Androgen Receptor Accelerates Premature Senescence of Human Dermal Papilla Cells in Association with DNA Damage

The dermal papilla, located in the hair follicle, expresses androgen receptor and plays an important role in hair growth. Androgen/Androgen receptor actions have been implicated in the pathogenesis of androgenetic alopecia, but the exact mechanism is not well known. Recent studies suggest that balding dermal papilla cells exhibit premature senescence, upregulation of p16INK4a, and nuclear expression of DNA damage markers. To investigate whether androgen/AR signaling influences the premature senescence of dermal papilla cells, we first compared frontal scalp dermal papilla cells of androgenetic alopecia patients with matched normal controls and observed that premature senescence is more prominent in the dermal papilla cells of androgenetic alopecia patients. Exposure of androgen induced premature senescence in dermal papilla cells from non-balding frontal and transitional zone of balding scalp follicles but not in beard follicles. Overexpression of the AR promoted androgen-induced premature senescence in association with p16INK4a upregulation, whereas knockdown of the androgen receptor diminished the effects of androgen. An analysis of γ-H2AX expression in response to androgen/androgen receptor signaling suggested that DNA damage contributes to androgen/androgen receptor-accelerated premature senescence. These results define androgen/androgen receptor signaling as an accelerator of premature senescence in dermal papilla cells and suggest that the androgen/androgen receptor-mediated DNA damage-p16INK4a axis is a potential therapeutic target in the treatment of androgenetic alopecia.


Introduction
Androgenetic alopecia (AGA) is an androgen-mediated disorder that causes hair thinning in a defined pattern [1,2]. The prevalence of AGA increases with aging, from 31% at age 40-55 years to 53% at age 65-69 years [3]. However, the pathogenetic mechanisms underlying AGA are not fully understood. The skin and the pilosebaceous unit are androgen target tissues. In dermal papilla cells (DPCs), the major circulating androgen, testosterone (T), can be locally metabolized to dihydrotestosterone (DHT) by steroid 5a-reductase. Based on its affinity for binding to the androgen receptor (AR), DHT is ten-fold more potent than testosterone [4]. According to the observation that male subjects with genetic deficiency of type 2 5a-reductase do not develop scalp hair loss [5], DHT has been suggested to be a major determinant in the pathophysiology of AGA in men. In AGA, DPCs from androgen-sensitive frontal scalp contain more AR and steroid 5a-reductase than those from androgen-insensitive occipital scalp [6]. AGA is characterized by miniaturization of hair follicles in the androgen-sensitive frontal scalp. The volume of dermal papilla (DP) depends on its number of DPCs and the amount of extracellular matrix, and correlates with the size of the hair fiber produced [7]. Whereas previous studies have shown that androgens do not alter the proliferation of DPCs [8,9]. In clinical observations, blocking the conversion of T to DHT with finasteride, a 5a reductase inhibitor, does not reverse miniaturized follicles to thick hair fibers in advanced AGA [10], suggesting that androgens/AR might irreversibly cause the damage of hair follicles.
Premature senescence is thought to accelerate the appearance of senescent phenotypes in cells upon exposure to sublethal stressors [11]. Several studies have shown that senescent cells are morphologically altered (enlarged and flattened) [12], and express more extracellular matrix-degrading proteases, collagenase, and matrix metalloproteinases [13,14]. In addition to its ability to deplete the renewal capacity of tissues by causing proliferative arrest, senescence may contribute to aging by influencing neighboring cells be means of secretory molecules thus disrupting the integrity and homeostasis of tissues [15][16][17]. A recent study indicated that balding DPCs undergo premature senescence in vitro in association with expression of senescence-associated bgalactosidase (SA b-gal) and p16 INK4a protein, and markers of oxidative and DNA damage [18]. It has been known that the DNA-damage response is a central mediator in triggering cellular senescence [19] and androgens act as DNA-damaging agents that generate DNA double-strained breaks (DSBs) and thus facilitate chromosomal translocation in androgen-sensitive prostate cancer cell lines [20]. However, how androgen/AR signaling leads to DNA damage in DPCs has not been elucidated.
To examine whether androgens may contribute to premature senescence by promoting DNA damage in DPCs of AGA patients, we cultured DPCs obtained from the frontal scalp of AGA and non-AGA individuals and determined their senescence phenotypes and investigated the effects of androgen/AR signaling on the development of premature senescence. We also studied the p16 INK4a protein and the relationship between androgen/AR signaling and DNA damage markers to elucidate the possible mechanisms of androgen/AR-accelerated premature senescence in DPCs.

Ethics Statements
The Chang Gung Medical Foundation institutional review board approved all described studies (protocol number 99-1933B). The study was conducted according to the Declaration of Helsinki Principles. Informed written consent was obtained from all patients.

Isolation and Culture of Human DPCs
Specimens were taken from beard, transitional zone of balding, balding or non-balding frontal scalps of ten males undergoing surgical excision of benign cutaneous tumors. The donors of beard specimen were aged 28 and 34 years, and the donor of transitional zone of balding scalp was aged 38 and 45 years. Four were AGA patients aged 20, 24, 27 and 40 years, and the others were agematched (20, 24, 27 and 40 years) non-AGA individuals. DPs were isolated from the bulbs of dissected hair follicles, transferred onto plastic dishes coated with 0.1% gelatin, and cultivated in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 IU/mL penicillin/100 mg/mL streptomycin and 0.4 mM L-glutamine (Sigma) in a humidified 95% atmosphere with 5% CO 2 at 37uC. DPCs from the second to sixth passages of subcultures were used.

SA b-gal Staining
The SA-b-Gal activity was determined by using a SA-b-Gal staining kit (Sigma). SA-b-Gal staining was detected in cultured DPCs seeded at a density of 10 5 cells in six-well plates and frozen human hair follicles in slide-mounted sections.

Measurement of Cell Size
Photomicrographs of DPCs were captured using a Nikon TE300 microscope equipped with a Nikon DSFi1 camera (Nikon Japan), and cell size was analyzed with Image-Pro Plus 6.1 software.

Cell Doubling Time Estimation
DPCs were trypsinized in exponential growth phase. Equal numbers of cells (1610 4 ) were plated in six-well culture plates (day 0) and cultured in 5% CO2 at 37uC. Viable cells were determined by the trypan blue exclusion test and counted in triplicate on days 1, 3, 5. Cell doubling time, DT, was estimated by the following formula: DT = (t1-t0 )log 2/(log N12log N0 ) where N1 and N0 are the number of cultured cells at the current (t1) and previous (t0) measurements.

Co-culture of DPCs and Keratinocytes (KCs)
Primary human hair follicular KCs were purchased and maintained in keratinocyte growth medium (KGM) (ScienCell research laboratories, California, USA). For the experiment, KCs in the third to fourth passage of subculture were used. DPCs in the second passage were treated with or without 0.1 mM of DHT in 10 cm culture dishes for 5 days. Then DPCs were seeded at a density of 5610 4 cells/well in the lower compartment of transwell multiplates (six-wells, Corning, NY, USA). At the same time, KCs (5610 4 cells/well) were also seeded with KGM into the transwell inserts with type I and III collagen-coated (24 mm diameter, 0.4 mm pore size, Corning). After incubation overnight, the medium was changed to MCDB153 (Sigma) without growth factors and bovine pituitary extract and DPCs and follicular KCs were cocultured and 0.1 mM of DHT or ethanol was added to these cultures. After they were incubated for 4 days, we counted the number of KCs in each well.

Transfection of DPCs
The AR expression plasmid, pcDNA3-hAR [21,22], pcDNA3-hERa [22] has been described previously. For overexpression of the AR or estrogen receptor a (ERa), DPCs were transiently transfected with pcDNA3-hAR or pcDNA3-hERa using the calcium phosphate method. AR expression was knocked down by infection with lentiviral vector pLKO.1-puro carrying a small hairpin RNA (shRNA) specific for AR-shRNA (59-CAC-CAATGTCAACTCCAGGAT-39). Reagents for RNA interference (RNAi) experiments were acquired from the National RNAi Core Facility at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica, Taiwan. Infected cells were selected by incubating with puromycin (2 mg/mL; Sigma) for 4 days. Cells infected with a scrambled shRNA (59-TCAGTTAAC-CACTTTTT-39) were used as a control.

Detection of Senescence Associated Heterochromatin Foci (SAHF)
DPCs were seeded onto eight-chamber slide at a density of 3610 3 cells per well. After fixing cells with cold methanol, SAHF were detected by staining with 0.13 mg/mL 49,6-diamidino-2phenylindol (DAPI) for 2 minutes at room temperature as described [26], and foci-positive DPCs were counted.

Statistical Analysis
All values are presented as means 6 standard deviations (SDs) of replicate samples, and experiments were repeated a minimum of three times. Differences between two groups were assessed using unpaired two-tailed Student's t test. In all statistical comparisons, P,0.05 was defined as significant. SPSS statistics software (Version 15.0) was used for all calculations.

Results
Balding DPCs from AGA Patients are more Senescent than Non-balding DPCs A previous study has shown that DPCs from balding (frontal, androgen sensitive) scalps of AGA patients display a more premature-senescence phenotype than DPCs from non-balding (occipital, androgen insensitive) scalps of AGA patients [18]. To further compare senescence phenotypes between balding and nonbalding DPCs isolated from androgen-sensitive frontal scalps, we harvested DPCs from frontal scalps of eight males (four AGA patients with frontal baldness and the others were age-matched normal individuals) and stained second-passage primary DPCs for SA-b-Gal activity. Balding DPCs isolated from AGA patients showed broader and polygon-shaped morphology (Fig. 1B) in contrast with the elongated, fibroblast-like appearance of DPCs isolated from the normal individuals (Fig. 1A). We also observed stronger positive SA-b-Gal staining and larger cell size in balding DPCs than non-balding DPCs of the same passage ( Fig. 1C and  D). The average of cell doubling time of balding DPCs was 56.3 hours compared to 32.5 hours of the non-balding DPCs. While the cell doubling time was variable among the same passage of balding DPCs from different AGA patients, the balding DPCs have relative longer cell doubling time compared to non-balding DPCs (Fig. 1E).

Androgen Treatment Promotes Senescence in Earlierpassage DPCs from Frontal Scalp
To determine whether the senescence phenotypes in DPCs of different passages were differentially affected by androgen treatment, non-balding DPCs isolated from frontal scalp of normal individuals were cultured in the absence or presence of 0.1 mM of DHT used in previous AGA studies [27,28]. We found that non-balding DPCs of various passages exhibited increased SA-b-Gal activity compared with vehicle (ethanol)-treated cells ( Fig. 2A). Quantification of these results revealed a significant increase in the percentage of SA-b-Gal-positive cells after DHT stimulation (Fig. 2B). Moreover, cells exposed to DHT were increased in size and had broader morphology. DHT-induced premature senescence was more significant at earlier passages: in second-and third-passage cells, the premature senescence phenotype, including positive-SA-b-Gal staining and cell-size enlargement, both reached statistical significance (Fig. 2B, C). Although more DPCs exhibited senescence phenotypes and induction of premature senescence by DHT persisted in passages four to six, statistical significance was gradually lost. To further investigate the linkage between the DHT-accelerated premature senescence of DPCs and the pathogenesis of AGA, we compared the DHT effects on premature senescence of DPCs isolated from transitional zones of balding scalp, beard, and androgenunresponsive human prostate cancer cell line, DU145. We found that DHT induced premature senescence significantly in the DPCs from both transitional zone of balding scalp and non-balding frontal scalp. In contrast, neither beard DPCs (passage 2) nor DU145 cells became senescent after exposed to DHT 0.1 mM (Fig. 2D, E and F). To access the functional defect of DHTmediated senescent DPCs in the interaction with hair follicular KCs, we employed an in vitro co-culture system [29]. DPCs was first cultured in the presence or absence of 0.1 mM of DHT for 5 days to induce premature senescence and then co-cultured with follicular KCs for 4 days in the presence or absence of DHT. As shown in Fig. 2G, the growth of KCs cocultured with DHTpretreated DPCs was decreased. Together, we suggest that DHTaccelerated premature senescence is a specific action in earlierpassage DPCs from frontal scalp and DHT-mediated senescent DPCs may have the functional defect to communicate with KCs and play an important role in AGA pathophysiology.

AR is Required for Androgen-induced Premature Senescence in DPCs
To further confirm the role of AR expression in androgenaccelerated premature senescence of DPCs and mimic balding DPCs, which contain higher levels of AR, we manipulated AR expression levels in DPCs. Non-balding DPCs isolated from the frontal scalp were transfected with an AR expression plasmid or vector control in the absence or presence of 0.1 mM of DHT. Premature senescence of DPCs was analyzed by staining for SA-b-Gal and measuring cell size. Overexpression of the AR increased both percentage of SA-b-Gal-positive cells and cell size and DHT enhanced the AR effects ( Fig. 3A-C). To firmly establish the relationship of the AR with the androgen-induced senescence phenotype of DPCs, we examined SAHF, a nuclear marker of senescence characterized by punctuate intranuclear foci in DAPIstained cells. A quantitative analysis showed that DHT induced SAHF formation in DPCs, and overexpression of the AR reinforced the effects of DHT compared with vector controltransfected cells (Fig. 3D). To gain insight into the mechanism of premature senescence induction in DPCs, we focused on the relationship between the AR and p16 INK4a protein, which is known to be involved in premature senescence and is upregulated in balding DPCs [18]. Although we saw an increase in DHT-induced p16 protein expression in DPCs isolated from the frontal scalp, quantification of these data showed that the effect of DHT on p16 expression was not significantly different and only p16 levels in DPCs with AR overexpression compared with empty vector cell were statistically significant (Fig. 3E, F). The p16 INK4a protein is also up-regulated in DHT-treated DPCs isolated from transitional zone of balding scalp but unaltered in DPCs isolated from beard (Fig. 3G). This indicates that androgens/AR signaling is functionally linked to premature senescence that is region-specific phenomenon (androgen sensitive scalp) in DPCs and associated with the pathogenesis of AGA. One of the important considerations of premature senescence in DPCs is whether this is a specific androgen/AR action. Thus, ERa was employed as a negative control in these experiments. Increasing percentage of SA-b-Galpositive cells, cell size, SAHF formation and induction of p16 INK4a protein levels were not observed in DPCs overexpressing ERa with 0.01 mM of 17b-estradiol stimulation (Supplemental Figure 1).
To further confirm the role of AR in the induction of premature senescence in DPCs, we knocked down the AR with a lentivirus expressing an AR-specific shRNA and evaluated the senescencepromoting effect of DHT. SA-b-Gal activity assays revealed that knockdown of AR expression led to suppression of DHT-induced premature senescence. AR-knockdown DPCs were more resistant to morphological alteration under androgen-stimulation conditions (Fig. 4A). A quantitative analysis showed that knockdown of AR expression led to suppression of DHT-induced SA-b-Gal activity, premature senescence cell size and the percentage of SAHF-containing cells (Fig. 4B, C and D). Similar results were observed as the expression of p16 protein is significantly decreased in AR-knockdown DPCs, compared with the control cells regardless to DHT treatment (Fig. 4E, F). Taken together, these results show that androgen-induced premature senescence was significantly augmented in DPCs overexpressing AR, and diminished in AR-knockdown cells.

Androgen/AR Action Leads to DNA Damage in DPCs
To investigate the DNA damage status after androgen treatment, we transiently expressed the AR in DPCs and employed c-H2AX as a sensor of double-strained DNA breaks (DSBs). A subsequent immunofluorescence analysis revealed a marked increase in c-H2AX foci in the nuclei of DPCs after exposure to DHT (Fig. 5A, B). A further increase in c-H2AX foci (i.e., DSBs) was detected in DPCs stably transfected with the AR upon androgen stimulation (Fig. 5A, B). Moreover, Western blot analyses showed that H2AX had been converted to its phosphorylated form (c-H2AX) in DPCs in response to DHT, and the c-H2AX/total H2AX ratio was further increased in cells overexpressing the AR (Fig. 5C,D).

Discussion
Clinically, patterned hair loss has been reported in non-AGA individuals with iatrogenic androgen stimulation [30], and in patients suffering from diseases characterized by excessive androgen, such as polycystic ovary syndrome and androgensecreting tumors [31]. It is known that balding DPCs already showed varying degrees of premature senescence in early passages [18]. Moreover, DPCs isolated from non-balding areas of normal individuals have been demonstrated to be androgen-responsive and have been used as a cell model of AGA [32]. AR expression level is a crucial factor in determining DPC sensitivity to androgens in AGA [33]. It is known that AR mRNA [34,35] and protein [36] are expressed in DPCs. In addition, DPCs isolated from balding area express more AR compared to those in non-balding areas [6,37]. AR expression is stronger in earlypassage DPCs and gradually decreasing during subcultivation [28]. Our results showed that earlier-passage DPCs with higher AR expression were more sensitive to androgen-mediated premature senescence. It is well known that primary DPCs spontaneously undergo replicative senescence during subcultivation. Indeed, we also observed that late-passage DPCs were more senescent (Fig. 2). Loss of responses to androgen-induced senescence in late-passage DPCs could reflect relatively lower AR expression in these cells as well as masking of androgen effects by replicative senescence. In our study, we also demonstrated that androgen-accelerated premature senescence was affected by AR expression level (Fig. 3 and 4). We thus concluded that blocking androgen/AR actions could play an important role in suppressing premature senescence in DPCs.
In prostate cancer cell lines, androgen signaling has been reported to induce recruitment of the AR-topoisomerase II beta (TOP2B) complex, which catalyzes DSBs at regulatory regions of AR target genes [20]. AR also acts in concert with genotoxic stress to induce alterations in local chromatin architecture. These events are permissive for sensitizing these regions to undergo chromosomal translocation through activation of ORF2 endonuclease [38]. In our study, DSBs were also induced in response to androgen/AR signaling, and c-H2AX foci and expression levels of c-H2AX proteins were further increased with AR overexpression ( Figure 5). These results support previous findings that two important DNA damage sensors involved in the phosphorylation of H2AX-the active form of ATM (ataxia-telangiectasia-mutated kinase) and ATR (ATM and Rad3-related)-were detected only in balding DPCs [18]. Although much of this DNA damage can be repaired and the cell can then re-enter the cell cycle, some of the aberrantly enhanced DSBs might destabilize the genome and potentially trigger premature senescence in DPCs. The roles of TOP2B and ORF2 in DNA damage leading to premature senescence of DPCs need further investigation.
The effects of androgen/AR signaling on senescence in prostate cancer cells remain a matter of controversy [39,40]. It has been reported that androgen depletion induces senescence in prostate cancer cells via down-regulation of Skp2 [39,40]. In contrast, androgen/AR signaling has also been reported to drive cellular senescence without the involvement of DNA damage and p16 INK4a upregulation in both prostate cancer and normal immortal prostate epithelial cell lines [39]. In DPCs, we found that p16 INK4a protein levels were upregulated in response to androgen, and AR overexpression may further enhance the expression level of p16 INK4a ( Figure 3E). These results suggest that androgen/AR signaling promotes senescence through the p16 INK4a pathway in DPCs and are in agreement with the results of a previous study, which showed increased expression of p16 INK4a in balding DPCs from AGA patients with premature senescence [18]. The discrepant reports of androgen/AR actions in senescence could reflect the diversity of biological responses to androgen/AR signaling in different cell types. While these latter studies utilized the prostate cancer cell lines, PC3 and PC3-AR, and immortalized normal prostate RWPE-1 cells as experimental models, it is well known that PC3 cells are AR-, p16 INK4a -and p53-null [39], and RWPE-1 cells are p53-and Rb-null [41]. Therefore, the senescence response and the pathway that mediates it in these cells might be different from that in primary DPCs. The signaling pathways activated by DNA damage converge on p53/p21 pathway-mediated replicative senescence caused by telomere shortening, and the p16 pathway is thought to mediate premature senescence [19]. Expression of p16 INK4a is induced by numerous stressors, including oxidative stress [42], overexpression of oncogenes [43,44], and DNA damage [45,46]. Nuclear expression of oxidative stress and DNA damage markers has been reported in balding DPCs [18]. Increased DSBs and upregulation of p16 INK4a in response to androgen/AR signaling suggest that DNA damage might be important in the androgen-accelerated premature senescence of DPCs, although we cannot exclude the possibility that oxidative stress is also induced by androgen/AR signaling. Androgen inducible molecules, such as Interleukin-6 (IL-6) and transforming growth factor (TGF)-b1 have been shown to inhibit hair growth in paracrine manner [27,47] and could be contrib-uting factors of cellular senescence. Whether the androgen/ARaccelerated premature senescence of DPCs is also mediated via autocrine manner by androgen inducible IL-6 and TGF-b1 needs more investigation.
The previous studies showed low passage DPCs could sustain epidermal cell proliferation; however, high passage DPCs could not [48]. In addition, DPCs after multiple passaging also reduced hair growth-promoting capabilities in vivo [30,49]. These evidence supports the senescent DPCs may have functional defect on promoting epithelial-mesenchymal interactions. It has been shown that androgen/AR regulates the interaction of DPCs and follicular KCs by androgen-inducible factors secreted from DPCs [27,28,47,50]. Recently, an impairment of hair follicle stem cells to differentiate into progenitor cells was reported to play a key role in the pathogenesis of AGA [51]. Miniaturization of hair follicles, the hallmark of AGA, displays thinner hair fibers and smaller DP size. It is possible that the androgen/AR-induced senescence in DPCs may not only lead to diminished DP size but also deregulate the communication between DPCs and hair follicle stem cells to differentiate to progenitor cells.
Here, we showed a previously unidentified relationship between androgen/AR signaling and induction of premature senescence in association with DNA damage and p16 INK4a upregulation in DPCs. Our study highlights the importance of androgen/ARaccelerated premature senescence in DPCs, a process that is thought to reflect irreversible cell growth arrest in the progression of AGA. The acceleration of premature senescence of DPCs by androgen/AR signaling may explain the miniaturization of hair follicles shown in AGA patients. These results provide novel impacts of androgen/AR signaling in balding DPCs and offer the potential therapeutic targets on combating for AGA. Figure S1 Estrogen/ERa signaling did not cause premature senescence in DPCs. (A) Non-balding DPCs of frontal scalp were transfected with pcDNA3-hERa or pcDNA3 empty vector and cultured in the presence of 0.01 mM of 17b-estradiol or ethanol (vehicle control) for 3 days. Premature senescence of DPCs was evaluated on day 5. Scale bar = 100 mm. SA-b-Gal activity (B), cell size (C), and the number of SAHF-containing DPCs (D) were unaltered. Values are means 6 SDs from three independent experiments (E) A representative immunoblot of cell lysates of DPCs after treatment with 17b-estradiol or vehicle for 84 hours. The numbers indicate p16 INK4a /GAPDH and ERa/GAPDH ratios. GAPDH (glyceraldehyde 3-phosphate dehydrogenase) was used as an internal standard. (TIF)