Reprint requests: Dr. M. J. Thornton, Medical Biosciences, School of Life Sciences, University of Bradford, Bradford, West Yorkshire BD7 1DP, UK. Tel: 44 0 1274 235517; Fax: 44 0 1274 309742; Email: M.J.Thornton@bradford.ac.uk
Improved wound healing of hairy skin may involve mesenchymal hair follicle cells with stem cell potential and enhancement by estrogen therapy. How estrogen affects follicular dermal papilla (DP) and dermal sheath (DS) cells in wound healing is unknown. Therefore, a comparison of estradiol action on DP, DS, and corresponding interfollicular dermal fibroblasts (DF) in a scratch-wound assay was performed using matching primary cultures established from female temporo-occipital scalp. All three cell types expressed mRNA transcripts and protein for estrogen receptors α (ERα) and β (ERβ). DF ERα transcripts were half that of DP and one-third of DS cells, while DF ERβ transcripts were two-thirds of DP and DS cells. In the scratch-wound assay all three cells types migrated at similar rates, but only the rate of DF was enhanced by estradiol. Mechanical wounding increased DNA synthesis rates of all three cell types and increased the secretion of collagen by DF and DS cells. All three secreted similar basal levels of vascular endothelial growth factor (VEGF), which was increased by wounding DF and DS cells, but not DP cells. DP cells required estradiol to increase VEGF secretion; by contrast VEGF secretion was decreased by estradiol in wounded DS cells. These results highlight differences in the responses of DF, DP, and DS cells to estradiol in a scratch-wound assay, providing further support for the dichotomy of cellular functions in the hair follicle. Further understanding of the role of estrogen in cutaneous wound healing may have important implications for the management of chronic wounds and scarring.
Skin appendages play a key role in epithelial regeneration following cutaneous wounding. Specifically, the regenerative properties of the adult hair follicle suggest that hair follicle stem cells may play an important role in the wound healing process. While epidermal stem cells of the hair follicle allow for re-epithelialization,1 the hair follicle mesenchyme, namely the dermal papilla (DP) and dermal sheath (DS), may play a part in dermal repair following cutaneous wounding.2
The importance of the hair follicle in cutaneous wound healing is provided by animal models. For example, full-thickness wounds heal more quickly in hairy animals than in humans, with greater contraction and less scar formation. Furthermore, a relationship between wound healing and the hair cycle has been demonstrated, with skin containing anagen (growing) follicles healing more rapidly. In humans, it has also been suggested that cutaneous wounds heal better in hair-bearing skin. When used for split-thickness skin grafting, rapid donor site healing has been reported in scalp skin with low infection rates and minimal scar formation.3 Furthermore, in a prospective study it was shown that following split-thickness skin graft harvesting, donor sites on hair-bearing scalp skin heal significantly better than those on nonhairy thigh skin.4
Evidence for the stem cell potential of both DP and DS cells comes from a study demonstrating that cells derived from both mouse and rat have hematopoietic stem cell activity in vitro. Furthermore, the same study was able to show that both DP and DS cells contribute to the regeneration of the entire hematopoietic system of lethally irradiated mice in vivo.5
Follicular DP cells have regenerative and inductive properties and have an important role in directing epithelial differentiation. Early studies showed that if the DP of individual rat whisker follicles are injured using a microsurgical needle, they heal without scar formation and normal hair growth resumes.6 DP cells secrete vascular endothelial growth factor (VEGF);7 a growth factor which plays a key role in angiogenesis. Although this occurs as part of the normal hair cycle, the release of VEGF by DP cells may also be important in wound healing.
The cells of the follicular DS, which is continuous with the DP at its base, also have a progenitor or stem cell role.2 Their regenerative potential has been shown by experiments where following amputation of the lower rat vibrissae hair follicle, incorporation of DS cells can regenerate the DP with subsequent hair growth.8 Similar results are seen when terminal human hair follicles are implanted into athymic nude mice following bulb amputation.9 Furthermore, DS tissue transplanted from one human to another induced hair growth and was not subject to rejection normally associated with allografts.10 These studies not only provide evidence of a mesenchymal lineage transition from DS to DP, but suggest that the DS displays a certain degree of immunoprivilege.
Evidence for the role of the DS in wound healing is suggested due to the high expression of α smooth muscle actin (α-SMA) in rat and human DS cells, both in vivo and in vitro.11,12 In vitro, the greatest expression is seen in DS cells derived from the lower sheath, which is considerably greater than corresponding DF. The expression of α-SMA is characteristic of the wound healing myofibroblasts found in granulation tissue, which are thought to play a key role in wound contraction.13
More recently, Gharzi et al.14 demonstrated using labeled cells and fluorescence microscopy that rat DF and vibrissa DS cells from the lower and upper part of the follicle were not only incorporated into healing ear wounds, but also migrated into the adjacent dermis.
Estrogen is known to play a role in skin homeostasis as shown by the changes seen in postmenopausal women, which include thinning of the skin and reduced collagen content.15 In addition, both human and animal studies have shown a reduced rate of wound healing associated with estrogen deficiency.16,17 Estrogen replacement therapy reverses the above effects with more rapid re-epithelization, increased collagen deposition and increased wound strength. Furthermore, topical estrogen applied before wounding has been shown to improve healing rates in elderly males and females.18 Studies to date appear to confirm that estrogen plays a key role in wound healing and may affect the inflammatory process, collagen deposition and proteolysis.
Estrogen can mediate its action via two nuclear receptors estrogen receptor (ER)—alpha (ERα) or beta (ERβ), or via membrane receptors.15 Human embryonic stem cells express both ERα and ERβ, and treatment with estradiol induces the expression of genes involved in differentiation, suggesting that estrogens can control differentiation of human embryonic stem cells into various cell types.19 Furthermore, estrogen significantly improves the osteogenic and adipogenic differentiation of bone marrow mesenchymal stem cells in vitro.20 A recent study using ER knockout models has shown that estrogen mediates the activation and tissue incorporation of bone marrow-derived endothelial progenitor cells following myocardial infarction.21 In addition, ERα was shown to be important in up regulating the secretion of VEGF.21
Because the DP and DS cells of the hair follicle may have a stem cell role in wound healing and also express estrogen receptors,22 the aim of this study was to establish the characteristics of corresponding DF, DP, and DS cells derived from the same patients, and using an in vitro wound healing assay, to determine whether 17β-estradiol significantly modulates important cellular responses including cell migration and proliferation, and collagen and VEGF secretion.
MATERIALS AND METHODS
Human scalp skin samples were obtained from healthy individuals following face-lift surgery (seven female patients, age range 49–60 years, median 50 years) with full consent and ethical approval and conformed to the guidelines contained within the Declaration of Helsinki Principles. To establish primary cultures of DF, the skin samples were dissected through the dermis, divided into 3 by 3 mm pieces, transferred to a 25 cm2 tissue-culture flask (Corning, New York, NY) and cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco, Paisley, Scotland, UK) supplemented with glutamine (2 mmol/mL), penicillin (100 U/mL), streptomycin (100 μg/mL), and amphotericin B (2.5 μg/mL), with 20% fetal calf serum (FCS) (Sigma-Aldrich, Poole, Dorset, UK) at 37 °C in 5% CO2 in air. To establish primary cultures of DP and DS cells (see Figure 1), terminal hair follicles were individually dissected as previously described,23 individual DP and the lower portion of the DS were transferred to separate 35-mm dishes (Falcon, Bectan, France) and cultured in DMEM with 20% human serum (Blood Transfusion Service, Sheffield, UK). Cultures were established from all seven of the DF explants, from five of the DP explants and from four of the DS explants, giving three matched cell lines that were derived from the same patient.
Once explant cultures were established, the medium was supplemented with 10% FCS. All cells were assayed at passage four to six. They were seeded in 35-mm dishes at a density of 1 × 105 per well in 2 mL growth medium and grown to confluence. Cells were then washed three times with phosphate-buffered saline (PBS) and the medium was replaced with serum-free, phenol red-free DMEM (Gibco). In assays where cells were cultured for >48 hours, 5% charcoal-stripped FCS (Sigma-Aldrich) was added to the media. Parallel dishes were incubated in triplicate containing either 10 nM 17β-estradiol or with vehicle control (absolute ethanol 0.0001%).
Reverse-transcriptase polymerase chain reaction (RT-PCR) for ERα, ERβ, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH)
Cells were seeded into six-well plates at a density of 1 × 105 per well in 2 mL normal growth media and grown for 2 days. Subconfluent cultures were washed with PBS (2 ×) and incubated in serum-free, phenol red-free DMEM for 24 hours before harvesting. Total cellular RNA was prepared using the Genelute mammalian total RNA kit (Sigma-Aldrich) according to the manufacturer's instructions, with the eluted RNA immediately treated with RNase-free DNase I (Promega Corp., Southampton, UK) for 1 hour before being purified by phenol:chloroform extraction and elution into PCR-grade distilled water. The concentration, purity and integrity of the resulting RNA was determined spectrophotometrically before the presence of mRNA for ERα, ERβ, and GAPDH was determined using end-point RT-PCR.
Reverse transcription and end-point PCR
First strand synthesis was performed on 250 ng of total cellular RNA using AMV-RT (Promega Corp.) according to the manufacturer's instructions in the presence of 25 Units of RNase inhibitor (RNasin; Promega Corp.) 0.4 mM dNTP mixture and 2.8 μM anchored dT primer (Sigma-Aldrich) for 1 hour at 42 °C, followed by 2 min at 95 °C to denature the reverse transcriptase enzyme. Amplification of GAPDH and ERα was performed in a Genius thermal cycler (Techne Corp., Duxford, UK) using 1 μL of cDNA, whereas ERβ was amplified using 4 μL cDNA. Reactions included 5 μL of 10 × AJ Buffer (450 mM Tris-HCl (pH8.8), 110 mM (NH4)2SO4, 45 mM MgCl2, 2 mM of each dNTP, 1.1 mg/mL acetylated BSA (Roche Diagnostics Ltd., Lewes, UK; 110 mM β-mercaptoethanol; 4.4 μM EDTA), 10 pmol of GAPDH-, ERα-, and ERβ-specific primers combined with 1.0 Unit of GO-Taq polymerase (Promega Corp.) diluted in 1 × AJ Buffer. The thermal cycler conditions for GAPDH were an initial denaturation step at 95 °C for 2 minutes, followed by 35 cycles of 95 °C for 45 seconds, 60 °C for 1 minute, 72 °C for 1 minute with a final extension time of 10 minutes at 72 °C. The absence of genomic DNA was confirmed using samples where the AMV-RT enzyme had been omitted. Optimal cycle conditions for ERα were 94 °C, for 30 seconds, 60 °C for 30 seconds, 68 °C for 1 minute, for an initial 10 cycles followed by another 25 cycles with increased extension times of 5 seconds/cycle followed by a final extension for 5 minutes at 68 °C. The amplification conditions for ERβ were the same as those for ERα with the exception that 40 total cycles were required. The amplification profiles for each gene were all within the linear range of detection (data not shown). After PCR, the reaction mixtures were separated through 3% TAE-agarose gels impregnated with ethidium bromide at ∼5 V/cm for 1 hour for UV visualization and images captured on a Gene Genius gel documentation and analysis system (Syngene, Cambridge, UK). Deoxyribonucleotide primers were obtained from Sigma-Genosys, (Pampisford, Cambridgeshire, UK) and purified by gel electrophoresis prior to use. The mRNA specific primer sets are described in Table 1. The expected sizes for the amplicons were 381 bp for ERα, 279 bp for ERβ and 347 bp for GAPDH. The level of the GAPDH house-keeping gene was used to evaluate any variation in the mRNA content and cDNA synthesis in the different preparations. No PCR products were detected when the reverse transcriptase was omitted (Figure 2). The relative levels of ERα and ERβ transcripts in the three cell types were quantified using quantitative real-time RT-PCR.
Table 1. Primer design and expected amplicons sizes
Amplification of ERα ERβ and GAPDH cDNA was performed in a Roche Lightcycler thermal cycler (Roche Diagnostics Ltd., Lewes, UK) using 1 μL of cDNA for ERα and GAPDH, and 4 μL of cDNA for ERβ with the Lightcycler FastStart DNA master SYBR Green I kit (Roche Diagnostics). Mastermix reactions were prepared as per the manufacturer's instructions on ice with 10 pmol of each primer pair (Table 1) and the cDNA added last. Controls included a water blank, minus RT control for each cell type, a series of diluted cloned human ERα, ERβ, (pCMV5-hERα24 and pcDNA3.1ERβ respectively25) and GAPDH (BioChain Institute Inc., Hayward, CA) cDNA targets adjusted so that a standard curve from 1 to 10,000,000 pmol of cDNA could be constructed. Additionally, a series of diluted DF, DP, and DS cell cDNA (1/10–1/10,000 dilution) were used to calculate amplification efficiencies for each cell type and each gene amplified. The cycle conditions were 95 °C for 10 minutes, followed by 45 cycles of 95 °C for 10 seconds, 60 °C for 5 seconds, 72 °C for 15 seconds for GAPDH, 95 °C for 12 seconds, 60 °C for 12 seconds, 72 °C for 24 seconds for ERα and EPβ. The crossing points were then used to measure the transcript concentrations from the standard curves and the predicted melting temperatures of the amplified products used to confirm that only the target DNA was amplified.
Expression of ERα and ERβ by immunocytochemistry
DS, DF, and DP cells were seeded into Lab-tek® 8 well chamber slides (Nalge Nunc International, Rochester, NY) at a density of 10,000 cells/well and allowed to attach for 24–48 hours. Cells were gently washed twice with PBS and fixed in ice-cold methanol (BDH Chemical Ltd., Dorset, UK) for 10 minutes at 4 °C. Cells were rehydrated gently with 200 μL of PBS and incubated at room temperature (RT) (25 °C) for 5 minutes, before blocking in 10% goat serum (secondary antibody host serum) for 90 minutes at RT. Primary antibodies at their predetermined optimal dilutions [ERα (HC-20) rabbit polyclonal antibody (1 : 20 dilution), sc-543, Santa Cruz Biotechnology Inc. (Santa Cruz, CA) and ERβ (H-150) rabbit polyclonal antibody (1 : 10 dilution), sc-8974, Santa Cruz Biotechnology] were prepared in 1% goat serum and centrifuged at 12,000 g for 3 minutes. The cells were washed in PBS for 5 minutes and 200 μL of the primary antibody was added into the appropriate wells and incubated at RT for 90 minutes. The wells were washed three times with PBS, each wash lasting 5 minutes. The secondary antibody (biotinylated goat anti-rabbit/anti-mouse IgG ready-to-use [Dako, Glostrup, Denmark]) was added into the wells and incubated at RT for 30 minutes. The washing procedure was repeated and then two to three drops of strepavidin peroxidase (AEC Dako) was added into each well before incubating at RT for 25 minutes. The washing procedure was repeated and 200 μL of distilled water was added to each well and incubated for 2 minutes at RT. Two to three drops of ready to use AEC chromogen (Dako) was added into each well and incubated at RT for 4 minutes. During this incubation period the immunoreaction was monitored under the light microscope on Bright Field at × 4 magnification. The wells were then washed with 200 μL distilled water to stop the chromogen reaction. The chamber slide wells were removed and the slide was mounted with prewarmed glycerol gelatine (Sigma-Aldrich, Poole, UK) and mounted with a cover-slip. The substitution of the primary antibody with 1% serum from the host of the secondary antibody was included as a negative control.
Scratch wound assay
To produce an in vitro wound healing assay, cells were grown to confluence and mechanically scratched to remove a fixed area of cells as previously described.26 For the migration assays a single scratch wound was created along the diameter of the dish according to a predesigned template. This was done using a standard plastic paper clip (800 μm wide). For all other assays, a different predesigned template was used to establish multiple scratches resulting in a cross-hatch pattern using a P200 Gilson pipette tip.
Cell migration assay
Following wounding cells were washed three times with PBS to remove damaged and nonadherent cells. Triplicate dishes of DF (n=7), DP (n=5), and DS (n=4) cells were cultured in 2 mL serum-free, phenol red-free medium with either vehicle control or 17β-estradiol. A previous dose–response study showed that 10 nM 17β-estradiol was effective at inducing migration of human breast DF, therefore all experiments used a concentration of 10 nM 17β-estradiol. In each dish the distance between the wound edges was measured at six fixed points (3 mm apart) along the length of the wound using a standard template. The migratory difference at six time points (0, 4, 8, 12, 24, and 48 hours) was recorded for each dish.
Tritiated thymidine assay
Wounded and nonwounded DF (n=7), DP (n=5), and DS (n=4) monolayers were washed three times with PBS to remove damaged and nonadherent cells. Triplicate dishes were incubated for 24 hours with serum-free, phenol red-free medium containing 17β-estradiol or vehicle control. Because a previous dose–response study showed that 10 nM 17β-estradiol was effective at inducing DNA synthesis of human breast DF, a concentration of 10 nM 17β-estradiol was used in all experiments.
The cells were incubated for six hours with methyl-3H-thymidine (S.A. 0.925 TBq/mmol; 0.5 μCi/dish). Three dishes containing nonwounded cells and three containing wounded cells incubated without 3H-thymidine were included for cell counting by hemocytometry to allow 3H-thymidine uptake to be correlated to cell number. After six hours the supernatant was aspirated and replaced with 1 mL 10% trichloroacetic acid (TCA) for 10 minutes at 4 °C. This was then aspirated and replaced by 400 μL 1 M NaOH solution and incubated at 37 °C for 16 hours. The aspirate from each dish was mixed with 4 mL scintillation cocktail (Ultima Gold XR, Perkin Elmer Life and Analytical Sciences, Boston) and the radioactivity (disintegrations per minute, d.p.m.) was measured using an LKB liquid scintillation spectrometer with a counting efficiency of 50%.
Total collagen secretion
Parallel dishes of mechanically wounded and nonwounded DF (n=7), DP (n=5), and DS (n=4) cells were prepared. Because the cells were incubated for up to four days, they were cultured in phenol red-free DMEM containing 5% charcoal stripped fetal calf serum, in the presence or absence of 17β-estradiol (10 nM). Ascorbic acid (50 μg/mL) was also added to the culture medium. Culture supernatant was collected from triplicate dishes after 2 days (2-day experiment) and the medium was replaced. Cells were incubated for a further 2 days after which culture supernatant was again collected (4-day experiment). Total soluble collagen was measured using the Sircol assay (Biocolor, Belfast, UK). Sirius Red reagent was added to 100 μL test sample according to the manufacturer's recommendations. This was mixed for 30 minutes, centrifuged at 13,000 ×g for 5 minutes and then following removal of the original solution, dissolved in 0.5 M NaOH. The absorbance was measured on a spectrophotometer at 570 nm and a calibration curve plotted according to the standards provided. Each assay was performed in triplicate, and a mean value calculated for each individual sample.
Parallel dishes of mechanically wounded and nonwounded DF (n=5), DP (n=5), and DS (n=4) cells were prepared and triplicate dishes were incubated in serum-free, phenol red-free DMEM in the presence or absence of 17β-estradiol (10 nM). After 24 hours the conditioned medium was aspirated from each dish and stored at −20 °C until required. VEGF levels within the conditioned medium were measured in triplicate using a Quantikine ELISA kit (R & D Systems, Abington, UK) following the manufacturer's recommendations.
Patient-specific data are presented as the mean ± SEM. The difference between each patient mean was analyzed using the either the paired Student's t-test, if normally distributed or the Mann–Whitney U-test. A p-value of <0.05 was considered significant. Statistical values were obtained using GraphPad Prism 2.01. Graphpad Software Inc and Analysis Toolpak, Microsoft Excel 97, Microsoft Office XP.
Data from the quantitative real-time PCR method were corrected for GAPDH levels, and normalized to the mean control ratio relative to the dermal fibroblast cell extracts according to the method of Pfaffl.40 Comparison of mean differences was performed using one-way ANOVA with Tukey's honestly significant difference test on the log-transformed ratios. A p-value of <0.05 was considered significant. Statistical values were obtained using GraphPad InStat version 3.01 for Windows 95/NT (GraphPad Software, San Diego, CA http://www.graphpad.com).
Establishment of the basal expression levels of ERα and ERβ transcripts in cultured dermal fibroblasts, dermal papilla cells, and dermal sheath cells
All three cell types expressed transcripts for ERα, ERβ and GAPDH (Figure 2). There was no evidence of contaminating gDNA and all cell types generated amplified sequences of expected molecular masses of 381, 279, and 347 bp for ERα, ERβ, and GAPDH, respectively (Figure 2). Quantitation of the relative levels of transcripts for the ER isoforms indicated that there were significant differences in the levels of ERα and ERβ transcripts with the highest levels in dermal fibroblast and dermal sheath cells, respectively. By contrast, the levels of GAPDH transcripts were approximately 3,000, 6,000, and 1,700 times greater than ERα transcript levels in the DP, DF and DS cells, respectively (Table 2). In relative terms, therefore, ERα expression in DF was approximately half that of DP cells and one third of DS cells (Table 2). Similarly, the expression of ERβ in DF was approximately two thirds of that found in both DP and DS cells (Table 2).
Table 2. Expression levels of ERα ERβ and GAPDH
ERα nmol/ng RNA
ERβ nmol/ng RNA
GAPDH nmol/ng RNA
ERα Relative Expression
ERβ Relative Expression
Data are the mean ± sem of 4 measurements of the transcript levels from 2 separate cultures. Relative expression levels were calculated using the Pfaffl's method40 using the extracts from the dermal papilla cell cultures as reference.
p<0.001 compared to the dermal papilla cell extracts;
p<0.001 compared to dermal fibroblast extracts; one-way ANOVA with Tukey's honestly significance difference test.
Cultured DF, DP, and DS cells all exhibited a similar expression of ERα and ERβ (Figure 3). All cells expressed strong cytoplasmic staining for both estrogen receptors and there appeared to be little variation between the level of expression for ERα and ERβ either within the same cell type, or between the different cell types.
Dermal fibroblasts, dermal papilla, and dermal sheath cells migrate into the mechanically created wound in vitro
In all cell populations, cells were observed migrating into the mechanically created wound (Figure 4). The distance between the wound edges was measured at 6 fixed points in triplicate dishes at 0, 4, 8, 12, 24, and 48 hours. An obvious movement of cells 4 hours following scratching was observed and this movement continued throughout the time period monitored. By 48 hours, the cells were almost touching in the middle of the “wound.” The DF and DS cells had an elongated appearance that is typical of migrating cells and in some cells there was evidence of membrane ruffling suggesting the presence of lamellopodia. In contrast, DP cells tended to maintain their rounded appearance while migrating (Figure 4).
Scalp dermal fibroblasts show an increased migratory rate in response to 17β-estradiol in an in vitro wound-healing assay
There was no significant difference in the rate of migration between scalp DF (n=7), DP (n=5), and DS (n=4) cells at any of the time points assessed (4, 8, 12, 24, and 48 hours) under serum-free, phenol red-free conditions. However, 17β-estradiol significantly (p<0.05) increased the migration of scalp DF after 48 hours by 10%, but had no significant effect on DP and DS cell migration at any time point assessed.
Scalp DF, DP, and DS cell cultures show different basal levels of DNA synthesis, which is increased by mechanical wounding, but is not altered by 17β-estradiol
Basal DNA synthesis of scalp DF (n=7), DP (n=5), and DS (n=4) cells was assessed after 24 hours using a 3H-thymidine assay (Figure 5). DF and DP cells showed a significantly (p<0.05) greater level of DNA synthesis (204 and 211 dpm/104 cells, respectively) than DS cells (164 dpm/104 cells) when all cell cultures were compared (Figure 5). Mechanical wounding significantly increased the DNA synthesis of all three cell types, although the relative increase in DNA synthesis was significantly (p<0.01) greater in DF (72%) and DS (78%) cells compared with DP (26%) cells (p<0.05) (Figure 5). However, 17β-estradiol did not significantly alter DNA synthesis by the nonwounded or mechanically wounded DF, DP, or DS cells (data not shown).
Scalp DF, DP, and DS cell cultures show different basal levels of total collagen secretion, which is increased by mechanical wounding, but is not altered by 17β-estradiol
The Sircol assay was used to measure basal total soluble collagen levels in cell supernatant collected from cell cultures of scalp DF (n=5), DP (n=5), and DS (n=4) cells after 2 and 4 days. Collagen secretion by nonwounded monolayers of DP and DS cells was significantly (p<0.05) greater (31 and 20 μg/104 cells, respectively) than that of DF (12 μg/104 cells) after 2 days (Figure 6). Furthermore, total collagen secretion by DP cells was significantly (p<0.05) greater than the DS cells (Figure 6). Similar results were seen in the medium collected after 4 days in culture (data not shown).
In response to mechanical wounding, a significant (p<0.05) increase in total soluble collagen secretion was shown by DF (150%) and DS (90%) cells after 2 days (Figure 6). In contrast, mechanical wounding did not significantly alter the secretion of total soluble collagen secretion by DP cells after 2 days (Figure 6). Similar results were seen 4 days after wounding (data not shown). Incubation with 17β-estradiol did not alter total soluble collagen secretion by either nonwounded or mechanically wounded cell populations after 2 or 4 days (data not shown).
Scalp DF, DP, and DS cell cultures secrete similar amounts of VEGF, but show differential effects in response to mechanical wounding and incubation with 17β-estradiol
An ELISA was used to measure VEGF levels in medium collected from cultures of scalp DFs (n=5), DP (n=5), and DS (n=4) cells. The secretion of VEGF by nonwounded monolayers of all cell types was similar (Figure 7). Incubation with 17β-estradiol did not alter the secretion of VEGF levels by nonwounded monolayers of DF, DP, or DS cells (Figure 7).
However, a significant (p<0.05) increase in VEGF secretion was shown in response to mechanical wounding by DF (68%) and DS (98%) cells, but not by DP cells (Figure 7). Although 17β-estradiol did not significantly alter VEGF secretion by mechanically wounded DF (Figure 7), mechanically wounded DP cells showed a significant (p<0.05) increase (55%) in VEGF secretion in response to 17β-estradiol (Figure 7), while mechanically wounded DS cells showed a significant (p<0.05) decrease (52%) in VEGF secretion in response to 17β-estradiol (Figure 7).
Although previous studies have shown that cultured human DF27 and DP28 cells express both ERα and ERβ this study now confirms that human DS cells also express both intracellular estrogen receptors (Figures 2 and 3). Furthermore, quantitation of the relative levels of transcripts for the ER isoforms using real-time RT-PCR indicated that there were significant differences in the levels of ERα and ERβ transcripts (Table 2).
An in vitro wound-healing assay was used to compare cell migration, proliferation, collagen and VEGF secretion between populations of scalp DF, DP and DS cells in response to 17β-estradiol. Mechanically wounding or “scratching” of cell monolayers is a recognized method of assessing cell migration in vitro.26 The same technique has also been used to investigate the proliferative response of cultured human DF to mechanical wounding in vitro.29
Because serum contains multiple growth factors and phenol red has been reported to have estrogenic effects, all experiments with 17β-estradiol were performed in their absence. Under serum-free and phenol red-free conditions, there was no difference in the rate of cell migration between DF, DP and DS cells. The increase in DF migration in response to 17β-estradiol concurs with our previous study on DF derived from breast skin. However, 17β-estradiol did not alter the migration of DP or DS cells under the same conditions, suggesting that although migration of DF is estrogen-dependent, migration of DP and DS cells is not. Because the expression of both ER isoforms in the DP and DS cells were confirmed by RT-PCR and immunohistochemistry then the migration rates of DP and DS cells in the presence of 17β-estradiol must be modulated via different mechanisms to that of DF cells.
This study also showed a significantly greater basal rate of DNA synthesis by DF and DP cells after 24 hours in serum-free, phenol red-free medium, compared with DS cells (Figure 5). Although previous studies have reported that DF cells proliferate at a greater rate than DP cells in culture,30 many studies assess proliferation in the presence of serum, while we assessed cells in the absence of serum and phenol-red. A higher proliferative response to serum by cultured DF compared with DP cells has been reported31 and we have previously demonstrated that DP cells secrete autocrine factors that stimulate their own DNA synthesis in the absence of serum.32 Therefore, in the absence of serum proliferation of DP cells may be similar to corresponding DFs.
Notwithstanding, following mechanical wounding, although all three cell types showed an increase in DNA synthesis (Figure 5), the relative increase was significantly greater in DF and DS cells when compared with DP cells. This response to mechanical wounding would concur with the proliferative capacity of these cell populations in vivo, because DS cells proliferate in early anagen and contribute to the anagen DP. The increased proliferative response of DS cells reflects a similar response seen by DF, highlighting a parallel between these cell types, distinct from DP cells.
The presence of 17β-estradiol did not alter the proliferation of nonwounded or mechanically wounded DF, DP, or DS cells (data not shown). This concurs with a previous study on nonwounded DF and DP cells where similar concentrations of 17β-estradiol had no effect on proliferation.33 Similarly androgens do not stimulate DNA synthesis of human beard DP cells in vitro, but exert their effects on DP cells via indirect mechanisms.32 It could be hypothesized that estrogens also work via similar mechanisms, stimulating the secretion of extracellular matrix components such as collagen, or increasing the secretion of specific growth factors and cytokines.
The amount of total collagen secreted per cell by nonwounded monolayers of dermal and hair follicle fibroblasts differed significantly (Figure 6). Basal collagen secretion by DP cells was significantly higher than DS cells, with both secreting significantly more collagen than corresponding DF (Figure 6). Although a significant increase in collagen secretion was seen in response to mechanically wounding DF and DS cells, there was no increase in total collagen secretion by mechanically wounded DP cells (Figure 6). This suggests that in line with DF, factors secreted in response to mechanical wounding stimulate collagen secretion by DS cells, providing evidence of further similarities between the two cell populations.
In vivo studies have shown an increase in collagen deposition in response to estrogen in normal and wounded skin in elderly individuals,18 while similar effects were also shown in skin of young men treated with topical estrogen.34 A previous study has also shown increased radiolabelled proline incorporation by intact cultured human DF in response to 1–10 nM estradiol,27 although this is the first study to investigate the effect of 17β-estradiol on collagen secretion by mechanically wounded human DF in vitro. Similarly, no previous studies have investigated the effect of 17β-estradiol on collagen secretion by DP or DS cells. However, there was no significant change in total collagen secretion in response to 17β-estradiol by either the nonwounded or wounded DF, DP, or DS cells after either 2 or 4 days in culture in the present study (data not shown).
Notwithstanding, because only total collagen secretion was determined, more relevant information may be gained by investigating the effect of estrogens on individual collagens. In vivo, although collagen types I and III are both important in wound healing, the relative proportion of collagen type III is increased in early wounds. Furthermore, estrogen has been shown to alter the ratio of types I to III collagen in the skin of postmenopausal women. Therefore, although the present study showed that total collagen secretion was not altered in response to 17β-estradiol, a change in the secretion of specific collagen subtypes was not established.
An important growth factor in wound healing is VEGF. All three cell types secreted similar levels of VEGF in vitro (Figure 7), which is in agreement with a previous study demonstrating VEGF mRNA expression is comparable between human DF, DP, and DS cells in vitro.36 Although mechanical wounding significantly increased VEGF secretion by DF and DS cells, it did not alter VEGF secretion by DP cells (Figure 7). Because increased secretion of VEGF is important in promoting angiogenesis, this highlights further similarities between DF and DS cells in a wound-healing assay.
Although 17β-estradiol did not significantly alter secretion of VEGF by intact monolayers of DF, DP, or DS cells, following mechanical wounding 17β-estradiol significantly increased VEGF secretion by DP cells (Figure 7). In contrast, in mechanically wounded DS cells, a significant decrease in VEGF secretion in response to 17β-estradiol was observed (Figure 7).
Opposing responses of wounded DP and DS cells imply that these hair follicle cells, with potential stem cell capability, have different roles in a wound-healing scenario. Plucking appears to mechanically wound the hair follicle, which induces proliferative activity and initiates the start of anagen (new hair growth). An important scenario in this event is angiogenesis of the hair follicle. Because the human hair cycle may be modulated by estrogens; the stimulation of DP cells by 17β-estradiol to secrete VEGF may be important in the initiation of anagen and the angiogenesis of the hair follicle.
However, because estrogen down-regulates the secretion of VEGF by wounded DS cells, this suggests that they have a role distinct from the DP cells. If DS cells are incorporated into the healing dermis, as previously suggested,14 it is possible that VEGF secretion is not required at this early time point of 24 hours. Because angiogenesis is a later event of the wound healing response, estrogen may dampen down the early secretion of VEGF by DF and DS cells. If this is the case, the time-dependent effects of estrogen on VEGF secretion by dermal and follicular fibroblasts warrants further investigation.
In summary, this study has shown that DS cells bear stronger similarities to corresponding DF in terms of migration, proliferation, collagen and VEGF secretion in a wound-healing assay, than to DP cells. In addition, estrogen, in the form of 17β-estradiol, differentially regulates the secretion of VEGF by DP and DS cells following mechanical wounding in vitro, suggesting that these mesenchymal hair follicle cells have separate and distinct roles. The modulation of other growth factors by 17β-estradiol warrants further investigation, because it may help to clarify the role of estrogen not only in wound healing, but also in the regulation of the hair cycle.
We would like to thank the Burns and Plastic Surgery Unit, University of Bradford, UK, for providing the funding for this study.