Dried aerial parts of E. alba were collected from the Jecheon Medicinal Herb Association, Korea and authenticated by Dr Ki Hwan Bae (College of Pharmacy, Chungnam National University, Daejeon, Korea) where the voucher specimens were deposited. The sample was ground into powder and extracted 3 times with petroleum ether at 40°C for 4 h under reflux, then filtered and concentrated under a vacuum evaporator (Serial No. 41440910; EYELA, N-N Series, Rikakikai Co. Ltd. Tokyo, Japan) to yield the corresponding petroleum ether extract (PEE; yield 0.89% w/w). The resulting residue was then extracted 3 times with methanol at 70°C for 4 h then filtered and evaporated. The dried MeOH extract residue was suspended in distilled water and the resulting aqueous suspension was fractionated sequentially with the hexane fraction (HeF) and n-butanol fraction (BuF) at 1:1 (v/v) ratio 3 times at room temperature. The resulting 2 fractions and remaining water fraction (WaF) were evaporated under a vacuum (extraction yield, HeF 6.19%, BuF 1.12% and WaF 6.74% w/w).

Athymic male nude (nu/nu) mice of BALB/c origin at 7 weeks of age, were purchased from Dae-Han Biolink, Inc. (Eumseong, Korea). They were kept in autoclaved cages with filter bonnets in a laminar flow unit under 12 h light:dark periods at 24±2°C in a humidified atmosphere and were fed sterilized food and distilled water. The experiments were performed in the Animal Center of Chungnam National University under aseptic conditions in accordance with the NIH guidelines for the care and use of laboratory animals. The authorization code number is CNU-00244 (Chungnam National University).

The mice were divided in to 6 groups; 5 males were allocated to each of the 6 groups. The animals in group 1 received 0.4 ml of the vehicle mixture (propylene glycol:ethanol:dimethyl sulfoxide, 67:30:3% v/v) (Sigma, St. Louis, MO, USA), and the animals in group 2 received minoxidil 2% (Mino). The animals in groups 3, 4, 5 and 6 received a 5 mg sample of PEE, HeF, BuF and WaF of E. alba with the vehicle formulation. Treatment was performed by topical application once per day on the backs of nude mouse skin for 20 consecutive days.

The mice were evaluated for hair coverage area by a score of 0 to 8 as described in Table I. Hair scores were taken on day 0, 5, 7, 12, 16 and 20. To evaluate the change in hair density, digital images of each mouse were randomly acquired on experimental days 8 and 16 in the same region (3.6 mm²) of interscapular skin. The change in hair density was evaluated by analyzing the images (x200 magnification; actual area, 3.6 mm²) using Kong, Bom-Viewer Plus software (Bomtech Electronics Co., Ltd., Seoul, Korea).

Skin samples were fixed in 10% neutral buffered formalin for histological analysis. Paraffin-embedded 4 μm sections were stained with Mayer’s hematoxylin and eosin (H&E; Sigma). The morphology and structure of the HFs of the nude mice were evaluated microscopically in the H&E-stained sections of dorsal skin at a magnification of x1,000. Five fields per section (magnification, x100) were used for counting the number of dermal and subcutaneous HFs with respect to the total number of HFs. Histological processing and digital photomicrographs were acquired using the Leica application suite, version 4.0.0 (Leica Microsystems Ltd., Heerbrugg, Switzerland).

Keratinocyte proliferation was evaluated by an intraperitoneal injection of BrdU (50 mg/kg body weight; Sigma) 1 h before the mice were sacrificed. Dorsal skin from both the treated and control animals was collected on experimental day 16 and fixed with 4% paraformaldehyde, dehydrated and embedded in paraffin. After sectioning, the sections were dewaxed and denatured in 1.5 mol/l of HCl for 30 min and neutralized with phosphate-buffered saline for 1 h. BrdU incorporation was detected by immunohistochemical (IHC) staining of the paraffin-embedded sections with mouse anti-BrdU primary antibody (1:200; Cat. no. SC-32323) in a moist chamber, at room temperature for 3 h (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). After washing 3 times, the sections were incubated with secondary antibodies [Biotinylated secondary antibody (Life Technologies, Carlsbad, CA, USA)] for 15 min. The assessment of follicular and epidermal keratinocyte and sebaceous gland epithelial cell BrdU labeling was performed by an observer blinded to the treatments using the original magnification of x400.

Western blot analysis was performed to evaluate the expression level of TGF-β1 during follicular morphogenesis. The skin samples were homogenized and lysed in protein extraction buffer (Pro-Prep; Intron Biotechnology, Inc., Seongnam, Korea). The protein concentration was measured by Bradford assay (Bio-Rad, Boston, MA, USA); equal amounts of protein (20 mg/sample) were separated by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; Mini Protean II; Bio-Rad) and then transferred onto a PVDF membrane (Immobilon). The membrane was blocked for 12 h at 4°C with 5% skimmed milk (Uppsala, Sweden) in 1X TBS (10 Mm Tris pH 7.5, 100 mM NaCl and 0.5% Tween-20). Immunodetection was performed by incubation at appropriate dilutions (1:500, TGF-β1 polyclonal antibody) at 4°C overnight then incubated with the secondary antibodies goat anti-rabbit IgG-HRP (both from Santa Cruz Biotechnology, Inc.) conjugate for 1 h at room temperature. After washing, the blots were detected by ECL western blotting detection reagents (Santa Cruz Biotechnology, Inc.). The band intensities were quantified using NIH ImageJ software.

The experimental data are expressed as the means ± standard deviation (SD). The Student’s t-test or one-way ANOVA were used for the assessment of significance between the different treatment groups. Statistical analysis was performed using SAS 9.2 software. A value of p<0.05 was considered to indicate a statistically significant difference.

The changes occurring in hair growth patterns in athymic nude mice were documented after the initiation of topical application for 20 consecutive days. The effects of PEE of E. alba were evident as early as 7 days following the commencement of treatment, and became obvious on the dorsal body surface. Hair growth on the PEE-treated mice first became evidient on the heads then consistently extended to the caudal region of the tail on experimental day 16 (Fig. 1B). On the other hand, the control mice exhibited relatively sparse and shorter hair growth that was roughly distributed on different regions of the body on day 10; this phenomena noticeably corresponds to the ‘wave like pattern’ of the nude phenotype (Fig. 1A) (19). Eventually, the control mice became nearly complete ‘nude’ on experimental day 16 (Fig. 1A and C), while PEE induced a distinctly smoother, thicker hair shaft, leading to dense and marked hair coverage on days 8 and 16 (Fig. 1B and D).

The stimulatory effects of PEE of E. alba on the hair coverage area of the nude mice were evaluated as they received specific concentrations of PEE, HeF, BuF, WaF vs. the vehicle and or 2% minoxidil. The effects of PEE on the hair coverage area of the mice in all treatment groups were precisely estimated for each mouse by giving them a score from 0 to 8 (Table I). The maximum hair growth score was significantly (p<0.001) increased in the mice treated with PEE of E. alba than in the mice in the other groups from day 8 and consistently covered the maximum area of the body on day 16 (Fig. 1E). By contrast, in the nude mice, rapid hair loss was observed in the minoxidil and other treatment groups, which then returned to baseline levels. In terms of hair density, on days 8 and 16 following the initiation of treatment, the mice in the PEE-treated group exhibited a significant increase (p<0.0001) in hair density compared to the other groups (Fig. 1G). Although, minoxidil had a significant effect (p<0.001) on sustaining hair density on day 8, progressive hair loss decreased and hair density thus also decreased on day 16. This is a characteristic pattern observed in the hair growth patterns of athymic nude mice.

The skin specimens of the control mice exhibited numerous dystrophic HFs, that twisted and coiled within the follicular infundibulum (Fig. 2C) and failed to penetrate the epidermis. However, the hair shafts that penetrated the epidermis were heavily twisted and frequently fractured before achieving a substantial length (Fig. 2A), as is normally evident in nude mouse skin. This striking characteristic of nude mice indicates that the hair fibers suffer from abnormal keratinization (20). On the other hand, the skins of the PEE-treated nude mice had relatively normal follicles containing well differentiated straight hair shafts which continued through the follicular ostia to the skin’s surface (Fig. 2B and D). In nude mice, the structure of the HFs, inner root sheath and hair shaft exhibit abnormalities (8), the most striking of which is the cuticle of the HF which is either discontinuous or, more often, totally absent (Fig. 2E). By contrast, the HFs of the PEE-treated mice were regularly formed and intact, coated by a clearly discernible hair cuticle (Fig. 2F). Furthermore, the control mice exhibited abortive HFs which revealed histological signs of the late catagen stage (Fig. 2A and G). On the other hand, treatment with PEE of E. alba resulted in a histological pattern identical to that observed in the late anagen phase of cycling hair and induced a marked increase in the number of HFs compared with the control mice (Fig. 2H).

To determine the difference in the follicular keratinocyte proliferation rate, we labeled the proliferating cells with BrdU in vivo on day 16 and subsequently stained them with an antibody to BrdU. Our results revealed that in the PEE-treated mice, the number of BrdU-labeled keratinocytes per anagen follicle increased significantly, particularly in the follicular matrix and outer root sheath compared to the control mice (Fig. 3A and B). The mean number of BrdU-positive cells per anagen follicle was 31.2±2 in the PEE-treated mice as compared with 14.9±1.7 in the control mice. This increase was statistically significant (p<0.001). Moreover, the PEE-treated nude mice also exhibited a significant increase in the number of BrdU-labeled proliferating epidermal keratinocytes (p<0.001) and BrdU-positive epithelial cells per sebaceous gland (p<0.01) (Fig. 3D, F and G).

The expression level of TGF-β1 was quantified by western blot analysis at different time point (on days 8 and 16) to determine the comparative expression of TGF-β1 in the vehicle- and PEE-treated mouse skins. Quantitative analysis of the western blots revealed a significant (p<0.001) decrease in the expression levels of TGF-β1 in the PEE-treated mice during early anagen (day 8) and during the late anagen or anagen-catagen transition (day 16) compared with the control mice (Fig. 4). Quantitative analysis indicated that the delayed hair regression in the PEE-treated mouse skins was associated with an altered expression of TGF-β1.