Biochemical Journal

Research article

Ligand-activated PPARδ inhibits UVB-induced senescence of human keratinocytes via PTEN-mediated inhibition of superoxide production

Sun Ah Ham, Jung Seok Hwang, Taesik Yoo, Hanna Lee, Eun Sil Kang, Chankyu Park, Jae-Wook Oh, Hoon Taek Lee, Gyesik Min, Jin-Hoi Kim, Han Geuk Seo

Abstract

UV radiation-mediated photodamage to the skin has been implicated in premature aging and photoaging-related skin cancer and melanoma. Little is known about the cellular events that underlie premature senescence, or how to impede these events. In the present study we demonstrate that PPARδ (peroxisome-proliferator-activated receptor δ) regulates UVB-induced premature senescence of normal keratinocytes. Activation of PPARδ by GW501516, a specific ligand of PPARδ, significantly attenuated UVB-mediated generation of ROS (reactive oxygen species) and suppressed senescence of human keratinocytes. Ligand-activated PPARδ up-regulated the expression of PTEN (phosphatase and tensin homologue deleted on chromosome 10) and suppressed the PI3K (phosphatidylinositol 3-kinase)/Akt pathway. Concomitantly, translocation of Rac1 to the plasma membrane, which leads to the activation of NADPH oxidases and generation of ROS, was significantly attenuated. siRNA (small interfering RNA)-mediated knockdown of PTEN abrogated the effects of PPARδ on cellular senescence, on PI3K/Akt/Rac1 signalling and on generation of ROS in keratinocytes exposed to UVB. Finally, when HR-1 hairless mice were treated with GW501516 before exposure to UVB, the number of senescent cells in the skin was significantly reduced. Thus ligand-activated PPARδ confers resistance to UVB-induced cellular senescence by up-regulating PTEN and thereby modulating PI3K/Akt/Rac1 signalling to reduce ROS generation in keratinocytes.

  • peroxisome-proliferator-activated receptor δ (PPARδ)
  • phosphatase and tensin homologue deleted on chromosome 10 (PTEN)
  • phosphatidylinositol 3-kinase (PI3K)
  • senescence
  • reactive oxygen species (ROS)
  • UVB

INTRODUCTION

Cellular senescence is the permanent and irreversible growth arrest of cells cultured in vitro, which is characterized by phenotypic changes in gene expression, function and morphology [1,2]. Primary cultured cells undergo replicative senescence, which is characterized by a shortened telomere length that eventually results in incomplete chromosomal replication via telomeric fusion or loss of telomere-bound factors [3]. In contrast, SIPS (stress-induced premature senescence) is triggered by factors that cause cellular stress, such as UV radiation [4], chemical agents that induce DNA damage [5,6] and inappropriate oncogene activation [7].

Solar UV radiation causes a range of acute and chronic damage, including sunburn, immune suppression, premature senescence (photo-aging) and skin cancer, depending on the strength and duration of exposure [8]. Acute sunburn and immune suppression occur in response to excessive exposure to the sun, whereas skin cancer and premature senescence result from the accumulated damage caused by repeated exposures over a long period of time [8]. UV radiation consists of three components, UVA, UVB and UVC. Whereas UVA and UVB reach the earth in sufficient amounts to damage the skin, UVC is almost completely absorbed by the ozone layer [9]. UVB is particularly damaging, as it penetrates the epidermis and the upper part of the dermis, where it damages keratinocytes and leads to sunburn, photo-aging and skin cancer [4,8].

Several lines of evidence suggest that ROS (reactive oxygen species) may play (an) important role(s) in mediating UV-induced cellular senescence. ROS break DNA strands and modify bases of DNA to elicit both replicative senescence and SIPS [2,10]. Two-photon fluorescence imaging and biochemical assays detect increased levels of ROS in UVB-irradiated skin [11,12]. UVB radiation of keratinocytes in vitro also stimulated a dose-dependent increase in the intracellular production of hydrogen peroxide [13], suggesting a role for UVB in cellular senescence. In addition, NADPH oxidase has been implicated in UV-dependent ROS generation [14,15], indicating that NADPH oxidase may also take part in cellular senescence of the skin.

PPARδ (peroxisome-proliferator-activated receptor δ), a member of the PPAR nuclear receptor family, is a ligand-inducible transcription factor that modulates multiple biological functions in energy metabolism and skin homoeostasis [16,17]. This receptor heterodimerizes with retinoid X receptor and regulates gene expression via PPREs (PPAR-response elements) located in the regulatory regions of its target genes [16,17]. PPARδ is ubiquitously expressed in a variety of cell lineages, including keratinocytes. It has been postulated that ligand-activated PPARδ exerts anti-tumorigenic effects by modulating differentiation and proliferation in the skin [18,19]. PPARδ also appears to be a key mediator of the epidermal effects in the wound-healing process, as it converts the extracellular inflammatory signal into an organized pattern of gene expression, which eventually leads to the survival, migration and differentiation of keratinocytes [17,20]. The co-ordinated interplay between PPARδ and TGF (transforming growth factor)-β1 accelerates wound closure, with TGF-β1 acting as a regulator of inflammation-induced PPARδ expression [21]. Furthermore, we recently demonstrated that ligand-activated PPARδ promotes wound healing by up-regulating TGF-β1-mediated expression of the extracellular matrix proteins [22]. On the basis of its beneficial properties, including acceleration of wound healing and regulation of inflammatory responses [20,21], PPARδ is a promising new target for the treatment of skin disorders. However, the role of PPARδ in photoaging has not been exploited as a therapeutic strategy.

We and others have shown that PPARδ regulates skin homoeostasis by modulating inflammatory responses and extracellular matrix expression [20,21]. A senescent cell phenotype has been identified in human keratinocytes and fibroblasts exposed to UVB radiation [4,23]. Accordingly, a question was raised as to whether PPARδ induces anti-senescent responses in skin cells exposed to UVB. In the present study we demonstrate that ligand-activated PPARδ prevents premature senescence and reduces the generation of ROS in both cultured human keratinocytes and in vivo mouse skin. Interestingly, up-regulation of PTEN (phosphatase and tensin homologue deleted on chromosome 10) is involved in this process.

MATERIALS AND METHODS

Materials

GW501516, CM-H2DCF-DA (chloromethyl-2′,7′-dichlorofluorescein diacetate), and LY294002 [2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one] were obtained from Calbiochem. NAC (N-acetylcysteine) and DPI (diphenyleneiodonium) were from Sigma–Aldrich.

Cell culture and UVB irradiation

Human primary keratinocytes (donor age 11, passage 3) were obtained from Welskin (Seoul, Korea). The cells were cultured in KGM (keratinocyte growth medium) containing keratinocyte growth supplement at 37°C in an atmosphere of 95% air and 5% CO2, on the basis of the manufacturer's recommendations. The cells were incubated for the indicated times in the presence or absence of the indicated reagents, and then washed twice with PBS before UVB exposure to avoid the photosensitization effect of components in culture medium. UVB irradiation was performed on serum-starved monolayer cultures using an FS20 Lamp (National Biological). The UVB source was a bank of two FS20 lamps emitting a continuous spectrum from 270 to 390 nm, with a peak emission at 313 nm; approximately 65% of the radiation was within the UVB (280–320 nm) wavelength. UVB strength was monitored using a Goldilux model 70234 photometer (Lynntech).

Senescence-associated β-galactosidase staining

Senescent cells were detected using a SA β-gal (senescence-associated β-galactosidase) staining kit, according to the manufacturer's instructions (Sigma–Aldrich). Briefly, human keratinocytes pretreated with or without the indicated reagents for 30 min were incubated in the presence or absence of GW501516 for 24 h, followed by exposure to UVB irradiation in 6-mm-diameter dishes. The cells were washed with ice-cold PBS and fixed in fixation solution for 7 min at room temperature (25°C). After washing twice with ice-cold PBS, the cells were incubated with staining solution (5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, 2 mM MgCl2, 40 mM citric acid, 40 mM sodium phosphate and 1 mg/ml X-Gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside, pH 6.0) at 37°C in the absence of CO2 overnight. Senescent cells were visualized using an Olympus JP/1×71 fluorescence microscope. For detection of senescent cells in the dorsal skin of UVB-irradiated hairless mice, serial cryosections (7 μm) mounted on gelatin-coated slides were incubated in staining solution and SA β-gal was visualized as described above.

siRNA (small interfering RNA)-mediated gene silencing

Cells were seeded into 100-mm-diameter culture dishes 18–24 h prior to transfection. Cells were transfected with control siRNA (Ambion), human PPARδ siRNA (Ambion) or human PTEN siRNA (Cell Signaling Technology) in serum-free medium using Welfect-Q (WelGENE). Following incubation for 6 h, the transfection medium was replaced with fresh medium, and the cells were allowed to incubate for an additional 48 h, at which point they were treated with the indicated reagents for the indicated periods of time.

Immunoblot analysis

Cells treated with the indicated reagents were washed in ice-cold PBS and lysed in PRO-PREP Protein Extraction Solution (iNtRON Biotechnology). An aliquot of the cell lysate was subjected to SDS/PAGE and transferred on to a Hybond-P+ PVDF membrane (GE Healthcare). Membranes blocked overnight at 4°C with 5% (w/v) non-fat dried skimmed milk powder in TBS (Tris-buffered saline; 50 mM Tris/HCl, 150 mM NaCl, pH 7.6) containing 0.1% Tween 20 were allowed to react overnight at 4°C with the indicated specific antibodies in TBS containing 1% BSA and 0.05% Tween 20, and then incubated with horseradish-peroxidase-conjugated goat antibody diluted to 1:3000 for 2 h at room temperature. After extensive washing in TBS containing 0.1% BSA and 0.1% Tween 20, immunoreactive bands were detected using West-ZOL Plus (iNtRON Biotechnology). For analysis of protein expression in the dorsal skin of UVB-irradiated hairless mice, the excised tissues were homogenized using the FastPrep-24 instrument with ceramic spheres (MP Biomedicals). After centrifugation at 15000 g for 10 min, an aliquot of the protein was subjected to immunoblot analysis, as described above. Polyclonal antibodies specific for p53, p21, phospho-Akt and Akt were obtained from Cell Signaling Technology. Polyclonal antibodies specific for PPARδ and Rac1, monoclonal antibodies specific for p16, PML (promyelocytic leukaemia protein), PTEN and horseradish peroxidase-conjugated IgG were purchased from Santa Cruz Biotechnology. Polyclonal rabbit anti-β-actin antibody was purchased from Sigma–Aldrich.

RNA blot analysis

Aliquots of 5 μg of total RNA were heat-denatured at 65°C for 15 min in gel-running buffer (40 mM Mops, 10 mM sodium acetate and 1 mM EDTA, pH 7.0) containing 50% formamide and then subjected to electrophoresis on a 1% agarose gel containing 2.2 M formaldehyde. Size-fractionated RNA was transferred overnight on to a Hybond-N+ nylon membrane (GE Healthcare) by capillary action and then hybridized with the 32P-labelled PTEN cDNA probe at 68°C in QuikHyb solution (Stratagene). The membrane was washed, and the radioactivity on the membrane was detected using a Fuji BAS-2500 Bioimaging Analyzer. The blots were stripped and rehybridized with a 32P-labelled GAPDH (glyceraldehyde-3-phosphate dehydrogenase) cDNA probe. The cDNA probe was generated by PCR using primers that were specific for nucleotides 1083–1518 of human PTEN.

Measurement of intracellular ROS

To assess the levels of intracellular accumulation of ROS, the fluorescent probe CM-H2DCF-DA (Molecular Probes) was used. Cells were seeded on to 35-mm-diameter coverglass bottom dishes (SPL Life Sciences) and transfected with the siRNA molecules described above. After incubation for 48 h, the cells were treated with 10 nM GW501516 or DMSO for 24 h, and then exposed to 40 mJ/cm2 UVB irradiation for 30 min. The cells were incubated with 10 μM CM-H2DCF-DA for 30 min at 37°C, and the green fluorescence corresponding to the levels of intracellular ROS was detected through a 520-nm long-pass filter using an Olympus FV-1000 laser fluorescence microscope.

Confocal immunofluorescence microscopy

Cells seeded on 35-mm-diameter coverglass bottom dishes were pre-treated with 10 μM LY294002 for 30 min or transfected with the indicated siRNA molecules for 48 h. The cells were treated with 10 nM GW501516 or vehicle (DMSO) for 24 h, and then exposed to 40 mJ/cm2 UVB radiation. After incubation for 30 min, the cells were fixed using neutral buffered solution (4% formaldehyde) for 7 min. After permeabilization with PBS containing 0.1% Tween 20 for 3 min, the fixed cells were stained overnight at 4°C using anti-γ-H2A.X (Cell Signaling Technology), anti-p53, anti-p21, anti-p16, anti-PML or anti-Rac1 antibody (Santa Cruz Biotechnology) at a dilution of 1:200. Secondary goat anti-rabbit IgG conjugated to Cy3 (indocarbocyanine; Invitrogen) or Alexa Fluor® 488-conjugated goat anti-mouse IgG (Invitrogen) was applied for 2 h at room temperature. For immunofluorescent histochemistry, tissues were boiled in 10 mM sodium citrate buffer (pH 6.0) for 10 min to retrieve antigens. Each section that was incubated with primary antibody overnight at 4°C was incubated with Alexa Fluor® 488-conjugated secondary antibody for 2 h at room temperature. Confocal imaging analysis was performed using an Olympus FV-1000 confocal laser fluorescence microscope.

Rac1 pull-down assays

Levels of GTP-bound Rac1 were determined using the Rac1 Activation Assay Kit (Upstate Biotechnology) according to the manufacturer's instructions. Briefly, cells pre-treated with LY294002 for 30 min in 100-mm-diameter culture dishes were treated with GW501516 for 24 h and exposed to UVB radiation for 30 min. After washing twice with MLB [Mg2+ lysis/wash buffer; 125 mM Hepes, pH 7.5, 750 mM NaCl, 5% Igepal CA-630, 50 mM MgCl2, 5 mM EDTA and 10% (v/v) glycerol], cell lysates were pulled down using the Rac-binding domain of PAK (p21-activated kinase 1) linked to glutathione–agarose beads for 6 h at 4°C, washed, and then eluted with SDS sample buffer. GTP-bound Rac1 was analysed by immunoblot using monoclonal anti-Rac1 antibody.

Measurement of PtdIns(3,4,5)P3 levels

PtdIns(3,4,5)P3 levels were quantified using a PtdIns(3,4,5)P3 mass strip assay kit according to the manufacturer's protocol (Echelon Biosciences). Briefly, human keratinocytes plated in 100-mm-diameter culture dishes were transfected with 200 pM human PTEN siRNA for 48 h and then pre-treated with GW501516. After incubation for 24 h, cells were exposed to UVB radiation for 90 min. Cells were collected using cold 0.5 M TCA (trichloroacetic acid) and washed with 5% (v/v) TCA solution containing 1 mM EDTA. After extraction of neutral lipids with methanol/chloroform (2:1, v/v), acidic lipids were extracted with chloroform/methanol/12 M HCl (40:80:1, by vol.) and vacuum dried. Samples were dissolved in chloroform and spotted on to nitrocellulose membrane. After blocking in PBS containing 3% (w/v) fatty acid-free BSA, chemiluminescent signals corresponding to PtdIns(3,4,5)P3 levels were detected.

Animal study

All animal studies were approved by the Institutional Animal Care Committee of Konkuk University. Six-week-old hairless mice (HR-1) were obtained from Japan SLC Inc. and maintained under controlled environmental conditions with a 12 h/12 h light/dark cycle. The mice received food and tap water ad libitum. Mice were intraperitoneally injected with GW501516 (10 mg/kg body weight) in DMSO or vehicle (DMSO) and exposed to UVB radiation 24 h later. Control mice received DMSO only. UVB irradiation (100 J/m2) was performed as described previously [24]. At 24 h later, mice were killed and squares of dorsal skin (approximately 1 mm×1 mm) were excised from each mouse and cut in half for histological examination and protein analysis. For H&E (haematoxylin and eosin) counterstaining, following fixation with 4% (w/v) paraformaldehyde and cryoprotection in 20% (w/v) sucrose, tissues were embedded in OCT™ (optimal cutting temperature; Sakura Finetech) and snap-frozen in 2-methylbutane, prechilled in liquid nitrogen. Serial cryosections (7 μm) were incubated sequentially at room temperature in 50% haematoxylin solution for 20 min and then differentiated in 1% HCl solution for 1 s. The sections were transferred directly to 0.5% eosin solution and stained for 1 min. The sections were then washed once with distilled water and incubated in a graded alcohol series ending with xylene to dehydrate the tissue. Finally, the tissue sections were coverslipped with Permount (Fisher Scientific).

Assay for intracellular superoxide

Intracellular superoxide production was measured using a fluorescent indicator, DHE (dihydroethidium), as described previously [25]. Briefly, serial cryosections (7 μm) of the frozen tissue samples obtained from the dorsal skin of the UVB-irradiated HR-1 hairless mice were mounted on to gelatin-coated slides, washed with ice-cold PBS, and incubated for 30 min at 37°C with 10 μM DHE in PBS. Following incubation in a humidified chamber protected from light, the red fluorescence signal corresponding to the levels of intracellular superoxide was detected through a 580-nm long-pass filter using an Olympus FV-1000 fluorescence microscope.

Statistics

Data are expressed as means±S.E.M. Statistical significance was determined by Student's t test or ANOVA with a post-hoc Bonferroni test. A value of P<0.05 was considered statistically significant.

RESULTS

Ligand-activated PPARδ inhibits UVB-induced senescence in keratinocytes

Since cellular senescence results from the accumulated damage caused by repeated exposure to UV light [26], we examined whether ligand-activated PPARδ affects the premature senescence of cultured human keratinocytes exposed to UVB. Keratinocytes exposed to increasing doses of UVB radiation exhibited significant increases (of 6- to 8-fold) in SA β-gal activity, a biomarker for cellular senescence, relative to the unexposed control cells. Similar results were obtained in immunofluorescent histochemical analyses using p53, p21, p16 or PML, which are also known biomarkers of cellular senescence (Figures 1A and 1B). This increase, however, was significantly suppressed in the presence of GW501516, suggesting the involvement of PPARδ in the inhibition of UVB-induced premature senescence (Figures 1A–1C and Supplementary Figure S1 at http://www.BiochemJ.org/bj/444/bj4440027add.htm). Furthermore, pretreatment with NAC, a thiol antioxidant, caused a similar reduction, as did GW501516 in UVB-induced SA β-gal activity (Figure 1D and Supplementary Figure S2 at http://www.BiochemJ.org/bj/444/bj4440027add.htm). These findings suggest that ROS mediates UVB-induced senescence in human keratinocytes.

Figure 1 Ligand-activated PPARδ inhibits UVB-induced senescence in human keratinocytes

(AD) Detection of senescent cells. Cells pre-treated for 24 h with (+) or without (−) GW501516 were exposed to 0, 40, 80 or 100 mJ/cm2 UVB irradiation. After incubation for 24 h, SA β-gal staining and fluorescent immunoblot analyses were performed (A) and SA β-gal, p53, p21, p16 or PML-positive cells were quantified (B). In (A), scale bars=100 μm; in (B), open bars, control; closed bars, GW501516-treated groups. (C, D and E) Cells pre-treated with increasing doses (0, 0.1, 0.5, 1, 5 and 10 nM) of GW501516 (C) or NAC (D) for 24 h were exposed to UVB. Following incubation for 24 h, SA β-gal staining was performed and quantified. (E) Expression of p53 and p21. Whole-cell lysates of keratinocytes pre-treated and irradiated as above were subjected to immunoblot analyses. Representative images from four independent experiments are shown. Results are expressed as the means±S.E.M. (n=4). *P<0.01 compared with untreated group; #P<0.01 and ##P<0.05 compared with UVB-irradiated vehicle group.

To characterize the molecular mechanisms whereby GW501516 modulates cellular senescence, we examined the expression levels of p53 and p21, key proteins in the senescence pathway [27]. As shown in Figure 1(E), a marked increase in the levels of p53 and p21 was observed in keratinocytes exposed to UVB, whereas pre-treatment with either GW501516 or NAC diminished the effect of UVB on these proteins. Furthermore, combined treatment with both GW501516 and NAC almost completely abolished the expression of these proteins, suggesting the involvement of ROS and a modulatory role of PPARδ in the premature senescence of keratinocytes induced by UVB radiation.

To verify the role of PPARδ in blocking UVB-induced senescence of keratinocytes, we examined the effect of GW501516 on cells treated with a siRNA against PPARδ. The level of PPARδ in keratinocytes was markedly reduced upon transfection with PPARδ siRNA in the presence of GW501516 and/or UVB radiation, whereas control siRNA, consisting of a pool of non-specific sequences, had no effect on PPARδ levels (Figure 2A). As expected, the siRNA-mediated down-regulation of PPARδ significantly suppressed the GW501516-mediated inhibition of premature senescence and γ-H2A.X foci of keratinocytes induced by UVB radiation (Figures 2B–2D).

Figure 2 siRNA against PPARδ abolished the effect of GW501516 on UVB-induced cellular senescence in human keratinocytes

(A) Expression of PPARδ in cells transfected with siRNA against PPARδ or control siRNA. Transfected cells incubated for 24 h in the presence (+) or absence (−) of GW501516 were exposed to UVB and incubated for an additional 24 h. (B) Percentage of SA β-gal-positive cells. Cells transfected with siRNA against PPARδ or control siRNA were incubated with GW501516 for 24 h. After exposure to UVB, the cells were further incubated for 24 h. (C and D) Assessment of γ-H2A.X foci. Cells transfected with the indicated siRNA were incubated with GW501516 for 24 h and then exposed to UVB irradiation. After incubation for 24 h, images are taken from four independent experiments and representative images are shown (C). Scale bars indicate 50 μm. Cells were chosen from at least four randomly selected fields and scored for γ-H2A.X-positive nuclei (D). Results are expressed as means±S.E.M. (n=4). *P<0.01 compared with untreated group; #P<0.01 compared with UVB-exposed group; †P<0.01 compared with control siRNA UVB plus GW501516-treated group.

Ligand-activated PPARδ suppresses ROS generation induced by UVB radiation

The antioxidant NAC prevented UVB-induced premature senescence of keratinocytes. Since UVB has been known to induce ROS generation by NADPH oxidase [15], we examined the effects of GW501516 on ROS production in keratinocytes exposed to UVB. Although UVB radiation significantly increased ROS generation, pre-treatment with GW501516 for 24 h significantly suppressed the effect of UVB radiation (Figures 3A and 3B). The reduction in ROS generation in response to GW501516 was recovered in cells transfected with PPARδ siRNA, suggesting a PPARδ-dependent effect of GW501516 on ROS production (Figures 3A and 3B).

Figure 3 Ligand activation of PPARδ or inhibition of PI3K suppresses ROS generation and cellular senescence of human keratinocytes induced by UVB irradiation

(AC) ROS generation in human keratinocytes. Cells transfected with siRNA against PPARδ or control siRNA were pre-treated with GW501516 for 24 h and then exposed to UVB. After incubation for 30 min, the cells were treated with a peroxide-sensitive dye, CM-H2DCF-DA (10 μM), during the final 30 min of incubation. The intracellular ROS levels were detected (A) and quantified (B) by confocal laser fluorescence microscopy. (C) Cells pre-treated for 30 min with (+) or without (−) the indicated inhibitors were incubated in the presence or absence of GW501516 for 24 h, followed by exposure to UVB. After incubation for 30 min, the intracellular ROS levels were detected and quantified as above. (D) The percentage of SA β-gal-positive cells. Keratinocytes pre-treated with or without LY294002 for 30 min were incubated in the presence or absence of GW501516 for 24 h, exposed to UVB irradiation, and incubated for another 24 h. The results are representative of four seperate experiments. The results are expressed as the means±S.E.M. (n=4). *P<0.01 compared with untreated group; #P<0.01 compared with UVB-exposed group; †P<0.01 compared with UVB plus GW501516-treated group. Scale bars indicate 50 μm.

To characterize the signalling pathways involved in UVB-induced ROS production, we next examined the effects of kinase inhibitors on ROS production in cells exposed to UVB. As shown in Figure 3(C), UVB-induced ROS production was significantly reduced in the presence of LY294002, a PI3K (phosphatidylinositol 3-kinase) inhibitor, but not in the presence of GF109203X, a PKC (protein kinase C) inhibitor. The effect of LY294002 was further augmented in cells incubated with GW501516. Prior incubation with DPI, a non-specific inhibitor of NADPH oxidase, significantly attenuated UVB-induced ROS production in keratinocytes. In addition, LY294002 reduced the percentage of SA β-gal-positive cells following UVB irradiation to the same extent as did GW501516 (Figure 3D). Treatment with both GW501516 and LY294002, however, did not produce different results from treatment with either alone, indicating that the anti-senescence activity of PPARδ is mediated via the PI3K signalling pathway. These results suggest the involvement of ROS in UVB-induced premature senescence of keratinocytes.

Ligand-activated PPARδ attenuates the UVB-induced activation of Akt

We next analysed the involvement of the PI3K/Akt pathway in UVB-induced ROS generation and cellular senescence [28]. UVB exposure results in the time-dependent phosphorylation of Akt (Figure 4A). This UVB-induced activation of Akt was, however, markedly reduced in the presence of GW501516 and almost completely abolished in the presence of both GW501516 and LY294002 (Figure 4B). Down-regulation of PPARδ by siRNA reversed the GW501516-induced decrease in Akt phosphorylation in cells exposed to UVB (Figure 4C). These results suggest that PPARδ blocks the UVB-induced activation of Akt in human keratinocytes. As cross-talk between PPARδ and Akt has been demonstrated in several cell types (Supplementary Figure S3 at http://www.BiochemJ.org/bj/444/bj4440027add.htm; [24,29]), we further examined the effects of PPARδ activation by GW501516 at different pre-treatment times. Interestingly, a brief pre-treatment (30 min) did not affect the level of phosphorylated Akt, whereas a long pre-treatment (24 h) significantly suppressed the UVB-induced increase in phosphorylated Akt (Figure 4D). This result may suggest that ligand-bound PPARδ goes through the regulatory pathways involving transcription of its target genes to bring about its biological effects on downstream molecules, such as kinases and/or phosphatases, in human keratinocytes.

Figure 4 Ligand-activated PPARδ suppresses the activation of Akt induced by UVB irradiation in human keratinocytes

(A) Time-dependent activation of Akt in response to UVB irradiation. Cells were irradiated with UVB for the indicated periods of time. p-Akt, phosphorylated (activated) Akt. (B) Effects of GW501516 and LY294002 on UVB-induced activation of Akt. Cells pre-treated with (+) or without (−) LY294002 for 30 min were incubated in the presence or absence of GW501516 for 24 h. After exposure to UVB, the cells were incubated for an additional 3 h. (C) Effect of siRNA against PPARδ on UVB-induced activation of Akt. Cells transfected with or without PPARδ siRNA were treated with GW501516 for 24 h, exposed to UVB, and incubated for an additional 3 h. (D) Effects of GW501516 on UVB-induced activation of Akt at different pre-treatment times. Cells pre-treated with or without GW501516 for 30 min or 24 h were exposed to UVB and incubated for an additional 3 h. The whole cell lysates were subjected to immunoblot analysis and representative blots from four independent experiments are shown. Results are expressed as the means±S.E.M. (n=4). *P<0.01 compared with untreated group; ##P<0.05 compared with UVB-exposed group.

PTEN is up-regulated by ligand-activated PPARδ

We next examined the effects of actinomycin D and cycloheximide on UVB-induced activation of Akt. The level of phosphorylated Akt was markedly reduced in the presence of actinomycin D or cycloheximide (Figure 5A). Thus UVB-induced phosphorylation of Akt involves de novo synthesis of mRNA and protein(s) in keratinocytes.

Figure 5 PTEN is up-regulated by ligand-activated PPARδ in human keratinocytes

(A) The effects of cycloheximide (CHX) and actinomycin D (AcD) on UVB-induced activation of Akt. Cells pre-treated with CHX or AcD for 30 min were exposed to UVB radiation and then incubated for 3 h. (B) Time-dependent up-regulation of both PTEN mRNA and protein induced by GW501516. Cells were treated with GW501516 for the indicated periods of time and subjected to RNA blot or immunoblot analysis. (C) Effects of siRNA against PPARδ on PTEN mRNA and protein expression. Cells transfected with PPARδ or control siRNA were incubated with or without GW501516 for 24 h, and subjected to RNA blot or immunoblot analysis. (D and E) Effects of GW501516 (0, 0.1, 1, 10, 50 and 100 nM) on the activation of Akt and expression of PTEN in UVB-irradiated human keratinocytes. Cells pre-treated with increasing doses of GW501516 for 24 h were exposed to UVB irradiation and then incubated for an additional 3 h before immunoblot analysis. Representative blots from four independent experiments are shown (D). Band intensities were quantified using an image analyser, and fold changes are plotted (E).

To investigate the molecular mechanism underlying the PPARδ-mediated inhibition of Akt phosphorylation, we analysed the expression of PTEN, a dual protein/lipid phosphatase that dephosphorylates PtdIns(3,4,5)P3 into PtdIns(4,5)P2 and thus negatively regulates the PI3K/Akt pathway [30]. Treatment with GW501516 markedly increased the levels of both PTEN transcript and protein in a time-dependent manner (Figure 5B). The high levels of PTEN mRNA and protein observed 24 h after treatment with GW501516 are consistent with the observation that pre-treatment with GW501516 for 24 h, but not for 30 min, significantly suppressed the UVB-induced phosphorylation of Akt (Figure 4D). The up-regulation of PTEN by GW501516 was suppressed in the presence of siRNA against PPARδ, suggesting that PPARδ has a causal role in the up-regulation of PTEN (Figure 5C).

We next examined the effect of increasing doses of GW501516 on Akt phosphorylation and PTEN expression in UVB-irradiated human keratinocytes. GW501516 reduced the level of UVB-induced phosphorylation of Akt in a dose-dependent manner. In contrast, the GW501516-mediated increase in PTEN expression was correlated with a decline in phosphorylated Akt (Figures 5D and 5E). These data suggest that ligand-activated PPARδ regulates the activity of Akt at least in part through the up-regulation of PTEN.

UVB-induced membrane translocation and activation of Rac-1 are suppressed by ligand-activated PPARδ

Translocation of Rac-1 to the membrane is an essential step in the activation of NADPH oxidase during ROS production [31]. Thus we used confocal immunofluorescence microscopy to determine whether UVB irradiation could induce the translocation of Rac1 to the membrane in human keratinocytes, and then examined the effects of GW501516 and/or LY294002 on the UVB-induced translocation of Rac1. Although UVB irradiation caused the rapid translocation of Rac1 to the cell membrane, the addition of GW501516 and/or LY294002 almost completely abolished the UVB-mediated translocation of Rac1 (Figure 6A). Consistent with these results, immunoblot analysis using Rac1-specific antibody revealed that Rac1 accumulated in the membrane fraction following UVB irradiation and inhibition of UVB-induced membrane translocation by both GW501516 and LY294002 (Figure 6B). In addition, a significant proportion of Rac1 undergoes translocation within 5 min of UVB irradiation. Translocation is inhibited by GW501516 in a dose-dependent manner, as demonstrated by immunoblot analysis (Supplementary Figure S4 at http://www.BiochemJ.org/bj/444/bj4440027add.htm). These results suggest the involvement of PI3K and PPARδ in UVB-induced Rac1 signalling in human keratinocytes.

Figure 6 UVB-induced membrane translocation and activation of Rac1 are suppressed by ligand-activated PPARδ in human keratinocytes

(A and B) Effects of GW501516 and LY294002 on the translocation of Rac1. (A) Cells pre-treated with or without LY294002 for 30 min were incubated in the presence or absence of GW501516 for 24 h, and then exposed to UVB radiation. After 30 min, the cells were fixed and immunocytochemistry was performed using an anti-Rac1 antibody and Cy3-labelled anti-rabbit IgG. Images presented for each treatment were obtained from the same confocal microscopic section and are representative of three different sections (n=3). Scale bars indicate 20 μm. (B) Cells treated as above were fractionated into cytosolic and membrane fractions, and subjected to immunoblot analysis to determine the translocation of Rac1 using a Rac1-specific antibody. The band intensities were quantified and the ratios of Rac1 in the membrane to cytosolic fraction are plotted. (C) Effects of GW501516 and LY294002 on the activation of Rac1. Cells treated as above in (A) and (B) were harvested, and the cell lysates were processed for the effector domain pull-down assay. GTP-bound Rac1 was quantified using immunoprecipitation (IP) and immunoblotting (WB), and levels were normalized against total Rac1. Representative blots from three independent experiments are shown. The results are expressed as the means±S.E.M. (n=3). *P<0.01 compared with untreated group; #P<0.01 and ##P<0.05 compared with UVB-exposed group; †P<0.01 and ††P<0.05 compared with UVB plus GW501516-treated group.

To further confirm that Rac1 is activated upon translocation to the membrane, the activation state of endogenous Rac1 was detected as GTP-bound Rac1 and analysed by the effector domain pull-down assay. As shown in Figure 6(C), UVB irradiation activated Rac1, whereas treatment with either GW501516 or LY294002 suppressed UVB-induced activation of Rac1. The level of the active form of Rac1, Rac1–GTP, was lower in the presence of both GW501516 and LY294002 than in the presence of either compound alone. These results demonstrate that both PPARδ and PI3K are involved in UVB-induced Rac1 activation in human keratinocytes.

PTEN is required for the suppressive effects of PPARδ on ROS generation and the cellular senescence induced by UVB irradiation

To verify the functional significance of PTEN up-regulation by PPARδ, we examined the impact of siRNA-mediated PTEN knockdown on various points of the signalling pathways involved in UVB-induced ROS generation and the ensuing cellular senescence. Expression of PTEN in human keratinocytes was markedly reduced upon transfection with PTEN siRNA, but not with control siRNA, confirming the specificity and function of siRNA against PTEN (Supplementary Figure S5 at http://www.BiochemJ.org/bj/444/bj4440027add.htm). As PTEN is a PtdIns(3,4,5)P3 phosphoinositide 3-phosphatase, we first determined the impact of PTEN knockdown on cellular PtdIns(3,4,5)P3 levels. As shown in Figure 7(A), GW501516 suppressed the UVB-induced increase in PtdIns(3,4,5)P3; however, siRNA against PTEN abolished this effect, suggesting that ligand-activated PPARδ suppresses PtdIns(3,4,5)P3 production by inducing the expression of PTEN.

Figure 7 PTEN is required for the suppressive effects of GW501516 on ROS generation and cellular senescence induced by UVB irradiation in human keratinocytes

(A) Cellular PtdIns(3,4,5)P3 levels. (B) Levels of Akt phosphorylation and PTEN expression. (C) Translocation and (D) activation of Rac1. Representative blots from three independent experiments are shown. ROS generation in cells transfected with PTEN siRNA was detected (E) and quantified (F) by confocal laser fluorescence microscopy. (G) Effects of siRNA against PTEN on the number of SA β-gal-positive cells. Results are expressed as means±S.E.M. (n=4). Cells transfected with PTEN siRNA for 48 h were pre-treated with GW501516 for 24 h and exposed to UVB irradiation. The cells were then incubated for 3 h (A and B), 30 min (CF) or 24 h (G), as described in the Materials and methods section. *P<0.01 compared with untreated group; #P<0.01 compared with UVB-exposed group; †P<0.01 compared with UVB plus GW501516-treated group. IP, immunoprecipitation; WB, Western blot.

Knockdown of PTEN also reversed the PPARδ-mediated suppression of Akt phosphorylation/activation induced by UVB irradiation (Figure 7B). Furthermore, the increased level of PTEN expression correlated with a decrease in Akt phosphorylation in cells pre-treated with GW501516 and exposed to UVB irradiation (Figure 7B). Akt phosphorylation was induced by PTEN siRNA even in cells not exposed to UVB, suggesting that PTEN regulates the activation of Akt under normal physiological conditions (Figure 7B).

Next, we examined the effects of PTEN siRNA on the PPARδ-mediated suppression of both translocation and activation of Rac1 induced by UVB irradiation. PPARδ-mediated suppression of Rac1 translocation to the membrane fraction was recovered in UVB-treated keratinocytes by transfection with PTEN siRNA (Figure 7C and Supplementary Figure S6 at http://www.BiochemJ.org/bj/444/bj4440027add.htm). Interestingly, blocking PTEN expression alone did not induce Rac1 translocation when cells were not irradiated with UVB (Figure 7C and Supplementary Figure S6). This may suggest that signalling pathways other than PtdIns(3,4,5)P3/Akt activation are also required for the UVB-induced translocation of Rac1. A similar reversal of the effects of PTEN siRNA was also observed for the PPARδ-mediated suppression of Rac1 activation, as confirmed by the effector domain pull-down assay (Figure 7D).

Furthermore, we also examined the effects of PTEN siRNA on the PPARδ-mediated suppression of ROS generation and the subsequent cellular senescence induced by UVB irradiation. As expected from the above results, knockdown of PTEN completely abolished the suppressive effects of GW501516 on both ROS generation (Figures 7E and 7F) and cellular senescence (Figure 7G) in human keratinocytes following UVB irradiation.

Together, these data demonstrate that ligand-activated PPARδ regulates UVB-induced ROS generation and cellular senescence by increasing the expression of PTEN, which in turn suppresses the UVB-induced PI3K/Akt/Rac1 signalling pathways.

Ligand-activated PPARδ prevents UVB-induced cellular senescence in HR-1 hairless mice in vivo

To verify our findings in cultured cells, we further examined the in vivo cellular senescence of skin in HR-1 hairless mice exposed to UVB. Consistent with results from the in vitro studies, UVB irradiation induced both cellular senescence and the generation of intracellular superoxide (Figures 8A and 8B) in the dorsal skin of hairless mice, as determined by SA β-gal activity and DHE red fluorescence, respectively. In contrast, administration of GW501516 significantly attenuated UVB-induced cellular senescence and superoxide production. Similar results were also observed in the expression levels of p53, p21, p16 and PML (Figures 8A and 8B). These results suggest that UVB irradiation plays a physiologically relevant role in premature skin senescence and that PPARδ is an endogenous anti-senescence molecule that can be activated and therapeutically targeted by its agonistic ligand in the irradiated cells.

Figure 8 Ligand-activated PPARδ prevents UVB-induced cellular senescence in HR-1 hairless mice

(A) Representative cross-sections of the dorsal skin of UVB-exposed HR-1 hairless mice treated with or without GW501516 for 24 h. Frozen tissue sections were subjected to staining for SA β-gal, p53, p21, p16, PML, DHE or H&E as described in the Materials and methods section. Scale bars indicate 50 μm. Representative images from four independent experiments are shown. (B) Quantitative analysis of SA β-gal-positive cells, DHE fluorescence and expression levels of p53, p21, p16 and PML in the tissue sections used in (A). (C and D) Activation of Akt and expression of PTEN, p53, p21, p16 and PML in dorsal skin tissues. Tissue extracts of dorsal skin from mice treated and irradiated as above were subjected to immunoblot analysis. Representative blots from tissue extracts of four different animals are shown in (C). Relative band intensities were quantified using an image analyser. Band intensities are represented as fold change relative to the untreated control for various treatments (D). *P<0.01 compared with untreated group; #P<0.01 compared with UVB-exposed group.

The molecular signalling pathways that modulate UVB-induced cellular senescence via PPARδ identified in in vitro studies were confirmed in a physiological environment using an in vivo mouse model. When hairless mice were irradiated with UVB, the expression of PTEN was down-regulated, with a concomitant increase in Akt phosphorylation and expression of p53, p21, p16 and PML (Figures 8C and 8D). However, administration of GW501516 caused a marked increase in PTEN expression, with a corresponding decrease in Akt phosphorylation and expression of p53, p21, p16 and PML. These results suggest that ligand-activated PPARδ up-regulates PTEN expression and down-regulates PI3K/Akt signalling pathways to suppress UVB-induced ROS generation and cellular senescence in the skin.

DISCUSSION

Repeated exposure to UV radiation from sunlight can result in sunburn, premature senescence and skin cancer [32]. In particular, long-term UV exposure causes DNA damage and oxidative stress in irradiated keratinocytes [33], resulting in premature senescence and skin cancer. Cellular senescence is defined as the irreversible arrest of cellular replication [4], and can be characterized by the expression of cyclin-dependent kinase inhibitors and SA β-gal [34,35]. Anti-proliferative cellular senescence was considered to be a normal defence mechanism not only to protect keratinocytes from malignant transformation by suppressing the replication of mutated DNA, but also to maintain the integrity of the protective barrier of the epidermis by forming premature senescence before proceeding into apoptosis in skin exposed to stress-induced injury such as UVB irradiation [4].

The results in the present study strongly support the hypothesis that PPARδ prevents cellular senescence of keratinocytes and is thus a key target for therapeutic intervention in skin aging. A previous report demonstrated that pioglitazone-activated PPARγ, another member of the PPAR family, suppressed the angiotensin II-induced senescence of endothelial progenitor cells by down-regulating the angiotensin II type 1 receptor [36]. In contrast, another investigation showed that PPARγ activation accelerated replicative senescence by inducing p16, a cell-cycle inhibitor that prevents entry into the cell cycle [37]. Although the role of PPARγ in senescence is controversial, our results demonstrate that ligand-activated PPARδ prevents the cellular senescence induced by UVB irradiation. Since senescent cells have been identified among human keratinocytes and fibroblasts exposed to UVB [4,23], it may be possible to reduce aging of the skin by suppressing cellular senescence of keratinocytes through PPARδ activation.

ROS is known to be involved in UVB-induced cellular senescence [4,23,38]. Among the enzymes implicated in ROS formation in skin cells, NADPH oxidase was reported to be the main source of ROS generated in UV-induced keratinocytes [14,15]. The increase in ROS generation following UVB irradiation was significantly suppressed by LY294002, an inhibitor of PI3K, but not by GF109203X, an inhibitor of PKC. Whereas activation of the PI3K/Akt pathway leads to translocation of Rac and prolonged oxidase-dependent ROS generation, PKC is involved in the initial activation of NADPH oxidase [39]. Accordingly, the PI3K/Akt pathway seems to play a major role in UVB-induced ROS generation under the present experimental conditions. The present study shows that ligand-activated PPARδ significantly inhibited ROS production. Although we did not directly determine the effects of ligand-activated PPARδ on PI3K activity, the inhibition of Akt, a downstream target of PI3K, suggests that the biological effects of ligand-activated PPARδ may be achieved, at least in part, through modulation of the PI3K/Akt signalling pathway to inhibit UVB-induced ROS production. DNA damage by ROS was also shown to participate in UVB-induced cellular senescence [38]. In fact, ligand-activated PPARδ attenuated the UVB-induced frequency of γ-H2A.X foci, a marker for DNA strand breakage, in the keratinocytes. These findings therefore suggest that PPARδ suppresses UVB-induced cellular senescence by mechanisms relevant to DNA damage.

Induction of PTEN expression by PPARδ is a key event in the blockade of UVB-induced cellular senescence by GW501516. The tumour suppressor PTEN, a negative regulator of the PI3K/Akt signalling cascade, modulates a variety of cellular processes, including cell growth, survival, proliferation and migration [40]. Several studies suggest that the transcriptional regulation of PTEN may be complex and involve differential modulation in response to diverse biological stimulations. The involvement of transcription factors, such as p53 and Egr-1, has been demonstrated in the regulation of PTEN expression [41,42]. However, the transcriptional regulation of PTEN is poorly understood. PPARδ was originally shown to inhibit the expression of PTEN in keratinocytes during wound healing [24]. However, this change in PTEN expression is not consistently observed in response to ligand activation of PPARδ in human keratinocytes [43], suggesting that the up-regulation of PTEN expression by PPARδ is context specific [19]. On the other hand, transcription of PTEN was reported to be up-regulated by rosiglitazone, a specific ligand of PPARγ occurring in several types of human cells [44,45]. Induction of PTEN by rosiglitazone occurs at the transcriptional level via binding of PPARγ to the PPRE located in the 5′-flanking region of the PTEN gene [44]. The effect of rosiglitazone on PTEN is suggested to impact on various functions of PPARγ, such as tumour suppression, anti-inflammatory actions and cell migration, by regulating the PI3K/Akt pathway. Since the present study has shown that PPARδ regulates the expression of PTEN at the transcriptional level, the PPRE in the PTEN gene may also mediate the up-regulation of PTEN by PPARδ.

Consistent with the findings obtained in cultured human keratinocytes, the administration of GW501516 significantly attenuated UVB-induced cellular senescence and superoxide production in vivo. The molecular signalling pathways modulated by PPARδ were also confirmed in the in vivo physiological environment. That is, administration of GW501516 increased PTEN expression with a corresponding decrease in Akt phosphorylation. Therefore, these results led us to hypothesize that ligand-activated PPARδ up-regulates PTEN expression and down-regulates PI3K/Akt activation signalling pathways to suppress UVB-induced ROS generation and cellular senescence of the skin. The results of several previous studies support this hypothesis. Various humoral factors involved in skin aging have been shown to enhance Akt activity [46]. Activation of Akt was observed in aged human skin lesions, but not in normal skin [47]. Since inhibition of Akt extended the life span of primary cultured human endothelial cells [48], sustained activation of Akt may promote cellular senescence in aged skin. In this context, PPARδ may serve as an anti-senescent mediator in age-related skin changes, such as wrinkle formation [49].

In summary, we report that ligand-activated PPARδ regulates UVB-induced cellular senescence in keratinocytes by inducing PTEN expression and thereby modulating the ability of PI3K/Akt/Rac-1 signalling pathways to reduce cellular ROS generation. Our results support the hypothesis that PPARδ prevents cellular senescence of keratinocytes and is thus a key target for therapeutic intervention in skin aging. Thus the present study not only enhances our understanding of the molecular mechanisms that underlie the anti-senescent effect of PPARδ, but also suggests novel treatments for age-related skin disorders.

AUTHOR CONTRIBUTION

Sun Ah Ham and Han Geuk Seo conceived and designed the project. Sun Ah Ham, Jung Seok Hwang, Taesik Yoo and Hanna Lee performed the experiments. Sun Ah Ham and Eun Sil Kang analysed the results. Chankyu Park, Jae-Wook Oh, Hoon Taek Lee and Gyesik Min contributed reagents, materials and analysis tools. Jin-Hoi Kim and Han Geuk Seo contributed to the discussion and interpretation of results and drafting the paper.

FUNDING

This work was supported in part by a Mid-career Research Program through an NRF grant funded by the MEST [grant number 2011-0012427], and Next-Generation BioGreen 21 Program [grant number PJ007980], Rural Development Administration, Republic of Korea.

Abbreviations: CM-H2DCF-DA, chloromethyl-2′,7′-dichlorofluorescein diacetate; Cy3, indocarbocyanine; DHE, dihydroethidium; DPI, diphenyleneiodonium; H&E, haematoxylin and eosin; NAC, N-acetylcysteine; PAK, p21-activated kinase 1; PI3K, phosphatidylinositol 3-kinase; PKC, protein kinase C; PML, promyelocytic leukaemia protein; PPAR, peroxisome-proliferator-activated receptor; PPRE, PPAR-response element; PTEN, phosphatase and tensin homologue deleted on chromosome 10; ROS, reactive oxygen species; SA, β-gal, senescence-associated β-galactosidase; SIPS, stress-induced premature senescence; siRNA, small interfering RNA; TBS, Tris-buffered saline; TCA, trichloroacetic acid; TGF-β1, transforming growth factor-β1

References

View Abstract