approved by the Human research ethics committees, Faculty of Pharmacy, Chiang Mai University. The ethics approval number was 04/2563. The satisfaction on microemulsion containing C. militaris extract was investigated in 30 healthy volunteers after 14 days of application. Likert-scale questionnaires were used to evaluate the volunteer's satisfaction in terms of physical appearance (viscosity, color, and odor) and topical application (non-sticky, emollient, skin absorption, non-skin irritation, moisturizing, skin texture, and skin slippery).

Statistical analysis
Data was analyzed and reported as mean and standard division (S.D.). The statistical analysis was processed by t-test and ANOVA using SPSS program (SPSS Statistics 17.0, IBM Corporations, New York, USA). Statistically signi cant different was denoted when P < 0.05.
Results And Discussion C. militaris extracts The external appearance of each C. militaris extracts were different. CC, CA, and CE were brown-yellowish semisolid masses with a distinct odor on the outside, while CH was a lighter brown-yellowish semisolid mass. In contrast, CW was a light-brown dry powder.
Anti-wrinkle activities of C. militaris extracts Skin ageing is naturally occurred by a decline of the extracellular matrix (ECM), including collagen bers, elastin bers, and hyaluronan by the favor of MMP-1, elastase, and hyaluronidase, respectively [18].
These enzymes mostly presented in the dermis layer and played important roles in the degradation of ECM and resulting in the skin winkles [19]. Inhibitory activities against MMP-1, elastase, and hyaluronidase of C. militaris extracts are shown in Table 2. Among different C. militaris extracts, CW possessed the signi cantly highest MMP-1 and elastase inhibitory activities with the inhibition of 77.9 ± 5.3% and 84.4 ± 4.0%, respectively. Interestingly, the MMP-1 and elastase inhibition of CW was as potent as oleanolic acid and EGCG, a naturally occurring triterpenoid widely known for MMP-1 and elastase cascade [11]. Therefore, CW was suggested as a potent natural extract which exerted anti-wrinkle activities. On the other hand, cordycepin was noted as a major component of CW which responsible for MMP-1 and elastase inhibition since it exhibited potent MMP-1 and elastase inhibitory activities with the inhibition of 58.5 ± 4.9% and 76.5 ± 7.1%, respectively. Although, there are some previous reports revealed about the MMP-1 inhibitory activity of C. militaris, they suggested cordycepin as the key compound that inhibited the activator protein-1 and resulting in the suppression of MMP-1 expression [20,21]. However, the present study was the rst to remarked that both cordycepin and C. militaris extracts exerted a direct effect on the MMP-1 enzyme. Additionally, the present study was the rst to reveal anti-elastase of C. militaris extracts and suggested CW as the most potent anti-wrinkle ingredient for cosmetic formulations. However, the IC 50 value was preferable to comparing the strength of the inhibition of CW with the standard compound. Therefore, it was suggested for further study to investigate the IC 50 value of CW in a comparison with oleanolic acid and EGCG on the inhibition of collagenase and elastase, respectively.  In contrast to MMP-1 and elastase inhibition, CW possessed the least inhibitory activity against hyaluronidase. CA was noted as the most potent hyaluronidase inhibitor with the inhibition of 69.0 ± 3.9%. Interestingly, CA was as potent as EGCG which was well-known as anti-ageing compound. Although cordycepin possessed a potent inhibitory activity of 65.5 ± 2.1% against hyaluronidase, the cordycepin content was not related to the hyaluronidase inhibition. Therefore, cordycepin was not the only compound responsible for the hyaluronidase inhibitory activity. Previous studies suggested that C. militaris contained large quantity of polysaccharides [22,23], which has been reported for the hyaluronidase inhibition [24]. Since the retardation in hyaluronan degradation and increasing hyaluronan level in the skin layer by both putting hyaluronan in topical skin care products or using as a temporary dermal lling agent were suggested for plumping and youthful skin [25]. CA would be another natural extract potentially retard the skin ageing.
Although CA possessed the signi cantly highest hyaluronidase inhibition (p < 0.05), it had no effect on MMP-1 inhibition. Therefore, CW was selected for further formulations development since CW exerted the most potent MMP-1 and elastase inhibition (p < 0.05), as well as a moderate hyaluronidase inhibitory activity (14.3 ± 5.6%). The effective concentration of CW was suggested at 0.1 mg/ml because this concentration led CW to possess approximately 80% inhibitory activity against both MMP-1 and elastase. However, only some active compounds could release and deeply penetrate into the target site of the skin.
Therefore, approximately 100 times higher than the effective concentration, which was 10 mg/ml or 1% w/v, was suggested for further product development [6].
Chemical characteristic of CW CW, which was the most potent anti-wrinkle extract, was characterized for its chemical constituent. The HPLC chromatograms of cordycepin and CW are shown in Fig. 1. Cordycepin was detected at the retention time of 6.347 min, whereas, there was a major peak detected around 6.3 min in the HPLC chromatograms of CW. Therefore, it could be noted that cordycepin was a major chemical constituent of CW, which could be further used as a marker for the quantitative determination of CW in the further study.

Effect of oil type
Various oil types generated different microemulsion region in the pseudoternary phase diagrams as shown in Fig. S1. Sugar squalane gave the largest area of microemulsion (9.5%), followed by jojoba oil, argan oil, apricot kernel oil, canola oil, corn oil, avocado oil, and perilla oil, respectively. Since sugar squalane is a unique cosmetic ingredient which is a mobile, colorless, odorless, and has good physical and chemical stability, it was selected for the further studies. Additionally, squalene has been known as a natural triterpene hydrocarbon which is one of the most important lipids in human skin cell and accounted for up to 13% of total lipids synthesized from the sebaceous glands [26].

Effect of surfactant type
Pseudoternary phase diagrams constructed using various types of surfactant. Tween® 85 was the only one surfactant which could generate microemulsion with the region of 9.5% in the pseudoternary phase diagram. The results were well accordance with previous studies, which suggested Tween® 85 as the most suitable surfactant for microemulsion development [27,28]. Therefore, Tween® 85 was selected for the further studies.

Effect of co-surfactant type
The single surfactant is not enough for reducing interfacial tension between water and oil and generate microemulsion, of which the internal droplet size is in the range of 10-200 nm [29]. Therefore, cosurfactant is required in the development of microemulsion. Pseudoternary phase diagrams constructed using various types of co-surfactant are shown in Fig. S2. The results showed that propylene glycol gave the largest microemulsion area in pseudoternary phase diagram (11.1%), followed by butylene glycol. On the other hand, glycerin could not generate microemulsion. The likely explanation might be due to the presence of alkane triol which composed of three hydroxyl groups in the glycerin molecule and led to high hydrophilic property but lower e cacy in the interfacial tension reduction [30]. Therefore, propylene glycol was selected for the further studies.

Effect of surfactant to co-surfactant ratio
Various ratio of surfactant to co-surfactant showed different microemulsion region in the pseudoternary phase diagrams as shown in Fig. S3. The results represented that microemulsion region increased when the surfactant proportion in Smix increased from 1:2 to 2:1. Therefore, it could be concluded that there should be higher content of surfactant than co-surfactant in the Smix. However, microemulsion region tended to decrease after increasing of the surfactant from 1:2 to 5:1. The likely explanation might be due to the lower co-surfactant content which was not enough to reduce the interfacial tension. Therefore, Smix ratio of 2:1 was selected for further studies.

Microemulsion preparation
Since all parameters, including oil type, surfactant type, co-surfactant type, and Smix ratio affected the microemulsion region in the pseudoternary phase diagrams, the system that generated the largest region of microemulsion was selected for further microemulsion development. In brief, the selected system composed of sugar sqaulane, Tween® 85, propylene glycol, and DI water. The Smix ratio was 2:1. Four formulations (ME1 -ME4) along the line of 30% w/w water content as shown in Fig. 2 were developed and characterized to investigate the effects of different Smix and oil content.
All formulations (ME1 -ME4) were homogeneous transparent liquid with low viscosity which could be de ned as microemulsion [31]. Higher content of Smix led to darker yellow color and more viscosity. The internal droplet size of each microemulsions were ranging from 107.2 ± 2.9 nm to 366.3 ± 16.0 nm as shown in Fig. 3. Since all microemulsion contained 30% w/w of water phase, the increasing of Smix but decreasing of oil phase were used in ME1 to ME4, respectively. ME1, which contained the highest oil content (40% w/w), tended to be W/O microemulsion since the amount of oil phase was higher than the water phase. ME2, which contained an equal amount of oil and water phase (30% w/w), tended to be O/W microemulsion since Tween® 85 is hydrophilic surfactant with high HLB value (11.0). The results noted that ME2 had the smallest internal droplet size (107.2 ± 2.9 nm). However, the size of O/W microemulsion increased with the increasing amount of Smix although the oil content decreased. The likely explanation might be due to their higher viscosity since size of internal droplets has been reported to increase with an increase of the viscosity [32]. PDI of these microemulsions were ranging from 0.7 ± 0.1 to 1.0 ± 0.0, which were very high. The likely explanation might be due to the measurement of undiluted sample of which the internal droplets were not well dispersed and led to the detection of larger size than usual [33]. However, microemulsion could not be diluted before the size and PDI measurement because the microemulsion system would be changed. On the other hand, all microemulsion had the same pH value around 6.9 ± 0.1 to 7.1 ± 0.2. Additionally, all microemulsions were stable since no changing in external appearance, e.g., separation, sedimentation, or coagulation, were not detected after 8 cycles of heating-cooling condition. The internal droplet size was still remained the same. Therefore, ME1 was selected as a representative of W/O microemulsion and ME2 was selected as a representative of O/W microemulsion for further incorporation of C. militaris extract (CW).
Microemulsion containing C. militaris extract ME1 and ME2 were selected for the incorporation of CW. The external appearance of microemulsion containing C. militaris extract (ME1-CW and ME2-CW) were homogeneously transparent and brownyellow color with low viscosity. The internal droplet size of ME2-CW (146.1 ± 1.5 nm) was signi cantly smaller than that of ME1-CW (212.4 ± 1.5 nm) (p < 0.05). On the other hand, the PDI of ME2-CW (0.5 ± 0.0) was signi cantly narrower than that of ME1-CW (0.9 ± 0.1) (p < 0.05). Therefore, ME2-CW tended to be more stable than ME1-CW due to its narrower size distribution of the internal droplets. Interestingly, the internal droplet size of ME1-CW and ME2-CW were larger than their own blank microemulsions as shown Fig. 4. Hence, it could be assumed that CW was entrapped inside the droplet of microemulsion and led to larger internal droplet size after the C. militaris extract incorporation.
A basic analysis of the stability under heating-cooling conditions could be used as a fast tool to predict the stability of emulsions, such as cosmetic formulations [34,35]. After 8 cycles of heating-cooling condition, both ME1-CW and ME2-CW were physically stable since no changing in external appearance were detected. However, the internal droplet size of ME1-CW was signi cantly increased (p < 0.05). The results were well accordance with the previous PDI data noted that the internal droplet size distribution of ME1-CW was board. On the other hand, ME2-CW was stable since its internal droplet size was still remained unchanged. Although the internal droplet size of ME1-CW increased, the size was still in the nano scale (212.4 ± 1.5 nm). Therefore, ME2-CW were selected for further preparation of topical serum formulation.

Cytotoxicity effects on HaCaT cells of microemulsion containing C. militaris extract
The cytotoxicity effects on HaCaT cells of microemulsions with and without C. militaris extract (CW) are shown in Fig. 5. All formulations were found to be safe since the HaCaT cell viability was approximately higher than 80%. No signi cant difference was detected among these formulations. The HaCaT cells viable after being exposed to the aqueous solution of C. militaris water extract (CW), ME1, ME2, ME1-CW, and ME2-CW were 78. 5  Topical serum formulation from microemulsion containing C. militaris extract The external appearance of topical serum formulation from microemulsion containing C. militaris extract was homogeneously transparent yellow liquid. The viscosity was 0.8 ± 0.0 mPas. The formulation was stable since no changing in external appearance, e.g., separation, sedimentation, or coagulation, were not detected after 8 cycles of heating-cooling condition. Additionally, the viscosity was still remained the same. Therefore, it was a good formulation suggested for further applied in the human volunteers.

Microbial enumeration test of non-sterile serum formulation
Microbial contaminants may grow by using ingredients in the products causing physicochemical changes or spoilage and nally loss in product safety, quality, activity, and stability. In tested serum, the TAMC was ≤ 1 x 102 CFU/mL and the TYMC was not detected. The speci ed pathogens, E. coli, P. aeruginosa, S. aureus, Clostridium sp. and C. albicans were not detected in tested serum formulation. The tropical serum products were compliant with the accepted criteria according to guidelines of non-sterile and cosmetic products due to non-excessive microbial count and the absences of speci ed pathogens.
The compliant serum formulation was applied use in human volunteers.
Irritation properties of topical formulation from microemulsion containing C. militaris extract Thirty healthy human volunteers (77% of female and 23% of male), ranged in age from 21 to 60, were tested on the serum formulation. The results noted that the topical formulation from microemulsion containing C. militaris extract induced no irritation sign on human skin. Therefore, it was suggested as safe for using topically on the skin. E cacy of topical formulation from microemulsion containing C. militaris extract After two weeks of the application, the skin appearance improved as shown in Fig. 6. The results showed that the skin elasticity and skin moisture signi cantly improved after 1 week of serum products application and continuously increased after 2 weeks as shown Fig. 7. Interestingly, the topical formulation from microemulsion containing C. militaris extract could enhance the skin elasticity and skin moisture when compared to its own base formulations. The topical formulation from microemulsion containing C. militaris extract could improve the skin elasticity by 39.9 ± 7.3% from the baseline after 1 week and maintained at 38.5 ± 9.0% after 2 weeks of application. The potent inhibitory activities of CW on MMP-1 and elastase may be the likely explanation of the effective results on the skin since it could reduce the damage of collagen and elastin ber. Apart from the enhancement of skin elasticity, the skin moisture was also improved after the application of serum formulation. The topical formulation from microemulsion containing C. militaris extract could increase the skin moisture by 42.2 ± 14.2% from its baseline after 1 week and 38.3 ± 15.2% after 2 weeks of applications. The likely explanation might be due to the composition of various polysaccharide in the C. militaris extracts, which could reduce transdermal water loss (TEWL) and protect the skin barrier [36]. Although the skin moisturization measure in the present study was based on the electrical conductance, the reduction in TEWL would result in increased skin moisturization since water can be retained in the skin layer. Apart from C. militaris extracts, sugar squalane, which was used as an oil phase in the microemulsion, would be another component that would help to moisturize the skin. Sugar squalene is commonly used as an occlusive ingredient, so it could reduce TEWL and thereby in-creases the skin moisture. Therefore, the serum formulation not only effectively enhanced the skin moisture, but also improved the skin elasticity. Although the clinical studies regarding cosmetic formulations e cacy are usually performed for two or three months because the epidermal transit time is around one month, the skin improvement could be detected since the rst week of a serum formulation application. The likely explanations were not only from the inhibitory activity on extracellular matrix degradation in the dermis layer but also attributed to the enhancement of the skin moisturization.

Conclusions
CW was the most potent natural anti-wrinkle extract with potent inhibitory activities against MMP-1 (77.9 ± 5.3%) and elastase (84.4 ± 4.0%). Interestingly, CW exhibited comparable MMP-1 and elastase inhibition to that of oleanolic acid and EGCG. Cordycepin, which was a major chemical constituent of CW, was the compound responsible for its anti-skin ageing activities. Therefore, CW has a great potential to be used as active cosmetic ingredient for anti-wrinkle. CW was incorporated in the microemulsion and topical serum formulation. No cytotoxicity effect of the microemulsions was observed in HaCaT cells. The topical serum formulation from microemulsion containing C. militaris extract induced no skin irritation and signi cantly improved the skin moisture (42.2 ± 14.2%), as well as skin elasticity (39.9 ± 7.3%) from its baseline after 1 week of the applications. Therefore, this formulation was recommended for further use as anti-wrinkle cosmetic/cosmeceutical products.   Internal droplet size of microemulsions with and without C. militaris water extract (CW). Asterisk (*) denoted signi cant difference between the microemulsions with and without CW analyzed using t-test (p < 0.05).

Figure 5
The viability of HaCat cells after exposed to the aqueous solution of C. militaris water extract (CW), blank microemulsion (ME1 and ME2), and microemulsion containing C. militaris water extract (ME1-CW and ME2-CW).

Figure 6
Skin photographs before (a) and after (b) the application of blank serum formulation (Blank) for 14 days (volunteer age: 36 years). Skin photographs before (c) and after (d) the application of topical formulation Page 24/25 from microemulsion containing C. militaris extract (CW serum) for 14 days (volunteer age: 36 years). Skin photographs before (f) and after (f) the application of blank serum formulation (Blank) for 14 days (volunteer age: 60 years). Skin photographs before (g) and after (h) the application of topical formulation from microemulsion containing C. militaris extract (CW serum) for 14 days (volunteer age: 60 years). The photographs were taken using INSHERE 50x-1600x USB Microscope, Handheld Magni er with the magnitude of 1000x.

Figure 7
Skin elasticity change (a) and skin moisture change (b) after application of topical formulation from microemulsion containing C. militaris extract (CW serum) and blank serum (Blank). Asterisk (*) denoted signi cant difference between before and after application of the formulations, analyzed using t-test (p < 0.05).