Advertisement

Oxidative Stress–Associated Senescence in Dermal Papilla Cells of Men with Androgenetic Alopecia

      Dermal papilla cells (DPCs) taken from male androgenetic alopecia (AGA) patients undergo premature senescence in vitro in association with the expression of p16INK4a, suggesting that DPCs from balding scalp are more sensitive to environmental stress than nonbalding cells. As one of the major triggers of senescence in vitro stems from the cell “culture shock” owing to oxidative stress, we have further investigated the effects of oxidative stress on balding and occipital scalp DPCs. Patient-matched DPCs from balding and occipital scalp were cultured at atmospheric (21%) or physiologically normal (2%) O2. At 21% O2, DPCs showed flattened morphology and a significant reduction in mobility, population doubling, increased levels of reactive oxygen species and senescence-associated β-Gal activity, and increased expression of p16INK4a and pRB. Balding DPCs secreted higher levels of the negative hair growth regulators transforming growth factor beta 1 and 2 in response to H2O2 but not cell culture–associated oxidative stress. Balding DPCs had higher levels of catalase and total glutathione but appear to be less able to handle oxidative stress compared with occipital DPCs. These in vitro findings suggest that there may be a role for oxidative stress in the pathogenesis of AGA both in relation to cell senescence and migration but also secretion of known hair follicle inhibitory factors.

      Abbreviations

      4-MU-Gal
      methylumbelliferyl galactopyranosidase
      AGA
      androgenetic alopecia
      AR
      androgen receptor
      BDPC
      balding dermal papilla cell
      DHT
      dihydrotestosterone
      DP
      dermal papilla
      DPC
      dermal papilla cell
      GSH
      glutathione sulfhydryl (reduced state)
      GSSG
      glutathione disulphide (oxidized state)
      ODPC
      occipital dermal papilla cell
      PD
      population doubling
      pRB
      retinoblastoma protein
      ROS
      reactive oxygen species
      SA-β-Gal
      senescence-associated beta-galactosidase
      TGF-β1/2
      transforming growth factor beta 1/2

      INTRODUCTION

      Androgenetic alopecia (AGA) is the most common form of male hair loss affecting 31% of genetically predisposed men aged 40–55 years (
      • Hamilton J.B.
      Patterned loss of hair in man; types and incidence.
      ;
      • Nyholt D.R.
      • Gillespie N.A.
      • Heath A.C.
      • et al.
      Genetic basis of male pattern baldness.
      ). AGA is a polygenic disorder with multiple risk loci (
      • Hoffmann R.
      • Seidl T.
      • Neeb M.
      • et al.
      Changes in gene expression profiles in developing B cells of murine bone marrow.
      ;
      • Heilmann S.
      • Kiefer A.K.
      • Fricker N.
      • et al.
      Androgenetic alopecia: identification of four genetic risk loci and evidence for the contribution of WNT-signaling to its etiology.
      ) driven predominantly by mutations of the X-linked androgen receptor (AR) gene (
      • Ellis J.A.
      • Stebbing M.
      • Harrap S.B.
      Polymorphism of the androgen receptor gene is associated with male pattern baldness.
      ). Oxidative stress may also have a role, as hair loss is linked to a number of factors that increase cellular oxidative stress, including metabolic syndrome, alcohol consumption, smoking, and UV radiation (
      • Severi G.
      • Sinclair R.
      • Hopper J.L.
      • et al.
      Androgenetic alopecia in men aged 40-69 years: prevalence and risk factors.
      ;
      • Trueb R.M.
      Association between smoking and hair loss: another opportunity for health education against smoking?.
      ,
      • Trueb R.M.
      Is androgenetic alopecia a photoaggravated dermatosis?.
      ;
      • Su L.H.
      • Chen T.H.
      Association of androgenetic alopecia with metabolic syndrome in men: a community-based survey.
      ). However, it is not known whether oxidative stress is a pathogenic contributor to AGA.
      The dermal papilla (DP) contains a population of pluripotent stem cells (
      • Hunt D.P.
      • Morris P.N.
      • Sterling J.
      • et al.
      A highly enriched niche of precursor cells with neuronal and glial potential within the hair follicle dermal papilla of adult skin.
      ;
      • Driskell R.R.
      • Clavel C.
      • Rendl M.
      • et al.
      Hair follicle dermal papilla cells at a glance.
      ) and has a fundamental role in regulating hair follicle development and hair growth (
      • Oliver R.F.
      Whisker growth after removal of the dermal papilla and lengths of follicle in the hooded rat.
      ;
      • Jahoda C.A.
      • Horne K.A.
      • Oliver R.F.
      Induction of hair growth by implantation of cultured dermal papilla cells.
      ). ARs are expressed by DP cells (DPCs) and the actions of androgens on hair growth are believed to be mediated via androgen regulation of hair growth regulatory factors (
      • Randall V.A.
      • Thornton M.J.
      • Hamada K.
      • et al.
      Androgens and the hair follicle. Cultured human dermal papilla cells as a model system.
      ;
      • Choudhry R.
      • Hodgins M.B.
      • Van der Kwast T.H.
      • et al.
      Localization of androgen receptors in human skin by immunohistochemistry: implications for the hormonal regulation of hair growth, sebaceous glands and sweat glands.
      ;
      • Hibberts N.A.
      • Howell A.E.
      • Randall V.A.
      Balding hair follicle dermal papilla cells contain higher levels of androgen receptors than those from non-balding scalp.
      ). In balding DPCs (BDPCs), androgens stimulate the secretion of hair growth inhibitory factors such as transforming growth factor beta 1 and 2 (TGF-β1/β2) and DKK-1 (
      • Foitzik K.
      • Lindner G.
      • Mueller-Roever S.
      • et al.
      Control of murine hair follicle regression (catagen) by TGF-beta1 in vivo.
      ;
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      ;
      • Hibino T.
      • Nishiyama T.
      Role of TGF-beta2 in the human hair cycle.
      ;
      • Kwack M.H.
      • Kim M.K.
      • Kim J.C.
      • et al.
      Dickkopf 1 promotes regression of hair follicles.
      ). In contrast, the occipital DPCs (ODPCs) from occipital, nonbalding scalp regions are classically insensitive to androgens (
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      ;
      • Randall V.A.
      Hormonal regulation of hair follicles exhibits a biological paradox.
      ).
      Previously, we demonstrated that BDPCs underwent premature senescence in vitro compared with ODPCs (
      • Bahta A.W.
      • Farjo N.
      • Farjo B.
      • et al.
      Premature senescence of balding dermal papilla cells in vitro is associated with p16(INK4a) expression.
      ) with a concomitant elevation of p16INK4a and retinoblastoma protein (pRB) expression, both of which are known to mediate cell cycle arrest in response to environmental stress (
      • Chen Q.M.
      Replicative senescence and oxidant-induced premature senescence. Beyond the control of cell cycle checkpoints.
      ). A major cause of cell stress in vitro stems from oxidative stress caused by reactive oxygen species (ROS;
      • Alaluf S.
      • Muir-Howie H.
      • Hu H.L.
      • et al.
      Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
      ;
      • Grayson W.L.
      • Zhao F.
      • Izadpanah R.
      • et al.
      Effects of hypoxia on human mesenchymal stem cell expansion and plasticity in 3D constructs.
      ). Therefore, because the concentration of oxygen in the dermis has been measured at between 1 and 5% O2 (
      • Wang W.
      Oxygen partial pressure in outer layers of skin: simulation using three-dimensional multilayered models.
      ) and cell culture is usually carried out at atmospheric levels of oxygen (21% O2) known to cause senescence in vitro (
      • Alaluf S.
      • Muir-Howie H.
      • Hu H.L.
      • et al.
      Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
      ), we have suggested that the premature senescence of BDPCs in vitro may be in response to stress or “culture shock” of tissue culture and may be indicative that BDPCs are more sensitive than ODPCs to environmental—including oxidative—stress. We sought to investigate the effects of physiologically relevant oxygen and oxidative stress on the growth and cell signaling potential of both BDPCs and ODPCs. We demonstrate in vitro a previously unreported role for oxidative stress in the pathogenesis of AGA both in relation to cell senescence and secretion of known hair follicle inhibitory factors. Furthermore, these experiments also highlight the major impact that oxygen can have on the physiology and biochemistry of cultured DPCs.

      RESULTS

      Low oxygen culture increases BDPC and ODPC viability and migration

      Patient-matched DPCs cultured from balding and occipital/nonbalding scalp biopsies demonstrate clear morphological differences when culture is performed in atmospheric (normoxia) versus physiological (hypoxic) O2. Both ODPCs (Figure 1a) and BDPCs (Figure 1b) exhibited a morphology associated with stress at 21% O2. BDPCs displayed a particularly severe phenotype, growing in a dispersed manner with flattened morphology classically associated with senescence (
      • Alaluf S.
      • Muir-Howie H.
      • Hu H.L.
      • et al.
      Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
      ). ODPCs at 21% O2 also had a mildly flattened morphology, although they grew in organized clusters and did not exhibit full senescent morphology (Figure 1a).
      Figure thumbnail gr1
      Figure 1Dermal papilla cell (DPC) viability is directly affected by O2 conditions. Morphology and arrangement of DPCs under varying O2 conditions for the following: (a) ODPCs at 21% O2, (b) BDPCs at 21% O2, (c) ODPCs at 2% O2, and (d) BDPCs at 2% O2. Images are representative of cells from three individual patients. (e) Cell velocity was measured by culturing DPCs to 30–50% confluence and capturing images from 10 randomly selected areas at 10-minute intervals for 24 hours using an inverted light microscope contained within a cell incubator at atmospheric (95% air and 5% CO2) or low oxygen (97% N2, 5% CO2 and 2% O2) conditions. Metamorph software was used to track the velocity of 10 cells from each of the videos produced. Data are mean±SEM. Statistical analysis was carried out using one-way analysis of variance with Bonferroni’s post hoc test. **P<0.01, ***P<0.001. BDPCs, blading DPCs; BDPCs, DPC, dermal papilla cell; ODPCs, occipital DPCs. Bar=200 μm.
      ODPCs grown at 2% O2 maintained their healthy, spindle-like morphology and formed organized clusters with raised pseudopapillae (Figure 1c), a well-established characteristic of cultured DPCs (
      • Almond-Roesler B.
      • Schon M.
      • Schon M.P.
      • et al.
      Cultured dermal papilla cells of the rat vibrissa follicle. Proliferative activity, adhesion properties and reorganization of the extracellular matrix in vitro.
      ). BDPCs at 2% O2 maintained a spindle-like morphology typical of healthy fibroblasts (Figure 1d), although BDPC cultures did not form pseudopapillae.
      Oxygen conditions also mediated significant changes in the migratory function of DPCs (Figure 1e). Time-lapse analysis of cells cultured in 21% O2 revealed that ODPCs had a mean velocity of 0.12 μm per minute, which was over 8-fold faster than BDPCs (0.014 μm per minute) under the same conditions (P<0.001). At 2% O2, the mean velocity of both DPCs significantly increased to 0.58 μm per minute for ODPCs and 0.34 μmper minute for BDPCs. Notably, there was still a significant difference between BDPCs and ODPCs at 2% O2 (P<0.01), although low oxygen conditions appeared to reduce the relative difference in mean velocities between DPC types.

      Oxidative stress decreases population doubling and increases senescence in DPCs

      Measurement of population doubling (PD) and quantification of senescence-associated beta-galactosidase (SA-β-Gal) activity showed that ODPCs and BDPCs were more proliferative (P<0.001 for both ODPCs and BDPCs) and had reduced levels of senescence when cultured at 2% O2 versus 21% O2 (Figure 2). However, BDPCs underwent fewer PD than ODPCs under both oxygen conditions (12 vs. 16 PD at 2% O2 (P<0.001) and 5 vs. 9 PD at 21% O2P<0.001; Figure 2a and b); furthermore, senescence was significantly higher in BDPCs at 21% O2 than ODPCs (P<0.001; Figure 2c). There was no significant difference in senescence between ODPCs and BDPCs at 2% O2.
      Figure thumbnail gr2
      Figure 2O2 conditions have a direct effect on dermal papilla cell (DPC) proliferation and senescence. Population doubling rates measured at each passage for (a) ODPCs and (b) BDPCs from three individual patients in triplicate. DPCs were initially cultured at 21% O2 before being equally split between 2 and 21% O2 incubators at passage 2. At passage 4, DPCs were split again, and either maintained at the same condition or switched to the opposite O2 condition. (c) DPCs were taken from the end point (passage 6) of samples used in a and b. Lysates were quantified for senescence-associated β-Gal activity using 4-MU-Gal substrate. (d) ROS levels quantified via H2DCFDA fluorescence in DPCs from four individual patients in triplicate at 2 and 21% O2, and in the presence and absence of N-acetyl cysteine (e).(f) Senescence-associated β-Gal activity was quantified in DPCs in the presence and absence of N-acetyl cysteine. Data are mean±SEM. Statistical analysis was carried out using one-way analysis of variance with Bonferroni’s post hoc test. *P<0.05, **P<0.01, ***P<0.001. DPC, dermal papilla cell; H2DCFDA, hydroxyl-2-dichlorofluorescein diacetate; ODPCs, occipital DPCs; ROS, reactive oxygen species; 4-MU-Gal, methylumbelliferyl galactopyranosidase.
      To investigate whether the effects of oxygen on PD and senescence were reversible, we also switched cells at passage 4 from 21 to 2% O2 and from 2 to 21% O2 (Figure 2a and b). The slower growth rate of ODPCs and BDPCs at 21% O2 was rescued when cells were switched to 2% O2 (P<0.05 and P<0.001 for ODPCs and BDPCs, respectively), which correlated with the levels of cell senescence similar to those seen in cells grown continuously in 2% O2 (Figure 2c). In contrast, PD decreased in cells switched from 2 to 21% O2 (P<0.01 and P<0.001 for ODPCs and BDPCs, respectively) correlating with a significant increase in cell senescence (Figure 2c). BDPCs were particularly sensitive to senescence when transferred from 2 to 21% O2 (P<0.01). Thus, environmental O2 levels, which are routinely used in cell culture and known to cause oxidative stress in cultured cells, have a significant impact on DPC morphology, motility, proliferation, and senescence, and physiological (2%) O2 levels appear optimal to maintain healthy cells. Moreover, BDPCs are much more sensitive to the stress of environmental O2 than ODPCs.
      Hydroxyl-2-dichlorofluorescein diacetate was used to quantify ROS. Figure 2d shows that ROS levels were significantly higher in BDPCs at 21% O2 compared with ODPCs under identical conditions (P<0.01). At 2% O2, ROS levels were significantly lower in both BDPCs (P<0.001) and ODPCs (P<0.05) when compared with BDPCs and ODPCs at 21% O2, but there was no significant difference in ROS levels between BDPCs and ODPCs at 2% O2.
      We demonstrate that the higher levels of ROS in the BDPCs can be suppressed in a dose-dependent manner by treating BDPCs for 24 hours with the ROS-scavenger N-acetyl cysteine (NAC; Figure 2e). NAC was effective at significantly reducing the levels of ROS at 1 μM (P<0.05) and 5 and 10 μM (P<0.001; Figure 2e); moreover, analysis of SA-β-Gal showed that both 5 and 10 μM NAC also resulted in a significant (P<0.001) reduction in senescence in BDPCs (Figure 2f).

      Catalase and glutathione are influenced by oxygen and are higher in balding than nonbalding DPCs

      We also investigated the expression of catalase, an indicator of oxidative stress (
      • Clerch L.B.
      • Iqbal J.
      • Massaro D.
      Perinatal rat lung catalase gene expression: influence of corticosteroid and hyperoxia.
      ). Western blot analysis (Figure 3a) and densitometry (Figure 3b) showed that catalase expression was significantly lower in ODPCs at 2% O2 compared with BDPCs at the same passage under the same O2 conditions (P<0.001). At 21%, O2 catalase was upregulated in both ODPCs and BDPCs but remained higher in BDPCs compared with ODPCs (P<0.05). Catalase activity was also significantly higher in BDPCs (P<0.05) compared with ODPCs (Figure 3c) and was also higher at 21% O2 compared with 2% O2 (P<0.05). These data demonstrate that oxygen and cell passage have a significant impact on the stress response of DPCs but show that BDPCs appear to be more sensitive to stress than ODPCs.
      Figure thumbnail gr3
      Figure 3O2 conditions affect antioxidant response protein expression and activity. (a) Western blot analysis of catalase and β-actin loading control in ODPCs and BDPCs from three individual patients cultured at 2 and 21% O2. DPCs were lysed using RIPA buffer, and 5 μg of each lysate was run on a Sigma Nu-page electrophoresis gel. (b) Densitometric analysis was carried out on (n=3) blots using Image-J (NIH, Open source). Blots were stripped and reprobed with antibodies for β-actin to determine equal protein loading and to normalize densitometry values. (c) Activity of catalase quantified using Amplex Red assay in DPCs from three individual patients in triplicate at 2 and 21% O2. (d) Total glutathione concentration quantified using Ellman’s reagent, displaying reduced and oxidized fractions in DPCs cultured at 2 and 21% O2. Means±SEM. Statistical analysis carried out using the Bonferroni’s post hoc test. *P<0.05, **P<0.01, ***P<0.001. BDPCs, blading DPCs; ODPCs, occipital DPCs; RFU, relative fluorescence units; RIPA, radioimmunoprecipitation assay.
      Glutathione has a critical role in the elimination of ROS (
      • Maher P.
      The effects of stress and aging on glutathione metabolism.
      ). Total and oxidized glutathione was quantified in DPCs at 2 and 21% O2. Total glutathione (GSH and GSSG) concentration was significantly higher in BDPCs compared with ODPCs at 21% O2 (P<0.05; Figure 3d). Total glutathione concentration was also significantly higher in BDPCs cultured at 21% O2 compared with 2% O2 (P<0.01). There was no significant difference in total glutathione in BDPCs versus ODPCs at 2% O2 and between ODPCs at 21 and 2% O2. When we calculated the fraction of reduced glutathione (GSSG), we observed that BDPCs had a significantly higher concentration of GSSG at 21% O2 compared with ODPCs (P<0.01). No significant difference was observed between BDPCs and ODPCs at 2% O2. ODPCs at 21% O2 showed no statistical difference in GSH concentrations compared with those cultured at 2%. BDPCs cultured at 21% O2 had a significantly higher fraction of their total glutathione in its reduced form compared with those grown at 2% O2 (P<0.001). These data indicate that although the amount of active GSH was similar between ODPCs and BDPCs under all conditions, the overall fraction of active glutathione was smaller in BDPCs than in ODPCs and that active fraction of glutathione is lower in BDPCs than ODPCs under oxidative stress.

      Oxygen regulates p16INK4 and pRB

      Western blot analysis (Figure 4a) and densitometry showed that p16INK4a (Figure 4a and b) and the pRB(Figure 4a and c) were expressed at much higher levels in DPCs maintained at 21% O2 compared with 2% and that levels of p16INK4a and pRB were expressed at lower levels in ODPCs compared with BDPCs, thus confirming our previous data in which we showed that premature senescence of BDPCs was associated with p16INK4a and pRB (
      • Bahta A.W.
      • Farjo N.
      • Farjo B.
      • et al.
      Premature senescence of balding dermal papilla cells in vitro is associated with p16(INK4a) expression.
      ).
      Figure thumbnail gr4
      Figure 4O2 affects senescence-associated protein expression. (a) Western blot and (b) densitometric analyses for P16INK4a and pRB carried out on DPCs from three individual patients cultured at 2 and 21% O2. Data are mean±SEM. Statistical analysis carried out using one-way analysis of variance with Bonferroni’s post hoc test. *P<0.05. DPCs, dermal papilla cells; pRB, retinoblastoma protein.

      TGF-β and IGF-I secretion by balding and nonbalding DPCs is influenced by oxidative stress

      A critical function of the DP is its role in secreting growth factors that regulate the development and growth of the hair follicle. As TGF-β is a potent inhibitory hair growth factor and has previously been reported to be secreted by BDPCs in vitro in response to dihydrotestosterone (DHT;
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      ), we investigated the effect of oxidative stress and DHT on the secretion of TGF-β isoforms. Oxidative stress induced by hydrogen peroxide (100 μM H2O2) promoted the secretion of TGF-β1, and this was significantly higher (P<0.001) in BDPCs (Figure 5a). Oxygen conditions alone did not independently alter TGF-β secretion (Figure 5b and c). However, DHT (100 nM) significantly increased TGF-β1 and TGF-β2 secretion compared with the vehicle control (P<0.01) when BDPCs and ODPCs were cultured at 21% O2 (Figure 5b and c). Conversely, BDPCs cultured at 2% O2 showed no change in TGF-β2 secretion (Figure 5c) and a significant decrease in TGF-β1 secretion at both 1 nM (P<0.01) and 100 nM (P<0.001) DHT (Figure 5b). ODPCs cultured at 2% O2 also underwent a significant decrease in TGF-β1 secretion at 1 nM (P<0.01) and 100 nM (P<0.05) DHT (Figure 5b); however, no significant change in TGF-β2 secretion was observed (Figure 5c). This suggests that DPCs maintained in an environment of low oxidative stress (e.g., 2% O2) are protected from DHT-induced TGF-β secretion. IGF, one of the main growth factors involved in the maintenance of the follicle in the anagen stage of the growth cycle, was secreted at significantly lower levels by BDPCs compared with ODPCs and also suppressed in ODPCs when cultured at 21% O2 compared with 2% O2 (P<0.05), but not in BDPCs (Figure 5d).
      Figure thumbnail gr5
      Figure 5Oxidative stress induces changes to dermal papilla cell (DPC) growth factor secretion. (a) ELISA analysis was used to determine TGF-β1 secretion for DPCs from three individual patients at 2% O2 in response to H2O2-induced oxidative stress. Following this, BDPCs from three patients were cultured at both 2 and 21% O2 in the presence of DHT or ethanol vehicle control to quantify the secretion response of (b) TGF-β1, (c) TGF-β2, and (d) IGF. Data are mean±SEM. Statistical analysis carried out using one-way analysis of variance with Bonferroni’s post hoc test. *P<0.05; **P<0.01; ***P<0.001. DPCs, dermal papilla cells; TGF-β1/β2, TGF, transforming growth factor beta 1/2.

      DISCUSSION

      Here we show that environmental oxygen significantly alters DPC morphology, migration, proliferation, senescence, and TGF-β signaling. BDPCs were significantly more sensitive to oxidative stress than ODPCs, with reduced cell proliferation and migration together with increased ROS and senescence at 21% O2. Crucially, BDPCs that were protected from oxidative stress when cell culture was performed at 2% O2 did not secrete the negative hair growth factor TGF-β in response to DHT. This suggests that oxidative stress, as well as androgen signaling, may have an important role in the BDPC phenotype and AGA.
      Remodeling of the DP is essential for hair follicle cycling and is largely dependent on the migration of DPCs between the DP and the connective tissue sheath (
      • Tobin D.J.
      • Gunin A.
      • Magerl M.
      • et al.
      Plasticity and cytokinetic dynamics of the hair follicle mesenchyme: implications for hair growth control.
      ). Balding follicles exhibit shrinkage and rounding of the DP (
      • Miranda B.H.
      • Tobin D.J.
      • Sharpe D.T.
      • et al.
      Intermediate hair follicles: a new more clinically relevant model for hair growth investigations.
      ), suggesting restricted cell migration. Here we show that BDPCs have slower migratory velocity than ODPCs under both oxygen conditions. However, oxygen significantly altered cell motility—faster migration velocity was observed at 2% O2 in both DPC phenotypes. Therefore, these data suggest that reduced DPC migration caused by oxidative stress may inhibit hair follicle remodeling and could promote the balding phenotype.
      We previously reported that BDPCs undergo premature senescence in vitro, caused by elevated expression of p16INK4a and pRB, but not p53 or p21 (
      • Bahta A.W.
      • Farjo N.
      • Farjo B.
      • et al.
      Premature senescence of balding dermal papilla cells in vitro is associated with p16(INK4a) expression.
      ). We now show that the expression of p16INK4a and pRB is associated with high levels of oxygen. Environmental stress has been reported to trigger senescence via p16INK4a in dermal fibroblasts (
      • Chen Q.M.
      Replicative senescence and oxidant-induced premature senescence. Beyond the control of cell cycle checkpoints.
      ;
      • Jacobs J.J.
      • de Lange T.
      Significant role for p16INK4a in p53-independent telomere-directed senescence.
      ) and is considered a key factor of skin aging (
      • Alaluf S.
      • Muir-Howie H.
      • Hu H.L.
      • et al.
      Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
      ). Although cell passage did have an effect on the expression of senescence markers in DPCs, the differences between BDPCs and ODPCs cultured at 2 and 21% O2 were similar at low and high passage number. This suggests that the senescence that we observed was owing to oxidative stress as opposed to replicative senescence commonly observed in dermal fibroblasts as the result of passage-induced telomere shortening (
      • Itahana K.
      • Dimri G.
      • Campisi J.
      Regulation of cellular senescence by p53.
      ).
      Catalase was significantly higher in BDPCs than in ODPCs at passage 2, possibly in response to ROS. Catalase has previously been associated with hair follicle aging. However, in contrast to BDPCs, graying hair follicles express lower levels of catalase and higher levels of ROS compared with healthy, pigmented follicles (
      • Wood J.M.
      • Decker H.
      • Hartmann H.
      • et al.
      Senile hair graying: H2O2-mediated oxidative stress affects human hair color by blunting methionine sulfoxide repair.
      ;
      • Kauser S.
      • Westgate G.E.
      • Green M.R.
      • et al.
      Human Hair Follicle and Epidermal Melanocytes Exhibit Striking Differences in Their Aging Profile which Involves Catalase.
      ). ROS produced by the melanocytes in graying follicles may be responsible for creating an oxidative environment that could affect DPCs especially in balding scalp. The elevated levels of catalase present in the BDPCs did not translate to a reduction in total ROS nor reduced levels of senescence, suggesting that the BDPCs may be deficient in other antioxidants or in their ability to handle ROS.
      In addition to catalase, we also showed that glutathione levels were higher in BDPCs compared with ODPCs at 21% O2. Glutathione is an ROS-sensitive signal modulator that senses the oxidative equilibrium of the cell (
      • Maher P.
      The effects of stress and aging on glutathione metabolism.
      ). Reduced (active) GSH regulates intracellular signaling by blocking the promoter-binding sites of AP-1 and SP-1 (
      • Vayalil P.K.
      • Iles K.E.
      • Choi J.
      • et al.
      Glutathione suppresses TGF-beta-induced PAI-1 expression by inhibiting p38 and JNK MAPK and the binding of AP-1, SP-1, and Smad to the PAI-1 promoter.
      ). When ROS is abundant, GSH is oxidized forming a dimer (GSSG) with altered binding function of intracellular transcriptional mediators (
      • Maher P.
      The effects of stress and aging on glutathione metabolism.
      ). Therefore, the proportion of active GSH is not only critical as an antioxidant but also for regulating transcription in response to oxidative stress. There was no difference in GSH and GSSG levels between BDPCs and ODPCs at 2% O2, indicating limited oxidative stress and no associated change in intracellular signaling. However, BDPCs exposed to 21% O2 had significantly higher levels of GSSG and proportionally less GSH. There is therefore scope for further examination of the role of GSH and whether it is able to modulate DHT signaling in DPCs.
      Secretion of TGF-β in response to oxidative stress is the underlying pathophysiology of pulmonary fibrosis (
      • Cui Y.
      • Robertson J.
      • Maharaj S.
      • et al.
      Oxidative stress contributes to the induction and persistence of TGF-beta1 induced pulmonary fibrosis.
      ), heart disease (
      • Yeh Y.H.
      • Kuo C.T.
      • Chan T.H.
      • et al.
      Transforming growth factor-beta and oxidative stress mediate tachycardia-induced cellular remodelling in cultured atrial-derived myocytes.
      ), and photoaging of skin (
      • Debacq-Chainiaux F.
      • Borlon C.
      • Pascal T.
      • et al.
      Repeated exposure of human skin fibroblasts to UVB at subcytotoxic level triggers premature senescence through the TGF-beta1 signaling pathway.
      ). We chose to measure total TGF-β—as opposed to active TGF-β—as the most suitable measure of TGF-β, as it gives a more accurate indication of the overall bioavailability of the growth factor (
      • Koli K.
      • Saharinen J.
      • Hyytiainen M.
      • et al.
      Latency, activation, and binding proteins of TGF-beta.
      ). TGF-β1 and TGF-β2 are negative regulators of hair growth (
      • Foitzik K.
      • Lindner G.
      • Mueller-Roever S.
      • et al.
      Control of murine hair follicle regression (catagen) by TGF-beta1 in vivo.
      ;
      • Hibino T.
      • Nishiyama T.
      Role of TGF-beta2 in the human hair cycle.
      ), and H2O2-induced senescence of fibroblasts causes sustained overexpression of TGF-β1 and TGF-β2 via a pRB-regulated pathway (
      • Frippiat C.
      • Chen Q.M.
      • Zdanov S.
      • et al.
      Subcytotoxic H2O2 stress triggers a release of transforming growth factor-beta 1, which induces biomarkers of cellular senescence of human diploid fibroblasts.
      ). We showed that TGF-β1 and TGF-β2 secretion by BDPCs was stimulated by an acute dose of H2O2, and therefore that oxidative stress is able to stimulate secretion of known hair growth inhibitory factors. However, the fact that we observed no increase in TGF-β secretion under 21% O2 suggests that the conditions of oxidative stress experienced in vitro are insufficient to affect TGF-β secretion by cultured DPCs.
      We also investigated the effects of oxygen on secretion of IGF-I, a positive regulator of hair growth in vitro (
      • Philpott M.P.
      • Sanders D.
      • Westgate G.E.
      • et al.
      Effects of insulin and insulin-like growth factors on cultured human hair follicles: IGF-I at physiologic concentrations is an important regulator of hair follicle growth in vitro.
      ). BDPCs secreted significantly less IGF-1 than ODPCs. As IGF-1 has been shown to maintain in vitro–cultured human hair follicles in anagen, reduced IGF-1 secretion by BDPCs may result in impaired hair growth. Although oxygen had no effect on IGF-1 secretion in BDPCs, secretion of IGF-1 by ODPCs was significantly lower under 21% oxygen compared with 2%. This suggests that BDPCs’ secretion of IGF-I may be dependent on environmental stimuli.
      ARs are found in both balding and nonbalding DPCs, with balding cells expressing a higher number of receptors (
      • Hibberts N.A.
      • Howell A.E.
      • Randall V.A.
      Balding hair follicle dermal papilla cells contain higher levels of androgen receptors than those from non-balding scalp.
      ). Androgens have been shown to stimulate TGF-β1 secretion in AR-transfected BDPCs (
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      ). In our study, DHT stimulated TGF-β secretion from DPCs only at 21% O2, suggesting that oxidative stress is an essential component of androgen response in AGA. In addition, our observation that at 2% oxygen DHT had significantly reduced TGF-β1 secretion in ODPCs and BDPCs. The reasons for this effect are unknown, although it has been shown in the prostate that DHT inhibits TGF-β signaling and secretion and that this is mediated via stromal cells (
      • Kyprianou N.
      • Isaacs J.T.
      Expression of transforming growth factor-beta in the rat ventral prostate during castration-induced programmed cell death.
      ;
      • Placencio V.R.
      • Sharif-Afshar A.R.
      • Li X.
      • et al.
      Stromal transforming growth factor-beta signaling mediates prostatic response to androgen ablation by paracrine Wnt activity.
      ). AR-transfected BDPCs but not ODPCs respond to synthetic androgen R-1881 by producing TGF-β1 (
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      ). In contrast, we also show that at 21% O2 ODPCs also secrete TGF-β1 and TGF-β2 in response to DHT. ODPCs are classically described as androgen-insensitive; however, their insensitivity stems from a lack of 5α-reductase, which converts testosterone to the more active DHT, rather than from a lack of AR. Indeed,
      • Hibberts N.A.
      • Howell A.E.
      • Randall V.A.
      Balding hair follicle dermal papilla cells contain higher levels of androgen receptors than those from non-balding scalp.
      have shown that ODPCs express AR, although at lower expression levels than the BDPCs. The use of DHT in our experiments would bypass 5α-reductase testosterone metabolism and supports the theory that androgen sensitivity in the DP is predominantly controlled by 5α-reductase (
      • Kaufman K.D.
      Androgen metabolism as it affects hair growth in androgenetic alopecia.
      ). We observed significant effects of DHT on TGF-β secretion after only 1 hour of DHT treatment. Whether these rapid effects are being mediated via AR modulation of gene expression is not known. We ruled out the role of the nongenomic androgen response element GPCR6A (
      • Pi M.
      • Parrill A.L.
      • Quarles L.D.
      GPRC6A mediates the non-genomic effects of steroids.
      ) via quantitative reverse transcriptase in real time PCR analysis (Supplementary Figure S1 online). However, ARs are able to mediate changes in cell biology via nonclassical pathways that involve direct activation of mitogen-activated protein kinases, and this may explain the rapid response that we report here (
      • Lange C.A.
      • Gioeli D.
      • Hammes S.R.
      • et al.
      Integration of rapid signalling events with steroid hormone receptor action in breast and prostate cancer.
      for review).
      In our experiments, we required 100 nM DHT to observe any effect on TGF-β secretion. Although supraphysiological, similar levels of DHT have been used in other studies (
      • Kwack M.H.
      • Sung Y.K.
      • Chung E.J.
      • et al.
      Dihydrotestosterone-inducible dickkopf 1 from balding dermal papilla cells causes apoptosis in follicular keratinocytes.
      ;
      • Shin H.
      • Yoo H.G.
      • Inui S.
      • et al.
      Induction of transforming growth factor-beta 1 by androgen is mediated by reactive oxygen species in hair follicle dermal papilla cells.
      ). Indeed,
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
      showed that cultured DPCs lose AR expression in culture, and in their study they had to transfect their DPCs with AR to show stimulation of TGF-β secretion.
      In addition to elucidating the effect of oxidative stress on DPCs, these data demonstrate the necessity of optimal cell culture conditions to assess DPC signaling. Tissue culture is routinely carried out at atmospheric levels of 21% O2, vastly higher than the physiological levels of 1–5% O2 found within the dermis in vivo. These supraphysiological levels of oxygen accelerate the fibroblast transition into a postmitotic, senescence state (
      • Alaluf S.
      • Muir-Howie H.
      • Hu H.L.
      • et al.
      Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
      ;
      • Chen Q.M.
      Replicative senescence and oxidant-induced premature senescence. Beyond the control of cell cycle checkpoints.
      ). Our findings highlight that this is also the case for DPCs and demonstrate that these cells should be cultured under hypoxic conditions (2% O2). It is now well established that hypoxia is a characteristic of the stem cell niche (
      • Mohyeldin A.
      • Garzon-Muvdi T.
      • Quinones-Hinojosa A.
      Oxygen in stem cell biology: a critical component of the stem cell niche.
      ) and the DP is known to contain stem cells (
      • Jahoda C.A.
      Cell movement in the hair follicle dermis - more than a two-way street?.
      ), and, moreover, that normoxic conditions of 21% O2 promote stem cell differentiation (
      • Mathieu J.
      • Zhou W.
      • Xing Y.
      • et al.
      Hypoxia-inducible factors have distinct and stage-specific roles during reprogramming of human cells to pluripotency.
      ). This may explain why cultured DPCs rapidly lose their inductive capacity (
      • Jahoda C.A.
      • Horne K.A.
      • Oliver R.F.
      Induction of hair growth by implantation of cultured dermal papilla cells.
      ).
      In conclusion, we present a potential link between oxidative stress and impaired DP function. We also highlight the beneficial use of low oxygen environment for DPCs, which would aid the expansion and maintenance of viable cell cultures for a greater number of passages, a finding that may be of major benefit to researchers wanting to screen large numbers of compounds, or clinicians looking to expand allogeneic grafts for re-implantation.

      MATERIALS AND METHODS

      Cell culture

      Patient-matched punch biopsies (2 mm2) were taken from balding (frontal) and nonbalding (occipital) scalps of men at stages IV–VI on the Hamilton–Norwood scale (
      • Norwood O.T.
      Male pattern baldness: classification and incidence.
      ) undergoing hair transplant surgery for AGA. Ethical approval for this study was obtained from the East London and City Health Authority (LREC 09/H0704/40), and biopsies were taken with written, informed patient consent. All experiments adhered to the Declaration of Helsinki principles.
      Dermal papillae were microdissected from the hair follicles using a stereoscopic microscope, as described previously (
      • Bahta A.W.
      • Farjo N.
      • Farjo B.
      • et al.
      Premature senescence of balding dermal papilla cells in vitro is associated with p16(INK4a) expression.
      ). Identical numbers of BDPCs and ODPCs were explanted and maintained in DP medium: Williams E medium, supplemented with 2 mM L-glutamine, 10 μg ml-1 insulin, 100 ng ml-1 hydrocortisone, 100 units per ml of penicillin G, and 100 mg ml-1 streptomycin supplemented with 15% (v/v) fetal calf serum ( reagents from (Sigma-Aldrich, Poole, UK) at 37 °C in a humidified atmosphere of 5% CO2/95% air. At passage 2, DPCs were divided and either maintained at 21% O2 or switched to a Sci-tive Stem Cell Station (Ruskinn, Bridgend, UK). Cells were maintained under physiological oxygen conditions 5% CO2, 93% N2, and 2% O2. Proliferation rates were determined using Nucleocassette cell counters (Chemometec, Königswinter, Germany). Population doubling rates were calculated as follows: PD=log2 (NH/NS), where NS is the number of cells seeded and NH is the number of cells harvested upon passage.

      Cell migration assay

      DPCs were seeded at a density of 1 × 104 into 6-well plates. The plate was placed on a robotically controlled platform on a Nikon inverted light microscope inside a thermostatically controlled chamber maintained at either 21 or 2% O2 as above. Ten randomly selected points were chosen from each well and photographed using the Metamorph software (Molecular Devices, Wokingham, UK). The software then recorded images at these chosen points every 10 minutes for 24 hours. The resultant images were then sequenced into a time-lapse video, and image analysis was carried out using the Metamorph software to assess the velocity of individual cells’ movements.

      SA-β-Gal assay

      Cytochemical detection of SA-β-Gal activity was carried out using a modified version of the 4-methylumbelliferyl galactopyranosidase assay (4-MU-Gal;
      • Gary R.K.
      • Kindell S.M.
      Quantitative assay of senescence-associated beta-galactosidase activity in mammalian cell extracts.
      ). Briefly, DPCs were lysed in buffer (40 mM citrate, 40 mM sodium phosphate, 5 mM 3-((3-cholamidopropyl)dimethylammonio)-1-propanesulfonate, and protease inhibitor cocktail (Roche, Lewes, UK); adjusted to pH 6.0). Lysates were centrifuged for 5 minutes at 12,000 G, and the supernatant was mixed (1:1) with 4-MU-Gal reaction buffer (40 mM citrate, 40 mM sodium phosphate, 300 mM NaCl, 4 mM MgCl2, 10 mM β-mercaptoethanol, and 1.7 mM of 4-MU-Gal; adjusted to pH 6.0). Samples were then incubated at 37 °C for 1 h before 50μl aliquots were taken and mixed with sodium bicarbonate (400 mM) stop solution. Fluorescence was measured at 360/465 nm every 5 minutes for 1 hour using a Synergy HT microplate reader with KC4 software (BioTek, Bedfordshire, UK).

      Protein extraction and immunohistochemistry

      DPCs were homogenized in radioimmunoprecipitation assay buffer containing protease inhibitors (Roche, Lewes, UK), centrifuged, and the supernatants were stored at -80 °C until analysis. Samples were separated using a Nu-PAGE 10% (v/v) Bis-Tris gel (Invitrogen, Paisley, UK) according to the manufacturer’s instructions and then transferred to a HyBond nitrocellulose membrane (GE Healthcare, Buckinghamshire, UK) and incubated with mouse anti-p16INK4a (Santa Cruz Biotechnology, Wiltshire, UK), mouse anti-pRB (Millipore, Watford, UK), rabbit anti-catalase, or goat anti-β-actin (Abcam, Cambridge, UK), and thereafter with the appropriate horseradish peroxidase–conjugated secondary antibodies (DakoCytomation, Glostrup, Denmark). Membranes were imaged using ECL-Plus chemiluminesence solution and light-sensitive Hyperfilm (GE Healthcare). Densitometric analysis was carried out using the Image-J software (Public domain, NIH).

      ROS assay

      DPCs were seeded at a density of 1 × 104 per well in an opaque 96-well plate and left to adhere overnight. Cells were washed twice with phosphate-buffered saline and incubated with 50 μl of phenol red–free DP-specified medium containing 25 μM hydroxyl-2-dichlorofluorescein diacetate and placed in the incubator for 1 hour. Fluorescence was measured at 485/527 nm every 2 minutes for 30 minutes using Synergy HT microplate reader with the KC4 software (BioTek). Fluorescence was calculated as the slope of relative fluorescence units per minute.

      Catalase assay

      The catalase assay was carried out using the Molecular Probes: Amplex Red catalase assay kit (Invitrogen) according to the manufacturer’s instructions. Fluorescence was measured at 530/560 nm using Synergy HT microplate reader (BioTek).

      Glutathione assay

      Total and reduced glutathione were measured simultaneously using an Ellman’s reagent (
      • Ellman G.
      • Lysko H.
      A precise method for the determination of whole blood and plasma sulfhydryl groups.
      ) multiwell kit (Cayman Scientific, Ann Arbor, MI) according to the manufacturer’s instructions. Samples were colorimetrically quantified, and absorbance was measured at 405 nm at 5-minute intervals for 30 minutes using the Synergy HT microplate reader (BioTek).

      ELISAs

      IGF, TGF-β1, and TGF-β2 were measured using the Quantikine ELISA kits (R&D, Abingdon, UK) according to the manufacturer’s instructions. DPCs were incubated in DP medium containing charcoal-stripped fetal bovine serum (15%) for 24 hours before testing. Plates were read at an absorbance wavelength of 450 nm using a Synergy HT microplate reader (BioTek).

      ACKNOWLEDGMENTS

      JU was funded by a BBSRC CASE PhD studentship in collaboration with Unilever Research awarded to MP.

      SUPPLEMENTARY MATERIAL

      Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

      REFERENCES

        • Alaluf S.
        • Muir-Howie H.
        • Hu H.L.
        • et al.
        Atmospheric oxygen accelerates the induction of a post-mitotic phenotype in human dermal fibroblasts: the key protective role of glutathione.
        Differentiation. 2000; 66: 147-155
        • Almond-Roesler B.
        • Schon M.
        • Schon M.P.
        • et al.
        Cultured dermal papilla cells of the rat vibrissa follicle. Proliferative activity, adhesion properties and reorganization of the extracellular matrix in vitro.
        Arch Dermatol Res. 1997; 289: 698-704
        • Bahta A.W.
        • Farjo N.
        • Farjo B.
        • et al.
        Premature senescence of balding dermal papilla cells in vitro is associated with p16(INK4a) expression.
        J Invest Dermatol. 2008; 128: 1088-1094
        • Chen Q.M.
        Replicative senescence and oxidant-induced premature senescence. Beyond the control of cell cycle checkpoints.
        Ann NY Acad Sci. 2000; 908: 111-125
        • Choudhry R.
        • Hodgins M.B.
        • Van der Kwast T.H.
        • et al.
        Localization of androgen receptors in human skin by immunohistochemistry: implications for the hormonal regulation of hair growth, sebaceous glands and sweat glands.
        J Endocrinol. 1992; 133: 467-475
        • Clerch L.B.
        • Iqbal J.
        • Massaro D.
        Perinatal rat lung catalase gene expression: influence of corticosteroid and hyperoxia.
        Am J Physiol. 1991; 260: 428-433
        • Cui Y.
        • Robertson J.
        • Maharaj S.
        • et al.
        Oxidative stress contributes to the induction and persistence of TGF-beta1 induced pulmonary fibrosis.
        Int J Biochem Cell Biol. 2011; 43: 1122-1133
        • Debacq-Chainiaux F.
        • Borlon C.
        • Pascal T.
        • et al.
        Repeated exposure of human skin fibroblasts to UVB at subcytotoxic level triggers premature senescence through the TGF-beta1 signaling pathway.
        J Cell Sci. 2005; 118: 743-758
        • Driskell R.R.
        • Clavel C.
        • Rendl M.
        • et al.
        Hair follicle dermal papilla cells at a glance.
        J Cell Sci. 2011; 124: 1179-1182
        • Ellis J.A.
        • Stebbing M.
        • Harrap S.B.
        Polymorphism of the androgen receptor gene is associated with male pattern baldness.
        J Invest Dermatol. 2001; 116: 452-455
        • Ellman G.
        • Lysko H.
        A precise method for the determination of whole blood and plasma sulfhydryl groups.
        Anal Biochem. 1979; 93: 98-102
        • Foitzik K.
        • Lindner G.
        • Mueller-Roever S.
        • et al.
        Control of murine hair follicle regression (catagen) by TGF-beta1 in vivo.
        FASEB J. 2000; 14: 752-760
        • Frippiat C.
        • Chen Q.M.
        • Zdanov S.
        • et al.
        Subcytotoxic H2O2 stress triggers a release of transforming growth factor-beta 1, which induces biomarkers of cellular senescence of human diploid fibroblasts.
        J Biol Chem. 2001; 276: 2531-2537
        • Gary R.K.
        • Kindell S.M.
        Quantitative assay of senescence-associated beta-galactosidase activity in mammalian cell extracts.
        Anal Biochem. 2005; 343: 329-334
        • Grayson W.L.
        • Zhao F.
        • Izadpanah R.
        • et al.
        Effects of hypoxia on human mesenchymal stem cell expansion and plasticity in 3D constructs.
        J Cell Physiol. 2006; 207: 331-339
        • Hamilton J.B.
        Patterned loss of hair in man; types and incidence.
        Ann NY Acad Sci. 1951; 53: 708-728
        • Heilmann S.
        • Kiefer A.K.
        • Fricker N.
        • et al.
        Androgenetic alopecia: identification of four genetic risk loci and evidence for the contribution of WNT-signaling to its etiology.
        J Invest Dermatol. 2013; 133: 1489-1496
        • Hibberts N.A.
        • Howell A.E.
        • Randall V.A.
        Balding hair follicle dermal papilla cells contain higher levels of androgen receptors than those from non-balding scalp.
        J Endocrinol. 1998; 156: 59-65
        • Hibino T.
        • Nishiyama T.
        Role of TGF-beta2 in the human hair cycle.
        J Dermatol Sci. 2004; 35: 9-18
        • Hoffmann R.
        • Seidl T.
        • Neeb M.
        • et al.
        Changes in gene expression profiles in developing B cells of murine bone marrow.
        Genome Res. 2002; 12: 98-111
        • Hunt D.P.
        • Morris P.N.
        • Sterling J.
        • et al.
        A highly enriched niche of precursor cells with neuronal and glial potential within the hair follicle dermal papilla of adult skin.
        Stem Cells. 2008; 26: 163-172
        • Inui S.
        • Fukuzato Y.
        • Nakajima T.
        • et al.
        Androgen-inducible TGF-beta1 from balding dermal papilla cells inhibits epithelial cell growth: a clue to understand paradoxical effects of androgen on human hair growth.
        FASEB J. 2002; 16: 1967-1969
        • Itahana K.
        • Dimri G.
        • Campisi J.
        Regulation of cellular senescence by p53.
        Eur J Biochem. 2001; 268: 2784-2791
        • Jacobs J.J.
        • de Lange T.
        Significant role for p16INK4a in p53-independent telomere-directed senescence.
        Curr Biol. 2004; 14: 2302-2308
        • Jahoda C.A.
        Cell movement in the hair follicle dermis - more than a two-way street?.
        J Invest Dermatol. 2003; 121: ix-xi
        • Jahoda C.A.
        • Horne K.A.
        • Oliver R.F.
        Induction of hair growth by implantation of cultured dermal papilla cells.
        Nature. 1984; 311: 560-562
        • Kaufman K.D.
        Androgen metabolism as it affects hair growth in androgenetic alopecia.
        Dermatol Clin. 1996; 14: 697-711
        • Kauser S.
        • Westgate G.E.
        • Green M.R.
        • et al.
        Human Hair Follicle and Epidermal Melanocytes Exhibit Striking Differences in Their Aging Profile which Involves Catalase.
        J Invest Dermatol. 2011; 131: 979-982
        • Koli K.
        • Saharinen J.
        • Hyytiainen M.
        • et al.
        Latency, activation, and binding proteins of TGF-beta.
        Microsc Res Tech. 2001; 52: 354-362
        • Kwack M.H.
        • Kim M.K.
        • Kim J.C.
        • et al.
        Dickkopf 1 promotes regression of hair follicles.
        J Invest Dermatol. 2012; 132: 1554-1560
        • Kwack M.H.
        • Sung Y.K.
        • Chung E.J.
        • et al.
        Dihydrotestosterone-inducible dickkopf 1 from balding dermal papilla cells causes apoptosis in follicular keratinocytes.
        J Invest Dermatol. 2008; 128: 262-269
        • Kyprianou N.
        • Isaacs J.T.
        Expression of transforming growth factor-beta in the rat ventral prostate during castration-induced programmed cell death.
        Mol Endocrinol. 1989; 3: 1515-1522
        • Lange C.A.
        • Gioeli D.
        • Hammes S.R.
        • et al.
        Integration of rapid signalling events with steroid hormone receptor action in breast and prostate cancer.
        Ann Rev Physiol. 2007; 69: 171-199
        • Maher P.
        The effects of stress and aging on glutathione metabolism.
        Ageing Res Rev. 2005; 4: 288-314
        • Mathieu J.
        • Zhou W.
        • Xing Y.
        • et al.
        Hypoxia-inducible factors have distinct and stage-specific roles during reprogramming of human cells to pluripotency.
        Cell Stem Cell. 2014; 14: 592-605
        • Miranda B.H.
        • Tobin D.J.
        • Sharpe D.T.
        • et al.
        Intermediate hair follicles: a new more clinically relevant model for hair growth investigations.
        Br J Dermatol. 2010; 163: 287-295
        • Mohyeldin A.
        • Garzon-Muvdi T.
        • Quinones-Hinojosa A.
        Oxygen in stem cell biology: a critical component of the stem cell niche.
        Cell Stem Cell. 2010; 7: 150-161
        • Norwood O.T.
        Male pattern baldness: classification and incidence.
        South Med J. 1975; 68: 1359-1365
        • Nyholt D.R.
        • Gillespie N.A.
        • Heath A.C.
        • et al.
        Genetic basis of male pattern baldness.
        J Invest Dermatol. 2003; 121: 1561-1564
        • Oliver R.F.
        Whisker growth after removal of the dermal papilla and lengths of follicle in the hooded rat.
        J Embryol Exp Morphol. 1966; 15: 331-347
        • Philpott M.P.
        • Sanders D.
        • Westgate G.E.
        • et al.
        Effects of insulin and insulin-like growth factors on cultured human hair follicles: IGF-I at physiologic concentrations is an important regulator of hair follicle growth in vitro.
        J Invest Dermatol. 1994; 102: 857-861
        • Pi M.
        • Parrill A.L.
        • Quarles L.D.
        GPRC6A mediates the non-genomic effects of steroids.
        J Biol Chem. 2010; 285: 39953-39964
        • Placencio V.R.
        • Sharif-Afshar A.R.
        • Li X.
        • et al.
        Stromal transforming growth factor-beta signaling mediates prostatic response to androgen ablation by paracrine Wnt activity.
        Cancer Res. 2008; 68: 4709-4718
        • Randall V.A.
        Hormonal regulation of hair follicles exhibits a biological paradox.
        Semin Cell Dev Biol. 2007; 18: 274-285
        • Randall V.A.
        • Thornton M.J.
        • Hamada K.
        • et al.
        Androgens and the hair follicle. Cultured human dermal papilla cells as a model system.
        Ann NY Acad Sci. 1991; 642: 355-375
        • Severi G.
        • Sinclair R.
        • Hopper J.L.
        • et al.
        Androgenetic alopecia in men aged 40-69 years: prevalence and risk factors.
        Br J Dermatol. 2003; 149: 1207-1213
        • Shin H.
        • Yoo H.G.
        • Inui S.
        • et al.
        Induction of transforming growth factor-beta 1 by androgen is mediated by reactive oxygen species in hair follicle dermal papilla cells.
        BMB Rep. 2013; 46: 460-464
        • Su L.H.
        • Chen T.H.
        Association of androgenetic alopecia with metabolic syndrome in men: a community-based survey.
        Br J Dermatol. 2010; 163: 371-377
        • Tobin D.J.
        • Gunin A.
        • Magerl M.
        • et al.
        Plasticity and cytokinetic dynamics of the hair follicle mesenchyme: implications for hair growth control.
        J Invest Dermatol. 2003; 120: 895-904
        • Trueb R.M.
        Association between smoking and hair loss: another opportunity for health education against smoking?.
        Dermatology. 2003; 206: 189-191
        • Trueb R.M.
        Is androgenetic alopecia a photoaggravated dermatosis?.
        Dermatology. 2003; 207: 343-348
        • Vayalil P.K.
        • Iles K.E.
        • Choi J.
        • et al.
        Glutathione suppresses TGF-beta-induced PAI-1 expression by inhibiting p38 and JNK MAPK and the binding of AP-1, SP-1, and Smad to the PAI-1 promoter.
        Am J Physiol Lung Cell Mol Physiol. 2007; 293: L1281-L1292
        • Wang W.
        Oxygen partial pressure in outer layers of skin: simulation using three-dimensional multilayered models.
        Microcirculation. 2005; 12: 195-207
        • Wood J.M.
        • Decker H.
        • Hartmann H.
        • et al.
        Senile hair graying: H2O2-mediated oxidative stress affects human hair color by blunting methionine sulfoxide repair.
        FASEB J. 2009; 23: 2065-2075
        • Yeh Y.H.
        • Kuo C.T.
        • Chan T.H.
        • et al.
        Transforming growth factor-beta and oxidative stress mediate tachycardia-induced cellular remodelling in cultured atrial-derived myocytes.
        Cardiovasc Res. 2011; 91: 62-70