Evaluation of efficacy of Silybum marianum flower extract on the mitigating hair loss in vitro and in vivo

Natural components that can exert a wide range of anti‐hair loss activity with fewer side effects are in high demand. The objective of this study was to investigate the anti‐hair loss potential of Silybum marianum flower extract (SMFE) in vitro and in vivo.

androgenetic alopecia accounts for about 80% of all hair loss symptoms.Genetic factors and inadequate regulation of androgen contribute to a high concentration of dihydrotestosterone (DHT).
DHT can cause hair loss by binding to androgen receptors of DPCs in scalp.7][8] In women, female pattern hair loss (FPHL) is observed with a symptom of diffusing hair thinning.Hair thinning in women is commonly associated with menopause and other age-related changes in skin. 9cently, prevention of hair loss has been approached through various strategies, including activating dermal papilla cells (DPCs), reducing dihydrotestosterone (DHT) production, attenuating oxidative stress, inhibiting cellular senescence, and improving blood flow. 8,10,11noxidil and finasteride are representative hair loss treatments approved by the FDA.Minoxidil is an arteriovenous dilator that can modulate potassium channels and activate the β-catenin pathway to increase DPCs proliferation and VEGF mRNA expression. 12,13Finasteride is known to inhibit the conversion of testosterone into DHT by inhibiting 5α-reductase activity. 14Although both drugs are used in hair loss treatment, there is a problem in that hair loss can occur again if treatment is stopped.In addition, they can cause various side effects including sexual dysfunction, gynecomastia, and dermatitis. 15,16Therefore, there is a need to develop a hair loss treatment agent with minimal side effects using natural products.
Alleviation of hair loss requires a multimodal approach.Plant extracts containing multiple components such as polyphenols, flavonoids, saponin, and fatty acids have the potential to exert a broad range of benefits for hair follicle health.Silybum marianum L. (S. marianum) is a traditional medicinal plant that has been used for liver disease in Europe since about 2000 years ago.Substances found mainly in the seeds of S. marianum are collectively called silymarin, which includes isosilybin, silychristin, and silydianin, among which silibinin, called silybin, is known to be the main component accounting for about 50% of silymarin. 17According to a recent study, silibinin can induce hair growth by activating the Wnt/β-catenin signaling pathway and promoting the proliferation of dermal papilla cells. 18Since silymarin with such physiological activity is contained in a large amount in the seed of S. marianum, the remaining flower after harvesting the seed is discarded.In the cosmetic industry, for the development of sustainable eco-friendly products, there is a growing demand for upcycling raw materials with new value by adding utility to discarded by-products.Therefore, we have investigated the biological activity of Silybum marianum flower extract (SMFE) except seed on skin.In a previous study, the senolytic effect of SMFE in fibroblasts and apigenin was detected as a main peak while silymarin was not detected. 19Apigenin is also a bioflavonoid compound contained in S. marianum and has been found to be present in S. marianum flowers as well. 20 has been reported that apigenin can promote the proliferation of DPCs by suppressing TGFβ. 21TGF-β1 is known to induce catagen phase which leads to hair follicle regression.In a preliminary test, we confirmed that apigenin concentrated fraction of SMFE (apigenin 2%) showed higher activity on DPCs proliferation and senescence prevention than SMFE with 0.23% apigenin.Hence, this study was performed to elucidate the anti-hair loss effect of SMFE with 2% apigenin on DPCs and clinical models.

| High-performance liquid chromatography (HPLC)
The HPLC system used in this study was a Waters 2695 (Milford, MA, USA) equipped with a Waters 2996 Photodiode Array (PDA) Detector.The Empower 2 software was used to control the analytical system and perform the data collection and processing.HPLC-PDA was performed on a Phenomenex Luna C18 (4.6 × 150 mm, 5 μm) column reversed-phase column protected by a C18 guard column from Phenomenex, Inc. (Torrance, CA, USA).The sample injection volume was 10 μL.The signal was monitored at 335 nm.
The elution system used for the HPLC-PDA assay was a binary high-pressure gradient elution system with mobile phase A (0.1% TFA in H 2 O) and mobile phase B (acetonitrile).Elution gradients were: 35% organic phase B, hold for 10 min; from 35 to 100% organic phase B in 13 min (linear gradient), hold for 5 min; then back to the starting condition in 2 min and re-equilibration for 10 min.
The flow rate was 1.0 mL min −1 .Each analysis required 30 min, including the re-equilibration time.

| Cell culture
Human follicle dermal papilla cells (HFDPCs) were purchased from Promocell (Heidelberg, Germany) and cultured in a follicle dermal papilla cell growth medium mixed with a growth medium supplement mix (Promocell, Heidelberg, Germany) at 37°C in a 5% CO 2 incubator.Replicative senescent HFDPCs were constructed by longterm passaging over 14 passages.To confirm the effect of SMFE on oxidative stress-induced senescence of HFDPCs, HFDPCs were pretreated with 200 μM hydrogen peroxide (H 2 O 2 , Sigma-Aldrich) for 2 h and cultured with SMFE for 72 h.

| Quantitative detection of VEGF secretion
Secreted vascular endothelial growth factor (VEGF) was measured by VEGF ELISA (R&D System, Detroit, MI, USA).Cells were incubated with SMFE for 72 h and the conditioned medium was transferred to an ELISA plate.The sample was incubated at room temperature (RT) for 2 h and then mixed with VEGF conjugate for 2 h.The substrate was added and absorbance was measured at 450 nm with a spectrophotometer (Epoch, Bio-Tek Inc.).

| Quantitative real-time PCR
Total RNA was purified using a Trizol reagent (Thermo Fisher Scientific).The cells treated with Trizol were collected in a tube and then chloroform (Sigma-Aldrich) was added to separate the upper layer.After mixing the upper layer with 100% isopropyl alcohol (DAEJUNG, Siheung, Korea), the supernatant was removed and RNA was washed twice with 70% EtOH.Purified total RNA (1 μg) was synthesized into cDNA using amfiRivert cDNA Synthesis Plantinum Master Mix (GenDEPOT, Katy, TX, USA) through TaKaRa Thermal Cycler Dice Touch (TP350, TAKARA BIO INC., Shiga, Japan).
Gene expression was detected using 7500 Real-Time PCR System (Thermo Fisher Scientific, MA, USA) and AMPIGENE® qPCR Green Mix Hi-ROX (Enzo Life Sciences, Formingdale, NY, USA), and primers for PCR used were ordered from Bioneer Corp. (Daejeon, Korea).
Sequences are shown in Table 1.
The culture was incubated overnight after adding a staining solution: SAβ-gal in senescent cells stains blue.The proportion of senescent cells was determined as the number of senescent cells staining blue divided by the total number of cells.Average numbers of stained cells and total cells were determined for 10 fields (magnification: ×100).

| Quantitative detection of IL-6 secretion
Interleukin-6 (IL-6) secretion was detected by IL-6 ELISA (R&D System).After a conditioned medium was incubated in an ELISA plate at room temperature for 2 h, IL-6 conjugate was added and reacted for 2 h.The sample was reacted to substrate and absorbance was detected at 450 nm with a spectrophotometer.

| Clinical trial
To evaluate the efficacy and safety of alleviation of alopecia (hair loss) symptoms of SMFE, a randomized, double-blinded, placebo-controlled clinical study was performed.Written consent was obtained from all subjects before this study in accordance with guidelines for cosmetics used for the alleviation of alopecia (hair loss) symptoms.Forty-two subjects (male and female patients aged 18-54 years, mean age of 46.096 ± 6.60 years) who were diagnosed with androgenetic alopecia were chosen for this study.Those who had undergone surgical treatment for hair loss, such as hair transplantation or scalp reduction, and those who had taken dutasteride or finasteride orally within the last 6 months were excluded from this study.Shampoo containing 0.05% SMFE (0.001% apigenin) and placebo shampoo were applied on scalp and hair area over a period of 24 weeks (once a day).

| Evaluation method
Primary efficacy evaluation was performed through hair density measurement using Folliscope 5.0 (LeadM Corporation, Seoul, Korea).Secondary efficacy evaluation was conducted through visual evaluation of researchers and a subject survey.

Hair density measurement (Phototrichogram)
In order to measure hair density, the area to be evaluated was first designated and the hair was sheared to a certain area (about 1 cm 2 ).
After labeling with a hair dye (tattoo solution), the surface of the scalp was photographed.The scalp was taken with a Folliscope 5.0 at the designated test area before application and at 8 weeks, 16 weeks, and 24 weeks after application of the product.The total number of hairs in an area of 1 cm 2 of the image was analyzed and used as a hair density result.

Hair photography and researchers' visual evaluations
Subjects were photographed for the test area at an angle of 45° (the line of bangs) and 90° (top of the head) before and at 8 weeks, were conducted to determine whether there were differences between groups.

| Analysis of SMFE
The components identified in SMFE were flavonoids, and the dominant component was apigenin, which contained 2%.Chromatograms of SMFE and standards are shown in Figure 1.Peak identification was achieved by comparing the retention times with UV spectra obtained for individual standards.

| Effects of SMFE on proliferation and VEGF secretion of HFDPC
DPCs are major cells in hair follicles (HF).They can interact with hair stem cells.They are involved in hair growth by regulating the hair cycle. 27To confirm whether SMFE could regulate the proliferation of DPCs, HFDPCs were treated with SMFE at indicated concentrations for 72 h and an MTT assay was performed.Finasteride was used as a positive control for DPCs proliferation.
Finasteride can increase β-catenin and Akt, known as major proliferation signaling factors of DPC, and is also used as a positive control in DPC proliferation experiments. 28,29As a result, it was found that HFDPC proliferation was increased to 22% by 50 μg/ mL SMFE and 34% by 100 μg/mL SMFE compared with the control group (Figure 2A).
VEGF, a growth factor highly expressed in DPCs, is known to affect hair growth by regulating angiogenesis. 12,30In order to figure out whether SMFE can control VEGF production, the level of VEGF secreted in the medium was measured using ELISA.SMFE increased VEGF production to 100% at 100 μg/mL compared with the negative control (Figure 2B).2][33] We observed that SMFE decreased TGF-β1 and increased IGF-1 and FGF-7 gene expression levels (Figure 2C).These results suggest that SMFE can activate hair growth by increasing DPC proliferation and regulating several growth factors related to hair cycles.

| Effect of SMFE on senescent HFDPCs
Hair follicle (HF) aging shows main features of hair loss, anagen shortening, and miniaturization, which is associated with DPC senescence, which is known to affect HF inductivity due to reduced aggregation capacity. 34Bahta et al. have reported that DPC undergo premature senescence in vitro through high expression of p16INK4a/pRB, which contributes to loss of proliferative capacity. 35To observe the effect of SMFE on senescent DPC, we used a replicative senescent HFDPC through long-term passaging over 14 passages.In a preliminary study, senescence phenotype was confirmed after 14 passages.As shown in Figure 3A

| Effect of SMFE on oxidative stress-induced HFDPC senescence
Oxidative stress is one of the major factors that induce premature senescence of DPCs and hair loss. 41A previous study has reported that oxidative stress with a high concentration of O 2 can induce senescence of DPCs by increasing ROS levels and SAβ-gal activity. 8reover, lipid peroxide product as an oxidative stress biomarker contributes to early onset of catagen through induction of apoptosis of hair follicle cells. 42To determine whether SMFE could inhibit oxidative stress-induced HFDPC senescence, β-gal, and ROS generation assays were performed.EGCG was used as a positive control. 43 shown in Figure 4A, β

| Homogeneity test
The homogeneity of the "placebo product"group and the "test product" application group was checked.As a result of comparing the measured value of hair density before using the product, no significant difference between the two test groups was found (p = 0.977 > 0.05).
Therefore, it was confirmed that these two groups were homogeneous (Table 2).

| Measurement of hair density in the placebo product application group
Hair density of the placebo product application group was decreased to −2.136 ± 4.612 each/cm 2 (p = 0.124 > 0.05) after 16 weeks of application of the product and −0.364 ± 3.458 each/cm 2 (p = 1.000 > 0.05) after 24 weeks of application of the product compared with that before application (Table 3, Figure 5).However, such decreases were not statistically significant.

| Measurement of hair density in the test product application group
Hair density in the test product application group was statistically significantly increased to 3.700 ± 5.440 each/cm 2 (p = 0.020 < 0.05) after 16 weeks of application of the product and 5.450 ± 5.306 each/ cm 2 (p = 0.001 < 0.05) after 24 weeks of application of the product compared with that before application (Figure 5A, Table 4).As a result of comparing the change in hair density between the placebo product application group and test product application group, there were signification differences after 16 weeks (p = 0.001 < 0.05) and Table 4).In the hair density measurement, 95.45% of the placebo product application group showed no significant difference in the number of hair after 24 weeks, and 55% of the test product application group showed a significant above-average increase in the number of hair.

| Researchers' visual evaluations (ICC)
Visual evaluation was expressed as a score that evaluated the degree of hair distribution by two researchers (Dermatologists), and the re-

TA B L E 2
Measurement results of hair density before application of the product in the "placebo product" group and "test product" group.0.75-0.94][25][26] In the evaluation of the investigators in Table 5, researcher 2 seemed to consistently give a higher score than researcher 1, but as a result of analyzing the degree of agreement between the two researchers' visual evaluations, it was found that the degree of agreement between the two researchers at all times was 0.886 (Table 5).

| Researcher's visual evaluation for the placebo product application group
Visual evaluation scores of the two researchers for the placebo product application group were significantly decreased to −0.591 ± 0.590 (p = 0.001 < 0.017) after 8 weeks of application of the product, −0.864 ± 0.834 (p = 0.001 < 0.017) after 16 weeks of application of the product, and −1.182 ± 0.588 (p = 0.000 < 0.017) after 24 weeks of application of the product compared with those before application (Figure 6, Table 6).

| Researchers' visual evaluation for the test product application group
Visual evaluation scores of researchers for the test product application group were increased to 0.250 ± 0.639 (p = 0.096 > 0.017) after 8 weeks of application of the product compared with those before application (Figure 6, Table 7).However, the difference was not statistically significant.Visual evaluation scores of researchers for the test product application group were significantly increased to 0.900 ± 0.788 (p = 0.001 < 0.017) after 16 weeks and 0.950 ± 0.510 (p = 0.000 < 0.017) after 24 weeks of application of the product compared with those before application (Figure 6, Table 7).
TA B L E 3 Measurements of hair density in the "placebo product" application group.

TA B L E 4
Measurement of hair density in the "test product" application group.
As a result of comparing the visual evaluation scores of researchers for the "control product" group and the test product group, there were significant differences after 8 weeks (p = 0.001 < 0.05), 16 weeks (p = 0.000 < 0.05), and 24 weeks (p = 0.000 < 0.05) of application of the product (Table 7).As a result of confirming the improvement of the hair condition in researchers' visual evaluation, after 24 weeks, 59.09% of the placebo product application group showed 1-2 grade deterioration, 31.82%showed no grade change, and 4.54% showed 1-2 grade improvement.In comparison, after 24 weeks, 65.0% of the test product application group showed 1-2 grade improvement, and 35.0%showed no grade change.
All of these clinical results show that the number of hairs, the primary efficacy evaluation, and the expert visual evaluation score, the secondary efficacy evaluation, met the endpoint in the hair density change of the test product group compared with the placebo product group at week 24.

| DISCUSS ION
DPCs in the hair follicle bulb play a critical role in activating hair growth.It has been reported that the phenomenon of thinning and loss of hair occurs when the number of DPCs is reduced to less than a certain amount. 44 prolongs anagen, and inhibits the expression of TGF-b1 by regulating hair follicle cell proliferation. 32In addition, FGF-7 is a factor expressed in telogen and is involved in starting a new hair cycle by stimulating hair germ and activating MAPK signals. 33In this study, SMFE increased the proliferation of DPCs and regulated several factors related to hair follicle development.These results suggest that SMFE can promote hair growth through the activation of DPCs proliferation, perifollicular angiogenesis, and hair cycle regulation.
Various factors such as DHT, oxidative stress, and the neighboring environment of hair follicles can induce DPC senescence, which leads to hair loss through loss of proliferative capacity and inflammatory condition.It has been observed that DPCs from a balding scalp grow slower in vitro than DPCs from a nonbalding scalp due to DPC senescence. 46Cellular senescence of balding DPCs is mediated by p16INK4a and pRb pathways known to be induced by environmental stress. 35,47Oxidative stress, a well-known mediator of cellular senescence, is also involved in the induction of DPC senescence.It has been found that DHT, the main cause of androgen alopecia, can induce DPC senescence through the accumulation of mtROS.Phytochemicals with radical scavenging activity such as cyanidin 3-O-arabinoside can prevent DHT-induced senescence. 48ese findings indicate that reduction of DHT-induced DPC senescence is a critical target for androgen alopecia.In addition, direct effect of oxidative stress such as H 2 O 2 on DPCs appears to suppress hair growth through the downregulation of β-catenin and induction of cytotoxicity. 49In this study, we observed that SMFE decreased the level of SAβ-galactosidase in replicative and H 2 O 2 -induced senescence models.In addition, IL-6 secretion induced by senescence was inhibited by SMFE, indicating that SMFE could promote hair growth through inhibition of DPC senescence.A previous study has reported that the senescence pathway of replicative senescence through long-term passage is usually associated with the activation of p53 and p21Cip1 and that environmental stress-induced senescence is associated with the p16INK4a pathway. 50,51To elucidate the mechanism of SMFE on senescent DPC, further study needs to be performed to confirm the involvement of cell cycle-related signals such as p53, p21Cip1, and p16INK4a.
Recently, the relationship between dermal hair follicle environment and hair loss has been investigated.Cao et al. have reported that a young host dermal environment can rejuvenate transplanted aged follicles whereas an aged dermal environment could not induce the growth of hair shafts from transplanted young follicles. 52The SASP produced by senescent cells is a well-known factor that can exert a detrimental effect on anagen hair follicles.IL-6, a major inflammatory cytokine of SASP, can inhibit clonal keratinocyte growth in vitro and block telogen-to-anagen transition in vivo. 34In a previous study, we have confirmed that SMFE possesses an anti-aging activity through inhibition of IL-6 production in senescent fibroblasts.

TA B L E 7
Visual evaluation results for the "test product" application group.
16 weeks, and 24 weeks after application of the product under the same conditions.Based on the judgment criteria for the extent of hair condition accompanying it, two dermatologists made visual evaluations.If there was a difference in the evaluation between the two researchers, a low level was chosen.Visual evaluation grades according to the degree of hair distribution are shown as follows: very good, +3; good +2; a little better, +1; no change, 0; very bad, −3; bad, −2; and a little worse, −1.To analyze the consistency of visual evaluation of the two experts, the intraclass correlation coefficient (ICC) was analyzed.[23][24][25][26]2.11 | Statistical analysisAll in vitro data are expressed as mean ± standard deviation (SD).Differences between control and treatment groups were evaluated by Student's t-test using Statview software (Abacus Concepts, Piscataway, NJ, USA).A p < 0.05 was considered statistically significant.IBM SPSS Statistics 23.0 statistical analysis program (International Business Machine Corporation [IBM], Armonk, USA) was used to determine the significance of the change in hair measurement before and after product use and in the area of test product/control product use.Statistical analysis was performed after excluding the data of the dropout when a dropout occurred during the test, and the statistical significance was confirmed when the significance probability was p < 0.05 in the 95% confidence interval.Repeated measures ANOVA (or Friedman test) was conducted to determine whether there were differences according to measurement conditions within groups, and independent sample t tests (or Mann-Whitney U-tests) , replicative senescent HFDPC showed an increased level of SAβ-galactosidase staining compared with young HFDPC (at passages 2-4).Increased activation of SAβ-gal is a well-known indicator of cellular senescence.When senescent HFDPCs were treated with SMFE for 72 h, SMFE decreased the level of SAβ-gal staining in a dose-dependent manner compared with nontreated control.Cellular senescence is associated with low proliferative capacity.36Thus, we observed the effect of SMFE on the proliferation of senescent HFDPCs.SMFE increased the proliferation of senescent HFDPCs by 31% at 100 μg/ mL compared with nontreated control (Figure3B).It is widely known that senescence-associated secretory phenotype (SASP) production is increased with cellular aging which contributes to local and systemic low-grade inflammation that leads to degenerative phenotypes and impaired physical function.37,38IL-6 is a major inflammatory cytokine of SASP.It is increased in bald scalp.39,40In addition, senescent DPCs secrete higher amounts of IL-6 than nonsenescent DPCs, which reduces the entry into the anagen phase and inhibits shaft elongation.Based on these findings, we observed whether SMFE affected levels of IL-6 in senescent HFDPCs.As shown in Figure 3C, SMFE at 50 μg/mL and 100 μg/mL decreased levels of IL-6 production by 35% and 49%, respectively, to nontreated control.F I G U R E 1 HPLC chromatogram of Silybum marianum flower extract.These results suggest that SMFE can alleviate hair loss caused by hair follicle aging by suppressing DPC senescence and IL-6 secretion in senescent DPCs.

F I G U R E 2
(A) HFDPCs were cultured with indicated concentration of SMFE for 72 h.Proliferation was measured by MTT assay.(B) Secreted VEGF level was measured by ELISA.(C) The level of TGF-β1, IGF-1, and FGF-7 mRNA expressions were determined by qPCR.Data are presented as mean ± standard deviation.*p < 0.01 compared with SMFE nontreated control (n = 3).These results were obtained from three independent experiments.SMFE: Silybum marianum flower extract; FS: finasteride (20 μg/mL).

F I G U R E 3
Effects of SMFE on senescence and IL-6 secretion in replicative senescent HFDPCs.HFDPCs were pretreated with 20 μM EGCG for 24 h and cultured with SMFE for 72 h.(A) Cells were stained with β-gal.(B) Cell proliferation was measured by MTT assay.(C) Secreted IL-6 was measured by ELISA.Data are presented as mean ± standard deviation.*p < 0.01 compared with young cells, **p < 0.01 compared with vehicle-treated control (n = 3).These results were obtained from three independent experiments.YC; young cells; SMFE: Silybum marianum flower extract; EGCG: epigallocatechin gallate.
liability of visual evaluation between researcher 1 and researcher 2 was analyzed by ICC.ICC is a value between 0 and 1, where less than 0.4 indicates poor reliability, 0.5-0.75indicates medium reliability, F I G U R E 4 Effects of SMFE on senescence and IL-6 secretion induced by oxidative stress in HFDPCs.HFDPCs were pretreated with 200 μM H 2 O 2 for 2 h and cultured with SMFE for 72 h.(A) Cells were stained with β-gal.(C) Secreted IL-6 was measured by ELISA.(B) HFDPCs were cultured with SMFE for 72 h and 100 μM DCFDA was pretreated for 30 min.Then cells were incubated with 200 μM H 2 O 2 for 30 min, and the ROS level was measured by detecting fluorescence.Data are presented as mean ± standard deviation.*p < 0.01 compared with YC, **p < 0.01 compared with vehicle-treated control group (n = 3).These results were obtained from three independent experiments.YC; young cells; SMFE: Silybum marianum flower extract, EGCG; FS: finasteride.
Variation a = �∑ a k=0 Measured value of weeks after application (k) − Measured value of before application (k) � /a.Probability b (p value) **p < 0.05 repeated measures ANOVA post-hoc Bonferroni correction.F I G U R E 5 Comparing results of hair density between the "placebo product" application group and "test product" application group.(A) Absolute hair density at baseline/week 16/week 24 and (B) the changes in hair density.**p < 0.05, #p < 0.05.
Hence, activation of DPCs proliferation become one of the main strategies for hair loss treatment.VEGF can promote hair growth through regulation of perifollicular vascularization during hair cycle.Suppression of VEGF-mediated angiogenesis can impair hair growth. 45TGF-β1 is a factor expressed in dermal papilla, and its levels increase in late anagen and remain high during catagen, which hinders the development of hair follicles by inducing apoptosis of keratinocyte. 31IGF-1 increases the number of hair follicles, TA B L E 5 Researchers' visual evaluations.

F I G U R E 6
Visual evaluation results of the "test product" application group compared with those of the "placebo product" application group.(A) Visual evaluation grade and (B) the changes in evaluation grade.(C) Images of change in hair density following the application of scalp shampoo for 24 consecutive weeks.†p < 0.017, ‡p < 0.05.
TA B L E 6 Researchers' visual evaluation results for the "placebo product" application group.�∑ a k=0 Measured value of n weeks after application (k) − Measured value of before application (k) � /a.Compared within group b †p < 0.017(5%/3) by Friedman test, post hoc Wilcoxon signed rank test.Compared between group c : Comparison of "placebo product" group and "test product" application group, ‡p < 0.05 by Mann-Whitney U test.