Phenyl Propanoid 1’s-1′-Acetoxychavicol Acetate Protects Against Testosterone-Induced Alopecia by Inhibiting Nadph Oxidase

1. Introduction

Hair growth is composed of four phases: anagen (growth), catagen (regression), telogen (rest), and exogen (shedding).[1] Each hair follicle undergoes an independent cycle with ten to thirty cycles in its lifetime. Hair follicles are composed of cell types, including dermal papilla (DP), bulge, matrix, and root sheath cells. DP cells derived from mesenchymal stem cells are located at the bases of hair follicles and regulate hair growth by interacting with hair follicle epithelial cells. Matrix cells can differentiate into various cell lineages of the hair shaft and the inner root sheath, whereas the outer root sheath (ORS) is derived from epithelial progenitor cells.[2,3] Testosterone can regulate the cycle of hair follicles by stimulating the anagen-to-catagen transition, promoting hair loss. Typical characteristics of catagen and telogen phases are apoptosis of ORS and regression of DP, and ORS cell death is mediated by testosterone-mediated ROS generation.

Androgenetic alopecia (AGA), also known as hair loss, commonly affects both men and women and is regulated by several factors, including hereditary traits, testosterone-related action, hair follicle miniaturization, and shortened anagen phase.[4] Among these factors, the genomic action of testosterone within the DP is a well-known pathological mechanism of AGA. To be specific, type Ⅱ 5α-reductase converts testosterone into 5α-dihydrotestosterone (DHT), and DHT can bind to cytoplasmic androgen receptor (AR) to pass through the nucleus. The DHT-AR complex can stimulate the expression of various factors such as vascular endothelial growth factor (VEGF), transforming growth factor β1 (TGF-β1), and Dickkopf-related protein 1 (DKK-1).[5,6,7] Testosterone-dependent regulation of these factors are involved in premature anagen termination and early catagen induction in AGA.

In contrast to the genomic pathway of testosterone, several lines of evidence have indicated that G protein-coupled receptor family C group 6 member A (GPRC6A) can serve as a receptor for testosterone on the plasma membrane. GPRC6A has been identified as a receptor for L-amino acids, calcium, and osteocalcin.[8,9] We have previously reported that testosterone can stimulate GPRC6A and activate a cascade of Gq-PLCβ-IP₃-Ca²⁺ in epidermal keratinocytes.[10] On the other hand, nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (Nox) is a multicomponent enzyme complex that produces reactive oxygen species (ROS) and seven homologues (Nox1, Nox2, Nox3, Nox4, Nox5, Duox1, and Duox2) have been identified. Duox1 is a predominant isozyme in keratinocytes and testosterone-induced intracellular calcium mobilization through GPRC6A can directly activate Duox1, resulting in the generation of H₂O₂ and keratinocyte apoptosis. These explanations suggest a molecular mechanism between non-genomic testosterone action and cellular redox status related to AGA.

Based on these findings, we screened a library of natural compounds and identified 1'S-1'-acetoxychavicol acetate (ACA, 26) as having Nox inhibitory activity. The present study found that ACA could suppress testosterone-mediated ROS generation and cell death. Moreover, the addition of ACA into human hair follicle organ culture ameliorated testosterone-mediated hair growth inhibition. We provide a molecular mechanism by which ACA may be effective for treating AGA through Nox inhibition.

2. Results

2.1. Identification of ACA as an Inhibitor of Nox from a Library of Natural Compounds

Several lines of evidence indicate that testosterone-dependent activation of the GPRC6A cascade results in activation of Duox and intracellular ROS generation. These events can cause apoptosis of ORS cells and primary epidermal keratinocytes, leading to androgenic alopecia. Based on the hypothesis that testosterone-dependent apoptosis of ORS cells is mediated by Nox-induced ROS generation, a Nox inhibitor screening assay was performed by recording luminescence intensity derived from lucigenin oxidation. Mouse kidney membranes expressing Nox1, Nox2, Nox4, or Duox1 isozymes were incubated with natural compounds (1-38, Figure 1A), and diphenyleneiodonium (DPI) and Ewha-89403 (89403) were used as controls. ROS generation was then measured by lucigenin chemiluminescence (Figure 1B). Among the tested compounds, ACA (26) isolated from Alpinia galanga showed potent activity, with 60% inhibition of ROS generation at 10 μM.

ACA was first identified from A. galanga and A. conchigera, typically used as traditional spices in cooking and medicines. The structural characteristics of ACA are shown in Figure 2A. To investigate the inhibitory activity of ACA against each Nox isozyme, purified Drosophila membranes expressing human Nox1, Nox2, Nox4, and Duox1 were used to monitor ROS production by determining the luminescence of lucigenin in the absence or presence of ACA. The IC₅₀ values of ACA were 32.7 μM for hNox1, 42.3 μM for hNox2, 31.3 μM for hNox4, and 30.8 μM for hDuox1 membrane (Figure 2B). Therefore, ACA showed pan-Nox inhibitory activities against Nox1, Nox2, Nox4, and Duox1 isozymes. Dose-inhibition (%) curves of ACA against Nox1, Nox2, Nox4, and Duox1 are shown in Figure 2C.

2.2. Molecular Modeling of ACA with Nox5

As the structures of Nox enzymes are very similar, the crystal structure of Nox5 originating from Cylindrospermum stagnale (csNox5) was used as a template and the structure of hNox5-DH was modeled with the Swiss-Model server. The dehydrogenase domain (DH) of csNox5 consisted of flavin adenine nucleotide (FAD)-binding and NADPH-binding lobes.[11] The modeled hNox5-DH exhibited two distinct cavities: Cavity 1 and Cavity 2 (Figure 3A). Cavity 1 is located between two lobes and is likely to serve as the binding site for the nicotinamide moiety of NADPH, stacking against the isoalloxazine moiety of FAD.[11] The positively charged Cavity 2 is postulated to accommodate the phosphate group of NADPH. As shown in Figure 3B, the in silico docking model revealed that (S)-ACA lacked appropriate conformations for Cavity 1, although the phenyl ring in (S)-ACA was anticipated to stack against the isoalloxazine moiety of FAD. When (S)-ACA was superposed onto the in silico docking result of (R)-ACA, whose phenyl ring stacked against the isoalloxazine moiety of FAD, the acetoxyallyl group of the S form sterically clashed with FAD due to their different stereochemistries. Instead, it occupied Cavity 2 with the lowest free energy level (-8.2 kcal/mol), as demonstrated in Figure 3C. According to the in silico docking model of (S)-ACA in Cavity 2, the methyl moiety of the acetoxyallyl group interacted with the indole ring of Trp648 and the aliphatic chain of Arg509. Moreover, the carbonyl oxygen of the acetoxyallyl group formed two hydrogen bonds, one with the guanidino group of Arg509 and the other with the hydroxyl group of Ser456. The carbonyl oxygen of the acetoxy group engaged in hydrogen bonding interactions with the guanidino group of Arg641 and the hydroxyl group of Thr677. Additionally, the phenyl ring of (S)-ACA is likely to establish an H-π interaction with the aliphatic chain of Arg641. In conclusion, these findings suggest that (S)-ACA exhibits inhibitory activity against hNox5 by disrupting the binding of the phosphate group of NADPH within Cavity 2 of hNox5-DH.

2.3. ACA Suppresses Testosterone-Mediated ROS Generation in Skin Keratinocytes

To evaluate whether ACA could suppress testosterone-mediated ROS generation in the cell line, we measured intracellular levels of ROS in immortalized human keratinocytes (Ker-CT) by measuring the fluorescence of DCF-DA with a confocal microscope (Figure 4A). Stimulation of Ker-CT cells with testosterone increased ROS generation. Pre-incubation with ACA at different doses inhibited the generation of ROS in Ker-CT cells in a dose-dependent manner in response to testosterone. Based on the inhibitory pattern of ACA on ROS generation, the EC₅₀ value of ACA was 10.2 nM.

2.4. ACA Reduces Testosterone-Induced Apoptosis in Skin Keratinocytes

Previously, we reported that testosterone-dependent H₂O₂ generation is involved in the apoptosis of keratinocytes. To investigate whether ACA could inhibit testosterone-induced apoptosis, we used a TUNEL assay in hTERT-immortalized keratinocyte (Ker-CT) cells. Stimulation of Ker-CT cells with testosterone significantly increased TUNEL labeling compared to unstimulated cells, indicating that testosterone-dependent ROS generation induced apoptosis. Pretreatment of Ker-CT cells with the indicated concentration of ACA significantly suppressed testosterone-induced apoptosis (Figure 4B). The pattern of ACA inhibitory activity in testosterone-induced keratinocyte apoptosis was dose-dependent and the EC₅₀ value of ACA was 5.8 Nm (Figure 4C).

2.5. Function of ACA in Human Hair Follicle Organ Culture

To investigate the function of ACA in human hair follicle organ culture, hair follicles from male non-balding occipital scalps were cultured with testosterone in the absence or presence of ACA for 6 days (Figure 5A). Testosterone-treated human hair follicles showed suppression of hair growth. However, treatment with ACA mitigated this testosterone-dependent growth suppression (Figure 5B).

2.6. ACA Regulates Testosterone-Induced Hair Loss

To validate its therapeutic effect on hair loss, ACA was applied to the dorsal skin of a testosterone-induced hair loss mouse. The telogen phase was observed in the dorsal skin of C57BL/6 male mice ranging from 7 weeks old to 12 weeks old. C57BL/6 male mice (8 weeks old) were administered testosterone for 3 days after removing dorsal hair. Two concentrations of ACA (50 µM and 100 µM) were administered to the dorsal skin of each testosterone-treated mouse for 12 days. As shown in Figure 6A, hair follicles were observed to grow in the ACA-treated group from 6-8 days, changing the skin from pink to gray. Finally, the skin turned black (indicating hair regrowth) on day 12. The scheme of hair growth measurement from an androgenetic alopecia mouse model is shown in Figure 6B. Later, skin tissues were collected and stained with H&E to quantify the degree of hair growth (Figure 6C). Hair follicle lengths showed more significant increases in ACA-treated mice (50 and 100μM of ACA) than control mice (Figure 6D). These results indicated that ACA has good therapeutic activity in the androgenetic alopecia mouse model.

3. Discussion

It is well established that type Ⅱ 5α-reductase plays an important role in miniaturization of hair follicles and male hair loss, and thus inhibitors of type Ⅱ 5α-reductase including dutasteride and finasteride have been developed for conventional treatment of AGA.[12] However, long-term treatment has various adverse effects such as impotence, decreased libido, erectile dysfunction, testicular pain, and ejaculation disorders. Therefore, consumers suffering from hair loss desire novel target therapy without these side effects.

Several lines of evidence indicate that natural products can be applied to prevent and treat hair loss. Extracts of Camellia sinensis, Panax ginseng, and Hibiscus rosa-sinensis have been reported to stimulate hair growth in animal models.[13] In addition, decursin, a coumarin from Angelica gigas, showed a hair-growth promoting effect in male C57/BL6 mice by decreasing inflammatory cytokines, such as tumor necrosis factor α (TNF-α) and interleukin (IL)-1β, and stimulating anti-inflammatory cytokines including IL-4 and IL-13.[14] The widely known compounds in P. ginseng, ginsenosides Rb1 and Rg3, improved the testosterone-mediated growth suppression of hair matrix keratinocytes in hair follicles.[15] As natural products possess complexity from diverse constituents in the extracts, isolation of active compounds and elucidating activities of each constituent are required to find a novel mode-of-action in developing a therapy for hair loss.

Nox isozymes are one of the significant sources for ROS generation and expression levels of Nox isozymes, and their regulators are associated with various diseases, including diabetic nephropathy, non-alcoholic steatohepatitis, and hair loss.[16,17] In pathological hair loss, Nox isozymes are overexpressed in hair follicle cells, resulting in uncontrolled ROS generation. Therefore, uncontrolled Nox activation should be suppressed by treatment with Nox inhibitor as an emerging therapy for hair loss. On the other hand, levels of cellular antioxidants in the body are also associated with hair loss. Activation of the nuclear factor (erythroid-derived 2)-like 2 (Nrf2) pathway can regulate the expression of antioxidant enzymes, including glutathione reductase, peroxiredoxin 1, heme oxygenase-1, and NADH dehydrogenase.[18] Sulforaphane (SFN) from broccoli (Brassica oleracea var. italica) has been shown to have antioxidant activity through activation of Nrf2 and reduce anagen-to-catagen transition.[18,19] SFN can significantly ameliorate ROS-dependent apoptosis of matrix keratinocytes and reverse ROS-induced reduction in hair matrix proliferation. These results indicate that Nrf2 activators such as SFN might be another therapeutic target for reducing damage caused by excess cellular ROS levels.

Interaction of testosterone with GPRC6A can stimulate intracellular calcium mobilization, leading to generation of H₂O₂ through Duox1 in skin keratinocytes (Figure 7). Our results show that ACA inhibits such testosterone-induced ROS generation and apoptosis of keratinocytes. In human follicle organ culture, ACA improved testosterone-dependent growth suppression of hair length. Topical application of ACA on dorsal skins of testosterone-stimulated C57BL/6 suppressed testosterone-induced hair loss. These collective findings concerning the mode-of-action and proof-in-concept indicate that ACA is a good therapeutic candidate for male hair loss treatment.

4. Materials and Methods

4.1. Materials

Thirty-eight pure compounds were isolated from plant materials by one of the authors, Prof. Eun Kyoung Seo, at the Natural Product Chemistry Lab, College of Pharmacy, Ewha Womans University. ACA (26) isolated from A. galanga (L.) Will. (Zingiberaceae) was confirmed to be identical to a commercial compound (ab142417) obtained from Abcam (Cambridge, UK). Testosterone (17β-hydroxy-3-oxo-4-androstene; T1500) was purchased from Sigma-Aldrich (Milwaukee, USA) via Sam Eung Industrial Company (Seoul, Korea). β-Nicotinamide adenine dinucleotide phosphate hydrate (NADPH; N5755) was purchased from Sigma-Aldrich. 2',7'–Dichlorofluorescin diacetate (DCF-DA; D-399) was purchased from Molecular Probes.

4.2. Nox Inhibitor Screening

Nox1, 2, 4, or Duox1 inhibitory activities were determined in a novel assay involving natural compounds. To specifically inhibit Nox isozymes, transgenic Drosophila Duox knockdown lines were established that over-expressed human Nox1, 2, 4, or Duox1.[20] The genotypes used were human Nox1 (UAS-hNOX1/UAS-DUOX-RNAi; Da-GAL4/+), Nox2 (UAS-hNOX2/UAS-DUOX-RNAi; Da-GAL4/+), Nox4 (UAS-hNOX4/UAS-DUOX-RNAi; Da-GAL4/+), and Duox1 (UAS-hDUOX1/UAS-DUOX-RNAi; Da-GAL4/+). Transgenic flies were homogenated with ice-cold PBS containing protease inhibitors and membrane-enriched human Nox1, 2, 4, or Duox1 were harvested. Membranes containing human Nox1, 2, 4, or Duox1 served to monitor ROS production by lucigenin chemiluminescence in the absence or presence of natural compounds.[21] The reaction medium consisted in HEPES-buffered salt solution (145 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO₄, 1.0 mM KH₂PO₄, 1.75 µM CaCl₂, 0.03 Na₂EDTA, 5.5 mM glucose, and 10 mM HEPES, pH 7.4), 400 µM lucigenin (10,10-dimethyl-bis-9,9-bisacridinium nitrate) and 500 µM NADPH.

4.3. In Silico Model of the Dehydrogenase Domain of hNox5 in Complex with ACA

Using the crystal structure of Nox5 originating from Cylindrospermum stagnale (csNOX5) as a template, the hNox5-DH structure was modeled with the Swiss-Model server.[22] The hNox5/FAD complex was modeled by substituting csNox5 with hNox in the crystal structure of the csNox5/FAD complex (PDB code: 5O0X). The structure of the ACA was prepared using ChemDraw 3D software (Perkin Elmer Informatics, Waltham, USA). Molecular docking of ACA against the hNox5-DH was performed using AutoDock Vina software (Forte, S. AutoDock Tools (version 1.5.6 rc2), Molecular Graphics Laboratory, Scripps Research Institute (http://mgltools.scripps.edu)).[23]

4.4. Cell Culture

Ker-CT cells (RL-4048) were purchased from ATCC^® (Manassas, USA) and cultured in EpiLife^TM medium (MEPI500CA, Gibco, Waltham, USA) supplemented with EDGS (EpiLife^TM Defined Growth Supplement; S0125, Gibco) and 1% penicillin-streptomycin solutions (LS202-02, Welgene, Gyeongsan, Korea).

4.5. Measurement of Intracellular ROS by DCF-DA

After cells were starved overnight in serum-free medium, they were stimulated by the inducer, testosterone, with or without ACA. Cells were washed with Hanks’ balanced salt solution (HBSS) and incubated with HBSS containing 10 µM 2',7'-DCF-DA for 10 min. The fluorescence was measured using an LSM510 confocal microscope at an excitation wavelength of 488 nm and an emission wavelength of 515-540 nm. Five groups of each type of cell were randomly selected, and the mean relative fluorescence intensity was measured with a Carl Zeiss vision system (LSM510, version 2.3). All experiments were repeated at least three times.

4.6. TdT-UDP Nick end Labeling (TUNEL) Assay

Apoptotic cells were detected by the TUNEL technique using an In Situ Cell Death Detection Kit, Fluorescein (11684795910, Roche, Basel, Switzerland). Cells were incubated with testosterone with or without ACA for 24 h. They were fixed with 3.5% paraformaldehyde for 1 h at room temperature and permeabilized with pre-chilled 0.5% triton X-100 in PBS for 10 min at room temperature. These cells were then stained with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min and mounted with mounting solution (M1289, Sigma-Aldrich). Fluorescence was measured by confocal microscopy. Samples were randomly detected over four points, and the percentages of apoptotic cells were determined by counting numbers of positively stained cells using Image J software.

4.7. Function of ACA in Testosterone-Treated Human Hair Follicle Organ Culture

Non-balding scalp specimens were obtained from non-balding occipital scalps of male patients undergoing hair transplantation surgery for androgenetic alopecia at Kyungpook National University Hospital. To isolate hair follicles, subcutaneous fat portions were dissected from the epidermis and dermis. Isolated hair follicles were cultured in Williams E media (12551-032, Gibco) supplemented with 2 mM L-glutamine, 100 U/ml streptomycin, and 10 ng/ml hydrocortisone without insulin. All processes were performed following methods reported previously,[24] with minor modifications. Hair follicles were incubated with or without testosterone and ACA for 6 days.

4.8. ACA Efficacy Test in Testosterone-Treated Dorsal Skin of C57BL/6J Male Mice

To estimate the effects of ACA on androgenic alopecia, C57BL/6J male mice were anesthetized with isoflurane gas and their dorsal hair was shaved. Hairs were entirely removed with appropriate amounts of hair removal cream. The next day, scar-less mice were collected and treated with testosterone (200 µg/100 µl) for 3 days. Then, ACA (50 µM and 100 µM) and the same volume of DMSO were diluted with 70% ethanol and administered to each group of mice with testosterone. The treatment was repeated once per day for 12 days. Mice were then sacrificed and their dorsal skins were collected. Tissues were frozen in frozen section media (3801480, Leica) at -80℃. Sections were prepared at a thickness of 10-12 µm.

4.9. Statistical Analyses

All values are presented as mean ± SD or SEM. Significant differences between groups were determined using two-tailed Student’s t-test. P-values less than 0.05 were considered statistically significant.